|UFDC Home||myUFDC Home | Help|
This item has the following downloads:
1 INFLUENCE OF LANDSCAPE CHANGE ON THE NEARSHORE FISHERIES IN SOUTHERN CHILE By TRACY VAN HOLT A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DE GREE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA 2009
2 2009 Tracy Van Holt
3 To Sofia and Rodrigo
4 ACKNOWLEDGMENTS I thank Michael Binford and H. Russell Bernard for their advice throughout my dissert ation and for their willingness for me to take on an independent project that was at such a large scale, both theoretically and spatially. Dr. Binfords advice was instrumental in helping me secure my NASA Earth System Science Fellowship (ESSF/04R 00000080). His understanding of ecological processes amazes me and he has this ability to root out key problems with arguments that I lay out and helps make my work and writing better. Dr. Bernard exposed me to a whole new world and approach toward integrating my ecological background with quantitative social science methods. Dr. Bernard told me that I couldnt use (his help) up and Ive done a good job of almost doing so. He has read at times 20 versions of papers and has been instrumental in improving my writi ng and academic skills. Kenneth Portier helped transform my data analysis which was clunky at times in to a work that I am proud of. Tom Frazer introduced me to marine ecology and insisted that I took a course on invertebrate zoology, a course I am so grateful for because I could understand biologically what was happening in the nearshore system; he pushes you to think logically about marine ecology problems. Carlos Moreno welcomed me into his lab at Universidad Austral and made sure that I measured the parasites on the loco shellfish just right. Sandor Mulsow helped me get an army of student assistants who helped grind, cook, and analyze the lipid content in the shellfish; Sandors energy and great ideas motivated me throughout my time in Chile The Geos ciences Department at Universidad Austral probably will never allow another student to cook such stinky shellfish for almost a month straight ; I thank them for their patience Students at Universidad Austral helped me in the field especially Juan Paulo Tor res and Sandra Cuellar who assisted in analysis of loco health. I am gra teful to the fishers of FIPASUR
5 and the president Marco Ide. Also the fishers in Mehuin, Carrelmapu, the Federacion of Maullin, the Federacion of Estaquilla, those in Bahia San Pedro and Manquemapu, and Bahia Mansa (35 groups of fishers in total), who helped me harvest 1260 loco shellfish while they were working hard to harvest their quotas for the season. I especially thank Tito Gomez and Gaston Toro who welcomed me and spent hours ta lking with me about the fisheries in the region. Jose Alvarado from Purranque helped bring leaders of the region to me so I could present my research results when I was 7 months pregnant because I was terrified of driving on what I couldnt even call a roa d. Cayetano Augusto from the Servicio Nacional de Technica introduced me to many of the fishers and helped me enter their world. Of course, none would be possible without the support of Guillermo Quiroz, the fisheries director in the region. Not only did he spent hours talking with me, and did he find a walk in fr eezer where I could store the 1260 locos but he also met me sometimes at 12 in the morning to provide me the proper authorizat ion to transport the locos Magdalena Alid (Proyecto Asociativo de Fo mento Cerqueros de Valdivia) was an instrumental motivator to help jointly survey over 250 fishers in two months time. Leny Cares and other students from Universidad Austral tirelessly worked to help complete these surveys. Sandro Araneda who has grown up in Valdivia, and worked with the fishers for over 10 years, provided me an in depth understanding of the problems the fishers face and nuances of working with the fishing community Thanks are due to Antonio Lara, Eduardo Neira, and Patricio Romero for t heir advice and assistance with remote sensing and office space. In the US, I thank students in the Geography Department and the Land Use and Environmental Change Institute (LUECI) for the ir help. I especially thank Jaclyn Hall for her motivation, inspirat ion, and
6 support throughout my land cover change analysis. Jane Southworths lab manuals helped me re learn remote sensing and I am so grateful she made these manuals available online. The images and data used in this study were acquired using the GES DISC Interactive Online Visualization ANd aNalysis Infrastructure (Giovanni) as part of the NASA's Goddard Earth Sciences (GES) Data and Information Services Center (DISC). At the University of Florida, I thank the Department of Geography, Latin American Studi es, and Karen Bradley for helping me when my funding situation became unsure. I thank Pete Waylen for helping me secure funds and a teaching assignment. I thank Dave Hodell and Jason Curtis for assisting with the isotopic analysis. I also thank Rosana Resende for her advice as another mother and Ph.D. student at the University of Florida. Rodrigo Vergara, my main motivation for working in Chile, became my most valued field assistant. When I became pregnant, Rodrigo put his laboratory work on hold so he coul d stay with me. He was there for everything out on the fishing boats, driving, helping to logistically arrange the movement of the loco resources, collecting GPS points for me when I was too big to fit though the fences. He has helped me run countless anal yses on SAS into the wee hours of the morning and I cherish that I have a partner that I can discuss my work with. Sofia Vergara has been patient with her mommy working on her maps and Rodrigo has willingly taken up both mommy and daddy duties many times I also want to thank my mom, dad, and sister for supporting me emotionally and financially during my dissertation ; they knew when to stop asking, is it finished yet and they cherished my moments of success as I do. In addition to the NASA grant, this w ork was supported by the National Security Education Program, Rotary Ambassadorial Scholarships, Explorers Club, and the
7 Tropical Conservation and Development Program, the School of Natural Resources and Environment, and the Land Use and Environmental Cha nge Institute of the University of Florida. IRB approval (03031304) was obtained a t the University of Florida.
8 TABLE OF CONTENTS page ACKNOWLEDGMENTS ...............................................................................................................4 LI ST OF TABLES .........................................................................................................................11 LIST OF FIGURES .......................................................................................................................14 ABSTRACT ...................................................................................................................................17 CHAPTER 1 INTRO DUCTION ....................................................................................................................19 2 LAND COVER/LAND USE CHANGE AND FOREST PL ANTATION DEVELOPMENT PATTERNS IN A COASTAL TEMPERATE FOREST OF SOUTHERN CHILE AND ASSOCIATED VULNERABIL ITIES ......................................24 Introduction .............................................................................................................................24 Coastal Systems and Land Cover/ Land Use Change .....................................................24 Latin American Land Co ver Change Patterns .................................................................25 Globalization, subsidies, and local policies as landscape change drivers. ......................26 Forest Plantation Developmen t Effects ...........................................................................26 Objectives ...............................................................................................................................29 Methods and data sources .......................................................................................................29 Study Site .........................................................................................................................29 Image Preprocessing and Analysis ..................................................................................31 Land Cover Change Analysis ..........................................................................................32 Analysis of Landscape Change Patterns and Associated Migration ...............................35 Results and discussion ............................................................................................................36 Landscape Change in Southern Chile ..............................................................................36 Land cover classifications and accuracy assessments .............................................36 Major LCLUC patterns ............................................................................................36 Rural Migration Patterns ..........................................................................................39 Past and future development patterns .......................................................................40 Coastal management: predicting vulne rable sites and future management .....................43 3 CHLOROPHYLL A PATTERNS AND ISOTOPIC SIGNATURES OF LOCO ( CONCHOLEPAS CONCHOLEPAS ) shellfish in relation to forestplantation development in southern chi le ................................................................................................59 Introduction .............................................................................................................................59 Landscape Change and Nearshore Systems ....................................................................59 Chlorophyll a as an Indicator of Change .........................................................................62
9 Stable Isotopes: Tracers of Material Transport Pathways and FoodWeb Structure .......................................................................................................................63 Fis heries and Landscape Change .....................................................................................64 Methods and data sources .......................................................................................................65 Study Site .........................................................................................................................65 Satellite Image Analysis of Chlorophyll a and Sea Surface Temperature .....................66 Zone level analysis ...................................................................................................67 Watershed level analysi s ..........................................................................................69 Isotopic Analysis of Loco Shellfish .................................................................................70 R esults .....................................................................................................................................70 General C hlorophyll a Trends .........................................................................................70 Upwelling or Landscape Change? ...................................................................................71 Zone Scale .......................................................................................................................71 Watershed Scale Findings ...............................................................................................72 Isotopic Signatures of Loco Shellfish ..............................................................................73 D iscussion ...............................................................................................................................74 Conclusions and future research ..............................................................................76 4 MESO SCALE ANALYSIS OF THE Influences of Landscape Change on Loco ( Concholepas concholepas ) Health CONDITION Characteristics ........................................90 Introduction .............................................................................................................................90 Methods ..................................................................................................................................95 Study Site .........................................................................................................................95 Sampling and Data Processing ........................................................................................96 Statistical Analysis ..........................................................................................................97 Results .....................................................................................................................................98 LandscapeLevel Patterns of Loco Health Distribution ..................................................98 Loco Health Condition Groups ........................................................................................99 Relationships between Loco Health Related Characteristics and LCLUC Variables. ...................................................................................................................100 Discussion .............................................................................................................................101 Conclusions ...........................................................................................................................104 5 ADAPTIVE CAPACITY FOR FISHER SUCCESS ..............................................................118 Introduction ...........................................................................................................................118 Nearshore Systems and Landscape Change ..................................................................118 Fisheries Management in Chile .....................................................................................119 Study Site .......................................................................................................................123 Methods ................................................................................................................................125 Independent Variables ...................................................................................................125 Knowledge .............................................................................................................125 Environment ...........................................................................................................126 Personal variables ...................................................................................................128 Dependent Variables .....................................................................................................128 Analysis ................................................................................................................................129
10 Results ...................................................................................................................................130 Closed Fisheries .............................................................................................................130 Open Access C ongrio ...................................................................................................134 Adapting to Change .......................................................................................................135 Conclusions ...........................................................................................................................137 6 CONCLUSIONS .....................................................................................................................149 LIST OF REFERENCES .............................................................................................................152 BIOGRAPHICAL SKETCH .......................................................................................................170
11 LIST OF TABLES Tables: page 21 Accuracy assessment for the 2001 & 1985 image. ............................................................46 22 Area and percent of change and 2001 land use/cover of plantation, native forest, matorral agri culture, and clear land class. .............................................................47 23 Eigenvalues and eigenvectors of PCA1, PCA2, PCA3, and PCA4.. .................................48 24 Euclidean distance (m) from river or lake to plantation, native forest, agriculture, and cleared land classes. .................................................................................48 25 Area and percent of forest plantation (1985 and 2001) by soil use cla sses (CIREN and CORFO 1978). ..............................................................................................49 31 Monthly chlorophyll a (mg/m3) characteristics number of months with high and low chlorophyll a values and latitudinal and longitudinal distance covered by high and low chlorophyll a patches -in the Valdivia, Osorno, and Llanquihue zones from 1998 to 2005.. ..............................................................................79 32 The best regression models for 1, 2, and 3 independent variables. The best model was determined by R2, Adjusted R2, C(p), AIC, and BIC values. ........................80 41 Mean and standard deviation (in parenthesis) values of loco characteristics of each group in the hierarchical cluster an alysis for phoronids (PPHOR), shell boring bivalves (PBIVAV), polychaetes (PPOLYC), barnacles (PBARN), shell height (SHELLHT), loco meat weight (WGHMT), shell length (SHELLGT) and percent lipid in loco meat (PLIP). ........................................................106 42 Partial r square, significance levels, and average squared canonical correlations from the stepwise discriminant analysis of loco grouping using the loco characteristics, mean values for phoronids (PPHOR), shell boring bival ves (PBIVAV), polychaetes (PPOLYC), barnacles (PBARN), shell height (SHELLHT), loco meat weight (WGHMT), shell length (SHELLGT) and percent lipid in loco meat (PLIP). SHELLGT and PLIP were not statistically significant at p < 0.05. ..................................................................................107 43 Partial r square, significance levels, and average squared canonical correlation of stepwise discriminant analysis of LCLUC characteristics using percent change in plantations from 1985 to 2001 (PPL_CH) and ocean color characteristics (CHLA) in a canonical discriminant analysis of the hierarchical cluster analysis. ............................................................................................108 44 Mean and standard deviation (in parenthesis) values of LCLUC characterist ics, percent change in plantations from 1985 to 2001 (PPLCH) and ocean color characteristics (CHLA), for the corresponding upland
12 watershed for each group in the hierarchical cluster analysis of loco characteristics. ..................................................................................................................108 45 Number of observations, percent correctly classified (diagonal) and misclassified (off diagonal) estimates of hierarchical cluster analysis groups Llanquihue, Osorno, Valdivia, and Northern Valdivia using percent change in plantations from 1985 to 2001 (PLCH), percent native forest in 2001 (PFNF), and ocean color characteristics (CHLA) from canonical discriminant analysis. .............108 46 Probability of group membership of mis classified observations of the canonical discriminant analysis of hierarchical cluster analysis groups Llanquihue, Osorno, Valdivia, and Northern Valdivia using percent change in plantations from 1985 to 2001 (PLCH), percent native forest in 2001 (PFNF), and ocean color characteristics (CHLA). .........................................................................109 47 Cross validation summary and error count estimates of the canonical discriminant function of hierarchical cluster analysis groups Llanquihue, Osorno, Valdivia, and Northern Valdivia using percent change in plantations from 1985 to 2001 (PCPL) and ocean color characteristics (CHLA). .............................109 48 Standardized canonical coefficient s for the three canonical variables, LOCO1, LOCO2 and LOCO3 that describe the association between Loco Health Variables and loco groupings ..............................................................................110 49 Standardized canonical coefficients for the three canonical variables, LCLIC1, LCLUC2, and LCLUC3 that describe the importance of each landscape change variable on the canonical variable. .....................................................110 410 Pearsons Correlation among loco heal th condition variables where WGHMT = weight of loco meat (g), WGHV = weight of loco viscera (g), WGSH = shell weight (g), SHELGT = shell length (cm), SHELHGT = shell height (cm), SHELWDT = shell width (cm), FATIND = fat index (WGHMT/SHE LGT), PLIP = percent lipid (per g) in loco meat, PPOLYC = percent polychaete cover on shell, PPHOR = percent phoronid cover on shell, PBIVAV = percent bivalve cover on shell; PBARN = percent barnacle cover on shell; MALE = sex, male coded as 1. .........................................................................111 51 Dependent and independent variables tested in this study. ..............................................138 52 Descriptive statistics for variables used for the loco, congrio, and management area income analysis where fisher is the unit of analysis. ..........................139 53 Descriptive statistics for variables used for the loco, congrio, and management area income analysis where fishing syndicate is the unit of analysis. ............................................................................................................................140
13 54 Regression models of total locos caught in the 2003 season (TOTAL) and catch per unit effort (CPUE) for knowledge models. Independent variables used in the anal ysis shown in column 1. ........................................................................141 55 Regression models of catch per unit effort (CPUE) for environmental models. Independent variables used in the analysis shown in column 1. ........................141 56 Regression models of total locos caught in the 2003 season (total catch) and catch per unit effort (CPU).. ............................................................................................142 57 The best models for ea ch number of variables with R2, Adjusted R2, C(p), AIC, BIC, and the explanatory variables. ........................................................................143 58 Partial r square, significance levels, and average squared canonical correlation of stepwis e discriminant analysis of loco characteristics. .............................144 59 Number of observations and percent classified and error count estimates of successful and unsuccessful groups using stepwise discriminant analysis. .....................145
14 LIST OF FIGURES Figure s : page 21 Study site is the northern portion of Administration Region X of Chile.. ........................50 22 Digital elevation model of the study site based on Shuttle Radar Topography Mission (SRTM) Finished 3 arc second (90 m) raster elevation dataset. ..........................51 23 Supervised classification of October 5, 1985 (Landsat TM, path 233, rows 8789) and November 29, 2001 (Landsat ETM, path 233, rows 8789) images showing native forest, plantation, matorral agriculture, cleared land, sand, water, urban, snow, and wetland land cover/uses. .............................................................52 24 Principal components analysis biplot of the (A) 1985 and (B) 2001 land cover fraction variables. .............................................................................................................53 2.5 Plot of the change from 1985 to 2001 in the relative percent of A, native forest and agriculture/ranching B, plantation and native forest C, cleared land and agriculture/ranching and D, plantation and agriculture/ranching ...............................54 26 Conversion to the plantation class in 2001 from plantations, cleared land, native forest, and agriculture/agronomy in 1985. ..............................................................55 27 Conversion fr om the native forest class in 1985 to plantations, cleared land, native forest, and agriculture/agronomy in 2001.. .............................................................55 28 Vulnerable sites on land (regions with forest plantations and high los s of rural people in the county and in the nearshore region, in high plantation areas where rivers outlet to the ocean. ........................................................................................56 29 Agriculture and native forest lost to forest plantations from 1985 to 2001 and total plantations in 2001 (minus the area of agriculture and native forest converted) by percent loss of rural population per county as per the census data from 1994 to 2002. .....................................................................................................57 210 Forest plantations in 2001 by soil use maps. .....................................................................58 31 Study region in southern Chile showing zonal, watershed, and stable isotope analysis scales. ...................................................................................................................81 32 Supervised classification of October 5, 1985 (Landsat TM, path 233, rows 8789) and November 29, 2001 (Landsat ETM, path 233, rows 8789) images showing native forest, plantation, matorral agriculture, cleared land, sand, water, urban, snow, and wetland land cover/uses. .............................................................82 33 Latitudinal chlorophyll a (mg/m3) values (longitude averaged) derived from SeaWiFS satellite and seasurface temperatures from MODIS in the Valdivia
15 zone in 2003. Regions with high chlorophyll a values are not related to low (from 9C to 10C 9 C) and therefore upwelling probably does not play a role in chlorophyll a patterns in the area. ..................................................................................83 34 Monthly SeaWiFS chlorophyll a averages from 1998 to 2005 for the Valdivia, Osorno, and Llanquihue zones. ..........................................................................84 35 Chlorophyll a concentration (mg/m3) Hovmoller diagrams for the V aldivia, Osorno, and Llanquihue zones. ..........................................................................................85 36 Mean chlorophyll a values (mg/m3) as determined by the SeaWiFS satellite images from 19982005 in the winter (April to July), summer (November to February), and year round, in the Valdivia (39 to 40), Osorno (40 to 41), and Llanquihue (41 to 42) zones. ...................................................................................86 37 Scatter plot of mean chlorophyll a values at the watershed outlet (an av erage of one pixel for 8 years) from April to July (19982005) based on the SeaWiFS satellite images and area (km2) converted to plantation in each watershed (identified by numbers 113; see figure 3 1) from 1985 and 2001. ..................87 38 Scatter plot of mean SeaWiFS chlorophyll a values at the watershed outlet from April to July (1998 2005) based on SeaWiFS satellite images and area (km2) change native forest land cover in each watershed (identified by numbers 113) from 19852001. ........................................................................................88 39 1513C values for loco meat for native forest, mixed, and plantationinfluenced regions. See Figure 3 1 for location of sites. ...................................................89 41 Management sites where loco shellfish were sampled. ...................................................112 42 Average percent A) phoronid, B) polychaete, C) barnacle, and D) bivalve cover of on loco shells as well as E) lipid content and F) weight of loco meat in each management area and the percent forest plantation cover in 2001. (Averages obtained from a sample of 30 locos per management area). ..........................114 43 Loco shells showing A) phoronid, B) polychaete, C) bivalve and D) barnacles. A,B, and C are the ventral side of the loco shell. D is the dorsal side. ..................................................................................................................................115 44 Cluster analysis of loco shellfish in each management area based on health characteristics (percent phoronid, polychaete, and bivalve cover on loco shells, as well as weight and lipid content of meat, and shell length). Distinct levels of gray signify distinct clu sters. .............................................................................116 45 Scatterplot of LOCO1 and LAND1 from the canonical correlation analysis (R2 =0.881; p=0.000). ......................................................................................................117 51 Study Site in the Valdivia Province Coastal System, Chile. ...........................................146
16 52 Percent variance in price explained by environmental and knowledge factors for the loco and congrio fisheries. ...................................................................................147 53 The relationship between chlorophyll a and parasites on loco ( Concholepas concholepas ) shells in the management areas. If chlorophyll a increases above 2.4 mg/m3 then the % of parasites on the loco shells incre ases. ...........................148
17 Abstract of Dissertation Presented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy INFLUENCE OF LANDSCAPE CHANGE ON THE NEAR SHORE FISHERIES IN SOUTHERN CHILE By Tracy Van Holt August 2009 Chair: Michael W. Binford Cochair: H. Russell Bernard Major: Interdisciplinary Ecology Coastal systems provide livelihoods and food to over half of the worlds population. Land cover/land u se change (LCLUC) is one of the most significant global changes affecting aquatic systems. Human activities on the land deposit nutrients and sediments in the rivers that discharge into the nearshore environment and as a result coastal systems are becoming e utrophic and organisms are susceptible to parasites. This study examined the relationship between LCLUC especially forest plantation development, changes in nearshore chlorophyll a patterns, marine food webs, and the health of the loco ( Concholepas concholepas ) shellfish in southern Chile I also examined how fishers, who are restricted to harvest loco shellfish only in management areas, have adapted or are vulnerable to these environmental changes I used space for time substitution and worked across 13 watersheds (250 km of coastline) with over 30 fishing organizations and 41 management areas to provide a unique, geographical understanding of environmental change vulnerability, and adaptation to change Plantations increased 1,992 km2 from 1985 to 2001. The photosynthetic biomass from 1998 to 2005 across thirteen watersheds is higher 3.2 mg/m3 average chlorophyll a concentration values, in
18 nearshore regions that are influenced by forest plantations whereas native forest influenced areas have 2.0 mg/m3values. The loco shellfish harvested from management a reas influenced by forest plantations have approximately 30% more parasites on their shells when compared to locos from areas influenced by native forest land cover Fishers who harvested shellfish fr om the degraded areas lost $1 for each 10% increase of parasites on the loco shell. F isher knowledge and technology did not help fishers adapt to the environmental change. Instead fishers are moving to offshore fisheries such as congrio ( Genypterus sp.) w here their knowledge and technology can help fishers succeed and the environmental influences of landscape change are reduced
19 CHAPTER 1 INTRODUCTION This dissertation consists of four independent papers (Chapters 2 5) each of which is a paper written f or submission to a journal for publication. This dissertation is about land cover/land use change (LCLUC) in southern Chile the associated changes in chlorophyll a concentration levels in the nearshore marine ecosystem, the effect on the loco ( Concholepas concholepas ) shellfish health and how fishers adapt to these environmental changes. Th e second chapter will be submitted to The Geographical Journal and addresses the social and ecological vulnerabilities that stem from landscape change over the past 15 years in southern Chile and specifically asks : What are the patterns of landscape change over the past 15 years in southern Chile? What coastal areas are vulnerable to influences of landscape change over the past 15 years in southern Chile? Have people mi grated out of rural areas where forest plantations have been established in the past fifteen years? I identified whether plantation development resulted from afforestati on of grasslands or agriculture or substitution of native forest I then used principal components analysis to understand regional development patterns. Peasant agriculture has shifted to plantation forests and some remaining tracts of one of the worlds few temperate rain forests have been lost. Substitution of native forest by forest plantations accounted for 43% of the land converted to plantations. Afforestation accounted for 38% of the conversion-agriculture accounted for 27% and cleared land accounted for 11% of the conversion.
20 L andscape change stems from subsidies of plantation development Most native forest that has been converted is at the foothills of the mountains and therefore is likely not pristine forest. The nearshore environment in the Valdivian region likely receives nutrient inputs from forest plantations and these fishe rs are vulnerable to the effects of the environmental changes on the landscape. People living in the rural areas where forest plantations were established have migrated out. The third chapter will be submitted to Marine Ecology Progress Series and address es how the nearshore environment is influenced by landscape change and specifically addresses the following questions : Are nearshore chlorophyll a patterns related to landscape change patterns or are they related to upwelling? Are nearshore chlorophyll a concentration values at the watershed and zonal scale higher near plantation influenced watersheds? Are terrestrial derived nitrogen and carbon sources incorporated into the marine food web? I found high nearshore ocean chlorophyll a values (>10 mg/m3) near watersheds containing forest plantations that comprised up to 30% of their total land area. Low chlorophyll a values (<0.6 mg/m3) were found in the coastal areas near watersheds with high percentages (>80%) of native forest land cover. Loco tissue in nea rshore areas near plantation watersheds had enriched carbon and nitrogen ( 1 5N and 13C ) values, which indicate that terrestrial nitrogen entered the coastal waters and increased primary productivity which then increased atmospheric carbon fixation.
21 C hapt er 4 will be submitted to Ecological Applications I address how loco health was influenced by landscape change and associate d changes in the coastal system. I answered the following questions: Do locos located in management areas influenced by forest plan tations have more epibionts and shell boring organisms on their shells? Do locos located in management areas influenced by forest plantations have lower lipid content in their muscle tissue? Do locos located in management areas influenced by forest plantat ions weigh less than locos from other regions? I analyzed 1260 locos across 4 1 fisheries management areas. I used cluster analysis to group locos according to their height, meat weight, length, as well as the percent of phoronids, polychaetes, shell boring bivalves, and barnacles found on loco shells. I then used a discriminant analysis on the groups of the cluster analysis to identify the relationship between groups of locos and LCLUC. The discriminant analysis characterized 70% of the loco groups by percent forest plantation development from 19852001 and chlorophyll a values in the nearshore Canonical correlations found relationships between LCLUC and loco characteristics. The first canonical correlates LOCO1, representing loco parasites, and LAND1, representing a eutrophic system, showed a strong correlation 0.88, p=0.00. Loco shells from the northern Valdivia region a region with many forest plantations upland and high chlorophyll a values in the nearshore, we re infested with shell boring phoronids and barnacle epibionts and had low weight. Locos from the Osorno region, a predominantly pristine area, had shell boring bivalves and were heavier. Further examination of the social and ecological conditions in
22 sites that had locos with better or worse healt h than expected given the local conditions offered clues to how fishers can mitigate the influences of upland land use on their loco fisheries. In Chapter 5, to be submitted to Human Organization, I address how nearshore fishers on the coast of Chile adapt to environmental change. Specifically I answer the following questions: What factors (knowledge, technology, and environment) contribute to fisher success? Do fishers use the same strategies to adapt for success in the loco (closed access) and congrio (o penaccess) fishery? D ata include s reports by 279 fishers on the price they received for various resources, as well as on their knowledge of each resource, on the location of fisheries, on boat availability, and on their age. Data also include s satellite images (to measure plantation development and the productivity of marine phytoplankton) and GIS output (to characterize spatial relationships of management areas). Finally, 360 specimens of locos were surveyed for shell parasite load and other health chara cteristics. Environmental characteristics explained the most variance in price for the locos Specifically, shell parasites and chlorophyll a patterns accounted for th e most price variance for locos a restricted access shellfish Intensified technology (m ore boats) was detrimental to fisher success. Traditional ecological knowledge explained a smaller fraction of price variation. Fishers working in nearshore systems were heavily influenced by local environmental conditions: they benefit when the environmental quality is good and suffer losses from environmental change in their management areas. Fishers could use knowledge and
23 technology to receive higher prices for the congrio (openaccess) fishery where the environmental variables explain ed a lower fractio n of success. Fishers in eutrophic environments were more successful offshore, and those in noneutrophic environments were more successful in the closed access loco fishery.
24 CHAPTER 2 LAND COVER/LAND USE CHANGE AND FOREST PL ANTATION DEVELOPMENT PATTERNS IN A C OASTAL TEMPERATE FOREST OF SOUTHERN CHILE AND ASSOCIATED VULNE RABILITIES Introduction Coastal Systems and Land Cover/ Land Use Change The land sea interface -the coastal system is a rich environment for people and nature. Over half of the worlds population lives in coastal systems (Cohen et al. 1997) and the livelihoods of the worlds 13.1 million artisanal fishers depend on the coastal environment (FAO 2002). Landscape change and coastal development alter hydrological patterns, sediment and nutr ient composition, and biodiversity of rivers and the nearshore environment ( Milliman 1991; Diaz & Rosenberg. 1995; Smith et al. 1999; Bounoiua et al. 2002). People living in the coastal environment are affected by environmental change in the nearshore that alters photosynthetic biomass (Van Holt b in prep) the trophic dynamics of marine resources (Kemp et al. 2005), resource quality (Van Holt c in prep), and therefore fishing strategies and fisher success (Van Holt d in prep) Those people living in rural a reas on the land where change is occurring are al so affected by new development and some people relocate to urban areas (Kay 2006). By understanding development patterns and identifying areas where peoples livelihoods are likely to change, we can better u nderstand the critical pathways of change (Turner & Robbins 2008) and this can lead to a better understanding of adaptation and vulnerability to environmental change, which could help us develop a more sustainable future (Turner & Robbins 2008). This paper addresses the social and ecological vulnerabilities that stem from landscape change over the past 15 years in southern Chile. The specific research questions are:
25 What are the patterns of landscape change over the past 15 years in southern Chile? What coa stal areas are vulnerable to influences of landscape change over the past 15 years in southern Chile? Have people migrated out of rural areas where forest plantations have been established in the past fifteen years? Latin American Land Cover Change Pattern s Across Latin America globalization, local policies s ubsidies and population growth drive deforestation and afforestation (Grau & Aide 2008). Tropical forests are becoming deforested for cattle ranching and agriculture, especially across the Amazon (Hou ghton et al. 1991). The global soy bean market leads to deforestation of dry forests in Brazil, Bolivia, Paraguay, and Argentina by transforming forested areas into soybean fields (Dros 2004). Grasslands in Argentina and Uruguay are becoming afforested with plantations to support the global pulp industry (Farley et al. 2005). The foothills of mountains that were once used for agriculture are be ing abandoned as people move to the cities (Kay 2006) and these abandoned areas have regrown to forests ( Southworth & Tucker 2001). In southern Chile, where forest plantations account for almost 10% of the countrys exports (INFOR 2006); forest plantations for the global pulp market have replaced native forest and agricultural lands. Understanding the patterns of landscape change and its effects on the nearshore environment is particularly important in Chile, a coastal country where near shore fisheries provide a livelihood to over half a million people (Bernal et al. 1999 ).
