Citation
Transformation of Valley-Bottom Cultivation and Its Effects on Tanzanian Wetlands: A Case Study of Ndembera Wetland Area in Iringa Region

Material Information

Title:
Transformation of Valley-Bottom Cultivation and Its Effects on Tanzanian Wetlands: A Case Study of Ndembera Wetland Area in Iringa Region
Creator:
MAGEMBE, LUCY ( Author, Primary )
Copyright Date:
2008

Subjects

Subjects / Keywords:
Agricultural land ( jstor )
Agriculture ( jstor )
Cash crops ( jstor )
Crops ( jstor )
Farms ( jstor )
Highlands ( jstor )
Land use ( jstor )
Tillage ( jstor )
Villages ( jstor )
Wetlands ( jstor )

Record Information

Source Institution:
University of Florida
Holding Location:
University of Florida
Rights Management:
Copyright Lucy Magembe. Permission granted to University of Florida to digitize and display this item for non-profit research and educational purposes. Any reuse of this item in excess of fair use or other copyright exemptions requires permission of the copyright holder.
Embargo Date:
7/12/2007
Resource Identifier:
659860311 ( OCLC )

Downloads

This item has the following downloads:

magembe_l ( .pdf )

magembe_l_Page_117.txt

magembe_l_Page_018.txt

magembe_l_Page_082.txt

magembe_l_Page_042.txt

magembe_l_Page_108.txt

magembe_l_Page_084.txt

magembe_l_Page_005.txt

magembe_l_Page_111.txt

magembe_l_Page_006.txt

magembe_l_Page_116.txt

magembe_l_Page_024.txt

magembe_l_Page_002.txt

magembe_l_Page_075.txt

magembe_l_Page_069.txt

magembe_l_Page_015.txt

magembe_l_Page_097.txt

magembe_l_Page_031.txt

magembe_l_Page_001.txt

magembe_l_Page_090.txt

magembe_l_Page_035.txt

magembe_l_Page_008.txt

magembe_l_Page_047.txt

magembe_l_Page_085.txt

magembe_l_Page_012.txt

magembe_l_Page_027.txt

magembe_l_Page_068.txt

magembe_l_Page_037.txt

magembe_l_Page_007.txt

magembe_l_Page_100.txt

magembe_l_Page_011.txt

magembe_l_Page_093.txt

magembe_l_Page_089.txt

magembe_l_Page_113.txt

magembe_l_Page_070.txt

magembe_l_Page_112.txt

magembe_l_Page_029.txt

magembe_l_Page_036.txt

magembe_l_Page_019.txt

magembe_l_Page_061.txt

magembe_l_Page_057.txt

magembe_l_Page_055.txt

magembe_l_Page_092.txt

magembe_l_Page_110.txt

magembe_l_Page_104.txt

magembe_l_Page_107.txt

magembe_l_Page_052.txt

magembe_l_Page_114.txt

magembe_l_Page_062.txt

magembe_l_Page_044.txt

magembe_l_Page_004.txt

magembe_l_Page_121.txt

magembe_l_Page_083.txt

magembe_l_Page_022.txt

magembe_l_Page_009.txt

magembe_l_Page_065.txt

magembe_l_Page_026.txt

magembe_l_Page_017.txt

magembe_l_Page_014.txt

magembe_l_Page_039.txt

magembe_l_Page_079.txt

magembe_l_Page_071.txt

magembe_l_pdf.txt

magembe_l_Page_032.txt

magembe_l_Page_053.txt

magembe_l_Page_086.txt

magembe_l_Page_058.txt

magembe_l_Page_025.txt

magembe_l_Page_076.txt

magembe_l_Page_105.txt

magembe_l_Page_059.txt

magembe_l_Page_101.txt

magembe_l_Page_096.txt

magembe_l_Page_072.txt

magembe_l_Page_080.txt

magembe_l_Page_033.txt

magembe_l_Page_041.txt

magembe_l_Page_067.txt

magembe_l_Page_098.txt

magembe_l_Page_095.txt

magembe_l_Page_063.txt

magembe_l_Page_087.txt

magembe_l_Page_040.txt

magembe_l_Page_045.txt

magembe_l_Page_028.txt

magembe_l_Page_073.txt

magembe_l_Page_120.txt

magembe_l_Page_048.txt

magembe_l_Page_054.txt

magembe_l_Page_094.txt

magembe_l_Page_109.txt

magembe_l_Page_081.txt

magembe_l_Page_013.txt

magembe_l_Page_077.txt

magembe_l_Page_115.txt

magembe_l_Page_016.txt

magembe_l_Page_119.txt

magembe_l_Page_030.txt

magembe_l_Page_106.txt

magembe_l_Page_043.txt

magembe_l_Page_074.txt

magembe_l_Page_099.txt

magembe_l_Page_066.txt

magembe_l_Page_046.txt

magembe_l_Page_050.txt

magembe_l_Page_056.txt

magembe_l_Page_010.txt

magembe_l_Page_103.txt

magembe_l_Page_088.txt

magembe_l_Page_034.txt

magembe_l_Page_091.txt

magembe_l_Page_064.txt

magembe_l_Page_003.txt

magembe_l_Page_038.txt

magembe_l_Page_020.txt

magembe_l_Page_102.txt

magembe_l_Page_023.txt

magembe_l_Page_078.txt

magembe_l_Page_060.txt

magembe_l_Page_051.txt

magembe_l_Page_049.txt

magembe_l_Page_021.txt

magembe_l_Page_118.txt


Full Text





TRANSFORMATION OF VALLEY-BOTTOM CULTIVATION AND ITS EFFECTS ON
TANZANIAN WETLANDS: A CASE STUDY OF NDEMBERA WETLAND AREA INT
IRINGA REGION


















By

LUCY MAGEMBE


A THESIS PRESENTED TO THE GRADUATE SCHOOL
OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT
OF THE REQUIREMENTS FOR THE DEGREE OF
MASTER OF SCIENCE

UNIVERSITY OF FLORIDA


2007




























O 2007 Lucy Magembe




























To my son, George, and daughter, Iman, who fill my life with joy and whose presence
encourages me to fight my battles earnestly so as to be able to provide for them.









ACKNOWLEDGMENTS

Many people provided invaluable support. Since it is impossible to acknowledge each

person individually, I extend my sincere appreciation to all who played a part in this work.

However, I feel obligated to thank individually a few people who have given generously of their

time, knowledge, expertise, and moral support.

First and foremost, I extend my deep appreciation to my supervisory committee chair (Dr.

Michael Binford) for his invaluable assistance, advice, and encouragement. He practically

challenged me to work independently but continued to stimulate my analytical thinking, enabling

me to improve my writing skills.

I am indebted to my committee members Dr. Abe Goldman and Dr. Sandra Russo for

helping me grow intellectually and develop a more focused thesis. Their input to this work and

their attention are highly appreciated.

I extend my sincere gratitude to my sponsors, The United States Agency for International

Development (USAID) and the Africa-America Institute (AAI), who made my stay in the United

States content and without whose financial support this work would not have come to fruition.

In Tanzania, I would like to thank the people of Iringa and Mufindi Districts for their time,

support, and kindness. Sincerest appreciation is extended to the District Facilitation Teams and

to WWF (World Wide Fund for Nature) staff for their endless support during fieldwork.

I thank members of staff and students of the Geography Department at the University of

Florida for their moral support and for providing a social and academic atmosphere in which I

was able to grow.

My gratitude goes to my husband, George Magembe, my sister Annamarie Kiaga, and my

late mom, Mrs. Tabitha Kashaij a for their continued love, support, encouragement, and for

believing in me.










No words can express how grateful I am to the Mkanta family for allowing me to stay with

them for endless weeks while completing my thesis. A heartfelt gratitude goes especially to Dr.

William Mkanta for his assistance in statistical methods.

Anna Mushi, Tunu Mndeme, Simon Mwansasu and Juma Raj abu Mshana are

acknowledged for their assistance in perfecting the aerial photo interpretation and GIS analysis.

Lastly, I give thanks and praises to the Almighty God for the blessings in my life, but

mostly for enabling me endure the stresses associated with graduate schooling.











TABLE OF CONTENTS
page

ACKNOWLEDGMENT S .............. ...............4.....



LI ST OF T ABLE S .........__.. ..... .__. ...............9....



LIST OF FIGURES .............. ...............10....



AB S TRAC T ........._. ............ ..............._ 1 1...



CHAPTER



1 INTRODUCTION ................. ...............13.......... ......


1.1 Background and Statement of the Problem .............. ...............14....
1.2 Significance .............. ...............16....
1.3 Objectives .............. ...............17....
1.4 Study Questions ................. ...............17.......... .....

2 LITERATURE REVIEW ................. ...............21................

2. 1 Introducti on .................. ...............21........... ...
2.2 Conceptual Framework......................... ...............2
2.2. 1 Concepts of Land-Use and Land-Cover Change ................. ................ ....._.21
2.2.2 Human-settlement Models of Land-Use ................. ............. ......... .......24
2.2.3 Population Growth and Agricultural Land-Use .............. ...............26....
2.2.4 Drivers of Smallholders Farming Decisions ............... ... ....... ..............2
2.2.5 Decline in Agricultural Production and its Relation to Land Use ................... .......32
2.2.6 Wetlands: An Alternative Landscape Component for Crop Production ................34
2.2.7 Irrigation Practices and Wetland Use in Tanzania .............. .... ..................3
2.3 Summary of the Theoretical Background and Formulation of Hypotheses ...................40
2.3.1 Summary of the Theoretical Background............... ...............4
2.3.2 Justification of the Hypotheses .......__ ......... .........___.....__......43

3 STUDY OF FARMER AND HOUSEHOLD FACTORS INFLUENCING V7NYUNG;U
FARMING SY STEM ........._.._ ..... ._._ ...............45....


3.1 Study Area .............. ...............45....
3.1.1 Location ................. ......_ __ ...............45. ...
3.1.2 Physical Environment ....._.. __............. ... ......_._..........._. ....._..__.45
3.1.3 Demographic Characteristics............... ............4
3.1.4 Socioeconomic Situation ..........._.._. ......... ...............46....
3.1.5 Agriculture ................. ...............47..............












3.2 M material and M ethods .............. ... ......... ... ...............48.....
3.2. 1 Sample Profile and Sampling Procedure ........ ................. ............... ..4
3.2.2 Data Collection............... ...............5
3.2.3 Methods of Data Analysis ................. ...............54..............
3.2.3.1 Descriptive Statistics ......... ................ ...............55. ...
3.2.3.2 Measures of Associations ................. ...............55 .............
3.2.3.3 M ultivariable Analysis .............. .......... .. ...... ...... .......5
3.3 Results: Farmer and Household Characteristics of Valley-Bottom Cultivation
(vinyungu) around the Ndembera Swamp in Iringa and Mufindi Districts ................... ......56
3.3.1 Sociodemographic Characteristics .............. ...............56....
3.3.1.1 Land tenure/ownership ............... ... .......... ...............58 ..
3.3.1.2 Crop production, crop preference and land use............... ...................5
3.3.1.3 Inputs on vinyungu .............. ...............60....
3.3.1.4 Multivariable analysis .............. ......... .... ...........6
3.3.2 Environmental Issues Related to Vinyungu Cultivation ................. ............... ...64
3.3.3 Discussion. ..............._ ...............65......... ...


4 USE OF GEOGRAPHIC INFORMATION SYSTEMS AND REMOTE SENSING TO
DETERMINE THE MAGNITUDE AND RATE OF CONVERSION OF NDEMBERA
SWAMP AS A RESULT OF VINYUNG;UEXPANSION .....__.___ ........___ ..............79


4.1 D ata Sets .............. ...............80....
4.2 M ethods .............. ...............80....
4.3 Data Analysis............... ...............81
4.4 Results............... ...............8 1
4.5 Discussion............... ...............8


5 DISCUSSION AND CONCLUSION .............. ...............91....


5 .1 Introducti on ................. ...............91........... ...
5.2 Maj or Findings ........._.___..... .___ ...............91...
5.3 Conclusion ........._.___..... .__ ...............98....


APPENDIX



A SAMPLE SIZE DETERMINATION ........._._._ ...._. ...............102..



B SURVEY INSTRUMENT: INDIVIDUAL FARMER QUESTIONNAIRE.............._._. ....103

B.1 Location .............. ... ...............103..
B.2 Background Information ............_... .. ....__ .. ......__ ... ..........10
B.3 Land availability, crop preference, seasonality, and fallow period ............... ...............103
B.4 Labor, Input, and Tools on Vinyungu .............. ...............105....












C SURVEY INSTRUMENT : GROUP QUESTIONNAIRE ................. ................ ...._.107


C. 1 Background .............. ... ....._ ... .. ........ .............10
C.2 Land availability, cropping preference, and fallow period ................. .....................107
C.3 Input, Tools and Labor ................. ...............107..............
C.4 Wetlands vs. Uplands............... ...............108

LIST OF REFERENCES ................. ...............111........_......



BIOGRAPHICAL SKETCH ..............._ ...............120........_ ......










LIST OF TABLES


Table page

3-1 Sociodemographic and Land Use Characteristics of the Sample Farmers in the study
area (n=54) ................. ...............69......... ......

3-2 Maize yields in uplands and vinyungu between 1970s and 2000s (village responses,
n=9) (per year, i.e., one season in the uplands and two seasons in the vinyungu) .............71

3-3 Potential market price per unit output from cash crops grown in vinyunguand under
rain fed agriculture ................. ...............71.........._ ....

3-4 Results of Chi-Square Analysis of the Farmer and Household Factors influencing
Vinyungu-Farming Practice, i.e., size of vinyungu plots owned by household in
relation to sex, age, years of residence in the village, and number of potential
w workers ................ ...............72......... ......

3-5 Results of Chi-Square Analysis of the Farmer and Household Factors influencing
Vinyungu-Farming Practice, i.e., size of vinyungu plots grown in the dry season in
relation to sex, age, years of residence in the village, and number of potential
w workers ................ ...............72......... ......

3-6 Results of Chi-Square Analysis of the Farmer and Household Factors influencing
Vinyungu-Farming Practice, i.e., number of cash crops grown on vinyungu plots in
relation to sex, age, years of residence in the village, and number of potential
w workers ................ ...............73......... ......

3-7 Results of Chi-Square Analysis of the Farmer and Household Factors influencing
Vinyungu-Farming Practice, i.e., size of vinyungu plots grown cash crops in relation
to sex, age, years of residence in the village, and number of potential workers ...............73

3-8 Results of the Linear Regression Analysis of the Factors Affecting Vinyungu-
F arming Practice .............. ...............74....

3-9 Farmer' s perception on various environmental issues as discussed in the village
meetings: the current state compared to the 1970s (n = 40) ...........__... .......__.........75

3-10 Individual farmer' s perception on various environmental issues (n=54) ................... .......78

4-1 Map History (i.e., Data used) ...........__......___ ...............86..

4-2 Total area (ha) and area of change of land use types from 1977 to 1999 ................... .......90

4-3 Transition matrix of land use types from 1977 to 1999 (transition probabilities in %).....90










LIST OF FIGURES


Figure page

1-1 Map of Iringa region showing the study area (A). (Source: based on 1:2,000,0000
Tanzania Administrative Map of 1989, by Surveys and Mapping Division. Ministry
of Lands and Human Settlement). The inset (B) shows the location Iringa region in
Tanzania. ........._.. .. ....... ...............18.....

1-2 Typical vinyungu grown potatoes (A) and peas (B). Both pictures taken in
Usengelindete village, Iringa District of Iringa Region, June 2005. ................ ...............19

1-3 Enlarged map (A) shows Lyandembera swamp and surrounding villages. Subset
image (B) shows the location of Lyandembera swamp (The Study Area) in Iringa
region. ............. ...............20.....

3-1 A well is dug where water is not free flowing. Picture taken in Lumuli village in
Iringa District, June 2005 ................. ...............76................

3-2 A river is diverted to supply vinyungu with water. Picture taken in Usengelindete
village, Iringa District, June 2005 ................. ...............76........... ...

3-3 Picture showing Ndembera river behind which vinyungu cultivation is taking place in
Maduma village, Mufindi District. An example of how close to the river the
vinyungu are getting. Picture taken in June 2005 .............. ...............77....

4-1 Ndembera swamp in 1977 (A). The inset map (B) shows Ndembera swamp location
in Iringa region ................. ...............87........... ....

4-2 Ndembera swamp in 1999 (A). The inset map (B) shows tha location of
Lyandembera swamp in Iringa region .............. ...............88....

4-3 Land use/ cover change map (1977 and 1999) (A). The inset map (B) shows the
location of Lyandembera swamp in Iringa ................. ...............89...............









Abstract of Thesis Presented to the Graduate School
of the University of Florida in Partial Fulfillment of the
Requirements for the Degree of Master of Science

TRANSFORMATION OF VALLEY-BOTTOM CULTIVATION AND ITS EFFECTS ON
TANZANIAN WETLANDS: A CASE STUDY OF NDEMBERA WETLAND AREA IN
IRINGA REGION

By

Lucy Magembe

May 2007

Chair: Michael William Binford
Major: Geography

This study, conducted in southwestern Tanzania, examines the social and environmental

implications of conversions of wetlands to alternative land-uses, in particular agriculture. Using

the case of the vinyungu-faming system, a wetland form of agriculture practiced in Iringa region,

the study demonstrates complex ways in which ongoing pressure over food production is causing

transformations in the traditional cultivation methods. To do so, four questions were asked,

Research questions 1: How is vinyungu-farming practice related to household and
farmer characteristics?
Research questions 2: How has the vinyungu-farming system evolved over time?
Research questions 3: What have been the possible driving forces behind the changes in
vinyungu-farming practice?
Research questions 4: What is the effect of vinyungu-farming system?

To answer these questions, the study uses pre-designed questionnaires to examine farmers'

perceptions on how the system has changed over time and how this change may affect the

wetland resources. The study also uses Geographic Information Systems (GIS) to determine

whether or not there is a change in the wetland area used for agriculture in the study area.

Results from this study indicate a shift in agriculture whereby vinyungu-farming system

that supplemented household food obtained from upland fields has evolved to an economic

activity, producing cash crops for local, regional and national markets. In addition, this









traditional irrigation system is transforming from being predominantly a woman's activity to an

entire family's activity. An increase in the size of vinyungu cultivated, use of inorganic fertilizer

and encroachment of wetter parts of the wetland are some of the ways by which farmers have

modified their traditional irrigation system. Crop production is no longer confined to the dry

season but extends into the wet season, especially in areas that are not heavily inundated in

water. Results from the GIS analysis indicate an increased conversion of wetland area to

agriculture between 1977 and 1999. GIS analysis during the same period also indicate an

increased conversion of the wetland to grassland, a factor that farmers attributed to reduction of

streamflow into the wetland as a result of increased human activities.

Several factors have been found to influence the transformation of wetlands to vinyungu

fields and expansion of these fields. These include population growth, increased market

demands, government policies, and reduced production power in the uplands as a result of

rainfall unreliability, increased droughts and reduced soil fertility. All of these factors have

increased the demand for more arable land, in this case, the wetlands. In addition, most

households must have an adequate labor force to cultivate the wet and heavy soils that are

characteristic of wetlands. Improved communications networks and higher profitability in

vinyungu compared to upland plots also encourage modifications in this farming system.

Results of the study indicate that wetlands provide both social and economic benefits to the

people that live around or near them. On the other hand, these very resources that sustain

livelihoods and perform important functions such as flood control and water quality

improvement are increasingly being lost. It is important therefore for any future land-use regimes

in and around wetlands to consider human needs and wetlands sustainability simultaneously.









CHAPTER 1
INTTRODUCTION

Wetlands are important landscape components and critical natural ecosystems that provide

significant environmental and socio-economic benefits to humans (Roggeri, 1998; Silvius et al.,

2000; Stuip et al., 2002). Because of their characteristics (e.g., heavy soils, thick vegetation -

sometimes on extensive flat lands), wetlands store water (and nutrients) during wet periods and

maintain base flow during dry periods (Balek and Perry, 1973; Dugan, 1990; Roggeri, 1995).

Thus, wetlands can control floods, improve water quality, and provide unique habitat for

wildlife, including many rare and endangered species. These ecosystems provide numerous

benefits to humans including food, pharmaceuticals, and construction materials (Daily, 1997).

Rich soils and high moisture holding capacity make wetlands particularly attractive for

agriculture. However, wetland soils can be waterlogged and anaerobic, therefore inimical for

plants unless they are adapted to soggy soils or some means are used to raise the rooting zone

above the saturated layer.

For millennia, humans have cultivated wetlands to meet food security and livelihood needs

without necessarily affecting wetland structure and functions (Adams, 1993b; Banzi et al., 1992;

Erickson 1985; Roggeri, 1995; Scoones, 1991). In recent years, wetlands have come under

extreme pressure as many have been converted to agriculture, which raises concern over the

sustainability of wetland cultivation, especially in tropical developing countries where wetlands

are among the least protected ecosystems (Dixon and Wood, 2003; Gerakis and Kalburtji, 1998;

Hollis, 1990; Jensen et al., 1993; Jensen et al., 1995; Liu et al., 2004; Munyati, 2000; Ringrose et

al., 1988; Roggeri, 1995; Wang et al., 2006; Williams, 1991). This study is an effort to explore

changes in wetland farming systems and factors that are driving these changes. I used the

experience of an indigenous wetland farming system in Iringa region, Tanzania, to investigate










why and how the practices in this farming system have been transformed over time and

determine the social and environmental implications of that transformation.

1.1 Background and Statement of the Problem

Iringa region, southwest of Tanzania (Figure 1-1), is surrounded by semi-arid conditions.

Therefore, wetlands are the most ideal resource for agriculture. Farmers in wetland areas have

devised a traditional farming system called vinyungu that allows year-round cultivation of crops

(Majule and Mwalyosi, 2003). Vinyungu (or kinyungu in singular form) are ridges or raised beds

that are about 0.6 m high and 4-20 m wide with a cambered surface sloping down to the open

drain on either side. They are created by first clearing the land, then burning off the cleared

vegetation, followed by hand hoe plowing that is done simultaneously with the construction of

ditches and ridges, and lastly, harrowing, to smoothen the ridges. This farming system is highly

prevalent in Makete, Ludewa, Mufindi, Iringa and Nj ombe districts of lringa region and it is

practiced mainly by smallholder farmers of Bena and Hehe tribes. It is mostly a dry season

(June-October) agricultural activity that takes place almost exclusively in wet valley bottoms that

are characterized by heavy clayey soils. Green maize, beans, potatoes, and vegetables are the

most commonly grown crops. Figure 1-2 shows typical vinyungu grown crops.

Vinyungu farming, believed to have started as far back as the 1890s (Culwick, 1935), was

practiced mainly by women on small fields, with little or no economic contribution to the

livelihoods of the people that practiced it (Kuroda, 2001; Lema, 1996). Recently, however, a

combination of socio-economic factors has caused the extensification of the farming system, in

terms of size of vinyungu and the number of people involved in this farming practice. Among

these factors may be the increased market demand for food and other wetland products like

vegetables for urban centers in Iringa region as well as the distant cities of Dar es Salaam and

Mbeya. For example, Olindo (1992) observed the increased market demand for food and other









wetland products in some parts of Kenya as one of the maj or factors driving further conversion

of wetlands to agriculture. Similar to other cases in Eastern Africa (Denny and Turyastunga,

1992; Gichuki, 1992; Loevinsohn et al., 1992; Wood, 1996), the ability of farmers to harvest

crops twice or even thrice a year hence increasing food security especially in drought-prone areas

may also play an important role in vinyungu expansion in Iringa region. Burgeoning human

population, socio-economic changes, and government policies that call for use of more

agriculturally productive land to reduce food deficits, are considered some of the maj or factors

causing wetland conversions (Roggeri, 1995).

Generally, regardless of the motivation, unrestricted drainage and cultivation of wetlands

can have far-reaching and sometimes irreversible consequences. Unrestricted drainage and

cultivation of wetlands can affect wetland' s hydrological functions, causing a reduction in water

storage and quality, variable stream flows, and sometimes, complete dryness of wetlands (Denny

and Turyatunga, 1992; Dixon and Wood, 2003; Roggeri, 1995; MWLD, 2001). In addition,

riverbanks are eroded and sediments accumulate downstream, affecting the overall structure and

functions of wetlands (Kaswamila and Tenge, 1997; Olindo, 1992). These factors have serious

effects on the livelihoods of the local communities, especially those residing further downstream

- people walk long distances in search of water; land becomes uncultivable; livestock keepers are

forced to walk long distances in search of water and pasture; communal conflicts may arise due

to animal trampling of crops; and, fish and wildlife resources decline (MWLD, 2001). Despite

the growing awareness of wetland values and functions and the consequences of human

intervention, the issue of wetland loss and degradation has received lesser attention in many

African countries than other maj or environmental issues such as desertification and deforestation

(Acreman and Hollis 1996).









1.2 Significance

Understanding the impacts of draining and cultivating wetlands may encourage farmers to

adopt more sustainable practices to sustain their livelihoods. Previous studies in Tanzania have

tended to focus on maj or wetlands that are mostly associated with activities of greater economic

importance such as large-scale agriculture, transportation, and fisheries (Pallela, 2000). Sutton

(1969) and Lema (1996) who described vinyungu-farming systems and the associated

technologies reported that this system continues to receive little attention as it is largely

considered a side-line agricultural activity. Mkavidanda and Kaswamila (2001) looked at the role

of vinyungu in poverty reduction and concluded that vinyungu are a key factor in sustaining

livelihoods and reducing poverty. Majule andMwalyosi (2003) examined soil and water

characteristics and found that soils under vinyungu cultivation were acidic with a pH of 5.1 to 5.5

and the water samples downstream had traces of agrochemicals and pesticides, implying that in

the long-run vinyungu farming may reduce soil and water qualities as well as agricultural

production in both the vinyungu plots and further downstream.

In contrast, this study looks at the ways in which vinyungu farming has changed over the

years and how this change may have influenced social and environmental changes. It also looks

at the patterns and extent of wetland cover change that may have resulted from transformations

in vinyungu farming. Quantitative data derived from this study will generate knowledge on

changes in land use, the extent of wetland conversion and their socio-economic and

environmental consequences. This knowledge will be useful for a better understanding of the

relationships between humans and observed environmental changes and may encourage best

management practices around wetlands. No other known study has looked at the effect of

Tanzanian vinyungu in landscapes.









1.3 Objectives

The primary obj ective of this thesis is to study changes in land use, specifically wetland

conversion, over time and the environmental effects associated with those changes. I wish to

contribute to the knowledge on wetland loss, its causes and effects, all of which are essential for

policymaking and management of wetland resources.

Specific aim 1: To investigate what kind of transformations have occurred in vinyungu-
farming practice over time;
Specific aim 2: To propose the factors that may have influenced the recent
transformations in this farming practice;
Specific aim 3: To determine what is the effect of the vinyungu-farming system;
Specific aim 4: To identify and document change of Ndembera wetland area used for
agriculture between the period 1977 and 1999.

1.4 Study Questions

To meet the study obj ectives, I used information from farmers' interviews in Ndembera

wetland area, Iringa region, Tanzania (Figure 1-3). To identify long-term changes in land use in

this area, I sought data on land-use/cover changes over time. This information was used to

explore factors of wetlands transformation and changes in land cover over time. The following

questions were addressed:

Research questions 1: How is vinyungu-farming practice related to household and
farmer characteristics?
Research questions 2: How has the vinyungu-farming system evolved over time?
Research questions 3: What have been the possible driving forces behind the changes in
vinyungu-farming practice?
Research questions 4: What is the effect of vinyungu-farming system






























Make e' IN ~~be Ditrc omI

e- Imbe



8 $-Ludewa




?A 1n SQ 4 3fR S9 40 47
500000 600000 700000 800000 900000 1000000


Figure 1-1. Map oflIringa region showing the study area (A). (Source: based on 1:2,000,0000 Tanzania Administrative Map of 1989,
by Surveys and Mapping Division. Ministry of Lands and Human Settlement). The inset (B) shows the location Iringa
region in Tanzania.


































A B
Figure 1-2. Typical vinyungu grown potatoes (A) and peas (B). Both pictures taken in Usengelindete village, Iringa District of Iringa
Region, June 2005.















LYAN DE MBERA SWAM P .-
20 0 20 40 Km I





Ili~nga Distrc





Mgung Inthf ul~ nei


in Kilolo District
kngw ko es

te e si id~lembe '
I~g~e-_ Itimbo



, Nyoror


A70 000 720000 740000


r+


760000 790000 900000


LEGEND
o Towns

I~Bound Surface Road
Loose Surface Road
(VRiver


Lyand em bera Swamp

Country Boundary
Regional Boundary
District Boundary


I'


0 7 000


027 000


740000


67 0000


97 0000


900000


Figure 1-3. Enlarged map (A) shows Lyandembera swamp and surrounding villages. Subset image (B) shows the location of
Lyandembera swamp (The Study Area) in Iringa region.












CHAPTER 2
LITERATURE REVIEW

2.1 Introduction

This chapter reviews the literature that undertakes the role of putting together important

concepts describing land-use and land-use changes in Tanzania with a focus on the

transformations in vinyungu-farming system. The review is organized into two maj or parts. The

first part reviews some important concepts ofland-use/cover change as they relate to wetland

conversion to agriculture. It also reviews theories underlying transformations of agricultural

practices and their effect on wetland resources. These include human-settlement models of land

use, population growth and its implications for agricultural land use, and drivers of smallholders'

farming decisions. In addition, this part of the review also discusses the state of agricultural

production in Africa and reasons for its decline; use of wetlands as an alternative landscape

component for improving crop production; and irrigation practices as they relate to wetland-

related policies in Tanzania. The second part of the review summarizes the theoretical

background with a focus on land use at the household level and explores important relationships

and variables amenable to land-use changes.

2.2 Conceptual Framework

2.2.1 Concepts of Land-Use and Land-Cover Change

Wetlands conversion to agriculture falls under the larger topic of land-use/cover change

(Gerakis and Kalburtji, 1998; Jensen et al., 1995; Lambin et al., 2001; Liu et al. 2004; Meyer and

Turner, 1994; Meyer and Tumner II, 1996; Meyer and Turner II, 1992; Ringrose et al., 1988;

Turner et al. 1993; Wang et al.; 2006). Meyer and Turner II (1992) define the term "land use" as

human exploitation of the land and the term "land cover" as the physical and biotic character of










the land. Land use tends to cause land-cover change that, in turn, leads to further land-use

changes. Land-cover change may involve either land-cover conversion, i.e., "change from one

cover type to another," or land-cover modification, i.e., "alterations of structure or function

without a wholesale change from one type to another (Briassoulis, 2000).

Land-use change may involve "conversion from one type of use to another", i.e., "changes

in the mix and pattern of land uses in an area," or "modification of a certain type of land use"

i.e., "changes in the intensity of this use as well as alterations of its characteristic

qualities/attributes" (Briassoulis, 2000).

Land-use/cover change is driven by many factors including changes in the biophysical and

socio-economic drivers (Meyer and Turner II, 1992). The driving factors/forces are "a complex

set of actions and rationales that give rise to proximate causes" (Mertens et al., 2000, 984).

These forces could be "externally-driven, e.g., natural hazards, or internally-driven, e.g.,

population growth" (Mertens et al., 2000, 984). Turner et al. (1993) listed four examples of

major driving forces of land-use change: population, technology, political economy, and political

structure. Farming practices are listed as among the maj or causes of land-cover changes in

tropical Africa (Lele and Stone 1989; Meyer and Turner, 1994; Turner et al., 1994).

As driving forces of land-use/cover change interact differently to one another in different

spatial-temporal settings, it is important to examine causes of land-use/cover change within a

specific spatial-temporal setting. The method of examining both land-use/cover changes involves

selection of the land-use/cover types to be analyzed, determining the driving forces and process

of change that can be detected, and describing and explaining the linkages between land use and

land cover (Briassoulis, 2000). This study focused on a traditional farming system called

vinyungu, common in wetland areas of Iringa region, Tanzania, to determine how this farming










system has evolved over time, what are the driving forces behind this evolution, and what are its

effects on wetlands.

A challenge in analyzing land-use/cover change lies in linking the driving forces of change

to the observed land-use/cover changes. There is considerable variety of theoretical and

modeling frameworks and tools that are used to conceptualize and operationalize land-use/cover

change issues. The theoretical literature on land-use/cover change includes theories on social and

economic determinants of land-use/cover change.

Economic theories and models of land-use/cover change use concepts and procedures from

economics, e.g., "prices of the factors of production, of products and of services, transport cost,

marginal cost, and economies of scale, externalities, and, above all, utility" (Briassoulis, 2002).

"All behavioral assumptions made refer to the model of the rational, economic, utility

maximizing humans" (Briassoulis, 2002). Examples of economic theories and models include

von Thtmnen's agricultural land rent theory (1966); location models (see, for example, Alonso,

1964; Weber 1929; Losch, 1954); and settlement patterns models ( see, for example, Blaikie,

1971; Chisholm, 1962). Some of these theories have been elaborated in the subsequent sections

of this chapter to show how they relate to land-use/cover change and their relevance to this

study .

Theories on social determinants of land-use/cover change are based on ideas from the

social sciences and emphasize "the importance of human agency, social relationships, social

networks, and socio-cultural change in bringing about spatial, political, economic and other

changes" (Briassoulis, 2002). Theories and models used in the social approach include agent-

based theories which focus on the agents of change (i.e., users of land) and their

interdependencies as well as those theories that adopt perspectives in the nature-society









theorization traditions (see, for example, Liverman, 1994; Malthus 1960; Merchant, 1990; Sack,

1990). Some of these theories have been elaborated in the subsequent sections of this chapter to

show their relevance to this study.

Theories of land-use/cover change use models that are either descriptive or explanatory

(Briassoulis, 2000). According to Briassoulis (2000) "description of land-use change documents

changes from one type of land use or cover to another over a given time period and within a

given spatial entity" while "explanation attempts to address the question of 'why' these changes

have occurred (or, are occurring) and to uncover the factors or forces that bring about these

changes directly or indirectly, in the short- or long-run" (Briassoulis 2002).

Description of land-use/cover change can be achieved using GIS modeling approaches

whereby maps from different time periods are overlaid to identify the location and assess the

magnitude of change. Explanation of why these changes take place can be achieved through

socio-economic surveys and physical observations. Both methods have been employed in this

research. The descriptive part is covered in Chapter 4 while a review of theories that explain

"why" these changes occur (and their effects) is covered in the following sections.

2.2.2 Human-settlement Models of Land-Use

Different theories attempt to describe and explain the factors influencing human settlement

and land use. Most of these theories are based on economic concepts. One of the pioneering

economic theories of land use is von Thiinen's agricultural land rent theory, developed in 1826,

where land rent is defined as the price paid by the tenant for a particular land use on a particular

piece of land (von Thiinen 1966; original in 1826). Through this theory, von Thiinen modeled the

locational distribution of crops (or, land-use patterns) in a landscape as an algebraic function of

yields, market prices, production costs, transportation rates and distance to market.










According to von Thuinen, land rent decreases with increased distance from the center

(city/ town/ market). The rent one can afford to pay for a land-use activity on a parcel of land

depends on the value of the products produced on that land. This value decreases with increased

distance from the market. Therefore, a land-use activity with the highest value of output is found

on land with the highest rent, i.e., closest to the market (this would be the most valuable piece of

land in a village context). Transportation costs increase with increased distance from the center.

Therefore, transportation costs are inversely proportional to the land rent. Since farmers tend to

balance land cost, transportation costs, and profit, they produce the most cost-effective products

for market near the center, i.e., they cultivate the high market value-crops near the city/ town/

market (lest transportation costs reduce profit to zero) and lower market value-crops further

away from the city (meant for local use). Therefore, land-use patterns are influenced by distance

to roads, access to markets, and access to transportation. This may be true for wetlands use in

most of eastern Africa, including Iringa region in Tanzania. Dixon and Wood (2003) report

commercialization and communication networks that have spread throughout eastern Africa as

some of the factors increasing the market demand for food and other wetland products. Thus,

wetland use is increasingly shifting from subsistence agriculture to more commercialized

agriculture where locally marketable crops are grown (Olindo, 1992).

Chomitz and Gray (1995) tested the model in Belize and found that market access and

distance to roads strongly influence land use. According to the study, agriculture becomes less

attractive with increased distance from the market.

Interpreting von Thtmnen' s theories in a rural setting, Chisholm (1962) found that there

exists a relationship between land use, distance from market, and distance from the settlements.

Active farms require adequate labor and frequent attention and are therefore found closer to










settlements. The amount of labor required for a land use activity (e.g., agriculture) tends to

decrease with increased distance from the market. Therefore, closer to the market one expects to

find more labor-intensive crops, more labor use and more human settlement, hence intensive

land use.

Based on the frontier thesis (Richards 1990), people first search for unsettled yet desirable

lands to settle on. Later, secondary and tertiary frontier zones are created, land-use intensification

occurs, and eventually, settlement frontiers are expanded (Richards, 1990). The spatial scale of

penetration and land transformation is facilitated by such factors as expanding economic

demand, population growth and technological advancement (Richards 1990). Therefore, demand

for land and the ability to work on the land tend to influence expansion of settlement frontiers,

provided unoccupied land exists. As the population grows and land shortages in areas

traditionally farmed increase, more and more people in eastern Africa are forced into marginal

areas in search of agricultural land (Dixon and Wood, 2003). Wetlands may have become new

agricultural frontiers, replacing dry land margins that had been subj ect to spontaneous

agricultural settlement for decades.

In summary, access to roads and markets seems a common link among the different

models presented above. Construction of roads, distance to markets, and market demands

influence land-cover change. Roads increase access to land by enabling people to reach areas

that were not previously reachable. Markets accessibility and market demands encourage land

users to maximize the utility of surrounding land. These two factors tend to bring an economic

growth to an area that, in turn, attracts human settlement.

2.2.3 Population Growth and Agricultural Land-Use

Many researchers have linked population growth to increased demand for agricultural land

and consequently, agricultural extensification and intensification. Several theories have been









developed to examine this link and research has been conducted to test these theories, providing

insight and different perspectives on the issue.

One of the earliest theories on the effect of population growth on agriculture emanates

from Malthus (1960) who hypothesized that food production, growing at a linear rate, would

eventually be overtaken by populations growing exponentially. His original predictions have

materialized in some areas. According to Bilsborrow (1987), "Food production per capital in the

developing countries is barely keeping pace with population growth in the developing countries

despite the Green Revolution" (Bilsborrow 1987, 199).

Malthus also postulated that, population growth would negatively impact the environment,

potentially causing irreversible land degradation that would, in turn, hamper food production.

Bilsborrow and Ogendo (1992) showed that, population growth has indeed contributed to land

degradation in some areas, sometimes threatening food production. From a neo-Malthusian

perspective therefore, degradation is inevitable where levels of exploitation exceed the carrying

capacity of individual land resource such as individual wetlands. The disappearance of parts of

the Usangu wetland in Tanzania is one example of the effects of population pressure on wetlands

(MWLD, 2001). However, the Malthus hypothesis assumed unchanging methods of production

(i.e., constant technology). Today, economists argue that under well-functioning markets, as the

population grows and land resources become scarce, technologies are developed to improve

production as well as land management. As a nearby example, a study conducted in Machakos

District, Kenya, showed that population growth, market development and capital availability led

to technological advancement, which, in turn, led to increased agricultural productivity and

improved land and water management (Tiffen and Mortimore, 1994; Tiffen et al., 1994).









The economists' point of view emanates from Boserup's (1965, 1981) original ideas that

emphasize the importance of progressive adaptive management. Challenging Malthus's original

ideas, Boserup postulated that as the population grows, the demand for food increases and arable

land becomes scarce, forcing farmers to make adjustments in production so as to improve the

quality and productivity of land. These adjustments, which occur after the expansion of

cultivated land, include adoption of land-intensive technology, i.e., reduced fallow period,

adoption of new technologies, increased labor inputs per unit of land, and use of natural or

artificial fertilizer.

Generally, however, the impact of increasing population on land use and environmental

sustainability today appears more complex than what either Malthus or Boserup hypothesized,

and varies from system to system. Bilsborrow (1987) reports that, intensification/ technological

development is dependent upon environmental conditions of the area (soil fertility, rain). For

example, technological change, in terms of increased irrigation, happens when rainfall is

seasonal or irregular or when surface or underground water is available and can be tapped

(Bilsborrow, 1987). Netting (1968, 1993) reports that, farmers make several adjustments, such as

crop diversification, to cope with poor environmental conditions and maximize crop output.

Boserup (1965) neglected the possibility of an alternative economic response to increased

population pressure: an increase in the area of land under cultivation. This response, called

extensification, is achieved through clearing more of one' s own land, appropriation of

neighboring lands, or, migration to other areas with arable lands (Bilsborow and Ogendo 1992).

