Citation
Spring Focus on Social Science Research; Analysis of Rainfall Variability in Relation to Crop Production in Maun, Botswana

Material Information

Title:
Spring Focus on Social Science Research; Analysis of Rainfall Variability in Relation to Crop Production in Maun, Botswana
Series Title:
Journal of Undergraduate Research
Creator:
Muir, Carly
Place of Publication:
Gainesville, Fla.
Publisher:
University of Florida
Publication Date:
Language:
English

Subjects

Genre:
serial ( sobekcm )

Notes

Abstract:
The purpose of the study was to analyze patterns of the average rainfall characteristics during the growing season in Maun before and after a climatic shift that took place in the 1970’s and to assess how this variability affects risks of crop production at specified planting dates. Plots show that the majority of the water year experienced a decrease in the mean of both rainfall total and the count of rainy days. The graph developed can show probabilities of risk within a specified range and could help farmers make decisions about which types of crops to grow and when to plant as they adapt to the new conditions.

Record Information

Source Institution:
University of Florida
Holding Location:
University of Florida
Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
UF00091523_00602 ( sobekcm )

Downloads

This item is only available as the following downloads:


Full Text

PAGE 1

University of Florida | Journal of Undergraduate Research | Volume 1 6 , Issue 2 | Spring 2015 1 Analysis of Rainfall Variability in Relation to Crop Production in Maun, Botswana Carly Muir College of Liberal Arts and Sciences , University of Florida The purpose of the study was to analyze patterns of the average rainfall characteristics during the growing season in Maun before and after a climatic shift that took place in the 1970’s and to assess how this variability affects risks of crop production at s pecified planting dates. Plots show that the majority of the water year experienced a decrease in the mean of both rainfall total and the count of rainy days. The graph developed can show probabilities of risk within a specified range and could help farmers make decisions about wh ich types of crops to grow and when to plant as they adapt to the new conditions. INTRODUCTION As climate change manifests in the twenty first century , Africa will experience many alterations in its climate. Although there is some debate about whether precipitation will increase or decrease, and where, it is widely believed rainfall is becoming more variable and will decline (Muller). The impact of change in rainfall regime will be severe in semi arid to arid regions that already exhibit water scarcity. Sub Saharan Africa is expected to be one of the most impacted regions by the affects of climate change, particularly the rain fed agriculture farming systems that dominate the continent, and is responsible for about 90 percent of Africa’s staple food (Cooper, Dimes, Rao, Shapiro). Agriculture supports more than 60 percent of the over all African population, and 80 percent of people living in poverty (AQUASTAT). The advancement of agriculture, and ability to cope with changes in climate will be a defining element for Africa’s development. As population rises, intensifying water scarcity, th e importance of rain fed farming is increasing making it vital to understand rainfall variability and its affect. In the future it is anticipated that there will be a greater frequency of both extreme drought and excess rains, both of which can be deleteri ous to crops and soil. In the absence of reliable estimates of future climate, one may often turn to historic shifts in climate regime that were recorded in the historic record to provide possible analogs to future scenarios. In this paper, the shift in gl obal climate noted to have occurred in the late 1970s is examined in the historic values of two rainfall variables important to rain fed agriculture, the total rainfall and the number of rainy days during the growing season, at Maun Botswana, as it may hav e impacted the rainfall regime 195075 and 1976 92. The research is also part of a larger NASA funded research project in the Department of Geography, investigating the potential roles of climate change and human activity upon vegetation throughout a Botsw ana, Namibia, Angola and Zambia. STUDY AREA & PURPOSE Botswana is centrally located in southern Africa encompassing both semi arid and arid conditions. Botswana’s climate is known to vary locally. The northeast receives more than 400mm of precipitation annually and rainfall declines towards the southwest culminating in the Kalahari Desert climate conditions. The desert currently covers over two thirds of Botswana, making those parts unsuitable for agriculture (lund and Dahlberg) . Maun lies in the northwes t on the margins of the Kalahari as shown in Figure 1; unlike other regions of the country this area receives stream flow seasonally from the Okavango River on whose delta the city is located. Even with this additional potential source of water from the Okavango, less than 1% of Botswana’s agricultural lan d is irrigated (FAO). The purpose of the study is to analyze patterns of the average rainfall characteristics during the growing season in Maun before and after a climatic shift that took place in the 1970’s and to assess ho w this variability affects risks of crop production at specified planting dates. Data was collected from years 1950 to 1991, and split into two periods, as seen in F igure 2 , to signify the climatic shift in Figure 1. Kalahari Desert