26 Globalization, subsidies, and local policie s as landscape change drivers. Pine ( Pinus radiata) was introduced in the 1940s to protect degraded riparian areas (Lara & Veblen 1993; Clapp 1998; Toro & Gessel 1999) Plantation development exploded with the Decreto Ley 701 law of 1974 when the Pinoc het administration subsidized 75% of the cost of plantat ion establishment and eucalyptus ( Eucalyptus globulus and Eucalyptus nitens ) became an important pulp species (Lara & Veblen 1993). T o bring Chile into the world market economy and to make the country less vulnerable to the fluctuations in the copper market (Auty 1993, Gwynne 1996) t he government subsidized indemand, nontraditional agricultural exports such as pine trees In this process, pine trees are converted to wood chips and are exported for t he global pulp market. In some cases, pulp mills were built in Chile. The United Nations fostered commercial forestry by helping to establish the Instituto Forestal (Husch 1982), a research institute that provides techn ical support for plantation developme nt. By 1992, Chile was the worlds sixth most important wood pulp exporting country (Sedjo 1999). In 2005, plantations comprised 2.1 million hectares or 13.4% of the forest cover in Chile (INFOR 2006). Subsidies also led Brazil, Chile, China, India, and R ussia to become important forest plantation regions (Sedjo 1999; Barr & Cossalter 2004; Bull et al. 2006; Zhang & Song 2006) and these subsidies are a major driver of land cover/land use change (LCLUC) (Nagashima et al. 2001). Forest Plantation Developme nt Effects Identifying the location and extent of forest plantations, the geographical features of the landscape that have developed to plantations, and, whether plantations resulted from afforestation i.e., replacement of agricultural lands, ranchlands, grasslands, or shrublands (Blaschke et al. 1992; Barlow & Cocklin 2003; Farley & Kelly 2004) or
27 substitution, i.e., replacement of native forest (Lara & Veblen 1993; Ewers et al. 2006), helps to understand the effects of plantation development on the people and the environment and identify the vulnerable areas and people. In terms of ecological effects, researchers have identified that, at times, the effects of plantations on biodiversity, nutrient cycling, hydrology, and carbon sequestration, are distin ct in afforested and substituted lands, while other effects may depend on the area converted and not the historical land cover. Birds (Ca rnu s et al 2006; Vergara & Simonetti 2004) as wel l as mammals (Munoz et al. 1987; Saavendra & Simonetti 2005) have been negatively affected by substituted forests Biodiversity can increase or remain the same in afforested systems (Carnus et al. 2006). Lower diversity and abundance of macroinvertebrates were found in streams where upland plantations were present (Tierney et al. 1998). N utrients from plantations, for example nitrogen, are leached into streams (Binkley & Res h 1999) Nitrogen leaching and sedimentation is high at the onset of plantation establishment (Oyarzun & Pena 1995; Farley & K elly 2005; Oyarzun et al 2007) and nitrogen leaching has been identified after the plantation has been established as well ( Stevens et al. 1994). Parfitt et al. (2003) discovered that nitrate leaching was of a higher magnitude when native forests were first converted to agriculture and then to forest plantations. Plantations affect the hydrology of watersheds by increasing evapotranspiration and decreasing the amount of water that is supplied to s treams (Fahey & Jackson 1997; Gyenge 2003; He et al. 2003; Farley et al. 2005). Huber e t al. (2007) showed that afforested lands altered hydrology more than substituted lands, because less water was transpired by agricultural plants when compared with native forests. P lantations can decrease (Fearnside 1995; Byrne 2006; Glenday 2006; Nosetto 2006;
28 Betts et al. 2007) or increase (Binford et al. 2006) carbon sequestration, depending of the characteristics of the native forest that they are compared to Carbon sequestration can increase in afforested areas (Christie & Scholes 1995). With respect to the social effects of plantations, development patterns and geography can help to identify affected people. In other regions of Chile peasant farmers living on sloped mountains and using smallscale agriculture, domestic animals, and occasional season al work to survive eventually sold their land to plantation companies and move d to urban areas (Clapp 1995; 1998; 2001). These marginal agricultural areas are located on sloped mountainsides as in other parts of the world (Posner & MacPherson 1980) If ma rginal agricultural areas that may include forested landscapes are sold to plantation companies, localized rural to urban migration (Clapp 1998) might be observed. Fishers working in rivers and the nearshore areas can also have their livelihoods disrupted as plantation development has been shown to alter downstream aquatic systems (Iroume et al. 2006; Oyarzun et al. 2007) in Chile. In cases of extreme changes in the nearshore environment, key species may be lost and people face negative socioeconomic conseq uences ( Mee 1992; Kemp et al. 2005; Van Holt d in prep). Schlatter (1977) predicts where plantations should establish in the southern Chile based on the soil characteristics. Plantations are suited for type IV VI soils, which are soils that have a medium to moderate slope, cannot be plowed due to erosion risk, and have low draining; type VII soils, which are moderate to high sloped, can also be used for plantations, though there is a high risk of erosion and the best use of these areas is to preserve the n atural forest. Type I III soils, which drain easi ly, are best suited for agriculture. Wilson et al. (2005) predicted future plantation development patterns in
29 southern Chile They modeled future development patterns based on an analysis of a map that estimated forest cover prior to European settlement, which was derived from expert opinion and slope and elevation characteristics and a recent land cover map, which was based on aerial photography from 19951996. The study presented here is the first analysis in the region that measures the past with remotely sensed satellite images. The accuracy and quality of the satellite images are comparable in the past and present and offer a new perspective on regional development patterns. Objectives This study 1) cha racterizes LCLUC in the past 20 years across administrative region X of Chile in terms of area and percent of the watersheds converted 2) identifies whether forest plantations are a result of afforestation or substitution 3) examines whether rural migrat ion out of rural areas is associated with plantation establishment, and 4) identifies communities and areas vulnerable to recent land cover/landuse change (LCLUC). Methods and data sources Study Site The northern half of region X, an administrative region of Chile, contains three large watersheds with roughly one large watershed in each province (Figure 2 1). The names of watersheds on Fig. 2 1 are derived from the main river in the watershed: Valdivia (watershed # 3), Bueno (5), and Maulln (13), and ten small watersheds -Lingue (1), Bonifacio (2), Chaihun (4), Quilhue (6), Contaco (7), Hueyelhue (8), Cholguaco (9) San Luis (10), Llico (11) and Putratrn (12). High annual precipitation (2,0004,000 mm) and a mean annual temperature of 12.5C (Eldridge & Pacheco 1987) g i ve rise to a temperate rainforest ecosys tem, which prior to European settlement spanned across the
30 entire study site; today the remaining portion o f the forest are along the coastal and Andes mountains (Figure 22). Principal economic activities include forestry, fisheries, agriculture, and cattle ranching. Agriculture and ranching expanded in the valleys of the temperate rainforest ecosystem beginni ng in the late 1800s (Figure 2 2) Today, agriculture and ranching continue to be dominant land uses, although pine and eucalyptus plantations represent the most significant percentage of LCLUC in recent times (INFOR 2006) In 1997 farms covered approxima tely 290,000 ha in the valley with most of the land producing fodder crops (64%) and cereals (17%) such as wheat, oats, and barley (INE 1997). The region also is an important dairy producer accounting for almost 70% of the countrys dairy production. In 1997 (the latest figure), there were more than 1.5 million cows in the region most of which were raised on natural grasses and on small scale farms (INE 1997). Very few industrialized operations are present and fewer than one million people live in the regi on (INE 2002). Urban and commercial activities are not considered significant in terms of coastal systems, with the exception of the Celulosa Arauco ( CELCO ) pulp mill that began operating in 2003 and is located in the northern part of the study site that began operations in February 2004 and can process 680,000 tons/year (Oinonen 2005). The temperate rainforest offered available land and little competition for forest plantation development. Over the past fifteen years (1985 to 2001) plantations have encroached on the temperate rainforest, at times replacing both native forest and agricultural areas (Clapp 1998; Wilson et al. 2005). There are three main provinces in the region. The Valdivian province has extensive forest plantations and agriculture (Figure 21) The Osorno province has extensive native forest cover, agriculture activities, and a less extensive, but increasing, number of
31 plantations. The Llanquihue province has extensive agriculture and ranching, as well as matorral (new growth or shrubby forest) land cover and few plantations. Chile is a n export based economy where native forest and agricultural cover are relatively stabl e (no huge areas of deforestation or agricultural expansion is occurring) and therefore is a model for countries with similar landscape and economic characteristics Image Preprocessing and Analysis An October 5, 1985 Landsat 5 TM and a November 29, 2001 Landsat 7 ETM + scene (WRS II path 233, row 8789) were selected to correspond to the spring season when rains subside; these were the only available scenes with minimal clouds The 1985 scene represents the region 13 years after the 1972 forestry law that subsidized plantations ( Lara & Veblen 1993) and the 2001 image is from the time after Chile became an established, global level pulp producer (Sedjo 1999) A land cover map ( catastro) generated from aerial photographs in 1995 and 1996 (CONAF et al. 1997) was used to further understand specific land was used for training samples. I georectified the 2001 image by using training points from digital road maps from Chiles Ministerio de Obras Publicas which have accuracy that corresponds to both the Istituto Geographica Militar topographic maps and the catastro (CONAF et al. 1997) which both have a minimum mapped unit of 6.25 ha (and a 1:50,000 scale) and using the nearest neighbor sampling algo rithm. I then rectified the 1985 image to the 2001 image The root mean square errors for each image was less than 15 m. I then atmospherically corrected images using the C enter for the S t udy of I nstitutions, P opulations, and E nvironmental C hange (CIPEC) protocol that includes converting radiance values to surface reflectance values and dark object subtraction ( Markham & Barker 1986; Teillet & Fedos ejeus 1995). The low gain portion of band 6, the thermal band, was converted to temperature values. I
32 then topographically corrected all scenes using the Minnaert Correction, a modified cosine correction that helps determine the extent to which a pixel is Lambertian (Minnaert 1941). Teillet et al (1982) recommended modifying the cosine correction with a constant (k) that is derived in this case by regressing selected points in the native forest class of varying illuminations and slopes to determine to the extent the surface is Lambertian. I gener ated points for the regression of band k by selecting points in the native forest class of all illumination types. I then created additional layers by transforming the atmospherically and topographically corrected image with the tasseled cap transformatio n that converted bands to six new axes that represent brightness, greenness, wetness fourth, fifth, and sixth layers (Kauth & Thomas 1976). I also created a normalized dif ferential vegetation index layer (NDVI) (Rouse et al. 1974). Finally I ran a texture analysis on band 5 and created three new layers: a range, mean, and standard deviation texture layer. I used band 5 because it was most useful in separating forest plantations from native forest in trial tests. For the superv ised classification I created m ultiple layers that included all previously mentioned bands plus the Shuttle Radar Topography Mission (SRTM) Finished 3 a rc s econd (90 m) raster elevation dataset and the derived slope image. Land Cover Change Analysis I used the Maximum Likelihood super vised classification met hod to create nine classes (1) n ative forest, (2) plantation, (3) matorral (4) agriculture, (5) cleared land, (6) urban, (7) wetland, (8) water, and (9) snow classes (ITT 2008). I used the Chilean vegetation map and my knowledge of the spectral signatures when necessary to select training points (i.e. cases where an agricultural region was identified but wasnt characterized in terms of cleared land or active agriculture). I collected 109 accuracy
33 samples from July to August in 2004. I grew the locations of the accuracy samples to approximately 10 pixels using the spectral signatures to capture the overall signature in the training sample area and retained these samples for accuracy assessment of the 1985 and 2001 image. For the accu racy samples of each forest class, with the assistance of a Chilean forest engineer, I listed the main tree species present, percent ground cover (measured by taking a visual inspection of approximately a 10 x 10m plot and estimating the amount of ground c over) and canopy closure (measured by visually inspecting approximately a 30 x 30 m2 area and estimating the percentage of light that will go through the entire area) ; I also estimated the age of the forest plantations (measured by counting the rings of t he stumps, using a reference of growth of approximately 4 cm each year at breast height for pine and 2 cm each year for eucalyptus ( Toral et al. 2005; Munoz et al. 2005) and knowledge of land use in the region. I then conducted a second accuracy assessment by generating 30 random points in each land cover class to capture the accuracy of area that may not be easily accessible. In this case I used the national vegetation map (CONAF et al. 1997) and personal knowledge of the satellite image to assess accuracy (i.e. I could distinguish cleared land and agricultural areas on a 453 composite image; plantations were determined using the national vegetation map and visually inspecting a 4 5 3 composite of the satellite image). The forest plantations were visuall y obvious on the image, but difficult to separate by solely spectral means. It was essential to restrict the forest plantation analysis because of spectral confusion with disturbed native forest and because the study site was large (35,853 km2). To restric t the location of forest plantations I created a vector layer that included a 1km buffer around plantation polygons from the national vegetation map
34 (CONAF et al. 1997) which is based on aerial photos and visually obvious plantation regions that I identif ied T he one km buffer best accounted for expansion of plantations and eliminated the spectral confusion with native forest areas located in the middle of the extensive native forests of the Andes. I used the RuleGen extension in ENVI (ITT 2008) that was c reated by Loh and Shih (1997) and found that spatially reducing the area where plantations could be found, combined with the Maximum Likelihood classifier, provided better results as entire plantation areas were selected instead of just a few pixels within the plantation area I used the preliminary classification of 1985 and 2001 to further refine the classification. For example, I double checked regions that had cleared land in 1985 and native forest in 2001. Although plantation regions were accurately i dentified, some pixels within a plantation area were classified as native forest. I reclassified those plantation pixels (usually those pixels were inside areas already classified as plantations) and left the correct native forest pixels. I also sieved the urban class and made a minimum urban class consist of 10 pixels using an 8neighbor rule; this eliminated confusion between cleared land and urban areas. After the supervised classification, area (km2) and percent of each land cover/use class (km2 of LCL U/ total watershed area km2) in 1985 and 2001 were calculated in each of the thirteen watersheds (Figure 2 1). Then, the total area (km2) and percentage of each class that converted from the 1985 class to the 2001 class wa s calculated for each watershed I then calculated the Euclidean distance from the edge of lakes and rivers to the edge of each of the major land uses in 2001.
35 Analysis of Landscape Change Patterns and Associated Migration Watershed s were the unit of analysis to understand the ecological effects of landscape change because they are physical areas that drain the land into the rivers. I ran principal components analyse s on the percent of native forest, agriculture, plantations, and cleared land classes per watershed in 1985 and on the percent of each land cover/use class per watershed in 2001. Biplots (Gabrie l 1971 ; Gower & Hand 1996) of the first two principal components were used to identify watershed characteristics that were subsequently used to order watersheds and to understand regional patterns in land cove/use I characterized percent of land cover/use for each year because the area is an absolute quantity that can vary because of watershed size while proportion brings all the areas of the watersheds into the same value (100%). Bivaria te plots of percent cover with arrows indicating change from 1985 to 2001 were used to illustrate land cover /use composition change. To understand how plantations may be relat ed to rural to urban migration, I calculated the changes in rural populations pe r county I calculated the change in rural population from 1994 and 2002 per county using the Chilean Census Data (INE 1994; INE 2002) and compared these values to the landscape change data from 1985 to 2001 to evaluate whether rural people left regions where agriculture and native forest was converted to forest plantations and forest plantations were abundant. County level analysis is standard for understanding changes in population structure because they are important administrative boundaries in the regi on.
36 Results and discussion Landscape Change in Southern Chile Land cover classifications and accuracy assessments Figure 2 3 shows the landcover classifications for both 1985 and 2001. T he 2001 classification had an overall accuracy of 88.25% a kappa co efficient of 0.86, and no individual classes were less accurate than 75% (Table 21). The 1985 classification had an overall accuracy of 91%, a kappa coefficient of 0.90, and the accuracy of no individual classes was lower than 75%. I had high accuracy in detecting differences between native forest and forest plantations; the 2001 image had 87% producer s and 93% user s accuracy; native forest had an 85% producer s and 83% user s accuracy. The accuracy assessment with 30 randomly generated points of each l and cover class was similar The 2001 classificati on had an overall accuracy of 92 % with a kappa coefficient of 0.91 and the 1985 classification had an overall accuracy of 90 % with a 0.89 kappa coefficient; no individual classes were below 70% (Table 2 1) Major LCLUC patterns Plantations have becom e a more dominant part of the northern half of the Xth region of Chile. From 1985 to 2001 plantations and agriculture increased on average by 124 km2/yr and 53.4 km2/yr, respectively (see Figure 2 3, Table 22 ). Native forest and cleared land decreased by 81.1 km2/yr and 43.4 km2/yr, respectively More subtle changes in the landscape include increases matorral and urban land cover (3.7 km2/yr, and 1.9 km2/yr), and wetlands decreases 1.7 km2/yr A biplot of the PC A scores of the land cover fraction data show the 95% and 86% of the variability in the landcover fraction could be explained by two components in 1985 and 2001, respectively (Figure 24; Table 2 3). The main distinction between 2001
37 and 1985 PCAs is tha t plantations become an additional component on PCA2 in 2001. Native forest is the most influential land cover variable in PCA1, which explains the largest proportion of the data, 78.9% in 1985 and 70% in 2001. Watersheds 4,8,6,9, and 10 group together be cause these watersheds have high amounts of native forest; watersheds 1,3,5, 11, and 13 group together because they have low native forest. Watersheds 2, 7, and 12 have median amounts. Once watersheds are ordinated according to native forest cover, PCA2 e xplains and additional 16.5 % (1985) and 16 % (2001). Plantations covary with agriculture and cleared land in both years and they are opposite of native forest cover; so watersheds that are high on native forest are low on agriculture, plantations, and cle ared land. The watersheds group in similar patterns in 1985 and 2001 demonstrating that plantations have a small influence in the overall ordination of watersheds. Bivariate plots (Figure 2 5) of percent cover identify three distinct groups of watersheds. From 1985 to 2001, plantation watersheds ( 1, 3, 5, 7, and 11) gained a large to moderate increase in plantation fraction (3% to 35%) covering the watershed (watershed 5 is large so the fraction is relatively small). Also in 2001, these watersheds had low native forest fraction (70% or less) and high agriculture fraction (1035%) with moderate loss in fraction of native forest ( 313% loss, except watershed 5) from 1985 to 2001. Plantation watersheds have low percentage of native forest, high percentage of agriculture, and some fraction of native forest loss. Native forest watersheds (4, 6, 8, 9 and 10) started with a very high (>80 %) fraction of native forest which they maintained (>78%) in 2001. They had low (<5%) fractions of cleared land in 1985 and shi fted to low to moderate (6% 15%) fractions of
38 cleared land (watersheds 4 & 10 had moderate shifts) in 2001. Agriculture fractions ranged from 1% to 4% in 2001. W atersheds 4 & 10 had moderate shifts in native forest, however Watershed 4 lost 14% of its na tive forest (110 km2), and watershed 10 lost 13% native forest (97 km2). Of the native forest lost from 1985 to 2001 in watershed 4, 72% was converted to clear and 21 % to plantation. In watershed 10, 98% of the native forest converted transitioned to clea red land. In the m ixed watersheds (2, 12 and 13) the forest remained stable from 1985 to 2001. In 2001 the watersheds had moderate (60%) fractions of native forest; watershed 13 had only 24 % native forest but also had 6% mattoral which is similar to nativ e forest. From 1985 to 2001, t here was a low to moderate (2 11%) increase in the fraction of cleared land. Agriculture moderately increased and in 2001 720% of the watershed was covered in agriculture. L ittle to no plantations was present in either year. Substitution was the predominant method of transformation to plantations closely followed by afforestation (Figure 2 6; Figure 27) Substitution accounted for 43% or 984 km2 of the areas converted to forest plantations from 1985 to 2001. Most substituted forests were probably degraded native forest before conversion. These areas were degraded because these forests were located at the foothills of mountains and in fragmented areas or at the edges of large tracts of native forest where people graze cattle, plant small agricultural plots, and log the forest for firewood and other timber. Few areas in the coastal area or Andes mountains, where the forest is less disturbed, were converted (except for watershed 4). Agricultural/ranching areas accounted for 603.7 km2 or 26.6% of the area converted to plantations from 1985 to 2001 and cleared land explained 10.7% of the
39 conversion (462.9 km2) in total 37.3% of the plantations were afforested Watershed 1 lost the most agriculture (13.1% or 69.2 km2) mostly to plantations. Watersheds 3 and 5 lost the most cleared land 3.4% (345.7 km2) and 4.5% (688.6 km2), respectively, again with the majority converting to plantations. Watersheds 1, 3, 4, and 10 lost the most native forest, 17.1% (89.9 km2), 6.6% (677.3 km2), 14.4 % (109.6 km2) and 12.7% (96.6 km2), respectively (Figure 2 7) Most of the native forest lost in watersh ed 4 converted to cleared land and loss in watersheds 1 and 3 converted mainly to plantations. Watershed 5 had the most native forest that converted t o agriculture. Plantation regions were closer to rivers (43 m) than were native forest regions (256 m) (Table 2 4) and therefore nutrient inputs into the system from plantation systems have a more direct influence on the aquatic system. Rural Migration P atterns The livelihoods of farmers and fishers face vulnerabilities since regional development patterns have been altered. People living in counties with an increase in forest plantations have lost rural populations (Figure 28; Figure 2 9) In the norther nmost Mariquina County in 2001, plantations covered 300 km2. Since 1985 native forest and agricultural areas were converted to pine plantations. T he rural population also decreased by 20% from 1994 to 2002 (INE 1994; INE 2002). La Union and Los Lagos counties follow similar patterns, losing rural inhabitants as agriculture and native forest was lost in favor of plantations. O ther counties associated with plantations including, Mafil, Paillaco, Corral, and San Juan de la Costa also lost subsistence farmer s Not all plantation countie s lose rural people however, the Valdivia County which has a large amount of plantation activity, has had an increase in the rural population. Futrono and
40 Llanquihe have little plantation activity, but lost around 30% of the rural inhabitants so other factors can also influence rural migration. In total 11 plantation counties lost rural people; 8 nonplantation counties lost people; and 3 plantation counties remained the same or gained inhabitants (two of which are major metropolitan areas of the region Valdivia and Osorno). Past and future development patterns Most of the forest conversion was from substituted native forest that was likely degraded prior to conversion. Using the Chilean land cover map as a guide, substituti on of native forest was still an important component of plantation development, even after 19951996. For the total of the 983 km2 of native forest converted in all of Region X between 1985 and 2001, 308 km2 of native forest and 48 km2 of agriculture and c leared land was converted since 19951996 whereas and 562 km2 of native forest already was plantations in 19951996; other transitional phases account for 64 km2. Relatively undisturbed native forest was substituted mainly in watershed 4. Since most of th e substituted forest is located either in fragments (usually near agricultural lands) or at the edge of the native forest, it is likely that these forests were degraded before substitution occurred because forest edges have been shown to have higher mortal ity, damage, and turnover rates (Laurance et al. 1998). Native forests that were converted to plantations were at lower slopes and elevation (mean slope of 10 5 and 278 m 132 m s.d. elevation) whereas native forest that remained native forest were at higher slopes and elevations (13 9 slope and mean elevation of 562 m 353 s.d. m). Conversion of agricultural and cleared land areas in the past 15 years reflects t he development patterns in region IX ( Clapp 1995; 1998; 2001). I infer that in reg ion X, a s it
41 became more difficult to substitute forest, plantation companies capitalized on unsuccessful agricultural lands with limited yield and associated native forest as they did in the IXth region (Clapp 1998). Areas converted from agriculture to fo rest plantations had an average slope of 85 and elevation of 228 m 114 m whereas areas in sustained agricultural land covers were flatter and at lower elevations (average slope of 34 and elevation of 152m 98m). Based on the satellite image analys is, a griculture in the valleys remains stable most likely from the strong agricultural export market (Gwynne, 1996), reliable transportation (based on digital road maps from Chiles Ministerio de Obras Publicas and experience travelling in the region) and arable soil (Schlatter 1977). According to the soil use map (CIREN &CORFO 1983; Figure 2 10; Table 2 5) very little land useful for agriculture (Classes II and III), 11,170 km2 or 5% of the total plantation cover, was converted to plantations, in 2001. I n the 2001 Landsat image, large agricultural areas in watershed 5 on the eastern slope of the coastal mountains, had already been cleared and is a possible location for new plantations. These cleared soils range from type IV to type VII. Native forest tha t was degraded and agricultural lands on median slopes will probably continue to be locations for future plantation development. According to my analysis, the areas that experienced the most plantation development between 1985 and 2001 are located on the western portion of watersheds 3 and 5, which corresponds to the eastern mountain slopes of the coastal mountains and not along the western portion of the coastal forest (Figure 26). This contrasts with Wilson et als (2005) prediction of future development patterns. However, Wilsons prediction depends on road establishment and roads been developed only recently (2005) in the
42 coastal mountains and Wilson et als (2005) prediction may hold true in the future. Regions where coastal roads will be built could convert from native forest to plantations if undisturbed native forest is directly converted to plantations. More likely, however is that the native forest will become disturbed through coastal development and then plantations can become established. Bueno (5), San Luis (9), and Maullin (13) watersheds are possibly areas of future concern because I observed the government paving roads in the region in 2008 and if substitution of pristine forest occurs or the forests are degraded, then the Wilson et al (2005) patterns may hold true. Schlatters (1977) assessment of plantation establishment for pines agrees with the current trends. The least desirable soils (class VII) are mainly used for forest plantations, showing that available land that is best for pla ntations, type IV VI, is limited. Soil class VII accounts (Table 2 5) for most of the forest plantations in 2001 (63%) and 1986 (69%), respectively, followed by class IV, V, and VI, which accounted for (31%) of the plantations in 2001 and 24% of the planta tions in 1986 (Table 25). Figure 210 shows that the soil class IV, V, and VI are important in the Osorno region where this soil type corresponds to dairy farming. Plantations established on class VII soil mainly represents the Coastal and Andes Forest where plantations occur usually at the edges, probably because these areas have less steep slopes and are located closest to roads. The native forest cover in the coastal forest was v ery stable, at least from a regional perspective, over the past fifteen ye ar s. CONAF limits on substitution the geography of these watersheds, and the lack of connection to road networks probably reduced deforestation. Additionally large areas of the coastal native forest watersheds are located on the westward side of the co astal mountains, and the metamorphic soils are not ideal
43 for pine plantations because they are nutrient limited and have low potential depth for roots (Schlatter & Gerding 1995; Schlatter et al. 1995). Furthermore, the watersheds have little or no connect ion to road networks making plantations or agr iculture even less attractive. The native forests in the Andes mountains, which are located in the eastern portion of Valdivia (3) and Bueno (5), remain in highaltitude and sloped terrain that are either not feasible for plantation use due to the rugged terrain, or isolated and not connected to roads. Deforestation in Chaihun (4) and San Luis (10), represent two different development patterns. Prior to 1990 a company in the Chaihun watershed obtained appr oval to substitute Eucalyptus plantations for native forest. A portion of the land was substituted, another portion was clear cut, and a final portion remained native forest. According to local informants, t he company abandoned the project because of finan cial problems that developed in part from legal battles over substitution of native forest. Today a consortium of conservation organizations own the land-clear cutting ceased and the native forest will be restored. San Luis, located in the heart of the al erce ( Fitzroya cupressoides ) coastal forests (CONAF et al. 1997) was deforested most likely from illegal fires (Lara 1999) Dead alerce can be legally harvested, but local authorities cannot easily distinguish between a recently burned alerce and one th at died years ago. The most serious threat to the native forest watersheds in the alerce region is fire (Lara 1999) ; a fire could convert native forest watersheds to a mixed watershed and ultimately lead to conversion to plantations. Coastal management: predicting vulnerable sites and future management In general, region X has three development zones (Figure 2 1) The Valdivia zone (watersheds 1 4) is has changed from a system dominated by agriculture and to a lesser
44 extent native forest to a region where plantations dominate and people living in this area are most vulnerable to the influences of plantation development (Figure 28 ). The nearshore region may have increased nutrients, lower oxygen levels, and increased sedimentation affecting fisheries (Smith et al. 1999; Bounoua et al. 2002; Kemp et al. 2005, Van Holt c, d, in prep) (Figure 2 9). In many plantation counties the small scale farmers on the sloped mountains likely migrated to urban areas since forestry companies buy out small scale farmers and r eplace their land with forest plantations. The Osorno zone (59), which contains most of the small coastal native forest watersheds, remains relatively similar with extensive native forest and is the least affected. The Bueno watershed (watershed 5 in Fig ure 2 1) in the Osorno zone is an exception, with an extensive conversion to plantation and cleared land; however, the large size of the watershed and the surrounding native forest may buffer some of the effects. The Llanquihue zone (1013) is in a state o f transition with some native forest, matorral plantation, and agriculture, and potentially some ecological and social effects, but no clear indication as to what the future landscape pattern will be. According to Schlatter (1977), most of the region in w atershed 13, the area that is mainly mattoral cover, is not suitable for pine plantations because the adi soils have poor drainage and in the summer they become extremely dry; these thin soils are on flat lands and derived from volcanic ashes. In terms o f the coastal region, t he water resources along the rivers and coasts of areas near plantations need to be examined in more detail as future research hypotheses. The landscape analysis of this study can serve as a basis to design studies about nutrient dyn amics and conduct a more detailed analysis that includes history and extent of
45 conversion and possibly plantation age to further understand nutrient dynamics in plantation systems in southern Chile The relationship between loss of rural people and incre ase in plantations leads to future research hypotheses that address how rural livelihoods are shifting. I hypothesize that those who do remain in areas dominated by plantations adapt to a new rural life. Rural farmers who remain near plantations likely have fragmented social networks because their neighbors are gone -a rural farmer in the region reported that no one was left, just the plantations. Plantation companies buy out one family after another after another until whole communities disappear (Clapp 1998) Cattle ranching systems are also likely changing as forest plantations replace smallscale dairy farms located on marginal agricultural lands those plantation areas are located in IV, V, and VI soil types (pink areas in Figure 2 10). Some evidence of outmigration show that indeed changes are happening; Purranque, Rio Negro, Fresia and San Juan de la Costa, all ranching counties are losing rural people (between 9 and 22%) (INE 1994; INE 2002) and the loss of people can influence dairy production for Chile. Because Chile is a leader in plantation exports worldwide and has relatively stable native forest and agricultural areas, plantation development patterns can serve as a model for this type of development in other areas of the world that have simila r land constraints. Chiles relatively small size help make it a model country to identify effects of plantations on the ecosystem and peasant and fisher livelihoods.
