<%BANNER%>

Effect of Urbanization and Hurricanes on the Provision of Ecosystem Services in the Lower Suwannee River and Pensacola B...

Permanent Link: http://ufdc.ufl.edu/UFE0044694/00001

Material Information

Title: Effect of Urbanization and Hurricanes on the Provision of Ecosystem Services in the Lower Suwannee River and Pensacola Bay Watersheds, Florida, USA
Physical Description: 1 online resource (139 p.)
Language: english
Creator: Delphin, Sonia
Publisher: University of Florida
Place of Publication: Gainesville, Fla.
Publication Date: 2012

Subjects

Subjects / Keywords: carbon -- decision-tree -- drivers -- ecosystem-services -- ecosystems -- hurricanes -- land-cover-change -- pensacola -- services -- suwannee -- timber -- urbanization -- water
Forest Resources and Conservation -- Dissertations, Academic -- UF
Genre: Forest Resources and Conservation thesis, M.S.
bibliography   ( marcgt )
theses   ( marcgt )
government publication (state, provincial, terriorial, dependent)   ( marcgt )
born-digital   ( sobekcm )
Electronic Thesis or Dissertation

Notes

Abstract: Ecosystems can be impacted by differentdrivers that can alter their functionality. Thus, to understand the magnitudeof these impacts to Ecosystem Service (ES) provision, quantification of ES andthe drivers is important. This study estimated the effects of urbanization andhurricanes as direct drivers of carbon storage and timber volume, as well as wateryield in forest areas in the Lower Suwannee River and Pensacola Bay watersheds,in Florida, USA. A current 2003 to 2060 urbanization scenario using spatialdata and modeling, analyzed the effects of urbanization on ES. Results showed statisticallysignificant differences at 95% confidence level between 2003 and 2060 in termsof provision of the three ES analyzed. For the hurricane analysis, this studyestimated the potential loss of aboveground carbon storage and timber volume inhigh and low forest damage risk zones. Five factors were used in a decisiontree to determine potential for forest damage due to hurricanes. Results showedthat in the LowerSuwannee, 31% of the total aboveground carbon and 15% of the total timber werelocated in the high forest damage risk (HD) zone while the Pensacola Bay, had 0.5%of the total aboveground carbon and about the 0.7% of the total timber in theHD zone. Knowledge on how differentdrivers impact ecosystems and their services will provide a betterunderstanding to landowners, policy makers, about the consequences of land usechange and natural disturbance on the benefits provided by forests.
General Note: In the series University of Florida Digital Collections.
General Note: Includes vita.
Bibliography: Includes bibliographical references.
Source of Description: Description based on online resource; title from PDF title page.
Source of Description: This bibliographic record is available under the Creative Commons CC0 public domain dedication. The University of Florida Libraries, as creator of this bibliographic record, has waived all rights to it worldwide under copyright law, including all related and neighboring rights, to the extent allowed by law.
Statement of Responsibility: by Sonia Delphin.
Thesis: Thesis (M.S.)--University of Florida, 2012.
Local: Adviser: Escobedo, Francisco Javier.
Local: Co-adviser: Abd-Elrahman, Amr H.

Record Information

Source Institution: UFRGP
Rights Management: Applicable rights reserved.
Classification: lcc - LD1780 2012
System ID: UFE0044694:00001

Permanent Link: http://ufdc.ufl.edu/UFE0044694/00001

Material Information

Title: Effect of Urbanization and Hurricanes on the Provision of Ecosystem Services in the Lower Suwannee River and Pensacola Bay Watersheds, Florida, USA
Physical Description: 1 online resource (139 p.)
Language: english
Creator: Delphin, Sonia
Publisher: University of Florida
Place of Publication: Gainesville, Fla.
Publication Date: 2012

Subjects

Subjects / Keywords: carbon -- decision-tree -- drivers -- ecosystem-services -- ecosystems -- hurricanes -- land-cover-change -- pensacola -- services -- suwannee -- timber -- urbanization -- water
Forest Resources and Conservation -- Dissertations, Academic -- UF
Genre: Forest Resources and Conservation thesis, M.S.
bibliography   ( marcgt )
theses   ( marcgt )
government publication (state, provincial, terriorial, dependent)   ( marcgt )
born-digital   ( sobekcm )
Electronic Thesis or Dissertation

Notes

Abstract: Ecosystems can be impacted by differentdrivers that can alter their functionality. Thus, to understand the magnitudeof these impacts to Ecosystem Service (ES) provision, quantification of ES andthe drivers is important. This study estimated the effects of urbanization andhurricanes as direct drivers of carbon storage and timber volume, as well as wateryield in forest areas in the Lower Suwannee River and Pensacola Bay watersheds,in Florida, USA. A current 2003 to 2060 urbanization scenario using spatialdata and modeling, analyzed the effects of urbanization on ES. Results showed statisticallysignificant differences at 95% confidence level between 2003 and 2060 in termsof provision of the three ES analyzed. For the hurricane analysis, this studyestimated the potential loss of aboveground carbon storage and timber volume inhigh and low forest damage risk zones. Five factors were used in a decisiontree to determine potential for forest damage due to hurricanes. Results showedthat in the LowerSuwannee, 31% of the total aboveground carbon and 15% of the total timber werelocated in the high forest damage risk (HD) zone while the Pensacola Bay, had 0.5%of the total aboveground carbon and about the 0.7% of the total timber in theHD zone. Knowledge on how differentdrivers impact ecosystems and their services will provide a betterunderstanding to landowners, policy makers, about the consequences of land usechange and natural disturbance on the benefits provided by forests.
General Note: In the series University of Florida Digital Collections.
General Note: Includes vita.
Bibliography: Includes bibliographical references.
Source of Description: Description based on online resource; title from PDF title page.
Source of Description: This bibliographic record is available under the Creative Commons CC0 public domain dedication. The University of Florida Libraries, as creator of this bibliographic record, has waived all rights to it worldwide under copyright law, including all related and neighboring rights, to the extent allowed by law.
Statement of Responsibility: by Sonia Delphin.
Thesis: Thesis (M.S.)--University of Florida, 2012.
Local: Adviser: Escobedo, Francisco Javier.
Local: Co-adviser: Abd-Elrahman, Amr H.

Record Information

Source Institution: UFRGP
Rights Management: Applicable rights reserved.
Classification: lcc - LD1780 2012
System ID: UFE0044694:00001


This item has the following downloads:


Full Text

PAGE 1

1 EFFECT OF URBANIZATION AND HURRICA N ES ON THE PROVISION OF ECOSYSTEM SERVICES IN THE LOWER SUWANNEE RIVER AND PE N SACOLA BAY WATERSHEDS, FLORIDA, USA By SONIA ELSA ROCIO DELPHIN PEREZ A THESIS PRESENTED TO THE GRADUATE SCHOO L OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF MASTER OF SCIENCE UNIVERSITY OF FLORIDA 2012

PAGE 2

2 2012 Sonia Elsa Rocio Delphin Perez

PAGE 3

3 To my father and my mother

PAGE 4

4 ACKNOWLEDGMENTS I would like to tha nks my mother for being by my side despite the distance. I would also like to especially thank Dr. Francisco Escobedo for giving me the opportunity to patience, support, and fr iendship this ha s not been possible. His valuable comments were a key point for improving this thesis work. I would also want to thank Dr Amr Abd Elrahman and Dr. Wendell Cropper who provided good contributions that improve this study, and also for the ti me dedicated to read it All the comments provided by my committee were essential for improving this document; as well they will be very useful for my future works. Special thanks to the Florida Forest Service and the School of Forest Resources and Conserv ation University of Florida for the financial support that made this thesis possible. I would also like to thank Kimmel Chamat for being by my side during this time for supporting me in difficult times, for encouraging me to be better every day and for g iving me the strength to keep moving forward I also want to thank Natalia Nequi for trusting me and being part of this stage. I thank my friends and family in Gainesville: Martha Rosero, Claudia Penuela, Angelica Garcia, Oscar Murillo, Gisselle Guerra, Iv elisse Ruiz, Silvia Marino, Ronald Ca demus, Yin Hsuen Chen, Li Bai, Pinar Delilo lu, Esli Suarez, Maria Mendes, and Cristob a l Rodriguez, for all the support and all the fun times we have t ogether I would like to thank Lucy Aquino for trusting me and believing in me. I also want to thanks my best friends in Paraguay: Sandra Perrens, V eronica Rivas, Irma Santacruz, and Mabel Torrico for the distance support. Thanks to all people who joined me in this path.

PAGE 5

5 TABLE OF CONTENTS page ACKNOWLEDGMENTS ................................ ................................ ................................ .. 4 LIST OF TABLES ................................ ................................ ................................ ............ 7 LIST OF FIGURES ................................ ................................ ................................ .......... 9 ABSTRACT ................................ ................................ ................................ ................... 11 CHAPTER 1 INTRODUCTION ................................ ................................ ................................ .... 13 Conceptualizing Ecosystem Services ................................ ................................ ..... 13 Effect of Drivers on the Provision of Ecosystem Services ................................ ....... 17 Using Geographic Information Systems in Ecosystem Service Assessments ........ 18 Objectives ................................ ................................ ................................ ............... 20 2 EFFECTS OF URBANIZATION ON THE PROVISION OF ECOSYTEM SERVICES ................................ ................................ ................................ .............. 25 Background ................................ ................................ ................................ ............. 25 Carbon Storage and Urbanization ................................ ................................ .... 26 Timber and Urbanization ................................ ................................ .................. 28 Water Yield and Urbanization ................................ ................................ ........... 28 Ecosy stem Services Models ................................ ................................ ............. 30 Methods ................................ ................................ ................................ .................. 31 Study Areas ................................ ................................ ................................ ...... 31 Quantification of Ecosystem Services ................................ .............................. 34 Carbon storage ................................ ................................ .......................... 34 Forest Inventory and Analysis Program data ................................ ............. 34 Carbon pools analysis ................................ ................................ ................ 35 InVEST carbon storage model ................................ ................................ ... 39 Timber volume ................................ ................................ ........................... 39 Water yield estimation ................................ ................................ ................ 40 Water model input data process ................................ ................................ 42 InVEST water model assessment ................................ .............................. 44 2060 land use and land cover map ................................ ............................ 44 Results ................................ ................................ ................................ .................... 46 Carbon Storage in 2003 and Urbanized 2060 ................................ .................. 46 Timber Volume in 2003 and Urbanized 2060 ................................ ................... 48 Water Yield in 2003 and Urbanized 2060 ................................ ......................... 49 Water Yield Assessment for the Lower Suwannee River Watershed ............... 50 Tradeoffs Between Carbon Storage, Timber Volume, Water Yield, and Urbanizatio n ................................ ................................ ................................ .. 51

PAGE 6

6 Discussion ................................ ................................ ................................ .............. 51 Carbon Storage in 2003 and Urbanized 2060 ................................ .................. 52 Timber Volume in 2003 and Urbanized 2060 ................................ ................... 54 Water Yield in 2003 and Urbanized 2060 ................................ ......................... 54 Water Yield Assessment ................................ ................................ .................. 55 Tradeoffs Betwee n Carbon Storage, Timber Volume, Water Yield and Urbanization ................................ ................................ ................................ .. 56 Concluding Remarks ................................ ................................ ............................... 58 3 ASSESSING AND MAPPING POTENTIAL LOSS OF ABOVEGROUND CARBON STORAGE AND TIMBER VOLUME FROM HURRICANES ................... 82 Background ................................ ................................ ................................ ............. 82 Impacts of Hurricanes on Forests and Ecosystem Services ............................. 83 Factors Controlling Hurricane Damage on Forests ................................ .......... 85 As sessing and Mapping Hurricane Damage in Forests ................................ .... 87 Methods ................................ ................................ ................................ .................. 90 Study Areas ................................ ................................ ................................ ...... 90 Quantification of Aboveground Carbon Storage and Timber Volume ............... 92 Forest Inventory and Analysis Program Data ................................ ................... 92 Flo rida Vegetation and Land Cover 2003 ................................ ......................... 94 Decision tree Framework ................................ ................................ ................. 95 Potential Loss of Aboveground Carbon Storage and Timber Volume .............. 99 Results ................................ ................................ ................................ .................... 99 Potential Loss of Aboveground Carbon Storage Due to Hurricanes ................. 99 Potential Loss of Timber Volume Due to Hurricanes ................................ ...... 101 Discussion ................................ ................................ ................................ ............ 102 Potential Loss of Aboveground Carbon Stora ge Due to Hurricanes ............... 103 Potential Loss of Timber Volume Due to Hurricanes ................................ ...... 104 Concluding Remarks ................................ ................................ ............................. 105 4 CONCLUSIONS ................................ ................................ ................................ ... 119 Summary ................................ ................................ ................................ .............. 119 Effects of Urbanization on Carbon Storage, Timber Volume, an d Water Yield ..... 119 Potential Loss of Aboveground Carbon Storage and Timber Volume from Hurricanes ................................ ................................ ................................ ......... 120 Integrating the Concept of Ecosystem Services in Decisions ............................... 121 LIST OF REFERENCES ................................ ................................ ............................. 124 BIOGRAPHICAL SKETCH ................................ ................................ .......................... 139