PAGE 2

CARLY MUIR University of Florida | Journal of Undergraduate Research | Volume 1 6 , Issue 2 | Spring 2015 2 the 1970’s. The analysis is specifically design ed to assess these changes in a manner pertinent to determining risks to crop production. Historic daily precipitation data are available from 1923; however those data from 1950 to 1975 (period 1) appear to best reflect characteristics before the climate shift and those of 19761991 following the shift (period 2). The daily data were originally coded on the basis of a calendar year, however, given the strong seasonality of rainfall with a peak in the southern summer. These were subsequently recoded accordin g to a water year which spans July 1 to the following June 30. The water year now encompasses the entire rainy season and the year is divided at a time when rainfall is almost completely absent in the historic record. Figure 3 shows mean monthly precipit ation at Maun before and after 1975. Heavy horizontal lines represent means, thinner ones medians. The tops and bottoms of the boxes indicate 25th and 75th percentiles, the whiskers 10th and 90th, and the dots 5th and 95th. 30 60 90 120 150 180 210 240 270 300 30 35 40 45 50 55 60 30 60 90 120 150 180 210 240 270 300 30 35 40 45 50 55 60 Starting Date (Julian Days) 30 60 90 120 150 180 210 240 270 300 Duration (days) 30 35 40 45 50 55 60 30 35 40 45 50 55 60 Starting Date (Day in Water Year) 30 60 90 120 150 180 210 240 270 300 Duration (days) 30 35 40 45 50 55 60 205 193 202 154 168 203 203 197 150 150 96300 300 300 3001950 1951 1952 1950-75 1991 1992 1976-1992 Annual Variables Rainfall total (mm) Rainy days (days) 6 9 9 12 13 13 15 17 10 9 16 Figure 2 . Organization of D ata Figure 3 . Mean M onthly Pprecipitation B efore and A fter 1975

PAGE 3

ANALYSIS OF RAINFALL VARIABILITY IN RELATION TO CROP PRODUCTION IN MAUN, BOTSWANA University of Florida | Journal of Undergradua te Research | Volume 1 6 , Issue 2 | Spring 2015 3 Because each variable represents the sum of at least 30 days of daily rainfall totals or occasions when rain occurred or did not occur, it was initially assumed that their totals (total rainfall or total number of days with rain in the prescribed period) w ill approximate a normal distribution, as a result of the Central Limit Theorem (Burt, Barber, Rigby). The normal probability density function, f(x): ( ) = 1 2 . { ( ) } the standard deviation of the variable. The normal probability mass function, F(x
PAGE 4

CARLY MUIR University of Florida | Journal of Undergraduate Research | Volume 1 6 , Issue 2 | Spring 2015 4 0 100 200 300 400 500 Probability Density Function 0.000 0.002 0.004 0.006 0.008 0 100 200 300 400 500 0.000 0.002 0.004 0.006 0.008 Rainfall Total (mm) 0 100 200 300 400 500 Probability Density Function 0.000 0.002 0.004 0.006 0.008 Rainfall Total (mm) 0 100 200 300 400 500 0.000 0.002 0.004 0.006 0.008 Period 1 Period 2 1. 2. 3. 4. Figure 4 . Possible Outcomes o f Dif ferences Between Periods 1 a nd 2 Figure 5 . Difference i n Mean a nd Standard Deviation Between Period 1 And 2 Of Days With Rain Figure 6 . Difference Between t he Total Precipitati on Mean And Standard Deviation i n Period 1 a nd 2