46 Table 2 1. Accuracy assessment for the 2001 & 1985 image using the 109 accuracy assessm ent training points and 30 randomly selected points Training points 30 random points Class Name Producers Accuracy pixels User Accuracy Pixels Producer Accuracy % User Accuracy % Reference Totals Classified Totals Producer Accuracy % User Accuracy % 20 01 Image P lantations 315/362 315/337 87 94 30 31 93 90 Native Forest 242/286 242/292 8 5 8 3 30 27 80 89 Mattoral 207/266 207/217 7 8 9 5 30 35 97 83 Agriculture 159/159 159/188 100 8 5 30 27 87 96 Cleared land 191/210 191/226 9 1 8 5 30 23 70 91 W ater 50/50 50/50 100 100 30 32 100 94 Urban 272/292 272/272 93 100 30 32 100 94 Snow 46/46 46/46 100 100 30 30 100 100 Wetland 50/65 50/50 7 7 100 30 33 100 91 1985 Imag e P lantations 311/384 311/311 81 100 30 28 93 100 Native Forest 256/256 2 56/318 100 81 30 26 80 92 Mattoral 190/238 190/191 80 99 30 36 93 78 Agriculture 234/234 234/293 100 80 30 29 87 90 Cleared land 228/230 228/266 99 86 30 28 70 75 Water 242/246 242/250 98 97 30 30 100 100 Urban 76/97 76/76 78 100 30 30 97 97 Snow 253 /261 253/256 97 99 30 30 100 100 Wetland 181/223 181/184 81 98 30 33 90 82
47 Table 2 2. Area and percent of change and 2001 land use/cover of plantation, native forest, matorral agriculture, and clear land class. Area change was determined by calculati ng the difference in area between respective land cover class in the November 29, 2001 image (Landsat ETM, path 233, rows 8789) and the October 5, 1985 image (Landsat TM, path 233, rows 8789) for each class and each watershed. Area in any land cover type for 1986 can be calculated by subtracting the change from the value for 2001. Percent change was calculated by dividing the area change by the total watershed area. Plantation Native Forest Mattoral Agriculture Cleared Land Change 2001 Change 2001 Cha nge 2001 Change 2001 Change 2001 Area (km 2 ) 1 151.0 181.9 89.9 169.2 0.1 2.0 69.2 76.6 10.0 90.0 2 0.4 2.5 20.4 140.8 0.0 0.1 4.6 27.4 25.4 43.6 3 1264.0 1425.0 677.3 4483.5 7.5 73.4 61.0 2172.5 345.7 1365.5 4 27.1 27.1 109.6 591.0 0.5 0.5 2.0 21.0 78.3 101.3 5 417.2 489.5 56.8 5815.1 8.7 155.0 836.1 5403.4 688.6 1855.3 6 0.0 0.0 14.6 268.1 0.1 0.1 0.9 4.4 13.1 16.7 7 27.0 33.1 37.4 325.7 0.6 2.1 4.6 54.0 4.0 66.2 8 12.2 12.2 24.4 313.7 0.9 1.0 1.6 16.0 11.8 24.4 9 0.0 0.0 18.6 3 48.4 0.1 0.1 0.8 3.8 17.1 21.9 10 0.0 0.0 96.6 629.6 0.1 0.2 1.4 10.2 93.5 110.0 11 84.9 86.3 45.0 637.1 1.5 4.9 14.1 313.1 27.1 152.8 12 0.0 0.0 28.1 158.1 3.1 4.0 6.9 33.8 18.4 64.6 13 8.0 8.0 192.4 1141.2 68.5 308.7 30.6 990.9 94.9 634.9 Tota l 1991.8 2265.6 1297.5 15021.5 59.3 552.1 854.8 9127.1 694.9 4547.2 Percent 1 28.7 34.5 17.1 32.1 0.0 0.4 13.1 14.5 1.9 17.1 2 0.2 1.1 9.0 62.4 0.0 0.0 2.0 12.1 11.2 19.3 3 12.3 13.9 6.6 43.7 0.1 0.7 0.6 21.2 3.4 13.3 4 3.6 3.6 14.4 77.7 0. 1 0.1 0.3 2.8 10.3 13.3 5 2.7 3.2 0.4 37.9 0.1 1.0 5.4 35.2 4.5 12.1 6 0.0 0.0 5.0 91.2 0.0 0.0 0.3 1.5 4.4 5.7 7 5.6 6.8 7.7 67.1 0.1 0.4 1.0 11.1 0.8 13.6 8 3.3 3.3 6.6 84.6 0.2 0.3 0.4 4.3 3.2 6.6 9 0.0 0.0 4.9 91.9 0.0 0.0 0.2 1.0 4.5 5.8 1 0 0.0 0.0 12.7 83.0 0.0 0.0 0.2 1.3 12.3 14.5 11 6.1 6.2 3.2 45.7 0.1 0.3 1.0 22.5 1.9 11.0 12 0.0 0.0 10.6 59.6 1.2 1.5 2.6 12.7 6.9 24.3 13 0.2 8.0 4.0 23.8 1.4 6.4 0.6 20.7 2.0 13.3 Total 5.6 6.3 3.6 41.9 0.2 1.5 2.4 25.5 1.9 12.7
48 Table 23. Eigenvalues and eigenvectors of PCA1, PCA2, PCA3, and PCA4. Note: Standard deviations are the square root of the eigenvalues. Rotation contains the eigenvectors. PC1 PC2 PC3 PC4 2001 proportion data Plantation Area 0.19 0.92 0.23 0.25 Nati ve Area 0.91 0.04 0.17 0.38 Ag/Ranch Area 0.36 0.39 0.51 0.68 Clear Area 0.09 0.03 0.82 0.57 S.D. 25.51 8.82 5.27 3.75 Eigenvalues 650.76 77.79 27.77 14.06 1985 proportion data Plantation Area 0.02 0.21 0.18 0.96 Native Area 0.90 0.40 0.12 0.09 Ag/Ranch Area 0.38 0.59 0.67 0.26 Clear Area 0.20 0.67 0.71 0.02 S.D. 26.85 4.10 3.06 1.16 Eigenvalues 720.92 16.81 9.36 1.35 Table 2 4. Euclidean distance (m) from river or lake to plantation, native forest, agriculture, and cleared land classes. Mean Stdev Native forest 256 562 Plantation 43 248 Agriculture 139 408 Cleared land 77 336
49 Table 2 5. Area and percent of forest plantation (1985 and 2001) by soil use classes (CIREN and CORFO 1978). Soil classes I II are suitable for agriculture and have low slope and high drainage. Soil classes IV VI are best for plantations, the slope is higher and soils do not drain well. Soil classes VII are not ideal for plantations because the slope is even higher and drainage is poor. Soil class VIII are not suitable for plantations because they are located in rocks, snow, glaciers, dunes, or wetlands. Soil use class 2001 plantation area (km 2 ) 2001 plantation proportion 1985 plantation area (km 2 ) 1985 plantation prop ortion I 46 0 0 0 II 2,157 1 144 1 III 9,013 4 694 3 IV,V,VI 65,492 31 6,508 24 VII 132,126 63 18,438 69 VIII 1,757 1 830 3
50 Figure 2 1. Study site is the northern portion of Administration Region X of Chile. Thirteen watersheds for LCLUC analys is were delineated using the Direccin General de Aguas (the Chilean Ministry of Water) digital watershed map. In the map the Lingue watershed is represented by (1), Bonifacio (2), Valdivia (3), Chaihun (4), Bueno (5), Quilhue (6), Contaco (7), Hueyelhue (8), Cholguaco (9) San Luis (10), Llico (11), Putratrn (12) and Maulln (13). The names of the watersheds are derived from the main river in the area or the name of the region. The development patterns extend across three zones: Valdivia zone extends from 39S to 40S, the Osorno zone is from 40S to 41S and the Llanquihue zone is 41S to 42S.
51 Figure 2 2. Digital elevation model of the study site based on Shuttle Radar Topography Mission (SRTM) Finished 3 arc second (90 m) raster elevation dataset. T he remaining portion of the Coastal Temperate Forest and the Andes Forest is on high elevation areas (shown in white). Agricultural regions occur in the valley between the two mountain ranges.
52 Figure 2 3. Supervised classification of October 5, 1985 (Landsat TM, path 233, rows 8789) and November 29, 2001 (Landsat ETM, path 233, rows 8789) images showing native forest, plantation, matorral agriculture, cleared land, sand, water, urban, snow, and wetland land cover/uses. The numbers in white boxes corres pond to watersheds in Figure 21.
53 A. B. Figure 2 4. Principal components analysis biplot of the (A) 1985 and (B) 2001 land cover fraction variables. The numbers in correspond to watersheds in Figure 21 and Figure 22.
54 A. B C. D Figu re 2.5. Plot of the change from 1985 to 2001 in the relative percent of A, native forest and agriculture/ranching B, plantation and native forest C, cleared land and agriculture/ranching and D, plan tation and agriculture/ranching
55 Figure 2 6 Conversi on to the plantation class in 2001 from plantations, cleared land, native forest, and agriculture/agronomy in 1985. Based on supervised classification of October 5, 1985 (Landsat 5 TM) and November 29, 2001 (Landsat 7 ETM +), path 233, rows 8789. Figure 2 7 Conversion from the native forest class in 1985 to plantations, cleared land, native forest, and agriculture/agronomy in 2001. Based on supervised classification of October 5, 1985 (Landsat TM, path 233, rows 8789) and November 29, 2001 (Landsat ETM, path 233, rows 8789).
56 Figure 2 8. Vulnerable sites on land (regions with forest plantations and high loss of rural people in the county and in the nearshore region, in high plantation areas where rivers outlet to the ocean.
57 30 20 10 0 -10 -20 -30 -40 Percent loss of rural population per communa from 1994 to 2001 300 200 100 0 Agriculture and native forest lost (km2) to planatation from 1985 to 2001 & plantation cover (km2) in 2001 Valdivia San Pablo San Juan de la Costa Rio Negro Rio Bueno Puyehue Purranque Puerto Varas Puerto Octay Puerto Montt Panguipulli Paillaco Osorno Maullin Mariquina Mafil Los Muermos Los Lagos Llanquihue Lanco Lago Ranco La Union Futrono Frutillar Fresia Corral Calbuco Figure 2 9. Agricul ture and native forest lost to forest plantations from 1985 to 2001 and total plantations in 2001 (minus the area of agriculture and native forest converted) by percent loss of rural population per county as per the census data from 1994 to 2002 (INE 1994; INE 2002)
58 Figure 2 10. Forest plantations in 2001 by soil use maps (CIREN and CORFO 1983) see Table 2 6 for information on soil use classes Soil classes I II are suitable for agriculture and have low slope and high drainage. Soil classes IV VI are b est for plantations, the slope is higher and soils do not drain well. Soil classes VII are not ideal for plantations because the slope is even higher and drainage is poor. Soil classes that are not suitable for plantations include rocks, snow, glaciers, du nes, or wetlands.
59 CHAPTER 3 CHLOROPHYLL A PATTERNS AND ISOTOPIC SIGNATURES OF LOCO ( CONCHOLEPAS CONCHOLE PAS ) SHELLFISH IN RELAT ION TO FOREST PLANTATION DEVELOPME NT IN SOUTHERN CHILE Introduction Landscape Change and Nearshore Systems I examine here how nearshore coastal waters are influenced by landscape change at the meso scale. Human activities on land can alter coastal marine systems by affecting changes in hydrologic regimes, rates of nutrient loading, and sediment delivery of the rivers that dischar ge into the nearshore that, in turn, influence patterns of biological productivity and other key ecological processes (Smith et al. 1999; Bounoua et a l. 2002; Kemp et al. 2005). Terrestrial derived nutrients that are delivered to the nearshore environment can have profound influences on primary producers, phytoplankton in particular (Gomes et al. 2000; Shiomoto et al. 2002). Increased nutrient inputs to the nearshore environment can lead to changes in the material transport pathways and energy flow with consequences for foodweb structure and function (see e.g., Kemp et al. 2005). Extensive changes in the nearshore environment have been attributed to landscape change in the Black Sea, Chesapeake Bay, and Gulf of Mexico (Anderson et al. 2002; Kemp et al. 20 05; Diaz & Rosenberg 2008). Nearshore changes can have far reaching effects influencing organisms that live in the nearshore as well as offshore organisms that spend a part of their life cycle in the nearshore environment (Beck et al. 2001). In cases of e xtreme changes in the nearshore environment, key species may be lost and people are likely to face negative socioeconomic consequences ( Mee 1992; Kemp et al. 2005). This study addresses the following questions:
60 Are nearshore chlorophyll a patterns related to landscape change patterns or are they related to upwelling? Are nearshore chlorophyll a concentration values at the watershed and zonal scale higher near plantation influenced watersheds? Are terrestrial derived nitrogen and carbon sources incorporated into the marine food web? The social and environmental influence of landscape change on terrestrial systems has been the focus of much research (Foley et al. 2005). Less research has been carried out on the relationship between landscape change and the as sociated changes in coastal marine environment s because of complexities at the land sea interface (Talley et al. 2003). Studies that link landscape change to changes in nearshore marine systems tend to be small in scale (tens of kilometers along the coast ) and are generally focused on single estuaries with established connections to wellidentified watersheds (Camacho Ibar et al. 2002). Less information is available on how landscape change influences the nearshore environment at larger spatial scales (but see Bonsdorff et al. 1997; Alexander et al. 2000; Anderson 2002; and Matthews et al. 2004) because satellite images were not available until recently and ocean dynamics are complex i.e. geomorphology can influence species diversity (Fernandez et al. 2000) and large scale studies need to account for significant geomorphological, abiotic, and biotic differences (Talley et al. 2003). Today, however, satellite images and ancillary data on regional abiotic and biotic factors are available to help de sign studie s across mesoscales i.e., 100s of kilometers of coast. Meso scale studies across multiple watersheds allow for a greater understanding of processes that drive regional patterns. For example, a spacefortime substitution method
61 that compares watersheds with different landscape change patterns can help illustrate the environmental condition that may have existed in a watershed prior to the change (Fukami and Wardle 2005). Also, by comparing multiple watersheds, more evidence may be produced to determine whether environmental differences can be attributed to site specific upland activities, rather than to broad scale changes that might occur, for examples, as a consequence of climate change. Once these regional patterns are described, mechanistic studies c an be designed at the appropriate scale and location without bias due to accessibility Across Latin America globalization, population growth, and local policies and subsidies drive landscape change (Grau & Aide 2008). In Chile forest plantation development exploded with the Decreto Ley 701 law of 1974 when the Pinochet administration subsidized 75% of the cost of plantation establishment to support the global pulp industry (Lara & Veblen 1993); today pine ( Pinus radiata) and eucalyptus ( Eucalyptus globulus and Eucalyptus nitens ) are the main forest plantation species. By 1992, Chile was the worlds sixth most important woodpulp exporting country (Sedjo 1999). In 2005, forest plantations comprised 2.1 million hectares or 13.4% of the Chilean forest co ver (INFOR 2006). Plantations, especially Eucalyptus are fertilized with nitrogen and phosphorus to increase productivity (Guerra et al. 2007). Shlatters (1977) study shows that pine plantations do not hold nitrogen well in the soil and therefore also need to be fertilized. Recent studies of waters within the Valdivia watershed show that nitrogen concentrations are elevated relative to historical values and, in some cases, exceed legal limits (UACH 2005). Historical data indicate that nitrogen (N NO3) l evels in the rivers were low,
62 ranging from a minimum average of 0.065 mg/l to a maximum of 0.271 mg/l, from 1987 to 1992 (DGA 1992). Oyarzun et al.s (2007) study of small rivers in Chile showed an increase in nitrogen levels in plantation watersheds when compared with native forest watersheds. Little et al.s (2008) study of nitrogen levels in small rivers that had different land use strategies identified the highest nitrogen level in a watershed covered with forest plantations. P lantation development in Chile over the past 15 years has probably increased the amount of nitrogen being delivered to the nearshore environment via rivers Chlorophyll a as an Indicator of Change An increase of chlorophyll a in the nearshore environment can be an important indic ator of change in biological and ecological processes. Chlorophyll a is the primary light harvesting pigment found in all photosynthetic organisms including phytoplankton, which are generally the main contributors to primary production in the nearshore and oceanic systems. Chlorophyll a concentration is one of the best predictors of photosynthetic biomass available today (Huot et al. 2007). High chlorophyll a values indicate an increase in phytoplankton and could indicate that fertilization with nitrogen, phosphorus, or other nutrients has occurred, and initiated an increase in algal biomass in a n otherwise nutrient limited system. High chlorophylla values can also occur as a consequence of upwelling events that bring cool, nutrient rich water from deep waters to the ocean surface and stimulate photo plankton production (Ryther 1969). If high chlorophyll a values in coastal waters are accompanied by cold sea surface temperatures then upwelling is generally assumed to have contributed to the high chloroph yll a values (Ryther 1969; Moreno 1979). Spectral information obtained from satellites is often employed to quantify chlorophyll a levels in nearshore and oceanic waters and timeseries data, when available, can be used to identify changes in production characteristics.
63 If other abiotic and biotic factors are similar, space can substitute for time to understand how future landscape change may influence chlorophyll a values. Stable Isotopes: Tracers of Material Transport Pathways and Food Web Structure Loc o ( Concholepas concholepas ) is a murcid mollusk carnivore that preys predominately on filter feeding tunicates and barnacles (Stotz et al. 2003) and therefore is a good species for stableisotope analysis to understand material transport pathways and foodweb structures. The isotopic signatures of locos can indicate changes in the food web structure and how bottom up influences (nutri ent inputs ) can alter the pathways of flow within food webs Stable is otope analysis is based on the ratio of light and hea vy isotopes of an element; these ratios are compared to a standard and the ratios are either enriched, depleted, or equal to the standard (Fry 1988). Stable isotope ratios can help trace th e origi n of carbon and nitrogen because terrestrial, marine, and at mospheric nitrogen and carbon have distinct isotopic ratios (signatures) (Coleman & Fry 1991). Loco flesh should be lighter or heavier depending on the source of C and N and (McKinney et al. 2001; Freyer & Aly 1974; Kreitler et. al. 1978; Cole et al. 2004; Peterson & Fry 1987; Trumbore & Druffel ; Coleman & Fry 1991) If terrestrial nitrogen and carbon entered the system, the 1 5N values should be enrich ed or depleted slightly (Freyer & Aly 1974; Kreitler et. al. 1978; McKinney et al. 2001) and the and 13C values should be depleted (Trumbore & 13C values generally range from 32 to 20 (Coleman & Fry 1991) Known enrichments can also occur through the food web, 13C values are typically enriched 1 per trophic level and 14N values are typically enriched 3 4 per trophic level (Fry and Sherr 1984; Minagawa and Wada 1984). Shifts in the foodweb structure, such an increase or decrease in food we b length
64 (Cabana & Rasmussen 1994), or domination of some organisms (Bade et al. 2005) can also be identified with isotopes. Fisheries and Landscape Change The Chilean Fishing and Agriculture General Act of 1991, a Territorial User Rights Fishery (TURF) program, granted groups of artisanal fishers managed exploitation areas for benthi c resources (MEABR) in the nearshore. Fishers gain quasi property rights and exclusive rights to benthic resources within these areas (Bernal et al. 1999). The Act has been hailed as an ecological success because loco, among other nearshore marine species, have recovered from the effects of overharvest (Defeo and Castilla 2005; Gonzalez et al. 2006). Changes in the nearshore environment; however, can have consequences for those who depend on livingmarine resources for their livelihoods if changes in the la ndscape negatively influence the quality of the resources in the management areas and fishers cannot shift to other regions to harvest resources (Van Holt in prep d). Moreover, employees of Servicio Nacional de Pesca (SERNAPESCA), the Chilean fishery manag ement institution, report that in the near future fishers will be required to pay a fee to retain the MEABR s. As a result, fishers are becoming increasingly concerned that upland activities will affect their livelihoods. Indeed fishers living close to f orest plantations have locos with shells infested with parasites (Van Holt in prep c), and these fishers are paid lower prices per kg for the loco shellfish and earn a lower monthly income from the MAERBs. Some fishers working in these poor quality regions have given up on fishing altogether (Van Holt in prep d). Since Chile has a very long coast in relation to its total land area, the nearshore environment is important to its economy. As of 1992, fisheries provided 516,000 direct and indirect jobs (3.8% o f Chiles population in that year). Of those jobs, 86% were in
65 the artisanal (nearshore) fishery (Bernal et al. 1999). With 41 million fishers worldwide (FAO 2002), and 40% of the worlds population living along the coast (Cohen et al. 1997), the relationship between fisheries and landscape change need s to be more adequately addressed. Our specific objectives were to determine: (1) the spatial and temporal variability of chlorophyll a in nearshore coastal waters from 19982005; (2) if landscape change fa ctors or upwelling, or both, influence high chlorophyll a values ; (3) are uplandlandscape characteristics related to the nearshore environment with high chlorophyll a values ; and (4) if terrestrial nutrient inputs are discernable in the marine food web. M ethods and data sources Study Site In region X of Chile (Figure 31), along the central coast (Fernandez et al. 2000; Camus 2001), uniform wind patterns, topography, and temperature give rise to similar invertebrate an d fish communities in the rocky inter tidal ecosystem (Moreno et al. 1979; Fernandez et al. 2000; Lancellotti & Vasquez 2000; Camus 2001) Upwelling events are not considered major ecosystem drivers as in other regions of Chile (Hebbeln 2000); however, small, coastal upwelling events reportedly occur during summer months (January to March) in the northernValdivia region (Moreno et al. 1998) According to Hebbeln (2000), the surface sediment distribution and related photosynthetic biomass is related to either terrigenous inputs or ocean currents The temperate climate and high rainfall (approximate ly 20004000 m annually) ( Eldridge & Pacheco 1987 ) enhances erosion, surface water runoff, and sedimentation, especially in the winter (Fernandez et al. 2000). Additionally, beach sands interspersed throughout the rocky intertidal areas have mainly terrest rial rather than oceanic origins (Hebbeln 2000). Coastal
66 g eomorphology is also similar throughout the region--the continental shelf extends to approximately 6.54 km and to 200 m in depth; there are no fjords in the study area. (Fernandez 2000). The Humboldt Current is the main oceanographic feature that influences the region (Lancelloti & Vasquez 2000; Fernandez et al. 2000; Thiel et al. 2007) and typically flows north. In comparison to the extreme north and south of Chile, the diversity of fish and benthic organism s is relatively similar in the north and south (Lancellotti & Vasquez 2000) ; the extreme north has more diverse fishes and less diverse benthic organisms and the extreme south holds the opposite trend. H owever, diversity of benthic organisms does exhibit slight variation. In a rapid survey of benthic organisms, the Valdivian zone (Figure 3 1) had a low Shannon Index score for benthic organism s ( 0.2); the Osorno zone had the highest diversity score (2); and the Llanquihue zone had a median score (0 .7) (Moreno 2001) Important benthic fisheries include the Chilean abalone, loco ( Co ncholepas concholepas) mussels ( Mytilus sp.), tunicates ( Pyura sp.), limpets ( Fisserella sp.), and sea urchins ( Loxechinus albus ). Important exploited finfishes include congrio ( Genypterus sp.) corvina ( Cilus gilberti), and sierra ( Thyrsites atun). The main zone s in the study site include the Valdivian zone that has extensive forest plantations and agriculture the Osorno zone that has extensive native forest cover, agriculture activities, and a less extensive, but an i nc reasing number of plantations, a nd the Ll anquihue zone that has extensive agriculture and ranching, as well as matorral (new growth or shrubby forest) land cover (Figure 3 1) Satellite Image Analysis of Chlorophyll a and Sea Surface Temperature Spatial and temporal variability in chlor ophyll a concentration pattern was characterized using 9 km resolution Sea viewing Wide Field of view Sensors (SeaWiFS )
67 chlorophyll a product s Chlorophyll a (CHLA) concentration values that are obtained from the SeaWiFS satellite image are derived from an empirically based bio optical algorithm (OReilly et al., 2000) using reflectance values, which are derived from calibrated digital counts that have been atmospherically corrected (Hu et al. 2009). The SeaWiFS images provid e m onthly chlorophyll a concentr ation (mg/m3) averages from 1998 to 2005; monthly composites include every valid measurement of water leaving radiance that is spatially and temporally combined within a 9 by 9 km2 element (referred to as bins) (Thomas & Franz 2005). S ea surface temperatur e (SST) (4 micron night scales) were obtained at a 9 by 9 km2 scale using Moderate Resolution Imaging Spectroradiometer (MODIS ) images for each month from July 2002 to December 2005 to test whether upwelling influenced chlorophyll a patterns. Zone level a nalysis Three zones, i.e., Valdivia (39S to 40S and 73W to 74W), Osorno, (40S to 41S and 73.5W to 74.5W), and Llanquihue (41S to 42S and 73.5W to 74.5W) (Figure 3 1) were compared to assess the influence of landscape change on the nearshore sys tem. I described regional chlorophyll a patterns tested whether and identified whether upwelling or landscape change influenced chlorophyll a values. First monthly images from 1998 to 2005 were analyzed for each zone (Valdivia, Osorno, Llanquihue) using the GES DISC Interactive Online Visualization ANd aNalysis Infrastructure (Giovanni) from NASA's Goddard Earth Sciences (GES) Data and Information Services Center (DISC). In the zone level analysis, I generated Hovmoller diagrams to explain the latitudinal a nd longitudinal variability of chlorophyll a values. In the latitude Hovmoller diagrams, I averaged the longitudinal chlorophyll a values across one degree. In the longitude Hovmoller diagrams, I averaged the latitudinal chlorophyll a values across one deg ree.