PAGE 7

7 LIST OF TA BLES Table page 1 1 C ategories of ecosystem services ................................ ................................ ...... 22 1 2 Multiple Ecosystem Services assessment tools ................................ ................. 22 2 1 Land cover distribution in the Lower Suwannee River and Pensacola Bay watersheds, Florida based on the Florida Fish and Wildlife Conservation Commission ................................ ................................ ................................ ........ 61 2 2 Ecological subsection (Ecol. Subs.) descriptions ................................ ................ 62 2 3 Description of the Forest Inventory and Analysis (FIA) attributes used in the carbon and timber analysis ................................ ................................ ................. 63 2 4 Forest tree species used in the carbon storage and timber volume analysis by Land cover type ................................ ................................ ............................. 64 2 5 Carbon pools aggregation ba sed on the information available in the Forest Inventory and Analysis (FIA) Program database ................................ ................ 66 2 6 Reclassified land cover and carbon pools for the Pensacola Bay watershed ..... 66 2 7 Input data for the InVEST water yield model ................................ ...................... 69 2 8 Carbon storage descriptive statistics for the Lower Suwannee (LS) and Pensacola Bay (PB) Florida watersheds in Mg/ha ................................ .............. 71 2 9 Carbon storage in the four pools (Mg) for the Lower Suwannee River (LS) and Pensacola Bay (PB) watersheds, Florida ................................ .................... 71 2 10 Timber volume values for the Lower Suwannee (LS) and Pensacola Bay (PB) watersheds, year 2003 2060 ................................ ................................ ...... 71 2 11 Water yield for the Lower Suwannee (LS) and Pensacola B ay (PB) watersheds ................................ ................................ ................................ ......... 71 2 12 Average volume in m3 for the Gopher gage station and the InVEST water yield model ................................ ................................ ................................ ......... 72 2 13 Pearson Cor relation coefficients between carbon storage, timber volume, water yield and p ercentage (%) of forest change for the Lower Suwannee River and Pensacola Bay watersheds at the sub watershed level ..................... 72 3 1 Description of the Forest Inventory and Analysis (FIA) attributes used in the hurricane framework ................................ ................................ ......................... 108

PAGE 8

8 3 2 Range of values for the two defined classes for tree diameter, tree height an d basal area for the Lower Suwannee River and Pensacola Bay watersheds ..... 109 3 3 Expressions used in the Conditional function in ArcGIS Spatial Analyst to define the risk of forest damage by hu rricanes ................................ ................. 109 3 4 Potential aboveground carbon, timber net volume, and timber in the sawlog portion in the forest damage risk zones for the Lower Suwannee watershed ... 109 3 5 Potential aboveground carbon, timber net volume, timber in the sawlog portion in the forest damage risk zones for the Pensacola Bay watershed ...... 109

PAGE 9

9 LIST OF FIGURES Figure page 1 1 Framework for linkin g ecosystems to human wellbeing ................................ ...... 24 2 1 Study area: Lower Suwannee River and Pensacola Bay watersh eds ................ 73 2 2 Approach for the estimation of carbon pools and timber volume using the Forest Inventory and Analysis (FIA) program data ................................ ............. 73 2 3 Reclassification of the original land covers to include the ecological subsection and county data. ................................ ................................ ............... 74 2 4 Flow chart for estimating carbon and timber in the Lower Suwannee River and Pensacola Bay Florida watersheds to approximate the exact locations of the Forest Inventory and Analysis (FIA) plots ................................ ..................... 75 2 5 Conceptual framework for the InVEST water yield model ................................ .. 76 2 6 Total Carbon storage in the Lower Suwannee River, Florida. ............................ 77 2 7 Total carbon storage in the Pensacola Bay watershed ................................ ....... 78 2 8 Differences between 2060 and 2003 in the water yield per hectare for the Lower Suwannee River watershed ................................ ................................ ..... 79 2 9 Differences between 2060 a nd 2003 in the water yield (mean) and precipitation (ppt) for the Pensacola Bay watershed ................................ .......... 79 2 10 Normalized curves showing differences between average water yield volume from the InVEST model and the Gage station data (actual) for the Lower Suwannee River watershed ................................ ................................ ................ 80 2 11 Carbon storage, timber volume, and water yield in 2003 (current) and 2060 for the Lower Suwannee (LS) and Pen sacola Bay (PS) watersheds .................. 81 3 1 Decision tree scenario for determining forest damage risk due to hurricanes .. 110 3 2 American S 98 map with wind risks for the two watersheds: Lower S uwannee River and Pensacola Bay ............... 111 3 3 Flow chart for the decision tree analysis for determinin g forest damage risk zones ................................ ................................ ................................ ................ 112 3 4 Aboveground carbon storage in the Lower Suwannee River watershed in 2003 ................................ ................................ ................................ ................. 113

PAGE 10

10 3 5 Risk of fore st damage by hurricanes in the Lower Suwannee River watershed ................................ ................................ ................................ ......... 114 3 6 High risk of forest damage by hurricanes in the Lower Suwannee River watershed by sub watersheds ................................ ................................ .......... 115 3 7 Aboveground carbon storage in the Pensacola Bay watershed in 2003 ........... 116 3 8 Risk of forest damage by hurricanes in the Pensacola Bay watershed ............ 117 3 9 High risk of forest damage by hurricanes in the Pensacola Bay watershed by sub watersheds ................................ ................................ ................................ 118

PAGE 11

11 Abstract of Thesis Presented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Master of Science EFFECT OF URBANIZATION AND HURRICANES ON THE PROVISION OF ECOSYSTEM SERVICES IN THE LOWER SUWANNEE RIVER AND PENSACOLA BAY WATERSHEDS, FL ORIDA, USA By Sonia Elsa Rocio Delphin Perez August 2012 Chair: Francisco Escobedo Cochair: Amr Abd Elrahman Major: Forest Resources and Conservation Ecosystem s can be impacted by different drivers that can alter their functionality. Thus, t o understan d the magnitude of the se impacts to Ecosystem Service ( ES ) provision quantification of ES a nd the drivers is important This study estimate d the effects of urbanization and hurricanes as direct driver s of carbon storage and timber volume, as well as w ater yield in forest areas in the Lower Suwannee River and Pensacola Bay watersheds, in Florida, USA. A current 2003 to 2 060 urbanization scenario using spatial data and modeling, analyzed the effects of urbanization on ES. Results show ed statistically signifi cant differences at 95% confidence level between 2003 and 2060 in terms of provision of the three ES analyzed. For the hurricane analysis, t his study estimate d the potential loss of aboveground carbon storage and timber volume i n high and low forest damage risk zones F ive factors were used in a decision tree to determine potential for forest damage due to hurricanes Results show ed that in the Lower Suwannee 31% of the total aboveground carbon and 15% of the total timber were located in the high forest da mage risk (HD) zone while the Pensacola Bay had 0.5% of the total aboveground carbon and about the 0.7% of the

PAGE 12

12 total timber in the HD zone Knowledge on how different drivers impact ecosy stems and their services will provide a better understanding to land owners, policy makers about the consequences of land use change and natural disturbance on t h e benefits provided by forests

PAGE 13

13 CHAPTER 1 INTRODUCTION Conceptualizing Ecosystem Services Ecosystem services is becoming a mainstream concept in natural resourc es and environmental planning and management (Boyd and Banzhaf, 2007; Brown et al., 2007; de Groot et al., 2010; Farley and Costanza, 2010) The concept of ecosystem services has been increasingly used in the late 1990s (Costanza et al., 1997; Daily, 1997) There are several definitions of ecosystem services. One of the first uses of the term (1997) who defined them as the natural and benefit human life. These ecosystem services provide biodiversity and ecosystem goods, such as food, timber, fuel wood, among others. In addition to the production of goods, ES are based on ecological functions, such as providing clean water, and aesthetic and cultural benefits. Unfortunately, a coherent and integrated approach to practical applications of the concept in ecosystem and landscape planning, management and decision making is lacking (Daily et al., 2009; de Groot et al., 2010; ICSU et al., 2008) Daily et al. (2009) mentioned that finding a link between decision s and ecosystems is a huge challenge. De Groot et al. (2010) discussed some alternatives to integrate ecosystem services and values in planning, policy and decision making. One example is the communication about ecosystem services and their v alues in a clear way with the objective of making the concept accessible to policy makers (de Groot et al., 2010) Another alternative is to make databases related to ecosystem services accessible to land use planners (de Groot et al., 2010)

PAGE 14

14 Brown et al. (2007) refers to Daily (1997) and makes a distinction between ecosystem services and goods, and that both can produce benefits to human well being. Ecologists are likely to distinguish between ecosystem goods and services (Brown et al., 2007; Daily, 1997; de Groot et al., 2010; European Commission, 2009) however others are often combining both as ecosystem services and referring to ecosystem goods as one category of ecosystem services named provisioning ecosystem services (Costanza et al., 1997; de Groot et al., 2010; Millennium Ecosystem Assessment, 2003) The concept of ecosystem services a nd goods are usually used as synonymous with ecosystem functions or processes. However, ecosystem processes determine ecosystem services and goods. Therefore, ecosystem processes are an intermediate step, thus they often do not produce direct benefits to p eople; however, they are necessary for human life and their values are included in the final values (Figure 1 1) (Brown et al., 2007; Costanza et al., 1997; de Groot et al., 2010; Millennium Ecosystem Assessment, 20 03) The most commonly cited concept for ecosystem services is the one developed by the Millennium Ecosystem Assessment (2003) and is based on a combination of the definitions of Daily (1997) and Costanza et al (1997) According to the Millennium Ecosystem Assessment (2003) ES are the services provided by ecosystems with a direct influence on human well being. There are different categories of ecosystem services acc ording to the Millennium Ecosystem Assessment (2003) these include provisioning, regulating, cultural, and supporting services (Table 1 1). Provisioning services are the products people can get from the ecosystem; these are the ecosystem goods (e.g. food and timber). Regulating services refer to the benefits people receive

PAGE 15

15 from the regulation of ecosystem processes (e.g. flooding regulation and clean air). Cultural services are the nonmaterial benefits obtained from ecosystems (e.g. spiritual enrichment and recreation). Supporti ng services are those necessary for the production of other ecosystem services (e.g. primary production). Natural ecosystems (e.g. forests, savannas, and prairies) and human modified ecosystems (e.g. croplands and urban areas) should also be considered whe n assessing ecosystem services (Costanza et al., 1997; Millennium Ecosystem Assessment, 2003) Also, the four previously mentioned categories should be analyzed in order to have a more integrative approach because ecosystems can produce services in several categories. In doing so, ecosystems generate benefits to humans; some ecosystems provide more services than others. For example, forests provide a variety of services (Nasi et al., 2002; Yonavjak et al., 2011) in d ifferent categories: provisioning such as timber, pulpwood for paper, and non timber forest products; regulating such as climate, erosion, and water flow control, as well as, cultural services such as recreation and tourism, and some supporting services su ch as nutrient and water cycling, and primary production (Table 1 1). Specifically, southern United States (U.S.) forests are one of the most biologically diverse temperate forests in the world and consequently they produce a bundle of ecosystem services (i.e. a set of ecosystem services that are provided together as benefits to people) (Raudsep p Hearne et al., 2010) Therefore, people can get multiple ecosystem services; however they can decide which services are more relevant for them depending of what specific benefit they want from these ecosystems. For example, if a population wants to incr ease flood protection, or meet needs for supply of