PAGE 5

ANALYSIS OF RAINFALL VARIABILITY IN RELATION TO CROP PRODUCTION IN MAUN, BOTSWANA University of Florida | Journal of Undergradua te Research | Volume 1 6 , Issue 2 | Spring 2015 5 process of which crop to plant by allowing farmers to see how rainfall patterns have changed using precise planting dates and prescribed length of durations of the growing season. Positive values on the graphs signify that period 2 had a lower probability of rece iving the specified value of rainy days, and negative numbers imply an increase in probability. F igure 6a indicates the risk of experiencing less than 9 days of rain has increased. This is particularly noticeable (up to 0.4 increase in probability) in combinations of starting dates and durations that are influenced by rains in the period day 110150 (Oct. 19Nov 27) the very beginning of the rainy season. A similar, although not so large change is noted at day 245 260 (Mar. 3 – 18), the end of the rainy season. Only the earliest of starting dates for short durations, and dates 200225 (Jan 16 – Feb 12), at longer durations show a slight reduction in risk of such an outcome. Changes in the probabilities of experiencing between 15 and 20 days of rain, illus trated in figure 6b, are more complex. Zones that might previously have experienced this range in the number of rainy days experienced in the second period fewer rainy days, thus the probability of such occurrences declined (positive scores). Meanwhile, co mbinations of starting dates and durations (generally greater than 50 days) that had usually yielded larger numbers of daily rainfall occurrences before 1975, show a greater probability in period two (negative changes), because they have “slipped downwards ” into this class. Figure 6c shows that the probabilities of experiencing more than 25 rainy days have decreased throughout the graph. Changes are smallest in combinations of starting dates (early and late rainy season) and durations (less than 50 days), w hich had low probabilities of such events, even before 1975. The most marked declines in probabilities (>0.4) are found at longer durations (>60) associated with the starting dates that encompass the beginning or end of the rainy season . Figure 6a . Change In the Probability of Having Fewer Than 9 Rainy Days Figure 6b . Change in the probability of Having between 15 and 20 R ainy Days Figure 6c . Change i n t e Probability Of Having More Than 25 Rainy Days

PAGE 6

CARLY MUIR University of Florida | Journal of Undergraduate Research | Volume 1 6 , Issue 2 | Spring 2015 6 F igure s 7a, 7b, and 7c illustrate outcomes of similar steps taken, but for rainfall totals.To find points that were not normally distributed calculations were done using two times the standard deviation, and subtracting that value from the average. If the difference was greater than zero it could be confirmed that those points were normally distributed, vales less than zero implied points were not normally distributed. To resolve this problem, and improve the results Non Parametric tests could be performed on the data to account for the observations that do not qualify as normally distributed (Burt , Barber, Rigby). The graphs show that conditions are in fact becoming drier in Maun Botswana, particularly in the month of December where both rainfall total and count experienced major declines. Although a majority of the water year has undergone changes in rainfall characteristics. CONCLUSION Climate change will have damaging affects worldwide. Sub Saharan Africa has been identified as one of the regions that will be impacted most severel y, because of its current marginal semi arid to arid environment, its proximity to the anticipated expansion of the subtropical highs and the dependence upon rain fed agriculture. Arid environments already undergo stress from water scarcity and food short ages, putting them at high risk of experiencing harmful effects from increased variability in rainfall. Studies have already shown that southern Africa is experiencing a decline in mean annual precipitation, as well as an increase in variability, leading t o drier conditions. As rainfall characteristics become less consistent, the rain fed agriculture sector will need more investment and advancement. Making it important to assess the variation in rainfall regime due to climate change, and how this will impac t crop production. This study involved the production of plots to portray changes in rainfall before and after the climatic shift experienced in the mid 1970’s at Mann, Botswana. Rainfall Figure 7a . Change In The Probability Of Receiving Less Than 200 Mm Of Rainfall Figure 7b . Change In The Probability Of Receiving Between 300 And 400 Mm Of Rainfall Figure 7c . Change In The Probability Of Receiving More Than 500 Mm Of Rainfall