68 For example, to generate the value for December of 1998 at 35S latitude in the latitude Hovmoller diagram, I averaged the chlorophyll a values (9 x 9 km2 resolution) across all 35S pixels from 74W to 73W longitude to generate an average one degree value for 35S. To determine if upwelling explained high chlorophyll a patterns, I compared latitude Hovmoller diagrams of chlorophyll a values with latitude Hovmoller diagrams of MODIS generated sea surface temperature (SST) values for each month from J uly 2002 to December 2005. Regional SST values generally range between 12C to 14C, whereas cold, upwelling water ranges from 9C to 10C (Moreno et al. 1998). If low temperature values were found in the same areas with high chlorophyll a values, then upwelling is a likely explanation for the chlorophyll a patterns. Alternatively, low temperatures are not found with high chlorophyll a values then landscape change probably contributes to the high chlorophyll a values. I ran a test of independence between the presence or absence of unusually low SST and 24 high chlorophyll a patches (>10 mg/m3) to determine if high chlorophyll a values and low SST temperatures (9C to 10C) were dependent on each other. The effects of upwelling related to satellite images is immediate (i.e. cold temperatures immediately correspond to increase in chlorophylla values) so I did not quantitatively examine for lag time. I did, however, visually inspect temperature values two months before and after the increase in chlorophyll a and no cold temperatures were observed. I also calculated the chlorophyll a value for each zone b y averaging the chlorophyll a values in 30 pixels (9 by 9 km2 each) along the shore and then extending westward across one degree latitude for each zone. I compared mean values for all
69 months (January to December), summer months (November to February) and winter months (April to July) from 1998 to 2005. I used an ANOVA followed by a Tukey Kramer Multiple Comparison test to determine if mean chlorophyll a valu es for each zone (Valdiv ia, Osorno, and Llanquihue) w ere statistically different and to identify which specific zones were distinc t. I also conducted an ANOVA with a randomizedblock design using year as a block to account for the yearly variation in chlor ophyll a values Watershed level analysis For the watershed level analysis, I compared landscape change data in 13 watersheds with chlorophyll a concentration values at the 9 km pixel closest to the point where the river discharges into the coastal ocean (Figure 3 1). I averaged the SeaWiFS chlorophyll a data from April, May, June, and July (1998 2005) because these months did not have any reports of upwelling events (Moreno et al. 1998). Landscapechange data was obtained from an October 5, 1985 Landsat TM and a November 29, 2001 Landsat ETM sce ne (WRS II path 233, row 8789) I used the Maximum Likelihood supervised classification method to create nine classes (1) n ative forest, (2) plantation, (3) matorral (4) agriculture, (5) cleared land, (6) urban, (7) wetland, (8) water, and (9) snow classes (ITT 2008) (Figure 3 2; See Chapter 2 for details) For the analysis with the chlorophyll a data, I used the 1985 and 2001 landcover values as well as area and percent change in plantation, native forest, agri culture, cleared land, urban, and mattoral ( i.e. shrubby bushes and small trees). I then correlated the chlorophyll a av erage and landscape change data with a Bonferroni correction to account for statistically significant correlations by chance. Next I ra n a regression analysis to see if landscape variables explained the variance in average chlorophyll a values from 1998 to 2005. I selected three independent variables (percent
70 plantation change, area plantation change, and percent urban change) based on a scree plot of mean square error and a test of multicolinearity. Watershed 2 was excluded from the analysis because it was an outlier; the high chlorophyll a values were influenced by the Valdivia watershed (watershed 3). Watershed 2 rivers are small streams that have little input into the nearshore so watershed 3 values were related to chlorophyll a values near watershed 2; also watershed 2 doesnt have extensive landscape change. Isotopic Analysis of Loco Shellfish To investigate whether landscape activi ties are correlated with isotopic signatures of the carnivorous loco shellfish, I sampled six locos at four fisheries management areas that were influenced by native forest, plantation, and mixed landcover classes from July to August during the 2003 locoharvest season in Chile (Figure 3 1). Locos were dried, homogenized, and tissue (excluding viscera) was ground at the Mulsow Laboratory at Universidad Austral de Chile. At the University of Florida, the dried samples were loaded into a Micromass VG602 mas s spectrometer for analysis of stable carbon and nitrogen isotope ratios The results were then corrected and expressed in standard delta notation (PDB was the reference standard for carbon and nitrogen in air was the standard for nitrogen) (Fry 1988). I then used a scatterplot to examine the isotopic signatures. RESULTS General Chlorophyll a Trends Low chlorophyll a values were found from approximately April through August across all sites and chlorophyll a values decreased with distance offshore. High chlorophyll a values occurred from November to March and were located close to shore. Seasonal t emperature shifts played a role in seasonal chlorophyll a values. Lower SSTs (9.6C to 12C) and lower chlorophyll a values ( generally <2.5 mg/m3) were measured
71 from July to September; higher temperature values (12C to 16C) and higher chlorophyll a values ( generally >2.5 mg/m3) were recorded from January to April. Upwelling or Landscape Change? None of the >10 mg/m3 patches of chlorophyll a from the latitud e Hovmoller diagrams showed evidence of upwelling temperatures. For example, from February to March in 2003, the latitude Hovmoller diagram (Figure 3 3) shows a large >10 mg/m3 patch of chlorophyll a that spans from 39.40S to 39.75S. The corresponding l atitude Hovmoller diagram for SST showed no upwelling range temperatures (9C to 10C) at the same location during the same time. The test of independence shows that high chlorophyll a patches (>10 mg/m3) were independent and not related to low temperatur e (9C to 10C) events (p= 0.537) providing no evidence to support that upwelling temperatures are related to chlorophyll a and suggesting that landscape change could play a role in high chlorophyll a values. Zone Scale Consistent differences in chlorophyl l a were observed among the three zones (Table 3 1, Figure 3 4, Figure 35, and Figure 36). The chlorophyll a values derived from the SeaWiFS satellite images (area averaged) showed distinct patterns across zones from 1998 to 2005. Most notably the Valdivian zone had relatively high chlorophyll a values that ranged from 1 to 11.5 mg/m3, in comparison, the Osorno zone ranged from 0.7 to 6.6 mg/m3, and the Llanquihue zone, ranged from 0.7 to 9.2 mg/m3 (Figure 3 4). In the latitude and longitude averaged data, the Valdivian zone also had the most months with values above 10 mg/m3 when compared with the Osorno zone, which had the lowest chlorophyll a values, and the Llanquihue zone, which had median values (Table 31). The average span of high chlorophyll a values was the largest latitudinally and
72 longitudinally in the Valdivia zone, while the latitudinal and longitudinal coverage of low chlorophyll a values was the smallest in the Valdivia zone (Table 3 1; Figure 3 5). Osorno had the opposite trend and the Llanquihue zone had intermediate values. The area influenced by plantations was concentrated between 39.3S to 39.8S, with the highest chlorophyll a values located between 39.4S to 39.6S in t he Valdivia zone. Mixed landuse influences were seen from 41 .2S until 41.7S in the Llanquihue zone. The least influences were noted from 40.7S to 40.9S, which is related to the high a mount of native forest land cover in the Osorno Zone. The ANOVAs and Tukey Kramers multiple comparisons confirmed that chloroph yll a means were significantly different among zones (Figure 3 6). The Valdivian zone always had significantly higher chlorophyll a values than Llanquihue and Osorno. The most striking difference was between the Osorno and Valdivian zones in the summer months. The Valdivian zone had about twice the amount of chlorophyll a in the summer (5.2 compared to 2.2 mg/m3). Chlorophyll a in Llanquihue was statistically higher than Osorno in the summer and year round analysis, but no statistical difference was det ected between these two zones in the winter months. The same pattern was found even with yearly temperature fluctuations; s imilar patterns were shown when using year as a block in a randomized block design; across all years, the Valdivian zone had the hig hest chlorophyll a followed by Llanquihue, and then Osorno. Watershed Scale Findings C hlorophyll a values at the discharge point of the 13 watershed s were positively related to area (km2) of watershed covered in plantation in 1985 (r=0.726, p=0.007), are a (km2) of watershed converted to plantation from 1985 to 2001 (r=0.743, p=0.005), percent of watershed converted to plantation from 1985 to 2001 (r=0.744, p=0.006), area
73 (km2) of watershed converted to native forest from 1985 to 2001 (r= 0.841, p=0.001), and percent of watershed in urban land use in 1985 (r=0.706, p=0.010). The regression analysis showed that the area of plantation converted from 1985 to 2001 explained 55% of the variance in chlorophyll a values at the discharge point across sites; percen t of plantation converted from 1985 to 2001 explained an additional 17% of the variance, and percent urban areas explained an additional 7% (Table 3 2). The area of the watershed converted to forest plantations from 1985 to 2001 explained the high chlorophyll a values for Lingue (1) and Valdivia (3) watersheds (Figure 3 7). The Bonifacio (2) watershed, although classified as a mixed watershed (Chapter 2), did not have extensive plantations but did have high chlorophyll a values that resulted from the infl uence of the Valdivia watershed on the area. Interestingly, Bueno (5), which had substantial plantations, is similar to the native forest watersheds possibly because the Bueno watershed is so large and has extensive native forest coverage in the Andes Mountains that the influence of plantations could be mitigated. The native forest watersheds, Quilhue (6), Hueyelhue (8), Cholguaco (9) and San Luis (10), which all have few or no plantations, all exhibited the lowest chlorophyll a values; however Chaihuin (4 ) was more similar to the mixed watersheds, perhaps because the Valdivia watershed can influence parts of the Chaihuin watershed also because Chaihuin has some plantations in the watershed. The relationship between the area (km2) of native forest converted and chlorophyll a show the opposite trend with more native forest cover associated with lower chlorophyll a values (Figure 3 8). Isotopic Signatures of Loco Shellfish The raw 1 5N and 13C values in the loco shellfish were enriched (+2 3) in plantation influenced sites when compared with native forest and mixed landscape
74 influenced sites (Figure 3 9). Nitrogen likely originated from terrestrial sources; however, carbon exhibited a more characteristic marine signature. The enriched 1 5N values in the pl antation influenced area and depleted 1 5N values in areas near native forest, s uggest that nutrient s from forest plantations are a likely explanation for the relatively high phytoplankton biomass (as determined from chlorophyll a) observed with the SeaWi Fs satellite images. DISCUSSION N utrient dynamics in watersheds with forest plantations are driven by anthropogenic nitrogen inputs and the associated nearshore coastal waters show elevated chlorophyll a values The chlorophyll a patterns are consistent wi th Hebbelns (2000) findings that suggest upwelling is not the main driver of phytoplankton production in the region. The small upwelling areas identified by Moreno et al. (1998) in the summer months could provide additional nutrients to the system, but th ese small upwelling events do not explain the zonal differences in chlorophyll a patterns. Instead zonal differences are related to upland land use. Although Hebbeln (2000) mentioned that the Antarctic Circumpolar Current could influence the system, I find more evidence that the Humboldt Current brings nutrients north from the watershed outlet of the Valdivian river to the coastal region near the Bonifacio watershed. My findings also agree with Thiel et al. (2007) whom explain that in this region the Humbol dt Current influences nearshore physical dynamics yet the current is not responsible for major upwelling events. The increase in forest plantations in the past 15 ye ars both in area and percentage is the most important change that is influencin g chlorophy lla patterns. Urban area increase was also influential, but explained a much smaller portion of the variance in chlorophyll a patterns. The Valdivia and Lingue watersheds may provide some insights into the
75 percentage of land area converted req uired to aff ect chlorophyll a values. Using 2001 values, the Valdivia watershed had 1,425 km2 of plantations, covering almost 14% of this large 10,263 km2 watershed. Lingue had 182 km2 of plantations, covering 35% of the 527 km2 watershed. Both areas had elevated chl orophyll a values. The stableisotope analysis and increases in the chlorophyll a values show that nitrogen delivery in the rivers influences the ne arshore environment. The stable isotope findings confirm that nitrogen fertilizer that is used on plantati ons (Guerra et al. 2007) enters the rivers (Oyarzun et al. 2008; Little 2008) and enters the marine food web. Terrestrial nitrogen is an important source of nitrogen in the system since terrestrial nitrogen can be slightly enriched ( McKinney et al. 2001; F reyer & A ly 1974; Kreitler et. al. 1978). Marine nitrogen or atmospheric nitrogen did not explain the differences in isotopic signatures between native forest and plantationinfluenced watersheds because the nitrogen values would have been depleted (Peters on & Fry 1987). Terrestrial or marine carbon is not an influential source of carbon in the water near plantation watersheds because the 13C values would have been depleted (Trumbore & Druffel 1995; Coleman & Fry 1991) Instead atmospheric carbon is probably an influential component of carbon in the system since the 13C values were enriched (Coleman & Fry 1991) compared to native forest and mixed influenced areas A similar pattern of enriched 15N and 13C values was found in Frys (1999) study of clams where the highest chlorophyll a values were related to enriched carbon and nitrogen isotopic values Bade (2006) also showed a similar pattern; an increase in nitrogen and phosphorus was accompanied by an increase in chlorophyll a values and an elevated the uptake of CO2 in lakes. Schindler et al (1997) also show ed that an increase of
76 atmospheric carbon deposition wa s accompanied by increased primary productivity in lakes. A shift in the food web probably has not occurred for locos near plantationinfluenced watersheds compared to locos in native forest or mixed watersheds because although the nitrogen enriches according to Fry and Sherr (1984) and Miragawa and Wada (1987), the carbon does not. T he loco foodweb length may be different in plantation and native forest influenced sites, thereby causing higher 1513C values. As an organism consumes another organism at a higher trophic level, 15N and 13C values are generally enriched (DeNiro & Epstein 1978; Wada et al. 1987; Dauby et al. 1998). Likewise, foodweb studies have also shown that top carnivores with longer food chains have enriched 15N and 13C values (Cabana and Rasmussen 1994; Vander Zanden et al. 1999). If locos had a longer food web or consumed species at a higher trophic level, the locos probably consume a more diverse set of organism s. However, Morenos (2001) rapid survey of benthic organisms shows the opposite affect; a lower diversity score ( 0.2) was found in the forest plantation influenced area and a high diversity score (2) was found in the region associated with the native for est leaving landscape change as a more likely explanation of the 1513C values Conclusions and future research Forest plantations are an important factor in driving nutrient dynamics of the nearshore environment in southern Chile The photosynthe tic biomass patterns are related to upland land use; higher chlorophyll a concentrations are found in waters that are influenced by forest plantations. Isotopic signatures of the loco shellfish confirm that terrestrial nitrogen is a main source of nutrient s in the marine food web in areas influence d by forest plantations. Carbon is derived from the atmosphere in areas
77 influenced by plantations. The increase in photosynthetic organisms fix atmospherically derived carbon, and the distinct carbon signatures a re shown in the tissue of the loco shellfish from plantation regions. The changes in nutrient dynamics in plantation regions are associated with other changes in the nearshore fisheries. Fishers in the Valdivian region have reported that loco shellfish a re smaller in the region associated with landscape change and scientists (Moreno personal communication) have noted an increase in parasites on loco shells in the past 15 years. Indeed I found that loco shellfish that originated from the region with planta tion development are distinct and have more epibionts and boring organisms (Van Holt in prep c) and fishers get paid lower prices for these resources (Van Holt in prep d). In particular, the Chilean TURF management strategy could be in jeopardy if management areas do not yield the same quality and quantity of products because of landscape influences. Since fishers have no legal alternative areas to harvest their loco resources they face losing their fishing livelihood if the quality of the resources decrea ses substantially. To confirm the influence of plantation forests, a future study could analyze the isotopic signatures of carbon and nitrogen from sediment cores near shore waters adjacent to plantation areas and compare these to sediments obtained nears hore coastal waters This analysis would also provide a richer understanding of how nutrient dynamics may have shifted through time, something that is difficult to determine with satellite images because consistent datasets are lacking prior to 1998. Also, a more complete understanding of the isotopic changes in the food web could be undertaken by measuring additional components of the food webs in native forest and plantationinfluenced region.
78 Forest plantations located in coastal areas could influence nutrient dynamics in other systems as well where upwelling is not the main source of marine nutrients. We need to improve our understanding of linkages in the nearshore between the terrestrial, atmosphere, and marine environments Over half of the worlds population lives in coastal systems (Cohen et al. 1997) and extensive development occurs along the coast; the livelihoods of the worlds 13.1 million artisanal fishers depend on the nearshore environment (FAO 2002). Meso scale analyses of landscape change and related changes in the primary production characteristics are an important step toward understanding the nutrient dynamics in the nearshore With an understanding of the nutrient processes, we can begin to understand the underlying mechanisms that link landscape change to fisheries success in the nearshore.
79 Table 3 1. Monthly chlorophyll a (mg/m3) characteristics number of months with high and low chlorophyll a values and latitudinal and longitudinal distance covered by high and low chlorophyll a patch es -in the Valdivia, Osorno, and Llanquihue zones from 1998 to 2005. Area, longitude, and latitude Hovmoller diagrams were used to generate values. Measure Valdivia Osorno Llanquihue Component that was Averaged No. of months above 5 mg/m3 23 5 14 Area No. of months above 10 mg/m3 23 3 15 Longitude Mean latitudinal distance of values above 10 mg/m3 0.20 0.08 0.11 Longitude Maximum latitudinal distance of values above 10 mg/m3 0.7 0.1 0.3 Longitude No. of months below 0.6 mg/m3 2 24 20 L ongitude Mean latitudinal distance of values below 0.6 mg/m3 0.10 0.30 0.28 Longitude Mean latitudinal distance of values below 0.6 mg/m3 1 1.0 0.8 Longitude No. of months above 10 mg/m3 26 10 23 Latitude Mean longitudinal distance of values above 10 mg/m3 0.25 0.21 0.22 Latitude Mean longitudinal distance of values above 10 mg/m3 0.4 0.3 0.3 Latitude No. of months below 0.6 mg/m3 18 38 21 Latitude Mean longitudinal distance of values below 0.6 mg/m3 0.19 0.29 0.24 Latitude Mean longitudinal distance of values below 0.6 mg/m3 0.3 0.8 0.5 Latitude
80 Table 3 2. The best regression models for 1, 2, and 3 independent variables. The best model was determined by R2, Adjusted R2, C(p), AIC, and BIC values. The percent variation ex plained by each factor was calculated by comparing R2 values of the best models. Where PL_CH = area (km2) of the watershed converted to forest plantations from 1985 to 2001; PPL_CH = percent of the watershed converted to forest plantations from 1985 to 2001; PURB_CH = percent of the watershed converted to an urban land use from 1985 to 2001. Variables R 2 Adj.R 2 C(p) AIC BIC Variables in Model 1 0.55 0.50 9.50 25.76 25.21 PL_CH 2 0.73 0.66 4.78 29.57 26.92 PPL_CH PL_CH 3 0.80 0.72 4.00 31.15 25.65 PPL_CH PL_CH PURB_CH
81 Figure 3 1. Study region in southern Chile showing zonal, watershed, and stable isotope analysis scales. In the zone scale analysis, chlorophyll a values were analyzed in the Valdivia (39S to 40S and 73W to 74W), Osorno, (40S to 41S and 73.5W to 74.5W), and Llanquihue (41S to 42S and 73.5W to 74.5W) areas. The names of the zones correspond to the provinces (noted on the map). In the watershed analysis, the chlorophyll a values were obtained where the rivers in the corre sponding watersheds discharge into the sea. If more than one river discharged to the sea, the main river was selected. Samples for stable isotope analysis was collected at four sites (noted by black circle). In the map the Lingue watershed is represented b y (1), Bonifacio (2), Valdivia (3), Chaihun (4), Bueno (5), Quilhue (6), Contaco (7), Hueyelhue (8), Cholguaco ( 9) San Luis (10), Llico (11), Putratrn (12) and Maulln (13). The names of the watersheds are derived from the main river in the area or the n ame of the region.
82 Figure 32. Supervised classification of October 5, 1985 (Landsat TM, path 233, rows 8789) and November 29, 2001 (Landsat ETM, path 233, rows 8789) images showing native forest, plantation, matorral agriculture, cleared land, sand, water, urban, snow, and wetland land cover/uses. Watersheds are numbered as in Figure 3 1.
83 Figure 3 3. Latitudinal chlorophyll a (mg/m3) values (longitude averaged) derived from SeaWiFS satellite and seasurface temperatures from MODIS in the Valdivia zone in 2003. Regions with high chlorophyll a values are not related to low (from 9C to 10C 9 C) and therefore upwelling probably does not play a role in chlorophyll a patterns in the area.
84 Figure 3 4. Monthly SeaWiFS chlorophyll a averages from 199 8 to 2005 for the Valdivia, Osorno, and Llanquihue zones.
85 Figure 3 5. Chlorophyll a concentration (mg/m3) Hovmoller diagrams for the Valdivia, Osorno, and Llanquihue zones. These diagrams are based on monthly SeaWiFS satellite images and describe monthly chlorophyll a patterns from 1998 to 2005. Images were generated using NASAs Giovanni online software. The Valdivia zone has more frequent, high chlorophyll a patterns that extend across more degrees latitude, the Osorno zone has more frequent, low chl orophyll a patterns and no high chlorophyll a patterns. The Llanquihue zone has a mixture of high and low chlorophyll a patterns.
86 0 1 2 3 4 5 6 Winter Summer Year-Round Season Mean Chlorophyll-a (mg/m3) Valdivia Osorno Llanquihue a a a c c b b b b Figure 3 6. Mean chlorophyll a values (mg/m3) as determined by the SeaWiFS satellite images from 1998 2005 in the winter (April to July), summer (November to February), and year round, in the Valdivia (39 to 40), Osorno (40 to 41), and Llanquihue (41 to 42) zones. ANOVAs were highly significant for all analyses (p value<0.01) Lower case letters over the bars signify significant differences of mean chlorophyll a values as indicated by Tukey Kramers multiple comparison tests ( =0.05). All significant comparisons were highly significant (pvalue<0.005) except for the Valdivia Llanquihue in the winter (p value=0.024).
87 1400 1200 1000 800 600 400 200 0 Area (km2) plantation change from 1985 to 2001 2.5 2.0 1.5 1.0 Average chlorophyll-a (mg/m3) for April to July (1998 to 2005) 10 9 8 6 4 2 13 12 11 7 5 3 1 Native Forest Mixed Forest PlanationWatershed Type Figure 3 7. Scatter plot of mean chlorophyll a values at the watershed outlet (an average of one pixel for 8 years) from April to July (19982005) based on the SeaWiFS satellite images and area (km2) converted to plantation in each watershed (identified by numbers 113; see figure 31) from 1985 and 2001.
88 0 -200 -400 -600 Area (km2) native forest change from 1985 to 2001 2.5 2.0 1.5 1.0 Average chlorophyll-a (mg/m3) for April to July (1998 to 2005) 10 9 8 6 4 2 13 12 11 7 5 3 1 Native Forest Mixed Forest PlantationWatershed Type Figure 3 8. Scatter plot of mean SeaWiFS chlorophyll a values at the watershed outlet from April to July (1998 2005) based on SeaWiFS satellite images and area (km2) change native forest land cover in each watershed (identified by numbers 113) from 19852001.
89 Figure 3 9. 1513C values for loco meat for native forest, mixed, and plantationinfluenced regions. See Figure 3 1 for location of sites.
90 CHAPTER 4 MESO SCALE ANALYSIS OF THE INFLUENCES OF LANDSC APE CHANGE ON LOCO ( CONCHOLEPAS CONCHOLE PAS ) HEALTH CONDITION CHARACTERISTICS Introduction Fishery managers are concerned about shellfish health condition. Of particular concern is the influence of epibionts and shell boring parasi tes on shellfish health because these organisms can cause extensive financial losses in aquaculture ( Grabowski et al. 2007). Less information is known about the distribution and causes of epibiont and shell boring parasites and their influence on other hea lth related characteristics in nature. Epibont and shell boring parasites abundance has been attributed to eutrophication of the nearshore environment (Gabaev et al 2005; Belan 2003; Chang et al. 1992; Beukema 1991; Llanso 1992; Crema et al. 1991; Losovska ya 1988; Risk et al. 1995; and Perus and Bonsdorff 2004), which is related to human activities on the landscape, as well host species characteristics (Wahl 1989). Over half of the worlds population lives in coastal systems (Cohen et al. 1997) and extensive development occurs along the coast; the livelihoods of the worlds 13.1 million artisanal fishers depend on the nearshore environment (FAO 2002). Modeling the relationship between shellfish health condition, landscape change, and related changes in the n earshore nutrient dynamics at the meso scale provides a new perspective on the distribution, causes, and management of epibionts and shell boring organisms. In addition, today many fisheries are becoming closedaccess to help reduce overharvest of marine r esources (Defeo & Castilla 2005). If epibionts, shell boring parasites, and associated health problems are related to landscape change, fishers working in management areas that are close to landscape change are at a disadvantage. The management systems tha t have been established to help fishers sustainably harvest resources may be in jeopardy if resource quality declines and fishers lose money.
91 This study examines if the health condition of locos ( Concholepas concholepas ) in southern Chile is affected by upland land cover change, specifically forest plantation development, and the associated changes in phytoplankton biomass in the nearshore. Landscape change can alter nearshore systems (Foley et al. 2005). When people fertilize the soil, clear the land, or change the species that dominate the ecosystem, nutrient and sediment loading increases in the watersheds and rivers and hydrology patterns of the rivers change (Nixon 1995; Farley et al. 2005; Iroume et al. 2006; Binkley & Re s h 1999 ; Stevens et al. 1994; Farely & Kelly 2005; Oyarzun et al. 2007). The nearshore system usually responds to the changing nutrient dynamics, sediments, and hydrological patterns by becoming more eutrophic. Eutrophic systems are enriched with nutrients and may also have more turbidity and decreased oxygen levels (Smith et al. 1999; Kemp et al. 2005). Scientists have documented changes in the community structure and food webs of organisms in eutrophic systems ( Dougherty & Morgan 1991; Bini et al. 1999; Eyre 2000; Grigalunas et al. 2001; Kiirikki et al. 2001; Ludsin et al. 2001; Anderson et al. 2002; Diaz and Rosenberg 1995; Kemp et al. 2005). Eutrophication also compromises health of organisms in these environments (Johnson et al. 2007) and an abundance of generalist parasites can be found in eutrophied systems (Zander & Reimer 2002). The compromised health of organisms can lead to decreased growth (McDiarmid et al. 2004), which influences the food web structure and function ( Marcogliese 2005) Locos are economically the most importan t shellfish f or artisanal fishers in Chile. The environmental changes in the region are especially relevant to the 23,000 artisanal fishers (over a third of Chiles artisanal fishers) in the northern portion of region X (Figure 4 1) who fish predominantly in designated management areas along the coast (SERNAP 2003). Locos are the topcarn ivores in the benthic food web and consume tunicates, barnacles, and mussels as their
92 main prey source ( Stotz et al. 2003). To determine if loco health condition is affecte d by landscape change and associated changes in the nearshore system, I measured loco health conditionthe percentage of epibiont and shell boring organisms on loco shells, lipid content in the loco muscle tissue, and weight of the loco muscle tissue -and related these characteristics to landscape change patterns and addressed the following questions: Do locos located in management areas influenced by forest plantations have more epibionts and shell boring organisms on their shells? Do locos located in man agement areas influenced by forest plantations have lower lipid content in their muscle tissue? Do locos located in management areas influenced by forest plantations weigh less than locos from other regions? The landscape is rapidly changing in Chile, an i mportant international trade partner on the global pulp market. In 2005, 8.8% of all national export income came from forest plantations which covered 2.1 million hectares or 13.4% of the total forest area (INFOR 2006) From 1985 to 2001 in southern Chile (administrative region X), forest plantations increased by 1,992 km2 (Van Holt a in prep.) Signs of eutrophication are present in the nearshore environment (Van Holt b in prep.) The majority of plantations are north of the Rio Valdivia (Figure 4 1). Fo rest plantations are fertilized with nitrogen, phosphorus, potassium, and boron (Geldres & Schlatter 2004; Schlatter 1977; Guerra et al. 2007) to increase productivity and subsequently, increased nitrogen levels have been measured in Chilean the rivers of watersheds with forest plantations (Little et al. 2008; Oyarzun et al. 2007). Also, i ncr eases in forest plantations have lead to lower water levels in rivers because of increased evapotranspiration (Iroume et al. 2006).