One of the effects of agricultural expansion is deforestation, especially of tropical forests and

highland forested areas (Bilsborow and Ogendo 1992). Today, this expansion is affecting

wetlands as well. Up until the 1940s for example, the Sukuma people of Mwanza region,









Tanzania, cultivated on the hill slopes and left the valley bottomlands for cattle grazing (Pingali

et al., 1987). Today, the bottomlands are cultivated and the adj acent swamp vegetation is cleared

to provide more land for cultivation. A similar trend of converting wetlands to agricultural land

is taking place in Iringa region.

Boserup also neglected the influence of market demand on technological development.

Pingali et al. (1987) clearly show that local demands, including market demands, cause land-use/

cover change. Therefore, site-specific carrying capacities, based on often unique environmental

and socio-economic characteristics, are a maj or catalyst of change in the agricultural farming

practices and, consequently, the way wetlands are being utilized and managed.

2.2.4 Drivers of Smallholders Farming Decisions

Many regions of the developing world consist of substantial rural populations that rely

largely on farming as a principal source of income, food and employment. The most widespread

kind of farm unit is the small family farm or smallholding. Netting (1993) describes smallholder

farmers as rural cultivators practicing permanent, diversified agriculture on small farms.

Population density is generally high. Therefore, smallholders live under conditions of scarcity,

deriving their livelihood from an intensively cultivated holding, usually less than 2 to 5 ha.

Central to the small-hold intensive agriculture is the family unit or household. A household

is a cooperative work group engaged in production, distribution, transmission, biological and

social reproduction, and co-residence (Netting 1993; Wilk 1991) and may consist of a single

family, one person, an extended family, or any other group of related or unrelated persons

sharing living arrangements.

Farming decisions by a household are influenced by many factors including resource

availability (e.g., land, labor), economic status of the household, local traditions, and external

factors such as government policies and environmental factors. Differences among households










lead to differences in choices each household makes regarding what to cultivate, where and how

to cultivate, amount of land to cultivate, what crops to be marketed or preserved, whether or not

they should access additional land to cultivate, and how to use labor and cash resources.

Land tenure security, often with sellable, rentable, heritable rights (although such property

regimes can co-exist with communally managed resources) is important for small-hold

householders to have. This is especially important where intensive agriculture is practiced for

often this involves land improvements. Smallholders usually hold their most productive land as

private property while less intensively used land is more likely to be held as commons or rented

out.

Smallholder agriculture is sustainable where production is predictable, sufficient to feed

the producers, and stable over the long run. Factors like seasonality could affect agricultural

production and sustainability, hence affecting rural household food supply and income (Meertens

1999, Morgan and Solarz 1993). Scanty rainfall and droughts are among the major seasonality

factors affecting agricultural production (Alexandratos 1995). Technological innovations such as

irrigation may be employed in cases of droughts or seasonal rainfall (Bilsborrow 1987). Farmers

are therefore attracted to environments with a higher probability of sustaining production. This is

the case especially in drier and drought-prone areas where the availability of moisture in the soils

throughout the year and the ability to retain nutrients make wetland cultivation the most

sustainable risk management strategy.

Labor and economic resources available to the household influence household land-use

activities. Household members share different responsibilities to minimize costs associated with

assembling and scheduling labor. In most cases, very young and old members of the household

are excluded from agricultural activities. Landholders maximize their financial income by









allocating labor to activities they perceive will provide the greatest financial return on their labor

investment. When additional labor is required, farmers either hire laborers or alternate manpower

with other farmers (Meertens 1999). This may be the case especially for the harder tasks such as

clearing land on heavy wetland soils. In cases of limited labor availability or cash to hire

laborers, agricultural expansion or intensification is rare or minimized (Boserup, 1965).

Smallholders are not economically isolated. Income sources range from farm, off-farm, to

non-farm activities (Saith, 1992). Rural households depend mostly on farm activities as one of

several sources of income. Whether or not a household will diversify its income sources is

dependent upon agricultural production, market values, and economic status of the region

(Lipton and van der Gaag, 1993). Off-farm income opportunities may pull some labor off farm

due to low marginal return on intensive farm labor. Non-farm income provides cash for running

agricultural activities as well as provide for household needs, further pulling some labor off the

farm. Generally, there almost always remains farm labor for subsistence food production

(Morgan and Solarz, 1993).

There is lack of literature on the role of gender on the current trends in vinyungu farming.

Reports have shown that, previously, women were predominantly the vinyungu cultivators

(Lema, 1996) and that the role of men was limited to hired labor in land preparation. However,

due to the growing economic interest in wetland farming, this study will explore whether gender

influences some aspects of present-day vinyungu practice, including ownership, farm size, and

types of crops grown.

Therefore, agricultural production methods and land use patterns of small-hold

householders are potentially influenced by land quality and tenure policies; environmental









factors such as rainfall; available resources such as labor, land, and income; access to markets,

and gender roles..

2.2.5 Decline in Agricultural Production and its Relation to Land Use

Many Sub-Saharan African countries face economic stagnation, rapid population growth

and natural resource degradation. The economy of these countries suffers from accumulation of

foreign debt, weak agricultural growth, declining industrial output, dwindling export

performance, declining institutions, and deteriorating socioeconomic and developmental

conditions (World Bank Review 1989). These countries are among the poorest in the world with

the farming populations constituting both the majority and the poorest segments of society.

In an effort to sustain economic growth, agricultural development programs have been

promoted since the 1960s and agriculture remains the backbone of most of these countries'

economy and an important foreign exchange earner. Unfortunately, agricultural production in

most African countries is declining and food deficit is common among these countries. A

number of factors, including biophysical conditions, human resources, the economic

environment, infrastructure and government policies, contribute to this decline.

The soils of most of Africa (as well as those of some parts of Latin America and Asia), for

example, show great variability in quality and are, to a great extent, infertile. Wiebe (2003)

reports that, although the "quality of all land is lowest in the Middle East and North Africa, the

quality of cropland is lowest in Sub-Saharan Africa". Whereas 16% of cropland in Asia, 19% in

the Middle East and North Africa, 27% in Latin America, 29% in the developed countries, and

over 50% in Eastern Europe are considered to be of the highest quality, only about 6% of Sub-

Saharan Africa' s cropland is considered of high quality (Wiebe, 2003). This factor contributes to

high production costs and low productivity.









Global productivity gains (influenced by technological advancement) have been shown to

outweigh global productivity losses caused by processes such as soil erosion, nutrient depletion,

and salinization (Wiebe, 2003). However, this is not the case in many parts of Sub-Saharan

Africa where soils are either poor or fragile. In these areas, agricultural management practices

are not well developed, and, productivity levels are already low even though the need for growth

is high (Morgan and Solarz 1993; Wiebe, 2003). In fact, Sub-Saharan Africa makes the least use

of modern agricultural input in the world (Morgan and Solarz 1993). Scattered populations,

common in some parts of Africa, further suppress the development of modern agriculture in that

it becomes expensive for the government to support scattered populations and production may be

limited by many factors including labor, transportation costs, and access to knowledge and

technologies.

In most parts of this region, arable land is fast approaching limits of sustainable agriculture

while rainfall remains seasonal and insufficient (Alexandratos 1995). Where rainfall is plentiful,

most such areas are already under some form of protection hence inaccessible for agricultural

production (Alexandratos 1995). According to Morgan and Solarz (1993), Sub-Saharan Africa

has been more adversely affected by droughts than any other region in the world, especially

between 1982 and 1985 and later in the early 1990s.

The rate of population growth in Africa is known to be among the highest in the world.

Correspondingly, there is an increasing demand for food and agricultural land. In a few cases,

population density has been shown to be low or declining due to natural cause or out-migration.

In such cases, low population densities, coupled with land tenure system that guarantees access

to land, encourage extensive farming systems with low agricultural input (Morgan and Solarz

1993). Netting (1993) noted about the Koyfar of Nigeria that a decline in population density









leads to dis-intesification of agricultural production. Seasonality in labor demand and availability

limits farm sizes and productivity of farm systems, affecting land use and the overall agricultural

productivity (Morgan and Solarz 1993).

As a result of all the aforementioned factors, growth in agricultural production in the Sub-

Saharan African region is depressed and only in the range of 1.5% to 2% per year (Adams 1996),

reducing the amount of agricultural export and foreign exchange. Per capital food production in

the 1970s was high enough to meet the food and income requirements of the average household

in the region. Today, almost all southern African countries produce less and import more food

than they did in the 1970s (Pinstrup-Andersen et al., 1997). Although there has been recovery in

several places, agricultural and economic growth have not been able to keep pace with

population growth, posing a great challenge in addressing food, social, and environmental

insecurity problems (Pinstrup-Andersen et al., 1997).

To cope with this situation, more peasants have turned to a combination of off-farm

income generating activities, wage labor, and subsistence food cropping. Elmekki and Barker' s

study (1993) noted that, for example, Sudanese peasant families migrate in search of wage work.

The above review has shown that change in agricultural practices and land use in most

Sub-Saharan African countries is attributed to a number of factors including population trends

and human resources, biophysical factors, land tenure system, and the economic environment of

the country.

2.2.6 Wetlands: An Alternative Landscape Component for Crop Production

Seasonal rains and recurrent droughts have tremendously reduced Africa' s per capital food

production, reducing the availability of food and income in the region (Morgan and Solarz,

1993). Population growth is resulting in land shortages in areas traditionally farmed (Binns,

1994). Lack of adequate rainfall and population growth are maj or factors that have influenced









African countries to turn to wetlands as potential resources for overcoming the challenges of

poverty and hunger in the region (Adams, 1993a; Adams, 1993b; Adams, 1996; Dixon and

Wood, 2003). The rich soils with high moisture holding capacity and proximity to reliable water

sources make wetlands attractive for agriculture and are thus being converted to agriculture.

Because wetland soils can be waterlogged and anaerobic hence unfavorable for plants that are

not adapted to such conditions, farmers tend to raise the rooting zone above the saturated zone.

This fact dates many centuries back where wetland resources have been used to support

great ancient civilizations (such as those of the Maya and Inca) in the form of raised fields or

beds. Perhaps the most widely studied raised fields are the waru waru, originally made by pre-

Incan civilizations around Lake Titicaca, bordering Peru and Bolivia (Binford et al., 1997;

Camney et al., 1993; Denevan, 1970; Denevan and Tumner, 1974; Erickson, 1985; Erickson, 1988;

Erickson, 1999; Ortloff and Kolata, 1993; Tumner, 1994; Turner and Harrison 1981). These are

elevated earthen platforms that are 3 to 10 m wide and up to 200 m long, separated from each

other by seasonally flooded canals. Through this farming system, wetlands were converted to

cultivable land to increase land productivity so as to cope with food security needs of dense

human populations that lived in arduous environments. Raised fields were advantageous to

farmers for they raised the root zone above water-saturated soils; supplied reliable water;

conserved heat and protected crops against frost; retained dissolved and particulate nutrients;

enhanced nitrogen fixation; and mitigated soil salinity. These factors contributed to higher

productivity and more sustainable crop production compared to conventional methods of

cultivation (Binford et al., 1997; Carney et al., 1993).

Raised fields were also common among North American Indians prior the arrival of

settlers (Fowler 1969) and in Asia, especially China, in the fifth century B.C. (FAO 1980). In









Africa, raised beds were common in places like the Wahgi Valley of New Guinea prior to 350

B.C. (Lampert, 1967) and more recently, among the Koyfar of Nigeria (Netting, 1968). In most

of these cases, raised Hields were created not because of population or environmental stress but

rather may have developed as early forms of intensive agriculture.

In all cases though, it is the many benefits accrued from raised Hields that influenced

different societies around the world to develop and maintain this agricultural technology over

centuries. In the recent years, however, the outlook of raised Hields is somewhat different in a

number of places in that the negative impacts are on the verge of outweighing the benefits.

Increased population pressure and associated food scarcity in such places as East Africa

are driving farmers to drain and intensively cultivate wetlands (Dixon and Wood, 2003; Okeyo,

1992). Roggeri (1995) further supports this observation when he argued that although the

remaining large wetlands are those found in the tropical countries, the rate at which they are

being converted to non-natural state is increasing fast.

Unlike the Maya and Inca eras that were possibly sustainable, recent utilization of wetlands

is known to cause soil erosion, reduced water storage and quality, variable stream flows and

complete dryness of wetlands (Dixon and Wood, 2003; Thompson, 1976). These effects may

cause wetland loss, defined as loss of wetland area, due to its conversion to a non-wetland area

(Ramsar Convention Bureau, 1990). While "changes in land cover by land use do not necessarily

imply a degradation of the land" (Meyer and Turner, 1996:25), wetland loss tends to cause

wetland degradation, defined as the impairment of wetland functions (Ramsar Convention

Bureau, 1990). Intensive agriculture may also impair wetland importance as sources of building

material and food (mainly fish) suppliers. However, to most farmers, the immediate benefits

accrued through continued conversion of wetlands to agriculture may outweigh the long-term










ecological and socio-economic consequences. As such, although the spatial extent of wetland

conversion is limited by their characteristic wet conditions that make wetland soils potentially

anoxic, and most certainly, heavier and more difficult to cultivate than upland soils (McIntire et

al., 1992), farmers are known to develop techniques to cope with such situations. Therefore,

labor, time, seasons/climate, and technology are the factors that may limit or influence the rate

and amount of wetland conversion.

2.2.7 Irrigation Practices and Wetland Use in Tanzania

Tanzania continues to rely heavily on the agricultural sector for its economic growth and

improvement of the welfare of its people. Efforts have been made to promote agriculture,

including improvement of agronomic practices and introduction of improved crop varieties.

However, semi-arid climatic conditions and droughts in some parts of the country present severe

constraints on these efforts and agricultural productivity in general. As such, since colonial

times, there have been concerted efforts to promote irrigation practices at various scales (Majule

and Mwalyosi, 2003).

In Tanzania, irrigation, defined as a supply of water to cultivated plants by means other

than natural precipitation (Stern, 1989), is highly dependent on wetlands, e.g., rivers and

swamps. The country is estimated to have about 933,000 ha that have potential for irrigation.

However, only about 144,000 ha are reported to be under some form of irrigation (Ministry of

Agriculture, 1992). Based on scale and technology used, irrigation systems in Tanzania can be

grouped into three maj or categories, i.e., traditional or smallholder irrigation systems, the

modern small-scale/village irrigation system, and, medium to large-scale state farms/privately

owned irrigation estates (Mascarenhas et al., 1985; Mrema, 1984). Traditional small-scale

irrigation systems account for 120,378 ha of the irrigated land while large-scale estate farms

account for the remaining 23,622 ha.









The medium-to-large-scale farms or irrigation estates are state or privately owned. They

are centrally managed by either parastatals or private companies and generally have quite

efficient irrigation systems that also require large capital investment and well-trained manpower.

High value crops are grown for export and/or local consumption using these types of estates.

The modern small-scale/village irrigation systems were in most cases planned and

constructed throughout the country by the central or local government. Farmers were given the

responsibility over water distribution, land preparations, farm decisions, and scheduling of

activities. Unfortunately, almost all of them collapsed a few years after construction despite a

heavy investment in building them (Mkavidanda and Kaswamila, 2001). Two factors are

associated with the failure of these irrigation systems. First is the unfair distribution of water that

resulted in conflict among farmers who, in some cases, vandalized the facilities. Second is the

lack of or poor maintenance of the canals, which in turn, led to high rate of evaporation that

further hampered distribution of water.

The traditional or small-scale irrigation systems are the most common. They are owned by

individuals or a group of farmers and are usually small in size, often not more than 5 ha. Farmers

use low cost, temporary or semi-permanent intake structures to harness water from rivers,

springs and flood plains so as to produce food and cash crops.

Under this system, irrigation efficiency is very low as much of the diverted water is lost

due to seepage before reaching the farms. This farming system is associated with poor drainage

that causes siltation, and, poor infrastructure development that leads to unequal distribution of

water. Such problems are common in most parts of the country including Kilimanj aro (Banzi et

al., 1992), Lushoto (Kaswamila and Tenge, 1997), and currently, in Iringa region (Majule and

Mwalyosi, 2003).









The most common traditional irrigation system practiced in Iringa is called vinyungu

cultivation, a practice believed to have started in the 1890s (Culwick and Culwick, 1935).

Vinyungu are ridges or raised beds that are about 0.6 m high and 4 to 20 m wide created in

valley-bottoms and along streambeds. Smallholder farmers, mostly women, create vinyungu by

clearing the land, burning the cleared vegetation, plowing, creating ridges and ditches, and then

by harrowing to smoothen the ridges. Vinyungu are watered by surface run-off (a river or stream

originating from the upland areas) as well as subsurface run-off(springs). Since the valleys are

waterlogged during the wet season, hampering land preparation and cultivation, vinyungu are

cultivated mostly during the dry season. This farming system is common in Makete, Ludewa,

Mufindi, Iringa and Nj ombe districts oflIringa region and is practiced mainly by the Bena and

Hehe tribes of all social classes.

Overall, the area under irrigation in Tanzania is relatively small. However, reports indicate

that irrigated land in the country is diminishing in quality due to factors such as salinization and

sedimentation caused by erosion and perpetrated by poor management (Majule and Mwalyosi,

2003; Kaswamila and Tenge, 1997). This situation further diminishes the production potential of

the remaining irrigable land.

Unfortunately, unlike countries such as Uganda and Zambia for instance, Tanzania does

not currently have a national policy on wetlands, although efforts are underway to produce one.

The national water management strategy is also at the developmental stages. This means wetland

users currently lack guidance on best management practices over wetland resources. National

policies that can provide an enabling framework within which to improve the status of wetlands

sometimes promote actions which conflict in terms of improving the quality of wetlands. For

example, policies on agriculture, food security, and poverty reduction promote expansion of










irrigation, in most cases, regardless of where or how it is done or its consequences. Such actions

confuse policy implementers, i.e., local government and local resource users. Also, overlapping

responsibilities between policies are known to cause confusion as to who is to do what and can

end in no action in the areas of overlap. In summary, both existing and non-existing policies in

Tanzania currently make it difficult to manage wetlands effectively, and many may be

deteriorating.

Today, only three wetlands (the Malagarasi-Muyovozi, Lake Natron and the Kilombero

Valley) have some form of management following their designation as Ramsar sites, i.e.,

wetlands of international importance. The remaining wetlands are utilized for irrigation, among

other uses, without guidance from the government.

Due to poor technology, lack of capital, and, high maintenance costs that large irrigation

systems require, most rural farmers in Iringa (and Tanzania in general) confine themselves to

traditional irrigation practices.

2.3 Summary of the Theoretical Background and Formulation of Hypotheses

2.3.1 Summary of the Theoretical Background

This section summarizes the theoretical background with a focus on land use at the

household level and explores important relationships and variables amenable to land-use

changes.

Households make daily land-use choices (or, changes in land-use) with respect to their

perceived risk and expected financial returns. These choices are influenced by different factors

including household characteristics, ecological conditions, and economic conditions. Through

interviews, these factors were investigated to determine 1) how they influence wetland

conversion to agricultural land; and, 2) how they influence transformations in agricultural

practices in wetlands. Land-use changes eventually lead to land-cover changes that, in tumn, lead










to further land-use changes. GIS and remote sensing technologies were used to determine the

amount of change of the wetland area that is used for agriculture.

Analysis of drivers of smallholders farming decisions has shown that household

characteristics play an important role in land-cover change. In the case of this study, household

characteristics may influence the conversion or the non-conversion of wetlands to agriculture.

Such characteristics include resource availability, i.e. 1) land; 2) the amount of labor available

for agriculture (in terms of numbers and age both of which determine the labor force available

for agriculture and hence the amount of land that can be exploited); and 3) the amount of cash

available within the household for such purposes as labor hire, renting or purchase of land, and

technological improvement (e.g., use of fertilizer). Availability of cash and labor have also been

linked to agricultural expansion and intensification (see the analysis of population growth and

agricultural land-use, section 2.2.3). Under this theorization, two hypotheses were tested.

Hypothesis 1: Greater number of people within the working age in a household would be

associated with greater farm size and number of crops grown; and, Hypothesis 3: Farmers are

likely to use household income on farm inputs (i.e. fertilizer and labor) so as to improve crop

production.

Vinyungu-farming practice has, for many years, been predominantly a woman's activity that

supplemented upland food production. It is not anticipated, therefore, that gender and duration of

residence will have a strong association with vinyungu size or types of crops grown hence,

Hypothesis 2: Farmer' s gender or duration of residence in the village is not associated with the

size of vinyungu farming area or types of crops grown.

Ecological conditions, especially poor soil fertility and scanty rainfall, reduce per capital

food production (sections 2.2.4 and 2.2.5). A larger population and a high population density










reduce the amount of land available for agriculture (section 2.2.3). Shortage of arable land due to

poor ecological conditions or to a larger population, usually leads to agricultural intensification,

i.e., an increase in total input per unit of land area, or extensification, i.e., an increase in total area

cultivated (see, for example, the frontier thesis under section 2.2.2 and Bilsborow and Ogendo

theorization under section 2.2.3). As shortage of land in areas traditionally farmed increases,

more and more people are forced into marginal lands. Therefore, wetlands (historically

considered unproductive and unhealthy yet with rich soils and moisture-holding capacity) may

now be considered potential agricultural lands (sections 2.2.6 and 2.2.7). This transformation

may be further facilitated by construction of roads as roads tend to increase human access to

areas that were previously unreacheable. However, active farms require adequate labor and

frequent attention, they tend to be closer to human settlements (see Chisholm's findings in

section 2.2.2). Accordingly, in this research, the following additional hypotheses were tested:

Hypothesis 4: Members of households located within a short distance from the wetlands are

more likely to practice vinyungu farming system; and, Hypothesis 5: Greater expansion of

vinyungu farms would be associated with increased conversion of wetland area to agricultural

land.

Economic factors play an important role in influencing land-use/cover change. Based on

von Thtmnen theory, a land use activity with the highest value of output is usually found on land

closest to the market. On the other hand, transportation costs increase with increased distance

from the population center (including markets). To minimize production costs and maximize

profit, farmers produce high market value-crops near the population center and lower market

value-crops (or subsistence crops) further away from the population center. Therefore, distance

to roads (including access to transportation) and access to markets do influence land-use/cover










change (section 2.2.2). Based on this theorization, the following hypotheses were tested:

Hypothesis 6: Residing close to the main roads would be associated with greater proportion of

wetland converted to agriculture; Hypothesis 7: Residing close to a maj or market would be

associated with greater proportion of the wetland converted to agriculture; and, Hypothesis 8:

Residing close to maj or roads and markets would be associated with cultivation of a greater

proportion of cash crops than food crops.

To cope with food demands of growing populations, farmers make several production

adjustments so as to improve the quality and productivity of land. Such adjustments include the

adoption of land-intensive technology, e.g., adoption of new technologies, increased labor inputs

per unit of land, and use of natural or artificial fertilizer (section 2.2.3). In this research, the

following hypothesis was tested, Hypothesis 9: Households with greater proportion of land for

cash-crop cultivation would be associated with increasing costs of farm inputs (labor and

fertilizer).

2.3.2 Justification of the Hypotheses

This study uses information from a single time-point to assess factors of wetlands

conversion in the Ndembera basin. For this reason, I acknowledge that the causes of wetlands

conversion over the course of the past 30 years could not be completely determined through this

study design. However, there is overwhelming evidence showing trends of population growth in

these areas (and the country in general) during the same time period. Therefore, the basic

justification of the hypotheses tested in this thesis is that certain attributes of population pressure

in the wetland areas (for example, proximity to wetland areas, household size, and food

requirements) may have contributed to the increasing rate of wetlands conversions over time.

Consequently, I formulated hypotheses to test the current interrelationships in vinyungu farming

as a way of gaining knowledge of the possible factors of wetlands conversions over time. On









their own, the current interrelationships in vinyungu farming are expected to provide an

important input in any future policies aimed at sustainable wetlands use.









CHAPTER 3
STUDY OF FARMER AND HOUSEHOLD FACTORS INFLUENCING Y7NYUNGU
FARMING SYSTEM

3.1 Study Area

3.1.1 Location

Ndembera swamp is part of the Ndembera River system (also called Lyandembera) found

between Iringa and Mufindi districts of Iringa region, southwest of Tanzania (Figure 1-2).

The entire catchment occupies an area of 1,834.10 km2. The Ndembera river is perennial

and therefore of great importance, especially during the dry season, to farmers in the immediate

environments as well as those further downstream. The catchment is adj acent to the Tanzania-

Zambia (TANZAM) highway, one of the major highways in to the country that connects the

cities of Dar es Salaam and Mbeya.

3.1.2 Physical Environment

Topographically, dissected rolling hills and tangled mountain streams with numerous

shallow valleys of tectonic origin characterize Iringa region. Swamps are formed at valley

bottoms where rich organic matter accumulates, creating an ideal environment for agricultural

development.

The rainfall pattern in the region is mono-modal. The annual rainfall totals vary with years

but generally range between 500 and 1600 mm. The highland zone, mostly between 1,600 and

2,700 m above sea level (asl) on the eastern fringe of Iringa and Mufindi districts, has higher

rainfall that ranges between 1,000 and 1,600 mm per year. The midland zone, mostly between

1,200 and 1,600 m as1 on the maj ority of Mufindi district, has moderate to low rainfall, ranging

between 600 and 1,000 mm per year. The lowland zone, mostly between 900 and 1,200 m as1 on

the most westerly parts of Iringa and Mufindi districts, has low rainfall that ranges between 500

and 600 mm per year. Precipitation, the dominant factor governing wetland hydrology, is low









during the dry season (June to October) and high during the wet season (November to May) with

highest levels recorded between March and April. It is during the dry periods, i.e., June -

October, that farmers cultivate the vinyungu. The wet period, i.e., November to May, is spent

mostly in the upland plots. The annual temperature is typically lower (150 C) in hilly areas (with

extremes in June to July) and higher (between 250 C and 300 C) in lowlands (with extremes in

December).

3.1.3 Demographic Characteristics

In 2002, the human population of Iringa region was 1,490,892 (URT, 2003). The

population grew by 24% and 23% for Iringa and Mufindi districts respectively, over a period of

14 years between 1988 and 2002. These rates are on a high side when compared to the food

production rates in the region that are falling (URT, 1999). It is estimated that 20% of the

population of the entire region lives in the urban centers of Iringa, Mafinga, Makambako and

Nj ombe. The remaining 80% of the population lives in rural areas and engages in agriculture as

their main livelihood activity.

3.1.4 Socioeconomic Situation

The maj or contribution to the economy of Iringa and Mufindi districts comes from

agriculture. In Iringa district for example, agriculture contributes 81.7% to the district's Gross

Domestic Product (GDP) while other activities (livestock keeping, fishing, forestry, mining and

trading) collectively contribute the remaining 18.3%. However, much contribution to the 81.7%

comes from estates (including tobacco and tea estates). Otherwise, the majority of the

community consists of subsistence farmers who have a very low income and are generally

categorized as poor hence with less contribution to the GDP (URT, 2004a and b).









3.1.5 Agriculture

Iringa District has a total of 16,607.85 sq. km (or 1,660,800 ha) of arable land of which

only 120,612 ha (7.2%) are utilized for cultivation (URT, 1997). In Mufindi district, 133,200 ha

(i.e., 19.6% of the arable land) are cultivated. An even smaller percentage of the land suitable for

irrigation is irrigated. For example, an estimated 50,000 ha (or 500 sq. km) of land are suitable

for irrigation farming in Iringa but only 3,812 ha are actually under irrigated cultivation. The low

percentage is attributed to many factors including poor agriculture infrastructure and soil

infertility (URT, 1997).

Reduced soil fertility as well as frequent and prolonged droughts are some of the maj or

factors contributing to reduced production levels in the region in terms of yield per unit area.

According to the Danish Development Agency (DANIDA, 1982) droughts occur every after ten

years, causing food production to drop by more than half the normal production levels. Examples

include a very low maize productivity in 1996 where an average of 2 tons of maize per hectare

was harvested against the normal capacity of 6.5 tons per hectare, despite government subsidies

(URT, 1999).

Availability of water and fertile soils in wetland areas make them attractive for agriculture.

The numerous perennial streams in the region provide reliable water for both wet and dry season

farming and as such, the government of Tanzania is already promoting irrigation systems in such

ecosystems (DANIDA, 1982). While high soil water content of valley bottom swamps may make

cultivation difficult, farmers in Iringa region have devised vinyungu to overcome this problem

and improve crop production in the region. Over the years, this technology has been transformed

to cope with prevailing situations such as unreliable rainfall and increasing soil infertility. The

transformation, which includes further expansion of cultivated area, has not been examined to










determine its extent and implications on wetland resources and the overall sustainability of the

practice.

This study investigated the characteristics and nature of transformation of this technology

and evaluated the environmental implications of this transformation. Iringa and Mufindi Districts

were selected for this study for the following reasons: 1) they represent typical vinyungu

cultivation practice in Iringa; 2) there exist previous studies on the subj ect in Iringa district and it

was felt to be useful to conduct more research which builds upon the earlier work; 3) the

Tanzania-Zambia (TANZAM) highway that connects maj or cities of the country (Mbeya and

Dar es Salaam), traverses the two Districts and may have implications on both the marketing and

transportation of agricultural produce as well as on inputs used on the farms; 4) availability of

good quality aerial photos of the target area; and, 5) accessibility and logistics of sampling.

3.2 Material and Methods

In order to test the hypotheses, I examined the current trends in vinyungu-farming system

practices; explored the different factors that may be influencing farmers to change their

traditional farming system; and, analyzed how changes in this farming system may in turn be

influencing changes in the wetland cover in the study area. The main sources of data were

individuals and communities directly involved with vinyungu farming. Formal questionnaires

were used to gather information from individual farmers. The research also adopted a

Participatory Rural Appraisal (PRA) approach that has been proved to be effective in eliciting

detailed quantitative information where community-related studies are concerned (Chambers,

1994; Brace, 1995). PRA also facilitates discussion among community members as they analyze,

investigate, and present their experiences. Village meetings were held to validate the information

collected.









3.2.1 Sample Profile and Sampling Procedure

The sample population was obtained from nine villages in Iringa and Mufindi Districts,

i.e., Usengelindete, Wasa, Muwimbi, Kiponzero, Ihemi, Ilandutwa, and Ifunda in Iringa District

and Rungemba and Maduma in Mufindi District.

A list of adults known to cultivate/own vinyungu was obtained from village leaders of all

villages within the study area. Farmers were disaggregated by sex, village and ability to express

oneself as determined by the village leadership. Ability to express oneself was assessed by level

of education (mostly those who have attended at least primary school and can read and write)

and/or active participation in previous village meetings. From a list of 3,000 adult farmers who

were eligible for the study, a random sample of six (6) adult farmers from each village was

selected for interview, making a total of fifty-four (54) interviewees (refer to Appendix A for

sample size calculation). Information gathered from the study area through village leaders and

ward onfcials showed that of those engaged in vinyungu farming nearly 95% grow both food and

cash crops. From this information, I wanted to create a sample of vinyungu farmers with a 90%

confidence that it would include farmers who were growing both food and cash crops. Sample

size for this confidence interval was 51 farmers (Appendix A). I chose to select 54 farmers

because it was the closest number that gave me equal number of farmers, i.e., six farmers from

each of the nine villages. In addition to the interviewed farmers, other farmers from the same

village as interviewees were invited for group discussions to examine their perceptions of

various community issues pertaining to vinyungu farming and use of wetlands (some of the

maj or issues covered in group discussions are listed in Appendix C). The information gathered

from the group discussions was used to support the generalizability of the findings.









3.2.2 Data Collection

Primary data on vinyungu farming were collected in three main phases and involved both

formal and informal survey methods. The first phase involved a formal survey using a pre-

designed questionnaire (Appendix B). The questionnaire was used to collect information on

vinyungu farming from the 54 representative farmers.

During the individual farmer' s interviews, interviewees were provided an opportunity to

decide whether or not they were willing to participate, to ensure a participatory process. District

staff accompanied me to all the interviews and assisted with translations where necessary.

The questionnaire was read to the interviewees who also had a blank copy for reference. In

all cases, the research team, i.e., an interpreter (a District staff) and myself, recorded interviewee

responses on the questionnaire. The questionnaire was designed to generate information on the

factors driving farmers to cultivate vinyungu; the role of vinyungu in supporting livelihoods;

reasons for current transformation in the practice of vinyungu farming; and, on the environmental

implications of the farming system (Appendix A).

The questionnaire was divided into three sections. The first section focused on household

information, i.e., general demographics, migrant status and motivations for living in the area,

community identity, education level, main activities, and society affiliation. The second section

focused on land availability, crop preferences, fallow period, and trends in wetland use and

status. The last section focused on labor, input, tools and problems related to vinyungu farming.

The questionnaire has been used to test all the nine study hypotheses. For hypothesis 1,

"Greater number of people within the working age in a household would be associated with

greater farm size and number of crops grown", information on age, main occupation, land

ownership, land acquisition, land type owned, size of upland plots and vinyungu, and vinyungu

area cultivated food and cash crops was gathered through the questionnaire.









For hypothesis 2, "Farmer' s sex or duration of residence in the village is not associated

with the size of vinyungu farming area or types of crops grown", information on the number and

type of crops grown and on the proportion of land used for both cash and food crops as they

relate to sex and duration of residence in the village, was sought.

For hypothesis 3, "Farmers are likely to use household income on farm inputs (i.e.,

fertilizer and labor) so as to improve crop production", households were asked how much they

spent on fertilizer and labor hire and whether or not both labor and fertilizer were easily available

during the peak periods. Farmers were also asked how much money they accrue from cultivating

the wetlands.

For hypothesis 4, "Members of households located within a short distance from the

wetlands are more likely to practice vinyungu farming system", households were asked whether

they own, rent, rent-out or borrow land; the size of the land, type of tenure and length of

ownership; and distance traveled to the farm. They were also asked the location of their plots

(upland/wetland); the period of cultivation (wet/dry season); and the condition of their upland

fields in terms of production.

For hypothesis 5, "Greater expansion of vinyungu farms would be associated with

increased conversion of wetland area to agricultural land", farmers were asked about the status of

the wetland use (i.e., whether or not wetland utilization for cultivation has increased); status of

the wetland (i.e., whether or not wetland size and vegetation cover has declined since they

moved into the area); soil fertility; erosion; number of residents (i.e., whether or not there is a

population increase); factors that attracted farmers to cultivate vinyungu in this area; and whether

or not there were any management plans for the surrounding wetlands. Information on farmers'










perception of environmental problems associated with vinyungu farming was also sought. This

hypothesis was further tested using GIS and remote sensing technologies.

For hypothesis 6, "Residing close to the main roads would be associated with greater

proportion of wetland converted to agriculture", farmers were asked of the distance between their

plots and the main road. Most farmers have a clear understanding of the distance they walk to the

main road or their farms. Linear regression was used to determine whether or not distance to

main roads influenced greater conversion of wetlands to agriculture.

For hypothesis 7, "Residing close to a maj or market would be associated with greater

proportion of the wetland converted to agriculture", farmers were examined to determine

whether or not they develop land for vinyungu farming in relation to the distance to the market.

The distance from the market to the center of each village was measured as an estimate distance

from the farms within that village to the main market.

For hypothesis 8, "Residing close to maj or roads and markets would be associated with

cultivation of a greater proportion of cash crops than food crops", information on distance from

farms to maj or roads/ markets, types of crops grown, and proportion of land occupied by both

food and cash crops was sought.

For hypothesis 9, "Households with greater proportion of land for cash-crop cultivation

would be associated with increasing costs of farm inputs (labor and fertilizer)", information on

labor (hired workers) per day, the time of the year the hiring occurs, and tasks performed, was

gathered. In addition, information on use of inputs and tools in cultivation (types and costs of

fertilizer, types and number of agricultural tools) was also asked.

The second phase involved a PRA exercise (Chambers, 1994; Brace, 1995) that brought

together all the 54 interviewees to discuss issues that cut across village boundaries, e.g., water










flow and rate downstream. In addition to discussing vinyungu-farming practices in general,

information was sought on reasons for transformations in the practice as well as on their views

about the environmental implications of the practice. Semi structured discussions were core to

the exercise, but in addition, ranking techniques, Venn diagrams, and proportional piling (where

participants used stones to indicate the relative importance of different information) were used in

gathering information from farmers. Participants recorded key outputs on flip charts. In addition,

all the meetings were tape-recorded and research assistants who also attended the meetings

produced the transcripts.

Village meetings were conducted in all nine villages to 1) share the outcome of the PRA

exercise with all the other farmers; 2) gather farmers' views on the outcome of the PRA exercise;

and, 3) obtain any additional information and views about vinyungu. In trying to determine

farmer' s perception on environmental aspects associated with vinyungu farming, four to five

meetings were conducted per village, i.e., within four to five hamlets (sub-villages) per village.

A total of forty short (one hour each) meetings were held. This was done to allow a more

thorough discussion among villagers, in smaller groups, of the pertinent environmental issues.

Prior to all village meetings, a formal request was sent to the Village Chairmen of all nine

villages to call for a village meeting where a formal introduction was made to the village

community, the purpose of the research made known, and farmers cooperation in the interviews

requested. A proportionate number of adult (>18 years old) male and female participants was

requested.

A total of 2,302 farmers attended the meetings. Assisted by District staff, I moderated the

discussions to avoid bias or dominance by influential members of society. The discussions were










guided by a group questionnaire (Appendix C). These meetings were also tape-recorded and their

transcripts were compared to those of the 54 farmers.

All activities in the field were preceded by a formal introduction to the village

representatives. District officials introduced the researcher to the village government and study

participants as a student collecting information on the past and current vinyungu-farming

practices and their relation to recent wetland-cover changes.

3.2.3 Methods of Data Analysis

This study explored the extent of vinyungu-farming practice in wetland areas and its

perceived effects on the environment. A combination of quantitative and qualitative methods was

used to meet the obj ectives of the study. I have shown previously that little is known about the

effects of vinyungu-farming system on the environment. For this reason, data gathering was

made mainly qualitative for in-depth interviews and focus groups in order to generate detailed

data that are likely to offer the context under which vinyungu-farming practice changes over

time. Information gathered from the randomly selected maj or vinyungu cultivators was mainly

quantitative.

Methods of analysis are divided into three parts. The three parts of analyses used relevant

data to (i) describe the background information of the farmers, (ii) establish important

associations between the farmers and various aspects of vinyungu-farming practice, and (iii)

develop statistical models that can predict use of resources in vinyungu farming. In this chapter,

farmer' s perception on various environmental issues that may be associated with vinyungu

farming, was assessed. To assess and document the magnitude and extent of change of the

Ndembera wetland area used for vinyungu cultivation, GIS and remote sensing methods were

used (Chapter 4).









3.2.3.1 Descriptive Statistics

Within this part of the analysis, important demographic features of the farmers who

participated in the study are described. Descriptive statistics including means for continuous

variables and percentages for categorical variables are computed for all participating farmers.

Qualitative data explaining demographics of the villagers are also summarized in this part of

analy si s.

3.2.3.2 Measures of Associations

The effect of vinyungu-farming practice on the environment was partly assessed by the

association between some important farmer or household attributes and the environment, which

includes both the infrastructure, e.g., use of farm implements, and farm characteristics, e.g., size,

and location. In this part of analysis association tests were conducted to determine whether there

is an association between (i) sex and plot size, (ii) household size and plot size, (iii) distance

from home to plot and plot size, (iv) distance from the plot to major road/market and plot size,

(v) number of cash crops grown and plot size, (vi) population size and vinyungu farming area.

The Pearson' s Chi-square tests evaluated statistical significance of these relationships. Important

relationships obtained from the qualitative data are also narrated under this part of analysis. In

addition, prior to multivariable analysis, t-tests were run to compare mean levels of size of plots

(wetlands and uplands), years of vinyungu cultivation, size of vinyungu rent out, travel times, and

proportion of vinyungu planted cash/food crops.

3.2.3.3 Multivariable Analysis

Detailed analysis of some of the relationships examined in Part II was conducted using

multivariable analysis. Multivariable linear regression was employed in this section to determine

factors affecting plot sizes and types of crops grown (dependent variables) for both wetland and

upland farming. This analysis will indicate the importance of each independent variable (distance










to the plot, distance to the market, ownership, proximity to maj or river, number of household

members within working age, sex, farm labor, use of fertilizer, duration of vinyungu farming,

duration of residence, and sex) on the size of plot used for vinyungu and other farming practices.

For this exploratory study, all these factors were entered in the model without using any selection

criteria.

All statistical procedures in the three parts of analyses were conducted using SPSS for

Windows version 13.0 statistical package. The significance level of all tests was set at p < 0.05.

3.3 Results: Farmer and Household Characteristics of Valley-Bottom Cultivation
(vinzyunzgu) around the Ndembera Swamp in Iringa and Mufindi Districts

This section presents the major findings of the study on farmer and household factors

affecting vinyungu-farming system. The results are presented in accordance to the three parts of

methods of analysis: descriptive statistics, bivariate relationships between farmer and household

characteristics and factors influencing vinyungu-farming system, and the multivariate model of

the factors influencing vinyungu farming. Results on farmers' perception on the environmental

issues that may be related to vinyungu cultivation are also presented here.