PAGE 16

16 water, or an increase in small farm production, different services will be considered (McKenzie et al., 2011) So, the inclusion of the population in the proce ss of knowing about what ecosystem services is important for them and their region is key. Additionally, these ecosystem services can be identified at different scales, such as local (e.g. timber), regional (e.g. clean water), and global (e.g. climate regu lation) (Yonavjak et al., 2011) so this distinction in terms of scale is important for management purposes Although the fact that ecosystems provide services to people is already known, yet there are still several inf ormation gaps about ecosystem service dynamics, specifically how the provision of ecosystem services will change with changes in ecosystems (de Groot et al., 2010) Some aspects are still being studied and especially those related to the in teractions between ecosystem services (i.e. tradeoffs) and the effects o f land cover change on the provision of these ES. As a result of this lack of information, some land covers that are multifunctional (and consequently multiple ES provision) are conver ted often to single functional land covers (e.g. croplands). All ecosystems constantly change over time, and many of these changes are due to human impacts (e.g. urbanization) and natural processes (e.g. wildfire) (ICSU et al., 2008; Millennium Ecos ystem Assessment, 2003) Therefore, u nderstanding the causes of these changes, or drivers, is fundamental to minimize their impact on the provisioning of ecosystem services (Millennium Ecosystem Assessment, 2003) By definition, drivers are ecological or human factors that can affect ecosystem composition, functioning, and processes, that then result in increasing or decreasing provisioning of ecosystem services (Millennium Ecosystem Assessment, 2003) These

PAGE 17

17 drivers can be classified as direct or indirect drivers. Di rect drivers refer to those that directly affect ecosystem processes; on the other hand, indirect drivers are those affecting direct drivers. Examples of direct drivers are : land use change (e.g. forest to agriculture conversions), climate change, hurrican es pest and diseases, among others I ndirect drivers can include : population growth, land use policies, land values, and market values for provisioning services (ICSU et al., 2008; Millennium Ecosystem Assessment, 2003; Yonavjak et al., 2011) One of the challenges about ecosystem services is the identification of the main drivers of change, in other words, the drivers that affect specific ecosystems, as well as their consequences on the provision of ecosystem serv ices (M illennium Ecosystem Assessment, 2003) The definition of the key drivers for specific ecosystems will depend on several factors, such as historical information. For example, the most important direct drivers of change in terrestrial ecosystems over the pa st 50 years have been: land cover change (e.g. deforestation, cropland conversion) (Nelson, 2005) and the type of ecosystem (e.g. terrestrial and marine in that the main driver for marine ecosystems has been fishing). There are also economic, institutional technological, cultural, and demographic factors. Effect of D rivers on the Provision of Ecosystem S ervices In general, h uman beings modify ecosystems to increase the productivity of provisioning services such as food, timber, without paying attention to the tradeoffs, but increasing the use of one ecosystem service is resulting in a decrease in the use of another ecosystem service (Hancock, 2010; Millennium Ecosystem Assessment, 2003; Rodriguez et al., 2006) To b etter understand the influence of drivers in the form of land cover changes (e.g. forest to urban areas) and their effect on the provision of

PAGE 18

18 ecosystem services, it is becoming important to link this as a direct consequence to human wellbeing (ICSU et al., 2008) since human activities are producing changes in ecosystems rapidly (Metzger et al., 2006) The Millennium Ecosystem Assessment (2003) for example proposes a multi sectorial approach for evaluating changes in different ecosystem services (e.g. crop production versus water quality), as well as to better understand the interactions between them and the impacts on peopl e. The future of ecosystems to provide ecosystem services will depend on socio economic, clim ate, and land use changes among others drivers (Metzger et al., 2006) Various studies have been done to quantify the impact of land cover changes on the provision of ecosystem services. Reyers et al. (2008) found that ecosystem services experienced a decline ranging from 18% 44% as consequence of the land use change, specifically to croplands and urban areas in the Little Karoo in South Africa. Zang et al. (2011) quantified the decrease in ecosystem service values due t o land use changes (urbanization and croplands) in the HaDaQi industrial corridor in China, and they found a decrease of 29% from 1990 to 2005. McNulty (2002) found that a single hurricane can affect carbon sequestration in forests converting 10% of the total carbon sequestered by U.S forests into dead biomass. According to Zhang a nd Nagubadi (2005) urbanization has a negative impact on timberland use in general. However, there is still a need to do more ecosystem service studies to assess th e effects of land cover changes with more effective and simple methods. Using Geo graphic Information Systems in Ecosystem Service A ssessments Most of the works that quantify value and analyze trade offs among ecosystem services and land cover changes are b ased on Geographic Information Systems (GIS) methods. Since e cosystem s ervices and land cover changes can be quantified and

PAGE 19

19 mapped using GIS (Boyd and Banzhaf, 2007; Metzger et al., 2006) t his method is becoming w idely used because its simplicity for studies involved in evaluating and mapping ecosystem services (Chen et al., 2009; Egoh et al., 2008; Troy and Wilson, 2006) Increasing interest in ecosystem services has deriv ed the development of new tools to perform a multi ecosystem service assessments (Waage et al., 2008) Some examples are presented by Waage et al. (2008) and include : ( 1) Artificial Intelligence for Ecosystem Services ARIES, (2) Ecosystem Services Review ESR, (3) Integrated Valuation of Ecosystem Services and Tradeoffs InVEST, Multi scale Integrated Models of Ecosystem Services MIMES, and (5) Natural Value Initiative NV I (Table 1 2). Models such as InVEST use spatial data to estimate and map the quantity of ES (Tallis et al., 2011) The InVEST model is a GIS spatially explicit tool used for quantifying and valuing ecosystem services based on ec ological production functions and economic valuation methods (Nelson et al., 2009; Tallis et al., 2011) .The InVEST model is characterized by its repeatability and simplicity and because not much data is needed for running it; therefore it can be used in areas with poor data availability, and adjusted to the needs of the research because scenario analyses can be done (Tallis et al., 2011; Waage et al., 2008) Even though GIS tools can help in the assessment of ecosystem services, a combination of GIS and field data is recommended to improve accuracy in most analyses (Zheng et al., 2012) Therefore, the U.S. Department of Agriculture (USDA) Forest Service Inventory and Analysis (FIA) program data can be used for this type of analysis of some ecosystem services. The FIA data set contains database for the forest lands in all the U.S States including forest characteristics, such as extent, condition, and

PAGE 20

20 volume Several studies (Jenkins et al., 2003; Smith et al., 2004; Woodbury et al., 2007) have used this database for calculating fore st carbon stocks for example Objectives Given the above, t his study aims to identify and quantify the effects of ecosystem service drivers on land use and land cover change and the provision of ecosystem services in undeveloped areas in two watersheds l ocated in the State of Florida, the Lower Suwannee River and the Pensacola Bay watershed. To determine the effects of land cover change, two direct drivers were analyzed: urbanization and hurricanes, which constitute threats to stems (Leatherman and Defraene, 2006; Wear, 2002a) Urbanization through conver sion of natural ecosystems (Millennium Ecosystem Assessment, 2003) has been a key driver in the State, as a result of population growth, and because new urban areas are needed to allocate this new population. In Florida, according to Zwick & Carr (2006) between 2006 and 2060 the population of the State is projected to double. The second driver that will be analyzed are hurricanes, since according to several authors, one of the main consequences of climate change in Florida will be th and frequency (Florida Oceans and Coastal Council, 2009; Stanton and Ackerman, 2007) Therefore, the importance of understanding the possible consequences of these two drivers o n the provision of forest ecosystem services is becoming cr ucial. The following chapters addressed the two drivers previously mentioned: urbanization and hurricanes in two watersheds in the state of Florida, the Lower Suwannee River and the Pensacola Bay located in the Northeastern and Northwestern Florida respect ively. The 2 nd chapter analyzed the effects of urbanization on the provision of ecosystem services specifically carbon storage, timber volume and water

PAGE 21

21 yield by using the GIS spatial model InVEST to quantify the ecosystem services in a 2003 land cover and then compar ed it to a 2060 population distribution scenario (Zwick and Carr, 2006) The hypothesis and objectives we re: Hypothesis : Urbanization directly affects carbon storage, timber volume, and water yield. Carbon storage an d timber volume are negatively correlated with urbanization and water yield is positively correlated. Objective 1: Determine the effects of Land Use and Land Cover changes in undeveloped areas due to urbanization on the provision of ecosystem services such as carbon storage, timber volume, and water yield in the two watersheds before mentioned. Objective 2: For the years 2003 and 2060, to estimate the amount of carbon storage, timber volume, and water yield in the two watersheds based on the InVEST model an d the U.S.D.A Forest Inventory Data. The 3rd chapter used a decision tree to develop a framework to identify the risk of forest damage due to hurricanes and determined the potential loss of aboveground carbon storage and timber volume in high and low fores t damage risk zones The map was developed based on information from previous studies (McNulty, 2002; Oswalt and Oswalt, 2008; Stanturf et al., 2007) that determine d the main forest structure factors that increase the risk of hurricane damage to forests The objectives we re: Objective 1: Develop a hurricane framework based on a decision tree analysis, to identify high and low forest damage risk zones due to hurricanes based on the tree diameter and height, species t ype (e.g. softwoods and hardwoods), and distance to landfall. Objective 2 : Assess and map the potential loss of aboveground car bon storage and timber volume due to hurricanes in the two watersheds previously mentioned, based on the InVEST model and the U.S .D.A Forest Inventory Data in the high and low forest damage risk zones due to hurricanes. Finally, the 4 th chapter presented general conclusions including an analysis of the two drivers and their effects on ecosystem services, as well as alternatives to b etter include ecosystem services in decision making.

PAGE 22

22 Table 1 1 Categories of ecosystem services [ Adapted from Millennium Ecosystem Assessment. 2003. Ecosystems and Human Well being: A Framework for Assessment (Page 5, Figure 1). Island Press, Washington.] Categories Definition Examples Provisioning services Products obtained from ecosystems Food, fresh water, fuel wood. Fiber. Biochemical, genetic resources Regulating services Benefits obtained from regulation of ecosystem proce sses Climate regulation, disease regulation, water regulation, water purification, pollination Cultural services Non material benefits obtained from ecosystems Spiritual and religious, recreation and ecotourism aesthetic, inspirational, educational, sens e of place, cultural heritage Supporting services Services for the production of all other ecosystem services Soil formation, nutrient cycling, primary production Table 1 2. Multiple Ecosystem Services assessment tools [ Adapted from Waage, S., Stewart, E., and K. Armstrong 2008. Measuring Corporate Impact on Ecosystems: A Comprehensive Review of New Tools: Synthesis Report. Business for Social Responsibility, San Francisco, CA.] Model name Description Developer Type of results Artificial Intelligence for Ecosystem Services (ARIES) A computer model uses for estimating and forecasting ecosystem services provision and economic values in a specific area Ecoinformatics Conservation International, Earth Economics, and experts at Wageningen University Spatially explicit maps Ecosystem Services Review (ESR) A sequence of questions that helps managers develop strategies to manage risks and opportunities arising dependence on ecosystems World Resources Ins titute (WRI), Meridian Institute, and World Business Council for Sustainable Development (WBCSD) Risk and/or opportunit y analysis

PAGE 23

23 Table 1 2. Continued Model name Description Developer Type of results Integrated Valuation of Ecosystem Services and Trade offs (InVEST) A decision making aid to assess how distinct scenarios may lead to different ecosystem services and human well being related outcomes in particular geographic areas Natural Capital Project, a joint venture among Stanford In stitute for the Environment, The Nature Conservancy and World Wildlife Fund (WWF) Spatially explicit maps Multi scale Integrated Models of Ecosystem Services (MIMES) A multi scale, integrated suite of models that assess the true value of ecosystem service s, their linkages to human welfare, and how their function and value may change under various management scenarios Gund Institute for Ecological Economics Spatially explicit maps Natural Value Initiative (NVI) An evaluation benchma rk methodology for assessing biodiversity and ecosystem services related risks and opportunities in the food, beverage and tobacco sectors Fauna & Flora International, Brazilian business school FGV, and the United Nations Environment Programme Finance Init iative Risk and/or opportunit y analysis

PAGE 24

24 Figure 1 1. Framework for linkin g ecosystems to human wellbeing [ Adapted from de Groot, R.S., Alkemade, R., Braat, L., Hein, L., and L. Willemen. 2010 Challenges in Integrating the Concept of Ecosystem Ser vices and Values in Landscape Planning, Management and Decision Making (Page 264, Figure 2). Ecological Complexity ]

PAGE 25

25 CHAPTER 2 EFFECT S OF URBANIZATION ON T HE PROVISION OF ECOS YTEM SERVICES Background Land cover change affects ecosystem structure, process es and functions, and consequently their capacity to deliver ecosystem services (Daily, 1997; European Commission, 2009; Metzger et al., 2006; Millennium Ecosystem Assessment, 2003; Su et al., 2012) The changes in land cover are increasingly important, as they are projected to increase and accelerate in the next decades due to human activities (IPCC, 2001; Vitousek et al., 1997; Wear and Greis, 2002a) However, analyzing di fferent land cover change scenarios is not an easy task as land cover changes due to urbanization are determined by a variety of factors including socio economic and climate factors that can damage or modify natural ecosystems and determine their vulnerab ility to such changes (Metzger et al., 2006) One of the major threats to ecosystems is human population growth. World population is rapidly increasing and according to the U .S Department of Commerce, Census Bureau (2002) the population by 2050 is projected to be 9 billion. The result of this pop ulation rise will be more human influence s on these ecosystem services such as food, timber, and clean water, as well as changes in land cover (European Commission, 2009; Millennium Ecosystem Assessment, 2003; Vitousek et al., 1997) So, land cover change, driven by urbanization as well as agricultural expansion, deforestation, over exploitation of natural resources, and cl imate change (European Commission, 2009; Foley et al., 2005) have been identified as some of the most important drivers affecting terrestrial ecosystems (Reyers et al., 200 9; Vitousek et al., 1997) Although, land cover change can contribute to economic development, it is normally accompanied by