93 N ut rients that enter the water caus e an increase in nitrogen limited primary producers in the nearshore system which, in turn, increases photosynthetic biomass. Chl a values from 1998 to 2005 Seaviewing Wide Field of view Sensor ( SeaWiFS) measurements (an average of a onedegree by one deg ree summary) show that chlorophyll a values are higher (2.06 mg/m3) in the nearshore region near Valdivia, which has extensive forest plantations and lower (1.2 mg/m3) in the Osorno region that has extensive native forest (Van Holt b in prep; Figure 41). Chlorophyll a (CHLA) concentration values that are derived from the SeaWiFS satellite image are derived from an empirically based bio optical algorithm (OReilly et al., 2000) that is performed on reflectance values, which are derived from calibrated digit al counts that have been atmospherically corrected (Hu et al. 2009). Chlorophyll a concentration is one of the best predictors of photosynthetic biomass available today (Huot et al. 2007), although there are other photosynthetic pigments and biomass that s upports photosynthesis that are not related to chlorophyll a. Samples of loco shellfish muscle tissue from management areas near forest plantations have e nriched 1513C values (Van Holt b in prep). The enriched 1513C values indicate that terrestrial nitrogen is main source of nitrogen for the filter feeding organisms that locos consume. The increase in nitrogenlimited phytoplankton is also related t o an increase in atmospherically derived carbon in the plantation region. The variation in the 1513C values indicates that the nearshore waters of the Valdivian region are more eutrophic. In 2003, 1,572 tons of locos almost 20% of the countrys e ntire catch came from the Xth region; in 2003, fishers received up to US$ 1 for each loco. The Territorial User Rights Fisheries (TURF) (Bernal 1999) management program was established after the 1990 loco fishery collapse. Fishers gained fishing r ights to parcels of the ocean, called managed exploitation areas for benthic resources (MEABR), and the loco fisheries recovered. Consequently, fishers cannot
94 move to another management area if the environmental conditions degrade in their assigned region; therefo re upland activities and the consequences to fisheries are of concern to the fishers. In a pilot study, fishers reported and I noticed distinct differences between locos at various sites. Locos from the south are considered good, mainly because they ar e big and fishers are paid a premium price. Northern locos were considered bad and skinny. Indeed shells from the north had many epibionts and shell boring organisms some fishers reported that when they harvest a highly infected loco the shell cracks w hen they pull the loco from a rock rendering the meat useless for the market. In contrast, many shells in the south were clean with no shell boring organisms or epibionts. Because there was a notable difference in loco quality, I analyzed the percent cover of epibionts and shell boring organisms, which may represent a parasitic or commensal relationship, as a component of loco health condition. The shell boring organisms include phoronids ( Phoronis ssp.) polychaetes ( Class Polychaetea), most likely the nat ive Dodecaceria cf. opule ns (Moreno et a l 2006) and shell boring bivalves ( Hiatella solida ). B arnacles ( Verruca laevigata and Jehlius cirratus ) were the main epibiont. I expected that shells with many epibionts and shell boring organisms were from coasta l regions that were located near forest plantations and where the ocean is more eutrophic. Other studies have linked increased abundance of phoronids (Belan 2003; Chang et al. 1992), barnacles (Gabaev et al 2005) and bivalves (Risk et al. 1995) in compromi sed, eutrophic ecosystems that have high levels of chemicals, nutrients, and dissolved organic matter that originated from human disturbance of the landscape. Polychaetes are also abundant in eutrophic areas (Beukema 1991; Llanso 1992; Crema et al. 1991; L osovskaya 1988; and Perus and Bonsdorff 2004).
95 Methods Study Site The northern half of region X, an administrative region of Chile, contains three large watersheds with roughly one large watershed in each province The watershed names are derived from the main river in the watershed : Valdivia (watershed # 3), Bueno (5), and Maulln (13), and ten small watersheds -Lingue (1), Bonifacio (2), Chaihun (4), Quilhue (6), Contaco (7), Hueyelhue (8), Cholguaco (9) San Luis (10), Llico (11) and Putratrn (12) (Fi gure 41). High annual precipitation (2,0004,000 mm) and a mean annual temperature of 12.5C ( Eldridge & Pacheco 1987) gave rise to the temperate rainforest ecos ys tem. There are three provinces in the region. The Valdivian province (Figure 4 1) has extensive forest plantations and agriculture. The Osorno province has extensive native forest cover, agriculture activities, and a less extensive, but increasing, numbe r of plantations. The Llanquihue province has extensive agriculture and ranching, as well as matorral (new growth or shrubby forest) land cover and few plantations. Urban and commercial activities are not common in the area, with the exception of the Celul osa Arauco ( CELCO ) pulp mill, located in the northern part of the study site that began operating in late 2003. In the nearshore rocky intertidal ecosystem of Chiles central coast region, uniform wind patterns, topography, and temperature give rise to si milar invertebrate and fish communities (Moreno et al. 1979; Fernandez et al. 2000; Lancellotti & Vasquez 2000; Camus 2001) Terrestrial inputs are the main source of nutrients in the ecosystem (Hebbeln 2000; Van Holt b in prep). Small, coastal upwelling events reportedly occur in the summer months (January to March) in the northernValdivia region (Moreno et al. 1998) but these events dont contribute to regional differences in photosynthetic biomass. The photosynthetic biomass characteristics follow three zones, which are related to uplandland use (Van Holt in prep b). The Valdivia zone
96 (39S to 40S and 73W to 74W), had relatively high chlorophyll a values that ranged from 1 to 11.5 mg/m3, the Osorno zone (40S to 41S and 73.5W to 74.5W), had values ranging from 0.7 to 6.6 mg/m3, and the Llanquihue zone ( 41S to 42S and 73.5W to 74.5W) had values ranging from 0.7 to 9.2 mg/m3 (the ranges of the one degree by one degree blocks were created by averaging longitudinal data values from 1998 to 2005) (see Van Holt b for details). Also, chlorophyll a values at the discharge point of the watershed were positively related to percent and area (km2) of the watershed converted to plantation from 1985 to 2001 (Van Holt b in prep). The geomorphology is also similar throughout the region--the continental shelf extends to approximately 6.54 km and to 200 m depth (Gallardo 1984). Fjords, which have complex geomorphology and therefore have distinct ecological processes, are located further south of the study site ( Fernandez et al. 2000). The Humboldt Current is the main oceanographic system that influences the region (Lancelloti & Vasquez 2000; Fernandez et al. 2000) and typically flows north. Sampling and Data Processing Be tween June and August of 2003 (the loco harvest season) I sampled 30 locos from each of the 41 mana gement area for a total of 1260 locos (Figure 4 1) To characterize the health of each shell, I counted the presence of epibionts and shell boring bivalves. I generated a random sample of 25 points on the inner and outer shell surfaces. I identified and quantified phoronids (PPHOR), polychaetes (P POLYC), shell boring bivalves ( P BIVAV), and barnacles ( PBARN ) on each shell. Next I measured the weight of the muscle ( WGHMT ) and the viscera (WGHV ) with an Ohaus HH320 scale with a precision of 0.1g The length ( S HELGT ), weight ( WGHSH), height ( SHELHGT ), of each she ll was also measured. I also calculated the ratio of WGHV to SHELGT to create a fat index (FATIND); the purpose of this ratio was to characterize locos that were skinny (had a low weight) given their she ll size, which was a concern for some fishers.
97 Finally, on a subsample of six locos from each site, I analyzed the lipid content (P LIP) of the muscle tissue using a modified Folch et al. (1957) method for the isolation and purification of total lipids fro m animal tissues I then created distribution maps for the following variables: PPOLYC, PBIVAV, PPHOR, PBARN, PLIP, and WGHMT and used them in combination with the statistical analysis to describe regional loco health condition patterns. Statistical Analy sis I ran a cluster analysis on loco health and land cover change characteristics. The cluster analyses used polythetic, nonoverlapping, sequential agglomerative techniques. I f irst ran a hierarchical cluster analysis on the independent variables based on a correlation ma trix to eliminate highly multicollinear variables (SAS 2001 v8.2). Then a non hierarchical cluster analysis was used to generate the most stable, robust, cluster analysis, which was followed by a hierarchical cluster analysis that generated relationship trees (McGarigal et al. 2000). I tested four different fusion strategies (average linkage, single linkage, complete linkage and Wards minimum variance linkage ) and Wards best matched the groupings of the nonhierarchical analysis. I used the average value for each of the 41 management areas for the following variables: WGHMT, SHELGT, SHELHGT, PLIP, PPHOR, P POLYC, PBIVAV, and PBARN. For the LCLUC cluster analysis, 38 LCLUC variables that included 1985 and 2001 total area and percent area of each land cover as well as the area and percent change from 1985 to 2001 was used. I also included the chlorophyll a values for each management area. I then ran a stepwise discriminant analysis on the groupings of management areas obtained from the cluste r analysis of the health related characteristics to understand which health related variables helped to form the clusters (SAS 2001) I ran a canonical correlation to see if loco health condition characteristics were highly correlated to a combination of LCLUC characteristics (SAS 2001). T he squared canonical
98 correlation identified how well canonical variables correlated ; the standardized canonical coefficients identified which loco and LCLUC variables contributed the most to each canonical dimension. R esults Landscape Level Patterns of Loco Health Distribution There is a clear relationship between a high percentage of phoronids on loco shells and forest plantation cover (Figure 4 2; Figure 4 3). The loco shells that contained the greatest percentage cov er of phoronids (35.145%) were located north of Rio Valdivia (Watershed 3). Just south of Rio Valdivia, loco shells also contained some phoronids, up to 5% of the shell was covered. Locos had little to no phoronids south of the Valdivian region. Locos in the Valdivian region (both north and south of Rio Valdivia) had a high percentages of polychaetes (714%); however, a high polychaete percentage was also measured near the outlet of Rio Maullin (watershed 13), as well as near Bahia Mansa (Watershed 7), and Bahia San Pedro (Watershed 10); so other factors influence polychaete distribution as well (Figure 42). The locos from the Valdivian region (both north and south of Rio Valdivia) have a higher percentage of barnacle epibionts compared to areas south (Fig ure 4 2). The percent bivalve cover is higher along the coast of more remote watersheds with high percentage of native forest such as Bahia San Pedro (Watershed 10) and Bahia Mansa (Watershed 7), and just south of the Valdivian river (Watershed 4). Loco m eat in the northernValdivia region weighed less than locos from most other management areas; interestingly locos from some other remote regions also had low weight (Figure 4 2). Locos fro m the Llanquihue region had the highest weight, which follows report s from fishers that locos in the south are fatter; fishers hypothesize that locos in the south have more food resources and indeed the south is famous for the Austromegabalanus psittacus
99 barnacles that are a favored food choice for locos (Stotz et al. 2003) These barnacles appear to be abundant as I observed fishers harvesting these barnacles and also the barnacles were sold in local restaurants; normally fishers dont harvest resources that locos consume at large scales because the fishers want the loco f ood source to remain abundant. Loco meat further south of the northern Valdivia region had midrange weights. Lipid content in loco meat did not follow any specific patterns; although locos in the south and in remote places such as Manquemapu and El Manzan o had high lipid content in the loco meat, some management areas (Bonifacio A and Mississippi) in the northern Validivia region also had locos with high lipid content (Figure 4 2). Loco Health Condition Groups I identified four distinct groups of locos ( Figure 4 4). As expected, management areas that were geographically close together were in the same groups. The Northern Valdivia group contained locos from management areas from north of Rio Valdivia. The Valdivia group include s locos from management areas in the Valdivia province (below and a few above Rio Valdivia), but also include s locos from a few management areas ( Estaquilla B, Amortajado, and Cullinco) from the Osorno and Llanquihue regions. The Osorno group contains locos from the Osorno province, with the exception of Estaquilla and Guar Guar which are from Llanquihue. Within the Llanquihue group three subgroups exist two of which are the true Llanquihue group that has locos solely from Llanquihue and the first group in Figure 44 that has locos f rom all regions Percent phoronid infestation was the most influential variable that discriminated loco groups Percent shell boring bivalve cover was the next most important variable in discriminating groups followed by the height of the shell, weight of the loco meat, percent total barnacle cover and percent polychaet e cover (Tables 4 1 & 4 2). S hell length and percent lipid content of the loco meat were not significant discriminators. Northern Valdivia ha s at least five times more phoronids than any ot her group. Osorno ha s the m ost shell boring bivalves ( 3% total
100 cover) Northern Valdivia and Valdivia have at least 1.5 times more barnacles than Llanquihue and Osorno. Loco meat from Llanquihue weighs the most, on average at least 23 grams heavier. I ex cluded viscera weight, shell width, and fat index from the final cluster analysis model of locohealth characteristics because the variables were muli tcollinear with other variables In the LCLUC cluster analysis, I included only percent agricultural chang e, percent cleared land change, percent urban change, percent change in mattoral, area (km2) plantation change, percent plantation change, and chlorophyll a in the management area because the other variables were multicollinear. Relationships between Loco Health Related Characteristics and LCLUC Variables. The discriminant analysis of loco groups showed that percent change in plantations from 1985 to 2001 ( PLCH ) best explained the variation among groups, followed by the chlorophyll a c haracteristics (CHLA) (Table 4 3) The plantation and chlorophyll a patterns matched regional patterns the Valdivia province had the highest nearshore chlorophyll a and upland forest plantations and the Osorno province has the lowest amounts (Table 44). The overall misclassif ication rate from t he discriminant analysis is 3 7% but misclassifications are unevenly distributed among groups (Table 4 5 and 46). Osorno and Northern Valdivia, which represent loco health condition and LCLUC extremes, have the lowest misclassification rate (12%). In contrast, Valdivia ha s 89% of the management areas misclassified and Llanquihue has a 35% misclassification rate. C ross validation estimates of misclassification increased the estimate of the overall misclassification rate to 43 % (Table 4 7) Th e LOCO1 canonical correlation coefficient measures parasite presence because percent cover of phoronids and total barnacles have high weighting (Table 4 8). The LCLUC1 canonical coefficient represents eutrophication because the percentage of plantatio n change and chlorophyll a levels are positive have a high weight on LCLUC1 (Table 49). A high canonical
101 correlation between LOCO1 (the parasite measure) and LCLUC1 (the eutrophication measure) shows that locos with more phoronids and barnacles are associ ated with management areas that have high chlorophyll a values in the ocean, which stems from the nutrients input that flow from forest plantations to the nearshore environment. The squared canonical correlation for LOCO1 and LCLUC1 is 0.881 and a scatter plot of these factors demonstrates a linear and highly significant (p< 0.0001) relationship (Figure 4 5) All other canonical measures are more difficult to interpret because the correlations dont follow consistent patterns. Loco lipid content and loco wei ght were not important components of canonical correlates and therefore were not the driving factors that distinguished loco groups. Discussion The cluster analysis, discriminant analysis, and distribution maps show that locos in management areas north of Valdivia are distinctly less healthy than locos in other parts of the region. NorthernValdivia shells contain a high percent cover of barnacle epibionts, shell boring phoronids, and polychaetes, and they weigh less. The discriminant analysis demonstrates the link between high percent cover of phoronids and barnacles, upland plantations, and high chlorophyll a concentrations in the nearshore. Previous research (Van Holt b in prep) indicates that the distinct nutrient dynamics in the Northern Valdivia regio n are related to upland land use. The positive relationship I found between eutrophic environments (high chlorophyll a concentration) and presence of polychaetes, barnacles, and phoronids is demonstrated in many other studies worldwide (Belan 2003; Chang e t al. 1992; Korpinen et al. 2007; Beukema 1991; Llanso 1992; Crema et al. 1991; Losovskaya 1988; and Perus and Bonsdorff 2004; Zander & Reimer 2002) Percent cover of bivalves on loco shells does not relate directly to forestplantations. Instead, locos th at are larger in size and possibly older are infected with shell boring bivalves.
102 Locos that have bivalves also have a lower fat index or are locos flacos these locos have large shells but the meat weight is lower than expected given the shell length. Lo cos flacos occurred in some of the most remote sites. The misclassified observations in the discriminant analysis help to identify management areas that were not typical given the land cover/use and chlorophyll a characteristics. For example, Bonifacio A, was classified in the Valdivia group in the hierarchical cluster analysis using loco characteristics; however, if the locos were classified according to LCLUC characteristics, t hese loco would be classified with N orthern Valdivia. Bonifacio A has better l ocos than one would expect for the LCLUC and is considered by local fishers to be one of the best sites in northern Validiva, always producing abundant good quality locos Bonifacio A locos were probably classified in Valdivia because they h ave fewer phor onids than locos in Bonifacio B The Bonifacio B site, which is managed by the same group of fishers was not as productive Perhaps small scale upwelling as suggested by Moreno et al (1998) and Zamorano (pers. comm.) occurs in the Bonifacio A site and the upwelling diminishes the locos susceptibility to phoronids. Also local fishers and the scientist hired by the fishers (Zamorano pers. comm.) have reported that tunicates are abundant in Bonifacio A. Indeed locos from Bonifacio A have an unexpectedly high lipid content, which could be a result of more tunicates for the locos to consume. The abundant food source could help keep locos less susceptible to phoronids. I hypothesize t wo possible mechanisms to explain the expansion of epibionts and shell boring organisms on locos near plantation watersheds. Either the food web of the locos changed and locos bec a me we aker and susceptible to infestation, or the epibionts and shell boring organisms become more abundant because they have increased food resources or a hybrid of
103 both mechanisms are possible Gabaev et al. (2005) link increases of diatoms to an increase in barnacles. Enriched 1513C values of the loco shellfish meat suggest that high chlorophyll a values result from diatom blooms (Van Holt b). I hy pothesize that an increase in diatoms could also foster epibiont barnacle growth. Also it is possible that there are fewer food resources in Northern Valdivia. Moreno (2001) identified a lower diversity of organisms in Northern Valdivia. Fewer food resou rces could influence the locos ability to combat epibionts by not allowing loco to produce antifouling agents. Although the ability for locos to produce antifouling agents is purely hypothetical, antifouling agents are produced by other species and it is energetically costly to do so (Wahl 1989). So hypothetically locos with less food produce less antifouling agents and this causes increased infestation of epibionts. Percent phoronid cover is also correlated with percent polychaete cover, percent barnacle cover, and shell weight. Gallardo and Osorio ( 1978) hypothesized that shell boring bivalves originally penetrate loco shells after barnacles are present and, although I found no evidence of this phenomenon with shell boring bivalves (no significant correl ation exists between bivalve presence and barnacle presence), a significant positive correlation does exist between barnacle cover and phoronid and polychaete cover (Table 4 10). Gallardo and Osorio (1978) do have a promising hypothesis, however, that the presence of some epibionts fosters growth of other epibionts and boring organisms. I hypothesize that barnacle presence could leave locos susceptible to shell boring phoronids and polychaetes. Indeed Bers & Wahl (2004) show that substrate microtopography i nfluences epibiont presence and Wahl (1989) explains that the presence of some epibionts is dependent on another epibiont. Interestingly the length of locos does not correlate with presence of phoronids or polycheates, but the width is related to phoroni d presence (Table 4 10). Locos infested with
104 phoronids are narrower. Loco weight is negatively related to bivalves and not significantly related to phoronids or polychaetes. Since the management system restricts fisher movement, fishers are limited in ho w they can adapt to the changes in shellfish quality Fishers harvest in unsolicited areas outside of their management area, go further out to sea and harvest distinct species, or leave the fishery altogether. In the Mehuin region of Chile, an area that ha s poor quality locos fishers are leaving the fishery because of the low quality resources and a conflict with the Celco Arauco pulp mill. Fishers lose $1 per kg for each 10% increase in parasites (phoronids and polychaetes) (Van Holt in prep d). Plantati ons will likely increase in southern Chile and it is possible that locos in other regions will also experience changes in the shellfish quality as a consequence of this landscape change. An integrated coastal management approach is necessary to mitigate th e influence of epibionts and shell boring organisms, particularly phoronids, barnacles, and polychaetes on the loco shellfish. Conclusions In conclusion, landscape change and the associated changes in chlorophyll a values in the nearshore are the main factors that describe the distribution of phoronid borers and barnacle epibionts. Bivalve borers and lipid content are not linked to landscape change and chlorophyll a values. Boring polychaetes are slightly linked to the abovementioned environmental changes, but other factors also contribute to polychaete distribution. This mesoscale study shows that epibionts and boring organisms are abundant in areas where the nitrogen dynamics in the nearshore shift because of plantations or other landuse activities. If climate change influences the nearshore nutrient dynamics we can expect that similar changes in shellfish parasites can occur. Other regions of Chile with extensive land use change likely also have high levels of phoronids and barnacles, and possibly polyc haetes on loco shells. Additionally the presence of
105 one species barnacle, phoronid, or polychaete, may help another species colonize. In Chile and beyond, managers can monitor shellfish and use the epibiont and boring organism presence as indicators of environmental change over time. In Chile, the government could consider reducing the proposed cost associated with maintaining management areas that have high levels of parasites that affect a fishers income. Fishers lose $1 per kilo for each 10% increase of parasites on loco shellfish (Van Holt in prep d). We need to better understand how parasites influence resource quality. It is also important to better understand the relationship between harvest pressure and parasite abundance; it is difficult from the available data to tell whether harvest pressure or parasites or both factors reduce loco size. I hypothesize that harvest pressure accounts for only a small fraction of difference but an experimental study is needed to test this hypothesis. In Chile, the government may consider allowing fishers who have compromised management areas to solicit new management areas. Finally, this study demonstrates one reason why integrated co astal management is a necessary. Management systems in place have helped loco and other species recover but without addressing the relationship between landscape change, shifting nutrient dynamics in the nearshore, management areas, and shellfish parasites, the management system could be in jeopardy in Chile and in other managed fisheries in the nearshore
106 Table 4 1. Mean and standard deviation (in parenthesis) values of loco characteristics of each group in the hie rarchical cluster analysis for ph oronids ( PPHOR ), shell boring bivalves ( PBIVAV), polychaetes (PPOLYC ), barnacles (P BARN ), sh ell height ( SHELLHT ), loco meat weight ( WGHMT ), shell length ( SHELLGT ) and percent lipid in loco meat (P LIP ). GROUP PPHOR (%) P BIVAV (%) P POLYC (%) P BARN (%) SHELLHT (mm) WGHMT (g) SHELLGT (mm) PLIP (%) Llanquihue 0.25 (0.68) 0.64 (0.62) 4.42 (2.15) 27.34 (15.92) 53.18 (1.82) 121.4 (17.22) 112.21 (3.55) 7.11 (2.16) Osorno 0.00 (0.00) 3.03 (0.90) 5.23 (2.13) 12.71 (5.98) 55.63 (2.17) 91.4 (10.97) 115.23 (3.75) 7.95 (2.70) Valdivia 5.84 (9.61) 1.26 (0.49) 8.73 (3.47) 43.16 (16.74) 58.11 (2.38) 128.2 (14.70) 116.14 (1.60) 6.99 (2.18) Northern Valdivia 29.36 (7.49) 0.60 (0.57) 7.55 (1.78) 43.75 (11.71) 56.66 (1.35) 104.6 (14.76) 110.83 (3.10) 6.00 (1.59)
107 Table 4 2. Partial r square, significance levels, and average squared canonical correlation s from th e stepwise discriminant analysis of loco grouping using the loco ch aracteristics, mean values for ph oronids ( PPHOR ), shell boring bivalves (PBIVAV), polychaetes (PPOLYC ), barnacles (P BARN ), shell height ( SHELLHT ), loco meat weight ( WGHMT ), shell length ( SH ELLGT) and percent lipid in loco meat ( P LIP). SHELLGT and P LIP were not statistically significant at p < 0.05. Step Variable Partial R Square Pr>F Average squared canonical correlation 1 PPHOR 0.8199 <0.0001 0.273 2 P BIVAV 0.6728 <0.0001 0.488 3 SHELLHT 0.5033 <0.0001 0.620 4 WGHMT 0.3200 0.0034 0.672 5 PBARN 0.2225 0.0339 0.703 6 PPOLYC 0.2085 0.0497 0.737
108 Table 4 3. Partial r square, significance levels, and average squared canonical correlation of stepwise discriminant analysis of LCLUC characteristics using percent change in plantations from 1985 to 2001 (PPL_CH) and oc ean color characteristics (CHLA ) in a canonical discriminant analysis of the hierarchical cluster analysis. Step Variable Partial R Square Pr>F Average squared canonical correlat ion 1 PLCH 0.5917 <0.0001 0.197 2 CHLA 0.3370 0.0015 0.293 Table 4 4. Mean and standard deviation (in parenthesis) values of LCLUC characteristics, percent change in plantations from 1985 to 2001 (PPLCH) and ocean color characteristics (CHL A ), for the corresponding upland watershed for each group in the hierarchical cluster analysis of loco characteristics. Group PPLCH (%) CHL A (mg/m 3 ) Llanquihue 2 (3) 1.73 (0.47) Osorno 3 (3) 1.22 (0.12) Valdivia 6 (5) 2.06 (0.53) Northern Valdivia 16 (8) 2.45 (0. 18) Table 4 5. Number of observations percent correctly classified (diagonal) and misclassified (off diagonal) estimates of hierarchical cluster analysis groups Llanquihue, Osorno, Valdivia, and Northern Valdivia using percent change in plantations from 1985 to 2001 ( PLCH ), percent native forest in 2001 (PFNF), and ocean c olor characteristics (CHLA) from canonical discriminant analysis. Predicted Group Actual Group Llanquihue Osorno Valdivia Northern Valdivia Total Llanquihue 11 65% 5 29% 0 0% 1 6% 17 100% Osorno 1 13% 7 88% 0 0% 0 0% 8 100% Valdivia 4 44% 2 22% 1 11% 2 22% 9 100% Northern Valdivia 0 0% 0 0% 1 13% 7 88% 8 100% Total 16 38% 14 33% 2 5% 10 24% 42 100% Rate 0.35 0.13 0.89 0.13 0.37 Priors 0.25 0.25 0.25 0.25
109 Table 4 6. Probability of group membership of mis classified observations of the canonical discriminant analysis of hierarchical cluster analysis groups Llanquihue, Osorno, Valdivia, and Northern Valdivia using percent change in plantations from 1985 to 2001 (PLCH) percent native forest in 2001 (PFNF), and ocean color characteristics (CHL A ). Site From group To group Llanquihue Osorno Valdivia Northern Valdivia Amargos Llanquihue Northern Valdivia 0.00 0.00 0.19 0.81 Cholhuaco Llanquihue Osorno 0.06 0.87 0.07 0.00 Chaihuaco Llanquihue Osorno 0.34 0.53 0.13 0.00 Capitanes Llanquihue Osorno 0.24 0.69 0.08 0.00 Punta Colun Llanquihue Osorno 0.08 0.89 0.03 0.00 Roco Covodonga B Llanquihue Osorno 0.37 0.52 0.11 0.00 Estaquilla Osorno Llanquihue 0.57 0.21 0.22 0.00 A mortajado A Valdivia Llanquihue 0.70 0.00 0.30 0.00 Niebla Valdivia Llanquihue 0.61 0.00 0.39 0.00 Huape A Valdivia Llanquihue 0.48 0.00 0.34 0.18 Huiro Valdivia Llanquihue 0.59 0.00 0.41 0.00 Cullinco Valdivia Osorno 0.34 0.53 0.13 0.00 E staqu illa B Valdivia Osorno 0.06 0.87 0.07 0.00 Bonifacio A Valdivia Northern Valdivia 0.00 0.00 0.19 0.80 Los Molinos A Northern Valdivia Valdivia 0.01 0.00 0.55 0.44 I sla Mancera Valdivia Northern Valdivia 0.00 0.00 0.19 0.81 Table 4 7. Cross validation summary and error count estimates of the canonical discriminant function of hierar chical cluster analysis groups Llanquihue, Osorno, Valdivia, and Northern Valdivia using percent change in p lantations from 1985 to 2001 (PC PL ) and ocean color charact eristic s (CHLA) From Group Llanquihue Osorno Valdivia Northern Valdivia Total Llanquihue 9 55% 5 29% 2 12% 1 6% 17 100% Osorno 1 13% 7 88% 0 0% 0 0% 8 100% Valdivia 4 44% 2 22% 0 0% 3 33% 9 100% Northern Valdivia 0 0% 0 0% 1 13% 7 88% 8 100% Total 14 33% 14 33% 3 7% 11 26% 42 100% Rate 0.47 0.13 1.00 .13 0.43 Priors 0.25 0.25 0.25 0.25
110 Table 4 8. Standardized canonical coefficients for the three canonical variables, LOCO1, LOCO2 and LOCO3 that describe the association between Loco Health Variables and l oco groupings. Note: phoronids = PPHOR shell boring bivalves = PBIVAV, polychaetes = PPOLYC, barnacles = P BARN shell height = SHELLHT, loco meat weight = WGHMT, shell length = SHELLGT and percent lipid in loco meat = PLIP. Loco Variables Loco1 Eutroph ication Loco2 Loco3 WGHMT 0.253 0.452 0.173 SHELLGT 0.188 0.424 0.212 SHELLHT 0.069 0.478 0.066 P BARN 0.438 0.130 0.950 P POLYC 0.193 0.234 0.073 PPHOR 0.675 0.030 0.770 P BIVAV 0.240 0.680 0.236 P LIP 0.032 0.041 0.076 Table 4 9. Standar dized c anonical coefficients for the three canonical variables, LCLIC1, LCLUC2, and LCLUC3 that describe the importance of each landscape change variable on the canonical variable. Note land cover change is from 1985 to 2001 and PAG_CH = percent change in agriculture, PCL_CH = percent cleared land change, PURB_CH = percent urban land class change, PMAT_CH = percent mattoral change, PL_CH = area (km2) plantation change, and PPL_CH = percent plantation change LCLUC Variables LCLUC1 Plantation LCLUC2 LCLUC3 PAG_CH 0.187 0.218 0.832 PCL_CH 0.245 0.401 0.624 PURB_CH 0.028 0.303 0.474 PMAT_CH 0.036 0.582 0.627 PL_CH 0.294 0.565 1.264 PPL_CH 0.702 0.595 0.437 CHL A 0.399 0.393 1.316
111 Table 4 10. Pearsons Correlation among loco health condition vari ables where WGHMT = weight of loco meat (g), WGHV = weight of loco viscera (g), WG SH = shell weight (g), SHE LGT = shell length (cm), S HELHGT = shell height (cm), SHE LWDT = shell width (cm) FATIND = fat index (WGHMT/SHE LGT), PLIP = percent lipid (per xx g) in loco meat, PPOLYC = percent polychaete cover on shell, PPHOR = percent phoronid cover on shell, PBIVAV = percent bivalve cover on shell; PBARN = percent barnacle cover on shell; MALE = sex, male coded as 1. Note: ** is highly significant where p<0.01 a nd is significant where p<0.05. WGHMT W GHV W G HSH SHE LGT SHELHGT SHEL W D T FATIND PLIP PPOLYC PPHOR PBIVAV PBARN MALE WGHMT 1.00 0.78 ** 0.31 0.35 0.13 0.37 0.98 ** 0.15 0.10 0.23 0.39 0.22 0.34 W GHV 0.78 ** 1.00 0.09 0.08 0.10 0.01 0.80 ** 0.16 0.05 0.03 0.44 ** 0.11 0.26 W G SH 0.31 0.09 1.00 0.51 ** 0.50 ** 0.49 ** 0.23 0.17 0.54 ** 0.42 ** 0.13 0.71 ** 0.05 SHE LGT 0.35 0.08 0.51 ** 1.00 0.36 0.80 ** 0.17 0.03 0.14 0.31 0.43 ** 0.06 0.08 SHELHGT 0.13 0.10 0.50 0.36 1.00 0.16 0.07 0.03 0.35 0. 26 0.12 0.18 0.01 SHE LW D T 0.37 0.01 0.49 ** 0.80 ** 0.16 1.00 0.22 0.03 0.16 0.44 ** 0.28 0.27 0.17 FATIND 0.98 ** 0.80 ** 0.23 0.17 0.07 0.22 1.00 0.18 0.08 0.18 0.49 ** 0.22 0.37 PLIP 0.15 0.16 0.17 0.03 0.03 0.03 0.18 1.00 0.04 0.13 0.27 0.44 ** 0.07 PPOLYC 0.10 0.05 0.54 ** 0.14 0.35 0.16 0.08 0.04 1.00 0.46 ** 0.07 0.30 0.24 PPHOR 0.23 0.03 0.42 ** 0.31 0.26 0.44 ** 0.18 0.13 0.46 ** 1.00 0.19 0.40 ** 0.13 PBIVAV 0.39 0.44 ** 0.13 0.43 ** 0.12 0.28 0.49 ** 0.27 0.07 0.19 1.00 0.28 0.18 PBARN 0.22 0.11 0.71 ** 0.06 0.18 0.27 0.22 0.44 ** 0.30 0.40 ** 0.28 1.00 0.11 MALE 0.34 0.26 0.05 0.08 0.01 0.17 0.37 0.07 0.24 0.13 0.18 0.11 1.00 N 41 41 41 41 41 41 41 41 41 41 41 41 41
112 Figure 4 1. Management sites where l oco shellfish were sampled. Where AMG = Amargos, AMOA = Amortajado A, BMANA= Bahia Mansa A, BMANB = Bahia Mansa B, BONIA = Bonifacio A, BONIB = Bonifacio B, NPCHOC = Norte Punta Chochoi, BSP = Bahia San Pedro, CHAIA = Chaihuin A, CHAN = Chan Chan, CHOL = C holhuaco, CHUA = Chaihuaco, CULL = Cullinco, ESTAQ = Estaquilla, ESTAQB = Estaquilla B, FARE = Farellones, GUAR = Guar Guar, NIEB = Niebla, HUAPA = Huape A, HUAPB = Huape B, HUIR = Huiro, ISMAN = Isla Mancera, ISRE = Isla del Rey, CAP = Capitanes, LMOLA = Los Molinos A, LMOLB = Los Molinos B, MANQA = Manquemapu A, MANQB = Manquemapu B, MAQ = Maquillahue, MEHUB = Mehuin B, MISSI = Mississippi, MPUC = Manzana Pucatriuhue, PLBAN = Piedra Blanca, PCHOC = Punta Chocoi, PCOL = Punta Colun, PNUM = Punta Numpulli, POR = Punta Ortiga/Puga, PQUI = Punta Quillahue, PQUIB = Punta Quillahue B, PUP = Pupelde, ROCOB = Roco Solataria B, ROSO = Roco Solataria
113 A B C D
114 E F Figure 4 2. Average percent A) phoronid, B) polychaete, C) barnacle, and D) bivalve cove r of on loco shells as well as E) lipid content and F) weight of loco meat in each management area and the percent forest plantation cover in 2001. (Averages obtained from a sample of 30 locos per management area).