3.3.1 Sociodemographic Characteristics

The Hehe and Bena are the dominant tribes representing respectively 72% and 22% of the

total population. Minor tribes are the Jita, Kinga and Wawanji. The maj or tribes are from within

Iringa region while the minor tribes are from other regions in the country, indicating a minor

inmigration. The average number of people per household is six with approximately half of them

within the productive age of between 18 and 60 years old Based on the 2002 population and


i The International Labor Organization (ILO) recognizes children as those ranging between ages 5-17. The National
Youth Development Policy of Tanzania recognizes ages 15-24 as youths (below which are children). Age 18 is
considered "mature and able to vote" while 60 is a pensionable retirement age. For the purpose of this study, I used
"working ages" and not age distribution in the general population in my categorization. I have categorized ages 0-17
as children (potential labor force); ages 18-35 as youths (and the major labor force that is also dynamic and can own









housing census, the population distribution in the study villages stands at 20,695, i.e., 9,863

people (48%) between 0 and 14 years old; 5,880 people (28%) between 15 and 34 years old;

3,493 people (17%) between 35 and 59 years old; and 1,459 people (7%) at 60 and above years

old (URT, 2003). The maj or labor force lies between ages 15 and 59 which constitutes 45% of

the population. However, at the village level, 10 and 14 year olds (15%) are also involved in

some minor farm activities hence raising the farm labor force to 60% of the entire population in

the study villages.

Table 3-1 below summarizes the sociodemographic characteristics of the maj or vinyungu

farmers who participated in the survey. The results show that 63% of the farmers were men and

about 91% were married. The higher number of men in the sample may have been influenced by

the sampling methods where preference was given to a representative farmer that owned a

kmnyungu and who also had a proven ability to express oneself eloquently in the previous village

meetings. There was no significant age difference between men and women sampled. A maj ority

of the farmers had primary education and, as expected, almost all of them (96.3%) were engaged

in agriculture as their main occupation. The high proportion of residents with only primary

education is typical of rural areas in Tanzania, where agricultural activities are a maj or

occupation that starts at an early age and traditionally require no formal qualifications. Only a

small proportion of the farmers (27.8%) are members of farmers' cooperative unions, most of

which have been established only recently, i.e., in the early 1990s following a boom of paprika

as a cash crop. Cooperative unions prevalent in the 1970s became defunct in the late 1980s

mainly because most farmers felt exploited by them and also because they made more loss than

they did profit.


land); ages 36-60 as adults (and an important work force, perhaps more settled); and, ages 61 and above as the
ederly (and not a major labor force).









3.3.1.1 Land tenure/ownership

With regards to land ownership, Table 3-1 shows that almost all the farmers use their own

land for farming activities. While most of the farmers (79.6%) acquired their land through

inheritance, about 15% bought the land they currently own. The rest (5.6%) have acquired land

through the village government.

Through group discussions, farmers indicated that individuals from the same village who

are not landowners or who want to increase productivity, might request access to vinyungu

owned by colleagues or relatives both within and outside their village, especially from

landowners that are unable to cultivate their entire property due to lack of human or financial

resources. This pattern however, is becoming rare while renting is slowly increasing.

Information gathered through group discussions also indicated that vinyungu ownership

among households differed in numbers from one kmnyungu per household to as many as ten

vinyungu per household and that these vinyungu are mostly within the same valley the vinyungu

farmers live in. A few vinyungu may be scattered over several valleys, a factor driven by the size

of the valley and accessibility.

The data in Table 3-1 also show, as expected among farmers in this area, that typically

farmers in the study area would have both vinyungu and upland plots for their farming activities.

Group discussions confirmed that villagers may engage in other activities but all of them did

indeed cultivate the uplands, vinyungu, or both. However, there is a substantial difference in the

size of land owned for these types of plots. Nearly 70% of the farmers own at least three acres

(1.2 ha) of upland plots compared to only about 1 1% who own the same acreage for vinyungu

plots.









3.3.1.2 Crop production, crop preference and land use

In group discussions in all nine villages farmers reported that vinyungu farms have become

extremely important for both food and cash crop production especially when compared to upland

plots. For example, upland production of maize as the main food and cash crop has declined

from an average of 30 bags/ha in the 1970s/80s to about 15 bags/ha in the 1990s/00s, even after

application of inorganic fertilizer (Table 3-2). Farmers attributed this decline to a number of

factors including droughts, poor soils, lack of extension services and removal of government

subsidies that were key in providing adequate fertilizer and modern farm implements. On the

other hand, maize production in the vinyungu has increased from an average of 23 bags/ha in the

1970s/80s to an average of 33 bags/ha in the 1990s/00s (Table 3-2).

The most common food crops grown on vinyungu include maize, potatoes, beans, and

cassava. Maize, potatoes, beans, and vegetables such as peas, tomatoes, onions, and green pepper

are important sources of income. Farmers reported that, in most cases, vinyungu are cultivated at

least twice a year, mostly during the dry season when market prices are also higher hence

maximizing profit (Table 3-3). Maize, potatoes, beans, and some vegetables are sown around

August/ September and harvested in November/ December (potatoes, beans, and vegetables) and

later around January/February (maize and vegetables). The second round of planting takes place

in March/April and harvesting takes place in June/ July. Vegetables may be sown and harvested

thrice a year, depending on water availability. By juggling upland and bottomland cultivation,

farmers are able to produce enough food year-round as well as increase family income.

Table 3-1 shows that there are differences in the use of vinyungu plots with respect to the

types of crops grown. A larger proportion of the farmers (83.4% vs. 52.0%) use at least half the

size of their vinyungu plots for growing food crops compared to cash crops.









3.3.1.3 Inputs on vinzyungu

Data from Table 3-1 showed that on average, farmers used Tshs 22,245113,103

(approximately U.S. $22113) per bag of inorganic fertilizer per growing season per household.

At least 87% of the farmers interviewed use inorganic fertilizer despite the high costs and limited

availability. Most farmers buy between 1 and 3 bags per season for both upland and lowland

farms and use about half to 1 bag of inorganic fertilizer in their vinyungu. The individual farmers

that were interviewed attributed the use of fertilizer on vinyungu plots in the recent years (1990s

- 2005) to a decline in soil fertility.

While several sources of labor exist (e.g., hired labor, neighbors, friends), farmers reported

family members (i.e., husband, wife, children, dependents) as the most commonly used labor.

Neighbors or friends may be requested to help in cases where family labor is not adequate to

perform the task. The other alternative is hired labor. Data from Table 3-1 show that the mean

cost of hired labor is Tshs 9,983 & 3,975 (U. S. $10+4) per acre per household. Although most

farmers consider this rate high, at least 52% hire labor for the most laborious task of land

preparation and rely on family labor (or friends) for other farm activities such as planting,

weeding and harvesting.

Tables 4-7 show the results of Part II of the bivariate analysis that explored the association

between vinyungu farming practice and factors of influence that included sex, age, years of

residence in the village, and number of potential workers (aged 18 to 60 years) in the household.

Vinyungu farming was assessed as (i) size of vinyungu plots owned by household, (ii) size of

vinyungu plots grown in the dry season, (iii) number of cash crops grown on vinyungu plots, and

(iv) size of vinyungu plots grown cash crops.

The results in Table 3-4 show that vinyungu-farming practice is associated with the number

of working household members and age of the farmer. Specifically, the data indicated that









number of household members who can work is associated with the size of vinyungu. Almost all

households with more than three working members had large farming lands and, in addition,

over 54.8% such households had plots covering over an acre (> 0.4 ha) of land compared to

45.2% of those households with at most three members. Also, farmers in the 36-to-60-year-old

age group were more likely to have large vinyungu plots compared to other age groups.

The results in Table 3-7 showed that age is associated with the size of vinyungu plots

cultivated with cash crops (X2 = 4.432; p = 0.039). Here, farmers belonging to all age groups tend

to use less than 50% of their plots to grow cash crops.

None of the selected factors of influence was related to either size of vinyungu plots during

the dry season (Table 3-5) or number of cash crops grown (Table 3-6) at the bivariate level.

Overall, there was no difference between men and women in vinyungu-farming practices. That

is, men and women had similar patterns of land-use for vinyungu farming system in the study

area at this level of analysis.

3.3.1.4 Multivariable analysis

In this part of analysis a multivariable regression model was used to assess factors

affecting vinyungu-farming system. For this purpose, a general linear multivariable regression

model was fitted to the data collected from maj or vinyungu cultivators. Four variables, overall

size of vinyungu plots, size of vinyungu plots during the dry season, number of cash crops, and

size of vinyungu plots grown cash crops were selected as dependent variables describing

vinyungu-farming system. Four models, one for each dependent variable, were fitted to

determine how farmer and household characteristics influenced the dependent variables. The

independent variables included sex, distance from plot to maj or market, distance from plot to

maj or road, distance from house to plot, cost of hired labor, cost of fertilizer, number of workers

in the household and years lived in the village. Number of cash crops (described above as










dependent variable) was also used as independent variable in the model for overall size of

vinyungu and size of vinyungu plots in dry season. The problem of multicollinearity, i.e., high

correlations among the variables used in the model, was addressed prior to fitting of each of the

four models to avoid biased regression estimates.

The results of the linear regression models are presented in Table 3-8. Results from Model

I (overall size of vinyungu) show that the constant term is positive and significant. This means

regardless of the independent variables (or predictors) I have selected for this model, a farmer is

expected to have vinyungu plots with an average size of 1.558 ha. The model also showed that

the number of working persons in the household is related to the overall size of vinyungu plots

the household owns. Specifically, the data indicate that for each additional person of working age

group in a household, the size of vinyungu plot increases by 0.76 ha (p=0.029).

The relationship between distance from vinyungu plots to both maj or market and main road

appeared to have the hypothesized direction, but did not attain statistical significance. Also, sex,

costs of labor and fertilizer, duration of residence in the village and number of cash crops were

not related to the overall size of vinyungu plots owned by the farmers. Lastly, the R-Square for

model I was 0.342. This indicates that 34.2% of the variation in the size of vinyungu plots can be

attributed to the variations in the predictors selected to fit this model.

Model II was fitted to assess factors influencing size of vinyungu plots during the dry

season. My results revealed that the number of persons in a working group per household as well

as duration of residence in the village were associated with the size of vinyungu plots grown

during the dry season. On average, each additional person in the working group increases the

size of plots by 0.25 1 ha per household (p=0.024). Combined with the results of Model I, these

results support hypothesis 1, i.e., greater number of persons in a working age group in a









household is associated with greater size of the cultivated vinyungu plots. However, hypotheses 6

and 7 (i.e., residing close to the main roads and major market respectively, would be associated

with greater proportion of wetland converted to agriculture) are not supported. The results show

that residents with fewer years of residence in the villages are more likely to have larger sizes of

vinyungu plots in the dry season.

Model III assesses predictors of the number of cash crops grown in the vinyungu plots. I

found that distance from the farmer' s house to the plot is associated with the number of cash

crops grown. On the average, each additional kilometer of the distance between the house and

the plot is associated with a reduction of about 0.2 (or one out of five) possible cash crops that

can be grown in the vinyungu plots. In this model, the selected predictors explained only 28% of

the variation in the number of cash crops grown in the plots. That is, the data had the poorest fit

on Model III.

Lastly, Model IV was used to analyze factors of the proportion of vinyungu plots for

growing cash crops. The constant term indicates that when all predictors have no effect in the

model (i.e., when all coefficients are zero), we would expect that, on average, farmers would

have 61% of their vinyungu plots grown cash crops. Also, results show that sex and duration of

residence in the village were associated with the size of plots used to grow cash crops. Men had

16.5% more of their vinyungu plots grown cash crops compared to women. This finding

indicates that men, who were previously not involved in vinyungu cultivation except where hired

by women to prepare land for cultivation (Lema, 1996; Masija, 1993), are now increasingly

getting involved in this farming system and that they have interest in expanding their plots to

grow cash crops. Similar to the results of Model II, I found in Model IV that duration of

residence was associated to the size of vinyungu plots for growing cash crops. Those residents









with fewer years of residence in the villages were more likely to have larger proportions of their

vinyungu plots used to grow cash crops. All distance factors had the expected direction of

relationship with the size of vinyungu for growing cash crops, but were not statistically

significant.

Lastly, Table 3-8 shows that age, years of residence in the village, use of fertilizer have no

effect on the household's engagement in vinyungu-farming practice.

3.3.2 Environmental Issues Related to Vinzyungu Cultivation

This section presents farmer's perceptions on different environmental issues as discussed

during the village meetings (Table 3-9) and through individual farmers interviews (Table 3-10).

Descriptive statistics have been used to analyze the data.

Table 3-9 shows that land availability is not a problem in the area. Instead, farmer's

decision to cultivate vinyungu is driven by the availability of water (75%) and soil fertility

(20%). Figures 4 and 5 show the different sources of water used to make vinyungu cultivation

possible.

Data in Table 3-9 also show that the use of the wetlands for vinyungu cultivation has

increased when compared to the 1970s (97.5% of the responses). This is further supported by

data obtained through individual farmer' s interview whereby 63% of the farmers reported a

tremendous increase in the use of wetlands for vinyungu cultivation (Table 3-10). While wetland

size and soil fertility were reported to be declining, soil erosion was reported to be increasing as

a result of the expansion of vinyungu cultivation over the years. Farmers perceive reduced water

flow as the biggest problem resulting from the cultivation of vinyungu, followed by reduced soil

fertility, increased soil erosion, and drying of the wetland (Tables 9 and 10). At least 68% of the

responses associated soil erosion with the way vinyungu are cultivated, i.e., close to riverbanks

and catchment areas as well as massive clearance of wetland vegetation such as Cyperus papyrus










(madete) and some important catchment vegetation such as Syzygium cordatum (mivengi).

Through village discussions and individual farmers interviews, it was apparent that availability

of water and soil fertility are influencing farmers to clear more wetlands to create new vinyungu.

At least 93% of the respondents reported an increased number of households in the study

area. This increase may also be playing a part in influencing increased conversion of wetlands to

agriculture. Figure 3-3 shows Ndembera river only 1.5 meters behind the wetland vegetation,

farmers are cultivating vinyungu and they are expanding their vinyungu towards the river.

Farmers also reported that human conflict over use of wetland resources (land, water) was not

observed in the 1970s but is now growing. Although education on sustainable practices has been

provided and land use plans developed in many villages, farmers reported this to be a very recent

(2004/5) undertaking.

3.3.3 Discussion

This chapter examined the current trends in vinyungu-farming system practices and

explored the different farmer and household factors that may be influencing farmers to change

their traditional farming system as well as the implications of the new developments on wetland

resources.

Most farmers in the study area would have both vinyungu and upland plots for their

farming activities. However, the size of upland plots owned by farmers tends to be larger than

that of vinyungu. Land is acquired mostly through inheritance and is usually in close proximity to

the villages where farmers live. Therefore, members of households located within a short

distance from the wetlands do cultivate vinyungu. Although a few farmers purchased their

farmland, this trend is not common practice.

Farmers' decision to cultivate vinyungu and increase the size of their plots seems to be

partly driven by the availability of water, soil fertility and increased production. Due to declining









soil fertility in the vinyungu, however, most farmers (87.5%) use inorganic fertilizer. This

supports Hypothesis 3, i.e., Farmers are likely to use household income on farm inputs (i.e.,

fertilizer and labor) so as to improve crop production.

Family members, irrespective of sex, are the main source of labor in the construction of

vinyungu. The results in this section support Hypothesis 1. It has been observed that the greater

the number of people within a working age group in a household, the greater the size of vinyungu

the household is likely to own. Hired labor is considered expensive and is therefore rarely used

except for a tough task of land clearing.

The negative regression coefficients on distance factors values indicate that close

proximity to both markets and roads is likely to influence an increase in both the size of

vinyungu and the proportion of vinyungu cultivated cash crops (Hypotheses 6,7, and 8). Although

these relationships were not statistically significant, they give an indication of the validity of the

hypothesized relationships.

Sex, costs of labor and fertilizer, duration of residence in the village and number of cash

crops were not related to the overall size of vinyungu plots owned by the farmers as was

postulated in Hypothesis 2, i.e., farmer' s sex or duration of residence in the village is not

associated with the size of vinyungu farming area or types of crops grown.

Sex, age, years of residence, and number of people that can work did not seem to influence

the size of vinyungu plots cultivated during the dry season or number of cash crops grown, at the

bivariate level. However, at the multivariate level, the number of persons in a working age group

per household and duration of residence in the village were associated with the size of vinyungu

plots grown during the dry season. That is, households with a greater number of people within a

working age group are more likely to have more area in vinyungu during the dry season










(p=0.024) and residents with fewer years of residence in the villages are more likely to have

larger sizes of vinyungu plots in the dry season (p=0.048). An important implication of this

finding could be the movement of farmers coming to settle into the wetland areas. These farmers,

who are likely to migrate from more arid areas, seem to have tendency to cultivate larger

vinyungu plots during the dry season. These farmers tend to buy the land they cultivate and a few

rent.

There was no consistent difference between men and women in vinyungu-farming

practices. That is, men and women had similar patterns of land-use for vinyungu-farming system.

The only gender difference was observed in the proportion of land used to grow cash crops. Men

had a higher percentage of their land used to grow cash crops. This is an indication of a new

tradition in vinyungu i.e., having men more engaged with vinyungu farming and more

specifically in cultivation of cash crops.

Distance from the farmer' s house to the plot is inversely related to the number of cash

crops grown, i.e., the further the plot is from the house the less the cash crops that can be grown

in the vinyungu.

I found out that age, duration of residence, and use of fertilizer were not consistent factors

of vinyungu farming. With regards to age, I can explain that no differences occur in various

aspects of vinyungu farming because persons of all ages tend to be farmers once they reach

young adulthood. In this way, no specific age group may appear to have unique influence on

vinyungu farming.

The effect of duration of residence appeared to be limited to the growth of cash crops in

the wetlands. Thus, its effect was not detected in other areas, e.g., overall vinyungu area and










percentage of area cultivated in the dry season. This implies migration to the wetlands is mainly

influenced by cash crop production.

Use of fertilizer appeared to have no influence on vinyungu-farming practice probably

because all farmers are relatively at the same level of production, i.e., there is no difference in

resources that enable farmers to acquire farm inputs.

Increased soil erosion and reduced water flow, soil fertility and wetland size/ vegetation

are associated with both the expansion and methods of vinyungu cultivation. These results partly

support Hypothesis 5 which suggested that greater expansion of vinyungu farms would be

associated with increased conversion of wetland area to agricultural land. Parallel to increasing

environmental problems is growing human conflict over use of wetland resources, as was

reported by farmers.










Table 3 -1. Sociodemographic and Land Use Characteristics of the Sample Farmers in the study
area (n=54)
Variable Number Percent
Sex
Male 34 63.0
Female 20 37.0
Age
18 35 years 21 39.0
36 60 years 26 48.0
61+ years 7 13.0
Marital Status
Married 49 90.7
Education
Primary education 41 76.0
Primary and training 8 14.8
Secondary education 5 9.2
Ethnic Distribution
Bena 12 22.2
Hehe 39 72.2
Jita 1 1.9
Kinga 1 1.9
Wawanji 1 1.9
Age distribution by household
>18 years old (median = 3)
None 7 13
1-2 people 7 13
3-4 people 23 42.6
5-6 people 13 24.1
7+ people 4 7.4
18-60 years old (median = 3)
None 3 5.6
1-2 people 22 40.8
3-4 people 27 50.0
5-6 people 1 1.9
>6 people 1 1.9
>60 years old (median = 0)
None 47 87.0
1 person 5 9.3
2 people 2 3.7
Main Occupation
Agriculture 52 96.3
Small business (e.g., restaurant, local brew) 2 3.7
Cooperative Membership
Current member 15 27.8










Table 3-1 Continued
Variable
Land Ownership
Own land
Land Acquisition
Bought
Inherited
Given by the village government
Land Type Owned and Used
Both vinyungu and upland areas
Size of Upland Plots
Less than an acre (<0.4 ha)
1 2 acres (0.4 0.8 ha)
3 4 acres (1.2 1.6 ha)
Over 4 acres (>1.6 Ha)
Size of Vinyungu Plots
Less than an acre (<0.4 ha)
1 2 acres (0.4 0.8 ha)
3 4 acres (1.2 1.6 ha)
Over 4 acres (>1.6 Ha)
Proportion of Vinyungu Area for Food Crops
Less than 50%
About 50%
Over 50%
Proportion of Vinyungu Area for Cash Crops
Less than 50%
About 50%
Over 50%
Cost of fertilizer/bag/growing season/HH (87%
users ) Mean+ SD
Cost of labor/acre/HH
(52% hire labor) Mean+ SD


Number

53

8
43
3

52

2
15
20
17


Percent

98.1

14.8
79.6
5.6

96.3

3.7
27.8
37.0
31.5


50.0
38.9
3.7
7.4

16.6
39.0
44.4


26 48.0
21 39.0
7 13.0
TShs. 22,245113,103
($22 & 13)
TShs. 9,98313,975 ($1014)


Important cash crops include maize, potatoes, beans and vegetables such as tomatoes, onions and
peas. Fertlizer commonly used are Urea and CAN (top dressing) and DAP and TSP (during
planting). In this study, nutrient amounts were not calculated.










Table 3-2. Maize yields in uplands and vinyungu between 1970s and 2000s (village responses,
n=9) (per year, i.e., one season in the uplands and two seasons in the vinyungu)
Maize yields in the uplands Maize yields in the vinpungu

1970s /80s 1990s/00s 1970s/80s 1990s/00s

Bags/ha Mean+SD 30+6 15+7.8 23+11.5 33+9.3

# of bags/ha <10 bags 0 response 3 responses 0 response 0 response

farmers 10-19 bags 0 response 3 responses 6 responses 0 response

produced 20-29 bags 4 responses 3 responses 3 responses 0 response

between 30-39 bags 5 responses 0 response 0 response 7 responses

1970-2005 >40 bags 0 response 0 response 0 response 2 responses

Farmers expressed yields in bags/acre. Their estimates have been converted to bags/ha (1 bag=100kg when filled
with maize grains)


Table 3-3. Potential market price per unit output from cash crops grown in vinyunguand under
rain fed agriculture


Crop




Maize (bag ')

Shelled and dried

Beans (bag ')

Shelled and dried

Potatoes (bag ')

Peas (bag ')

In pods, not dried

Onion (bag ')

Tomatoes (tenga')

Cabbage (bag-')


Growing season (No.
of seasons/ year)

Rain fed Vinyungu

1 2


Average price
tenga)

Rain fed

28,000



54,000



25,000

20,000



40,000

3,000

4,000


(TShs/bag or


Vinyungu

36,000



54,000



30,000

15,000



50,000

5,000

5,000


Average income
(TShs/year)

Rain fed Vinyungu

28,000 72,000



54,000 108,000


25,000

20,000



40,000

3,000


60,000

30,000



100,000

15,000


4,000 10,000


1 bag = 100 kg when filled with dry maize grain;
grain; Tanzanian Shilling (TShs) 1,000 = $ 1


1 tenga = a tin filled with 16 kg of dry maize










Table 3-4. Results of Chi-Square Analysis of the Farmer and Household Factors influencing
Vinyungu-Farming Practice, i.e., size of vinyungu plots owned by household in
relation to sex, age, years of residence in the village, and number of potential workers
Factors of Influence Vinyungu size Vinyungu size X2 P
< 1 acre (< 0.4 ha) >1 acre (> 0.4 ha)


Female
Male


10 (43.5)
13 (56.5)

13 (56.5)
9 (39.1)
1 (4.3)

9 (39.1)
14 (60.9)

22 (95.7)
1 (4.3)


10 (32.3)
21 (67.7)

8 (25.8)
17 (54.8)
6 (19.4)

12 (38.7)
19 (61.3)

14 (45.2)
17 (54.8)


0.713 0.287



6.174 0. 046


0.001 0.598


10.963 0. 001


18-35 years
36-60 years
61 years or older
Years of residence
< 25 years
> 25 years
Number of workers
5 3 workers
> 3 workers


Table 3-5. Results of Chi-Square Analysis of the Farmer and Household Factors influencing
Vinyungu-Farming Practice, i.e., size of vinyungu plots grown in the dry season in
relation to sex, age, years of residence in the village, and number of potential workers


Factors of influence

Sex
Female
Male
Age
18-35 years
36-60 years
61 years or older
Years of residence
< 25 years
> 25 years
Number of workers
< 3 workers
> 3 workers


Vmnyungu size dry season
< 2 acres (< 0.8 ha)

17 (36.2)
30 (63.8)

18 (38.3)
24 (51.1)
5 (10.6)

17 (36.2)
30 (63.8)

34 (72.3)
13 (27.7)


Vinyungu size dry season
> 2 acres (> 0.8 ha)

3 (42.9)
4 (57.1)

3 (42.9)
2 (28.6)
2 (28.6)

4 (57.1)
3 (42.9)

5 (71.4)
2 (28.6)


0.117 0.519



2.184 0.336


1.128 0.256


0.003 0.637










Table 3-6. Results of Chi-Square Analysis of the Farmer and Household Factors influencing
Vinyungu-Farming Practice, i.e., number of cash crops grown on vinyungu plots in
relation to sex, age, years of residence in the village, and number of potential workers
Factors of influence No. of cash crops No. of cash crops X2 p
<3 >3


Female
Male


10 (30.3)
23 (69.7)

15 (45.5)
14 (42.4)
4 (12.1)

13 (39.4)
20 (60.6)

24 (72.7)
9 (27.3)


10 (47.6)
11 (52.4)

6 (28.6)
12 (57.1)
3 (14.3)

8 (38.1)
13 (61.9)

15 (71.4)
6 (28.6)


1.650 0.160



1.564 0.457


0.009 0.577


0.011 0.578


18-35 years
36-60 years
61 years or older
Years of residence
<25 years
> 25 years
Number of workers
5 3 workers
> 3 workers


Table 3-7. Results of Chi-Square Analysis of the Farmer and Household Factors influencing
Vinyungu-Farming Practice, i.e., size of vinyungu plots grown cash crops in relation
to sex, age, years of residence in the village, and number of potential workers
Factors of influence Vinystngst area planted Vinystngst area planted
cash crops cash crops X2 p
< 50% > 50%


Sex
Female
Male
Age
18-35 years
36-60 years
61 years or older
Years of residence
< 25 years
> 25 years
Number of workers
< 3 workers
> 3 workers


18 (37.5)
30 (62.5)

21 (43.8)
21 (43.8)
6 (12.5)

18 (37.5)
30 (62.5)

35 (72.9)
13 (27.1)


2 (33.3)
4 (66.7)

0 (0.0)
5 (83.30)
1 (16.7)

3 (50.0)
3 (50.0)

4 (66.7)
2 (33.7)


0.040 0.609



4.432 0. 039


0.351 0.431


0.104 0.539












Table 3-8. Results of the Linear Regression Analysis of the Factors Affecting Vinyungu-Farming Practice
Variable Size of Vinyungu Size of Vinyungu in Number of Cash Size of Vinyungu
Plot (I) Dry Season (II) Crops (III) for Cash Crops (IV)


Coefficient p
1.558 0. 016
0.218 0.563
-0.002 0.878
-0.010 0.664
0.001 0.979
-0.001 0.564
-0.001 0.396
0.760 0. 029
0.002 0.844
0.029 0.944


Coefficient P
61.013 0. 001
16.492 0. 031
-0.032 0.886
-0.254 0.547
-0.058 0.953
0.000 0.722
0.000 0.088
-1.768 0.778
-0.705 0. 010
NIa NIa
0.383


Coefficient
0.231
-0.001
0.002
0.000
NIa
0.000
0.000
0.251
-0.008
-0.031


P
0.248
0.993
0.532
0.987
NIa
0.646
0.876
0. 024
0. 048
0.355


Coefficient
4.369
-0.478
0.003
0.067
-0.219
0.000
0.000
0.169
-0.036
NIa
0.284


p
0. 007
0.505
0.878
0.136
0. 031
0.776
0.765
0.565
0.155
NIa


Constant
Sex
Distance to maj or market (km)
Distance from plot to main road (km)
Distance from house to plot (km)
Labor cost ($)
Cost of fertilizer ($)
No. of working household members (#)
No. of years lived in the village (#)
No. of cash crops (#)


R-Square 0.342
a Indicates variable "Not Included" in the model due to multicollinearity


0.353




















































V


Table 3-9. Farmer' s perception on various environmental issues as discussed in the village
meetings: the current state compared to the 1970s (n = 40)


Variable
Use of the wetland
Has increased
Wetland/wetland vegetation cover
Decreased/declined
Soil fertility
Soil fertility has increased
Soil fertility has declined
Soil erosion
No erosion
Some erosion
Great erosion
Motivation to cultivate vinyungu in their area
Lack of alternative land
Rich soil fertility
Availability of water
Vinyungu cause environmental problems
Yes
The biggest problem associated with vinyungu cultivation
Reduced water flow
Reduced soil fertility
Soil erosion
Human conflicts
Management plans for wetland resources
Exi st
Do not exist
Education on sustainable practices
Has been provided
Has not been provided
Suggestion to the Government on the future of vinyungu cultivation
Vinyungu cultivation be halted/ banned
Sustainable/ best management practices be promoted


Responses Percent

39 97.5

36 90


12.5
82.5

12.5
67.5
15

5
20
75


37 92.5


57.5
17.5
20
5

62.5
37.5

70
30

15
85
































Figure 3-1. A well is dug where water is not free flowing. Picture taken in Lumuli village in
Iringa District, June 2005


Figure 3-2. A river is diverted to supply vinyungu with water. Picture taken in Usengelindete
village, Iringa District, June 2005












Vinyungu cultivation taking
place just 1.5 meters behind the
wetland vegetation. Clearing
of vegetation was ongoing at
the time this photo was taken
SWetland vegetation




- Ndembera river


Wetland
vegetation


Figure 3-3. Picture showing Ndembera river behind which vinyungu cultivation is taking place in
Maduma village, Mufindi District. An example of how close to the river the vinyungu
are getting. Picture taken in June 2005










Table 3-10. Individual farmer' s perception on various environmental issues (n=54)
Variable Responses Percent
Upland plots for crop production
Fairly productive 37 68.5
Not very productive 17 31.5
Wetland plots for crop production
Very productive 28 51.9
Fairly productive 26 48.1
Factors That influenced farmers to own or rent vinyungu
They are cheap, easy to get 4 7.4
Close proximity from home 4 7.4
High soil fertility 30 55.6
Excellent dry season alternative (availability of water) 16 29.6
Trend in the utilization of wetlands for vinyungu cultivation
Use has increased tremendously 34 63.0
Use has increased slightly 15 27.8
Use has remained the same 1 1.9
Use has declined 4 7.4
Trend in the number of households
Number of households has increased 50 92.6
Number of households has decreased 1 1.9
Number of households has remained the same 3 5.6
Environmental problems associated with vinyungu cultivation
Soil erosion 4 7.4
Sedimentation 4 7.4
Reduced water flow 26 48.1
Drying of the wetland 6 11.1
Reduced wetland size 4 7.4
Flooding 4 7.4
Erosion of river banks 4 7.4
None 2 3.7









CHAPTER 4
USE OF GEOGRAPHIC INFORMATION SYSTEMS AND REMOTE SENSING TO
DETERMINE THE MAGNITUDE AND RATE OF CONVERSION OF NDEMBERA
SWAMP AS A RESULT OF VINYUNGU EXPANSION

In the previous section of the thesis I explored changes in vinyungu cultivation, factors

influencing the changes, and environmental implications as perceived by interviewed farmers.

According to the farmers, vinyungu cultivation has increased over time. However, no study has

been conducted to determine the extent and rate of wetlands conversion to agriculture.

Geographic Information Systems (GIS) and Remote Sensing technologies are useful

analytical tools for obtaining information on land-use/cover change (LUCC), including where

and when LUCC occurs and the rates at which they occur (Lambin et al., 1999; Turner et al.,

1993). GIS are a computer-based systems of retrieving, storing, manipulating, updating and

mapping spatially referenced data (Jones 1997). Remote Sensing is a method of collecting spatial

data using remote sensors (i.e., not in direct contact with the target of interest) such as satellites

and aerial photography (Jensen 1996). Remote sensing and GIS are often integrated and used to

analyze ecosystems on multiple scales with both spatial and temporal factors. These technologies

have been used to map and document changes in wetland use and cover (Burgi and Turner, 2002;

Jensen et al., 1995; Munyati, 2000; Wang et al., 2006). Such studies have also been able to

utilize these technologies to assess further the factors influencing the observed changes. Since

field studies can be time consuming, expensive and cover only a small area, remote sensing

offers a potentially cost-effective way to study ecological changes over time.

I used remote sensing and GIS to determine the land-use and land-cover changes

associated with agricultural activities around Ndembera swamp. My obj ectives were 1) to

determine the magnitude of wetland change between 1977 and -1999; and, 2) to determine










whether or not there exists a relationship between agricultural land expansion and a decline in

wetland area.

4.1 Data Sets

Two datasets from two different periods (1977 and 1999) were used to analyze land-use/

cover changes. Both datasets are based on un-rectified aerial photographs at the scale of

1:50,000.

4.2 Methods

Photographs were interpreted, using a Topcon Stereoscope. Different land cover types

were delineated on the set of un-rectified aerial photographs. Each land cover was symbolized at

its location as one land use class or polygon, e.g., "Ag" represented "agricultural land".

Graphical radial triangulation method (Schwidefsky, 1959) was used to control the horizontal

scale (vertical scale was controlled during the digitization process). Control points of common

areas were marked on all maps including the 1:50,000 topographic maps (i.e., the base control

map).

Each photograph had a point that was stereoscopically located. By using these stereo

located points, differences of scale between the first photograph to the second, third, fourth and

so on were controlled. Transparent overlays were used to transfer the interpreted land covers

from aerial photographs to the paper. The pieces of transparent paper were then j oined by

orientation technique (pass pointed) from the preliminary map. Tracing paper was laid on these

transparent overlays and traced as first base map to be relayed on as the base map with all

details. This was done for both 1977 and 1999.

The data on the tracing papers were digitized to produce land-use maps of the study area

for 1977 and 1999. Visual interpretation was preferred given the relatively small size of the study

area. The resulting interpretation was also digitized using Arc INFOC 3.5.1. To rectify the aerial










photographs, features such as road junctions and river confluence points were identified both on

the two sets on one hand, and geo-referenced topographic maps (sheet no. 232/1 and 232/2) on

the other hand. All photos were geometrically rectified and registered to a common UTM

(Universal Transverse Mercator) proj section based on 1:50,000 scale topographic maps of

Tanzania (i.e., UTM zone 36 south, datum Arc 1960). The identified points were distributed

throughout the study area. Using Arc INFO 3.5.1, the points were used to transform (proj ect) the

interpretation from aerial photographs.

4.3 Data Analysis

The coverage or layers produced in ArcE\NFO were used to produce land use/ cover maps

of the two sets in ArcView 3.2. Data analysis to produce the change detection matrix was done

using ArcView and Microsoft Excel. ArcView shape Hiles were exported (as dbf Eiles) to

Microsoft Excel where the pivot table function was used to produce the change detection matrix.

Quantitative data for the overall land use changes and gains and losses in each category were

compiled (Table 4-2 and Table 4-3). The change matrix provides information on the main types

of changes in the study area.

4.4 Results

Land use and land cover maps of 1977 and 1999 are displayed in Figure 4-1 and 4-2. The

overall land use changes from 1977 to 1999 are shown in Table 4-2. According to the table,

agricultural land was the largest land-use type, both in 1977 and 1999. Agricultural land took up

65.3 and 46.0%, respectively, of the total area. As per Table 4-2, densely populated areas as well

as areas of grassland and woodland increased from 1977 to 1999. On the contrary, agricultural

land and open valleys declined.

In 1977, agricultural land covered about 65% of the study area and with an estimated area

of 589.4 ha, but in 1999, the total agricultural area was estimated to have decreased by about










29.4% to 416 ha. The annual rate of decrease is estimated at 1.3% during 1977-1999. Valley

bottoms covered about 17.4% of the study area and with an estimated area of 157.4 ha, but in

1999, the total valley bottom area was estimated to have declined by about 18.1% to 128.9 ha.

The annual rate of decrease is approximately 1% during 1977-1999. Meanwhile, densely

populated areas, grassland and woodland increased by about 138, 308, and 79% respectively

with an annual growth rate of 6, 13 and 3% during 1977-1999.

Results from the transition matrix in Table 4-3 indicated the area increase or decline of

each land use type. It is clear that between 1977 and 1999, the transition replacement rates of

agricultural land, populated area, grassland and woodland were high at 42.6, 44.5, 43.4, and

49.6% respectively. That of valley bottoms was the lowest with a transition rate of 25.1%.

Between 1977 and 1999, about 17% of valley bottoms were transformed to agricultural land and

about 16% were transformed to grassland.

4.5 Discussion

Remote sensing and GIS technologies were used to determine the land-use/cover changes

associated with agricultural activities around the Ndembera swamp. Specifically, I evaluated the

magnitude of wetland cover change that took place between 1977 and 1999 around Ndembera

wetland.

Ndembera wetland has been experiencing various changes over the years. Data on Table 4-

2 indicate that densely populated areas, grassland, and woodland areas increased between 1977

and 1999 while agricultural land and wetland areas decreased during that same period. Although

the agricultural land decreased, it remains the largest land use type in both 1977 and 1999.

Table 4-2 shows changes of the wetland area between 1977 and 1999. Table 4-3 shows

that the total wetland area declined by about 18% between 1977 and 1999 and that at least 17%

of the wetland area was transformed to agriculture. This finding supports farmers' perceptions









observed in Chapter 3 (Table 3-10 ) that wetland utilization for vinyungu cultivation has

increased. At least 97% of the farmers interviewed reported this increase. In addition, 90% of the

farmers that were interviewed believe the Ndembera wetland has decreased in size and that the

rivers that drain into the wetland have dried the outcomes they attributed to increased

cultivation of vinyungu. Farmers attributed the increased utilization of vinyungu mainly to a

decline in crop production from the uplands that is caused by increased soil infertility and

droughts (Tables 9 and 10). The decline in crop production in the upland areas was further

compounded by the structural adjustment program that started in the mid 1980s that required the

government to stop subsidizing farmers for fertilizer and farm tools. Availability of water and

fertile soils in the wetlands therefore, attracted farmers as a way to increase their cultivation of

vinyungu. Today, vinyungu are not only attractive to farmers but also to the government. During

group discussions, farmers reported they were directed by the government to cultivate vinyungu

between 1992 and 1993 to overcome food shortage that was caused by prolonged droughts.

Table 3-3 in Chapter 3 clearly indicates an increased income from vinyungu cultivation

compared to rain fed agriculture. Farmers associated the increase in income with trade

liberalization that occurred in the 1990s that allowed for improved marketing of goods both in

and outside the country. The increase in income is likely to influence further conversion of

wetlands to agriculture.

Table 4-3 shows that human settlement in villages surrounding the Ndembera wetland has

increased over time by 137%. This increase in human settlement may be contributing to

increased conversion of wetlands to agriculture.

The percentage change of the wetland to agriculture seems small. However, recurrent

prolonged droughts and market forces that seem to contribute in driving vinyungu expansion









started only recently, meaning, the observed land cover changes took place within a relatively

short period of time. For example, during the group discussions, farmers reported that the most

popular crops, i.e., peas, tomatoes, onions, and potatoes started to be cultivated mostly in the

1990s. Paprika (that is becoming increasingly important to farmers in terms of financial returns)

started to be cultivated around 2004 and is now finding its way to vinyungu plots.

Based on these findings, it is evident that wetlands are potential agricultural lands, that

unexploited wetlands are likely to be continuously cleared as long as farmers have access to

them, and that markets for goods obtained from vinyungu remain viable. These findings also

confirm the hypothesis that expansion of vinyungu farming is decreasing the size of wetlands.

About 25 ha (or 16%) of the wetland area has been transformed to grassland. The grassland

in this area, as was observed during field survey, consists of short, dry (yellow) grass that

remains so until during the rainy months (March to April) when the area may be inundated and

the grass turns green. This is to say that the wetland vegetation has been replaced by vegetation

that can withstand prolonged dry spells. This change may be caused by reduced water flows into

the wetland that was reported by farmers (Table 3-9). It may also be due to accumulation of soils

resulting from soil erosion further upstream that was reported by farmers during the interview

(Table 3-9). Soil erosion upstream is known to cause sedimentation downstream that may be

colonized by vegetation.

Table 4-3 also shows that about 23% of the agricultural land changed to a woodland area.

Farmers were asked during the group discussions the reasons for the agricultural land to be

replaced by the woodlands. Farmers attributed this change to a ban made by the Government in

the mid 1980s to stop farmers from cultivating finger millet in the woodlands, allowing a

recolonization of woodlands in the previously cultivated areas.