PAGE 26

26 degradation of ecosystem services due to unsustainable human practices resulting in a reduction of the benefits for future generat ions (Millennium Ecosyst em Assessment, 2003; Vitousek et al., 1997) The State of Florida is not an exception to this trend, according to Zwick & Carr (2006) the population will double by 2060. Urbanization, defined as the development of new urban ar eas from non urban lands, is one important human induced disturbance that affects natural ecosystems (Bengston et al., 2005; Zhang et al., 2012) More specifically, urbanization can alter the structure, composition processes and functions of the ecosystems, in doing so it can (Grimm et al., 2008; Xu et al., 2007) Urbanization is also predicted to alter significant areas of forest and ot her natural systems as well as agricultural areas in the next few years with forests being the most subject to fragmentations (Alig et al., 2004; Smail and Lewis, 2009; Wear and Greis, 2002a) According to Bengston et al. (2005) during the latter ha lf of the 1990s there was an increased concern in unplanned urbanization, especially because of the resulting environmental impacts. This unplanned urbanization was identified as the most significant factor affecting forest ecosystems in the southern U S (Wear, 2002b) since declin ed since 1992 (Wear and Greis, 2002a) Urbanization can alter both biotic and abiotic ecosystem properties (Grimm et al., 2008) and have significant impacts on (Xu et al., 2007) as well as other ecosystem s ervices such as clean water by increasing point and non point pollutants (Grimm et al., 2008; Wear and Greis, 2002a; Zheng et al., 2012) Carbon Storage and U rbanization It is estimated that urbanization has emitte d about 0.21 Pg c arbon between 1945 through 2007 in the Southern U S. (Zhang et al., 2012) However, urban areas can also

PAGE 27

27 store carbon in planted and reforested trees and can reach the pre urbanization amounts in 70 100 years (Zhang et al., 2012) According to Zheng et al. (2012) forest carbon sequestration can be reduced by 1.7 9.3% as a result of urban ization in the forested wildland urban interface areas in New England USA They also mentioned that carbon sequestration c an increase by 40% by 2030. Carbon storage in ecosystems is of interest because of climate change mitigation policies (Davies et al., 2011) The Kyoto Protocol for example pro poses to take carbon storage and sequestration as a valid mechanism for emission reduction to mitigate global climate change (Gorte, 2009; Newell and Stavins, 2000; Stavins, 1999) Forests are key for carbon storag e and critical to the global carbon cycle as they comprise a large amount of carbon comparing to other terrestrial ecosystems (Gorte, 2009; Harmon, 2001; Jandl et al., 2007; Yonavjak et al., 2011) Of the total car bon stored in terrestrial vegetation and soil s (2,200 Giga tons Gt), forests comprise over half (53%) of this value (Braatz, 2001) Gorte (2009) presented a comparison of carbon stocks for different land covers and the highest value corresponded to forests ranging from 68 182 tons per acre and the lowest to croplands with 37 tons per acre. In forests, carbon can be stored in the : soil, live tree, standing dead tree, forest floor, down dead wood, and understory vegetation (Harmon, 2001; Smith et al., 2004; Tawil, 2007) However, in some cases, only fore st and forest products are considered relevant carbon pools (Harmon, 2001) Southern forests represent 29% of the U.S. forests and play an important role in the sequestration of carbon at a global scale. Approxima tely one third of the carbon sequestered in U.S. forests corresponds to Southern forests, hence they can help in mitigating the effect of

PAGE 28

28 climate change (Hanson et al., 2010; Wear and Greis, 2002a) .Therefore, when forests are degraded or their area reduced the carbon stored is directly affected (Hanson et al., 2010) and when forests are degraded or converted ; the carbon stored is directly affected (Hanson et al., 2010) Timber and U rbanization Timber products are considered provisioning ecosystem service and are used to meet a variety of needs (Millennium Ecosystem Assessment, 2005) To consider timber as a provisioning ecosystem service, it should provide direct benefits to people (e.g. wood for furniture or fuel) (Millennium Ecosystem Assessment, 2005) and hence should be expressed in terms of production. The benefits of this ecosystem service are widely recognized and are more easily valued than other ecosystem services because market prices exis t for timber and non timber forest products (de Groot et al., 2010) US Southern forests have an important potential for timber production (Prestemon and Abt, 2002a) The South produces more than 60% of the timber products in the U.S. (Prestemon and Abt, 2002a) and according to Bystriakova et al. (2005) timber harvest at the global scale has been increased by 60% in the last four decades and the trend is to continue increasing but at a slower rate. Zhang & Nagubadi (2005) found that urbanization is one the drivers generating the decline in timberland area Cubbage et al. (1995) also found that urbanization reduced the timber supply especially that of softwoods. Additionally Hodges et al. (1998) found that in St. Tammany Parish i n Louisiana, forests were reduced by urban development. Water Yi eld and U rbanization Water is another important resource that provides services to human. Water enters ecosystems through precipitation and leaves through evaporation and runoff

PAGE 29

29 (Chapin III et al., 2002) Precipitation can be affected in different ways in terrestrial ecosystems. It can be: (1) intercepted by vegetation, (2) stored in the soils where it can be taken up by vegetation roots or discharge d into the aquifer and surface streams, (3) evaporated by soil and (4) transpired by veg etation. As water moves through the land, human activities and land use/cover types determine its physical and biochemical characteristics (Conte et al., 2011) Land use changes can alter water quantity and quality (Foley et al., 2005; Sun et al., 2005) Therefore, determining the effect of land use/cover on water yield and quality becomes important. Water yield, defined in this study as all annual precipitation that does not evaporate from soil and water or transpire s from vegetation (Mendoza et al., 2011) Forests can reduce runoff as a result of increased evapotranspiration and infiltration, as well as improve water quality and maintain str eam levels thus minimizing stormwater runoff (Hanson et al., 2010; Nasi et al., 2002) Several studies have determined the relationship between water and forests (Bosch and Hewlett, 1982; Sun et al., 2005) and urban areas. For example, w ater flow can change after urbanization since Hollis (1977) found that the median flow was increased 0.057 m 3 to 0.142 m 3 per se (1987) also analyzed the impact of urbanization on water yield in a Dallas, Texas, USA watershed, and found that it changed from rural to about 77 percent urban. Based on a model called Simulator of Water Resources in Rural Basins (SWRRB) they also calculated that annual surface runoff with no increase in urbanized area would be 135 mm rather than 151 mm about 10% increase Boggs and Sun (2011) compared t w o areas in North Carolina, one more urbanized and the other more forested, and concluded that urbanization can increase

PAGE 30

30 annual discharge volumes due to the decrease in evapotranspiration. Some studies were done in Florida, where Sun et al. (2008) fou nd that land use and land cover changes will have little effect on water quantity and the water demand relationship in 2020 in the southeastern U.S. They also mentioned that land use changes affect s water yield and evapotranspiration loss. Many studies in Florida have analyzed the effect of urbanization on water quality (Cooley an d Martin, 1979; Frick, 1998) and other consequences on water such as impact on insect communities (Jones and Clark, 1987) and water supply (Boggess, 1968) However more information about the impact of urbanization on water yield is lacking in the State of Florida. Ecosystem Services M odels The previously mentioned e cosystem s ervices can be quantified and mapped using Geographic Information S ystems ( GIS ) models (Boyd and Banzhaf, 2007; Metzger et al., 2006) Models such as the Integrated Valuation of Ecosystem Services and Tradeoffs (InVEST) use spatial data to estimate and map the quantity of ES (Tallis et al., 2011) The InVEST model is a GIS spatially explicit tool used for quantifying and valuating ecosystem services based on ecological production functions and economic valuation methods (Nelson et al., 2009; Tallis et al., 2011) The model can determine the quantity of ecosystem services in a current land use/cover and how these ecosystem services can change in the future with different scenarios such as new policies, and programs. Model outputs are in biophysical terms (e.g. tons of carbon stored) or economic terms (e.g. net present value of carbon sequestered) (Tallis et al., 2011) In this chapter, three ecosystem services were analyzed: two regulating services, (1) carbon storage as an indicator of climate regulation and (2) water yield as an

PAGE 31

31 indicator of flooding regulation, as well as one provisioning service, (3) timber volume as an indicator of timber production The objective s were to determine th e effects of urbanization on these three ecosystem services (carbon storage, timber volume, and water yield) in the Lower Suwannee River and Pensacola B ay watersheds in Northwestern Florida, respectively. This specific study aim was to quantify three key E S in a current and a modeled urbanized scenario using the InVEST model, the U S D epartment of A griculture (USDA) Forest Inventory and Analysis (FIA) field data and an urbanization land cover map for 2060 using the population distribution map produced by the GeoPlan Center at the University of Florida. The hypothesis tested was that urbanization directly affects carbon storage, timber volume, and water yield; c arbon storage and timber volume are negatively correlated with urbanization and water yield is po sitively correlated. This quantification of ecosystem services can then contribute to the valuation of ecosystems, as well as to understanding the magnitude of urban development on ES in subtropical Florida, thus promoting better understanding by landowner s urban planners and policy makers and for better assessment s of E S values. Methods Study A rea s The Lower Suwannee River and the Pensacola B ay watersheds in the State of Florida were selected for this study. The Lower Suwannee River watershed is located i n Northeastern Florida and the Pensacola Bay in Northwestern Florida (Figure 2 1). These two watersheds are different in terms of land use/cover characteristics with the Lower Suwannee River watershed being more forested and the Pensacola B ay being more ur banized (Table 2 1).

PAGE 32

32 The Lower Suwannee watershed consists of an area of 408,828 ha, and is one of five watersheds that comprise the Suwannee River watershed (Katz et al., 1997) The Lower Suwannee watershed is roughly center e d at 29 West and e ight counties are within the watershed boundary: Col umbia, Dixie, Gilchrist, Lafayette, Levy, Madison, Suwannee, and Taylor. This watershed was chosen for the analysis due to its hydrological nature as it comprises much of the Upper Floridian aquifer springs and the highest concentration in the world of fir st magnitude springs (Obreza and Means, 2006) The Suwannee River is the second largest river in Florida in terms of average discharge (Light et al., 2002) Several studies have been done in the Lower Suwa nnee, because in the past years the main threat to this area is the transition from natural ecosystems to more intense land use practices (Martin 2007) This area is defined by Obreza & Means (2006) as an area with unique biota and water resources that need to be protected. The Suwannee watershed area is characterized by the presence of karst features, wetlands, lowland topography, and a sm all number of tributary streams (Katz et al., 1997) Interactions betw een groundwater and surface water occur within the Suwannee River watershed due to the presence of karst features Five different ecological subsections are within its boundaries : 232Jf, 232Bj, 232Ka, 232Lc, and 232Db (Table 2 2) The land cover type in th e watershed is predominantly forest, agriculture, and wetlands (Katz et al., 1997) The Lower Suwannee River watershed is divided into 63 sub watersheds. The watershed has subtropical climate with warm summers and moderate winters (Katz et al., 1997) The average precipitation in the watershed is 1,270 mm per year (Obreza and Means, 2006)

PAGE 33

33 The Pensacola Bay watershed comprises an area o f 140,825 ha and is part of the Pensacola Bay system that is composed by five watersheds (Schwenning et al., 2007) The watershed is roughly center e d at 30 25 43 6 59 29 The counties of Escambia, Okaloosa, and Santa Rosa are within the watershed boundaries. This watershed includes the Pensacola Bay es tuary and parts of four major rivers: Escambia, Blackwater, Shoal and Yellow Rivers, and some small tributaries (Florida Department of Environmental Prote ction, 2007; Thorpe et al., 1997) The watershed was chosen because of its difference land cover characteristics compared to the Lower Suwannee River watershed. The Pensacola Bay watershed is more urbanized than the Lower Suwannee with around 30% of its a rea covered by urban areas. Three different ecological subsections are within its boundaries: 232Bj, 232Bg, and 232Lb (Table 2 2) The Pensacola Bay system has unique features such as its morphology and the hydrodynamics that are considered key points for the natural ecosystems (Hays, 2009) However, over the last decades it was impacted by human activities such as habitat alteration, urban development, and point and non point source pollution (Thorpe et al., 1997) such as industrial and domestic waste and non point sources from urban, suburban, and agricultural areas (Thorpe et al., 1997; U.S. Environmental Protection Agency, 2004) The most pollution in the watershed occurs around urban areas (Thorpe et al., 1997) Natural ecosystems such as wetlands and aquatic habitats have been affected by deposition and sedimentation (Thorpe et al., 1997) as well as through conversion and fragmentation (Northwest Florida Water Management District, 2006) Another threat to the watershed are hurricanes, which occur approximately every 5 to 10 years (NOAA