115 A B C D Figure 4 3. Loco shel ls showing A) phoronid, B) polychaete, C) bivalve and D) barnacles. A,B, and C are the ventral side of the loco shell. D is the dorsal side. Phoronids leave light brown marks on shell A. The dark black lines in B are polychaetes. The circular marks on shel l C reflect the damage from the bivalves..
116 Figure 4 4. Cluster analysis of loco shellfish in each management area based on health characteristics (percent phoronid, polychaete, and bivalve cover on loco shells, as well as weight and lipid content of meat and shell length) Distinct levels of gray signify distinct clusters.
117 2 1 0 -1 LAND1 2 1 0 -1 LOCO1 ROSO ROCOB PUP PQUIB PQUI POR PNUM PCOL PCHOC PBLAN MPUC MISSI MEHUB MAQ MANQB MANQA LMOLB LMOLA CAP ISRE ISMAN HUIR HUAPB HUAPA NIEB GUAR FARE ESTAQB ESTAQ CULL CHUA CHOL CHAN CHAIA BSP NPCHOC BONIB BONIA BMANB BMANA AMOA AMG R Sq Linear = 0.881 Figure 4 5. Scatterplot of LOCO1 and LAND1 from the canonical correlation analysis (R2 =0.881; p=0.000). Abbreviations as in Fig. 4 1)
118 CHAPTER 5 ADAPTIVE CAPACITY FOR FISHER SUC CESS Introduction Nearshore Systems and Landscape Change This study examines how nearshore fishers on the coast of Chile adapt to environmental change. Human activities that include agriculture, deforestation, plantation development, animal production, a nd urbanization can alter nearshore systems, which receive outflows from the upland ecosystem (Nixon 1995). To develop the landscape, people clear the land and fertilize the soils and these activities increase nutrient loading and sediment delivery to the rivers (Nixon 1995). Land use/land cover change (LULCC) has a major influence on the hydrologic regimes of the rivers since evapotranspiration rates and the amount of water that flows from the watershed to the rivers depends on the dominant species in the landscape (Farley et al. 2005; Iroume et al. 2006) Nearshore systems respond to nutrient increases by becoming more eutrophic. Eutr ophic systems are enriched with nutrients and may also be more turbid and have decreased oxygen levels (Smith et al. 1999; Kemp et al. 2005). Consequently phytoplankton blooms in eutrophic areas and extreme changes in the phytoplankton dynamics can alter t he benthic community structure (Diaz and Rosenberg 1995) and foodweb structure and function (Kemp et al. 2005). Organisms living in eutrophic environments are also more susceptible to parasites ( Zander & Reimer 2002; Johnson et al. 2007; Van Holt in prep c) This cascade of effects may cause problems for nearshore fishers in Chile because they are restricted by recent management laws to specific areas, and cannot move to new areas if upland LULCC alters the nearshore environment in ways that degrade the
119 fish eries. To understand how nearshore fishers adapt to environmental change, I answered the following questions: What factors (knowledge, technology, and environment) contribute to fisher success? Do fishers use the same strategies to adapt for success in th e loco (closed access) and congrio (openaccess) fishery? In southern Chile, the nearshore environment is influenced by forest plantation development, which is the most significant landscape change that occurred from 1985 to 2001. Forest plantations are f ertilized (Schlatter 1977; Guerra et al. 2007) to increase productivity and subsequently, increased nitrogen levels have been measured in watersheds with forest plantations (Little et al. 2008; Oyarzun et al. 2007). I ncr eases in forest plantations have als o lead to increased evapotranspiration and consequently a reduction in the amount of water that flows to the rivers (Iroume et al. 200 6) The nearshore environment that receives outflows of nutrients from plantation areas has higher photosynthetic activity, which has been measured by high chlorophyll a concentrations in data derived from Sea viewing Wide Field of view Sensor ( SeaWiFS) satellite images (Van Holt b in prep). The loco ( Concholepas concholepas ) shellfish from the regions influenced by plantations have a higher parasite load (Van Holt c in prep) and weigh less. Fisheries Management in Chile Recently Chile implemented a T err itorial User Rights Fisheries (TURF) management system to protect benthic species from overharvesting by providing fishers
120 quasi property rights to parcels of the ocean called managed exploitation areas for benthic resources (MEABR) and exclusive rights to the benthic resources in those areas (Bernal et al. 1999) Fishers usually only harvest loco and refrain from harvesting other benthic species in management areas to avoid disrupting the valuable locos and also because fishers must hire a scientific expert to create quotas for all species harvested in the MEABRs an expense that few fishers want to incur. If fishers harvest another species, they usually harvest lapa ( Fisserella sp) at a very small scale. The MEABRs were implemented to control harvest of the loco fishery, a benthic fishery that collapsed in the mid 1980s from overfishing. In addition to successfully bringing back the loco fishery, the new management system changed how fishers can respond to environmental change because the fishers can no longer move to another region if the resource quality is poor in their management area. Fishers can legally harvest locos only from their management area. Today most of the coast in the study site (Figure 5 1) is registered to fishing syndicates that consist of local village level groups. Fishers cant shift to harvest in regions other than their management area to escape the consequence of environmental changes in their management area and in the coastal system. Fisher knowledge varies greatly because of how the MEABRs were established. Essentially anyone that is a certified diver, fisher, or coastal resource gatherer can join a fisher syndicate and solicit a management area from Servicio Nacional de Pesca (SERNAPESCA). People were attracted to government assistance and the potential to earn money from the coast and as a result, many new people entered the fishery. Many of these new fis hers are part timers, have less experience than traditional fishers, and have
121 other sources of income. When prices are low, these part timers stay on shore and await better prices. In contrast, those who depend entirely on fishing, and have traditionally f ished marine resources with boats, have better boats fishing knowledge, and organizational skills, yet they cannot always wait for a better price. These fishing skills, technology, and greater marketing experience, allow the full time fishers to capture s pecies that are beyond the reach of the part timers. More experienced divers often work in the rivers as well as the management areas and harvest multiple species of benthic shellfish not only locos More experienced fishers also spend more time offshore and harvest multiple species of fin fish. In contrast, the newer fishers and those with other livelihoods tend to harvest resources only in the management areas. To understand how fishers adapt to environmental change, I used a space for time substitution ( Fukami 2005) and measured the success of fishers who worked in environments that had varying levels of eutrophication and associated natural resource degradation (Figure 5 1). Adaptation takes time, so instead of following a group of fishers over time, I examined cases that stand in for different times along the adaptation process. I also compared the factors that predicted success across a closed access ( loco shellfish) and openaccess [ congrio ( Genypterus sp.) ] fishery to determine if fishery type and ma nagement regimes (closed or open access) affected success. Success was my proxy for adaptation I defined successful fishers as those who caught more fish, made more money, or gained a higher price per kg of fish than unsuccessful fishers. Studies that mea sure fisher success (usually measured as catch per unit effort), find that successful fishers have more knowledge and technology compared to
122 unsuccessful fishers (Acheson 1977; Palsson 1994; Palsson & Dur renberger 1982) Because new fishers were introduced to the system when the management areas became established, the amount of time that fishers spend in the open ocean, rivers, and management areas are important indicators of knowledge because they help separate new and experienced fishers; fish species harvested and other livelihood activities also help isolate new and experienced fishers. Presence of boats was the most relevant indicator of technology since established fishers usually had more boats and novices had fewer boats. Knowledge, however, is dependent on the environment and therefore the environment is a relevant factor in success; Harris (1979) demonstrates that the derivatives of culture or knowledge stem from the environment, production, and infrastructure. Including the environment was particularly important in this study because the environmental conditioned varied throughout some fisheries management areas were influenced by landscape change. These management areas had higher chlorophyll a concentrations in the water and the locos harvested from those management areas had higher parasites on their shells. I included environmental variables that measured the level of eutrophication in the system (chlorophyll a concentration in the ocean and t he percentage of parasites on the loco shellfish) because these factors likely affect fisher success and can isolate how fishers adapted to environmental change. Spatial measures that address infrastructural features of the landscape such as management are a size, distance from the management area to the fishing headquarters, and distance to Valdivia, the central market, were also included because these factors can influence success because they are derivatives of
123 knowledge and fisher movement is restricted under the new management system and those fishers who have less time to travel may gain an advantage. To measure success, I included catch per unit effort (CPUE) measures, when available, because CPUE has traditionally been used to measure success in the literature. Monthly income, the total profit that each fisher gained, also measured success; this income measure took into account the costs associated with fishing such as fuel costs, payment for management area scientists, and equipment costs. Finally, to compare success in loco and congrio fisheries, I used price per kg, which measures broadscale regional differences in success. Since the interviews with fishers were conducted over two months when no major fishing events happened, the prices reported were not subject to seasonal influences. Study Site The Valdivian province of Chile contains four large watersheds (Figure 5 1). The names of watersheds on Figure 5 1 are derived from the main river in the watershed: Lingue (1), Bonifacio (2), Valdivia (w atershed 3), and Chaihun (4). High annual precipitation (2,0004,000 mm) and a mean annual temperature of 12.5C ( Eldridge & Pacheco 1987 ) g i ve rise to a temperate rainforest ecosys tem, today the remaining forest is located along the coastal and Andes mountains (Figure 51). Principal economic activities include forestry, fisheries, agriculture, and cattle ranching. Forest plantations represent the most significan t land cover change in the past 15 years (Van Holt a in prep). Forest plantations are extensive in the Lingue and Valdivia watershed. Chaihuin is covered by mainly native forest, but also has some forest plantations. Some agriculture occurs in the
124 Lingue, Bonifacio, and Valdivia watershed. The nearshore environment from the outlet of the Validiva river to the outlet of the Lingue river has high chlorophyll a levels that indicate that nitrogen from the forest plantations are leaching into the rivers and even tually affecting the chlorophylla concentration in the nearshore environment. Across the Valdivian province, located in the northern part of administrative Region X (Figure 51), artisanal fishers dive for benthic locos in the management areas yet fishe rs are not restricted to harvest congrio only in the management areas. Harvesting locos in the management areas is a group effort a president of each syndicate helps process the legal paper work and organizes the group; however, the money that individual f ishers receive are not always divided equally and the way that money is divided varies from syndicate to syndicate. The fishers tend to work in smaller groups for other fisheries, such as congrio Although prices may differ slightly within a fishing syndic ate because fishers have distinct rules on how to divide the money; the going price for a resource is relatively similar in the same region. On average, in management areas across Region X, artisanal fishers harvested 1,979 tons of locos each year betwee n 2003 and 2005. The finfish catches are very small; from 20032005 the entire purse seine and artisanal fishers together, across region X, reported selling, on average, 116 tons of congrio annually (SERNAP 2003, 2004, 2005) Forest plantations have increased differentially across the study site and environmental change in the nearshore environment follows forest plantation development patterns (Van Holt d in prep.). From 1985 to 2001, in the northern portion of the study site (across watersheds 1 3), 1,415 km2 of plantations have become
125 established. In comparison, across in the southern portion of the study site (watershed 4), only 27 km2 of land has converted to plantations. Chlorophyll a concentration values are also higher (3.7 mg/m3) in the nearshore region with plantations as opposed to 2.0 mg/m3 across all months between 1998 2005 (Van Holt c in prep). Finally, loco shellfish parasites are also higher in plantationinfluenced areas with an average of 30% of the shell covered with phoronid ( Phoronis ssp.) and polychaete ( Class Polychaetea) parasites in the plantation areas and about 5% cover in nonplantation areas (Van Holt d in prep). Methods Independent Variables Knowledge A team that included employees of Proyecto de Fomento (PROFO) Cerqueros, students from Universidad Austral, and myself interviewed 2 79 fishers from 11 fishing syndicates for the first F isher C ensus of fishers from the Federacin Provincial de Pescadores Artesanales del Sur (FIPASUR). I identified 21 variables to test whether they explained success (Table 5 1). To measure local knowledge fishers were asked to select the b enthic invertebrates (BENT) and fish (FISH) they harvested from a picture list of 43 resources. Fishers then reported how the proportion of their time fishing distributed across working in the management area (MAT), open ocean (OCT), and rivers ( RIVT). Fis hers reported on the number of other fisheries (purse seine, diving, and fishing) they pursued (FAC) and the number other livelihood activities that they pursued (OAC) that included: tourism, raising domestic animals, agriculture, forestry, and commercialization
126 of products. Fishers also reported how many years they had fished (YEAR), and how long they studied in school (EDU). Environment Thirty locos were collected by fishers in each of the 11 management areas; locos were sampled from the north, south and mid sections of each management area from July to August during the 2004 harvesting season. The percentage of parasites (PARA) -phoronids and polychaetes -were recorded with assistance of marine biology students from Universidad Austral. To measure pa rasites, first I generated thirty randomly selected points on the shell, then each point was assessed for presence or absence of a parasite (a polychaete, for example), finally, I created a proportional relationship. If three out of 30 points were covered with polychaetes, then 10% of the shell was covered. The percent cover analysis was repeated for phoronids. Then both percentages were summed to create PARA; a simple summation was possible because the sampling point cannot have both a polychaete and phor onid. Loco shell length and meat weight (WEIGH) was also recorded. SeaWiFS satellite image s from April 9, 2003 and May 1999 were used to calculate change (1999 to 2003) in chlorophyll a in the water. These dates were selected because the images were cloud free, upwelling does not usually occur during this time period, and chlorophyll a values are likely related to landscape change rather than upwelling (Moreno et al. 1998). C hlorophyll a concentration values were extracted from SeaWiFS satellite images using SeaDAS 5.3 (Baith et al. 2002). The NASA generated chlorophyll a concentration values are derived from an empirically based bio optical algorithm
127 (OReilly et al., 2000) that is performed on reflectance values, which are derived from calibrated digital counts that have been atmospherically corrected (Hu et al. 2009). Chlorophyll a concentration is one of the best predictors of photosynthetic biomass available today (Huot et al. 2007), although there are other photosynthetic pigments and biomass that supports photosynthesis that are not related to chlorophyll a. One pixel was selected to represent chlorophyll a concentration in each management area (PHOT); if fishers harvested in more than one management area, the values from the management areas were aver aged. Informants reported how far they traveled from home to the fishing headquarters (DISHQ ) which is the area where they keep their boats. I calculated area of the management area (AREMA) by digitizing management areas using published coordinates from the Diario Oficial of Chile and calculating the distance from the fishing headquarters to the center of the management area (DISMA) (ArcGIS 9.0). Distances were averaged if two or more management areas were present. Two or more management areas were usually located next to each other making this calculation simple. I also measured the distance (km) from the fishing headquarters to Valdivia, the nearest city (ROADV). Landscape change, and specifically plantation land cover, was measured from 1985 to 2001 in four watersheds near the management areas (PLANT) (Van Holt a in prep). The influence of plantations in the nearshore environment was not related to the four separate watersheds, but rather across two distinct zones; the coastal region near watershed 4 was the first zone where few plantations were established, and the second
128 zone was the coastal area n ear watersheds 1 3, where extensive landscape change occurred. Other variables chlorophyll a, loco parasites, and the distance from the fishing headquarters to the Valdivian marketplace were correlated to landscape change and had more refined measurements (not only two zones) and therefore those variables were used in the final analysis. To measure available technology (BOAT) I totaled the number of boats that registered with Servicio Nacional de Pesca for each syndicate and verified the registration recor ds with fishers. Each fisher in the same fishing group, therefore, had the same value for boat availability. This syndicate level variable BOAT best represented how many boats were available to harvest resources because fishers share boats and not owning a boat doesnt necessarily signify that you dont actively dive or fish. Personal variables Age (AGE) was measured by direct interviews. Dependent Variables I tested, five separate dependent variables (1) the total catch of locos by the syndicate in the 2004 harvest season (TOTAL), (2) the number of locos harvested per day by the syndicate for the 2004 loco harvest (Catch per Unit Effort, or CPUE), (3) the price each fisher reported receiving for one kilogram of loco (4) the price each fisher reported re ceiving for one kg of congrio, and (5) the total monthly income each fisher reported from the management area (MAI). Since fishers provided the price for their top five resources harvested, not all informants were included in the individual scale analysi s [ loco (N=108) and congrio (N=74)]. The individual level data were obtained through
129 personal interviews and the syndicate level data were obtained from the Chilean fisheries agency, the SERNAPESCA. The management area income was a rank order estimate of inc ome generated from the loco fishery but in a few cases includes a small fraction of the income from the lapa fishery as well. Analysis For the syndicate level analysis (catch per unit effort and total catch), I ran a stepwise regression. Since the degrees of freedom were low, I selected specific variables to use and had to create separate environmental and knowledge technology models. The environmental model included loco parasites, chlorophyll a concentration in the management area, and loco weight. The knowledge technology model included time spent in the ocean, in the management areas, and the river, as well as total number of boats. For the individual fisher analysis, ( loco congrio, and management area income), I first tested for and removed the following multicollinear independent variables : PLANT, WEIGH, YEAR, and DISTMA because they were not found to be important by preliminary analysis. I then ran a stepwise multiple regression for all remaining independent variables with the dependent variable for each analysis. I compared all models using, in this order, Mallows' Cp statistic, the Bayesian information criterion (BIC), a scree plot of mean square error (MSE) vs. model size, and R2 (Mallows 1973; Spiegelhalter et al. 2002). In the final model, the condition index, tolerance values, and variation inflation index were used to check if any multicollinearity issues remained. I plotted the residuals from the regression and the dependent variables to check for
130 normality of the error term. I also tested fo r interactions between EDU, YEAR and AGE, and between DISTMA and PHOT. Including interactions did not improve the models. To understand general trends for the fisheries, I then plugged in average independent variable numbers in the regressionmodel equati ons to solve for the dependent variable. I used the averages for all fishers for the following variables: FISH, MAT, RIVT, YEARS, and EDU (Table 52). Fisher syndicate averages were used for PARA, PLANT, PHOT, DISTMA, and ROADV variables since the syndicat e level average was used in the individual analysis for these variables (Table 5 3). An exchange rate of $1USD = 600 pesos was used. To confirm the findings with an alternative analysis, I ran a s tepwise discriminant analysis ( PROC STEPDISC in SAS 2001) o n groups of successful and unsuccessful individuals. I divided informants into two groups: those with median prices and above for their catches were the successful group and those with prices below the median were the not successful group. I split the data using mode, median, and 50 percentile; median best split the data and was used to generate the two groups. In the discriminant analysis, independent variables enter ed and remained in the model if p value Results Closed F isheries A larger fraction of time spent working in the rivers, management areas, or the ocean increased total catch of locos for the syndicate (Table 5 4). Environmental variables however, were n ot statistically significant in explaining total catch. The number of locos harvested per day was positively influenced by the proportion of time
131 spent fishing in the river negatively influenced by the amount of time working in the open ocean; and positively influenced if fishers had more boats The only environmental factor that influenced the number of locos harvested per day was increased chlorophyll a concentration in the management area (Table 5 5). The importance of knowledge in total catch and catch per unit effort agrees with other studies (Acheson 1977; Palsson 1994; Palsson & Durrenberger 1982) Chlorophyll a concentration also explains a large fraction of the variation in catchper unit effort, but knowledge explain a much larger fraction of catch per unit effort. Time spent working in the river accurately reflects diving experience since it was so influential in determining catch per unit effort. For locos parasites alone explained 43% of the variance in price and chlorophyll a concentration explained another 21% (Tables 5 6 and 57; Figure 5 2). Lack of boats explained an additional 5% in the model. Knowledge explained only 6% of the price variance--i f fishers were experienced, harvested multi ple fishes; however they did not spend a lot of time in the open ocean, and had specific experience in rivers, they reported higher prices for locos The resource quality as measured by parasites, which is influenced by environmental change associated with eutrophication, overwhelmingly predicted price. I used the loco regression model to predict price changes using $7.20 as the average pri ce for a kilo of loco (Table 5 6) A ten percent increase in parasites had a dramatic affect and decreased loco price per kilo by $0.99 or 17%. A 0.2 mg/m3 increase in chlorophyll a concentration in the management area increased price per kilo by $0.28 Chlorophyll a concentration and parasites are related. If the chlorophyll a concentration
132 is higher than 2.4 mg m3, th e benefits of the chlorophyll a may diminish and the parasites influence dominates (Figure 5 3) To adapt to environmental changes, either environmental problems are mitigated or fishers use their knowledge or technology to adapt If the fisher group had one boat less the price of loco would increase $0.08. To completely compensate for a 10% increase in parasites, fishers would need to have fewer than four boats which is an unrealistic scenario A fishers knowledge is the most likely the way that fishe rs can adapt to environmental change since boats negatively influence the price of locos however price gains from knowledge were not as substantial as the environmental effects of increased parasites. The loco price increases $0.05 for each additional fis hspecies harvested by an individual; spending 10% more time working in river based fisheries adds $0.09; and spending 10% less time working in the open ocean fisheries adds $0.11 per kg. F ishers who work in rivers gain an advantage (and those that spend m ore time in the open ocean do not) with the loco harvest because fishers who work in the rivers have more diving experience. W orking in additional offshore fisheries also adds to the price individuals are paid for loco s The discriminant analysis of successful (> median price) vs. nonsuccessful (< median price) loco fishers confirmed that fishers with few parasites on their locos were more successful (Table 5 8). Weight of the loco meat best explained the successful groups of fishers in the discriminant a nalysis since fishers are paid higher prices/kg for heavier locos Parasites were the most important success factor in both the regression and discriminant analyses. Chlorophyll a concentration continued to be important
133 successful fishers had management a reas with high chlorophyll a concentration and low parasite loads, which confirms the regression analysis and illustrates the complex relationship between parasites and chlorophyll a. Also fishers had an advantage if they lived further away from the Valdivian market probably because of the associated environmental changes close to Valdivia. Fishers with more years experience had a slight advantage. The discriminant analysis model was highly accurate for successful (94.34%) and unsuccessful fishers (83.05%) (Table 5 9 ). The monthly income that fishers rec eive from the management areas ( consist ing of mainly loco resources ) shared similar findings with the loco price data (Table 5 6; Table 57). Management area income was dependent on the number of benthic fis heries harvested ( working in multiple benthic fisheries was better) ; weight of the loco resources (the heavier the loco the better); the size of the management area (larger areas were better); and how close a fisher lived to the fishing headquarters (clos er was better); chlorophyll a concentration negatively influenced income. The discriminant analysis has similar results ; successful fishers had locos with a low number of parasites (associated with higher weight), larger management areas, more benthic spe cies and fewer boats F ishers who worked in multiple fishing activities (diver, purse seine fishery, and the smallscale open fishery) had a n advantage, but the mean difference in successful and unsuccessful fishers was very small. Also distance to the hea dquarters had the opposite relationship when compared with the multiple regression analysis ( being located further away from the headquarters was better in the discriminant analysis ), but only a small fraction of success was explained by distance to headqu arters