PAGE 7

ANALYSIS OF RAINFALL VARIABILITY IN RELATION TO CROP PRODUCTION IN MAUN, BOTSWANA University of Florida | Journal of Undergradua te Research | Volume 1 6 , Issue 2 | Spring 2015 7 totals and the count of the days experiencing rain were used for this analysis. The data ranged from likely to be encountered, both frequently and infrequently, in a growing season of prescribed length following planting on a particular day. We tested t he null hypothesis that there is no significant difference in the valu es of these parameters between the two periods. Daily rainfall within each water year were aggregated from starting dates with fiveday intervals of a durations of 30 to 90 days, also at five day intervals, yielding 533 possible combinations of starting da tes and durations annually for each variable, and necessitating over 2000 statistical tests. The means and standard deviation were calculated for each unique coordinate, allowing for a t test and F test to be carried out on the data on the parameters estim ated from data collected before and after the reported climatic shift. Producing graphs for both variables showing the difference in the means and variances between the two periods allowed for a rapid visualization of the most impacted times of the water y ear, and the magnitude of those changes. Overall, the majority of the water year has experienced a decrease in the mean of both rainfall total and the count of rainy days. Although the graphic portrayals of these changes are useful and instructive in descr ibing the nature of the changes in daily rainfall regime associated with the climatic shift of the late 1970s, the results have far more powerful implications and uses for, planners, decision makers, agriculturalists and managers. A combination of the assu mption of normally distributed variables and estimates of their respective means and variances during each period, permits taking these analyses one stage further. It was possible to use any specified value of interest for a variable and find the probabili ty or risk of obtaining this quantity in each time period. These probabilities can be expressed in terms of the risks of experiencing outcomes that do or do not exceed some critical level, or fall within some specified range of interest, the changes in whi ch can be investigated between periods. The method is sufficiently flexible that many objectives can be handled by the same approach. For instance, different crops may have different minimum and maximum requirements of both total and number of days for gro wth. So long as these requirements are known the approach can provide estimates of current probabilities and the changes that occurred at the time of the shift. These differences in probability or risk can be plotted in a similar fashion to show which peri ods and combinations of starting dates and growing seasons, had higher chances of receiving the particular value. Likewise, planners who may have acceptable levels of risk can translate that into equivalent rainfall totals or numbers of days with rain, and the values of these variables can be evaluated before and after the shift. This type of graph could help farmers make decisions about which types of crops to grow and when to plant, as they adapt to the new conditions. Some of the negative impacts climate change will induce are already known. In regions that already face environmental and developmental concerns, such as providing irrigation for crops, it is critical that the affects be studied and mitigated. REFERENCES Aquastat. (2015). Countr y profile Botswana. Retrieved December 12, 2012, from http://www.fao.org/nr/water/aquastat/countries_ regions/ BWA/index.stm Burt, J., Barber, G., & Rigby, D. (Eds.). (1990). Elementary statistics for geographers (third ed.). New York: Guilford Press. Cooper, P. J. M., Dimes, J., Rao, K. P. C., Shapiro, B., Shiferaw, B., & Twomlow, S. (2008). Coping better with current climatic variability in the rain -fed farming systems of sub -saharan africa: An essential first step in adapting to future climate change? Agriculture Ecosystems & Environment, 126(1-2), 24 -35. doi:10.1016/j.agee. 2008.01.007 Kalahari desert map (2009). Environmental Maps. Retrieved December 12, 2012, from http://enviro-map.com/karahari -desert map Muller, C. (2009). Climate change impact on sub -saharan africa? an overview and analysis of scenarios and models. (Discussion Paper). Tulpenfeld: German Development Institute. lund, G., & Dahlberg, E. (2008). Botswana Environmental and Climate Change Analysis. Retrieved December 12, 2012, from http ://www.sida.se/globalassets/global/countries -and -regions/africa/ botswana/environmental -policy -brief -botswana.pdf