PAGE 34

34 National Hurricane Center, 201 2 a ) The Pensacola Bay watershed has humid subtropical weather with warm temperatures. The annual precipitation ranges from 730 to 2,280 mm (Hays, 2009) with an average of 1,625 mm (Thorpe et al., 1997) Quantification o f Ecosystem S ervices Carbon storage Carbon storage was analyzed using two data sets: the USDA Forest Inventory and Analysis (FIA) data and the Florida Vegetation and Land Cover 2003 (Florida Fish and Wildlife Conservation Commission, 2004) in addition to modeled urbanized land cover for the y ear 2060 (Zwick and Carr, 2006) Zheng et al. (2012) mentioned that the combination of field data (e.g. FIA) and remote sensing products (e.g. land cover maps) are necessary to improve carbon estimations accuracy. The InVEST model, devel oped for the Natural Capital Project (Natural Capital Project, 2011) was used for analy zing the amount of carbon storage in the 2003 land cover and th e 2060 urbanized land cover. These datasets were used as input data for the model, which estimates carbon storage in Mega grams (Mg) p er hectare (ha) for each land cover. Forest I nventory and Analysis Program data The U.S. Department of Agriculture (USDA) Forest Service Inventory and Analysis (FIA) program data was used for analyzing forest carbon storage. The FIA characterizes forest lands in all the U.S States including forest data and characteristics, such as extent, condition, and volume. The FIA Progra m started in 1929 with periodic forest inventories of nation wide network of plots in all U.S. States. Since 1999, those inventories are conducted annually (Jenkins et al., 2003; Woudenberg et al., 2010) The plot level information is based on t he individual tree measurement.

PAGE 35

35 One limitation with the FIA data is that exact geographic coordinates (latitude and longitude) of the plots are not provided bec plot integrity. However, t he location of the plots is with in a buffer of approximately 1 mile or mile from their true position. Moreover, 20% of the plots are swapped, which means that some private plots are exchanged with other similar private plot in the same county (Woudenberg et al., 20 10) This situation creates an uncertainty wh en plots need to be overlaid with other GIS data. Sabor et al. (2007) found that the potential error of using the altered location of the plots in combination with other layers, for example for a 30 meters by 30 mete r raster layer results in plot misclassification that ranged from 32% to 66%. However, different approaches, based on other FIA attributes, can be applied to better approximate true plot location. For example, only the 1 10 % of plots are misclassified wh en analyzing for ecological subsection data (Sabor et al., 2007) As such, includ ing the ecological subsection strata FIA survey unit and County attributes can better help specify the true location of plots. For example in the case of county, only the plo ts located within a specific county can be extracted f rom the FIA database (i.e. Tree table or Condition table). Carbon pools analysis As mentioned before, carbon storage analyses were performed based on the different forest types according to the 2003 Flo rida Vegetation and Land cover map produced by the Florida Fish and Wildlife Conservation Commission (Florida Fish and Wildlife Conservation Commission, 2004) This map was developed based on Landsat Enhanced Thematic Mapper satellite imagery with a 30 meter spatial resolution and detailed di stinction between different ecosystems. It contains 43 classes of vegetation

PAGE 36

36 and land cover, 26 being natural and semi natural vegetation types, 16 types of disturbed lands (e.g., agriculture and urban), and a single water class (Stys et al., 2004) For the carbon pool analysis, 11 forest cover classes were analyzed. Th e classes are shown in Table 2 1 as well as the presence or absence in each watershed. To deal with the problem of the true FIA plot location, a different approach was developed to try to approximate the carbon values for the different types of forests. So me attributes available at the FIA database were considered, specifically the FIA Survey Unit, ecological subsection, county, condition status, and stand age (Figure 2 2). The description of each attribute is shown in Table 2 3 The original 11vegetation c lasses in the map were reclassified using ArcGIS Spatial Analyst (ESRI, 2011a) to add new values to the raster file. New classes wer e created using the FIA attributes: Ecological subsection (U.S.D.A Forest Service ECOMAP Team, 2007) and County. Both vector files were converted to raster format and combined using ArcGIS Spatial Analyst. Th e result was one raster file showing the unique combination of classes for each forest ecosystem, ecological subsection and county (Figure 2 3 ) Once forest ecosystems were identified for each watershed, the main forest tree species for each of them were i dentified based on vegetation descriptions prepared by Gilbert and Stys (2004) that contains the list of forest tree species observed in each forest ecosystem. Then, the forest tree species found in the FIA database were considered and the species codes were identified (Table 2 4 ). The four carbon pools, measured in metric tons per hectare, were aboveground biomass, belowground biomass, soil organic matter and dead biomass (Tallis et al., 2011) Carbon abo veground biomass refers to all living materials above the soil (e.g.

PAGE 37

37 bark, leaves, trunks), belowground biomass corresponds to the living root systems of the aboveground biomass, soil organic matter comprises the organic component of the soil, and dead bio mass includes the litter, and downed and standing dead wood (Tallis et al., 2011) Table 2 5 shows the FIA attributes used for each carbon pool calculation. The FIA data corresponding to cycle 8 (2002 2007) were used in the analy sis as they correspond to the 2003 land cover classification. To extract the c arbon aboveground and belowground from the FIA data, the Tree table was used (Woudenberg et al., 2010) which contain ed individual tree information. A tree expansion factor, which can be found in the Tpa_unadj attribute in the Tre e table was used. The expansion factor was used to scale each tree on a plot to a per acre basis, considering the number of trees per acre (TPA) (Woudenberg et al., 2010) Per tree values for aboveground and belowground carbon were multiplied by the corresponding Tpa_unadj to get the plot level information in pounds per acre. Using the Pivot table function in Microsoft Excel (Microsoft, 2010) these values were summed by plot (PLT_CN) considering specie code (SPCD), FIA survey unit (UNITCD), ecological subsection (ECOSUBCD), county (COUNTYCD), condition st atus (COND_STATUS_CD), and stand age (STDAGE). The carbon value calculated was the total carbon aboveground and belowground for each plot. The total values were converted to tons per hectare t hen; the average was calculated using these plot values to get t he value for each species, considering the ecological subsection and county. ranging from 1 to 72 years were considered in the analysis. Based on a previous study where the exact locations of the FIA plots were obtained for the Forest Stewardship

PAGE 38

38 Properties located in the State of Florida (Abd Elrahman et al., 2012) the range of stand ages were extracted. To assign the carbon values to a raster file (i.e. reclassified land cover ) f rom the FIA plot locations the forest type were identified in both data sets so the carbon values in the FIA data for one specific forest type one ecological subsection, and one county were assigned to the each forest ecosystem in the raster file that co rresponded to the same attributes (i.e. ecological subsection and county). T he final carbon above and belowground for each forest ecosystem was the result of the sum of all the species within that land cover type. No timber harvest effects were considered in the analysis. For the down ed dead, standing dead, soil, understory aboveground, understory belowground, and litter carbon the approach used was similar to the approach described above. Since these C values are in the FIA Condition table that provides pl ot level information, the forest type attribute was used. The values for these carbon pools are by plot and presented in tons per acre, which were then converted to tons per hectare. Once all carbon values were calculated from the FIA data, the different v alues were aggregated into the four pools (Table 2 5 ) In table 2 6 the resulting carbon pool for the Pensacola Bay watershed are presented. The carbon pools for the urbanized 2060 scenario were estimated based on the assumption that the forest carbon pool in 2060 will be 5 percent less when compared to the pool in 2010 (Huggett et al., 2002) Several authors refer to the uncertainty in the forest carbon projections, so this 5% assumption was made (Heath and Smith, 2000; Huggett et al., 2002; Moore et al., 2002)

PAGE 39

39 InVEST c arbon storage model The InVEST c arbon storage model summed the amount of carbon stored in the four aforementioned pools based on the land cover map (Tallis et al., 2011) The carbon values were estimated for each pixel considering the previously mentioned land cover type s Model output is the total carbon storage, which sums the four pools. Intermediate results are the carbon aboveground, belowground dead, and soil (Tallis et al., 2011) Model output is expressed in Mg of carbon per pixel units and then converted to Mg per hectare. A flow chart illustrating the carbon estimation approach is shown in Figure 2 4 Timber volume The approach used for analyzing timber production is similar to the one used for the carbon estimation. T he timber volume and the timber in the sawlog portion were calculated and are basically the potential timber available for harvesting Timber volume w as processed using FIA survey unit, ecological subsection, and county. Only the condition status defined as forest land and the stand age ranging from 1 to 72 years were considered in the analysis. From the Tree table, two attributes were analyzed: the net volume (volnet) and the net volume in the saw log portion (volsaw). These attributes contain tree level information, so the expansion factor was used to convert to plot level data in cubic feet per acre. The final results were converted to cubic meter per hectare and volume values were assigned to individual pixels in the map. Figure 2 4 presents the timber estimation process flow chart adopted in this study. For the timber volume projections for 2060, the volume values presented in Prestemon and Abt (2002b) based on the Subregional Timber Supply Model (SRTS) of Abt et al. (2000) were used. An average per year increase in timber volume was used to determine the timber volume for the year 2060.

PAGE 40

40 Water yield estimation The InVEST water purification model estimates the annual average runoff (water yie ld) or the effect of land cover type on the annual water yield (Tallis et al., 2011) The model was used to estimate the water yield from each pixel in the form of precipitation minus the evapotranspiration fraction The model sum s the pixel water yield information to get the sub watershed level value and then the sub watershed values are summed to get the watershed level value (Tallis et al., 2011) The model is for surface water and does not account for groundwater (Mendoza et al., 2011) However, there is no differentiation between surface, subsurface and baseflow water in the InVEST model (Tallis et al., 2011) According to Zektser and Loaiciga (1993) groundwater comprises on average 30% of the stream flow around the world and in some areas can reach up to 90% (Winter, 1999) Therefore, in areas where groundwater is influential, the InVEST model can produce low water yield values when compared to observed stream f low levels (Mendoza et al., 2011) Figure 2 5 shows the conceptual framework of the model. This model is based on the Budyko curve and annual average precipitation (Tallis et al., 2011) The Budyko curve describes the relationship between precipitation and actual and potential evapotranspiration (Mendoza et al., 2011) This approach for calculating the evapotranspiration depends on water availability (i.e. precipitation) and atmospheric demand (i.e. potential evapotranspiration (Mendoza et al., 2011; Zhang et al., 1999; Zhang e t al., 2001) The annual water yield is estimated based on precipitation and actual evapotranspiration. The model analyzes each pixel considering each land cover ; Eq u ation 2 1 shows the calculation. (2 1)

PAGE 41

41 where Y xj is the annual water yield, AET xj is the annual actual evapotranspiration on pixel x with LULC j and P xj is the annual precipitation in pixel x with LULC j, and A xj is the area of x in LULC j. The evapotranspiration p artition of the water balance is also calculated. This is an approximation of the Budyk o curve presented by Zhang et al. (2001) who defined this equation by using potentia l evapotranspiration and plant available water content. The maximum difference in this ratio (AET/P) corresponds to forest cover areas, and in this case trees will use more soil water storage (Zhang et al., 2001) and is described as: (2 2) where R xj is the ratio of potential evapotranspiration to precipitation, named the Budyko dryness ratio and w xj is a modified dimensionless ratio of plant accessible water storage to expected precipitation during the year (Mendoza et al., 2011) Therefore, w xj is given by Eq uation 2 3: (2 3) where AWC x is the volumetric (mm) plant available water content, Z is the seasonality factor showing the rainfall distribution (Mendoza et al., 2011) .Finally, the ratio of potential evapotranspiration to precipitation (R xj ) is calculated by (2 4) where ETo x that is the reference evapotranspiration from pixel x and k xj is the evapotranspiration coefficient associated with LULCj on pixel x.