About 10% of the agricultural land is transformed to a grassland area between 1977 and

1999. Again, the grassland consists of short dry grass. These areas were observed as abandoned

agricultural land.

Other changes include the transformation of woodlands to agricultural land and to

grassland. Farmers attributed these changes to encroachment into woodlands where woodlands

are cleared for cultivation purposes or to obtain wood for home consumption.

About 33% of the grassland areas changed to agricultural land between 1977 and 1999.

Farmers reported these areas to be wetter than they used to be, probably due to change in the

configuration of rivers that drain in to the Ndembera wetland. Availability of water makes them

attractive for cultivation.










Table 4-1. Map History (i.e., Data used)
S/N Date of Photograph Film No.

1 July 1977 (dry season) 1828

1844


Scales

1:50,000







1:50,000


Exposures

111-114

137-142

144-151

106-113

2998-3004

2587-2595

2661-2670


1809

1809

RUN 15

RUN 16

RUN 17


2 September 1999 (dry
season)
















LEGEND


LAND USE/COVER
Woodland
Grassland
Wetland Which can Flood
Agricultural Land
Densely Populated Area


n Study Village

a Other Village

IVAll Weather Road

/VRiver


Figure 4-1. Ndembera swamp in 1977 (A). The inset map (B) shows Ndembera swamp location in Iringa region.
















LEGEND


LAND USE/C OVER
Woodland
Grassland
Wetland Which can Flood
Agricultural Land
Densely Populated Area


o Study Village

a Other Village

(VAll Weather Road

/VRiver


Figure 4-2. Ndembera swamp in 1999 (A). The inset map (B) shows tha location of Lyandembera swamp in Iringa region


IRINGA REGION












Legend
8 CHANGES IN WETLAND 1977 AND 1999 Wtadi 97ad19
4 0 4 8 12 16 20 Kilometers O
D Wetland In 1977 only
g ~Wetland in 1999 only
~I II I ~ ~All Weather Road
Dry Weather Road
'~ Motorable Tracke Road
Footpath

-1 ..t~/ Ruder~ilg


...i ~~ ,. ex Other Village


Figure 4-3. Land use/ cover change map (1977 and 1999) (A). The inset map (B) shows the location of Lyandembera swamp in Iringa











Table 4-2. Total area (ha) and area of change of land use types from 1977 to 1999
Total Area (ha) Total Area (%) 1977 1999
(change)


Land use
Type
Densely Populated Areas
Agricultural Land
Grassland
Wetland
Woodland
Estimated total


1977 1999 1977 1999 (ha)
16.4 39.1 1.8 4.3 22.6
589.4 416.0 65.3 46.0 -173.5
29.8 121.8 3.3 13.5 91.9
157.4 128.9 17.4 14.3 -28.5
110.2 197.6 12.2 21.9 87.4
903.3 903.3 100.0 100.0


%/yr
6.0
-1.3
13.4
-0.8
3.5


137.7
-29.4
308.3
-18.1
79.27


"Wetland" is synonymous to "Open valley seasonally flooded"





Table 4-3. Transition matrix of land use types from 1977 to 1999 (transition
probabilities in %)
Land
use type 1999


Densely


Agricultural


populated areas Land
Area (Ha) % Area (Ha)
9.1 55.5 4.4


Grassland
% Area (Ha)
26.8 0.9


Wetland
% Area (Ha)
5.6 0.3


Woodland
Area (Ha) %
1.7 10


1977
Densely
populated areas
Agricultural land
Grassland
Wetland
Woodland


27.3 4.6
0.1 0.4
1.8 1.2
0.3 0.2


338.3
10
27.2
36.1


60.5 10.3
16.7 56.5
25.4 16.1
18.3 16.6


24.9 4.2 138.4 23.5


1.3
102.2
0.1


1.4
0.8
55.5


4.7
0.5
50.3









CHAPTER 5
DISCUSSION AND CONCLUSION

5.1 Introduction

This thesis examined the growth of wetland farming and its impact to the environment.

The first four chapters presented the background and introduction to the study, literature review,

analyses of farmer and household characteristics within the study area, and the assessment of

land use and land cover change. In this fifth and last chapter, discussion of the maj or findings is

covered. This chapter relates research findings to the research questions and hypotheses posed in

Chapter 1. Policy and practice implications of these findings are also presented. A study

conclusion that also discusses study limitations and recommendations for future research is

covered in the last section of this thesis.

5.2 Major Findings

Analysis of farmer and household characteristics within the study area used a set of

questionnaires to determine how vinyungu-farming practice is related to household and farmer

characteristics; how vinyungu-farming system evolved between 1970s and 1990s; what have

been the driving forces behind the changes in the vinyungu-farming practice; and, what have

been the social and environmental implications of changes in vinyungu-farming system. The

primary objective of this thesis was to study changes in land use over time and the environmental

effects associated with those changes.

Many farmers involved in vinyungu farming around Ndembera wetland in Iringa and

Mufindi Districts of Iringa region, Tanzania, provided information for this study regarding their

wetland use and their perceptions of changes. Analysis involved both quantitative and qualitative

data from participating farmers and group discussions conducted in the study area.









I used the information gathered to test the hypotheses that members of households located

within a short distance from the wetlands are more likely to practice vinyungu-farming system

and that short distance from the wetlands to roads and markets would be associated with greater

proportion of wetland converted to agriculture. I also tested the hypotheses that road and market

accessibility would be associated with cultivation of a greater proportion of cash crops than food

crops; greater cash crop production would be associated with increased costs of farm inputs

(labor and fertilizer); greater number of people within the working age in a household would be

associated with greater farm size and number of crops grown; and that farmers are likely to use

household income on farm inputs (i.e., fertilizer and labor) so as to improve crop production.

Farmer' s sex or duration of residence in the village was not expected to be associated with the

size of vinyungu-farming area or types of crops grown. Some of these hypotheses have been

supported while others have been rej ected as has been illustrated in the following sections.

The findings show that almost all farmers in the study area cultivate both vinyungu and

upland plots. However, the size of their upland plots is larger than that of vinyungu, i.e., 70% of

farmers own > 1.2 ha in the uplands compared to 1 1% who own > 1.2 ha of vinyungu. Although

most farmers appear to own relatively small vinyungu plots, the overall area used for this farming

practice is large because almost all farmers in this area own vinyungu. Ownership of both

vinyungu and upland plots means that vinyungu cultivation is not there to replace upland

cultivation but that both vinyungu and upland cultivation are important sources of livelihoods for

the people of Iringa. Generally, a farmer' s decision to cultivate vinyungu is driven by the

availability of labor, water, and soil fertility and not by lack of land.









At least 63% of those interviewed were men. Recalling that women previously

predominantly undertook vinyungu cultivation, the higher proportion of men' s involvement

indicates a growing interest in wetland use among men, possibly for highly profitable cash crops.

Although almost all farmers cultivate on land they have acquired through inheritance,

usually, in close proximity to the villages they live in, at least 15% bought the land they

cultivate. The relatively large proportion of land purchases in these areas, where native farmers

would mostly acquire land through traditional, non-monetary transactions, might be an indicator

that there is a growing need and willingness to acquire and cultivate natural wetland areas among

individuals from other areas of the region or country.

Farmers may request access to vinyungu owned by colleagues or relatives both within and

outside their village, especially from landowners that are unable to cultivate their entire property

due to lack of human and financial resources. This trend however, is becoming rare while renting

is slowly increasing. Currently though, only a few farmers do rent vinyungu plots from fellow

farmers because such plots are not easily available and are relatively expensive. In most cases,

vinyungu are held as a precious and limited resource by families and therefore, all farmers

interviewed in this study reported that they do not rent out their plots to other farmers. Where the

wetland area is large enough to accommodate more farmers, farmers from within the village

could access more land without any formal request. Water resources are communally managed to

ensure equal access among farmers. However, where the wetland resources are limited, there is

an increase in conflicts over access to vinyungu land and to water resources. These conflicts were

traditionally uncommon in the study area.

Family members (i.e., husband, wife, children, dependents) are the most commonly used

labor. This further explains that vinyungu cultivation is no longer a predominantly women's










activity but rather an activity of the entire family. In fact, the linear regression revealed that both

men and women cultivated vinyungu and that men had more of their vinyungu plots grown cash

crops compared to women. The linear regression further revealed that households with a greater

number of people within a working age group are more likely to have more area in vinyungu

during the dry season. This is mainly because construction of vinyungu is both time-consuming

and labor-intensive due to wet and heavy soils that are typical of wetlands. The time required to

make a kinyungu is determined by the size of kinyungu desired and the labor available to

accomplish the task. Labor hire is considered an expensive undertaking and it is therefore not

often used.

Crop production in the uplands has declined due to decline in soil fertility and rainfall. On

the other hand, crop production in the vinyungu has increased mainly because farmers cultivate

the vinyungu twice or thrice a year, maximizing use of the available water resources and fertile

soils. Vinyungu are therefore considered important for both food and cash crop production. In

this case, water, soil fertility, and increased production seem to influence farmers' decisions to

cultivate vinyungu and/or increase the size or number of plots.

The bivariate analysis revealed that more farmers use at least half of their vinyungu plots

for growing food crops compared to a few who use the same hectareage for growing cash crops.

However, the linear regression, which included all factors influencing vinyungu farming, showed

that farmers would have at least 61% of their vinyungu grown cash crops. This result suggests a

growing importance of wetland use among the residents, i.e., an activity historically known to

supplement food production from upland plots is now increasingly used to produce cash crops.

Close proximity to both markets and roads was shown to influence an increase in both the

size of vinyungu and the proportion of cultivated cash crops. However, in my study, these did not









show statistical significance. Various factors may explain why proximity to maj or markets and

roads may not affect cash crop production. One, there have been concerted national efforts to

improve communication networks in the country making most remote places easily accessible;

and two, due to good road networks and market demand in maj or cities, middlemen buy crops

from the farm and transport the produce to maj or markets where crops are sold at almost twice or

thrice the original price. This eases sale of produce by farmers at the same time enables

middlemen to make profit.

Residents with fewer years of residence in the villages are more likely to have larger sizes

of vinyungu plots in the dry season (p=0.048). It is not clear what could be the reason for this

finding, but one possible explanation could be the influx of new residents into the wetland areas

in recent years due to better crop production in the wetlands. These farmers did not have access

to inherited crop land.

Greater distance from the house to the plot was associated with lesser cash crop growing in

the vinyungu. That is, plots that are further from home are used to grow mostly food crops. These

farms are therefore not very active because active farms require adequate labor and frequent

attention and as such, they are usually close to settlements.

Duration of residence was associated with the size of vinyungu plots for growing cash

crops, i.e., residents with fewer years of residence in the villages were more likely to have larger

proportions of their vinyungu plots used to grow cash crops. It appears that new resident farmers

are those that are interested in increasing income through the cultivation of wetlands. Thus, most

recently settled farmers may be more inclined to growing cash crops than food crops.

Expansion of vinyungu and cultivation near sources of water may be associated with

farmers' perceptions of reduced water flow, reduced soil fertility, reduced wetland size and










vegetation as well as increased soil erosion. As a result, at least 87.5% of farmers use inorganic

fertilizer on vinyungu currently (1990s-00s) compared to the 1970s when they did not use

fertilizer. As water resources are becoming limiting in Iringa region farmers reported increasing

human conflict over use of these resources.

As discussions with farmers clearly indicated an increased use of wetland resources for the

cultivation of vinyungu, I used GIS and remote sensing technologies to determine the magnitude

of wetland conversion to agriculture. GIS and remote sensing technologies are an economically

feasible way of gathering information with high spatial, spectral, and temporal resolution over

large areas (Verstraete et al., 1996). Lack of recorded historical data may limit the use of

remotely sensed data to detecting land use changes due to difficulty in estimating uncertainties

about the land use classification. However, Hield survey on land use, interviews with local

farmers and district staffs to gather historical land cover data, and simple classification systems,

make the use of remotely sensed data an effective means of acquiring information on land use

changes, and these methods were employed in this research.

Figure 4-2 shows the Ndembera wetland has experienced land use changes over time. A

transition matrix was developed to test Hypothesis 5, "greater expansion of vinyungu farms

would be associated with increased conversion of wetland area to agricultural land". Table 4-3

indicates a transition probability of 17% of the wetland area to agriculture.

Farmers linked their interest on the wetland for crop production to lack of alternative land,

soil fertility, and availability of water. In addition to these, they related the increased use of

wetlands for agricultural purposes to changes that took place in the country about ten to twenty

years ago. These include the structural adjustment program initiated in the mid 1980s that,

among other things, took government subsidies away from farmers. Inputs such as inorganic









fertilizer became limited and expensive. Also, in the 1990s and especially 1992 and 1993,

Tanzania experienced serious and prolonged droughts that affected crop production in the

country. To many farmers, wetlands (typically with rich soils and wet conditions) offered a

solution. Farmers also reported receiving government directives to cultivate vinyungu to

overcome food insecurities.

Farmers use at least 61% of their vinyungu to grow cash crops. This is to say vinyungu,

historically known to supplement food production from the uplands, are now increasingly used to

produce cash crops. Farmers associated this change to increased markets for their goods that was

in turn influenced by trade liberalization that took place in the mid 1980s and was stronger in

1990s. Trade liberalization replaced the former planned economic system with a market-driven

economic system. It opened doors for Tanzanians to market their goods (both inside and outside

the country) unlike during the former system where the government controlled all the marketing.

This increased the ability of Tanzanians to purchase goods. It also made markets for Tanzanian

goods (including food) available. The increased purchase power and markets encourage farmers

to increase production, both in the uplands and the wetlands.

As noted in chapter 3, farmers can cultivate vinyungu twice, even thrice a year. This

increases production and financial gains, encouraging further conversion of wetlands to

agriculture especially if the household has adequate labor to cultivate the arduous wetland soils.

The percentage change of the wetland to agriculture seems small. However, some of the

factors that seem to contribute in driving vinyungu expansion (droughts, poor soil conditions, and

market forces) intensified in the late 1980s early 1990s, meaning, the observed land cover

changes took place within a short period of time.









5.3 Conclusion

This study was conducted in Iringa and Mufindi districts, both of Iringa region in

southwest Tanzania. Nine villages were studied. The goal of this study was to investigate the

nature of transformation of the vinyungu-farming practice in the two districts and to determine

factors influencing the transformations in this farming system and environmental implications

associated with this farming system. Two approaches were used. First, I used structured

questionnaires to determine farmer and household characteristics associated with vinyungu-

farming system as well as farmers' perceptions on how the system has changed over time, factors

influencing the change, and environmental implications associated with these changes. I then

used remote sensing and GIS to investigate the relationship between wetland change and

agricultural expansion.

Several factors have been found to influence the transformation of wetlands to vinyungu

fields. These include soil infertility of the upland plots making the rich and moist wetland soils

more attractive for agriculture. In the 1970s, vinyungu were small gardens cultivated by women

to supplement household diets. However, with population growth, reduced soil fertility in the

upland plots, rainfall unreliability and increased droughts, a demand for more arable land

increased. Availability of surface water and the high water table of Iringa provide vital moisture

that allows farmers to cultivate more than once in a year, hence maximizing household annual

crop production. Generally, wetlands in the study area can still accommodate more vinyungu.

This, plus the availability of family labor encourage further conversion of wetlands. Accessibility

to maj or markets and roads did not appear to be strongly associated with the size of vinyungu

cultivated or the proportion of vinyungu cultivated cash crops. However, given the improvement

of the road systems, most farmers do not have to transport their produce to the maj or markets, as

buyers drive to the farms to collect the produce to be sold in maj or markets. Lastly, farmers










reported that sometimes they cultivate vinyungu following government directives during severe

droughts.

This research also looked at how vinyungu farming practices have changed over the years.

The results indicate a shift in agriculture in the area i.e. vinyungu farming system that

supplemented household food obtained from uplands fields has evolved to an economic activity,

producing cash crops for the internal market as well as for maj or towns and cities both in and

outside the region. Also, due to a decline of production power in the uplands and increased

profitability in the vinyungu, this traditional irrigation system is transforming from being

exclusively a woman's activity to being an activity for both sexes. A slight increase in the size of

vinyungu cultivated, use of fertilizer, year-round production of crops, and encroachment on to

wetter parts of the wetland are other ways by which farmers have modified their traditional

irrigation system.

Modifications in the vinyungu-farming system have been influenced by a number of

factors. These include social change in terms of population growth that required technological

methods which would augment productivity so as to cope with growing populations and market

demands that require fast, massive and efficient crop productivity (hence the use of fertilizer and

repeated cultivation). Another factor is the decline of production power in the uplands due to

various reasonssuch asreduced soil fertility and droughts. Improved accessibility allows the

middlemen to reach farmers easily, increasing farmer' s ability to market both perishable and

non-perishable crops. Individual needs to increase personal income to overcome social

challenges and access social amenities such as education, health and clothing, also influence the

modifications of vinyungu-farming system. Other reasons for modification of vinyungu-farming

system may be the great profit accrued from sale of vinyungu produce over a short period given









the repeated production over a year. This encourages farmers to increase their energy on wetland

plots. As such, use of the wetlands for vinyungu cultivation has increased when compared to the

1970s as was indicated by nearly 98% of the respondents in this studyas well as through the GIS

analysis that indicated the wetland area declined between 1977 and 1999 with agriculture being

the maj or reason for that decline.

Overall, good progress has been made in achieving the aims of the study. However, a few

factors may have influenced the outcome of the study, for example, the small sample size and

study area. Important data such as historical census at the village level is either lacking or

incomplete, hence limiting the analysis of population change at the village level. Future research

is needed to improve the predictive ability of the models used by including other factors, such as

water quality, water flow and rate as well as better measurement of distance from the plots to

markets and roads. I also recommend that future studies involve a much larger sample of farmers

or follow the same group of farmers over time to establish the validity of the relationship

between distance to maj or markets/ roads and growth of vinyungu farming as well as the

relationship between women' s access to the middlemen and growth of vinyungu farming. Future

studies may need to determine whether or not women compete with men in vinyungu farming

and whether or not women have changed their roles following men's participation.

Finally, it is apparent that wetlands are of a great social and economic importance to the

people living around them. On the other hand, farmers admitted an increased deterioration of

wetland resources, an outcome they linked to a growing level of human activities. More

importantly, farmers clearly understand that if the current use of wetland resources remains

unchecked they may end up losing the very resource they depend upon for their livelihoods. As

such, farmers welcome any intervention that would balance human needs and wetland




Full Text

PAGE 1

TRANSFORMATION OF VALLEY-BOTTOM CULTIVATION AND ITS EFFECTS ON TANZANIAN WETLANDS: A CA SE STUDY OF NDEMBER A WETLAND AREA IN IRINGA REGION By LUCY MAGEMBE A THESIS PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLOR IDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF MASTER OF SCIENCE UNIVERSITY OF FLORIDA 2007

PAGE 2

2 2007 Lucy Magembe

PAGE 3

3 To my son, George, and daught er, Iman, who fill my life with joy and whose presence encourages me to fight my battles earnestly so as to be able to provide for them.

PAGE 4

4 ACKNOWLEDGMENTS Many people provided invaluable support. Sinc e it is impossible to acknowledge each person individually, I extend my sincere appreciati on to all who played a part in this work. However, I feel obligated to th ank individually a few people who ha ve given generously of their time, knowledge, expertise, and moral support. First and foremost, I extend my deep apprecia tion to my supervisory committee chair (Dr. Michael Binford) for his invalu able assistance, advice, and encouragement. He practically challenged me to work independently but continue d to stimulate my analytical thinking, enabling me to improve my writing skills. I am indebted to my committee members Dr. Abe Goldman and Dr. Sandra Russo for helping me grow intellectually and develop a more focused thesis. Their input to this work and their attention are hi ghly appreciated. I extend my sincere gratitude to my sponsors, The United States Agency for International Development (USAID) and the Africa-America Institu te (AAI), who made my stay in the United States content and without whose financial support this work would not have come to fruition. In Tanzania, I would like to thank the people of Iringa and Mufindi Districts for their time, support, and kindness. Sincerest appreciation is ex tended to the District Facilitation Teams and to WWF (World Wide Fund for Nature) staff for their endless support during fieldwork. I thank members of staff and students of the Geography Department at the University of Florida for their moral support and for providing a social and academic atmosphere in which I was able to grow. My gratitude goes to my husband, George Mage mbe, my sister Annamarie Kiaga, and my late mom, Mrs. Tabitha Kashaija for their co ntinued love, support, encouragement, and for believing in me.

PAGE 5

5 No words can express how grateful I am to the Mkanta family for allowing me to stay with them for endless weeks while completing my thesis A heartfelt gratitude goes especially to Dr. William Mkanta for his assistance in statistical methods. Anna Mushi, Tunu Mndeme, Simon Mwansasu and Juma Rajabu Mshana are acknowledged for their assistance in perfecting the aeria l photo interpretation and GIS analysis. Lastly, I give thanks and prai ses to the Almighty God for th e blessings in my life, but mostly for enabling me endure the stresse s associated with graduate schooling.

PAGE 6

6 TABLE OF CONTENTS page ACKNOWLEDGMENTS...............................................................................................................4 LIST OF TABLES................................................................................................................. ..........9 LIST OF FIGURES................................................................................................................ .......10 ABSTRACT....................................................................................................................... ............11 CHAPTER 1 INTRODUCTION..................................................................................................................13 1.1 Background and Statement of the Problem......................................................................14 1.2 Significance............................................................................................................... .......16 1.3 Objectives................................................................................................................. ........17 1.4 Study Questions............................................................................................................ ....17 2 LITERATURE REVIEW.......................................................................................................21 2.1 Introduction............................................................................................................... ........21 2.2 Conceptual Framework.....................................................................................................21 2.2.1 Concepts of Land-Use and Land-Cover Change....................................................21 2.2.2 Human-settlement Models of Land-Use................................................................24 2.2.3 Population Growth and Agricultural Land-Use.....................................................26 2.2.4 Drivers of Smallholders Farming Decisions..........................................................29 2.2.5 Decline in Agricultural Producti on and its Relation to Land Use..........................32 2.2.6 Wetlands: An Alternat ive Landscape Component for Crop Production................34 2.2.7 Irrigation Practices and Wetland Use in Tanzania.................................................37 2.3 Summary of the Theoretical Backgr ound and Formulation of Hypotheses.....................40 2.3.1 Summary of the Theoretical Background...............................................................40 2.3.2 Justification of the Hypotheses...............................................................................43 3 STUDY OF FARMER AND HOUS EHOLD FACTORS INFLUENCING VINYUNGU FARMING SYSTEM.............................................................................................................45 3.1 Study Area................................................................................................................. .......45 3.1.1 Location................................................................................................................. .45 3.1.2 Physical Environment.............................................................................................45 3.1.3 Demographic Characteristics..................................................................................46 3.1.4 Socioeconomic Situation........................................................................................46 3.1.5 Agriculture..............................................................................................................47

PAGE 7

7 3.2 Material and Methods.......................................................................................................48 3.2.1 Sample Profile and Sampling Procedure................................................................49 3.2.2 Data Collection.......................................................................................................50 3.2.3 Methods of Data Analysis......................................................................................54 3.2.3.1 Descriptive Statistics....................................................................................55 3.2.3.2 Measures of Associations.............................................................................55 3.2.3.3 Multivariable Analysis.................................................................................55 3.3 Results: Farmer and Household Charact eristics of Valley-Bottom Cultivation ( vinyungu ) around the Ndembera Swamp in Iringa and Mufindi Districts.........................56 3.3.1 Sociodemographic Characteristics.........................................................................56 3.3.1.1 Land tenure/ownership.................................................................................58 3.3.1.2 Crop production, crop preference and land use............................................59 3.3.1.3 Inputs on vinyungu .......................................................................................60 3.3.1.4 Multivariable analysis..................................................................................61 3.3.2 Environmental Issues Related to Vinyungu Cultivation.........................................64 3.3.3 Discussion...............................................................................................................65 4 USE OF GEOGRAPHIC INFORMATION SYSTEMS AND REMOTE SENSING TO DETERMINE THE MAGNITUDE AND RATE OF CONVERSION OF NDEMBERA SWAMP AS A RESULT OF VINYUNGU EXPANSION.....................................................79 4.1 Data Sets.................................................................................................................. .........80 4.2 Methods.................................................................................................................... ........80 4.3 Data Analysis.............................................................................................................. ......81 4.4 Results.................................................................................................................... ...........81 4.5 Discussion................................................................................................................. ........82 5 DISCUSSION AND CONCLUSION....................................................................................91 5.1 Introduction............................................................................................................... ........91 5.2 Major Findings............................................................................................................. .....91 5.3 Conclusion................................................................................................................. .......98 APPENDIX A SAMPLE SIZE DETERMINATION...................................................................................102 B SURVEY INSTRUMENT: INDIVIDUAL FARMER QUESTIONNAIRE.......................103 B.1 Location................................................................................................................... ......103 B.2 Background Information................................................................................................103 B.3 Land availability, crop preferen ce, seasonality, and fallow period...............................103 B.4 Labor, Input, and Tools on Vinyungu ............................................................................105

PAGE 8

8 C SURVEY INSTRUMENT : GROUP QUESTIONNAIRE..................................................107 C.1 Background................................................................................................................. ...107 C.2 Land availability, cropping preference, and fallow period............................................107 C.3 Input, Tools and Labor...................................................................................................107 C.4 Wetlands vs. Uplands.....................................................................................................108 LIST OF REFERENCES.............................................................................................................111 BIOGRAPHICAL SKETCH.......................................................................................................120

PAGE 9

9 LIST OF TABLES Table page 3-1 Sociodemographic and Land Use Characteris tics of the Sample Farmers in the study area (n=54).................................................................................................................... .....69 3-2 Maize yields in uplands and vinyungu between 1970s and 2000s (village responses, n=9) (per year, i.e., one season in the uplands and two seasons in the vinyungu ).............71 3-3 Potential market price per unit output from cash crops grown in vinyungu and under rain fed agriculture........................................................................................................... ..71 3-4 Results of Chi-Square Analysis of th e Farmer and Household Factors influencing Vinyungu -Farming Practice, i.e., size of vinyungu plots owned by household in relation to sex, age, years of residence in the village, and number of potential workers........................................................................................................................ .......72 3-5 Results of Chi-Square Analysis of th e Farmer and Household Factors influencing Vinyungu -Farming Practice, i.e., size of vinyungu plots grown in the dry season in relation to sex, age, years of residence in the village, and number of potential workers........................................................................................................................ .......72 3-6 Results of Chi-Square Analysis of th e Farmer and Household Factors influencing Vinyungu -Farming Practice, i.e., number of cash crops grown on vinyungu plots in relation to sex, age, years of residence in the village, and number of potential workers........................................................................................................................ .......73 3-7 Results of Chi-Square Analysis of th e Farmer and Household Factors influencing Vinyungu -Farming Practice, i.e., size of vinyungu plots grown cash crops in relation to sex, age, years of residence in the v illage, and number of potential workers................73 3-8 Results of the Linear Regression Analysis of the Factors Affecting Vinyungu Farming Practice............................................................................................................... .74 3-9 Farmers perception on various environmen tal issues as discussed in the village meetings: the current state compared to the 1970s (n = 40)..............................................75 3-10 Individual farmers perception on various environmental issues (n=54)..........................78 4-1 Map History (i.e., Data used).............................................................................................86 4-2 Total area (ha) and area of change of land use types from 1977 to 1999..........................90 4-3 Transition matrix of land use types fr om 1977 to 1999 (transition probabilities in %).....90

PAGE 10

10 LIST OF FIGURES Figure page 1-1 Map of Iringa region showing the st udy area (A). (Sour ce: based on 1:2,000,0000 Tanzania Administrative Map of 1989, by Su rveys and Mapping Division. Ministry of Lands and Human Settlement). The inset (B) shows the locati on Iringa region in Tanzania....................................................................................................................... ......18 1-2 Typical vinyungu grown potatoes (A) and peas (B). Both pictures taken in Usengelindete village, Iringa Dist rict of Iringa Region, June 2005..................................19 1-3 Enlarged map (A) shows Lyandembera swamp and surrounding villages. Subset image (B) shows the location of Lyandemb era swamp (The Study Area) in Iringa region......................................................................................................................... ........20 3-1 A well is dug where water is not free flowing. Picture taken in Lumuli village in Iringa District, June 2005...................................................................................................76 3-2 A river is diverted to supply vinyungu w ith water. Picture taken in Usengelindete village, Iringa District, June 2005......................................................................................76 3-3 Picture showing Ndembera river behind which vinyungu cultivation is taking place in Maduma village, Mufindi District. An ex ample of how close to the river the vinyungu are getting. Picture taken in June 2005..............................................................77 4-1 Ndembera swamp in 1977 (A). The inse t map (B) shows Ndembera swamp location in Iringa region............................................................................................................... ....87 4-2 Ndembera swamp in 1999 (A). The inset map (B) shows tha location of Lyandembera swamp in Iringa region...............................................................................88 4-3 Land use/ cover change map (1977 and 1999) (A). The inset map (B) shows the location of Lyandembera swamp in Iringa........................................................................89

PAGE 11

11 Abstract of Thesis Presen ted to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Master of Science TRANSFORMATION OF VALLEY-BOTTOM CULTIVATION AND ITS EFFECTS ON TANZANIAN WETLANDS: A CA SE STUDY OF NDEMBER A WETLAND AREA IN IRINGA REGION By Lucy Magembe May 2007 Chair: Michael William Binford Major: Geography This study, conducted in southwestern Tanzania examines the social and environmental implications of conversions of wetlands to altern ative land-uses, in partic ular agriculture. Using the case of the vinyungu -faming system, a wetland form of ag riculture practiced in Iringa region, the study demonstrates complex ways in which ongoing pressure over food production is causing transformations in the traditi onal cultivation methods. To do s o, four questions were asked, Research questions 1: How is vinyungu -farming practice rela ted to household and farmer characteristics? Research questions 2: How has the vinyungu -farming system evolved over time? Research questions 3: What have been the possible dr iving forces behind the changes in vinyungu -farming practice? Research questions 4: What is the effect of vinyungu -farming system? To answer these questions, the study uses pr e-designed questionnaires to examine farmers perceptions on how the system has changed over time and how this change may affect the wetland resources. The study also uses Geographi c Information Systems (GIS) to determine whether or not there is a change in the wetla nd area used for agricultu re in the study area. Results from this study indicate a shift in agriculture whereby vinyungu -farming system that supplemented household food obtained from upland fields has evolved to an economic activity, producing cash crops for local, regi onal and national markets. In addition, this

PAGE 12

12 traditional irrigation system is transforming from being predominantly a womans activity to an entire familys activity. An increase in the size of vinyungu cultivated, use of inorganic fertilizer and encroachment of wetter parts of the wetla nd are some of the ways by which farmers have modified their traditional irri gation system. Crop production is no longer confined to the dry season but extends into the wet season, especially in areas that are not heavily inundated in water. Results from the GIS analysis indicate an increased conversi on of wetland area to agriculture between 1977 and 1999. GIS analysis during the same period also indicate an increased conversion of the wetla nd to grassland, a factor that fa rmers attributed to reduction of streamflow into the wetland as a re sult of increased human activities. Several factors have been found to influence the transformation of wetlands to vinyungu fields and expansion of these fields. Thes e include population grow th, increased market demands, government policies, and reduced produ ction power in the uplands as a result of rainfall unreliability, increased droughts and reduced soil fertil ity. All of these factors have increased the demand for more arable land, in this case, the wetlands. In addition, most households must have an adequa te labor force to cultivate th e wet and heavy soils that are characteristic of wetlands. Improved communica tions networks and higher profitability in vinyungu compared to upland plots also encourag e modifications in this farming system. Results of the study indicate that wetlands provi de both social and economic benefits to the people that live around or near them. On the other hand, these very resources that sustain livelihoods and perform important functions such as flood control and water quality improvement are increasingly being lost. It is im portant therefore for any future land-use regimes in and around wetlands to consider human need s and wetlands sustainability simultaneously.

PAGE 13

13 CHAPTER 1 INTRODUCTION Wetlands are important landscape components and critical natural ecosystems that provide significant environmental and socio-economic benef its to humans (Roggeri, 1998; Silvius et al., 2000; Stuip et al., 2002). Because of their charact eristics (e.g., heavy soils, thick vegetation sometimes on extensive flat lands), wetlands st ore water (and nutrients) during wet periods and maintain base flow during dry periods (B alek and Perry, 1973; Dugan, 1990; Roggeri, 1995). Thus, wetlands can control floods, improve wate r quality, and provide unique habitat for wildlife, including many rare and endangered species. These ecosystems provide numerous benefits to humans including food, pharmaceutic als, and construction materials (Daily, 1997). Rich soils and high moisture holding capacity make wetlands particularly attractive for agriculture. However, wetland so ils can be waterlogged and anaer obic, therefore inimical for plants unless they are adapted to soggy soils or some means are used to raise the rooting zone above the saturated layer. For millennia, humans have cultivated wetlands to meet food securi ty and livelihood needs without necessarily affecting wetland structure and functions (Adams, 1993b; Banzi et al., 1992; Erickson 1985; Roggeri, 1995; Scoones, 1991). In recent years, wetlands have come under extreme pressure as many have been converted to agriculture, which ra ises concern over the sustainability of wetland cultivat ion, especially in tropical deve loping countries where wetlands are among the least protected ecosystems (Di xon and Wood, 2003; Gerakis and Kalburtji, 1998; Hollis, 1990; Jensen et al., 1993; Jensen et al., 19 95; Liu et al., 2004; Mu nyati, 2000; Ringrose et al., 1988; Roggeri, 1995; Wang et al., 2006; Williams, 1991). This study is an effort to explore changes in wetland farming systems and factors that are driving these changes. I used the experience of an indigenous wetland farming syst em in Iringa region, Tan zania, to investigate

PAGE 14

14 why and how the practices in this farming system have been transformed over time and determine the social and environmental implications of that transformation. 1.1 Background and Statement of the Problem Iringa region, southwest of Tanzania (Figur e 1-1), is surrounded by semi-arid conditions. Therefore, wetlands are the most ideal resource for agriculture. Farmers in wetland areas have devised a traditional farming system called vinyungu that allows year-round cultivation of crops (Majule and Mwalyosi, 2003). Vinyungu (or kinyungu in singular form) are ridges or raised beds that are about 0.6 m high and 4-20 m wide w ith a cambered surface sloping down to the open drain on either side. They are created by first clearing the land, then burning off the cleared vegetation, followed by hand hoe plowing that is done simultaneously with the construction of ditches and ridges, and lastly, harrowing, to smoothen the ridges. This farming system is highly prevalent in Makete, Ludewa, Mufindi, Iringa a nd Njombe districts of Iringa region and it is practiced mainly by smallholder farmers of Bena and Hehe tribes. It is mostly a dry season (June-October) agricultural activity that takes place almost exclusively in wet valley bottoms that are characterized by heavy clayey soils. Green ma ize, beans, potatoes, and vegetables are the most commonly grown crops. Figure 1-2 shows typical vinyungu grown crops. Vinyungu farming, believed to have started as fa r back as the 1890s (Culwick, 1935), was practiced mainly by women on small fields, wi th little or no economic contribution to the livelihoods of the people that practiced it (Kuroda, 2001; Le ma, 1996). Recently, however, a combination of socio-economic factors has caused the extensification of the farming system, in terms of size of vinyungu and the number of people involved in this farming practice. Among these factors may be the increased market de mand for food and other wetland products like vegetables for urban centers in Ir inga region as well as the distant cities of Dar es Salaam and Mbeya. For example, Olindo (1992) observed the increased market demand for food and other

PAGE 15

15 wetland products in some parts of Kenya as one of the major factors driving further conversion of wetlands to agriculture. Similar to other cases in Eastern Africa (Denny and Turyastunga, 1992; Gichuki, 1992; Loevinsohn et al., 1992; W ood, 1996), the ability of farmers to harvest crops twice or even thrice a y ear hence increasing food security especially in drought-prone areas may also play an important role in vinyungu expansion in Iringa region. Burgeoning human population, socio-economic changes, and governme nt policies that call for use of more agriculturally productive land to reduce food deficits, are consider ed some of the major factors causing wetland conversions (Roggeri, 1995). Generally, regardless of the motivation, unrestr icted drainage and cultivation of wetlands can have far-reaching and sometimes irreversib le consequences. Unrest ricted drainage and cultivation of wetlands can affect wetlands hydrological functions, causi ng a reduction in water storage and quality, variable str eam flows, and sometimes, complete dryness of wetlands (Denny and Turyatunga, 1992; Dixon and Wood, 2003; Roggeri, 1995; MWLD, 2001). In addition, riverbanks are eroded and sediments accumulate downstream, affecting the overall structure and functions of wetlands (Kaswamila and Tenge, 1997; Olindo, 1992). These factors have serious effects on the livelihoods of the local communities, especially those residing further downstream people walk long distances in search of water; land becomes uncultivable; livestock keepers are forced to walk long distances in search of wa ter and pasture; communal conflicts may arise due to animal trampling of crops; and, fish and w ildlife resources decline (MWLD, 2001). Despite the growing awareness of wetland values and functions and the consequences of human intervention, the issue of wetla nd loss and degradation has rece ived lesser attention in many African countries than other major environmental i ssues such as desertific ation and deforestation (Acreman and Hollis 1996).

PAGE 16

16 1.2 Significance Understanding the impacts of draining and cul tivating wetlands may encourage farmers to adopt more sustainable practices to sustain their livelihoods. Previ ous studies in Tanzania have tended to focus on major wetlands that are mostly associated with activities of greater economic importance such as large-scale agriculture, transportation, and fi sheries (Pallela, 2000). Sutton (1969) and Lema (1996) who described vinyungu -farming systems and the associated technologies reported that this sy stem continues to receive litt le attention as it is largely considered a side-line agricultura l activity. Mkavidanda and Kasw amila (2001) looked at the role of vinyungu in poverty reduction and concluded that vinyungu are a key factor in sustaining livelihoods and reducing poverty. Majule andM walyosi (2003) examined soil and water characteristics and found that soils under vinyungu cultivation were acidic with a pH of 5.1 to 5.5 and the water samples downstream had traces of agrochemicals and pesticides, implying that in the long-run vinyungu farming may reduce soil and water qualities as well as agricultural production in both the vinyungu plots and further downstream. In contrast, this study looks at the ways in which vinyungu farming has changed over the years and how this change may have influenced so cial and environmental changes. It also looks at the patterns and extent of wetland cover chan ge that may have resulted from transformations in vinyungu farming. Quantitative da ta derived from this study will generate knowledge on changes in land use, the extent of wetla nd conversion and their socio-economic and environmental consequences. This knowledge w ill be useful for a bette r understanding of the relationships between humans and observed en vironmental changes and may encourage best management practices around wetlands. No ot her known study has looked at the effect of Tanzanian vinyungu in landscapes.

PAGE 17

17 1.3 Objectives The primary objective of this thesis is to st udy changes in land use, specifically wetland conversion, over time and the environmental effect s associated with those changes. I wish to contribute to the knowledge on wetland loss, its causes and effects, all of which are essential for policymaking and management of wetland resources. Specific aim 1: To investigate what kind of transformations have occurred in vinyungu farming practice over time; Specific aim 2: To propose the factors that ma y have influenced the recent transformations in this farming practice; Specific aim 3: To determine what is the effect of the vinyungu -farming system; Specific aim 4: To identify and document change of Ndembera wetland area used for agriculture between the period 1977 and 1999. 1.4 Study Questions To meet the study objectives, I used informa tion from farmers interviews in Ndembera wetland area, Iringa region, Tanzania (Figure 1-3). To identify long-term changes in land use in this area, I sought data on land-use/cover cha nges over time. This information was used to explore factors of wetlands transformation and changes in land cover over time. The following questions were addressed: Research questions 1: How is vinyungu -farming practice rela ted to household and farmer characteristics? Research questions 2: How has the vinyungu -farming system evolved over time? Research questions 3: What have been the possible dr iving forces behind the changes in vinyungu -farming practice? Research questions 4: What is the effect of vinyungu -farming system

PAGE 18

18 Figure 1-1. Map of Iringa region showing th e study area (A). (Source: based on 1:2,0 00,0000 Tanzania Administrative Map of 198 9, by Surveys and Mapping Division. Ministry of Lands and Human Settlement). Th e inset (B) shows the location Iringa region in Tanzania. A B

PAGE 19

19 A B Figure 1-2. Typical vinyungu grown pot atoes (A) and peas (B). Both pictures taken in Usengelindete village Iringa District of Iringa Region, June 2005.