134 Open Access Congrio The multiple regressions show that chlorophyll a concentration explain ed 13% of the price variation in congrio. Fisher knowledge and experience explained 16% of the variance and formal education explained 11% of the variance ( Tables 5 6 and 57; Figure 52). Fishers who worked in multiple fin fisheries, had spent many years fishing and those who did not spend time working in the rivers were most successful. More formal schooling also benefited price. The congrio model (Table 56) (usi ng a base price of $1.45) shows that knowledge was most influential in price adding $0.02 per kg for each additional species fished; $0.05 per kg for each additional year spent in school, $0.01 for each additional year fishing and a loss of $0.05 per kg for each additional 10% of a fishers time that is dedicated to fishing in the river. An additional 0.2 mg/m3 increase in chlorophyll a concentration in the management area added $0.03 per kg. Knowledge and experience explain a larger fraction of success when compared with the environment. The discriminant analysis (Table 5 8) confirmed that successful fishers dedicated themselves to artisanal fishing and included fewer activities outside of the fisheries and worked in fewer different types of fis heries i.e. they did not work in purse seine, benthic fisheries, and offshore small scale fisheries. These fishers also had more boat power The m ost influential factor however, was how close the fi shers were to the market. Interestingly fishers located n ear the market were probably located in plantation watersheds had a degraded management area, and found a new niche by using their knowledge and technology to work in fisheries that were more resilient to environmental
135 change. Successful fishers had a sli ghtly higher formal education that probably helped them negotiate higher prices for their resources Smaller management areas benefitted the fishers, perhaps because they require less work to monitor and care for year round. The model predicted successful fishers with 81.25% and unsuccessful fishers with 88.57% accuracy (Table 5 9). Adapting to C hange Fishers cannot easily adapt to be successful if their management areas contain locos with high parasite loads While fishers can reduce fleet size and increase knowledge, these factors do not mitigate the overwhelming influence of parasites. While boats are important in explaining the catch per unit effort having more boats does not translate to higher prices. For success in these nearshore fisheries where ch anges in chlorophyll a concentration and parasites affect the locos conservation efforts must address the influences of eutrophication in the system Fishers, especially those who are new to the management system, could be vulnerable if their resources become affected by eutrophication. New fishers are not rewarded for having more boats and they are only slightly rewarded for focused diving skills so these new fishers are not developing the skills that will help them adapt to environmental change The congrio fishery demonstrates how fishers may adapt to e nvironmental change. Th e fishers who were successful in the congrio fishery had extensive fishing knowledge, technology, and they were more connected to the market. The offshore fisheries may be a means th at fishers can adapt to the environmental change in the management areas. To help fishers be less
136 vulnerable to environmental changes associated with eutrophication, fishers should continue to develop their offshore fishing skills and improve technology. I n the management areas with higher concentration of chlorophyll a fishers spent more time working offshore; they harvested more fishes (r=012, p=0.05), and fewer benthic resources (r=0.38, p=0.00). These fisher s also spent a lower proportion of their time diving in the rivers (r= 0.42, p=0.00) and a higher proportion of time in the open ocean (r=0.15, p=0.02). T he fishers still spend a large fraction of their time working in the management areas (r=0.16, p=0.01), possibly because the economic benefits are highest in the management areas In the open access congrio fishery, knowledge and technology was more influential and could help fishers adapt to environmental change. Knowledge alone can help fishers adapt to change in nearshore environmental conditions, specifically, changes in chlorophyll a concentration and parasite loads in the benthic fisheries Boats also helped fishers adapt to environmental change ( discriminant analysis). These findings agree with other studies on fisher success that explain th at knowledge and technology help fishers succeed. To foster success in nearshore fisheries, environmental condition must be improved. Fishers mitigate poor environmental conditions in closed access areas nearshore by harvesting offshore. Ecological knowle dge and boats help make offshore fishers successful. A recent fishing net diversification program supported by the Chilean Government in the Valdivia Province is an example of a program that could foster adaptability to environmental change. Fishers were g iven different sized fishing nets to
137 diversify catches. Based on our analysis, fishers working in the offshore fisheries would benefit and those working in nearshore fisheries gain a smaller benefit from such a program. Becoming skilled in diverse fishing strategies however, could prove very useful if the environmental condition in the he althy management areas degrades, because the fishers trained in using other gear types will be able to harvest other open access resources that are not as influenced by th e immediate, nearshore environmental effects. C onclusions Environmental features best explained price differences for the benthic, closed access, loco fisheries because these organisms are relatively stationary compared to fin fish and are directly influen ced by local environmental conditions. Knowledge and technology, which help fishers catch more fish and perhaps find the better resources, dont compensate for environmental change in these fisheries. Those fishers in management areas with poor quality res ources are vulnerable to changes in the environment because the traditional means by how fishers adapt to change (knowledge and technology) does not foster success. Local knowledge explained a smaller fraction of success for the loco fishery and a larger f raction of success for the congrio fishery. Diving experience helped loco fishers succeed and offshore fishery experience helped the congrio fishers succeed. Only in the openaccess fisheries w ere fisher knowledge powerful enough and the influence of the e nvironment low enough that fishers could adapt to change s in the environment
138 Table 5 1. Dependent and independent variables tested in this study. Type Variable Measure Method Code Independent variables Knowledge Benthic species harvested # of benth ic species harvested interview BEN Fish harvested # of species of fish harvested interview FISH Management area time Proportion of time spent in MA interview MAT Open ocean time Proportion of time spent in ocean interview OCT River time Proportion of time spent in river interview RIVT Livelihood # of activities outside the fishery interview OAC Livelihood # of fishing activities (purse seine, diving, fishing) interview FAC Fisher experience # of years fishing interview YEAR Years of educatio n Years of education Interview EDU Environment Carnivore Health % Loco shell w/ parasites* ecological survey PARA Carnivore Health Weight loco meat (g) ecological survey WEIGH Carnivore Health Length loco meat (cm) ecological survey LENGTH Chlorophy ll a Change in chlorophyll a (mg/m 3 ) in management area (1999 2003) satellite image PHOT Distance to market Distance (km) from fishing headquarters to Valdivia city GIS ROADV Conversion to plantations % new plantation in watershed satellite image PLAN T Distance to headquarters Distance to fisher headquarters interview DISTHQ Land owned Area of land owned (m 2 ) interview LAND Distance to MA Distance to management area (m) GIS DISTMA Size MA Size (ha) of management area GIS AREAMA Boats availabi lity Boats (#) interview BOAT Personal Fisher age Age Interview AGE Dependent Variables Economy Price of loco Price per kilo (pesos/kg) Interview loco Price of congrio Price per kilo (pesos/kg) Interview congrio Management area income Rank order of income per month in management area Interview MAI Catch per unit effort locos # locos / day per syndicate in 2004 SERNAPESCA CPUE Total catch # locos caught in the 2004 season per syndicate SERNAPESCA TOTAL
139 Table 52. Descriptive statistics for variables used for the loco congrio, and management area income analysis where fisher is the unit of analysis Variable N Mean Std Dev Minimum Maximum PLANT (km 2 ) 279 896.0 566.5 27.1 1264.0 ROADV (km) 279 68.4 45.9 17.9 152.9 PHOT (mg/m 3 ) 279 2.1 0.5 1. 2 2.7 PARA (%) 255 19.2 15.5 3.7 42.6 WEIGHT (g) 255 112.1 16.2 88.4 134.8 BOAT (#) 279 17.8 6.8 7.0 28.0 EDU (y) 277 9.0 3.1 0.1 17.0 YEAR (y) 244 24.7 12.1 1.0 63.0 DISTHQ (m) 278 1622.1 3057.9 0.0 25000.0 AREA AM (ha) 279 126.1 65.4 49.0 352.0 MA T (%) 279 35.7 25.7 0.0 100.0 OC T (%) 279 49.7 31.4 0.0 100.0 RIV T (%) 279 18.0 27.8 0.0 100.0 FISH (#) 279 7.8 5.9 0.0 23.0 BEN (#) 279 5.8 3.6 0.0 14.0 Loco ($/kg) 142 4787.0 863.8 3000.0 6000.0 Congrio ($/kg) 92 947.3 172.0 400.0 1300.0 MAI (rank order) 279 1.4 0.9 0.0 4.0 TOTAL (#) 279 64659.3 60331.2 3723.0 262703.0 CPUE (catch/day) 279 16395.5 11206.7 3016.0 39633.0 Note for MAI, monthly management area income (UD$/month): 1: <$50 2: $50 $99 3: $100 $149 4: $150 $200
140 Table 53. Descriptive statistics for variables us ed for the loco congrio, and management area income analysis where fishing syndicate is the unit of analysis. Variable N Mean Std Dev Minimum Maximum PLANT (km 2 ) 11 926.7 577.8 27.1 1264.0 ROADV (km) 11 70.6 48.2 1 7.9 152.9 PHOT (mg/m 3 ) 11 2.3 0.5 1.2 2.7 PARA (%) 11 21.5 16.4 3.7 42.6 WEIGHT (g) 11 114.1 16.2 88.4 134.8 BOAT (#) 11 16.2 6.6 7.0 26.0 EDU (y) 11 8.9 0.9 7.2 10.2 YEAR (y) 11 24.6 3.7 18.5 31.1 DISTHQ (m) 11 1806.8 1388.4 353.1 4200.0 AREA AM ( ha) 11 134.9 78.6 70.0 352.0 MAT (%) 11 36.3 14.2 17.8 62.0 OC T (%) 11 51.2 17.2 27.3 79.9 RIV T (%) 11 15.9 18.0 0.0 52.8 FISH (#) 11 8.4 3.2 4.9 16.4 BEN (#) 11 5.5 1.8 2.7 9.1 Loco ($/kg) 11 4362.9 953.5 3000.0 5800.0 Congrio ($/kg) 11 922.7 133.3 700.0 1133.3 MAI (rank order) 11 1.4 0.5 0.4 2.1 TOTAL (#) 11 71261.6 73860.1 3723.0 262703.0 CPUE (#/day) 11 13774.4 10480.1 3016.0 39633.0
141 Table 5 4. Regression models of total locos caught in the 2003 season (TOTAL) and catch per unit effort (CPUE) for knowledge models. Independent variables used in the analysis shown in column 1. Note: ** is highly significant where p<0.01, is significant where p<0.05, and + is slightly significant where p<0.1. Dependent Variables Total Catch Knowledge ( B ) # of locos Total Catch Knowledge (F) CPUE Knowledge ( B ) Locos /day CPUE Knowledge (F) R 2 0.82 0.88 Intercept 507267** 16.05 19250* 11.63 MAT 6660** 2209 OCT 4963** 14.55 335* 10.07 RIVT 5354** 27.22 211+ 4.55 BOAT 612* 7.51 N 11 11 Unit o f analysis Syndicate Syndicate Table 5 5. Regression models of catch per unit effort (CPUE) for environmental models. Independent variables used in the analysis shown in column 1. Note: ** is highly significant where p<0.01, is significant where p<0.05, and + is slightly significant where p<0.1. Dependent Variables CPUE Environment ( B ) Locos /day CPUE Environment (F) R 2 0.50 Intercept 50339** 16.68 PARA P HOT 16212* 9.13 WEIGH N 11 unit of analysis Syndicate
142 Table 5 6. Regression models of total locos caught in the 2003 season (total catch) and catch per unit effort (CPU). Independent variables used in the analysis shown in column 1. Dependent Variables Loco ( B ) pesos/kg Loco (F) Congrio ( B ) pesos/kg Congrio (F) Management Area Income ( B ) Units Management Area Income (F) Units R 2 0.7 5 0.45 0.35 I ntercept 4518.44** 273.73 0.00 0.08 1.63 11.39 Knowledge BEN 0.05 8.12 FISH 32.41** 12.34 11.39** 14.3 MAT OCT 6.51** 12.17 RIVT 5.17* 6.25 3.10** 7.37 YEARS 3.7* 5.29 EDU 30.60** 20.54 Environment PARA 59.32** 251.78 PHOT 801.30** 51.65 220.16** 19.18 0.62 28.41 WEIGHT 0.03 55.90 DISTHQ 0.00 20.97 AREAMA 0.00 14.46 BOAT 46.55** 36.91 N 109 75 218 Unit o f analysis Fisher Fisher Fisher
143 Table 5 7. The best models for each number of variables with R2, Adjusted R2, C(p), AIC, BIC, and the explanatory variables. Percent variation explained by each factor was calculated by comparing R2 values. Variabl es R 2 Adj.R 2 C(p) AIC BIC Variables in Model Loco Price 1 0.43 0.43 149.72 1413.80 1412.85 PARA 2 0.64 0.64 58.24 1365.95 1365.80 PHOT PARA 3 0.70 0.69 35.27 1349.61 1349.93 PHOT PARA BOAT 4 0.71 0.70 30.65 1346.18 1346.54 PHOT PARA BOAT RIVT 5 0.74 0.72 21.57 1338.44 1339.46 PHOT PARA BOAT OCT FISH 6 0.75 0.74 16.72 1333.96 1335.62 PHOT PARA BOAT OCT RIVT FISH 7 0.76 0.75 14.27 1331.53 1333.79 PHOT PARA BOAT YEARS OCT RIVT FISH 8 0.77 0.75 13.19 1330.36 1333.18 PHOT PARA BOAT AGE YEARS OCT RIV T FISH Congrio Price 1 0.14 0.13 35.35 777.55 778.22 PHOT 2 0.23 0.20 26.38 771.50 772.14 PHOT FISH 3 0.34 0.31 14.12 761.72 762.72 PHOT EDU FISH 4 0.40 0.36 9.10 756.54 758.66 PHOT EDU RIVT FISH 5 0.45 0.40 6.00 753.20 756.21 PHOT EDU YE ARS RIVT FISH M onthly 1 0.08 0.07 87.0 69.50 68.75 DISTHQ M anagement 2 0.16 0.15 61.74 88.23 87.53 PHOT WEIGH Area Income 3 0.23 0.22 40.78 105.33 104.41 PHOT WEI GH DISTHQ 4 0.29 0.27 24.81 119.53 118.18 PHOT WEIGH DISTHQ AREAMA 5 0.33 0.31 12.13 131.74 129.74 PHOT WEIGH DISTHQ AREAMA BEN Total Catch 1 0.27 0.19 22.10 246.15 244.89 RIVT (know) 2 0.42 0.32 16.55 244.85 243.43 MAT RIVT 3 0.82 0.75 4.00 234.48 240.40 OCT MAT RIVT CPU Know 1 0.70 0.67 11.20 19 2.14 192.24 RIVT 2 0.81 0.76 6.55 189.14 190.85 BOAT OCT 3 0.88 0.83 4.00 185.63 191.54 BOAT OCT RIVT CPU Env 1 0.50 0.45 0.98 198.91 202.13 PHOT
144 Table 5 8. Partial r square, significance levels, and average squared canonical correlation of stepwise discriminant analysis of loco characteristics. Step Variable Partial R Square Pr>F Average squared canonical correlation Mean Values Successful Mean Values Unsuccessful Loco 1 WEIGH (g) 0.43 <0.001 0.44 126 104 2 PARA (%) 0.17 <0.001 0.53 10 19 3 PHOT (mg/m 3 ) 0.15 <0.001 0.65 2 2 4 ROADV (km 2 ) 0.10 <0.001 0.68 96 49 5 BEN (#) 0.04 0.03 0.69 4 6 6 YEAR (y) 0.03 0.07 0.70 23 23 Congrio 1 ROADV (km 2 ) 0.23 <0.001 0.24 47 92 2 FAC (#) 0.05 0.05 0.28 1 2 3 AREAMA (ha) 0 .04 0.10 0.30 95 115 4 BOAT (#) 0.06 0.04 0.34 20 16 5 OAC (#) 0.07 0.02 0.39 1 2 6 PARA (%) 0.08 0.02 0.43 22 23 7 EDU (y) 0.05 0.06 0.47 9 8 Management area 1 PARA (%) 0.16 <0.001 0.16 13 25 2 AREAMA (ha) 0.07 <0.001 0.21 144 112 3 D ISTHQ (km 2 ) 0.05 <0.001 0.25 1887 1418 4 FAC (#) 0.04 <0.001 0.28 2 1 5 BOAT (#) 0.02 0.03 0.29 16 19 6 BEN (#) 0.04 0.01 0.32 6 6
145 Table 5 9. Number of observations and percent classified and error count estimates of successful and unsuccessful groups using stepwise discriminant analysis. Successful Unsuccessful Total Loco Successful 50 94 3 66 53 100 Unsuccessful 10 17 49 83.05 59 100 Total 60 54 52 46.43 112 100 Rate 0.06 0.17 0.11 Priors 0.50 0.50 Congrio Successful 39 81 9 19 48 100 Unsuccessful 4 11 31 89 35 100 Total 43 52 40 48 83 100 Rate 0.19 0.11 0.15 Priors 0.50 0.50 Management Area Successful 93 77.5 27 23 120 100. Unsuccessful 55 41.04 79 59 134 100 Total 148 58.27 106 42 254 100 Rate 0.23 0.4 1 0.32 Priors 0.50 0.50
146 Figure 5 1. Study Site in the Valdivia Province Coastal System, Chile. The management areas are located in black along the coast. In some cases two management areas are listed due to space constraints (i.e. Huape A and Huape B). Watersheds are numbered from 1 to 4. Some fisher syndicates have more than one management area. The influence of forest plantations (high chlorophyll a concentration values in the nearshore and loco shellfish with high levels of parasites) extends f rom the outlet of Rio Valdivia (the river from the Valdivian watershed, #3) northward (management areas Punta Numpulli northward).
147 Figure 5 2. Percent variance in pr ice explained by environmental and knowledge factors for the loco and congrio fisheries. 0 10 20 30 40 50 60 70 Environmental Knowledge % Variance Explained Factors loco congrio
148 Figure 5 3. The relationship between chlorophyll a and parasites on loco ( Concholepas concholepas ) shells in the management areas. If chlorophyll a increases above 2.4 mg/m3 then the % of parasites on the loco shells increases.
149 CHAPTER 6 CONCLUSIONS In region X of Chile, forest plantations represent the most recent development threat to the coastal system Native forest is substituted and marginal agricultural areas are being converted to forest plantations; fishers and farmers livelihoods are therefore changing. The Valdivia zone has changed from a syst em dominated by agriculture and to a lesser extent native forest to a region where plantations dominate P eople living in counties with forest plantations are migrating. T he smallscale farmers on the sloped mountains have likely migrated to urban areas since forestry companies buy out small scale farmers and replace their land with forest plantations. Fishers living in regions that are influenced by forest plantations are vulnerab le to the associated changes in the nutrient dynamics in the nearshore. The chlorophyll a patterns in the nearshore show that chlorophyll a concentration is influenced by landscape change The Valdivia region has the highest chlorophyll a concentration in the nearshore. Th e high chlorophyll a values are the result of nutrients that enter the water and increase the presence of nitrogen limited primary producers in the system. The enriched 1 5N and 13C values of the locoshellfish tissue confirm that terre strial nitrogen is entering the nearshore that is influenced by forest plantations The nitrogen is increasing the photosynthetic biomass in the region and these new photosynthetic organisms are fixing atmospheric carbon. In contrast, the native forest in fluenced watersheds have low chlorophyll a values and depleted isotopic signatures. Landscape change and the associated changes in chlorophyll a values in the nearshore are the main factors that describe the distribution of phoronid shell borers and barna cle epibionts across loco shellfish in region X E pibionts and boring organisms are
150 abundant in areas where the nitrogen dynamics in the nearshore have shift ed because of plantation development Fishers have been adapting to the environmental changes in t he region. Environmental features best explained price differences for the benthic, closed access, loco fisheries Knowledge and technology, which help fishers catch more fish and perhaps find the better resources, dont compensate for environmental change in these fisheries. Those fishers in management areas with poor quality resources are vulnerable to changes in the environment because the traditional means by how fishers adapt to change (knowledge and technology) does not foster success. Local knowledge explained a smaller fraction of success for the loco fishery and a larger fraction of success for the congrio fishery. Diving experience helped loco fishers succeed and offshore fishery experience helped the congrio fishers succeed. Only in the openacces s fisheries w ere fisher knowledge powerful enough and the influence of the environment low enough that fishers could adapt to change s in the environment We need to continue to understand the terrestrial influences on the nearshore ecosystem beyond the traditional scope. Companies and governments need to proactively understand the potential impacts of plantations at landscape and seascape levels. Economic models that incorporate environmental variables need to characterize other environmental costs not pre viously considered or seen. For example, if landscape change and the associated changes in the chlorophyll a influence the growth of the loco shellfish in Valdivia by 20 grams, these 20 grams trans late into 2,000 kg for fishers given an average harvest of 100,000 locos during the year. If of six locos sell per kilo, and locos cost $1 per loco, a fisher group may lose $12,000 per year, and each fisher could lose
151 $400, the equivalent of about two months salary. Over half of the worlds population lives in co astal systems (Cohen et al. 1997) and extensive development occurs along the coast; the livelihoods of the worlds 13.1 million artisanal fishers (FAO 2002) depend on our continued understanding of the nearshore environment.
152 LIST OF REFERENCES Acheson, J. M. 1977. Technical Skills and fishing success in the Maine Lobster Industry. Pages 111138 in H. Lechtman and R. Merrill, editors. Material Culture Styles: Organization and Dynamics of Technology. West Publications, St. Paul, MN. Alexander, R. B., R. A. Smith, and G. E. Schwarz. 2000. Effec t of stream channel size on the delivery of nitrogen to the Gulf of Mexico. Nature. 403:758761. Anderson, D. M., P. M., Glibert, and J. M. Burkholder. 2002. Harmful algal blooms and eutrophication: Nutrient sources, composition, and consequences. Estuaries 25: 704726. Auty, R. M. 1993. Sustaining Development in Mineral Economies: the resource curse thesis. Routledge London. Bade, D.L., M. L. Pace, J.J. Cole, and S. R. Carpenter. 2005. Can algal photos ynthetic inorganic carbon isotope fractionation be predicted in lakes using existing models? Aquatic Science 68: 142153. Baith, K., R. Lindsay, G. Fu, and C.R. McClain, 2001: SeaDAS, a data analysis system for ocean color satellite sensors Eos, Tr ansactions, American Geophysical Union 82: 202. Barlow, K. and C. Cocklin. 2003. Reconstructing rurality and community: plantation forestry in Victoria, Aust ralia. Journal of Rural Studies 19: 503519. Barr, C. and C. Cossalter. 2004. Chinas development of a plantation based wood pulp industry: government policies, financial incentives, and investment trends International Forestry Review 6: 267281. Beck, M W. Heck Jr., K. L., Able, K.W., Childers, D.L, Eggleston, D. B., Gillanders, B. M., Halpern, B., Hays, C. G., Hoshino, K., Minello, T. J., Orth, R. J., Sheridan, P. F., and M. P. Weinstein. 2001. The Identification, Conservation, and Management of Estuarine and Marine Nurseries for Fish and Invertebrates. BioScience 51 ( 8): 633641. Belan, T. A. 2003. Benthos abundance pattern and species composition in conditions of pollution in Amursky Bay (the Peter the Great Bay, the Sea of Japan) Marine Pollution Bulletin 46:11111119.
153 Bernal, P. A., D. Oliva, B. Aliaga, and C. Morales 1999. New regulations in Chilean fisheries and Aquaculture: ITQ's and Territorial Users Rights. Ocean and Coastal Management 42:119142. Bers A.V., and M. Wahl. 2004. The influence of natural surface m icrotopographies on fouling Biofouling. 20 (1):4 351. Betts, R. A., P. D. Falloon, K. K. Goldewijk, and N. Ramankutty. 2007. Biogeophysical effects of land use on climate: Model simulations of radiative forcing and large scale temperature change. Agricultu ral and Forest Meteorology 142: 216233. Beukema, J.J. 1991. Changes in composition of bottom fauna of a tidal flat area during a period of eutrophication. Marine Biology 111:293301 Binford M.W ., H. L. Gholz G. Starr and T.A. Martin. 2006. Regional carbon dynamics in the southeastern U.S. coastal plain: Balancing land cover type, timber harvesting, fire, and environmental variation. Journal of Geophysical Research 111 (24) : D24S92. Bini, L. M., S, M. Thomaz, K. J. Murphy, and A. F. M. Camargo. 1999. Aquatic macrophyte distribution in relation to water and sediment conditions in the Itaipu Reservoir, Brazil. Hydrobiologia 415:147154. Binkley, D., and S. C. Resh. 1999. Rapid changes in soils following eucalyptus afforestation in Hawaii. Soil Science Society of America Journal 63: 222225. Blaschke, P. M., N. A. Trustrum, N. A., and R. C. Derose. 1992. Ecosystem Processes and Sustainable Land Use in New Zealand S teeplands. Agriculture Ecosystems & Environment 41: 153178. Bonsdorff E ., E.M. Blomqvist J. Mattila and A. Norkko. 1997. Long term changes and coastal eutrophication. Examples from the land Islands and the Archipelago Sea, N. Baltic Sea. Oceanologica Acta 20:319329. Bounoiua, L., R. DeFries, G.J. Collatz P. Sellers, and H. Khan. 2002. E ffects of land cover conversion on surface clima te. Climatic Change. 52 (1 2) 2964.
154 Bull, G. Q., M. Bazett O. Schwab, S. Nilsson, A. White, and S. Maginnis 2006. Industrial forest plantation subsidies: i mpacts and implicatio ns. Forest Policy and Economics 9: 1331. Byrne, K. A. and R. Milne. 2006. Carbon stocks and sequestration in plantation forests in the Republic of Ireland. Forestry 79: 361369. Cabana, G. and J.B. Rasmussen. 1994. Modeling foodchain structure and contaminant bioaccumulation using stable nitrogen isotopes. Nature 372 (6503): 255257. Camacho Ibar, V., V. Dupra F. Wulff S. V. Smith, J. I. Marshall Crossland, and C.J. Cr ossland. 2002. Estuarine systems of the Latin American Region (Regional Workshop V) and estuarine systems of the Arctic Region: Carbon, nitrogen and phosphorus fluxes. LandOcean Interactions in the Coastal Zone) ( LOICZ) Reports & Studies No. 23, Net herlands: LOICZ International Project Office. Camu s P. A. 2001. Biogeographa of cont inental Chile. Revista Chilena d e Historia Natural 74(3) : 587617. Carnus, J.M., J. Parrotta, E. Brockerhoff M. Arbe H. Jactel A. Kreme D. Lamb K. O'Hara, and B. Walters, B. 2006. Planted forests and bi odiversity. Journal of Forestry 104: 6577. Chang S., Steimle, F.W., Reid, R. N., Fromm, S. A., Zdanowicz, V.S., and R. A. Pikanowski. 1992. Association of benthic macrofauna with habitat types and quali ty in the New York Bight. Marine Ecology Progress Series 89:237251. Christie, S. I. and R. J. Scholes 1995. Carbon storage in eucalyptus and pine plantations in South Africa. Environmental Monitoring and As sessment 38: 231241. CIREN (Centro de I nformacion de recursos naturales) and CORFO (Corporacion de Fomento). 1983. Materiales y simbolos utilizados por el proyecto aerofotogrametrico CHILE/OEA/BID. 3a Edicion. 53 paginas. Clapp, R. A. 1995. Creating Competitive Advantage Forest Policy a s Industrial Policy in
155 Chile. Economic Geography 71:273296. Clapp, R. A. 1998. Regions of refuge and the agrarian question: Peasant agriculture and plantation forestry Chile an araucania. World Development 26: 571589. Clapp, R. A. 2001. Tree farmin g and forest conservation in Chile: Do replacement forests leave any originals behind? Society & Natural Resources 14: 341356. Cohen, J. E., C. Small, A. Mellinger, J. Gallup, and J. Sachs. 1997. Estimates of coastal populations. Science 278 :12111212. Cole, M. L., I. Valielia, K.D. Kroeger G.L. Tomasky J. Cebrian C. Wigand, R. A. McKinney, S. P. Grady, and M. H Carvalho de Silva C. 2004. Assessment of delta 15N isotopic method to indicate anthropogenic eutrophication in aquatic ecosystems. Jo urnal of Environmental Quality 33 :124132. Coleman, D.C., and B. Fry., editors, 1991. Carbon Isotope Techniques (Isotopic Techniques in Plant, Soil, and Aquatic Biology Series, Vol 1. Academic Press San Diego, CA. 274. P p. CONAF, CONAMA, and P B irf. 1997. Catastro y evaluacin de los recursos vegetacionales nativos de Chile. CONAF, Santaigo, Chile. Crema, R., Castellia, A., and D. Prevedelli. 1991. Long term eutrophication effects on macrofaunal communities in the northern Adriatic Sea. M arine Pollution Bulletin 22(10): 503508 Dauby, P., A. Khomsi, and J.M. Bouquegneau. 1998. Trophic relationships within intertidal communities of the Brittany Coasts: a stable carbon isotope analysis. 14(4):12021212. Defeo O., and J.C. Castilla. 2005. More than one bag for the world fishery crisis and keys for comanagement successes in selected artisanal Latin American shellfisheries. Reviews in Fish Biology and Fisheries 15(3): 265283. DeNiro,M.J. and S. Epstien. 1978. Influence of diet on the distribution of carbon isotopes in animals. Geochimica et Cosmochimica Acta 42(5): 495 506.