PAGE 42

42 Water model input dat a process The input parameters for the InVEST water yield m odel are summarized in T able 2 7 as well as parameter descriptions, units of measure, and data sources. T he process to get the data for each of the inputs needed for the water yield model is descr ibed in the next section, as well as the data source. Precipitation Annual precipitation data for the years 1998 2008, were obtained from the PRISM Climate Group (formerly SCAS) at Oregon State University and an average precipitation was calculated using the 10 year values For the year 2060 precipitation data, the projections of Coulson et al. (2010) based on the Australian Commonwealth Scientific and Industrial Research Organization CSIRO Mk3 Climat e System, were used The CSIRO Mk3 climate model is a representation of the components of the climate system: atmosphere, land surface, oceans, and sea ice with the main objective of investigating the climate system to be able to develop predictions (Gordon, 2002) Average values for the years 2055 2065 were used. Soil depth The source data was the U.S. General Soil Map (State Soil Geographic Statsgo2) developed by the USDA Natural Resources Conservation Service. The Soil Data Viewer 1 tool was used to produce the soil thematic maps. The tool generated spatial information directly in a vector format that was converted to raster format for use in the mode l. The maximum soil depth and water table depth were generated using this tool and the final input is the combination of these two data sets. Average annual potential e vapotranspiration (PET) The original data (organized by county in an ASCII format with 2 km spatial resolution) was obtained from 1 Soil Data Viewer is a tool built as an extension to ArcMap that allows a user to create soil based thematic maps

PAGE 43

43 U. S. Geological Survey's Florida Integrated Science Center (U.S.Geological Survey Florida Water Science Center, 2008) and converted to ra ster format. Preparing the PET data involved the following steps: (a) getting the annual data for each cell (original data is daily) for the years 1998 2008, (b) converting tabular data to spatial data using the latitude and longitude coordinates of each d ata point, and (c) converting point data to a raster format using the Inverse Distance Weighting (IDW) interpolation method. The IDW interpolation use d a weight to determine cell values, this weight is a function of inverse distance, and so closer data poi nts exert more influence than farther data points (ESRI, 2012a) The annual ras ter layers were averaged to get the average annual potential evaporation layer. For the 2060 PET, the projections of Coulson et al. (2010) based on the CSIRO model were used. An average value for the years 2055 2065 was used. Plant available water c ontent The InVEST model requires a GIS raster dataset with a plant available water content value for each cell. The source data is the General Soil Map STATSGO2 obtained from the USDA NRCS. The Soil Data Vi ewer tool was used to get the vector data set, which was converted to raster format and used in the model. Maximum root depth Values for vegetated land cover types, except for wetlands and grasslands species were collected from Canadell et al. (1996) To assign maximum root depth values to each land use/land cover type the main species for each land cover were identified. For the non v egetative land use/land cover classes such as, urban, extractive and Sand/Beach, a value of 1 was assigned as suggested by Tallis et al. (2011)

PAGE 44

44 Zhang c onstant This is a factor used to characterize the seasonality of precipitat ion in an area (Tallis et al., 2011) The Lower Suwannee River and Pensacola Bay watersheds are located in a subtropical ecoregion, where most rainfall occurs during the summer months similar to tropical ecoregions. According to Tallis et al. (2011) the value for tropical or sub tropical watersheds was 4; hence this value was used. Evapotranspiration coefficient This coefficient is needed when the reference evapotranspiration is used in the model; how ever if the PET is used, this coefficient is not required (Eq. 2.4), thus a value of 1 was assigned to all land cover types. InVEST w ater model assessment An assessment of the water yield model was done based on the comparison of the average annual water y ield from the model and the average annual streamflow data from the Gopher River gage station (Site 02323592) located at a downstream point in the watershed. A 10 year average value should be used for the stream flow data (Tallis et al., 2011) However, the streamflow data found at the U.S. Geological Service (USGS) National Water Information System (NWIS) ( http://waterdata.usgs.gov/nwis/ ) had information for the period 2000 to 2008 NWIS has information about the stream discharge for 8 sites along the Suwannee River (Nielsen and Norris, 2007) The assessment was done only for the Lower Suwannee River watershed 2060 l and use and l and cover map The 2060 urbanization scenario land use and land cover map was developed based on the 2003 Florida Vegetation and Land Use map and the 2060 Florida population distribution scenario (Zwick and Carr, 2006) The population distribution map for 2060 is based on the population projections f rom the Bureau of Economic and

PAGE 45

45 Business Research (BEBR) and GIS suitability analysis (Zwick and Carr, 2006) Based on population projections, the amount of people that need to be allocated can be identified, as well as the land area in hectares. The Florida 2060 map was developed by combining spatial data to model land use suitability considering conservation (forest), urban and agricultural lands. Three main concepts we re considered in the analysis: suitability, preference, an d conflict mapping (Carr and Zwick, 2007) Suitability refers to the fitness of one specific use (e.g. conservation, urban, and agricultural lands) in an area. Preference reflects the opinion of the community in other words, what the community wants Conflictive areas are the ones that cannot be urbanized for several reasons (e.g. forest conservation lands) (Carr and Zwick, 2007) Based on these three concepts the most suitable areas for urban development were identified by Zwick and Carr (2006) The final output was a raster file with the predicted urban ar eas for 2060. The 2060 Population distribution map was combined with the 2003 Florida vegetation map to update and produce the 2060 urbanization scenario Land cover map using ArcGIS Spatial Analyst For the carbon and timber model, the 2060 Florida vegeta tion map was used; this was the result of the extraction of the forest covers from the 2060 Florida vegetation and land use. This raster was combined with the ecological subsection and county layers. However, for the water yield model all land cover types were considered in the analysis. Statistical differences between the 2003 and the 2060 urbanization map at the sub watershed level were de termine d using a two tail t test with two paired samples for means a t 95% confidence level. A Pearson correlation coef ficient was calculated between carbon storage, timber volume, and water yield, as well as between each of

PAGE 46

46 them and the amount of forest areas. The analys e s w ere done at the sub watershed level and Microsoft Excel (Microsoft, 2010) was used for both calcul ations. Results Carbon S torage in 2003 and U rbanized 2060 In 2003, the total carbon storage in the Lower Suwannee River watershed was 36,586,750 Mg. The total carbon storage ranged from 72 to 342 Mg/ha with an average of 196 Mg/ha. However, in 2060, the to tal carbon storage was 30,638,160 Mg with a minimum and maximum value of 69 and 326 Mg/ha, respectively. The average total carbon for the watershed was 187 Mg/ha (Table 2 8 ). A reduction of 5,948,590 Mg of total carbon (19%) occurred due to urbanization T he highest carbon values in 2003 2060 correspond to the soil and aboveground pools (Table 2 9 ). In 2003, the total carbon aboveground was 10,193,638 Mg with an average value of 55 Mg/ha. In 2060, the total aboveground carbon was 8,467,580 Mg with an averag e of 52 Mg/ha. The highest average values of total carbon were concentrated in the northern part of the wat ershed (Table 2 9 and Figure 2 6 ), which corresponded to the mixed wetland forest and hardwood swamp in the Gulf Coastal Lowlands (232Db) and Flo rida Northern Highlands (232Ka) ecological subsections respectively. The lowest value corresponded to the sandhill ecological subsection 232Ka. There was a reduction in the mean carbon value due to urbanization across all 63 sub watersheds that ranged from 4 to 26 Mg/ha. The largest difference was observed in the Waccasassa Slough sub watershed (Id 49), where the mean carbon storage in 2003 was 183 Mg/ha, and in 2060 this value decreased to 156 Mg/ha with a reduction of 26 Mg/ha. In this sub watershed there wa s a decrease in forest area of 14% and an increase of 40% in urban areas. The largest changes were observed in the hardwood

PAGE 47

47 hammock forest s and the pineland s The smallest difference was observed in the sub watershed id 39 (Rock Bluff Spring), which had a forest area reduction of 7% and an increase of 18% in urban areas (Figures 2 6, 2 8 ). Differences in mean total carbon storage between sub watersheds for the period 2003 and 2060 were statistically significant at 95% confidence level (Table 2 9 ). In the Pe nsacola Bay in 2003, the total carbon storage in forest was 7,998,265 Mg with an average of 148 Mg/ha, with 66 and 229 Mg/ha as minimum and maximum values (Figure 2 7) However, in 2060, the total carbon storage decreased to 5,866,804 Mg with a minimum and maximum value of 64 and 218 Mg/ha, respectively. The average total carbon for the watershed was 137 Mg/ha. The forest area was reduced by 11,212 ha, with a resulting reduction of 2,131,4 61 Mg of total carbon (Table 2 8 ). The carbon aboveground in the wate rshed was 1,396,430 Mg, representing the 17% of the total carbon storage in the Pensacola Bay watershed (Table 2 9 ); however, in 2060 this value was 1,015,730 Mg. The highest total carbon was observed in the Bay Swamp within the ecological subsection denom inated 232Bg. The lowest value was for the Hardwood Hammocks and Forest in the ecological subsection named Florida Gulf Coast Flatwoods (232Lb). At the sub watershed level in the Pensacola Bay watershed there was a reduction in the mean total carbon in t he 15 sub watersheds from 3 to 18 Mg/ha The largest difference was observed in the Fundy Bayou Williams Creek Frontal sub watershed (Id 11), where the mean carbon was 162 Mg/ha in 2003 and 144 Mg/ha in 2060 with a reduction of 18 Mg/ha. Regarding the land use change in this sub watershed, the forest area was reduced by 34% and the urban area increased by 36%. The largest changes

PAGE 48

48 were found the Pineland areas (Figures 2 9, 2 10 ). A t test using two paired samples for means was conducted to find differences in mean total carbon between sub watersheds due to urbanization and differences were significant at 95% confidence level (Table 2 9 ). Timber V olume in 2003 and U rbanized 2060 The average timber net volume (volnet) for the Lower Suwannee in 2003 was 109 m 3 / ha and a total timber net volume of 20,402,395 m 3 approximately over an area of 186,275 ha of forest. However, for 2060, the average timber net volume was 110 m 3 /ha within an area of 163,692 ha, giving an approximate total net volume 18,100,416.54 m 3 Ther efore, the urbanization in the Lower Suwannee watershed produced a decrease of 2,301,978.47 m 3 in the timber net volume, as a result of the loss of 22,583 ha of forest ; this de spite of an increase in timber volume of 1.72% (Prestemon and Abt, 2002b) The timber volume in the sawlog portion, in 2003 was 15,4 77,193.37 m 3 with an average of 83 m 3 /ha However, i n 2060, the timber volume in the sawlog portion was 13,540,937.07 m 3 with an average of 83 m 3 /ha a nd a reduction of 1,936,256.24 m 3 of timber i n the sawlog portion (Table 2 1 0 ). Higher values were observ ed in the 2003 compar ed to 2060. The highest timber net volume corresponded to the northeast part of the watershed and to the Pinelands located in the Flor ida Northern Highlands (232Ka) ecological subsection At the sub watershed level in 2003 the largest decrease (17%) in mean timber net volume was for the Waccasassa Slough sub watershed (Id 49). This sub watershed had a decrease in forest area of 14% and an increase in urban areas of 40%. A t test using two paired samples for means was conducted to find t hat differences between the mean

PAGE 49

49 timber net volume in 2003 and 2060 were statistically significant at 95% confidence level (Table 2 1 0 ). For Pensacola Bay, the average timber net volume in 2003 was 69 m 3 /ha in an area of 54,381 ha and total timber net volu me was approximately 3,730,642 m 3 However, for the year 2060, there was a re duction of 11,514 ha of forest that produced a decrease of 800,909 m 3 of timber net volume. For sawlog timber volume, in 2003 the volume was approximately 2,980,362.85 m 3 with an average of 68.344 m 3 /ha. However, in 2060, the sawlog timber volume was approximately 2,309,369 m 3 (Table 2 1 0 ). The highest timber net volume corresponded to the Pinelands located in the ecological subsection named Florida Gulf Coast Flatwoods (232Lb). At the sub watershed level, in 20 60 there was a decrease in the mean timber net volume in 13 out of 15 sub watersheds. The largest difference (6 m 3 /ha) in the mean timber net volume was observed in the Williams Creek Oriole Beach Frontal sub watershed (Id 11 ). In this sub watershed there was a decrease of 34% of forest and an increase of 36% of urban areas. The t test s showed statistically significant differences between sub watersheds for 2003 and 2060 at 95% confidenc e level (Table 2 1 0 ). Water Y ield in 200 3 and Ur banized 2060 M ean annual water yield for the Lower Suwannee River watershed in 200 3 was 1,293 mm and water yield volume of 12,932 m 3 /ha. However, in 2060, the mean precipitation and PET as was mean annual water yield ( 1,342 mm ) (Table 2 1 1 ) and vol ume ( 13,422 m 3 /ha ) were greater than 2003 The highest values of water yield were located in the south part of the watershed and lowest values were located in the north part of the watershed, where most of the forest areas are located.