PAGE 20

20 Figure 1-3. Enlarged map (A) shows Lyandembera swamp and su rrounding villages. Subset imag e (B) shows the location of Lyandembera swamp (The Study Area) in Iringa region. A B

PAGE 21

21 CHAPTER 2 LITERATURE REVIEW 2.1 Introduction This chapter reviews the literature that undert akes the role of putting together important concepts describing land-use and land-use changes in Tanzania with a focus on the transformations in vinyungufarming system. The review is orga nized into two major parts. The first part reviews some important concepts of la nd-use/cover change as they relate to wetland conversion to agriculture. It al so reviews theories underlying tr ansformations of agricultural practices and their effect on we tland resources. These include hu man-settlement models of land use, population growth and its implications for ag ricultural land use, and dr ivers of smallholders farming decisions. In addition, this part of the re view also discusses the state of agricultural production in Africa and reasons for its decline; use of wetlands as an alternative landscape component for improving crop production; and irri gation practices as they relate to wetlandrelated policies in Tanzania. The second part of the review summarizes the theoretical background with a focus on land use at the household level and expl ores important relationships and variables amenable to land-use changes. 2.2 Conceptual Framework 2.2.1 Concepts of Land-Use and Land-Cover Change Wetlands conversion to agricultu re falls under the larger to pic of land-use/cover change (Gerakis and Kalburtji, 1998; Jens en et al., 1995; Lambin et al., 2001; Liu et al. 2004; Meyer and Turner, 1994; Meyer and Turner II, 1996; Meye r and Turner II, 1992; Ringrose et al., 1988; Turner et al. 1993; Wang et al.; 2006). Meyer and Turner II (1992) define the term land use as human exploitation of the land and the term land cover as the physical a nd biotic character of

PAGE 22

22 the land. Land use tends to cause land-cover change that, in tur n, leads to further land-use changes. Land-cover change may involve either land-cover conversion, i.e., change from one cover type to another, or land-cover modifica tion, i.e., alterations of structure or function without a wholesale change from one type to another (B riassoulis, 2000). Land-use change may involve conversion from one type of use to another, i.e., changes in the mix and pattern of land uses in an area, or modification of a certain type of land use i.e., changes in the intensity of this use as well as alterations of its characteristic qualities/attributes (Briassoulis, 2000). Land-use/cover change is driven by many fact ors including changes in the biophysical and socio-economic drivers (Meyer and Turner II, 1 992). The driving factors/forces are a complex set of actions and rationales that give rise to proximate cause s (Mertens et al., 2000, 984). These forces could be exter nally-driven, e.g., natural hazards or internally-driven, e.g., population growth (Mertens et al., 2000, 984). Turn er et al. (1993) liste d four examples of major driving forces of land-use change: populat ion, technology, political economy, and political structure. Farming practices are listed as among the major causes of land-cover changes in tropical Africa (Lele and Stone 1989; Meyer and Turner, 1994; Turner et al., 1994). As driving forces of land-use/cover change in teract differently to one another in different spatial-temporal settings, it is important to ex amine causes of land-use/cover change within a specific spatial-temporal setting. The method of examining both land-use/cover changes involves selection of the land-use/cover t ypes to be analyzed, determining the driving forces and process of change that can be detected, and describing and explaining the linkages between land use and land cover (Briassoulis, 2000). This study focu sed on a traditional farming system called vinyungu common in wetland areas of Iringa region, Ta nzania, to determine how this farming

PAGE 23

23 system has evolved over time, what are the drivin g forces behind this evolution, and what are its effects on wetlands. A challenge in analyzing land-use/ cover change lies in linking the driving forces of change to the observed land-use/cover ch anges. There is considerable variety of theoretical and modeling frameworks and tools that are used to conceptualize and opera tionalize land-use/cover change issues. The theoretical literature on land-u se/cover change includes theories on social and economic determinants of land-use/cover change. Economic theories and models of land-use/cover change use concepts and procedures from economics, e.g., prices of the f actors of production, of products a nd of services, transport cost, marginal cost, and economies of scale, external ities, and, above all, util ity (Briassoulis, 2002). All behavioural assumptions made refer to th e model of the rational, economic, utility maximizing humans (Briassoulis, 2002). Examples of economic theories and models include von Thnens agricultural land rent theory (1966); lo cation models (see, for example, Alonso, 1964; Weber 1929; Losch, 1954); and settlement patte rns models ( see, for example, Blaikie, 1971; Chisholm, 1962). Some of these theories have been elaborated in the subsequent sections of this chapter to show how they relate to la nd-use/cover change and th eir relevance to this study. Theories on social determinants of land-us e/cover change are based on ideas from the social sciences and emphasize the importance of human agenc y, social relationships, social networks, and socio-cultural ch ange in bringing about spatial, political, economic and other changes (Briassoulis, 2002). Theo ries and models used in the social approach include agentbased theories which focus on the agents of change (i.e., users of land) and their interdependencies as well as t hose theories that adopt persp ectives in the nature-society

PAGE 24

24 theorization traditions (see, for example, Live rman, 1994; Malthus 1960; Merchant, 1990; Sack, 1990). Some of these theories have been elaborated in the subsequent secti ons of this chapter to show their relevance to this study. Theories of land-use/cover change use models that are either descriptive or explanatory (Briassoulis, 2000). According to Briassoulis (20 00) description of land-use change documents changes from one type of land use or cover to another over a given ti me period and within a given spatial entity while explanation attempts to address the question of why these changes have occurred (or, are occurring) and to uncover the factors or forces that bring about these changes directly or indirectly, in the shortor long-run (Briassoulis 2002). Description of land-use/cover change can be achieved using GIS modeling approaches whereby maps from different time periods are ov erlaid to identify the location and assess the magnitude of change. Explanation of why thes e changes take place can be achieved through socio-economic surveys and physical observations. Both methods have been employed in this research. The descriptive part is covered in Chapter 4 while a re view of theories that explain why these changes occur (and their effect s) is covered in the following sections. 2.2.2 Human-settlement Models of Land-Use Different theories attempt to describe and e xplain the factors influe ncing human settlement and land use. Most of these theories are ba sed on economic concepts. One of the pioneering economic theories of land use is von Thnens agricultural land rent th eory, developed in 1826, where land rent is defined as the price paid by th e tenant for a particular land use on a particular piece of land (von Thnen 1966; original in 182 6). Through this theory, von Thnen modeled the locational distribution of crops (o r, land-use patterns) in a lands cape as an algebraic function of yields, market prices, production costs, tr ansportation rates and distance to market.

PAGE 25

25 According to von Thnen, land rent decreases with increased distance from the center (city/ town/ market). The rent one can afford to pay for a land-use activ ity on a parcel of land depends on the value of the products produced on that land. This value de creases with increased distance from the market. Therefore, a land-use ac tivity with the highest value of output is found on land with the highest rent, i.e., closest to the market (this would be the most valuable piece of land in a village context). Trans portation costs increase with increased distance from the center. Therefore, transportation costs are inversely prop ortional to the land rent. Since farmers tend to balance land cost, transportation costs, and profit, they produce the most cost-effective products for market near the center, i.e., they cultivate the high market value-crops near the city/ town/ market (lest transportation costs reduce profit to zero) and lower market value-crops further away from the city (meant for loca l use). Therefore, land-use patt erns are influenced by distance to roads, access to markets, and access to tran sportation. This may be true for wetlands use in most of eastern Africa, including Iringa re gion in Tanzania. Dixon and Wood (2003) report commercialization and communicati on networks that have spread throughout easte rn Africa as some of the factors increasing the market demand for food and other wetland products. Thus, wetland use is increasingly shifting from subsistence agriculture to more commercialized agriculture where loca lly marketable crops are grown (Olindo, 1992). Chomitz and Gray (1995) tested the model in Belize and found that market access and distance to roads strongly influence land use. According to the study, ag riculture becomes less attractive with increased distance from the market. Interpreting von Thnens theories in a rural setting, Chisholm (1962) found that there exists a relationship between land use, distance fr om market, and distance from the settlements. Active farms require adequate labor and frequent attention and are ther efore found closer to

PAGE 26

26 settlements. The amount of labor required for a land use activity (e.g., agriculture) tends to decrease with increased distance from the market. Therefore, closer to the market one expects to find more labor-intensive crops, more labor use and more human settlement, hence intensive land use. Based on the frontier thesis (Richards 1990), pe ople first search for unsettled yet desirable lands to settle on. Later, secondary and tertiary frontier zones are created, land-use intensification occurs, and eventually, settlement frontiers are expanded (Richard s, 1990). The spatial scale of penetration and land transformation is facil itated by such factors as expanding economic demand, population growth and technological adva ncement (Richards 1990). Therefore, demand for land and the ability to work on the land tend to influence expansion of settlement frontiers, provided unoccupied land exists. As the population grows and land shortages in areas traditionally farmed increase, more and more peopl e in eastern Africa are forced into marginal areas in search of agricultural land (Di xon and Wood, 2003). Wetlands may have become new agricultural frontiers, replacing dry land margins that had been subject to spontaneous agricultural settlement for decades. In summary, access to roads and markets seems a common link among the different models presented above. Construction of road s, distance to markets, and market demands influence land-cover change. Roads increase a ccess to land by enabling people to reach areas that were not previously reachable. Markets acc essibility and market demands encourage land users to maximize the utility of surrounding land. These two factors tend to bring an economic growth to an area that, in turn, attracts human settlement. 2.2.3 Population Growth and Agricultural Land-Use Many researchers have linked population growth to increased demand for agricultural land and consequently, agricultural ex tensification and intensificati on. Several theories have been

PAGE 27

27 developed to examine this link and research has been conducted to test these theories, providing insight and different pers pectives on the issue. One of the earliest theories on the effect of population growth on agriculture emanates from Malthus (1960) who hypothesized that f ood production, growing at a linear rate, would eventually be overtaken by popul ations growing exponentially. Hi s original predictions have materialized in some areas. According to B ilsborrow (1987), Food produc tion per capita in the developing countries is barely keeping pace with population growth in the developing countries despite the Green Revolu tion (Bilsborrow 1987, 199). Malthus also postulated that, population growth would negative ly impact the environment, potentially causing irreversible land degrada tion that would, in tur n, hamper food production. Bilsborrow and Ogendo (1992) showed that, populat ion growth has indeed contributed to land degradation in some areas, sometimes thre atening food production. From a neo-Malthusian perspective therefore, degradati on is inevitable where levels of exploitation exceed the carrying capacity of individual land resour ce such as individual wetlands The disappearance of parts of the Usangu wetland in Tanzania is one example of the effects of population pressure on wetlands (MWLD, 2001). However, the Malthus hypothe sis assumed unchanging methods of production (i.e., constant technology). Today, economists ar gue that under well-functioning markets, as the population grows and land resources become scar ce, technologies are developed to improve production as well as land management. As a n earby example, a study conducted in Machakos District, Kenya, showed that popu lation growth, market developmen t and capital availability led to technological advancement, which, in tur n, led to increased agri cultural productivity and improved land and water management (Tiffen and Mortimore, 1994; Tiffen et al., 1994).

PAGE 28

28 The economists point of view emanates from Boserups (1965, 1981) original ideas that emphasize the importance of progressive adaptive management. Challenging Malthuss original ideas, Boserup postulated that as the population gr ows, the demand for food increases and arable land becomes scarce, forcing farmers to make adjustments in production so as to improve the quality and productivity of land. These adjust ments, which occur af ter the expansion of cultivated land, include adopti on of land-intensive technolog y, i.e., reduced fallow period, adoption of new technologies, incr eased labor inputs pe r unit of land, and use of natural or artificial fertilizer. Generally, however, the impact of increas ing population on land use and environmental sustainability today appears more complex than what either Malthus or Boserup hypothesized, and varies from system to system. Bilsborrow (1 987) reports that, intens ification/ technological development is dependent upon environmental cond itions of the area (soil fertility, rain). For example, technological change, in terms of in creased irrigation, happens when rainfall is seasonal or irregular or when surface or unde rground water is available and can be tapped (Bilsborrow, 1987). Netting (1968, 1993) reports that, farmers make several adjustments, such as crop diversification, to cope with poor envi ronmental conditions and maximize crop output. Boserup (1965) neglected the possibility of an alternative economic response to increased population pressure: an increase in the area of land under cultivation. This response, called extensification, is achieved through clearing more of ones own land, appropriation of neighboring lands, or, migration to other areas with arable lands (Bilsborow and Ogendo 1992). One of the effects of agricultural expansion is deforestation, especially of tropical forests and highland forested areas (Bilsborow and Oge ndo 1992). Today, this expansion is affecting wetlands as well. Up until the 1940s for ex ample, the Sukuma people of Mwanza region,

PAGE 29

29 Tanzania, cultivated on the hill slopes and left th e valley bottomlands for cattle grazing (Pingali et al., 1987). Today, the bottomlands are cultivated and the adjacent swamp vegetation is cleared to provide more land for cultivation. A similar tr end of converting wetlands to agricultural land is taking place in Iringa region. Boserup also neglected the influence of market demand on technological development. Pingali et al. (1987) clearly show that local dema nds, including market demands, cause land-use/ cover change. Therefore, site-specific carrying capacities, based on often unique environmental and socio-economic characteristics, are a major ca talyst of change in the agricultural farming practices and, consequentl y, the way wetlands are being utilized and managed. 2.2.4 Drivers of Smallholders Farming Decisions Many regions of the developing world consist of substantial rural populations that rely largely on farming as a principal source of inco me, food and employment. The most widespread kind of farm unit is the small family farm or smallholding. Netting (1993) describes smallholder farmers as rural cultivators practicing perman ent, diversified agriculture on small farms. Population density is generally high. Therefore, smallholders live under c onditions of scarcity, deriving their livelihood from an intensively cultivated holding, usually less than 2 to 5 ha. Central to the small-hold inte nsive agriculture is the fam ily unit or household. A household is a cooperative work group engaged in producti on, distribution, transmission, biological and social reproduction, and co-residence (Netting 1993; Wilk 1991) and may consist of a single family, one person, an extended family, or any other group of related or unrelated persons sharing living arrangements. Farming decisions by a household are influe nced by many factors including resource availability (e.g., land, labor), economic status of the household, local traditions, and external factors such as government policies and envir onmental factors. Differences among households

PAGE 30

30 lead to differences in choices each household makes regarding what to cultivate, where and how to cultivate, amount of land to cultivate, what cr ops to be marketed or preserved, whether or not they should access additional land to cultivate, and how to use labor and cash resources. Land tenure security, often with sellable, rentable, heritable rights (although such property regimes can co-exist with communally manage d resources) is important for small-hold householders to have. This is especially importa nt where intensive agriculture is practiced for often this involves land improvements. Smallholde rs usually hold their most productive land as private property while less intensively used land is more likely to be held as commons or rented out. Smallholder agriculture is sustainable where pr oduction is predictable, sufficient to feed the producers, and stable over th e long run. Factors like seasonal ity could affect agricultural production and sustainability, he nce affecting rural household food supply and income (Meertens 1999, Morgan and Solarz 1993). Scanty rainfall and droughts are among the major seasonality factors affecting agricu ltural production (Alexandratos 1995). Technological innovations such as irrigation may be employed in cases of drought s or seasonal rainfall (Bilsborrow 1987). Farmers are therefore attracted to environments with a hi gher probability of sustaining production. This is the case especially in drier and dr ought-prone areas where the availabi lity of moisture in the soils throughout the year and the ability to retain nutrients make wetland cultivation the most sustainable risk management strategy. Labor and economic resources available to the household influen ce household land-use activities. Household members share different res ponsibilities to minimize costs associated with assembling and scheduling labor. In most cases very young and old members of the household are excluded from agricultural activities. La ndholders maximize their financial income by

PAGE 31

31 allocating labor to activities they perceive will pr ovide the greatest financial return on their labor investment. When additional labor is required, farm ers either hire laborer s or alternate manpower with other farmers (Meertens 1999). This may be th e case especially for the harder tasks such as clearing land on heavy wetland soils. In cases of limited labor availability or cash to hire laborers, agricultural expans ion or intensification is ra re or minimized (Boserup, 1965). Smallholders are not economically isolated. Income sources ra nge from farm, off-farm, to non-farm activities (Saith, 1992). Rural households depend mostly on farm activities as one of several sources of income. Whether or not a hou sehold will diversify its income sources is dependent upon agricultural pr oduction, market values, and economic status of the region (Lipton and van der Gaag, 1993). Off-farm income opportunities may pull some labor off farm due to low marginal return on intensive farm labor. Non-farm income provides cash for running agricultural activ ities as well as provide for household n eeds, further pulling some labor off the farm. Generally, there almost always remain s farm labor for subsistence food production (Morgan and Solarz, 1993). There is lack of literature on the role of gender on the current trends in vinyungu farming. Reports have shown that, previously, women were predominantly the vinyungu cultivators (Lema, 1996) and that the role of men was limite d to hired labor in land preparation. However, due to the growing economic interest in wetla nd farming, this study will explore whether gender influences some aspects of present-day vinyungu practice, including owne rship, farm size, and types of crops grown. Therefore, agricultural production methods and land use patterns of small-hold householders are potentially influenced by land quality and tenure policies; environmental

PAGE 32

32 factors such as rainfall; available resources such as labor, land, and income; access to markets, and gender roles.. 2.2.5 Decline in Agricultural Production and its Relation to Land Use Many Sub-Saharan African countries face ec onomic stagnation, rapid population growth and natural resource degradation. The economy of these countries suffers from accumulation of foreign debt, weak agricultural growth, d eclining industr ial output, dwindling export performance, declining institutions, and deteriorating socioeconomic and developmental conditions (World Bank Review 1989). These count ries are among the poores t in the world with the farming populations constituting both the ma jority and the poorest segments of society. In an effort to sustain economic growth, ag ricultural development programs have been promoted since the 1960s and agriculture remain s the backbone of most of these countries economy and an important foreign exchange earn er. Unfortunately, agri cultural production in most African countries is declining and f ood deficit is common among these countries. A number of factors, includ ing biophysical conditions, human resources, the economic environment, infrastructure and government policies, contribute to this decline. The soils of most of Africa (as well as those of some parts of Latin America and Asia), for example, show great variability in quality and are, to a great extent, infertile. Wiebe (2003) reports that, although the quality of all land is lowest in the Middle East and North Africa, the quality of cropland is lowest in Sub-Saharan Af rica. Whereas 16% of cropland in Asia, 19% in the Middle East and North Africa, 27% in Latin America, 29% in the developed countries, and over 50% in Eastern Europe are co nsidered to be of the highest quality, only about 6% of SubSaharan Africas cropland is consid ered of high quality (Wiebe, 2003) This factor contributes to high production costs and low productivity.

PAGE 33

33 Global productivity gain s (influenced by technological advancement) have been shown to outweigh global productivity losse s caused by processes such as soil erosion, nutrient depletion, and salinization (Wiebe, 2003). However, this is not the case in many parts of Sub-Saharan Africa where soils are either poor or fragile. In these areas, ag ricultural management practices are not well developed, and, productiv ity levels are already low ev en though the need for growth is high (Morgan and Solarz 1993; Wiebe, 2003). In fact, Sub-Saharan Africa makes the least use of modern agricultural input in the world (Morgan and Sola rz 1993). Scattered populations, common in some parts of Africa, further suppress the development of modern agriculture in that it becomes expensive for the government to support scattered populations and production may be limited by many factors including labor, trans portation costs, and access to knowledge and technologies. In most parts of this region, arable land is fast approaching limits of sustainable agriculture while rainfall remains seasonal and insufficient (Alexandratos 1995). Where rainfall is plentiful, most such areas are already unde r some form of protection hen ce inaccessible for agricultural production (Alexandratos 1995). According to Mo rgan and Solarz (1993), Sub-Saharan Africa has been more adversely affected by droughts th an any other region in the world, especially between 1982 and 1985 and later in the early 1990s. The rate of population growth in Africa is know n to be among the highest in the world. Correspondingly, there is an increasing demand fo r food and agricultural land. In a few cases, population density has been shown to be low or d eclining due to natural cause or out-migration. In such cases, low population densities, coupled with land tenure system that guarantees access to land, encourage extensive farm ing systems with low agricultural input (Morgan and Solarz 1993). Netting (1993) noted about the Koyfar of Nigeria that a decline in population density

PAGE 34

34 leads to dis-intesification of ag ricultural production. Seasonality in labor demand and availability limits farm sizes and productivity of farm system s, affecting land use and the overall agricultural productivity (Morgan and Solarz 1993). As a result of all the aforementioned factors, growth in agricultural production in the SubSaharan African region is depres sed and only in the range of 1.5% to 2% per year (Adams 1996), reducing the amount of agricultural export and fo reign exchange. Per cap ita food production in the 1970s was high enough to meet the food and in come requirements of the average household in the region. Today, almost all southern Af rican countries produce le ss and import more food than they did in the 1970s (Pinstrup-Andersen et al., 1997). Although there has been recovery in several places, agricultural and economic grow th have not been able to keep pace with population growth, posing a great challenge in addressing foo d, social, and environmental insecurity problems (Pinstrup-Andersen et al., 1997). To cope with this situation, more peasants have turned to a comb ination of off-farm income generating activities, wage labor, and subsistence food cropping. Elmekki and Barkers study (1993) noted that, for example, Sudanese peas ant families migrate in search of wage work. The above review has shown that change in agricultural practices and land use in most Sub-Saharan African countries is attributed to a number of fa ctors including population trends and human resources, biophysical factors, land te nure system, and the economic environment of the country. 2.2.6 Wetlands: An Alternative Landscape Component for Crop Production Seasonal rains and recurrent droughts have tr emendously reduced Africas per capita food production, reducing the availabil ity of food and income in th e region (Morgan and Solarz, 1993). Population growth is result ing in land shortages in areas traditionally farmed (Binns, 1994). Lack of adequate rainfall and population gr owth are major factors that have influenced

PAGE 35

35 African countries to turn to wetlands as poten tial resources for overcoming the challenges of poverty and hunger in the region (Adams, 1993a ; Adams, 1993b; Adams, 1996; Dixon and Wood, 2003). The rich soils with high moisture holding capacity and proxi mity to reliable water sources make wetlands attractive for agriculture and ar e thus being converted to agriculture. Because wetland soils can be waterlogged and an aerobic hence unfavorable for plants that are not adapted to such conditions, farmers tend to ra ise the rooting zone a bove the saturated zone. This fact dates many centuries back where we tland resources have been used to support great ancient civilizations (such as those of the Maya and Inca) in the form of raised fields or beds. Perhaps the most widely studied raised fields are the waru waru originally made by preIncan civilizations around Lake Titicaca, borde ring Peru and Bolivia (Binford et al., 1997; Carney et al., 1993; Denevan, 1970; Denevan and Turner, 1974; Erickson, 1985; Erickson, 1988; Erickson, 1999; Ortloff and Kolata, 1993; Turn er, 1994; Turner and Harrison 1981). These are elevated earthen platforms that are 3 to 10 m wide and up to 200 m long, separated from each other by seasonally flooded canals. Through this farming system, wetlands were converted to cultivable land to increase land pr oductivity so as to cope with food security needs of dense human populations that lived in arduous environments. Raised fi elds were advantageous to farmers for they raised the root zone above water-saturated soils; s upplied reliable water; conserved heat and protected crops against frost; retained dissolv ed and particulate nutrients; enhanced nitrogen fixation; and mitigated soil salinity. These factors contributed to higher productivity and more sustainable crop producti on compared to conventional methods of cultivation (Binford et al., 1997; Carney et al., 1993). Raised fields were also common among Nort h American Indians prior the arrival of settlers (Fowler 1969) and in Asia, especially China, in the fifth century B.C. (FAO 1980). In

PAGE 36

36 Africa, raised beds were common in places lik e the Wahgi Valley of New Guinea prior to 350 B.C. (Lampert, 1967) and more recently, among th e Koyfar of Nigeria (Netting, 1968). In most of these cases, raised fields were created not because of population or environmental stress but rather may have developed as early forms of intensive agriculture. In all cases though, it is the many benefits accru ed from raised fields that influenced different societies around the wo rld to develop and maintain th is agricultural technology over centuries. In the recent years, however, the outlook of raised fields is somewhat different in a number of places in that the negative impacts are on the verge of outweighing the benefits. Increased population pressure and associated f ood scarcity in such pl aces as East Africa are driving farmers to drain and intensively cultivate wetlands (Di xon and Wood, 2003; Okeyo, 1992). Roggeri (1995) further supports this obs ervation when he argue d that although the remaining large wetlands are t hose found in the tropical countries, the rate at which they are being converted to non-natura l state is increasing fast. Unlike the Maya and Inca eras that were possibl y sustainable, recent utilization of wetlands is known to cause soil erosion, reduced water storage and quality, variable stream flows and complete dryness of wetlands (Dixon and Wood, 2003; Thompson, 1976). These effects may cause wetland loss, defined as loss of wetland ar ea, due to its conversion to a non-wetland area (Ramsar Convention Bureau, 1990). While changes in land cover by land use do not necessarily imply a degradation of the land (Meyer a nd Turner, 1996:25), wetland loss tends to cause wetland degradation, defined as the impairmen t of wetland functions (Ramsar Convention Bureau, 1990). Intensive agriculture may also impa ir wetland importance as sources of building material and food (mainly fish) suppliers. Howeve r, to most farmers, the immediate benefits accrued through continued conversion of wetlands to agriculture may outweigh the long-term

PAGE 37

37 ecological and socio-economic consequences. As such, although the spa tial extent of wetland conversion is limited by their char acteristic wet conditions that make wetland soils potentially anoxic, and most certainly, heavier and more diffic ult to cultivate than up land soils (McIntire et al., 1992), farmers are known to develop techniques to cope with such situations. Therefore, labor, time, seasons/climate, and technology are the factors that may limit or influence the rate and amount of wetland conversion. 2.2.7 Irrigation Practices and Wetland Use in Tanzania Tanzania continues to rely heavily on the ag ricultural sector for its economic growth and improvement of the welfare of its people. Effo rts have been made to promote agriculture, including improvement of agronomic practices and introduction of im proved crop varieties. However, semi-arid climatic conditions and drought s in some parts of the country present severe constraints on these efforts and agricultural produ ctivity in general. As such, since colonial times, there have been concerted efforts to promot e irrigation practices at various scales (Majule and Mwalyosi, 2003). In Tanzania, irrigation, defined as a supply of water to cultivated plants by means other than natural precipitatio n (Stern, 1989), is highly depende nt on wetlands, e.g., rivers and swamps. The country is estimated to have about 933,000 ha that have potential for irrigation. However, only about 144,000 ha are reported to be under some form of i rrigation (Ministry of Agriculture, 1992). Based on scale and technology used, irrigation systems in Tanzania can be grouped into three major categorie s, i.e., traditional or smallholder irrigation systems, the modern small-scale/village irri gation system, and, medium to la rge-scale state farms/privately owned irrigation estates (Mascarenhas et al ., 1985; Mrema, 1984). Traditional small-scale irrigation systems account for 120,378 ha of the irri gated land while large-scale estate farms account for the remaining 23,622 ha.

PAGE 38

38 The medium-to-large-scale farms or irrigation estates are state or privately owned. They are centrally managed by either parastatals or private companies and generally have quite efficient irrigation systems that also require large capital invest ment and well-trained manpower. High value crops are grown for export and/or loca l consumption using thes e types of estates. The modern small-scale/village irrigation systems were in most cases planned and constructed throughout the country by the central or local governm ent. Farmers were given the responsibility over water distri bution, land preparations, farm decisions, and scheduling of activities. Unfortunately, almost all of them collapsed a few years afte r construction despite a heavy investment in building them (Mkavidanda and Kaswamila, 2001). Two factors are associated with the failure of these irrigation syst ems. First is the unfair di stribution of water that resulted in conflict among farmers who, in some cases, vandalized the fa cilities. Second is the lack of or poor maintenance of the canals, whic h in turn, led to high ra te of evaporation that further hampered distribution of water. The traditional or small-scale irrigation systems are the mo st common. They are owned by individuals or a group of farmers and are usually small in size, of ten not more than 5 ha. Farmers use low cost, temporary or semi-permanent inta ke structures to harness water from rivers, springs and flood plains so as to produce food and cash crops. Under this system, irrigation efficiency is very low as much of the diverted water is lost due to seepage before reaching the farms. This fa rming system is associated with poor drainage that causes siltation, and, poor infrastructure development that leads to unequal distribution of water. Such problems are common in most parts of the country including Kilimanjaro (Banzi et al., 1992), Lushoto (Kaswamila and Tenge, 1997), a nd currently, in Iringa region (Majule and Mwalyosi, 2003).

PAGE 39

39 The most common traditional irrigation system practiced in Iringa is called vinyungu cultivation, a practice believed to have started in the 1890s (Culwick and Culwick, 1935). Vinyungu are ridges or raised beds that are abou t 0.6 m high and 4 to 20 m wide created in valley-bottoms and along streambeds. Smallholder farmers, mostly women, create vinyungu by clearing the land, burning the clear ed vegetation, plowing, creating ridges and ditches, and then by harrowing to smoothen the ridges. Vinyungu are watered by surface run-off (a river or stream originating from the upland areas) as well as subs urface run-off (springs). Since the valleys are waterlogged during the wet season, hampering land preparation and cultivation, vinyungu are cultivated mostly during the dr y season. This farming system is common in Makete, Ludewa, Mufindi, Iringa and Njombe districts of Iringa region and is practiced mainly by the Bena and Hehe tribes of all social classes. Overall, the area under irrigation in Tanzania is relatively small. However, reports indicate that irrigated land in the country is diminishing in quality due to factors such as salinization and sedimentation caused by erosion and perpetrate d by poor management (Majule and Mwalyosi, 2003; Kaswamila and Tenge, 1997). This situation fu rther diminishes the production potential of the remaining irrigable land. Unfortunately, unlike countries such as Uganda and Zambia for instance, Tanzania does not currently have a national policy on wetlands although efforts are underway to produce one. The national water management strategy is also at the developmental stages. This means wetland users currently lack guidance on best management practices over wetland resources. National policies that can provide an enab ling framework within which to improve the status of wetlands sometimes promote actions which conflict in terms of improving the quality of wetlands. For example, policies on agriculture, food securi ty, and poverty reduction promote expansion of

PAGE 40

40 irrigation, in most cases, regardless of where or how it is done or its consequences. Such actions confuse policy implementers, i.e., local government and local resource us ers. Also, overlapping responsibilities between policies are known to cause confusion as to who is to do what and can end in no action in the areas of overlap. In su mmary, both existing and non -existing policies in Tanzania currently make it difficult to ma nage wetlands effectively, and many may be deteriorating. Today, only three wetlands (the Malagarasi -Muyovozi, Lake Natron and the Kilombero Valley) have some form of management follo wing their designation as Ramsar sites, i.e., wetlands of international importance. The remain ing wetlands are utilized for irrigation, among other uses, without guida nce from the government. Due to poor technology, lack of capital, and, hi gh maintenance costs that large irrigation systems require, most rural farmers in Iringa (and Tanzania in general) confine themselves to traditional irrigation practices. 2.3 Summary of the Theoretical Backgr ound and Formulation of Hypotheses 2.3.1 Summary of the Theoretical Background This section summarizes the theoretical b ackground with a focus on land use at the household level and explores important relati onships and variables amenable to land-use changes. Households make daily land-use choices (or, changes in land-use) with respect to their perceived risk and expected financial returns. These choices are influenced by different factors including household character istics, ecological conditions, and economic conditions. Through interviews, these factors were investigated to determine 1) how they influence wetland conversion to agricultura l land; and, 2) how they influen ce transformations in agricultural practices in wetlands. Land-use chan ges eventually lead to land-cove r changes that, in turn, lead

PAGE 41

41 to further land-use changes. GIS and remote se nsing technologies were used to determine the amount of change of the wetland area that is used for agriculture. Analysis of drivers of smallholders fa rming decisions has shown that household characteristics play an important role in land-cover change. In the case of this study, household characteristics may influence the conversion or the non-conversion of wetlands to agriculture. Such characteristics include resour ce availability, i.e. 1) land; 2) the amount of labor available for agriculture (in terms of numb ers and age both of which determine the labor force available for agriculture and hence the amount of land that can be exploited); and 3) the amount of cash available within the household for such purposes as labor hire, renting or purchase of land, and technological improvement (e.g., use of fertilizer). Availability of cash and labor have also been linked to agricultural expansion and intensification (see the anal ysis of population growth and agricultural land-use, section 2.2.3). Under this theorizati on, two hypotheses were tested. Hypothesis 1: Greater number of people within the working age in a household would be associated with greater farm size and number of crops grown; and, H ypothesis 3: Farmers are likely to use household income on farm inputs (i.e fertilizer and labor) so as to improve crop production. Vinyungu -farming practice has, for many years, been predominatly a womans activity that supplemented upland food production. It is not anti cipated, therefore, that gender and duration of residence will have a st rong association with vinyungu size or types of crops grown hence, Hypothesis 2: Farmers gender or duration of reside nce in the village is no t associated with the size of vinyungu farming area or types of crops grown. Ecological conditions, especially poor soil fertility and scanty rainfall, reduce per capita food production (sections 2.2.4 and 2.2.5). A larger population and a high population density

PAGE 42

42 reduce the amount of land available for agriculture (section 2.2.3). Shortage of arable land due to poor ecological conditions or to a larger populati on, usually leads to agricultural intensification, i.e., an increase in total input per unit of land area, or ex tensification, i.e., an in crease in total area cultivated (see, for example, the frontier thesis un der section 2.2.2 and Bilsborow and Ogendo theorization under section 2.2.3). As shortage of land in areas traditionally farmed increases, more and more people are forced into margin al lands. Therefore, wetlands (historically considered unproductive and unhealthy yet with rich soils and moisture-holding capacity) may now be considered potential agricultural lands (sections 2.2.6 and 2.2.7). This transformation may be further facilitated by c onstruction of roads as roads tend to increase human access to areas that were previously unreacheable. Howe ver, active farms require adequate labor and frequent attention, they tend to be closer to human settlements (see Chisholms findings in section 2.2.2). Accordingly, in this research, the following additional hypotheses were tested: Hypothesis 4: Members of households located w ithin a short distance from the wetlands are more likely to practice vinyungu farming system; and, Hypothesis 5: Greater expansion of vinyungu farms would be associated with increased conversion of wetland area to agricultural land. Economic factors play an important role in influencing land-use /cover change. Based on von Thnen theory, a land use activity with the hi ghest value of output is usually found on land closest to the market. On the other hand, transp ortation costs increase with increased distance from the population center (including markets). To minimize production costs and maximize profit, farmers produce high market value-crops near the population center and lower market value-crops (or subsistence crops ) further away from the populati on center. Therefore, distance to roads (including access to tr ansportation) and access to markets do influence land-use/cover

PAGE 43

43 change (section 2.2.2). Based on this theorizat ion, the following hypotheses were tested: Hypothesis 6: Residing close to th e main roads would be associat ed with greater proportion of wetland converted to agriculture; Hypothesis 7: Residing close to a major market would be associated with greater proportion of the wetla nd converted to agricultu re; and, Hypothesis 8: Residing close to major roads and markets would be associated with cu ltivation of a greater proportion of cash crops than food crops. To cope with food demands of growing populations, farmers make several production adjustments so as to improve the quality and pr oductivity of land. Such adjustments include the adoption of land-intensive techno logy, e.g., adoption of new tec hnologies, increased labor inputs per unit of land, and use of natural or artificial fertilizer (section 2.2.3). In this research, the following hypothesis was tested, H ypothesis 9: Households with greater proportion of land for cash-crop cultivation would be associated with increasing costs of farm inputs (labor and fertilizer). 2.3.2 Justification of the Hypotheses This study uses information from a single time-point to assess factors of wetlands conversion in the Ndembera basin. For this r eason, I acknowledge that the causes of wetlands conversion over the course of the past 30 years c ould not be completely determined through this study design. However, there is overwhelming ev idence showing trends of population growth in these areas (and the country in general) dur ing the same time period. Therefore, the basic justification of the hypothese s tested in this thesis is that certain attribut es of population pressure in the wetland areas (for example, proximity to wetland areas, household size, and food requirements) may have contributed to the increasing rate of wetlands conver sions over time. Consequently, I formulated hypotheses to test the current in terrelationships in vinyungu farming as a way of gaining knowledge of the possible factors of wetlands conversions over time. On

PAGE 44

44 their own, the current interrelationships in vinyungu farming are expected to provide an important input in any future policie s aimed at sustainable wetlands use.

PAGE 45

45 CHAPTER 3 STUDY OF FARMER AND HOUS EHOLD FACTORS INFLUENCING VINYUNGU FARMING SYSTEM 3.1 Study Area 3.1.1 Location Ndembera swamp is part of the Ndembera River system (also called Lyandembera) found between Iringa and Mufindi dist ricts of Iringa region, southwes t of Tanzania (Figure 1-2). The entire catchment occupies an area of 1,834.10 km2. The Ndembera river is perennial and therefore of great importanc e, especially during the dry season, to farmers in the immediate environments as well as those further downstream The catchment is adjacent to the TanzaniaZambia (TANZAM) highway, one of the major hi ghways in to the countr y that connects the cities of Dar es Salaam and Mbeya. 3.1.2 Physical Environment Topographically, dissected rolling hills and tangled mountain streams with numerous shallow valleys of tectonic or igin characterize Iringa region. Swamps are formed at valley bottoms where rich organic matter accumulates, creating an ideal environment for agricultural development. The rainfall pattern in the regi on is mono-modal. The annual rain fall totals vary with years but generally range between 500 and 1600 mm. Th e highland zone, mostly between 1,600 and 2,700 m above sea level (asl) on the eastern fringe of Iringa and Mufindi districts, has higher rainfall that ranges between 1,000 and 1,600 mm pe r year. The midland zone, mostly between 1,200 and 1,600 m asl on the majority of Mufindi dist rict, has moderate to low rainfall, ranging between 600 and 1,000 mm per year. The lowland zone, mostly between 900 and 1,200 m asl on the most westerly parts of Iringa and Mufindi di stricts, has low rainfall that ranges between 500 and 600 mm per year. Precipitation, the dominan t factor governing wetland hydrology, is low

PAGE 46

46 during the dry season (June to October) and high during the wet season (Nove mber to May) with highest levels recorded between March and April. It is during the dry periods, i.e., June October, that farmers cultivate the vinyungu The wet period, i.e., November to May, is spent mostly in the upland plots. The annual temperature is typically lower (150 C) in hilly areas (with extremes in June to July) and higher (between 250 C and 300 C) in lowlands (with extremes in December). 3.1.3 Demographic Characteristics In 2002, the human population of Iring a region was 1,490,892 (URT, 2003). The population grew by 24% and 23% fo r Iringa and Mufindi districts respectively, over a period of 14 years between 1988 and 2002. These rates are on a high side when compared to the food production rates in the region th at are falling (URT, 1999). It is estimated that 20% of the population of the entire region lives in the ur ban centers of Iringa, Mafinga, Makambako and Njombe. The remaining 80% of the population lives in rural areas and enga ges in agriculture as their main livelihood activity. 3.1.4 Socioeconomic Situation The major contribution to the economy of Iringa and Mufindi districts comes from agriculture. In Iringa di strict for example, agriculture cont ributes 81.7% to the districts Gross Domestic Product (GDP) while other activities (liv estock keeping, fishing, forestry, mining and trading) collectively co ntribute the remaining 18.3%. Howeve r, much contribution to the 81.7% comes from estates (including tobacco and tea estates). Otherwise, the majority of the community consists of subsistence farmers w ho have a very low income and are generally categorized as poor hence with less cont ribution to the GDP (URT, 2004a and b).

PAGE 47

47 3.1.5 Agriculture Iringa District has a total of 16,607.85 sq. km (or 1,660,800 ha) of arable land of which only 120,612 ha (7.2%) are utilized for cultivation (URT, 1997). In Mufindi district, 133,200 ha (i.e., 19.6% of the arable land) are cultivated. An even smaller percentage of the land suitable for irrigation is irrigated. For example, an estim ated 50,000 ha (or 500 sq. km ) of land are suitable for irrigation farming in Iringa but only 3,812 ha are actually under irriga ted cultivation. The low percentage is attributed to many factors incl uding poor agriculture infrastructure and soil infertility (URT, 1997). Reduced soil fertility as we ll as frequent and prolonged droughts are some of the major factors contributing to reduced pr oduction levels in the region in terms of yiel d per unit area. According to the Danish Development Agency (DANIDA, 1982) droughts occur every after ten years, causing food production to drop by more th an half the normal production levels. Examples include a very low maize productivity in 1996 wher e an average of 2 tons of maize per hectare was harvested against the normal capacity of 6.5 tons per hectare, despite government subsidies (URT, 1999). Availability of water and fertil e soils in wetland areas make them attractive for agriculture. The numerous perennial streams in the region prov ide reliable water for both wet and dry season farming and as such, the government of Tanzania is already promoting irrigation systems in such ecosystems (DANIDA, 1982). While high soil water content of valley bottom swamps may make cultivation difficult, farmers in Iringa region have devised vinyungu to overcome this problem and improve crop production in the region. Over the years, this technology has been transformed to cope with prevailing situati ons such as unreliable rainfall an d increasing soil infertility. The transformation, which includes further expansion of cultivated area, has not been examined to

PAGE 48

48 determine its extent and implications on wetland resources and the overall sustainability of the practice. This study investigated the char acteristics and nature of tr ansformation of this technology and evaluated the environm ental implications of this transf ormation. Iringa and Mufindi Districts were selected for this study for the fo llowing reasons: 1) they represent typical vinyungu cultivation practice in Iri nga; 2) there exist previous studies on the subject in Iri nga district and it was felt to be useful to conduct more resear ch which builds upon the earlier work; 3) the Tanzania-Zambia (TANZAM) highway that connect s major cities of the country (Mbeya and Dar es Salaam), traverses the two Districts and may have implications on both the marketing and transportation of agricultural produ ce as well as on inputs used on the farms; 4) availability of good quality aerial photos of the target area; a nd, 5) accessibility and logistics of sampling. 3.2 Material and Methods In order to test the hypotheses, I examined the current trends in vinyungu -farming system practices; explored the different factors that may be influencing farmers to change their traditional farming system; and, analyzed how changes in this farming system may in turn be influencing changes in the wetland cover in th e study area. The main sources of data were individuals and communities directly involved with vinyungu farming. Formal questionnaires were used to gather information from indivi dual farmers. The research also adopted a Participatory Rural Appraisal (PRA ) approach that has been proved to be effective in eliciting detailed quantitative information where commun ity-related studies are concerned (Chambers, 1994; Brace, 1995). PRA also facilitates discussion among community members as they analyze, investigate, and present their experiences. Village meetings were held to validate the information collected.