156 DGA (Dirrecion General de Aguas). 1992. Analysis redes de vigilancia caldidad agues terrestres estadisticas hidroquimica nacional. Decima Region. Rep ort to DGA. Diaz, R. J. and R. Rosenberg. 1995. Marine benthic hypoxia: a review of its ecological effects and the behavioral responses of benthic macrofauna. Oceanography and Marine Biology 33:245303. Diaz, R. J. and R. Rosenberg. 2008. Spreadin g dead zones and consequences for marine ecosystems. Science 321 (5891): 926929. Dougherty, J. E., and M. D. Morgan. 1991. Benthic community response (primarily Chironomidae) to nutrient enrichment and alkalinization in shallow, soft water humic lakes. Hydrobiologia 215 : 7382. Dros, J. M. 2004. Managing the soy boom: two scenarios of soy production expansion in South America. Aideenvironment, Amsterdam, The Netherlands. Eldridge, W. D., and N. V. Pacheco. 1987. Activity patterns and habitat utilization of pudus ( Pudu pudua) in SouthCentral Chile. Biology and Management of the Cervidae. Ewers, R. M., A. D. Kliskey, S. Walker, D. Rutledge J. S. Harding, and R. K. Didham 2006. Past and future trajectories of forest loss in New Ze aland. B iological Conservation 133: 312325. Eyre, B. D. 2000. Regional evaluation of nutrient transformation and phytoplankton growth in nine river dominated sub tropical east Australian estuaries. Marine Ecology Progress Series 205: 6183. Fahey, B., and R. Jackson. 1997. Hydrological impacts of converting native forests and grasslands to pine plantations, South Island, New Zealand. Agricultural and Forest Meteorology 84: 6982. FAO (Food and Agriculture Organization). 2002. The State of World Fis heries and Aquaculture. FAO, Rome, Italy.
157 Farley, K. A., E. G. Jobbagy, and R. B. Jackson. 2005. Effects of afforestation on water yield: a global synthesis with implications fo r policy. Global Change Biology 11: 15651576. Farley, K. A. and E. F Kelly. 2004. Effects of afforestation of a paramo grassland on soil nutrient status. For est Ecology and Management 195: 281290. Fearnside, P. M. 1995. Global warming response options in Brazil's forest sector: Comparison of project level costs and benefits. Biomass & Bioenergy, 8: 309322. Fernandez, M., E. Jaramillo P.A. Marquet, C.A. Moreno, S.A Navarrete, F.P. Ojeda, C.R. Valdovinos and J.A. Vasquez. 2000. Diversity, dynamics and biogeography of Chilean benthic nearshore ecosystems: an overview and guidelines for conservation. Revista Chilena De Historia Natural 73:797830. Folch, J., Lees, M. and G. H. Sloane Stanley. 1957. A simple method for isolation and purification of total lipids from animal tissues. Journal of Biological C hemistry. 226:297509 Foley, J. A ., R. DeFries G. P Asner, C. Barford, G. Bonan, S. R. Carpenter, F. S. Chapin, M. T. Coe, G. C. Daily, H. K. Gibbs, J. H. Helkowski, T. Holloway, E. A. Howard, C. J. Kucharik C. Monfreda, J. A. Patz, I. C. Prentice, N. Ramankutty and P. K. Snyder 2005. Global consequences of land use. Science 309 (5734): 570574. Freyer, H.D. and A.I.M. Aly. 1974. Nitrogen15 variations in fertilizer nitr ogen. J ournal of Environmental Quality 3:405406. Fry, B. 1988. Food web structure on Georges Bank from stable C, N, and S isotopic compositions. Limnology and Oceanography 33:11821190. 13C measurements as indictors of carbon flow in marine and freshwater ecosystems. Contributions to Marine Science 27: 1347. Fry, B. 1999. Using stable isotopes to monitor watershed influences on aquatic trophodynamics. Can adian Journal of Fisheries and Aquatic Sciences. 56:21672171. Fukami, T., and D.A. Wardle. 2005. Long term ecological dynamics: reciprocal insights from
158 natural and anthropogenic gradients. Proceedings of the Royal Society B Biological Sciences 272(1577): 21052115. Gabaev, D., Taupek, N. a nd N. Kolotukhina. 2005. Specificity of life conditions for commercial invertebrates on artificial substrates in eutrophicated Amur Bay, the Sea of Japan. Russian Journal of Ecology 36(5): 336 342. Gabriel, K. R. 1971. The biplot graphical display of matrices with application to principal component analysis. Biometrika 58: 453467. Gallardo, C.S., and C. Osorio. 1978. Hiatella solida (Bivalvia, Hiatellidae) on Concholepas concholepas and other substrates The Veliger 20 (3); 274278. Geldres, E., and J.E. Schlatter. 2004. Crecimiento de las plantaciones de Eucalyptus globules sobre suelos rojo arcillosos de la provincial de Osorno, Dcima Regin. Bosque 25(1): 95101. Glenday, J. 2006. Carbon stor age and emissions offset potential in an East African tropical rainforest. Forest Ecology and Management 235: 7283. Gomes, H. R., J. I. Goes, and T. Saino. 2000. Influence of physical processes and freshwater discharge on the seasonality of phytopl ankton regime in the Bay of Bengal. Continental Shelf Research 20(2000):313330. Gonzalez, J., W. Stotz, J. Garrido, J. M. Orensanz, A. M. Parma, C. Tapia and A. Zuleta. 2006. The Chilean TURF system: How is it performing in the case of the loco fishery? Bulletin of Marine Science 78(3): 499527. Gower, J. C. and D. J. Hand, 1996. Biplots Chapman & Hall, New York. Grabowski, J.H., Peterson, C.H., Bishop, M.J., and R. Conrad. 2007. The bioeconomic feasibility of culturing triploid Crassostrea ariakensis in North Carolina. Journal of Shellfish Research 26 (2):529
159 Grau, H. R., and M. Aide. 2008. Globalization and landuse transitions in Latin America. Ecology and Society 13(2): 1623. Grigalunas, T., J. J. Opaluch, and M. F. Luo. 2001. The economic costs to fisheries from marine sediment disposal: case study of Providence, RI, USA. Ecological Economics 38:4758. Guerra, E., Herrera, M.A., and F. Drake. 2007. Profitability of fertilization at establishment of Eucalyptus globules planta tions. Agrociencia 41 (7):7 97 804. Gwynne, R. N. 1996. Direct foreign investment and nontraditional export growth in Chile: the case of the forestry sector. Bulletin of Latin American Research 15(3): 341357. Gyenge, J. E., M. E. Fernaandez and T. M. Schlichter. 2003. Water relations of ponderosa pines in Patagonia Argentina: implications for local water resources and individual growth. Trees Structure and Function 17 : 417423. Harris, M. 1979. Cultural Materialism: the struggle for the science of culture. Random House, New York. He, X.B., Z.B. Li, M. D. Hao, K. L. Tang, and F. L. Zheng 2003 Down scale analysis for water scarcity in response to soil water conservation on Loess Plateau of China. Agriculture Ecosystems & Environment 94: 355361. Hebbeln, D., Marchant, M., Freudenthal, T., and G. Wefer. 2000. Surface sediment distribution along the Chilean continental slope related to upwelling and productivity. Marine Geology 164 (2000): 119137. Houghton, R. A., D. S. Leftwowitz, and D. C. Skole. 1991. Changes in the landscape of Latin America between 1850 and 1985: I. Progressive loss of forests. Forest Ecology and Management 38:143172. Hu C. Lin, I., and C.C. Lien. 2009. Introduction to a SeaWiFS chlorophyll database and an analysis tool Pages 269276 in Liu, K.K., Atkinson, L., Quiones, R., and L. Talaue McManus in, editors. Carbon and Nutrient Fluxes in Continental Margins A Global Synthesis Series: Global Change The IGBP Series Springer, NY.
160 Huber A., A. Iroum e, and J Bathurs t. 2007. Effect of Pinus radiata plantations on water balance in Chile. Hydrological Processes 22: 142148. Huot, J, Babin, M, Bruyant, F., Grob, C., Twardowski, S., and H. Claustr e. 2007. Relationship between photosynthetic parameters and different proxies of phytoplankton biomass in the subtropical ocean. Biogeosciences 4: 853868. Husch, B. 1982. Forestry in Chile. Journal of Forestry November 735 737. INFOR (Instituto Fo restal) (2006). Estadistical Forestales Chilenas 2005. Boletin Estadistico. No. 111. Instituto Forestal, Santiago, Chile. INE (Insituto N acional de Estadsticas). 1994. Censo 1994. Insituto Nacional de Estadsticas, Santiago, Chile. INE (Insitut o Nacional de Estadsticas) 1997. Censo Agropecuario 1997. Insituto Nacional de Estadsticas, Santiago, Chile. INE (Insit uto Nacional de Estadsticas). 2002. Censo 2002. Insituto Nacional de Estadsticas, Santiago, Chile. Iroume, A., O. Mayen, and A. Huber. 2006. Runoff and peak flow responses to timber harvest and forest age in southern Chile. Hydrological Processes 20:3750. ITT. 2008. ENVI 4.5/ IDL Version 7.0 Users Guide: ITT Visual Information Solutions, Boulder, Colorado. Johnson P. T. J. J. M. Chase, K. L. Dosch, R. B. Hartson, J. A. Gross, D. J. Larson, D. R. Sutherland, and S. R. Carpenter. 2007. Aquatic eutrophication promotes pathogenic infection in amphibians. Proceedings of the National Academy of Sciences of the Unit ed States of America 104(40):15781 -15786. Kauth R.J. and G.S. Thomas. 1976. The tasseled Cap -a graphic description of the spectral temporal development of agricultural crops as seen by LANDSAT. Proceedings of the Symposium on Machine Processing of Remotely Sensed Data, Purdue University of West Lafayette, Indiana. Pp. 4B 41 to 4B 51.
161 Kay, C. 2006. Rural poverty and development strategies in Latin America. Journal of Agrarian Change 6(4), 455 508. Kemp W M ., W. R. Boynton, J. E. Adolf D. F. Boesch W. C. Boicourt G. Brush J. C. Cornwell T. R. Fisher P. M. Glibert, J. D. Hagy L. W. Harding E. D. Ho ude D. G. Kimmel, W. D. Miller R. I. E. Newell M. R. Roman E. M. Smith and J. C. Stevenson JC 2005. Eutrophication of Chesapeake Bay: historical trends and ecological interactions. Marine Ecology Progress Series 303:1 29. Kiirikki, M., A. Inkala, H. Kuosa H. Pitkanen, M. Kuusisto, and J. Sarkkula 2001. Evaluating the effects of nutrient load reductions on the biomass of toxic nitrogenfixing cyanobacteria in the Gulf of Finland, Baltic Sea. Boreal Environment Research 6:131146. Kreitler, C. W., S. Ragone and B.G. Katz. 1978. 15N/14N ratios of ground water nitrate, Long Island, New York. Ground Water 16:404409. Lancellotti, D. A., and J. A. Vasquez. 2000. Zoogeographa de macroinvertebrados bentnicos de la costa de Chile: contribucin para la conservacin marina. Revista Chilena De Histor ia Natural 73:99129. Lara, A., Fraver, S., Aravvena, J.C., and A. Wolodarsky Franke. 1999. Fires and the dynamics of Fitzroya cupressoides (alerce) forests of Chile Cordillera Pelada. Ecoscience 6(1):100109 Lara, A., and T. Veblen. 1993. Forest plantations in Chile: a successful model? Pages 118 139 i n A. Mather, editor Afforestation: Policies, planning, and progress. Belhaven Press, Great Bri tain Laurance, W.F., Ferriera, L.V., Rankin De Merona, J.M. and S.G. Laurance. 1998. Rain forest fragme ntation and the dynamics of Amazionian tree communities. Ecology 79(6): 20322040. Llanso, R. J. 1992. Effects of hypoxia on estuarine benthos: the lower Rappannock River (Chesapeake Bay), a case study. Estuarine, Coastal, and Shelf Science 35:491515.
162 Little, C., D. Soto, A. Lara, and J.G. Cuevas. 2008. Nitrogen exports at multiple scales in a southern Chilean watershed (Patagonian Lakes district) Biogeochemistry 87 :297309 Loh, W.Y. and Y.S. Shih. 1997. Split selection methods for classification trees Statistica Sinica 7: 815840. Losovskaya. G. V., 1988. Long terms changes in the composition and distribution of polychaetes in the north western part of the Blac k Sea. Gidrobiologicheskii Zhurnal 24(4): 2125. Ludsin, S.A., M. W. Kershner, K. A. Blocksom, R. L. Knight, and R. A. Stein. 2001. Life after death in Lake Erie: Nutrient controls drive fish species richness, rehabilitation. Ecological Applications 11:731746. Mallows, C.L. 1973. Some Comments on Cp Technometrics 15: 661675. Marcogliese, D. J. 2005. Parasites of the superorganism: Are they indicators of ecosystem health? International Journal of Parasitology 35 (7):705716. Markham, B.L ., and J.L. Barker. 1986. Landsat MSS and TM post calibration dynamic ranges, exoatmospheric reflectances and at satellite temperatures. EOSAT Technical Notes, August 1986. Matthews, H. D., A. J. Weaver, K. J. Meissner, N. P Gillett and M. Eby. 20 04. Natural and anthropogenic climate change: incorporating historical land cover change, vegetation dynamics, and the global carbon cycle. Climate Dynamics 22 (5): 461479. McDiarmid, H, R. Day, and R. Wilson. 2004. The ecology of polychaetes that infest abalone shells in Victoria, Australia. Journal of Shellfish Research 23 (4): 11791188. McGarigal, K., Chushman, S., and S. Stafford. 2000. Multivariate Statistics for Wildlife and Ecology Research. New York: Springer. 283.pp McKinney, R. A., W. G. Nelson, M. A. Charpentier, and C. Wigand. 2001. Ribbed mussel nitrogen isotope signatures reflect nitrogen sources in coastal salt marshes. Ecological
163 Applications 11(1): 203214. Mee, L. D. 1992. The Black Sea in Crisis: A Need for Concert ed International Action. Ambio 21( 4): 278286. Milliman, J. D 1991. Flux and fate of fluvial sedime nt and water in coastal seas. Pages 6989 in R.F.C. Mantoura, J.M. Martin, and R. Wollast editors Ocean Margin Processes in Global Change. John Wiley and Sons, Hoboken, New Jersey Minagawa, M. and E. Wada. 1984. Stepwise enrichment of 15N along food chains. Further evidence and the relation between 15N and animal age. Geochimica et Cosmochimica Acta 48: 11351140. Minnaert, N. 1941. The reciprocity principle in lunar photometry. Astrophysics Journal 93:403 410. Moreno, C. 2001. Informe Final: Estudio de areas potenciales de reservas marinas y parques marinos entre la VIII y X regiones. Universidad Austral, Valdivia. Moreno, C.A., G. Asencio, W. E. Duarte, and V. Marin. 1998. Settlement of the muricid Concolepas concholepas (Brugire) and its relationship with El Nino and coastal upwelli ngs in Southern Chile. Marine Ecology Progress Series 167:171175. Moreno, C. A., W. E. Duarte and J. H. Zamorano. 1979. Latitudinal changes in the number of fish species in the rocky sublittoral system An ecological approach. Arc hivos de Biologia y M edicina Experimentales 12:169178. Moreno, R.A., Neill, P.E., and N. Rozbaczylo. 2006. Native and nonindigenous boring polychaetes in Chile: a threat to native and commercial mollusc species Revista Chilena de Historia Natural. 79: 263278. Munoz A., R. Murua, and C. Jofre. 1987. Small Mammals Diversity in a Pinus Radiata Plantation in the ViiiRegion of Chile. Archivos de Biologia y Medicina Experimentales, 20: R227R227.
164 Munoz, F., Espinosa, M., Herrera, M. A., and J. Cancino. 2005. Growt h characteristics in diameter, height, and volume of a Eucalyptus nitens plantation with different silvicultural treatment for pruning and thinning. Bosque 26 (1): 9399. Nagashima, K., R. Sands A. G. D. Whyte, E. M. Bilek and N. Nakagoshi. 2001. Fo restry expansion and landuse patterns in the Nelson Region, New Zealand. Landscape Ecology 16: 719729. Nixon S W 1995. Coastal marine eutrophication: a definition, social causes, and future concerns. Ophelia 41:199219 Nosetto, M. D., E. G. Jobbagy, and J. M. Paruelo. 2006. Carbon sequestration in semi arid rangelands: Comparison of Pinus ponderosa plantations and grazing exclusion in NW Patagonia. J ournal of Arid Environments 67: 142156. Oinonen, H. 2005. Arauco Valdivia produces eucalypt us and pine pulp. Paperi Ja PuuPaper and Timber 87(6): 372 O'Reilly, J.E., S. Maritorena, D.A. Siegel, M.C. O'Brien, D.Toole, F.P. Chavez, P. Strutton, G.F. Cota, S.B. Hooker, C.R. McClain, K.L. Carder, F. Muller Karger, L. Harding, A. Magnuson, D Phinney, G.F. Moore, J. Aiken, K.R. Arrigo, R.Letelier, and M. Culver, 2000. Ocean Chlorophyll a Algorithms for SeaWiFS, OC2, and OC4: Version 4. In O'Reilly, J.E., and 24 Coauthors, SeaWiFS Postlaunch Calibration and Validation Analyses, Part 3. N ASA Tech. Memo. 2000206892, Vol. 11, 919. Oyarzun, C., C. Aracena, P. Rutherford, R. Godoy and A. Deschrijver. 2007. Effects of land use conversion from native forest to exotic plantations on nitrogen and phosphorus retention in catchments of sout hern Chile. Water Air Soil Pollution 179:341 350. Oyarzun, C. E., and L. Pena. 1995. Soil Erosion and OverlandFlow in Forested Areas with Pine Plantations at Coastal Mountain Range, Centra l Chile. Hydrological Processes 9: 111118. Palsson, G. 1994. Enskilment at Sea. Man 29:901 927. Palsson, G., and P. Durrenberger. 1982. To Dream of Fish the Causes of Icelandic Skippers
165 Fishing Success. Journal of Anthropological Research 38:227242. Parfitt, R. L., N. A. Scott D. J. Ross, G. J. Salt, a nd K. R. Tate. 2003. Land use change effects on soil C and N transformations in soils of high N status: comparisons under indigenous forest, pasture and pine plantation. Biogeochemistry 66: 203221. Perus, J. and E. Bonsdorff. 2004. Longterm changes in macrozoobenthos in the Aland archipelago, northern Baltic Sea. Journal of Sea Research 52(1): 45 56. Peterson, B. J. and B. Fry. 1987. Stable isotopes in ecosystem studies. Annual Review of Ecology and Systematics 18:293320. Posner, J ., and M. MacPherson. 1980. Agriculture on the steep slopes of tropical Ame rica. World Development 10(5): 341353. Risk M J ., Sammarco P .W ., and E.N. Edinger. 1995. Bioerosion in Acropora across the continental shelf of the Great Barrier Reef. Coral Reefs 14:7986 Rouse, J. W., R. H. Hass, J. A. Schell, and D. W. Deering. 1974. Monitoring vegetation systems in the Great Plains with ERTS. Proceedings, Third Earth Resources Technology Satellite 1 Symposium, Greenbelt: NASA SP 351, 3010317. Ryther, J. H. 1969. Photosynthesis and Fish Production in the Sea. Science 166: 7276. Saavedra, B. and J. A. Simonetti, J.A. 2005. Small mammals of Maulino forest remnants, a V anishing ecosystem of so uthcentral Chile. Mammalia 69: 337348. SAS. 2001. SAS/ STAT. SAS Institute Inc, Cary, NC. Schindler, D.E., S. R. Carpenter, J. J. Cole, J. F. Kitchell, and M. L. Pace 1997. Influence of food web structure on carbon exchange between lakes and the atmosphere. Science 277(5323): 248251. Schlatter, J. E. 1977. La relacion ente suelo y plantaciones de Pinus radiate D. Don en Chile Central: Analisis de la situacion actual y planteamientos para su future manejo. Bosque
166 2(1): 12 31. Schlatter, J.E., and V. Gerding. 1995. Metodos de clasificacion de s itios para la produccion forestal, ejemplo en Chile. Bosque 16:13 20. Schlatter, J.E., Gerding, V., and H. Huber. 1995. Sistema de Ordenamiento de la Tierra: Herramienta para la planificacion forestall aplicado a la X Region. Universidad Austral de Chile Faculdad de Ciencias Forestales, Valdivia, Chile. Sedjo, R. A. 1999. The potential of highyield plantation forestry for me eting timber needs. New Forests 17: 339359. SERNAP (Servicio Nacional de Pesca). 2003. Anuarios Estadisticos 2003. Servici o Nacional de Pesca Valparaiso, Chile. SERNAP. 2004. Anuario estadistico de pesca 2004. Servicio National de Pesca, Ministerio de Economia, Fomento y Reconstruccion Chile, Valparaiso, Chile. SERNAP. 2005. Anuario estadistico de pesca 2005. Servici o National de Pesca, Ministerio de Economia, Fomento y Reconstruccion Chile, Valparaiso, Chile. Shiomoto, A., S. Saitoh, K. Imai, M. Toratani Y. Ishida and K. Sasoaka. 2002. Internannual variation in phytoplankton biomass in the Bering Sea basin i n 1990s. Progress in Oceanography 55: 147163. Smith, V. H., G. D. Tilman and J. C. Nekola. 1999. Eutrophication: impacts of excess nutrient inputs in freshwater, marine, and terrestrial ecosystems. Environmental Pollution 100:179196. Stev ens, P. A., D. A. Norris, T. H. Sparks, and A. L. Hodgson. 1994. The Impacts of Atmospheric N Inputs on Throughfall, Soil and Stream Water Interactions for Different Aged Forest and Moorland Catchments in Wales. Wa ter Air and Soil Pollution 73: 297317. Southworth, J., and C. Tucker. 2001. The influence of accessibility local institutions, and
167 socioeconomic factors on forest cover change in the mountains of Western Honduras. Mountain Research and Development 21:276283. Spiegelhalter, D. J., N. G Best, B. P. Carli n, and A. van der Linde. 2002. Bayesian Measures of Model Complexity and Fit (With Discussion) Journal of the Royal Statistical Society B 64:583639. Stotz, W. B, S. A.Gonzalez L. Caillaux and J. Aburto. 2003. Quantitative evaluation of the diet and feeding behavior of the carnivorous gastropod, Concholepas concolepas (Burguiere, 1789) (Murcidae) in subtidal habitats in a southeastern pacific upwelling system. Journal of Shellfish Research 22 (1): 147164. Talley, D. M. Northb, E W., Juhlc, A. R., Timothyd, D. A., Condee, D., deBrouwerf, J. F.C., Browng, C. A., Campbellh, L. M., Garsteckii, T., Halla, C. J., Meysmanf, F. J.R., Nemersonj, D.M., Souza Filhok, P.W., and R. J. Woodl. Research challenges at the landsea interf ace. Estuarine, Coa stal and Shelf Science 58: 699702. Teillet, P.M., and G. Fedosejeus, G. 1995. On the Dark Target Approach to Atmospheric Correction of Remotely Sensed Data. Canadian Journal of Remote Sensing 21(4):374 387. Teillet, P. M., Guin don, B., and D.G. Goodenough. 1982. On the slope aspect correction of multispectral scanner data. Canadian Journal of Remote Sensing 8: 84 106. Thiel, M., Macaya, E.C., Acuna, E., Arntz, W.E., Bastias, H., Brokordt, K., Camus, P.A., Castilla, J.C., Cast ro, L.R., Cortes, M., Dumont, C.P., Escribano, R., Fernandez, M., Gajardo, J.A., Gaymer, C.F., Gomez, I., Gonzalez, A.E., Gonzalez, H.E., Haye, P.A., Illanes, J.E., Iriarte, J.L., Lancellotti, D.A., Luna Jorquerai, G., Luxoroi, C.,Manriquez, P.H., M arin, V., Munoz, P., Navarrete, S.A., Perez, E., Poulin, E., Sellanes, J., Sepulveda, H.H., Stotz, W., Tala, F. Thomas, A. Vargas,C.A., Vasquez, J.A., Vega, J.M.A. 2007. The Humboldt Current System of northern and central Chile O ceanography and Marine Biology. 45:195344. Thomas, D. and B. Franz. 2005. Overview of SeaWiFS Data Processing and Distribution. Accessed April 6th, 2008 at http://oceancolor.gsfc.nasa.gov/DOCS/SW_proc.html Tierney, D., M. Ke lly Quinn, and J. J. Bracken. 1998. The faunal communities of upland streams in the eastern region of Ireland with reference to afforestation impacts. Hydrobiologia 389:
168 115130. Toral I., M., Fratti B., A., and L.A. Gonzalez R. 2005. Seasonal growth and yield of radiata pine commercial forest plantation s on volcanic ("trumao") soils. Bosque 26(1):43 54. Toro, J., and S. Gessel. 1999. Radiata pine planta tions in Chile. New Fore sts 18: 3344. Trumbore, S. E., and E. R. M. Druffel 1995. Carbon isotopes for characterizing sources and turnover of nonliving organic matter. Pages xxx in Z epp, R.G. and C. Sonntag editors, Role of Nonliving Organic Mat ter in the Earths Carbon Cy cle. John Wiley. Hoboken, New Jersey. Turner II, B. L. and P. Robbins. 2008. Land change science and political ecology: similarities, differences, and implications for sustainability science. Annual Review of Environmental Resources 33:295316. UACH ( Un iversidad Austral de Chile). 2005. Esudio sobre origin de mortalidades y disminucion poblacional de aves acuaticas en el santuario de la naturaleza Carlos Anwandter, en la provincial de Valdivia. Final Report. Convenio Complementario Esepecifico N 121001203/20041214, Direccion regional CONAMA Xth Region de Los Lagos Universidad Austral de Chile. April 18, 2005. Van Holt T. a, In prep. Land Cover/Land Use Change and Forest Plantation Development Patterns in a coastal te mperate forest of s outhern Chile. Van Holt, T. b, In prep. Mesoscale analysis of cholorophyll a patterns across Chiles central coast and forest plantation development in the t emperate r ain f orest ecosystem. Van Holt, T., c. In prep. A mesoscale analysis of the inf lue nces of landscape change on l oco ( Concholepas concholepas ) health condition characteristics Van Holt, T., d. In prep. Adaptive Capacity for Fisher Success. Vander Zanden, M. J. B. J. Shuter N. Lester and J. B. Rasussen. 1999. Patterns of food ch ain length in lakes: a stable isotope study. American Naturalist. 154(4):406416. V ergara, P. M. and J. A. Simonetti, J.A. 2004. Avian responses to fragmentation of the Maulino
169 Forest in central Chile. Oryx 38 : 383388. Wada, E., M. Terazaki, Y. Kabaya and T. Nemoto. 1987. N 15 and C 13 abundances in the Antarctic Ocean with emphasis on the biogeochemical structure of the food web. Deep Sea Research Part A Oceanographic Research Papers 34 (5 6): 829841 Wahl, M. 1989. Marine epibiosis. I. Fouling and antifouling: some basic aspects. Marine Ecology Progress Series. 5: 175189. Wilson, K., A. Newton C. Echeverria, C. Weston, and M. Burgman. 2005. A vulnerability analysis of the temperate forest of south central Chile. Biological Conse rvation 122: 921. Zander, C. D., and L. W. Reimer. 2002. Parasitism at the ecosystem level in the Baltic Sea. Parasitology 124:S119S135. Zhang, Y.X., and C.H. Song. 2006. Impacts of afforestation, deforestation, and reforestation on forest cover in China from 1949 to 2003. Journal of Forestry 104: 383387.
170 BIOGRAPHICAL SKETCH Tracy Van Holt is a Landscape Socio Ecologist finishing her Ph.D. at the University of Florida. Her research focuses on the ecological and social factors that influence natur al resource harvest and marketing of these products with a focus on Latin American Conservation and Development. For her dissertation, Tracy has studied the effects of land use/land cover change on the nearshore fisheries in southern Chile. She has identif ied how forest plantations influenced the nearshore environment and the impacts on the loco shellfishery. Her work also addressed fisher success and the social and ecological (at the small scale and landscape level) factors that influence price for various artisanal fisheries in Chile. Her past work with hunters in Bolivia has a ddressed how researchers can gain a consensus about a communitys opinion on natural resources. For her MS at the University of Florida, Tracy worked in the Brazilian Pantanal and ex amined how wildlife movements were influenced by palm fruits and spatial arrangement of forest islands. Prior to grad school, Tracy ran the Research Fellowship Program at the Wildlife Conservation Society for three years. Her B.S. degree is in Biochemistry from SUNY at Stony Brook. Her research has been supported by NASA, Fulbright, NSEP, Rotary, and the TCD program at the University of Florida.