PAGE 50

50 At the sub watershed level, there were increases in water yield, precipitation, and PET annual water yield between 2003 and 2060 was observed in the Barnett Creek sub watershed (Id 1), where there was an increase of 152 mm/year and the largest increase in precipitation in 2060 but no changes in land use/cover. In the 63 sub watersheds there was more water yield per hectare in 2060 (Figure 2 8 ). The t test showed that the differences in water yield volu me were statistically significant at 95% confidence level (Table 2 1 1 ). In 2003 the Pensacola Bay the annual water yield was 1,520 mm and t he average water yield per hectare was 15,319 m 3 However, in 2060 the precipitation and PET were higher as were annu al water yield at 1,578 mm and water yield per hectare at 15,776 m 3 At the sub watershed level, the largest increase in mean annual water yield was observed in the Bayou Cico Bayou Garcon Frontal sub watershed (id 15) that registered the largest inc rease in precipitation in 2060 and there is a high positive correlation ( r= 0.969) between the difference in annual water yield and difference in precipitation (Figure 2 9 ). There was also a decrease in the forest area and an increase in the urban areas. The t te st showed statistically differences in the water yield volume between sub watersheds for 2003 and 2060 at 95% confidence level (Table 2 1 1 ). Water Yield A ssessment for the Lower Suwannee River W atershed R esults show that there we re differences between the average water volume from the Gopher gage station and the average water yield estimated from the model. The average value from the model was 5,184,553,019 m 3 and the average from the gage station was higher 6,722,705,162 m 3 (Table 2 1 2 ). However, when the values we re

PAGE 51

51 normalized the trend s between both curves ( i.e. model ed and measured water yield ) were similar (Figure 2 1 0 ). Tradeoffs Between C a rbon Storage, Timber V olume, Water Y ield and U rbanization Tradeoffs were evident between carbon storage, timber volume and water yield in both watersheds. In the 2060 urbanized scenario carbon storage and timber volume were lower than the 2003; however, water yiel d was higher in 2060 ( Figure 2 1 1 ). Also a positive correlation was observed between carbon storage an d timber volume, and a negative correlation with wat er yield (Table 2 13 ). Because the loss of forest areas in this study is the result of urbanization, it can thus be surmised that carbon storage and timber volume decrease when urban areas increase H owev er water yield increase when forest areas decrease or urban areas in crease, but the correlation was weakly negative (Table 2 1 3 ) Discussion The study hypothesis was accepted, thus urbanization ignoring the ecosystem service provision in urban areas dir ectly affects carbon storage, timber volume, and water yield. Carbon storage and timber volume were high negative ly correlated to urbanization in both watersheds and water yield was high positively correlated in the Lower Suwannee watershed and weak positi vely correlated in the Pensacola Bay watershed Results showed that urbanization will convert forest areas by the year 2060, and this will result into the decrease of the provision of carbon storage and timber volume and into an increase in water yield th at will statistically be signi ficant differences at 95% confidence level.

PAGE 52

52 Carbon S torage in 2003 and U rbanized 2060 Aboveground and soil carbon accounted for about 50 80% of the total carbon storage in both watersheds (Table 2 9 ) and stor e the largest amou nt of carbon in forests (Huggett et al., 2002; Lal, 2005; Moore et al., 2002) The mean total carbon storage for the Lower Suwannee and Pensacola Bay watershed were 196 and 148 Mg/ha respectiv ely for the year 2003 (Table 2 8 ). These values were within the range reported by Heath et al. (2011) for forest areas in the United States, where the mean was 162 Mg of carbon per ha, and ranged from 150 Mg/ha to 192 Mg/ha. According to Timilsina et al. (2011) carbon pools average for Florida forests were 165 Mg of carbon per ha in cold spots (i.e. clusters of lower carbon storage values) and 177 Mg of carbon per ha in hot spots (i.e. clusters of h igher carbon storage values) Lal (2005) mentioned an aver age of 60 130 Mg C/ha in temperate forests and 120 194 Mg of carbon per ha in tropical forests. However, carbon storage decreased by 16% and 26% in the 2060 land use scenario in the Lower Suwannee and Pensacola Bay watersheds, respectively. Escobedo et al (2009a) presented a comparison of average carbon storage by urban forests in Florida that ranged from 9.3 Mg per hectare in Miami Dade County, 27.4 Mg per hectare in So uthern Escambia County, and 30.8 Mg per hectare in Gainesville. Using this study's 2003 average forest carbon storage values and average carbon storage for Southern Escambia's urban forests (Escobedo et al., 2009a) the total c arbon storage for the P ensacola B ay watershed will be about 175.4 Mg/ha. Therefore, the amount of c arb on stored in the P ensacola B ay watershed 's urban forests is about 16% of the total carbon storage (i.e. 175 .4 Mg/ha). Using the former total c arbon storage value that includes both urban and natural forests, total c arbon storage in the urbanized 2060 scen ario in P ensacola B ay watershed decreased 26%

PAGE 53

53 when compared to the 2003 total carbon storage. So, even if carbon stored in urban forests were included in this analysis; there would still be an overall decrease in total carbon storage in the P ensacola B ay w atershed from 2003 to 2060. This reduction in C storage was a combination of the increased urbanization as well as the 5% reduction in the carbon pools consi dered in the analysis for 2060. Since the decreases in both study areas were greater than 5% urban ization and not the assumed 5% carbon reduction loss, was the main driver for this decrease In Pensacola Bay, the forest area losses due to urbanization were 21% and corroborates studies that show that l and use change as an important driver affecting car bon storage and sequestration in ecosystems (Birdsey et al., 2006; Moore et al., 2002; Reyers et al., 2009) as well as diseases pests, hurricanes, and fires (McNulty, 2002) Also urbanization is the dominant land use change that affect s southern forests ( Conner and Sheffield, 2006; Nowak et al., 2005; Wear and Greis, 2002b) and Wear and Greis (2002a) mentioned that 30 43 million acres of land are likely to be converted to urban land uses by 2060 in the Southern United States. At the sub watershed level, both watersheds showed reductions of carbon storage and the largest differences between 2003 and 2060 were observed in sub watersheds where forest ecosystems have decreas ed and urban areas increased. This indicates that urbanization generates forest losses and carbon emissions (Wear and Greis, 2002b) Huggett et al. (20 02) developed some scenarios assuming high and low urbanization and the highest amount of carbon storage in 2060 was for the low urbanization scenario, and the lowest carbon storage was for the high urbanization scenario. Th two watersheds were very different in terms of distribution of land

PAGE 54

54 cover in the year 2003 since Pensacola Bay was more urbanized than Lower Suwannee (30% and 6%) (Table 2 1), therefore the percentage of carbon storage in the Pensacola Bay was lower because of a higher percen tage of urbanization, and was thus the most impacted watershed in 2060. Timber V olume in 2003 and U rbanized 2060 The Lower Suwannee had 11% less timber volume in 2060 compared to 2003; however, the Pensacola Bay had a larger decrease (21%) (Table 2 1 0 ). As mentioned before, the Pensacola Bay is more impacted by urbanization. The same situation applied to the timber in the sawlog portion. At the sub watershed level, both watersheds reported the largest changes between 2003 and 2060 in areas where the forest area has been decrease and urban area has been increased. Two main factors are considered important in terms of forest loss: urbanization and timber prices, because the first one comprises land use change and the second one can increase levels of timber ha rvest (Huggett et al., 2002) According to Conner & Hartsell (2002) Florida has lost most of the forest areas due to urbanization; however, timber prices are projected to increase. Hodges et al. (1998) found that the forest area was reduced by urbanization b y 25% % and 21% in the Lower Suwannee and Pensacola Bay watersheds respectively (Table 2 1 0 ) Munn et al. (2002) mentioned that urbanization affects timber harvesting possibilities because of the decrease in the forest area. Water Y ield in 2003 and U rbanized 2060 In both watersheds the water yield was higher in 2060; as well as the precipitation and potentia l evapotranspiration. Based on data produced by t he CSIRO Mk3 model precipitation and evapotranspiration will increase in the next years and according to

PAGE 55

55 several studies (Alvarez et al., 2001; Florida Oceans and Coa stal Council, 2009; Furniss et al.) I ncrease in water yield is due to two main factors: increase in precipitation the primary driver of water yield (Dymond et al., 2012; Sun et al., 2005) and changes in land use /cover A high correlation factor (0.9) between precipitation and water yield can be observed In the case of changes in land use/cover, many studies proved that water yield is lower in forested areas (Bosch and Hew lett, 1982; Dymond et al., 2012; Hibbert, 1965) Bosch & Hewlett (1982) found that forests decrease water yield, for ex ample for deciduous hardwoods and coniferous there is a 25 mm and 40 mm change in water yield respectively per 10% change in cover. This value was close to the one obtained in this study; in the Lower Suwannee there was a 12% change in cover and a 49 mm ch ange in water yield. However in the Pensacola Bay there was a 21% change in forest cover and a 57 mm change in water yield. However, Swank et al. (2001) identified a 28% increase in water yield after forest harvesting due to reduction in total ecosystem evapotranspiration and the increase in r unoff. Even, Dymond et al. (2012) mentioned that there is a reduction between 30% 50% in the water yield bec ause of new forest areas. Since water yield is highly influenced by precipitation, the percentage of increase will depend on the characteristics of the site Swank et al. (2001) did the study in western North Carolina where annual precipitation ranges from 1,700 mm to 2,500 mm. In many sub wate rsheds, the increase in water yield was the result of the increase in precipitation, decrease in potential evapotranspiration, and decrease in forest areas (Dymond et al., 2012) Water Yield A ssessment The difference s in absolute values for the InV EST model and the gauging station were large in most analyzed years (Table 2 1 2 ). These differences could be related to

PAGE 56

56 the groundwater discharge, according to Katz et al. (1997) this watershed is characterized by the presence of karst features that enable the interaction between the groundwater and surface water. Therefore, groundwater is an import ant component in this watershed; however the InVEST model does not account for it. This situation can be considered a problem in the case where absolute values are important for example if the water will be used for irrigation, so in this case the model should be calibrated. Tradeoffs Between C arbon Storage, Timber Volume Water Y ield and Urbanization Tradeoffs arise when the increase in use of one ecosystem service can decrease the use of another service (Rodriguez et al., 2006) Therefore, when forests are managed for increased carbon storage, tradeoffs can arise as seen in the two studied watersheds. Carbon storage and timber volume w ere highly positive correlated (Table 2 1 3 ), so the increase in carbon storage results in an increase in timber volume (Figure 2 1 1 ). However if timber production is taken into account, this relationship was negative; an increase in timber harvesting will result in a decrease in carbon storage in live trees ( in situ ). There is also carbon storage in wood products (ex situ) and this pool can store considerable amounts of carbon (Seidl et al., 2007) T he ex situ carbon was not considered in this analysis. Timber harves ting can provide positive benefits in economic terms but also there are some negative effects in ecological terms because of the loss in forest area, such as release of carbon to the atmosphere, increase in flooding, and climate regulation (Dymond et al., 2012; Raudsepp Hearne et al., 2010) W ater yield was however, inversely related to carbon storage and timber volume (Table 2 1 3 ) (Dymond et al., 2012) There was an increase in water yield with decrease in forest areas and this can considered as both a negative and a positive thing

PAGE 57

57 depending on the characteristics of the watershed. For example if there is a demand for irriga tion water, less water yield is a negative thing; however, the increase in water yield can result in an increase in flooding and in soil erosion, because more water is moving faster over the surface (Dymond et al., 2012) In general terms, increase in provisioning services will result in a decrease in regulating services (Dymond et al., 2012; Raudsepp Hearne et al., 2010) Ur banization, as a result of this study is the responsible of the forest area losses. So, urbanization was negatively correlated to carbon storage and timber volume and positively correlated to water yield (Table 2 13). According to Escobedo et al. (2009a) the average carbon storage for urban forests in Southern Escambia County in the Pensacola Bay watershed is 27.4 Mg per hectare, so this carbon value is much lower when com pared to forest areas analyzed in this study. The key point is to reduce tradeoffs between ecosystem services, as well as between ecosystem services and urbanization. T his is complicated because the interactions between ecosystem services are complex (Dymond et al., 2012) in addition to the difficulty in finding the balance between the use of different ecosystem services (DeFries et al., 2004) DeFries et al. (2004) referred to the problem of finding the balance between human needs (e.g. new u rban areas) and ecological functions of ecosystems. Therefore, the starting point for assessing tradeoffs is to quantify ecosystem services and to know about the response of ecosystems to different land use changes (e.g. urbanization) (DeFries et al., 2004) This can be understood as the inclusion of ecological knowledge in the land use decisions (DeFries et al., 2004) One alternative p roposed by DeFries et al. (2004) and Gimona and van der Horst (2007) is to identify the neither