PAGE 49

49 3.2.1 Sample Profile and Sampling Procedure The sample population was obtained from nine v illages in Iringa a nd Mufindi Districts, i.e., Usengelindete, Wasa, Muwimbi, Kiponzero, Ih emi, Ilandutwa, and Ifunda in Iringa District and Rungemba and Maduma in Mufindi District. A list of adults known to cultivate/own vinyungu was obtained from village leaders of all villages within the study area. Farmers were di saggregated by sex, village and ability to express oneself as determined by the village leadership. Ability to express oneself was assessed by level of education (mostly those who have attended at least primary school and can read and write) and/or active participation in pr evious village meetings. From a list of 3,000 adult farmers who were eligible for the study, a random sample of six (6) adult farmers from each village was selected for interview, making a total of fiftyfour (54) interviewees (refer to Appendix A for sample size calculation). Information gathered from the study area through village leaders and ward officials showed that of those engaged in vinyungu farming nearly 95% grow both food and cash crops. From this information, I wanted to create a sample of vinyungu farmers with a 90% confidence that it would include farmers who were growing both food and cash crops. Sample size for this confidence interval was 51 farmer s (Appendix A). I chose to select 54 farmers because it was the closest number that gave me equal number of farmers, i.e., six farmers from each of the nine villages. In a ddition to the interviewed farmers, other farmers from the same village as interviewees were invited for group discussions to examine their perceptions of various community issues pertaining to vinyungu farming and use of wetlands (some of the major issues covered in group discussions are li sted in Appendix C). The information gathered from the group discussions was used to s upport the generalizabili ty of the findings.

PAGE 50

50 3.2.2 Data Collection Primary data on vinyungu farming were collected in three main phases and involved both formal and informal survey methods. The first phase involved a formal survey using a predesigned questionnaire (Appendix B). The questi onnaire was used to collect information on vinyungu farming from the 54 representative farmers. During the individual farmers interviews, in terviewees were provided an opportunity to decide whether or not they were willing to participate, to ensure a participatory process. District staff accompanied me to all the interviews and assisted with translat ions where necessary. The questionnaire was read to the interviewees who also had a blank copy for reference. In all cases, the research team, i.e., an interpreter (a District staff) and myself, recorded interviewee responses on the questionnaire. The questionnaire was designed to generate information on the factors driving farmers to cultivate vinyungu ; the role of vinyungu in supporting livelihoods; reasons for current transfor mation in the practice of vinyungu farming; and, on the environmental implications of the farming system (Appendix A). The questionnaire was divided into three sec tions. The first section focused on household information, i.e., general demographics, migrant status and motivations for living in the area, community identity, education level, main activit ies, and society affiliation. The second section focused on land availability, crop preferences, fallow period, and trends in wetland use and status. The last section focused on labor input, tools and problems related to vinyungu farming. The questionnaire has been used to test a ll the nine study hypotheses. For hypothesis 1, Greater number of people with in the working age in a household would be associated with greater farm size and number of crops grown information on age, main occupation, land ownership, land acquisition, land type owned, size of upland plots and vinyungu and vinyungu area cultivated food and cash crops wa s gathered through the questionnaire.

PAGE 51

51 For hypothesis 2, Farmers sex or duration of re sidence in the village is not associated with the size of vinyungu farming area or types of crops gr own, information on the number and type of crops grown and on the proportion of la nd used for both cash and food crops as they relate to sex and duration of resi dence in the village, was sought. For hypothesis 3, Farmers are likely to use household income on farm inputs (i.e., fertilizer and labor) so as to improve crop pr oduction, households were asked how much they spent on fertilizer and labor hire and whether or not both labor and fertilizer were easily available during the peak periods. Farmers were also aske d how much money they accrue from cultivating the wetlands. For hypothesis 4, Members of households locat ed within a short distance from the wetlands are more likely to practice vinyungu farming system, households were asked whether they own, rent, rent-out or borrow land; the si ze of the land, type of tenure and length of ownership; and distance traveled to the farm. They were also asked the location of their plots (upland/wetland); the period of cultivation (wet/d ry season); and the condition of their upland fields in terms of production. For hypothesis 5, Greater expansion of vinyungu farms would be associated with increased conversion of wetland area to agricultural land, farmers were asked about the status of the wetland use (i.e., whether or not wetland utili zation for cultivation has increased); status of the wetland (i.e., whether or not wetland size and vegetation c over has declined since they moved into the area); soil fertil ity; erosion; number of residents (i.e., whether or not there is a population increase); factors that attracted farmers to cultivate vinyungu in this area; and whether or not there were any management plans for the surrounding wetlands. Information on farmers

PAGE 52

52 perception of environmental problems associated with vinyungu farming was also sought. This hypothesis was further tested using GI S and remote sensing technologies. For hypothesis 6, Residing close to the main roads would be associated with greater proportion of wetland converted to agriculture, farmers were aske d of the distance between their plots and the main road. Most farm ers have a clear understanding of the distance they walk to the main road or their farms. Linear regression was used to determine whethe r or not distance to main roads influenced greater conversion of wetlands to agriculture. For hypothesis 7, Residing close to a major market would be associated with greater proportion of the wetland converted to agricult ure, farmers were examined to determine whether or not they develop land for vinyungu farming in relation to the distance to the market. The distance from the market to the center of each villag e was measured as an estimate distance from the farms within that village to the main market. For hypothesis 8, Residing close to major roads and markets would be associated with cultivation of a greater proportion of cash crops than food crops, information on distance from farms to major roads/ markets, types of cr ops grown, and proportion of land occupied by both food and cash crops was sought. For hypothesis 9, Households w ith greater proportion of la nd for cash-crop cultivation would be associated with increa sing costs of farm inputs (labor and fertilizer), information on labor (hired workers) per day, the time of the year the hiring occurs, and tasks performed, was gathered. In addition, information on use of input s and tools in cultivat ion (types and costs of fertilizer, types and number of ag ricultural tools) was also asked. The second phase involved a PRA exercise (C hambers, 1994; Brace, 1995) that brought together all the 54 interviewees to discuss issu es that cut across villag e boundaries, e.g., water

PAGE 53

53 flow and rate downstream. In addition to discussing vinyungu -farming practices in general, information was sought on reasons for transformati ons in the practice as well as on their views about the environmental implicati ons of the practice. Semi struct ured discussions were core to the exercise, but in addition, ranking techniques, Venn diagrams, and pro portional piling (where participants used stones to indicate the relative im portance of different information) were used in gathering information from farmers. Participants recorded key outputs on f lip charts. In addition, all the meetings were tape-recorded and research assistants who also attended the meetings produced the transcripts. Village meetings were conducted in all nine villages to 1) share the outcome of the PRA exercise with all the other farmers; 2) gather farmers views on the outcome of the PRA exercise; and, 3) obtain any additional information and views about vinyungu In trying to determine farmers perception on environmen tal aspects associated with vinyungu farming, four to five meetings were conducted per village, i.e., within four to five haml ets (sub-villages) per village. A total of forty short (one hour each) meetings were held. This was done to allow a more thorough discussion among villagers, in smaller gr oups, of the pertinent environmental issues. Prior to all village meetings, a formal request was sent to the Village Chairmen of all nine villages to call for a village meeting where a formal introduction was made to the village community, the purpose of the research made kno wn, and farmers cooperation in the interviews requested. A proportionate number of adult (>18 y ears old) male and female participants was requested. A total of 2,302 farmers attended the meetings. Assisted by District staff, I moderated the discussions to avoid bias or dominance by influe ntial members of society. The discussions were

PAGE 54

54 guided by a group questionnaire (Appendix C). These m eetings were also tape-recorded and their transcripts were compared to those of the 54 farmers. All activities in the field were preceded by a formal introduction to the village representatives. District offici als introduced the researcher to the village government and study participants as a student collecting information on the past and current vinyungu -farming practices and their relation to recent wetland-cover changes. 3.2.3 Methods of Data Analysis This study explored the extent of vinyungu -farming practice in wetland areas and its perceived effects on the environment. A combina tion of quantitative and qualitative methods was used to meet the objectives of the study. I have shown previously that little is known about the effects of vinyungu -farming system on the environment. For this reason, data gathering was made mainly qualitative for in-d epth interviews and focus groups in order to generate detailed data that are likely to offer the context under which vinyungu -farming practice changes over time. Information gathered from the randomly selected major vinyungu cultivators was mainly quantitative. Methods of analysis are divided into three parts. The three part s of analyses used relevant data to (i) describe the bac kground information of the farmer s, (ii) establish important associations between the farm ers and various aspects of vinyungu -farming practice, and (iii) develop statistical models that can predict use of resources in vinyungu farming. In this chapter, farmers perception on various environmental issues that may be associated with vinyungu farming, was assessed. To assess and document th e magnitude and extent of change of the Ndembera wetland area used for vinyungu cultivation, GIS and remote sensing methods were used (Chapter 4).

PAGE 55

55 3.2.3.1 Descriptive Statistics Within this part of the analysis, importan t demographic features of the farmers who participated in the study are de scribed. Descriptive statistics including means for continuous variables and percentages for categorical variable s are computed for all participating farmers. Qualitative data explaining demographics of the v illagers are also summarized in this part of analysis. 3.2.3.2 Measures of Associations The effect of vinyungu -farming practice on the environm ent was partly assessed by the association between some important farmer or household attributes and the environment, which includes both the infrastructure, e .g., use of farm implements, and fa rm characteristics, e.g., size, and location. In this part of analysis associati on tests were conducted to determine whether there is an association between (i) sex and plot size, (ii) household si ze and plot size, (iii) distance from home to plot and plot size (iv) distance from the plot to major road/market and plot size, (v) number of cash crops grown and plot size, (vi) population size and vinyungu farming area. The Pearsons Chi-square tests evaluated statisti cal significance of these relationships. Important relationships obtained from the qua litative data are also narrated unde r this part of analysis. In addition, prior to multivariable analysis, t-tests were run to compare mean le vels of size of plots (wetlands and uplands), years of vinyungu cultivation, size of vinyungu rent out, travel times, and proportion of vinyungu planted cash/food crops. 3.2.3.3 Multivariable Analysis Detailed analysis of some of the relationshi ps examined in Part II was conducted using multivariable analysis. Multivariable linear regressi on was employed in this section to determine factors affecting plot sizes and types of crops grown (dependent variables) for both wetland and upland farming. This analysis will indicate the im portance of each independent variable (distance

PAGE 56

56 to the plot, distance to the market, ownership, pr oximity to major river, number of household members within working age, sex, farm labor, use of fertil izer, duration of vinyungu farming, duration of residence, and sex) on the size of plot used for vinyungu and other farming practices. For this exploratory study, all th ese factors were entered in the model without using any selection criteria. All statistical procedures in the three parts of analyses were conducted using SPSS for Windows version 13.0 statistical package. The si gnificance level of all te sts was set at p < 0.05. 3.3 Results: Farmer and Household Charac teristics of Valley-Bottom Cultivation ( vinyungu ) around the Ndembera Swamp in Iringa and Mufindi Districts This section presents the major findings of the study on farmer and household factors affecting vinyungu -farming system. The results are presente d in accordance to the three parts of methods of analysis: desc riptive statistics, bivariate relati onships between farmer and household characteristics and factors influencing vinyungufarming system, and the multivariate model of the factors influencing vinyungu farming. Results on farmers perception on the environmental issues that may be related to vinyungu cultivation are also presented here. 3.3.1 Sociodemographic Characteristics The Hehe and Bena are the dominant tribes representing respectivel y 72% and 22% of the total population. Minor tribes are th e Jita, Kinga and Wawanji. The major tribes are from within Iringa region while the minor tr ibes are from other regions in the country, indicating a minor inmigration. The average number of people per household is six with approximately half of them within the productive age of between 18 and 60 years old1. Based on the 2002 population and 1 The International Labor Organization (ILO) recognizes children as those ranging between ages 5-17. The National Youth Development Policy of Tanzania recognizes ages 15-24 as youths (below which are children). Age 18 is considered mature and able to vote while 60 is a pensi onable retirement age. For the pu rpose of this study, I used working ages and not age distribution in the general population in my categorization. I have categorized ages 0-17 as children (potential labor force); ages 18-35 as youths (and the major labor force that is also dynamic and can own

PAGE 57

57 housing census, the population distribution in the study villages st ands at 20,695, i.e., 9,863 people (48%) between 0 and 14 years old; 5,880 pe ople (28%) between 15 and 34 years old; 3,493 people (17%) between 35 and 59 years old; and 1,459 people (7%) at 60 and above years old (URT, 2003). The major labor force lies betw een ages 15 and 59 which constitutes 45% of the population. However, at the village level, 10 and 14 year olds (15%) are also involved in some minor farm activities hence raising the farm labor force to 60% of th e entire population in the study villages. Table 3-1 below summarizes the sociodem ographic characteristics of the major vinyungu farmers who participated in the survey. The results show that 63% of the farmers were men and about 91% were married. The higher number of men in the sample may have been influenced by the sampling methods where preference was given to a representative farmer that owned a kinyungu and who also had a proven abil ity to express oneself eloquen tly in the previous village meetings. There was no significant age differenc e between men and women sampled. A majority of the farmers had primary education and, as exp ected, almost all of them (96.3%) were engaged in agriculture as their main occupation. The high proportion of reside nts with only primary education is typical of rural areas in Tanzania, where agricu ltural activities are a major occupation that starts at an early age and trad itionally require no formal qualifications. Only a small proportion of the farmers (27.8%) are member s of farmers coopera tive unions, most of which have been established only recently, i.e., in the early 1990s following a boom of paprika as a cash crop. Cooperative unions prevalent in the 1970s became defunct in the late 1980s mainly because most farmers felt exploited by them and also because they made more loss than they did profit. land); ages 36-60 as adults (and an important work force, perharps more se ttled); and, ages 61 and above as the ederly (and not a major labor force).

PAGE 58

58 3.3.1.1 Land tenure/ownership With regards to land ownership, Table 3-1 show s that almost all the farmers use their own land for farming activities. While most of the farmers (79.6%) acqui red their land through inheritance, about 15% bought th e land they currently own. The rest (5.6%) have acquired land through the village government. Through group discussions, farmers indicated that individuals from the same village who are not landowners or who want to incr ease productivity, might request access to vinyungu owned by colleagues or relatives both within and outside their villag e, especially from landowners that are unable to cult ivate their entire prope rty due to lack of human or financial resources. This pattern however, is becomi ng rare while renting is slowly increasing. Information gathered through group di scussions also indicated that vinyungu ownership among households differed in numbers from one kinyungu per household to as many as ten vinyungu per household and that these vinyungu are mostly within the same valley the vinyungu farmers live in. A few vinyungu may be scattered over several valleys, a factor driven by the size of the valley and accessibility. The data in Table 3-1 also show, as expected among farmers in this area, that typically farmers in the study area would have both vinyungu and upland plots for th eir farming activities. Group discussions confirmed that villagers may e ngage in other activities but all of them did indeed cultivate the uplands, vinyungu or both. However, there is a substantial difference in the size of land owned for these types of plots. Nearly 70% of the fa rmers own at least three acres (1.2 ha) of upland plots compared to onl y about 11% who own the same acreage for vinyungu plots.

PAGE 59

59 3.3.1.2 Crop production, crop preference and land use In group discussions in all nine villages farmers reported that vinyungu farms have become extremely important for both food and cash crop production especially when compared to upland plots. For example, upland production of mai ze as the main food and cash crop has declined from an average of 30 bags/ha in the 1970s/80s to about 15 bags/ha in the 1990s/00s, even after application of inorganic fertiliz er (Table 3-2). Farmers attributed this decline to a number of factors including droughts, poor so ils, lack of extension servi ces and removal of government subsidies that were key in providing adequate fertilizer and modern fa rm implements. On the other hand, maize production in the vinyungu has increased from an average of 23 bags/ha in the 1970s/80s to an average of 33 bags/h a in the 1990s/00s (Table 3-2). The most common food crops grown on vinyungu include maize, potatoes, beans, and cassava. Maize, potatoes, beans, and vegetables such as peas, tomatoes, onions, and green pepper are important sources of income. Farmers reported that, in most cases, vinyungu are cultivated at least twice a year, mostly duri ng the dry season when market prices are also higher hence maximizing profit (Table 3-3). Maize, potatoes, beans, and so me vegetables are sown around August/ September and harvested in November/ D ecember (potatoes, beans, and vegetables) and later around January/February (mai ze and vegetables). The second round of planting takes place in March/April and harvesting takes place in June / July. Vegetables may be sown and harvested thrice a year, depending on water availability. By juggling upla nd and bottomland cultivation, farmers are able to produce enough food yea r-round as well as increase family income. Table 3-1 shows that there ar e differences in the use of vinyungu plots with respect to the types of crops grown. A larger proportion of the farmers (83.4% vs. 52.0%) use at least half the size of their vinyungu plots for growing food crops compared to cash crops.

PAGE 60

60 3.3.1.3 Inputs on vinyungu Data from Table 3-1 showed that on average, farmers used Tshs 22,245,103 (approximately U.S. $22) per bag of inorgani c fertilizer per growing season per household. At least 87% of the farmers interviewed use inor ganic fertilizer despite the high costs and limited availability. Most farmers buy between 1 and 3 bags per season for both upland and lowland farms and use about half to 1 bag of inorganic fer tilizer in their vinyungu The individual farmers that were interviewed attributed the use of fertilizer on vinyungu plots in the recent years (1990s 2005) to a decline in soil fertility. While several sources of labor exist (e.g., hire d labor, neighbors, friends ), farmers reported family members (i.e., husband, wife, children, de pendents) as the most commonly used labor. Neighbors or friends may be requested to help in cases where family labor is not adequate to perform the task. The other alternative is hired labor. Data from Table 3-1 show that the mean cost of hired labor is Tshs 9,983 3,975 (U.S $10) per acre per hous ehold. Although most farmers consider this rate high, at least 52% hire labor for th e most laborious task of land preparation and rely on family labor (or friends ) for other farm activi ties such as planting, weeding and harvesting. Tables 4-7 show the results of Part II of the bivariate analysis that explored the association between vinyungu farming practice and factors of influe nce that included sex, age, years of residence in the village, and num ber of potential workers (aged 18 to 60 years) in the household. Vinyungu farming was assessed as (i) size of vinyungu plots owned by household, (ii) size of vinyungu plots grown in the dry season, (i ii) number of cash crops grown on vinyungu plots, and (iv) size of vinyungu plots grown cash crops. The results in Table 3-4 show that vinyungu -farming practice is associ ated with the number of working household members and age of the fa rmer. Specifically, the data indicated that

PAGE 61

61 number of household members who can wo rk is associated with the size of vinyungu Almost all households with more than three working memb ers had large farming lands and, in addition, over 54.8% such households had pl ots covering over an acre (> 0.4 ha) of land compared to 45.2% of those households with at most three members. Also, fa rmers in the 36-to-60-year-old age group were more likely to have large vinyungu plots compared to other age groups. The results in Table 3-7 showed that age is associated with the size of vinyungu plots cultivated with cash crops ( 2 = 4.432; p = 0.039). Here, farmers belonging to all age groups tend to use less than 50% of their plots to grow cash crops. None of the selected factors of infl uence was related to either size of vinyungu plots during the dry season (Table 3-5) or number of cash crop s grown (Table 3-6) at the bivariate level. Overall, there was no difference between men and women in vinyungu -farming practices. That is, men and women had simila r patterns of land-use for vinyungu farming system in the study area at this level of analysis. 3.3.1.4 Multivariable analysis In this part of analysis a multivariable re gression model was used to assess factors affecting vinyungu -farming system. For this purpose, a general linear multivariable regression model was fitted to the data collected from major vinyungu cultivators. Four variables, overall size of vinyungu plots, size of vinyungu plots during the dry season, number of cash crops, and size of vinyungu plots grown cash crops were selected as dependent variables describing vinyungu -farming system. Four models, one for e ach dependent variable, were fitted to determine how farmer and household characteris tics influenced the dependent variables. The independent variables included se x, distance from plot to major market, distance from plot to major road, distance from house to plot, cost of hired labor, cost of fert ilizer, number of workers in the household and years lived in the village. Number of cas h crops (described above as

PAGE 62

62 dependent variable) was also used as independent variable in the model for overall size of vinyungu and size of vinyungu plots in dry season. The problem of multicollinearity, i.e., high correlations among the variables used in the model, was addressed prior to fitting of each of the four models to avoid biased regression estimates. The results of the linear regression models are presented in Table 3-8. Results from Model I (overall size of vinyungu ) show that the constant term is positive and significant. This means regardless of the independent variab les (or predictors) I have selected for this model, a farmer is expected to have vinyungu plots with an average size of 1.558 ha. The model also showed that the number of working persons in the hous ehold is related to the overall size of vinyungu plots the household owns. Specifically, the data indicate that for each additional person of working age group in a household, the size of vinyungu plot increases by 0.76 ha (p=0.029). The relationship between distance from vinyungu plots to both major market and main road appeared to have the hypothesized direction, but did not attain st atistical signifi cance. Also, sex, costs of labor and fertilizer, dur ation of residence in the village and number of cash crops were not related to the overall size of vinyungu plots owned by the farmers. Lastly, the R-Square for model I was 0.342. This indicates that 34.2% of the variation in the size of vinyungu plots can be attributed to the variations in the pred ictors selected to fit this model. Model II was fitted to assess factors influencing size of vinyungu plots during the dry season. My results revealed that the number of persons in a work ing group per household as well as duration of residence in the village were associated with the size of vinyungu plots grown during the dry season. On average, each additio nal person in the working group increases the size of plots by 0.251 ha per household (p=0.024). Comb ined with the results of Model I, these results support hypothesis 1, i.e., greater numbe r of persons in a working age group in a

PAGE 63

63 household is associated with greater size of the cultivated vinyungu plots. However, hypotheses 6 and 7 (i.e., residing close to the main roads and major market respectively, would be associated with greater proportion of wetland converted to agriculture) are not supported. The results show that residents with fewer years of residence in the villages are more likely to have larger sizes of vinyungu plots in the dry season. Model III assesses predictors of the number of cash crops grown in the vinyungu plots. I found that distance from the farmers house to the plot is associated with the number of cash crops grown. On the average, each additional kilometer of the distance between the house and the plot is associated with a reduction of about 0.2 (or one out of five) possible cash crops that can be grown in the vinyungu plots. In this model, the selected predictors explained only 28% of the variation in the number of cash crops grown in the plots. That is, the data had the poorest fit on Model III. Lastly, Model IV was used to analyze factors of the proportion of vinyungu plots for growing cash crops. The constant te rm indicates that when all predictors have no effect in the model (i.e., when all coefficients are zero), we would expect that, on average, farmers would have 61% of their vinyungu plots grown cash crops. Also, resu lts show that sex and duration of residence in the village were associated with th e size of plots used to grow cash crops. Men had 16.5% more of their vinyungu plots grown cash crops compared to women. This finding indicates that men, who were previously not involved in vinyungu cultivation except where hired by women to prepare land for cultivation (Lem a, 1996; Masija, 1993), are now increasingly getting involved in this farming system and that they have interest in expanding their plots to grow cash crops. Similar to the results of M odel II, I found in Model IV that duration of residence was associated to the size of vinyungu plots for growing cash crops. Those residents

PAGE 64

64 with fewer years of residence in the villages were more likely to have larger proportions of their vinyungu plots used to grow cash crops. All dist ance factors had the expected direction of relationship with the size of vinyungu for growing cash crops, but were not statistically significant. Lastly, Table 3-8 shows that age, years of resi dence in the village, use of fertilizer have no effect on the households engagement in vinyungu -farming practice. 3.3.2 Environmental Issues Related to Vinyungu Cultivation This section presents farmers perceptions on different environmental issues as discussed during the village meetings (Table 3-9) and thro ugh individual farmers interviews (Table 3-10). Descriptive statistics have b een used to analyze the data. Table 3-9 shows that land availability is not a problem in the area. Instead, farmers decision to cultivate vinyungu is driven by the availability of water (75%) and soil fertility (20%). Figures 4 and 5 show the differ ent sources of water used to make vinyungu cultivation possible. Data in Table 3-9 also show th at the use of the wetlands for vinyungu cultivation has increased when compared to the 1970s (97.5% of the responses). This is further supported by data obtained through individual farmers inte rview whereby 63% of the farmers reported a tremendous increase in the use of wetlands for vinyungu cultivation (Table 3-10). While wetland size and soil fertility were reporte d to be declining, soil erosion was reported to be increasing as a result of the expansion of vinyungu cultivation over the years. Farmers perceive reduced water flow as the biggest problem resulting from the cultivation of vinyungu followed by reduced soil fertility, increased soil erosion, and drying of th e wetland (Tables 9 and 10). At least 68% of the responses associated soil erosion with the way vinyungu are cultivated, i.e., close to riverbanks and catchment areas as well as massive cl earance of wetland vegetation such as Cyperus papyrus

PAGE 65

65 (madete) and some important catchment vegetation such as Syzygium cordatum (mivengi). Through village discussions and individual farmers interviews, it was appare nt that availability of water and soil fertility are influencing farmers to clear more wetlands to create new vinyungu At least 93% of the respondent s reported an increased numbe r of households in the study area. This increase may also be playing a part in influencing increased co nversion of wetlands to agriculture. Figure 3-3 shows Ndembera river only 1.5 me ters behind the wetland vegetation, farmers are cultivating vinyungu and they are expanding their vinyungu towards the river. Farmers also reported that human conflict over use of wetland resources (land, water) was not observed in the 1970s but is now growing. Alt hough education on sustainable practices has been provided and land use plans developed in many villag es, farmers reported this to be a very recent (2004/5) undertaking. 3.3.3 Discussion This chapter examined the current trends in vinyungu -farming system practices and explored the different fa rmer and household factors that may be influencing farmers to change their traditional farming system as well as the implications of the new developments on wetland resources. Most farmers in the study area would have both vinyungu and upland plots for their farming activities. However, the size of upland pl ots owned by farmers tends to be larger than that of vinyungu Land is acquired mostly through inheritanc e and is usually in close proximity to the villages where farmers live. Therefore, members of house holds located within a short distance from the wetlands do cultivate vinyungu Although a few farmers purchased their farmland, this trend is not common practice. Farmers decision to cultivate vinyungu and increase the size of their plots seems to be partly driven by the availa bility of water, soil fertility and increased production. Due to declining

PAGE 66

66 soil fertility in the vinyungu however, most farmers (87.5%) us e inorganic fertilizer. This supports Hypothesis 3, i.e., Farmers are likely to use household income on farm inputs (i.e., fertilizer and labor) so as to improve crop production. Family members, irrespective of sex, are the main source of labor in the construction of vinyungu The results in this section support Hypothes is 1. It has been obser ved that the greater the number of people within a working age gr oup in a household, the greater the size of vinyungu the household is likely to own. Hire d labor is considered expensive and is therefore rarely used except for a tough task of land clearing. The negative regression coefficients on dist ance factors values indicate that close proximity to both markets and roads is likely to influence an increase in both the size of vinyungu and the proportion of vinyungu cultivated cash crops (Hypotheses 6,7, and 8). Although these relationships were not statistically significant, they give an indication of the validity of the hypothesized relationships. Sex, costs of labor and fertiliz er, duration of residence in th e village and number of cash crops were not related to the overall size of vinyungu plots owned by the farmers as was postulated in Hypothesis 2, i.e., farmers sex or duration of residence in the village is not associated with the size of vinyungu farming area or types of crops grown. Sex, age, years of residence, and number of pe ople that can work did not seem to influence the size of vinyungu plots cultivated during the dry season or number of cash crops grown, at the bivariate level. However, at th e multivariate level, the number of persons in a working age group per household and duration of residence in the village were associated with the size of vinyungu plots grown during the dry season. That is, households with a great er number of people within a working age group are more like ly to have more area in vinyungu during the dry season

PAGE 67

67 (p=0.024) and residents with fewer years of residence in the villag es are more likely to have larger sizes of vinyungu plots in the dry season (p=0.048). An important implication of this finding could be the movement of farmers coming to settle into the wetland areas. These farmers, who are likely to migrate from more arid areas seem to have tendenc y to cultivate larger vinyungu plots during the dry season. These farmers tend to buy the land they cultivate and a few rent. There was no consistent differe nce between men and women in vinyungu -farming practices. That is, men and women had similar patterns of land-use for vinyungu -farming system. The only gender difference was obs erved in the proportion of land used to grow cash crops. Men had a higher percentage of their land used to grow cash crops. This is an indication of a new tradition in vinyungu i.e., having men more engaged with vinyungu farming and more specifically in cultiv ation of cash crops. Distance from the farmers house to the plot is inversely related to the number of cash crops grown, i.e., the further the plot is from the house the less the cash crops that can be grown in the vinyungu I found out that age, duration of residence, and use of fertilizer were not consistent factors of vinyungu farming. With regards to age, I can expl ain that no differences occur in various aspects of vinyungu farming because persons of all ages tend to be farmers once they reach young adulthood. In this way, no specific age group may appear to have unique influence on vinyungu farming. The effect of duration of residence appeared to be limited to the growth of cash crops in the wetlands. Thus, its effect was not detected in other areas, e.g., overall vinyungu area and

PAGE 68

68 percentage of area cultivated in the dry season. This implies migration to the wetlands is mainly influenced by cash crop production. Use of fertilizer appeared to have no influence on vinyungu -farming practice probably because all farmers are relatively at the same level of production, i.e., there is no difference in resources that enable farmers to acquire farm inputs. Increased soil erosion and reduced water flow soil fertility and wetland size/ vegetation are associated with both the expansion and methods of vinyungu cultivation. These results partly support Hypothesis 5 which suggest ed that greater expansion of vinyungu farms would be associated with increased conversion of wetland area to agricultural land. Parallel to increasing environmental problems is growing human c onflict over use of wetland resources, as was reported by farmers.

PAGE 69

69 Table 3-1. Sociodemographic and Land Use Characteristics of the Sample Farmers in the study area (n=54) Variable Number Percent Sex Male Female 34 20 63.0 37.0 Age 18 35 years 36 60 years 61+ years 21 26 7 39.0 48.0 13.0 Marital Status Married 49 90.7 Education Primary education Primary and training Secondary education 41 8 5 76.0 14.8 9.2 Ethnic Distribution Bena Hehe Jita Kinga Wawanji 12 39 1 1 1 22.2 72.2 1.9 1.9 1.9 Age distribution by household >18 years old (median = 3) None 1-2 people 3-4 people 5-6 people 7+ people 18-60 years old (median = 3) None 1-2 people 3-4 people 5-6 people >6 people >60 years old (median = 0) None 1 person 2 people 7 7 23 13 4 3 22 27 1 1 47 5 2 13 13 42.6 24.1 7.4 5.6 40.8 50.0 1.9 1.9 87.0 9.3 3.7 Main Occupation Agriculture Small business (e.g., restaurant, local brew) 52 2 96.3 3.7 Cooperative Membership Current member 15 27.8

PAGE 70

70 Table 3-1 Continued Variable Number Percent Land Ownership Own land 53 98.1 Land Acquisition Bought Inherited Given by the village government 8 43 3 14.8 79.6 5.6 Land Type Owned and Used Both vinyungu and upland areas 52 96.3 Size of Upland Plots Less than an acre (<0.4 ha) 1 2 acres (0.4 0.8 ha) 3 4 acres (1.2 1.6 ha) Over 4 acres (>1.6 Ha) 2 15 20 17 3.7 27.8 37.0 31.5 Size of Vinyungu Plots Less than an acre (<0.4 ha) 1 2 acres (0.4 0.8 ha) 3 4 acres (1.2 1.6 ha) Over 4 acres (>1.6 Ha) 27 21 2 4 50.0 38.9 3.7 7.4 Proportion of Vinyungu Area for Food Crops Less than 50% About 50% Over 50% 9 21 24 16.6 39.0 44.4 Proportion of Vinyungu Area for Cash Crops Less than 50% About 50% Over 50% 26 21 7 48.0 39.0 13.0 Cost of fertilizer/bag /growing season/HH (87% users ) Mean SD TShs. 22,245,103 ($22 13) Cost of labor/acre/HH (52% hire labor) Mean SD TShs. 9,983,975 ($10) Important cash crops include maize, potatoes, be ans and vegetables such as tomatoes, onions and peas. Fertlizer commonly used are Urea and CAN (top dressing) and DAP and TSP (during planting). In this study, nutrient amounts were not calculated.

PAGE 71

71 Table 3-2. Maize yiel ds in uplands and vinyungu between 1970s and 2000s (village responses, n=9) (per year, i.e., one season in the uplands and two seasons in the vinyungu ) Maize yields in the uplands Maize yields in the vinyungu Bags/ha Mean+ SD 1970s /80s 30+ 6 1990s/00s 15+ 7.8 1970s/80s 23+ 11.5 1990s/00s 33+ 9.3 <10 bags 0 response 3 res ponses 0 response 0 response 10-19 bags 0 response 3 respons es 6 responses 0 response 20-29 bags 4 responses 3 res ponses 3 responses 0 response 30-39 bags 5 responses 0 res ponse 0 response 7 responses # of bags/ha farmers produced between 1970-2005 40 bags 0 response 0 respons e 0 response 2 responses Farmers expressed yields in bags/acre. Their estimates ha ve been converted to bags/ha (1 bag=100kg when filled with maize grains) Table 3-3. Potential market price per unit output from cash crops grown in vinyungu and under rain fed agriculture Crop Growing season (No. of seasons/ year) Average price (TShs/bag or tenga) Average income (TShs/year) Rain fed Vinyungu Rain fed Vinyungu Rain fed Vinyungu Maize (bag-1) Shelled and dried 1 2 28,000 36,000 28,000 72,000 Beans (bag-1) Shelled and dried 1 2 54,000 54,000 54,000 108,000 Potatoes (bag-1) 1 2 25,000 30,000 25,000 60,000 Peas (bag-1) In pods, not dried 1 2 20,000 15,000 20,000 30,000 Onion (bag-1) 1 2 40,000 50,000 40,000 100,000 Tomatoes (tenga-1) 1 3 3,000 5,000 3,000 15,000 Cabbage (bag-1) 1 2 4,000 5,000 4,000 10,000 1 bag = 100 kg when filled with dry maize grain; 1 tenga = a tin filled with 16 kg of dry maize grain; Tanzanian Shilling (TShs) 1,000 = $ 1

PAGE 72

72 Table 3-4. Results of Chi-Square Analysis of the Farmer and Household Factors influencing Vinyungu -Farming Practice, i.e., size of vinyungu plots owned by household in relation to sex, age, years of residence in the village, and number of potential workers Factors of Influence Vinyungu size < 1 acre (< 0.4 ha) Vinyungu size 1 acre ( 0.4 ha) 2 Sex Female Male 10 (43.5) 13 (56.5) 10 (32.3) 21 (67.7) 0.713 0.287 Age 18-35 years 36-60 years 61 years or older 13 (56.5) 9 (39.1) 1 (4.3) 8 (25.8) 17 (54.8) 6 (19.4) 6.174 0.046 Years of residence 25 years > 25 years 9 (39.1) 14 (60.9) 12 (38.7) 19 (61.3) 0.001 0.598 Number of workers 3 workers > 3 workers 22 (95.7) 1 (4.3) 14 (45.2) 17 (54.8) 10.963 0.001 Table 3-5. Results of Chi-Square Analysis of the Farmer and Household Factors influencing Vinyungu -Farming Practice, i.e., size of vinyungu plots grown in the dry season in relation to sex, age, years of residence in the village, and number of potential workers Factors of influence Vinyungu size dry season < 2 acres (< 0.8 ha) Vinyungu size dry season 2 acres ( 0.8 ha) 2 Sex Female Male 17 (36.2) 30 (63.8) 3 (42.9) 4 (57.1) 0.117 0.519 Age 18-35 years 36-60 years 61 years or older 18 (38.3) 24 (51.1) 5 (10.6) 3 (42.9) 2 (28.6) 2 (28.6) 2.184 0.336 Years of residence 25 years > 25 years 17 (36.2) 30 (63.8) 4 (57.1) 3 (42.9) 1.128 0.256 Number of workers 3 workers > 3 workers 34 (72.3) 13 (27.7) 5 (71.4) 2 (28.6) 0.003 0.637

PAGE 73

73 Table 3-6. Results of Chi-Square Analysis of the Farmer and Household Factors influencing Vinyungu -Farming Practice, i.e., number of cash crops grown on vinyungu plots in relation to sex, age, years of residence in the village, and number of potential workers Factors of influence No. of cash crops 3 No. of cash crops > 3 2 Sex Female Male 10 (30.3) 23 (69.7) 10 (47.6) 11 (52.4) 1.650 0.160 Age 18-35 years 36-60 years 61 years or older 15 (45.5) 14 (42.4) 4 (12.1) 6 (28.6) 12 (57.1) 3 (14.3) 1.564 0.457 Years of residence 25 years > 25 years 13 (39.4) 20 (60.6) 8 (38.1) 13 (61.9) 0.009 0.577 Number of workers 3 workers > 3 workers 24 (72.7) 9 (27.3) 15 (71.4) 6 (28.6) 0.011 0.578 Table 3-7. Results of Chi-Square Analysis of the Farmer and Household Factors influencing Vinyungu -Farming Practice, i.e., size of vinyungu plots grown cash crops in relation to sex, age, years of residence in the v illage, and number of potential workers Factors of influence Vinyungu area planted cash crops 50% Vinyungu area planted cash crops > 50% 2 Sex Female Male 18 (37.5) 30 (62.5) 2 (33.3) 4 (66.7) 0.040 0.609 Age 18-35 years 36-60 years 61 years or older 21 (43.8) 21 (43.8) 6 (12.5) 0 (0.0) 5 (83.30) 1 (16.7) 4.432 0.039 Years of residence 25 years > 25 years 18 (37.5) 30 (62.5) 3 (50.0) 3 (50.0) 0.351 0.431 Number of workers 3 workers > 3 workers 35 (72.9) 13 (27.1) 4 (66.7) 2 (33.7) 0.104 0.539

PAGE 74

74Table 3-8. Results of the Linear Regre ssion Analysis of the Factors Affecting Vinyungu -Farming Practice Variable Size of Vinyungu Plot (I) Size of Vinyungu in Dry Season (II) Number of Cash Crops (III) Size of Vinyungu for Cash Crops (IV) Coefficient Coefficient Coefficient Coefficient Constant 1.558 0.016 0.231 0.248 4.369 0.007 61.013 0.001 Sex 0.218 0.563 -0.001 0.993 -0.478 0.505 16.492 0.031 Distance to major market (km) -0.002 0.878 0.002 0.532 0.003 0.878 -0.032 0.886 Distance from plot to main road (km) -0.010 0.664 0.000 0.987 0.067 0.136 -0.254 0.547 Distance from house to plot (km) 0.001 0.979 NIa NIa -0.219 0.031 -0.058 0.953 Labor cost ($) -0.001 0.564 0.000 0.646 0.000 0.776 0.000 0.722 Cost of fertilizer ($) -0.001 0.396 0.000 0.876 0.000 0.765 0.000 0.088 No. of working household members (#) 0.760 0.029 0.251 0.024 0.169 0.565 -1.768 0.778 No. of years lived in the village (#) 0.002 0.844 -0.008 0.048 -0.036 0.155 -0.705 0.010 No. of cash crops (#) 0.029 0.944 -0.031 0.355 NIa NIa NIa NIa R-Square 0.342 0.353 0.284 0.383 a Indicates variable Not Included in the model due to multicollinearity

PAGE 75

75 Table 3-9. Farmers perception on various enviro nmental issues as discussed in the village meetings: the current state compared to the 1970s (n = 40) Variable Responses Percent Use of the wetland Has increased 39 97.5 Wetland/wetland vegetation cover Decreased/declined 36 90 Soil fertility Soil fertility has increased 5 12.5 Soil fertility has declined 33 82.5 Soil erosion No erosion 0.5 12.5 Some erosion 27 67.5 Great erosion 6 15 Motivation to cultivate vinyungu in their area Lack of alternative land 2 5 Rich soil fertility 8 20 Availability of water 30 75 Vinyungu cause environmental problems Yes 37 92.5 The biggest poblem associated with vinyungu cultivation Reduced water flow 23 57.5 Reduced soil fertility 7 17.5 Soil erosion 8 20 Human conflicts 2 5 Management plans for wetland resources Exist 25 62.5 Do not exist 15 37.5 Education on sustainable practices Has been provided 28 70 Has not been provided 12 30 Suggestion to the Government on the future of vinyungu cultivation Vinyungu cultivation be halted/ banned 6 15 Sustainable/ best management practices be promoted 34 85

PAGE 76

76 Figure 3-1. A well is dug where water is not free flowing. Picture taken in Lumuli village in Iringa District, June 2005 Figure 3-2. A river is diverted to supply vinyungu with water. Pi cture taken in Usengelindete village, Iringa District, June 2005

PAGE 77

77 Figure 3-3. Picture showing Nd embera river behind which vinyungu cultivation is taking place in Maduma village, Mufindi District. An ex ample of how close to the river the vinyungu are getting. Picture ta ken in June 2005 N dembera rive r Wetland vegetation Wetlan d vegetation Vinyungu cultivation taking p lace just 1.5 meters behind the wetland vegetation. Clearing of vegetation was ongoing a t the time this photo was taken

PAGE 78

78 Table 3-10. Individual farmers perception on various environmen tal issues (n=54) Variable Responses Percent Upland plots for crop production Fairly productive Not very productive 37 17 68.5 31.5 Wetland plots for crop production Very productive Fairly productive 28 26 51.9 48.1 Factors That influenced farmers to own or rent vinyungu They are cheap, easy to get Close proximity from home High soil fertility Excellent dry season alternative (availability of water) 4 4 30 16 7.4 7.4 55.6 29.6 Trend in the utilization of wetlands for vinyungu cultivation Use has increased tremendously Use has increased slightly Use has remained the same Use has declined 34 15 1 4 63.0 27.8 1.9 7.4 Trend in the number of households Number of households has increased Number of households has decreased Number of households has remained the same 50 1 3 92.6 1.9 5.6 Environmental problems associated with vinyungu cultivation Soil erosion Sedimentation Reduced water flow Drying of the wetland Reduced wetland size Flooding Erosion of river banks None 4 4 26 6 4 4 4 2 7.4 7.4 48.1 11.1 7.4 7.4 7.4 3.7

PAGE 79

79 CHAPTER 4 USE OF GEOGRAPHIC INFORMATION SYSTEMS AND REMOTE SENSING TO DETERMINE THE MAGNITUDE AND RATE OF CONVERSION OF NDEMBERA SWAMP AS A RESULT OF VINYUNGU EXPANSION In the previous section of the thesis I explored changes in vinyungu cultivation, factors influencing the changes, and environmental imp lications as perceived by interviewed farmers. According to the farmers, vinyungu cultivation has increased ove r time. However, no study has been conducted to determine the extent and rate of wetlands conve rsion to agriculture. Geographic Information Systems (GIS) and Remote Sensing technologies are useful analytical tools for obtaining information on la nd-use/cover change (LUCC), including where and when LUCC occurs and the rates at which th ey occur (Lambin et al ., 1999; Turner et al., 1993). GIS are a computer-based systems of retrieving, storing, manipulating, updating and mapping spatially referenced data (Jones 1997). Remo te Sensing is a method of collecting spatial data using remote sensors (i.e., not in direct contact with the target of interest) such as satellites and aerial photography (Jensen 1996). Remote sensi ng and GIS are often integrated and used to analyze ecosystems on multiple scales with both sp atial and temporal factors. These technologies have been used to map and document changes in wetland use and cover (B urgi and Turner, 2002; Jensen et al., 1995; Munyati, 2000; Wang et al., 2006). Such studies have also been able to utilize these technologies to assess further the factors influencing the observed changes. Since field studies can be time consuming, expensiv e and cover only a small area, remote sensing offers a potentially cost-effective way to study ecological changes over time. I used remote sensing and GIS to determine the land-use and land-cover changes associated with agricultural activities around Ndembera swamp. My objectives were 1) to determine the magnitude of wetland change be tween 1977 and -1999; and, 2) to determine

PAGE 80

80 whether or not there exists a relationship betw een agricultural land expa nsion and a decline in wetland area. 4.1 Data Sets Two datasets from two different periods ( 1977 and 1999) were used to analyze land-use/ cover changes. Both datasets are based on un -rectified aerial photogra phs at the scale of 1:50,000. 4.2 Methods Photographs were interpreted, using a Topc on Stereoscope. Different land cover types were delineated on the set of un -rectified aerial photographs. Each land cover was symbolized at its location as one land use class or polygon, e.g., Ag represente d agricultural land. Graphical radial triangulation method (Schwidefsky, 1959) was used to control the horizontal scale (vertical scale was controlled during the digitization process). C ontrol points of common areas were marked on all maps including the 1:50,000 topographic maps (i.e., the base control map). Each photograph had a point that was stereosc opically located. By using these stereo located points, differences of scale between the first photograph to the second, third, fourth and so on were controlled. Transparen t overlays were used to transf er the interpreted land covers from aerial photographs to the paper. The pieces of transparent paper were then joined by orientation technique (pass point ed) from the preliminary map. Tracing paper was laid on these transparent overlays and traced as first base ma p to be relayed on as the base map with all details. This was done for both 1977 and 1999. The data on the tracing papers were digitized to produce land-use maps of the study area for 1977 and 1999. Visual interpretation was preferre d given the relatively small size of the study area. The resulting interpretation was also digitized using Arc IN FO 3.5.1. To rectify the aerial

PAGE 81

81 photographs, features such as ro ad junctions and river confluen ce points were identified both on the two sets on one hand, and geo-referenced to pographic maps (sheet no. 232/1 and 232/2) on the other hand. All photos were geometrically rectified and registered to a common UTM (Universal Transverse Mercator) projecti on based on 1:50,000 scale topographic maps of Tanzania (i.e., UTM zone 36 south, datum Arc 1960). The identified points were distributed throughout the study area. Using Arc INFO 3.5.1, the points were used to transform (project) the interpretation from aerial photographs. 4.3 Data Analysis The coverage or layers produced in ArcINFO were used to produce land use/ cover maps of the two sets in ArcView 3.2. Data analysis to produce the change detection matrix was done using ArcView and Microsoft Excel. ArcView sh ape files were exported (as dbf files) to Microsoft Excel where the pivot table function was used to produ ce the change detection matrix. Quantitative data for the overall land use change s and gains and losses in each category were compiled (Table 4-2 and Table 4-3). The change matrix provides information on the main types of changes in the study area. 4.4 Results Land use and land cover maps of 1977 and 1999 are displayed in Figure 4-1 and 4-2. The overall land use changes from 1977 to 1999 are s hown in Table 4-2. According to the table, agricultural land was the larges t land-use type, both in 1977 and 1999. Agricultural land took up 65.3 and 46.0%, respectively, of the total area. As per Table 4-2, densely popul ated areas as well as areas of grassland and woodland increased from 1977 to 1999. On the contrary, agricultural land and open valleys declined. In 1977, agricultural land covered about 65% of the study area and with an estimated area of 589.4 ha, but in 1999, the total agricultural area was estimated to have decreased by about

PAGE 82

82 29.4% to 416 ha. The annual rate of decrease is estimated at 1.3% during 1977-1999. Valley bottoms covered about 17.4% of the study area a nd with an estimated area of 157.4 ha, but in 1999, the total valley bottom area was estimated to have declined by about 18.1% to 128.9 ha. The annual rate of decrease is approxima tely 1% during 1977-1999. Meanwhile, densely populated areas, grassland and woodland incr eased by about 138, 308, and 79% respectively with an annual growth rate of 6, 13 and 3% during 1977-1999. Results from the transition matrix in Table 43 indicated the area incr ease or decline of each land use type. It is clear that between 1977 and 1999, the transition replacement rates of agricultural land, populated ar ea, grassland and woodland were high at 42.6, 44.5, 43.4, and 49.6% respectively. That of valley bottoms was th e lowest with a transition rate of 25.1%. Between 1977 and 1999, about 17% of valley bottoms were transformed to agricultural land and about 16% were transformed to grassland. 4.5 Discussion Remote sensing and GIS technologies were us ed to determine the land-use/cover changes associated with agricultural ac tivities around the Ndembera swamp. Specifically, I evaluated the magnitude of wetland cover change that took place between 1977 and 1999 around Ndembera wetland. Ndembera wetland has been experiencing variou s changes over the years. Data on Table 42 indicate that densely populat ed areas, grassland, and woodl and areas increased between 1977 and 1999 while agricultural land and wetland areas decreased during that same period. Although the agricultural land decreased, it remains the largest land use t ype in both 1977 and 1999. Table 4-2 shows changes of the wetland area between 1977 and 1999. Table 4-3 shows that the total wetland area declined by about 18% between 1977 and 1999 and that at least 17% of the wetland area was transformed to agricultur e. This finding supports farmers perceptions

PAGE 83

83 observed in Chapter 3 (Table 3-10 ) that wetland utilization for vinyungu cultivation has increased. At least 97% of the farmers interviewe d reported this increase. In addition, 90% of the farmers that were interviewed believe the Ndembe ra wetland has decreased in size and that the rivers that drain into the wetland have dried the outcomes they attributed to increased cultivation of vinyungu Farmers attributed the increased utilization of vinyungu mainly to a decline in crop production from the uplands th at is caused by increase d soil infertility and droughts (Tables 9 and 10). The decline in cr op production in the upland areas was further compounded by the structural adjustment program th at started in the mid 1980s that required the government to stop subsidizing farmers for fertiliz er and farm tools. Availability of water and fertile soils in the wetlands therefore, attracted farmers as a way to increase their cultivation of vinyungu Today, vinyungu are not only attractive to farmer s but also to the government. During group discussions, farmers reported they were directed by the government to cultivate vinyungu between 1992 and 1993 to overcome food shortage that was caused by prolonged droughts. Table 3-3 in Chapter 3 clearly in dicates an increased income from vinyungu cultivation compared to rain fed agriculture. Farmers a ssociated the increase in income with trade liberalization that occu rred in the 1990s that allowed fo r improved marketing of goods both in and outside the country. The increase in income is likely to influen ce further conversion of wetlands to agriculture. Table 4-3 shows that human settlement in villages surrounding the Ndembera wetland has increased over time by 137%. This increase in human settlement may be contributing to increased conversion of wetlands to agriculture. The percentage change of the wetland to ag riculture seems small. However, recurrent prolonged droughts and market forces th at seem to contribute in driving vinyungu expansion

PAGE 84

84 started only recently, meaning, the observed land cover changes took place within a relatively short period of time. For example, during the group discussions, farmers reported that the most popular crops, i.e., peas, tomatoes, onions, and pota toes started to be cultivated mostly in the 1990s. Paprika (that is becoming increasingly important to farmers in terms of financial returns) started to be cultivated around 2004 and is now finding its way to vinyungu plots Based on these findings, it is evident that wetl ands are potential agricultural lands, that unexploited wetlands are likely to be continuously cleared as l ong as farmers have access to them, and that markets for goods obtained from vinyungu remain viable. These findings also confirm the hypothesis that expansion of vinyungu farming is decreasing the size of wetlands. About 25 ha (or 16%) of the wetland area has be en transformed to grassland. The grassland in this area, as was observed dur ing field survey, consists of s hort, dry (yellow) grass that remains so until during the rainy months (March to April) when the area may be inundated and the grass turns green. This is to say that the wetland vegetation has been replaced by vegetation that can withstand prolonged dry spells. This change may be cau sed by reduced water flows into the wetland that was reported by farmers (Table 39). It may also be due to accumulation of soils resulting from soil erosion further upstream th at was reported by farmers during the interview (Table 3-9). Soil erosion upstream is known to cause sedimentation downstream that may be colonized by vegetation. Table 4-3 also shows that about 23% of the agricultural land changed to a woodland area. Farmers were asked during the group discussions the reasons for the agricultural land to be replaced by the woodlands. Farmers attributed this change to a ban made by the Government in the mid 1980s to stop farmers from cultivati ng finger millet in the woodlands, allowing a recolonization of woodlands in th e previously cultivated areas.

PAGE 85

85 About 10% of the agricultural land is tran sformed to a grassla nd area between 1977 and 1999. Again, the grassland consists of short dry grass. These areas were observed as abandoned agricultural land. Other changes include the transformation of woodlands to agricu ltural land and to grassland. Farmers attributed these changes to encroachment into woodlands where woodlands are cleared for cultivation purposes or to obtain wood for home consumption. About 33% of the grassland areas changed to agricultural land between 1977 and 1999. Farmers reported these areas to be wetter than th ey used to be, probably due to change in the configuration of rivers that drai n in to the Ndembera wetland. Av ailability of water makes them attractive for cultivation.

PAGE 86

86 Table 4-1. Map Histor y (i.e., Data used) S/N Date of Photograph Film No. Exposures Scales 1828 111-114 1844 137-142 1809 144-151 1 July 1977 (dry season) 1809 106-113 1:50,000 RUN 15 2998-3004 RUN 16 2587-2595 2 September 1999 (dry season) RUN 17 2661-2670 1:50,000

PAGE 87

87 Figure 4-1. Ndembera swamp in 1977 (A). The inset map (B) shows Ndembera swamp location in Iringa region. A B

PAGE 88

88 Figure 4-2. Ndembera swamp in 1999 (A). The inset map (B) shows tha location of Lyandembera swamp in Iringa region A B

PAGE 89

89 Figure 4-3. Land use/ cover change map (1977 and 1999) (A). The inset map (B) shows th e location of Lyandembera swamp in Iringa A B

PAGE 90

90 Table 4-2. Total area (ha) and area of ch ange of land use types from 1977 to 1999 Total Area (ha) Total Area (%)1977 1999 (change) Land use Type 1977199919771999(ha)% %/yr Densely Populated Areas 16.439.11.84.322.6137.7 6.0 Agricultural Land 589.4416.065.346.0-173.5-29.4 -1.3 Grassland 29.8121.83.313.591.9308.3 13.4 Wetland 157.4128.917.414.3-28.5-18.1 -0.8 Woodland 110.2197.612.221.987.479.27 3.5 Estimated total 903.3903.3100.0100.0 Wetland is synonymous to Open valley seasonally flooded Table 4-3. Transition matrix of land use types from 1977 to 1999 (transition probabilities in %) Land use type 1999 Densely populated areas Agricultural Land Grassland Wetland Woodland 1977 Area (Ha) % Area (Ha) % Area (Ha) % A rea (Ha) % Area (Ha)% Densely 9.1 55.5 4.426.80.95.60.3 2.1 1.710 populated areas Agricultural land 27.3 4.6 338.357.4 60.510.324.9 4.2 138.423.5Grassland 0.1 0.4 1033.9 16.756.5 1.3 4.4 1.44.7Wetland 1.8 1.2 27.217.325.416.1 102.2 65 0.80.5Woodland 0.3 0.2 36.132.718.316.60.1 0.1 55.550.3

PAGE 91

91 CHAPTER 5 DISCUSSION AND CONCLUSION 5.1 Introduction This thesis examined the growth of wetland farming and its impact to the environment. The first four chapters presented the backgr ound and introduction to the study, literature review, analyses of farmer and household characteristic s within the study area, and the assessment of land use and land cover change. In this fifth and last chapter, di scussion of the major findings is covered. This chapter relates research findings to the research questions and hypotheses posed in Chapter 1. Policy and practice implications of these findings are also presented. A study conclusion that also discusses study limitations and recommendations for future research is covered in the last section of this thesis. 5.2 Major Findings Analysis of farmer and household characteri stics within the study area used a set of questionnaires to determine how vinyungu -farming practice is relate d to household and farmer characteristics; how vinyungu -farming system evolved between 1970s and 1990s; what have been the driving forces behind the changes in the vinyungu -farming practice; and, what have been the social and environmenta l implications of changes in vinyungu -farming system. The primary objective of this thesis was to study cha nges in land use over time and the environmental effects associated with those changes. Many farmers involved in vinyungu farming around Ndembera wetland in Iringa and Mufindi Districts of Iringa regi on, Tanzania, provided information for this study regarding their wetland use and their perceptions of changes. An alysis involved both quantitative and qualitative data from participating farmers and group discussions conducted in the study area.

PAGE 92

92 I used the information gathered to test the hypotheses that members of households located within a short distance from the we tlands are more likely to practice vinyungu -farming system and that short distance from the wetlands to ro ads and markets would be associated with greater proportion of wetland converted to agriculture. I also tested the hypotheses that road and market accessibility would be associated with cultivation of a greater proportion of cash crops than food crops; greater cash crop production would be associated with in creased costs of farm inputs (labor and fertilizer); greater number of people within the work ing age in a household would be associated with greater farm size and number of crops grown; and that farmers are likely to use household income on farm inputs (i.e., fertilizer and labor) so as to improve crop production. Farmers sex or duration of reside nce in the village was not expect ed to be associated with the size of vinyungu -farming area or types of crops grow n. Some of these hypotheses have been supported while others have been rejected as ha s been illustrated in the following sections. The findings show that almost all farmers in the study area cultivate both vinyungu and upland plots. However, the size of their upland plots is larg er than that of vinyungu i.e., 70% of farmers own > 1.2 ha in the uplands co mpared to 11% who own > 1.2 ha of vinyungu Although most farmers appear to own relatively small vinyungu plots, the overall area used for this farming practice is large because almost all farmers in this area own vinyungu Ownership of both vinyungu and upland plots means that vinyungu cultivation is not there to replace upland cultivation but that both vinyungu and upland cultivation are impor tant sources of livelihoods for the people of Iringa. Generally, a farmers decision to cultivate vinyungu is driven by the availability of labor, water, and so il fertility and not by lack of land.

PAGE 93

93 At least 63% of those interviewed were men. Recalling that women previously predominantly undertook vinyungu cultivation, the higher proportion of mens involvement indicates a growing interest in wetland use am ong men, possibly for highl y profitable cash crops. Although almost all farmers cultivate on land th ey have acquired through inheritance, usually, in close proximity to the villages th ey live in, at least 15% bought the land they cultivate. The relatively large pr oportion of land purchases in th ese areas, where native farmers would mostly acquire land through tr aditional, non-monetary transa ctions, might be an indicator that there is a growing need and willingness to acquire and cultivate natural wetland areas among individuals from other areas of the region or country. Farmers may request access to vinyungu owned by colleagues or re latives both within and outside their village, especially from landowners that are unable to cultivate thei r entire property due to lack of human and financ ial resources. This trend however, is becoming rare while renting is slowly increasing. Currently though, only a few farmers do rent vinyungu plots from fellow farmers because such plots are not easily availabl e and are relatively expensive. In most cases, vinyungu are held as a precious and limited resource by families and therefore, all farmers interviewed in this study reported that they do not rent out their pl ots to other farmers. Where the wetland area is large enough to accommodate more farmers, farmers from within the village could access more land without a ny formal request. Water resour ces are communally managed to ensure equal access among farmers. However, wh ere the wetland resources are limited, there is an increase in conflicts over access to vinyungu land and to water resources. These conflicts were traditionally uncommon in the study area. Family members (i.e., husband, wife, childre n, dependents) are the most commonly used labor. This further explains that vinyungu cultivation is no longer a predominantly womens

PAGE 94

94 activity but rather an activity of the entir e family. In fact, the linear regression revealed that both men and women cultivated vinyungu and that men had more of their vinyungu plots grown cash crops compared to women. The linear regression fu rther revealed that hou seholds with a greater number of people within a working age gr oup are more likely to have more area in vinyungu during the dry season. This is mainly because construction of vinyungu is both time-consuming and labor-intensive due to wet and heavy soils th at are typical of wetlands. The time required to make a kinyungu is determined by the size of kinyungu desired and the labor available to accomplish the task. Labor hire is considered an expensive undertaking and it is therefore not often used. Crop production in the uplands has declined due to decline in soil fertil ity and rainfall. On the other hand, crop production in the vinyungu has increased mainly because farmers cultivate the vinyungu twice or thrice a year, maxi mizing use of the available water resources and fertile soils. Vinyungu are therefore considered important for both food a nd cash crop production. In this case, water, soil fertility, and increased production seem to influence farmers decisions to cultivate vinyungu and/or increase the size or number of plots. The bivariate analysis revealed that mo re farmers use at least half of their vinyungu plots for growing food crops compared to a few who us e the same hectareage for growing cash crops. However, the linear regression, whic h included all factors influencing vinyungu farming, showed that farmers would have at least 61% of their vinyungu grown cash crops. This result suggests a growing importance of wetland use among the reside nts, i.e., an activity historically known to supplement food production from upland plots is now increasingly used to produce cash crops. Close proximity to both markets and roads was shown to influence an increase in both the size of vinyungu and the proportion of cultivated cash crops However, in my study, these did not

PAGE 95

95 show statistical significance. Various factors may explain why proximity to major markets and roads may not affect cash crop production. One, th ere have been concerted national efforts to improve communication networks in the country making most remote places easily accessible; and two, due to good road networks and market demand in major cities, middlemen buy crops from the farm and transport the pr oduce to major markets where crops are sold at almost twice or thrice the original price. This eases sale of produce by farmers at the same time enables middlemen to make profit. Residents with fewer years of residence in the villages are more likely to have larger sizes of vinyungu plots in the dry season (p= 0.048). It is not clear what c ould be the reason for this finding, but one possible explanatio n could be the influx of new re sidents into the wetland areas in recent years due to better crop production in the wetlands. Th ese farmers did not have access to inherited crop land. Greater distance from the house to the plot was associated with lesser cash crop growing in the vinyungu That is, plots that are further from home are used to grow mostly food crops. These farms are therefore not very active because activ e farms require adequate labor and frequent attention and as such, they are usually close to settlements. Duration of residence was associated with the size of vinyungu plots for growing cash crops, i.e., residents with fewer year s of residence in the villages we re more likely to have larger proportions of their vinyungu plots used to grow cash crops. It appears that new resident farmers are those that are interested in increasing income through the cultivation of wetlands. Thus, most recently settled farmers may be more inclin ed to growing cash crops than food crops. Expansion of vinyungu and cultivation near sources of water may be associated with farmers perceptions of reduced water flow, re duced soil fertility, reduced wetland size and

PAGE 96

96 vegetation as well as increased soil erosion. As a result, at least 87.5% of farmers use inorganic fertilizer on vinyungu currently (1990s-00s) compared to the 1970s when they did not use fertilizer. As water resources are becoming limiti ng in Iringa region farm ers reported increasing human conflict over use of these resources. As discussions with farmers clearly indicated an increased use of wetland resources for the cultivation of vinyungu I used GIS and remote sensing tec hnologies to determine the magnitude of wetland conversion to agriculture. GIS and remote sensi ng technologies are an economically feasible way of gathering information with high spatial, spectral, and temporal resolution over large areas (Verstraete et al., 1996). Lack of recorded hist orical data may limit the use of remotely sensed data to detecting land use chan ges due to difficulty in estimating uncertainties about the land use classification. However, field survey on land use, interviews with local farmers and district staffs to ga ther historical land cover data, and simple classification systems, make the use of remotely sensed data an effective means of acquiring information on land use changes, and these methods were employed in this research. Figure 4-2 shows the Ndembera wetland has experienced land use changes over time. A transition matrix was developed to test Hypothesis 5, greater expansion of vinyungu farms would be associated with incr eased conversion of wetland area to agricultural land. Table 4-3 indicates a transition probability of 17% of the wetland ar ea to agriculture. Farmers linked their interest on the wetland for crop production to lack of alternative land, soil fertility, and availability of water. In additi on to these, they related the increased use of wetlands for agricultural purposes to changes that took place in th e country about ten to twenty years ago. These include the structural adjust ment program initiated in the mid 1980s that, among other things, took government subsidies aw ay from farmers. Inputs such as inorganic

PAGE 97

97 fertilizer became limited and expensive. Also, in the 1990s and especially 1992 and 1993, Tanzania experienced serious and prolonged dr oughts that affected crop production in the country. To many farmers, wetlands (typically wi th rich soils and wet conditions) offered a solution. Farmers also reported receiving government directives to cultivate vinyungu to overcome food insecurities. Farmers use at least 61% of their vinyungu to grow cash crops. This is to say vinyungu historically known to supplement food production from the uplands, are now increasingly used to produce cash crops. Farmers associated this change to increased markets for their goods that was in turn influenced by trade liberalization that took place in the mid 1980s and was stronger in 1990s. Trade liberalization replaced the former planned economic system with a market-driven economic system. It opened doors for Tanzanians to market their goods (both inside and outside the country) unlike during the former system wh ere the government contro lled all the marketing. This increased the ability of Tanzanians to pur chase goods. It also made markets for Tanzanian goods (including food) available. The increased purchase power and markets encourage farmers to increase production, both in the uplands and the wetlands. As noted in chapter 3, farmers can cultivate vinyungu twice, even thrice a year. This increases production and financial gains, enco uraging further conversion of wetlands to agriculture especially if the household has adequate labor to cultivate the arduous wetland soils. The percentage change of the wetland to agri culture seems small. However, some of the factors that seem to contribute in driving vinyungu expansion (droughts, poor soil conditions, and market forces) intensified in the late 1980s early 1990s, meaning, the observed land cover changes took place within a short period of time.

PAGE 98

98 5.3 Conclusion This study was conducted in Iringa and Mufindi districts, both of Iringa region in southwest Tanzania. Nine villages were studie d. The goal of this study was to investigate the nature of transformation of the vinyungu -farming practice in the two districts and to determine factors influencing the transformations in this farming system and environmental implications associated with this farming system. Two appr oaches were used. First, I used structured questionnaires to determine farmer and hous ehold characteristic s associated with vinyungu farming system as well as farmers perceptions on how the system has changed over time, factors influencing the change, and environmental implica tions associated with these changes. I then used remote sensing and GIS to investigat e the relationship betw een wetland change and agricultural expansion. Several factors have been found to influence the transformation of wetlands to vinyungu fields. These include soil infert ility of the upland plots making the rich and moist wetland soils more attractive for agriculture. In the 1970s, vinyungu were small gardens cultivated by women to supplement household diets. However, with p opulation growth, reduced soil fertility in the upland plots, rainfall unreliability and incr eased droughts, a demand for more arable land increased. Availability of surface water and the hi gh water table of Iringa provide vital moisture that allows farmers to cultivate more than once in a year, hence maximizing household annual crop production. Generally, wetlands in the study area can still accommodate more vinyungu This, plus the availability of family labor enco urage further conversion of wetlands. Accessibility to major markets and roads did not appear to be strongly a ssociated with the size of vinyungu cultivated or the proportion of vinyungu cultivated cash crops. However, given the improvement of the road systems, most farmers do not have to transport their produce to the major markets, as buyers drive to the farms to collect the produce to be sold in major markets. Lastly, farmers

PAGE 99

99 reported that sometimes they cultivate vinyungu following government directives during severe droughts. This research also looked at how vinyungu farming practices have changed over the years. The results indicate a shift in agriculture in the area i.e. vinyungu farming system that supplemented household food obtained from uplands fields has evolved to an economic activity, producing cash crops for the intern al market as well as for major towns and cities both in and outside the region. Also, due to a decline of production power in the uplands and increased profitability in the vinyungu this traditional irrigation system is transforming from being exclusively a womans activity to being an activity fo r both sexes. A slight increase in the size of vinyungu cultivated, use of fertilizer, year-round pr oduction of crops, and encroachment on to wetter parts of the wetland are other ways by wh ich farmers have modified their traditional irrigation system. Modifications in the vinyungu -farming system have been influenced by a number of factors. These include social change in terms of population growth that required technological methods which would augment productivity so as to cope with growing populations and market demands that require fast, massive and efficient crop productivity (hence the use of fertilizer and repeated cultivation). Another f actor is the decline of production power in the uplands due to various reasonssuch asreduced soil fertility and droughts. Improved acces sibility allows the middlemen to reach farmers easily, increasing farmers ability to market both perishable and non-perishable crops. Individual needs to in crease personal income to overcome social challenges and access social amenities such as e ducation, health and clothing, also influence the modifications of vinyungu -farming system. Other reas ons for modification of vinyungu -farming system may be the great pr ofit accrued from sale of vinyungu produce over a short period given

PAGE 100

100 the repeated production over a year. This encourag es farmers to increase their energy on wetland plots. As such, use of the wetlands for vinyungu cultivation has increased when compared to the 1970s as was indicated by nearly 98% of the re spondents in this studyas well as through the GIS analysis that indicated the we tland area declined between 1977 and 1999 with agriculture being the major reason for that decline. Overall, good progress has been made in achie ving the aims of the study. However, a few factors may have influenced th e outcome of the study, for example, the small sample size and study area. Important data such as historical census at the vill age level is either lacking or incomplete, hence limiting the analysis of population change at the village le vel. Future research is needed to improve the predictiv e ability of the models used by including other factors, such as water quality, water flow and rate as well as bett er measurement of distance from the plots to markets and roads. I also recommend that future studies involve a much larger sample of farmers or follow the same group of farmers over time to establish th e validity of th e relationship between distance to major markets/ roads and growth of vinyungu farming as well as the relationship between womens access to the middlemen and growth of vinyungu farming. Future studies may need to determine whethe r or not women compete with men in vinyungu farming and whether or not women have changed th eir roles following mens participation. Finally, it is apparent that we tlands are of a great social and economic importance to the people living around them. On the other hand, farm ers admitted an increased deterioration of wetland resources, an outcome they linked to a growing level of hu man activities. More importantly, farmers clearly understand that if the current use of we tland resources remains unchecked they may end up losing the very reso urcse they depend upon for their livelihoods. As such, farmers welcome any intervention th at would balance human needs and wetland

PAGE 101

101 sustaninability. These findings provide an opportu nity for the government to encourage farmers to adopt farming practices that conserve the wetlands. I hope that the results of this study would pr ovide a useful addition to the knowledge on vinyungu farming dynamics. I expect that in the long -run, results of this study, when combined with those of earlier investig ators on the evolvement of vinyungu farming, would provide an important input towards the making of policies governing the use of wetlands.

PAGE 102

102 APPENDIX A SAMPLE SIZE DETERMINATION 1. Reports from the village and other divisional leaders in the study area indicated that there were approximately 3,000 farmers in the nine villages that were selected for participation. 2. For the purpose of this study, I wanted to obtain the approximate number of farmers who practiced vinyungu farming. After a series of discussions with th e leaders and consultation of the village records, I determined that between 1,500 and 2,000 farmers were engaged in vinyungu farming. 3. Of those practicing vinyungu farming, it was indicated that nearly 95% grow both food and cash crops. This study targeted those farmers, in order to measure the extent to which wetlands are used for subsistence and commercial farming. 4. I wanted to be 90% confident that my sample in cluded at least 95% of fa rmers who grew both food and cash crop. With the 90% confidence level, = 0.1. I imposed an error margin of 5, i.e., I wanted to get this proportion within 5% of its true value. 5. With this information, I used a standard formula for computing sample size for proportion shown below: 2 2 2 / E pq z n Equation A-1 where z is the standard normal value corresponding to a (1) confidence coefficient, p=0.95 the population proportion, q=1-p=0.05, and E is the error margin. I have z /2 = z0.05 = 1.65, and n is the sample size for a two-sided test. Inserting these values in equation 1, I get: 2 2) 05 0 ( ) 05 0 ( ) 95 0 ( 65 1 n= 51.73 6. Thus my calculations indicated that a minimum of 51 respondents is required to be 90% confident that the sample consists 95% of farmers who grow both food and cash crops in wetlands. 7. To ensure that I obtain an equal number of farmers from each of the nine villages, I interviewed 54 farmers, i.e., six from each village.

PAGE 103

103 APPENDIX B SURVEY INSTRUMENT: INDIVI DUAL FARMER QUESTIONNAIRE B.1 Location 1(a) Village_____________ 1(b) Ward____________ 1(c) Division__________ (1d) District ___________ 1(e) Distance from interviewees house to the plot: (km) _________ 1(f) Distance from main/tarmac road to cent er of plot of interviewee: (km) _________ B.2 Background Information 2. Name_________________________________________ 3. Sex: i) M [ ] ii) F [ ] 4. Age: i) <18 [ ] ii) 19-35 [ ] iii) 36-60 [ ] iv) >61 [ ] v) DN [ ] 5. Marital Status: i) Married [ ] ii) Single [ ] iii) Divorced [ ] iv) Widowed [ ] v) Separated [ ] 6. Education level: i) Uneducated [ ] ii) Primary [ ] iii) Primary + Course [ ] iv) Secondary [ ] v) Higher Education [ ] vi) Univ./College [ ] 7. How many people live in this household through most of the year (including yourself)? i) (61+) a) M [ ] b) F [ ] ii) (18-60) a) M [ ] b) F [ ] iii) (Under 18) a) M [ ] b) F [ ] 8. Self-identified Cultural Group/Tribe_______________________ 9. How many years have you lived here in this community? (Years/ duration) _________ (if immigrant go to Q.10, if resident go to Q.11) 10. Where did you come from? i) (Village) _________________ ii) (Region) _______________ 11. What inspired you to live aroun d here? i) Easy access to land [ ] ii) Soil fertility [ ] iii) Family ties/ responsibilities [ ] iv) Inheritance [ ] v) Close to good markets [ ] vi) Close to my work place [ ] vii) Other:_ ____________ ______ 12. What do you do for a living/what is your main occupation? __________________ 13(a) Do you belong to a farmer coop erative society/ farmer association? i) Yes [ ] ii) No [ ] (if no go to #14) 13(b) If yes, what is the name of the cooperative society/ association? __________________ 13(c) When did you join? (Year) _________ 13 (d) What is the purpose of the society/association? _________________________________________ 13(e) Has your membership been benefici al? i) Yes [ ] ii) No [ ] Explain______________________________________________________________________________ B.3 Land availability, crop preference, seasonality, and fallow period 14(a) Do you own a piece of farmland? i) Yes [ ] ii) No [ ] (If No go to # 22) 14(b) How did you acquire this farmland? i) Bought it [ ] ii) Given/inherited [ ] iii) First here [ ] iv) Other [ ] __________________________________________________ 14(c) Do you have some type of title or papers to indicate ownership? i) Yes [ ] ii) No [ ] Mention: ________________________________________________________ (If No go to # 15) 14(d) If yes, how long is it valid for? (Years/Duration) _____ 15. Where is the farmland? i) Wetland [ ] ii) Upland [ ] iii) Both [ ] 16. If cultivating in the wetland, when did you start cultivating there? (Year) _________________ 17. What is the total estimated amount of land owned by the household? (a)Upland i) less than an acre (b) Wetland i) less than an acre ii) 1-2 acres ii) 1-2 acres iii) 3-4 acres iii) 3-4 acres iv) 4+ acres iv) 4+ acres 18. When do you farm them? (a) Upland i) Dry season [ ] ii) Wet season [ ] iii) Both [ ] iv) Never [ ]

PAGE 104

104 (b) Wetland i) Dry season [ ] ii) Wet season [ ] iii) Both [ ] iv) Never [ ] 19. How productive are the upland fields? i) Very produc tive [ ] ii) Fairly productive [ ] iii) Not very productive [ ] iv) Not at all productive [ ] v) Cant choose [ ] 20. How productive are the wetland fields? i) Very productiv e [ ] ii) Fairly productive [ ] iii) Not very productive [ ] iv) Not at all productive [ ] v) Cant choose [ ] 21(a) Do you rent out or lend out your wetland farm to other farmers sometimes? i)Yes [ ]ii) No [ ] 21(b) If yes, how many plots? _____ 21(c) Of what area/size? i) less than an acre [ ] ii) 12 acres [ ] iii) 3-4 acres [ ] iv) 4+ acres [ ] 21(d) If no, why?________________________________________________________________________ 22(a) Do you rent/borrow wetland plots from other farmers to plant your crops? i) Yes [ ] ii) No [ ] (if no, go to #23) 22(b) How many plots? __________________ 22(c) Of what size? i) less than an acre [ ] ii) 1-2 acres [ ] iii) 3-4 acres [ ] iv) 4+ acres [ ] 22(d) Do you use these plots differently than the plots you own? i) Yes [ ] ii) No [ ] 22(e) If yes, please describe how ____________________________________________________ 23. What influenced your decision to own or rent a wetland plot? i) cheap/easy to get [ ] ii) close to the market [ ] iii) close to home [ ] iv) high soil fertility [ ] v) dry season alternative (water availability) [ ] vi) other ____________________________ 24. How far is your /rented wetland farm from the Ndembera river tributaries? i) right in the river valley [ ] ii) 1-5m from the river [ ] iii) 6-10m from the river [ ] iv) 11-15m from the river [ ] v) 16+m from the river [ ] 25. How much of the land you own /rent is usually planted? (a) Upland i) Wet season? All__ ____acres (b) Wetland i) Wet season? All__ ___acres ii) Dry season? All__ ____acres ii) Dry season All__ ___acres ii) Both seasons? All__ ____acres iii) Both seasons? All__ ___acres 26. How far, roughly, do you have to travel to your plot(s)? (a) Upland i) around 30min (~4km) (b) Wetland i) around 30min (~4km) ii) around 1hr (~7km) ii) around 1hr (~7km) iii) around 2hrs (~14km) iii) around 2hrs (~14km) iv) > 2hrs (> 14km) iv) > 2hrs (> 14km) v) Do not know v) Do not know 27. How often do you travel to each plot? (a) Upland ____ (b) Wetland ____ 28. What 3 main crops are grown for food for the household? (Begin with the most important/ often eaten) Crop Rank Location Crop Rank Location Crop Rank Location _____ ____ _______ _____ ____ _______ _____ _____ _______ Others: _____________ 29. What are the 3 main crops you grow for sale? (Rank by total income earned for hh from crop) Crop Rank Location Crop Rank Location Crop Rank Location _____ ____ _______ _____ ____ _______ _____ _____ _______ Others: _______________ 30. What proportion of your wetland plot is planted with: (a) Food crops ____% (b) Cash crops ____% 31(a). If you sell the food you produce from a wetland plot, what proportion remains to meet the household needs and what proportion is sold? (i) Remaining _______ % (ii) Sold ______ % 31(b) If you sell the food you produce from upland plots, what proportion remains to meet the household needs and what proportion is sold? (i) Remaining _______ % (ii) Sold ______ % 31(c) What do you use the money for? ______________________________________________________ 32(a) Do you plant the same annual crops in each of your wetland plots year after year? i) Yes [ ] ii) No [ ] 32(b) If yes, why? _______________________________________________________________________ 32(c) If no, please describe how you usually rotate your annual crops (season to season, year to year)

PAGE 105

105 ______________________________________________________________________________________ 33(a) Do you purposely rest (fallow) any of your upland plots before planting them again? i) Yes [ ] ii) No [ ] 33(b) If yes, How many seasons/ years do you rest each of your upland plots before planting them again? i) every season [ ] ii) every after one season [ ] iii) every after two seasons [ ] iv) never [ ] v) other _____________________________________ 34(a) Do you purposely rest (fallow) any of your wetland plots before planting them again? ) Yes [ ] ) No [ ] 34(b) If yes, How many years/ seas ons do you rest each of your wetland plots before planting them again? i) every season [ ] ii) every after one season [ ] iii) every after two seasons [ ] iv) never [ ] v) other _____________________________________ 35. What was the status of the wetland when you moved here? i) in its natural state [ ] ii) inhabited [ ] iii) cultivated [ ] iv) Other ________________ 36(a) How would you describe the trends in wetland utilization? i) uses have increased tremendously [ ] ii) uses have incr eased slightly [ ] iii) uses have remained the same [ ] iv) uses have declined [ ] v) I dont know [ ] 36(b) How many households use the wetland for valley-bottom cultivation? __________________ 36(c) Has this number I) increased? [ ] ii) Decreased? [ ] iii) remained the same? [ ] B.4 Labor, Input, and Tools on Vinyungu 37(a) Do you work for other farmers as hired labor? i) Yes [ ] ii) No [ ] (If No, go to # 38) 37(b) If yes, how much money do you get (per day or task)? (TShs) ________________ 37(c) What is the main use of the money you earn from working? i) domestic use (food and clothing) [ ] ii) pay school fees for kids [ ] iii) re-invested on the farm [ ] iv) pocket money for head of household [ ] v) put to family business [ ] vi) other: ___________________________________ 38(a) Do you hire others to work for you? i) Yes [ ] ii) No [ ] (If No g to # 40) 38(b) If yes, for what tasks do you hire the labor? i) to cultivate [ ] ii) to clear land [ ] iii) to weed [ ] iv) to harvest [ ] v) others:_______________________________ 38(c) How much do you pay for labor (per day or task)? _______________________________ 39(a) Can you find workers whenever you need them? i) Yes [ ] ii) No [ ] 39(b) If yes, where do they come from? i) within the village [ ] ii) neighboring villages [ ] iii) other regions [ ] iv) Dont know Name: _________________________ 40(a) Do you usually use manure on any of your crops or plots in the wetland? i) Yes [ ] ii) No [ ] (If No, go to # 41) 40(b) If Yes, on which crops do you usually use manure? Crop ____________ ____________ ____________ _____________ Type of manure (animal/ peat?) ____________ ____________ ____________ _____________ 40(c) Have you always used manure since you started working a kinyungu ? i) Yes [ ] ii) No [ ] 40(d) If No, what might have triggered the use of manure r ecently? i) loss of fertility in the soil [ ] ii) need to grow more and faster [ ] iii) it is a habit to use manure [ ] iv) easy way to discard family manure [ ] v) other: ________________________________________________________ 40(e) Do you usually have enough manure from your own animals/ source for your crops? i) Yes [ ] ii) No [ ] 40(f) Do you ever buy manure? i) Yes [ ] ii) No [ ] (If no, go to question # 41) 40(g) If yes, how often do you buy manure? i) every season [ ] ii) every other season [ ] iii) once every two seasons [ ] iv)once every three seasons v) once every four seasons v) Other: ______________________ 40(h) How much did you buy the last time you bought manure? i) 1-5 bags [ ] ii) 6-10 bags [ ] iii) 11-15 bags [ ] iv) 16-20 bags [ ] v) other ________ 40(i) How much did you pay the last time you bought manure? i) <500 TShs per bag [ ] ii) 501-1000 TShs per bag [ ] iii) 1001-1500 TShs per bag [ ] iv) 1501-2000 TShs per bag [ ] v)2001-2500 TShs per bag [ ] vi) 2501-3000 TShs per bag [ ] vii) Other: __________________________ 41(a) Do you use fertilizer on any of your crops? I) Yes [ ] ii) No [ ] (If no go to # 43) 41(b) If Yes, how often do you use fe rtilizer? i) Regularly [ ] ii) Occasionally [ ] iii) Never [ ]

PAGE 106

106 41(c) Why did you begin to use fertilizer? i) loss of fertility in the soil [ ] ii) need to grow more faster [ ] iii) it is a habit to use fertilizer [ ] iii) other: _____________________________________________________________________ 41(d) What type(s) of fertilizer do you use? __________________________________________ 41(e) On which crop(s) do you use fertilizer? _____________________________________________ 41(f) Where do you purchase or obtain the fertilizer from? ____________________________________ 41(g) What amount do you purchase? i) 1-5 bags [ ] ii) 6-10 bags [ ] iii) 11-15 bags [ ] iv) 16-20 bags [ ] v) other ________ 41(h) How much did you pay the last time you bought fertilizer? i) <500 TShs per bag [ ] ii) 501-1000 TShs per bag [ ] iii) 1001-1500 TShs per bag [ ] iv) 1501-2000 TShs per bag [ ] v)2001-2500 TShs per bag [ ] vi) 2501-3000 TShs per bag [ ] vii) Other: _____________ 41(i) Is fertilizer always available if you have money to buy it? i) Yes [ ] ii) No [ ] 42. If you have never used fertilizer, please give the reasons: _____________________________________ 43(a) Are there any other methods (local/traditional) that you use to improve the soil fertility of your wetland plot/land? i) Yes [ ] ii) No [ ] 43(b) If yes, what other methods do you use? _________________________________________________ 44. How do you cultivate your fields? i) ox plow [ ] ii) hand hoe [ ] iii) tractor [ ] If more than one practice please explain____________________________________________________ 45. What would you say is the major agricultural problem associated with cultivating in the wetland? i) laborious [ ] ii) costly investment [ ] iii) conflicts iv) low return/low value for money [ ] v) it is only seasonal [ ] vi) Other: ______________________________________________ 46. What would you say is the major environmental prob lem associated with cultivating in the wetland? i) soil erosion [ ] ii) sedimentation [ ] iii) reduced river flow [ ] iv) drying of the wetland [ ] v) reduced wetland size [ ] vi) flooding [ ] vii) river banks erosion [ ] ix) Other ____________

PAGE 107

107 APPENDIX C SURVEY INSTRUMENT : GROUP QUESTIONNAIRE C.1 Background 1(a) Main tribe(s) _____________ ________________ _______________ ______________ 1(b) Number of farmers present: i) Male [ ] ii) Female [ ] C.2 Land availability, cropping preference, and fallow period 2(a). Do most farmers in this village i) own land [ ] ii) rent land [ ] iii) borrow land [ ] 2(b) What indicates ownership of land by a farmer in your villages? i) village government letter/certificate of ownership [ ] ii) local government letter/ certificate of ownership [ ] iii) title deed [ ] iv) nothing [ ] iv) Dont know [ ] 2(c) Do most farmers in your villages have as much land as they need? i) Yes [ ] ii) No [ ] 2(d) If No what do they do to cope with the situation? _________________________________________ 3(a) What are the three most important cr ops grown for food in your villages for: Food crops _______________ ________________ ________________ 3(b) What are the three most important crops grown for income in this village for: Income crops _______________ ________________ ________________ 3(c) Which other important crops are grown in this village? Crops ________________ _________________ ________________ 4(a) What are the main crops grown in/around the wetland in the dry season? Crops ____________ _____________ ______________ 4(b) What are the main crops grown in/around the wet season? Crops ______________ ______________ ______________ 5. How long have vinyungu been cultivated? (Years) ____________ 6. What inspired/attracted people to cultivate vinyungu ? i) cheap/easy to get [ ] ii) close to the market [ ] iii) close to homes [ ] iv) high soil fertility [ ] v) dry season alternative (water availability) vi) close to peoples working places [ ] vi) other ________________ ___________ 7(a) Do farmers in your villages rotate crops in their wetland fields? i) Yes [ ] ii) No [ ] 7(b) If No why is that the case _____________________________________ 7(c) Do farmers in your villages rest their farms before being planted again? i) Yes [ ] ii) No [ ] 7(d) If No explain _____________________________________________________________________ 8(a) Are the wetland plots in your villages ir rigated? I) Yes [ ] ii) No [ ] 8(b) If Yes which methods of irrigatio n are used? ____________ _______________ ______________ 8(c) If No why is it the case? __________________________________________________________ C.3 Input, Tools and Labor 9. What is the most commonly used method of land preparation in the vinyungu in your villages? i) Hand hoeing [ ] ii) Ox plowing [ ] iii) Tractors [ ] 10(a) Do farmers in your villages use fertilizer in the vinyungu ? i) Yes [ ] ii) No [ ] 10(b) Have farmers always used fertili zer since they started working the vinyungu ? i) Yes [ ] ii) No [ ]

PAGE 108

108 10(c) If No, what might have triggered the use of fertilizer recently? i) loss of fertility in the soil [ ] ii) need to grow more and faster [ ] iii) it is become a habit to use fertilizer [ ] iii) other: ________________________________________________________ 10(d) What proportion would say use fertilizer in the vinyungu ? i) about of the farmers [ ] ii) about the farmers [ ] iii) about of the farmers [ ] iv) all [ ] v) none [ ] 10(e) For what crops is fertilizer used? Crops ____________ ______________ ______________ 10(f) How much does fertilizer cost in your villages? (TShs) ____________ 10(g) Can most people afford fertilizer in your villages? i) Yes [ ] ii) No [ ] 10(h) Was fertilizer available and enough for all fa rmers who were willing to pay the (higher) price? i) Yes [ ] ii) No [ ] Explain ____________________________________________ 11(a) Do farmers in your villages use manure in the vinyungu ? i) Yes [ ] ii) No [ ] 11(b) What proportion would you say use manure? i) about of the farmers [ ] ii) about the farmers [ ] iii) about of the farmers [ ] iv) all [ ] v) none [ ] 11(c) For what crops is the manure used? Crop ____________ ____________ ____________ _____________ Type of manure ____________ ____________ ____________ _____________ 11(d) Have farmers always used manure since they started working a kinyungu ? i) Yes [ ] ii) No [ ] 11(e) If No, what might have triggered the use of manure r ecently? i) loss of fertility in the soil [ ] ii) need to grow more and faster [ ] iii) it is a habit to use manure [ ] iv) easy way to discard family manure [ ] v) other: ________________________________________________________ 11(f) From where do farmers get most of the manure they use? i) own animals [ ] ii) buy [ ] 11(g) How much does manure cost in your villages? (TShs.) _________________ 12. Are there any other methods (local/traditional) that you use to improve the soil fertility of your land? i) Yes [ ] ii) No [ ] If yes, what other methods do you use? __________ ____________ 13(a) Do farmers in your villages use hired labor? i) Yes [ ] ii) No [ ] 13(b) Are hired laborers available whenever farmer s need them? ii) Yes [ ] ii) No [ ] 13(c) If no, explain why ______________________________________________________________ 14. From where do most farmers obtain the cash they need to pay farming expenses?________________ ____________________________________________________________________________________ C.4 Wetlands vs. Uplands 15. How important to you are the wetland fields for crop production cf. upland fields? i) Very important [ ] ii) Fairly Important [ ] iii) Not very important [ ] iv) Not at all important [ ] v) Cant choose [ ] 16. How important to you are the wetland fields for income cf. upland fields? i) Very important [ ] ii) Fairly Important [ ] iii) Not very important [ ] iv) Not at all important [ ] v) Cant choose 17. If cultivating in both wetland and upland, do most fa rmers have adequate time to work in both fields? i) Yes [ ] ii) No [ ] If No, what do you do about it? ______________________________ 18(a) Is there more or less renting of land now than in the past? I) More [ ] ii) Less [ ] iii) No change [ ] 18(b) What is the main reason? I) land unavailability [ ] ii) land very expensive [ ] iii) Other _________________ Cropping and Land Use about 35 years ago compared to today (relative to aerial photos of 1940s, 70s, 90s) Now I would like to know how the system has change d during the last 35 years that is from the late Mwalimu Julius Nyereres era to today the time of political and economic changes.

PAGE 109

109 19.Thirty-five years ago, what were the 3 main activitie s undertaken in the wetland by the villages around the Lyandembera? (in order of importance) ________________ _____________ ______________ 20(a) Thirty-five years ago, what were the 3 main crops for food in your villages (rank in order of priority)? _____________ ___________________ __________________ 20(b) Thirty-five years ago, what were the 3 main cr ops for income your villages (in order of priority)? ________________ _________________________ __________________ 21(a) Are there any new crops or varieties in this v illage now that were not grown thirty-five years ago? i) Yes [ ] ii) No [ ] 21(b) If yes, please list crops in order of importance? ____________ _____________ _____________ 21(c) When did farmers start to grow them? (Year) _________ 21 (d) What is the main driving force for farmers to grow the most popular new crop? i) quick return for money and effort [ ] ii) less laborious [ ] iii) fetch more money [ ] iv) needed to try different options [ ] v) strong marketing by outside sources [ ] vi) government directive [ ] vii) Other _____________ 22(a) Are there crops/varieties that you planted in the past that you no longer plant (or have sharply reduced planting)? i) Yes [ ] ii) No [ ] 22(b) If yes, which ones? ______________ _____________ ______________ 22(c) Why did you stop/reduce pl anting each of them? i) unmarketable [ ] ii) laborious [ ] iii) too costly to grow [ ] iv) seeds not easily available [ ] v) hard to store [ ] vi) take too long to grow [ ] vii) easily affected by diseases/pest/animals [ ] viii) Other ___________________ 23(a) How can upland yields be compared to the past? i) in creased [ ] ii) decreased [ ] iii) same [ ] 23(b) How much maize (or whatever is popularly grown in the uplands) do you produce today compared to the 1970s? Upland crop 1970s yield 2005 yields 24(a) How can wetland yields be compared to the past? i) increas ed [ ] ii) decreased [ ] iii) same [ ] 24(b) How much crop (popularly grown in the wetlands) do you produce today compared to the 1970s? Wetland crop 1970s yield 2005 yield 25(a) How can fertilizer cost be compared to the past? i) cheap er [ ] ii) more expensive [ ] ii) more or less the same [ ] 25(b) How has this affected use of fertilizer? i) reduced [ ] ii) increased [ ] iii) no change [ ] 25(c) (If use has declined) How do farmer s compensate for using less fertilizer? i) use manure [ ] ii) fallowing [ ] iii) crop rotation [ ] iv) Nothing [ ] 25(d) How has reduced use of fertilizer affected your crop ou tput? i) reduced [ ] ii) increased [ ] iii) No change [ ] 25(e) How has reduced use of fertilizer affected your choice of crops or cropping patterns? ______________________________________________________________________________________ 26. How would you describe the trends in wetland utilization? i) uses have increased tremendously [ ] ii) uses have incr eased slightly [ ] iii) uses have remained the same [ ] iv) uses have declined [ ] v) I dont know [ ] 27. How would you compare the current wetland size to that of the 1970s? i) Has decreased [ ] ii) Has increased [ ] iii) No changes [ ] iv) Dont know [ ] 28. How would you compare the current vegetation cover/ natural state of the wetland to that of the 1970s? I) has increased/improved [ ] ii) has decreased/declined [ ] iii) no change [ ] iv) Dont know [ ] 29. How would you compare the current soil fertility to that of the 1970s? I) more fertile [ ] ii) less fertile [ ] iv) very unfertile [ ] v) Do not know [ ]

PAGE 110

110 30. What would you say about the current state of the soil erosion in this area especially around the wetland? I) No erosion [ ] ii) some erosion [ ] iii) highly eroded [ ] iv) Dont know [ ] 31. What would you say about relations over use of water for cultivation in the wetland? I) no conflict [ ] ii) some/rare conflicts [ ] iii) rare but huge/deadly conflicts [ ] iv) conflicts are often but not huge [ ] v) conflicts are often and very huge [ ] vi) Dont know [ ] 32. What would you say about the population trends? I) has increased [ ] ii) has declined [ ] iii) no change [ ] iv) Dont know [ ] 33. Why do you think people chose this area for valley-bottom cultivation? i) lack of land [ ] ii) soil fertility [ ] iii) water availability [ ] iii) Other______________ 34(a) Do you perceive agricultural activities detrimental to the wetland? I) Yes [ ] ii) No [ ] 34(b) If Yes, what has been the bigg est impact on wetland resources? i) redu ced water flow [ ] ii) reduced soil fertility [ ] iii) soil erosion [ ] iv) human conflict [ ] v) other ______________ 35. What would you say is the trend with rainfall comparin g today and the 1970s? I) has increased [ ] ii) has decreased [ ] iii) has remained the same [ ] iv) Do not know [ ] 36(a) Are there any management plans in your village s with regards the future of the wetland resources? I) Yes [ ] ii) No [ ] 36(b) If No, what would you say should be the number one undertaking/consideration for the future of the wetlands? i) drained/ further opened to economic [ ] ii) be left alone to recover [ ] iii) a land use plan developed to balance human and environment needs v) maintain the current status quo [ ] iv) Other __________________________________________________________________ Explain ___________________________________________________________________________ 37(a) Have you ever been directed or advised by the government on wise use strategies for this wetlands? i) Yes [ ] ii) No [ ] If yes, what kind of directives/advise? _______________________________________________________ 37(b) Why do you think the government has not reduced or banned valley-bottom cultivation yet? ______________________________________________________________________________________ 37(c) What do you think the government should do with regards valley-bottom cultivation? I) ban it [ ] ii) promote sustainable cultivation through participatory management planning [ ] iii) should not interfere [ ] iii) Other ________________________________________________________

PAGE 111

111 LIST OF REFERENCES Acreman, M.C., and G.E. Hollis. 1996. Water Management and Wetlands in Sub-saharan Africa IUCN, Gland, Switzerland. Adams, W.M. 1993a. Agriculture, grazing and forestry. In The Hadejia-Nguru wetlands: Environment, economy and sustainable d evelopment of a Sahelian floodplain wetland, ed. G.E. Hollis, W.M. Adams, and M. Amini-Kanu, 89-96. IUCN, Gland, Switzerland. Adams, W.M. 1993b. Indigenous use of wetlands a nd sustainable development in West Africa. The Geographical Journal 159:209-218. Adams, W.M. 1996. Economics and hydrologica l management of African floodplains. In Water Management and Wetlands in Sub-saharan Africa ed. M.C. Acreman, and G.E. Hollis, 211-214. IUCN, Gland, Switzerland. Alexandratos, N. 1995. World agriculture towards 2010 FAO, Rome, Italy. Alonso, W. 1964. Location and land use: Towards a general theory of land rent Cambridge, Massachusetts: Harvard Univesrity Press. Balek, J., and J.E. Perry. 1973. Hydrology of s easonally inundated Afri can headwater swamps. Journal of Hydrology 19:227-249. Banzi, F.M., Ph.A. Kips, D.N. Kimaro, and J.D.J Mbogoni. 1992. Soil appraisal of four village irrigation schemes in Mwanga District, Kilimanj aro Region (Kirya, Kileo, Mvuleni, and Kigonigoni) National Soil Service (NSS) Report. Agri cultural Research Institute (ARI) Mlingano, Tanga, Tanzania. P.O. Box 5088, Tanga, Tanzania. Telephone: (+255) 27 2647647. Email: mlingano@iwayafrica.com Bilsborrow, R.E. 1987. Population pressures and agricultural devel opment in developing countries: A conceptual framework and recent evidence. World Development 15:183-203. Bilsborrow, R.E. and H.W.O. Ogendo. 1992. Population-driven cha nges in land use in developing countries. Ambio 21: 37-45. Binford, M.W., A.L. Kolata, M. Brenner, J. W. Janusek, M.T. Seddon, M. Abbott, and J.H. Curtis. 1997. Climate variation and the rise and fall of an A ndean civilization. Quarterly Research 47:235-248. Binns, T. 1994. Tropical Africa Routledge, London. Blaikie, P.M. 1971. Spatial organization of agri culture in some Indian village, Part II. Transactions of the Institute of British Geographers 53:15-30. Boserup, E. 1965. The conditions of agricultural grow th: The economics of agrarian change under population pressure Allen and Unwin, London.

PAGE 112

112 Boserup, E. 1981. Population and technological change: A study of long term trends University of Chicago Press, Chicago. Brace, S. 1995. Participatory rural appraisal: A significant step forward in understanding relationships between poor peopl e and their environment. In People and Environment in Africa ed. T. Binns, 39-46. Wiley and Sons, Chichester, UK. Briassoulis, H. 2000. Analysis of land use ch ange: Theoretical and modeling approaches. In The web book of regional science ed. S. Loveridge, Regional Research Institute, West Virginia University http://www.rri.wvu.edu/WebBook/ Briassoulis/contents.htm (August 2006). Briassoulis, H. 2000. Analysis of land use ch ange: Theoretical and modeling approaches. In The web book of regional science ed. S. Loveridge, Regional Research Institute, West Virginia University http://www.rri.wvu.edu/WebBook/Briasso ulis/Chapter1(introduction).htm#1.3 (August 2006). Briassoulis, H. 2000. Analysis of land use ch ange: Theoretical and modeling approaches. In The web book of regional science ed. S. Loveridge, Regional Research Institute, West Virginia University http://www.rri.wvu.edu/WebBook/Briasso ulis/Chapter3(Theories).htm#3.3 (August 2006). Briassoulis, H. 2000. Analysis of land use ch ange: Theoretical and modeling approaches. In The web book of regional science ed. S. Loveridge, Regional Research Institute, West Virginia University http://www.rri.wvu.edu/WebBook/Briasso ulis/Chapter3(Theories).htm#3.4 (August 2006). Burgi, M., and M.G. Turner. 2002. Factors a nd processes shaping land cover and land-cover changes along the Wisconsin River, USA. Ecosystems 5:184-201. Carney, H.J., M.W. Binford, A.L. Kolata, R.R. Marin, and C.R. Gold man. 1993. Nutrient and sediment retention in Andean raised-field agriculture. Nature 364:131-133. Chambers, R. 1994. Participatory rural appr aisal (PRA): Analysis of experience. World Develoment 22:1253-1268. Chisholm, M. 1962 Rural settlement and land use: An essay in location Hutchinson University Library, London. Chomitz, Kenneth M., and David A. Gray. 1995. Roads, lands, markets, and deforestation: A spatial model of land use in Belize World Bank Policy Research Working Paper 1444. World Bank, Policy Research De partment, Washington, D.C. Culwick, A.T., and G.M. Culwick. 1935. Ubena of the rivers Allen and Unwin, London. Daily, G. 1997. Introduction: What are ecosystem services?. In Natures services: Societal dependence on natural ecosystems ed. G. Daily, 1-19. Island Press, Washington, DC.

PAGE 113

113 DANIDA. 1982. Water master plan for Iringa, Ruvu ma and Mbeya Regions. Irrigation and Hydropower. Volume 11. Carlbro, Cowiconsult, Kampsax-Kruger. Dar es Salaam, Tanzania. Denevan, W. 1970. Aboriginal draine d-field cultivation in the Americas. Science 169:647-654. Denny, P., and F. Turyatunga. 1992. Ugandan wetlands and their management. In Conservation and development: The sustainable use of wetland resources, Proceedings of the Third International Wetlands Conference, Rennes, France, 19-23 September 1988 ed. E. Maltby, P.J. Dugan, J.C. Lefeuvre, 77-84. IUCN, Gland, Switzerland. Dixon, A.B., and A.P. Wood. 2003. Wetland cultiva tion and hydrological management in eastern Africa: Matching community and hydrological needs through sustainable wetland use. Natural Resources Forum 27:117-129. Dugan, P.J. 1990. Wetland conservation: A review of current issues and action IUCN, Gland, Switzerland. Elmekki, A., and J. Barker. 1993. Marketing, politics and agrarian crisis in Sudan. IDS Bulletin 24:68-74. Erickson, C.L. 1985. Applications of prehistoric Andean technology: Experiments in raised field agriculture, Huatta, Lake Titicaca, Peru, 1981-1983. In Prehistoric intensive agriculture in the tropics ed. I. Farrington, 209-232, British Archaeological Reports, International Series, No. 232, Oxford. Erickson, C.L. 1988. Raised field agriculture in the Lake Titicaca basin: Putting ancient agriculture back to work. Expedition 30:8-16. Erickson, C.L. 1999. Neo-environmental determ inism and agrarian c ollapse in Andean prehistory. Antiquity 73:634-642 Erickson, C.L. 2000. The Lake Titicaca basin: A pre-Columbian built landscape. In Imperfect balance: Landscape transformations in the pre-Columbian Americas ed. D. Lentz, 311356. Columbia University Press, New York. Food and Agriculture Organization (FAO). 1980 World Census of Agriculture FAO, Rome, Italy. Fowler, M.L. 1969. Middle Miss issippian agricultural fields. American Antiquity 34:365-375. Gerakis, A., and K. Kalburtji. 1998. Agricultural activities affec ting the functions and values of Ramsar wetland sites of Greece. Agriculture, Ecosystems, and Environment 70:119-128. Gichuki, F.N. 1992. Utilization and conservatio n of wetlands: An agricultural drainage perspective. In Wetlands of Kenya, Proceedings of the KWWG Seminar on Wetlands of Kenya, Nairobi, Kenya, 3-5 July 1991 ed. S.A. Crafter, S.G. Njuguna, G.W. Howard, 147154. IUCN, Gland, Switzerland.

PAGE 114

114 Hollis, G.E. 1990. Environmental impacts of deve lopment on wetlands in arid and semi arid lands. Hydrological Sciences Journal 35:411-428. Jensen, J.R., S. Narumalani, O. Weatherbee, and J.R Mackey. 1993. Measurement of seasonal and yearly cattail and waterlily changes using multi date SPOT panchromatic data. Photogrammetric Engineering and Remote Sensing 59:519-525. Jensen, J.R., K. Rutchey, M.S. Koch, and S. Narumalani. 1995. Inland wetland change detection in the Everglades Water Conservation Area 2A using a time series of normalized remotely sensed data. Photogrammetric Engineering and Remote Sensing 61:199-209. Jensen, J.R. 1996. Introductory digital image proce ssing: A remote sensing perspective Prentice-Hall, Inc., Upper Saddle River, New Jersey. Jones, C. 1997. Geographic Information System s and Computer Cartography Addison Wesley Longman Limited, London, Harlow, England. Kaswamila, A.L., and A.J.M. Tenge. 1997. The neglect of traditional agr oforestry and its effects on soil erosion and crop yield in the west Usambara uplands in Tanzania A Research on Poverty Alleviation (REPOA) report. Contact: 157 Migombani Street, Regent Estate, P O Box 33223, Dar es Salaam, Tanzania. Phone: (+255) 22 2700083 or 2772556. Fax: 25522-2775738.Email: repoa@repoa .or.tz. Kuroda, M. 2001. The development process of comm ercial tomato producti on in Iringa region, southern highland, Tanzania. Journal of African Studies 59: 33-52. Lampert, R.J. 1967. Horticulture in the New Guinea highlands, C 14 Dating. Antiquity 41:307309 Lambin, E.F., X. Baulies, N. Bockstael, G. Fisc her, T. Krug, R. Leemans, E.F. Moran, R.R. Rindfuss, Y. Sato, D. Skole, B.L. Turner II, and C. Vogel. 1999. Land-use and land-cover change (LUCC): Implementation strategy IGBP Report No. 48, IHDP Report No. 10, IGBP and IHDP, Stockholm and Bonn. Lambin, E.F., B.L. Turner, H.J. Geist, S.B. Agbo la, A. Angelsen, J.W. Bruce, O.T. Coomes, R. Dirzo, G. Fischer, C. Folke, P.S. George, K. Homewood, J. Imbernon, R. Leemans, X. Li, E.F. Moran, M. Mortimore, P.S. Ramakrishnan, j.F. Richards, H. Skanes, W. Steffen, G.D. Stone, U. Svedin, T.A. Veldkamp, C. Vogel, and J, Xu. 2001.The causes of land-use and land-cover change: Moving beyond the myths. Global Environmental Change 11:261-269. Lele, U., and S.W. Stone. 1989. Population pressure, the envi ronment, and agricultural intensification: Variations on the Boserup hypothesis MADIA Discussion Paper 4. World Bank, Washington, D.C. Lema, A.J. 1996. Cultivating the valleys: vinyungu farming in Tanzania. In Sustaining the soil: Indigenous soil and water conservation in Africa ed. C. Reij, I. Scoones, and C. Toulmin, 139-144. Earthscan, London.

PAGE 115

115 Lipton and J. van der Gaag. 1993. Including the Poor World Bank, Washington, D.C. Liu, H., S. Zhang,, Z. Li, X. Lu, and Q. Yang. 2004. Impacts on wetlands of large scale land-use changes by agricultural development: The small Sanjiang plain, China. Ambio 3:306-310 Liverman, D. 1994. Modeling social systems and th eir interactions with the environment: A view from Geography. In Integrated regional models: Inte ractions between humans and their environment ed. P.M. Groffman and G.E. Likens, 67-78. New York: Chapman and Hall. Loevinsohn, M.E., J. Mugarura, and A. Nkusi. 1992. Group innovation in utilizing land and water resources in Rwandan valleys ODI Irrigation Management Paper. 13:3-14 Losch, A. 1954. The economics of location New Haven, CT: Yale University Press/Oxford: Oxford University Press (original German edition, 1940). Majule, A.E. and R.B.B Mwalyosi, 2003. Enhancing agricultural productivity through sustainable irrigation: A case of vinyungu farming system in selected zones of Iringa Research report submitted to ENRECA (Enha ncing Research Capacity), Institute of Resource Assessment, University of Dar es Salaam, PO Box 35097, Dar es Salaam, Tanzania. Phone and Fax: (+255) 51 410393. Email: ira@udsm.ac.tz. Malthus, T.R. 1960. On population (from: An essay on the prin ciple of population. 1766-1834). New York: Modern Library for Random House. Mascarenhas, A., J. Ngana, and M. Yoshida. 1985. Opportunities for irri gation development in Tanzania Joint Research Program (JRP) Series 52. Institute of Developing Economies, Tokyo, Japan. Masija, E.H. 1993. Irrigation of wetlands in Tanzania. In Wetlands of Tanzania: Proceedings of a seminar on the wetlands of Tanzania, Morogoro. Tanzania, 27-29 November, 1991 ed. G.L. Kamukala and S.A. Crafter eds, 73-83. IUCN, Gland, Switzerland. McIntire, J., Daniel Bour zat, and Prabhu Pingali. 1992. Crop-livetsock interaction in SubSaharan Africa The World Bank, Washington D.C. Meertens, H.C.C. 1999. Rice cultivation in the farming sy stems of Sukumaland, Tanzania: A quest for sustainable production under structural adjustment programs Tanzania Royal Tropical Institute. KIT Press, Amsterdam, The Netherlands. Merchant, C. 1990. The realm of social rela tions: Production, reproduc tion, and gender in environmental transformations. In The earth as transformed by human action ed. B.L. Turner II, W.C. Clark, R.W. Kates, J.E. Ri chards, J.T. Mathews, and W.B. Meyer, 672684. Cambridge: Cambridge University Press. Mertens, B., W.D. Sunderlin, O. Ndoye, and E.F Lambin. 2000. Impact of macroeconomic change on deforestation in south Cameroon: In tegration of household survey and remotely sensed data. World Development 28:983.

PAGE 116

116 Meyer, W.B. and B.L. Turner II. 1994. Changes in land use and land cover: A global perspective Cambridge University Press, Cambridge. Meyer, W.B. and B.L. Turner II. 1992. Huma n population growth and global land-use/ cover change. Annual Review Ecol ogy and Systematics 23:39-62. Meyer, W.B. and B.L. Turner II. 1996. Land-Use/Land-Cover Change: Challenges for Geographers. Geojournal 39:237-240. Ministry of Agriculture. 1992. Luganga smallholder irrigation re habilitation project. Appraisal Report No. PPU 19/92. Project Preparation Mo nitoring Bureau. Contacts: P.O. Box 5384, Dar es Salaam, Tanzania. Phone: (+255) 22 2864069. Email: psk@kilimo.go.tz Mkavidanda, T.A.J. and A.L. Kaswamila. 2001. The role of traditional irrigation systems (v inyungu ) in alleviating poverty in Iringa rural district, Tanzania. Research on Poverty Alleviation Report No. 01.2. Contacts: 157 Mi gombani Street, Regent Estate, P O Box 33223, Dar es Salaam, Tanzania. Phone: 255-22-2700083 or 2772556. Fax: 255-222775738. Email: repoa@repoa .or.tz Morgan, W.B. and J.A. Solarz. 1994. Agricultural crisis in Sub-Saharan Africa: Development constraints and policy problems. Geographical Journal 160:57-73). Munyati, C. 2000. Wetland change detection on the Kafue flats, Zambia, by classification of a multitemporal remote sensing image dataset. International Journal of Remote Sensing. 9:1787-1806 Myers, N. 1994. Tropical deforestation: rates and patte rns. In The causes of tropical deforestation ed. K. Brown and D.W. Pearce, 27-41. University College London Press, London. Mrema, C.G. 1984. Development of smallholder i rrigation in Tanzania: Pr oblems and prospects. In African regional symposium on smallholde r irrigation, Harare, Zimbabwe, 5-7 September, 1984 ed. M.J. Blackie, 70-79. Hydraulics Research Ltd, London, UK. United Republic of Tanzania (URT). 1999. Iringa region socioeconomic profile Planning Commission, DSM and Regional Commi ssionals Officer, Iringa, Tanzania. United Republic of Tanzania (URT). 2003. 2002 Population and Housing Census: General report National Bureau of Statistics, Presidents Office, Planning and Pr ivatization, Dar es Salaam, Tanzania. United Republic of Tanzania (URT). 1997. Iringa region socioeconomic profile Planning Commission, DSM and Regional Commi ssionals Officer, Iringa, Tanzania. United Republic of Tanzania (URT). 2004a. Mufindi district socioeconomic profile Planning Commission, DSM and Mufindi dist rict council, Iringa, Tanzania.

PAGE 117

117 United Republic of Tanzania (URT). 2004b. Iringa district socioeconomic profile Planning Commission, DSM and Iringa distri ct council, Iringa, Tanzania. Netting, R. McC. 1968. Hill farmers of Nigeria: Cultural ec ology of the Koyfar of the Jos Plateau University of Washington Press, Seattle and London. Netting, R.McC. 1993. Smallholders, householders: Farm families and the ecology of intensive, sustainable agriculture Stanford: Stanford University Press Okeyo, D.O. 1992. Wetland fish of Kenya. In Wetlands of Kenya. Proc eedings of a seminar on wetlands of Kenya, Nairobi, Kenya, 3-5 July ed. S.A. Crafter, S.G. Njuguna and G.W. Howard, 47-53. IUCN, Gland, Switzerland. Olindo, P. 1992. Food policy and wetlands. In Wetlands of Kenya. Proc eedings of a seminar on wetlands of Kenya, Nairobi Kenya, 3-5 July 1991 ed. S.A. Crafter, S.G. Njuguna, G.W Howard, 174-180. IUCN, Gland, Switzerland. Ortloff, C. and A. Kolata. 1993. Climate and co llapse: Agroecological perspectives on the decline of the Tiwanaku state. Journal of Archaeological Science 20:195-221 Pallela, E. 2000. The impact of anthropogeni c factors on urban wetlands : The case of Msimbazi valley, Dar es Salaam A dissertation submitted in partial fulfillment of the requirements for the degree of Master of Arts (Geogra phy and Environmental Management), University of Dar es Salaam, Dar es Salaam, Tanzania. Pingali, P., Y. Bigot, and H.P. Binswanger. 1987. Agricultural mechanization and the evolution of farming systems in Sub-Saharan Africa John Hopkins University Press, Baltimore, M.D. Pinstrup-Andersen, P., R. Pandya-Lorch, and S. Babu. 1997. A 2020 vision for food, agriculture, and the environments in southern Africa. In Achieving food security in Southern Africa: New challenges, new opportunities ed. L. Haddad, 16-58. International Food Policy Research Institute (IFPRI), Washington, D.C. Ramsar Convention Bureau. 1990. Directory of wetlands of international importance Ramsar Convention Bureau, Gland, Switzerland. Richards, J.F. 1990. Agricultural impacts in tr opical wetlands: Rice pa ddies for mangroves in South and Southeast Asia. In Wetlands: A threatened landscape ed. M. Williams, 217233. Oxford, UK, Blackwell Publishers. Ringrose, S., W. Matheson, and T. Boyle. 1988. Differentiation of ecol ogical zones in the Okavango Delta, Botswana, by classification a nd contextual analyses of Landsat MSS data. Photogrammetric Engineering and Remote Sensing 54:601-608 Roggeri, H. 1995. Tropical Freshwater Wetlands: A Guide to Current Knowledge and Sustainable Management H.J. Dumont (ed.). Kluwer Academic Publishers, Drodrecht, The Netherlands.

PAGE 118

118 Saith, A. 1992. The rural non-farm economy: Process and policies International Labour Office, World Employment Programme Geneva:, Switzerland. Scoones, I. 1991. Wetlands in drylands: Key resour ces for agricultural and pastoral production in Africa. Ambio Vol. 20:366-371. Silvius, M.J., M. Oneka, A. Verhagen. 2000. Wetlands: Lifelines for people at the edge. Physics and Chemistry of the Earth Part B 25:645-652. Ministry of Water and Livestock Development (MWLD). 2001. Sustainable Management of the Usangu Wetland and Catchments (SMUWC): Talking about Usangu, Project Brochure. www.usangu.org (August 2006). Stern, P. 1989. Small-scale irrigation Intermediate technology publication. London, UK. Stuip, M.A.M., C.J. Baker, and W. Oosterberg. 2002. The socio-economics of wetlands Wetlands International and RIZA, The Netherlands. Sutton, J.E.G. 1969. Ancient civi lizations and modern agricultural systems in the southern highlands of Tanzania. AZANIA 4:1-14. Schwidefsky, K. 1959. An outline of photogrammetry Pitman Publishing Corporation, London, UK. Thompson, K. 1976. Swamp development in the h eadwaters of the White Nile. pp 177-196. In The Nile: biology of an ancient river, ed. J. Rzoska, Monographiae Biologicae, 29 W. Junk, b.v. The Hague, The Netherlands. Tiffen, M. and M. Martimore. 1994. Malthus controverted: the ro le of capital a nd technology in growth and environmental recovery in Kenya. World Development 22:997-1010. Tiffen, M., M. Mortimore, and F. Gichuki. 1994. More people, less erosion: Environmental recovery in Kenya. John Wile y & Sons Ltd., Chichester. Turner II, B. L. 1994. Prehistoric intensive ag riculture in the Mayan lowlands. Science 185:118124. Turner II, B. L. and P. D. Harrison. 1981. Pr ehistoric raised-field ag riculture in the Maya lowlands. Science 213: 399-405. Turner II, B.L., G. Hyden, and R.W. Kates. 1993. Population growth and agricultural change in Africa. University of Florida Press, Gainesville. Turner II, B.L., W.B. Meyer, and D.L. Skol e. 1994. Global land-use/ land-cover changes: towards an integrated study. Ambio 23: 91-95.

PAGE 119

119 Verstraete, M.M., B. Pinty, and R. Myneni. 1996. Potential and limitations of information extraction on the terrestrial biosphere from satellite remote sensing. Remote Sensing of Environment. 58:201-214. von Thnen, J.H. 1966. Von Thnens Isolated Stat e. P. Hall (Ed.). London: Pergamon Press. Originally published 1826 Walker, R. and A.K.O Homma. 1996. Land use and land cover dynamics in the Brazilian Amazon: an overview. Ecological Economics 18:67-80 Wang, Z., B. Zhang, S. Zhang, X. Li, D. Liu, K. Song, J. Li, F. Li, and H. Duan. 2006. Changes of land use and of ecosystem service valu es in Sanjiang plain, northeast China. Environmental Monitoring and Assessment 112:69-91 Weber, A. 1929. Alfred Webers theory of the loca tion of industries, translated by C.J. Friedrich. University of Chicago Press, Chicago (first German Edition 1909). Wiebe, K. 2003. Linking land quality, agricultu ral productivity, and food security. Agricultural Economic Report No. (AER823) Williams, M. 1991. Wetlands: A threatened landscape. Alden Press Ltd. Oxford, UK and Blackwell Publishers, Cambridge, Massachusetts. Wilk, R.R. 1991. Household ecology: Economic change and domestic life among the Kekchi Maya in Belize. Tucson: University of Arizona Press. Wood, A.P. 1996. Wetland drainage and mana gement in south-west Ethiopia: Some environmental experiences of an NGO. In The Sahel workshop 1996, ed. A. Reenburg, H.S. Marcusen, I. Nielsen, 119-136. Institut e of Geography, University of Copenhagen Copenhagen. World Bank Review. 1989. Sub-Saharan Africa: From crisis to sustainable growth: A long-term perspective study. World Bank, Washington D.C.

PAGE 120

120 BIOGRAPHICAL SKETCH Dismayed by the rate at which the beaches in her native Tanzania were being degraded, Lucy embarked on and dedicated her time to st udying biological scien ces. She graduated a 3year B.Sc. degree program in zoology and ma rine biology from the University of Dar-EsSalaam, Tanzania in 1993. Her intentions were to help in her countrys efforts of reversing the degradation of its aquatic resour ces that were receiving less c onservation attention then. Lucy was immediately employed by one of the world s largest conservation organizations, the World Wide Fund for Nature (WWF), to manage it s aquatic resources c onservation programs. With WWF, Lucy successfully facilitated the process of developing (and then manage) several conservation programs including establis hing Tanzania mainlands first marine Park (Mafia Island Marine Park) and Menai Bay Conservation Area in Zanzibar. Lucy also managed one of WWFs major and challenging programs in Tanzania, the Ruaha Water Program. Overall, her work involved integrating human-environment dimensions of natural resources management, to promote sustainable development. Having wo rked with WWF for over 9 years, Lucy has accumulated a wealth of knowledge and e xperience in developing and implementing conservation and development programs. She has a broad experience in strategic planning as well as in negotiating agreements related to natura l resources management with relevant sectors. Her membership in WWF's Marine Adviso ry Group (1993 to 1995) and WWF's Global Freshwater Team (2002 to 2005) enabled her to shar e lessons with the global network and gain a global perspective to water resources management further improving her management strategies. She has also taken the opportunity to gain ex tra qualifications (e.g., Integrated River Basin Management through the Training of Trainers Course in Wetla nds Management Wageningen, Netherlands; Coastal Zone Management University of Rhode Island, USA) to improve her management strategies.

PAGE 121

121 These experiences inspired Lucy to concen trate her further studi es on the ecology and conservation of aquatic ecosystems, especially on wetland ecology and management, as well as on the application of geographic technologies to environmental ma nagement. She graduated with a Master of Science degree in Geography from the University of Florida in Spring 2007.