PAGE 58

58 positive or negative consequences in the ecosystem services. Then, areas where the ecosystem services will suffer less impact should be selected for new development. Conclu ding Remarks Urb anization reduces forest area thus affecting its functions and the provision of services, and consequently human well being. Even though, there are tradeoffs between ecosystem services and between human needs and ecological functions, some management alter natives can be used for reducing these tradeoffs. Some key challenges in terms of ecosystem management are related to better knowing about the ecosystem services in biophysical terms (e.g. Mg of carbon storage ) and to understand how ecosystem services are linked, their tradeoffs, as well as the effects of drivers on them. However, to develop alternatives for reducing these tradeoffs and consequently to produce less impacts on ecosystems and people is not simple. Further, spatial mapping and quantifying ecos ystem services is not an easy task. T he use of the Integrated and Valuated Ecosystem Services and Tradeoffs ( InVEST ) model the USDA FIA field data and other Geographic Information System (GIS) data analysis provided a viable method to spatially map ecosy stem services such as carbon storage and water yield despite some limitation. The advantage of the InVEST model is that is simple, so little data is needed and can be used in areas with poor data availability. However, there are also some limitations, for example the water yield model does not account for the groundwater, and this situation generates a large difference in absolute water yield values when comparing to real data from gauge stations. Nevertheless, as shown in Figure 2 9 the model can predict t rends in water yield very well and most importantly models can not reflect t he complexity of the real world, but they can approximate processes

PAGE 59

59 T he FIA field data provide information that can be used in several approaches such as for determining the value s for different carbon pools and timber volume in the study area H owever, data limitation is the exact plot position (latitude/longitude), plots displacement of plus or minus or 1 mile because of privacy rights. A different approach was used in this ana lysis for dealing with this limitation; some attributes from the FIA data such as FIA Survey Unit, County, and Ecological subsection were used for specifying carbon and timber data The potential of this FIA database is large and several analysis based on GIS could be done if the FIA plot geographic coordinates were available. Future research can include other ecosystem services in the analysis to better have a complete idea of the interactions between ES. One ecosystem can provide multiple services at the same time; therefore all of them need to be considered in the analysis. These services can be water quality, erosion control, food, recreation, among others. Similarly, the analysis of other direct and indirect drivers is recommended. A climate change sce nario could be very interesting, as well, as differe nt forest management scenarios. The results provided in this study are useful for land planners for balancing ecological information with human needs and to include this ecological information in land dec isions. Land planners would know the quantity of ecosystem services that can be lost when a forest area is converted into an urban area. This information should be part of the decision making processes. Results can also be used for maximizing human well be ing by maximizing all the services provided by ecosystems, as well as reducing the tradeoffs between them, and the valuation of ecosystem services. The first step for

PAGE 60

60 valuing ecosystem services is t o quantify the ir production functions ; the second one is t o assign a n economic value to these functions Knowing the value of the ecosystem services and the effects of drivers influencing them is important when developing payments for ecosystem services as a conservation tool. Payment for ecosystem services is be coming an important mechanism to achieve the balance between human and natural well being. F orecasting land use change will also show how human needs (new urban areas) can alter ecosystem and its services and how ecosystems respond to these changes.

PAGE 61

61 Table 2 1. Land cover distribution in the Lower Suwannee River and Pensacola Bay watersheds, Florida based on the Florida Fish and Wildlife Conservation Commission Number L and cover Type Pensacola B ay Lower Suwannee Area (ha) % Area (ha) % 1 Coastal Strand 1,034 0.7 ----------2 Sand/Beach 2,173 1.5 1 0 3 Xeric Oak Scrub 121 0.1 ----------4 Sand Pine Scrub (*) S 489 0.4 ----------5 Sandhill (*) S 15,018 10.7 3,352 0.8 6 Dry Prairie ----------259 0.1 7 Mixed Pine Hardwood Forest (*) H 3,745 2.7 10,561 2.6 8 Hardwood Hammocks and Forest (*) H 2,144 1.5 38,880 9.5 9 Pinelands (*) S 18,475 13.1 97,118 23.8 12 Freshwater Marsh and Wet Prairie 2,124 1.5 2,599 0.6 15 Shrub Swamp 217 0.2 8,579 2.1 16 Bay Swamp (*) S 5,969 4.2 187 0.1 17 Cypress Swamp (*) S 399 0.3 12,724 3.1 18 Cypress/Pine/Cabbage Palm (*) S ----------0.27 0 19 Mixed Wetland Forest (*) H 4,486 3.2 9,325 2.3 20 Hardwood Swamp (*) H 3,629 2.6 13,538 3.3 21 Hydric Hammock (*) H ----------590 0.1 23 Salt Mar sh 136 0.1 2,148 0.5 26 Tidal Flat 17 0 0.09 0 27 Open Water 48,597 34.5 6,706 1.6 28 Shrub and Brushland 1,922 1.4 43,910 10.7 29 Grassland 38 0 11 0 30 Bare Soil/Clearcut 412 0.3 30,694 7.5 31 Improved Pasture 10 0.01 74,320 18.2 32 Unimpro ved Pasture ----------2,841 0.7 35 Row/Field Crops 1,098 0.8 21,098 5.2 36 Other Agriculture 25 0 3,824 0.9 41 High Impact Urban 24,181 17.2 12,705 3.1 42 Low Impact Urban 4,365 3.1 12,843 3.1 43 Extractive ----------17 0 Total 140,82 5 100 408,828 100 Land cover used in the carbon storage and timber volume analysis; H, hardwoods; S, softwoods

PAGE 62

62 Table 2 2 Ec ological subsection (Ecol. Subs.) descriptions [ Adapted from McNab, W. H., Cleland, D. T., Freeouf, J. A., Keys, J., J.E., Nowa cki, G. J., and C. A. Carpenter. 2007. Description of Ecological Subregions: Sections of the Conterminous United States. U.S. Department of Agriculture, Forest Service. Washington.] Ecol S ubs Name Description Mean annual Temperature ( F) Annual Precipita tion ( mm ) 232Bg Gulf Coastal Plains and Flatwoods Section Flat topography. Natural vegetation consists of longleaf slash pine, loblolly shortleaf pine, and oak hickory cover types, with oak gum cypress along rivers 67 1 587 232Bj Gulf Coastal Plains and Flatwoods Section F lat topography. Natural vegetation consists of longleaf slash pine, loblolly shortleaf pine, and oak hickory cover types, with oak gum cypress along rivers 65 1 553 32Db Florida Coastal Lowlands Gulf Section G enerally flat topography. Vegetation is mostly longleaf slash pine and oak gum cypress cover types. 72 1 303 232Jf Southern Atlantic Coastal Plains and Flatwoods section Terrain is weakly dissected irregular Vegetation is mainly a mixture of loblolly shortleaf pine, longleaf sl ash pine, oak pine, and oak gum cypress cover types. 66 1 262 232Ka Florida Coastal Plains Central Highlands Section Terrain is hilly with excessively drained. Forests consist of longleaf slash pine and oak hickory cover types. Much of the original forest vegetation has been removed and replaced by highly cultivated citrus groves. Section is ecologically significant because environments are drier than adjacent units. 68 1 393

PAGE 63

63 Table 2 2 Continued Ecol S ubs Name Description Mean annual Temperature ( F) Annual Precipitation ( mm ) 232Lb Gulf Coastal Lowlands Section Flat topography. Vegetation is mostly longleaf slash pine and oak gum cypress cover types. 67 1 627 232Lc Gulf Coastal Lowlands Section F lat topography. Vegetation is mostly longleaf sla sh pine and oak gum cypress cover types. 68 1 473 Table 2 3 Description of the Forest Inventory and Analysis (FIA) attributes used in the carbon and timber analysis [ Adapted from Woudenberg, S. W., Conkling, B. L., O'Connell, B. M., LaPoint, E. B., Tur ner, J. A., and K. L. Waddell. 2010. The Forest Inventory and Analysis Database: Database Description and Forest Service, Rocky Mountain Research Station. Fort Collins, CO.] Attribute c ode Attribute FIA Table name Attribute description UNITCD FIA Survey unit Condition, Tree, Plot Survey units are groups of counties within each State PLT_CN Plot sequence number Condition, Tree, Plot Unique number to identify plots ECOSUBCD Ecological S ubsection Plot An area of similar surficial geology, lithology, geomorphic process, soil groups, sub regional climate, and potential natural communities COUNTYCD County Condition, Tree, Plot The identification number for a county, a governmental unit in a State COND_STATUS_CD Condition status Condition The condition status indicates the basic land cover (e.g. forest land) STDAGE Stand age Condition The average age of the stand SPCD Species code Tree A FIA tree species code FORTYPCD Forest description C ondition Forest type code

PAGE 64

64 Table 2 4 Forest tree species used in the carbon storage and timber volume analysis by Land cover type Scientific name F orest I nventory and A nalysis code Land Cover number 4 5 7 8 9 16 17 18 19 20 21 Acer rubrum 316 X X X Carpinus caroliniana 391 X X Carya glabra 403 X X Carya tomentosa 400 X X Chamaecyparis thyoides 43 X Cornus florida 491 X Fagus grandifolia 531 X X Fraxinus americana 541 X X Fraxinus caroliniana 548 X X Gordonia lasianthus 555 X Ilex cassine 7455 X X Ilex opaca 591 X X Juniperus silicicola 57 X Liquidambar styraciflua 611 X X X Magnolia grandiflora 652 X X Magnolia virginiana 653 X X X X Morus rubra 682 X X Nyssa aquatic 691 X X Nyssa sylvatica var biflora 693 X X X X Ostrya virginiana 701 X X Persea palustris 720 X Pinus clausa 107 X X 4: Sand Pine Scrub, 5: Sandhill, 7: Mixed Pine Hardwood Forest, 8: Hardwood Hammocks and Forest, 9: Pinelands, 16: Bay Swamp, 17: Cypress Swamp, 18: Cypress/Pine/Cabbage Palm, 19: Mixed Wetland Forest, 20: Hardwood Swamp, 21: Hydric Hammock

PAGE 65

65 Table 2 4 Contin ued Scientific name F orest I nventory and A nalysis code Land Cover number 4 5 7 8 9 16 17 18 19 20 21 Pinus echinata 110 X X Pinus elliottii 111 X X X X X Pinus glabra 115 X X X Pinus palustris 121 X X X X X Pinus seroti na 128 X X X X Pinus taeda 131 X X X X Quercus chapmanii 800 X Quercus falcate 812 X X Quercus geminate 800 X X X Quercus incana 842 X X X Quercus laevis 819 X Quercus laurifolia 820 X X X Quercus lyrata 822 X X Quercus marilandica 824 X X Quercus myrtifolia 800 X Quercus nigra 827 X Quercus shumardii 834 X X Quercus stellata var. margaretta 835 X X Quercus virginiana 838 X X X Sabal palmetto 912 X X X X X X Taxodium ascendens 222 X X X Taxodium distichum var distichum 221 X X X X X Tilia americana 951 X X 4: Sand Pine Scrub, 5: Sandhill, 7: Mixed Pine Hardwood Forest, 8: Hardwood H ammocks and Forest, 9: Pinelands, 16: Bay Swamp, 17: Cypress Swamp, 18: Cypress/Pine/Cabbage Palm, 19: Mixed Wetland Forest, 20: Hardwood Swamp, 21: Hydric Hammock

PAGE 66

66 Table 2 5 Carbon pools aggregation based on the information available in the Forest Invent ory and Analysis (FIA) Program database Carbon pools FIA Carbon data Aboveground biomass Carbon aboveground + c arbon understory aboveground Belowground biomass Carbon belowground + c arbon understory belowground Soil organic matter Carbon soil Dead biom ass Carbon litter + carbon standing dead + c arbon down dead Table 2 6 Reclassified land cover and c arbon pools for the Pensacola Bay watershed Raster value Original LC Eco sub s County C above C below C soil C dead LC name 1 9 232Bj 113 44.63 9.91 107.5 9 17.39 Pinelands 2 7 232Bj 113 63.91 13.93 49.84 11.82 Mixed Pine Hardwood Forest 3 8 232Bj 113 14.68 2.72 49.84 11.82 Hardwood Hammocks and Forest 4 20 232Bj 113 10.92 2.08 174.04 13.62 Hardwood Swamp 5 19 232Bj 113 55.27 12.17 121.21 18.15 Mixed Wet land Forest 6 9 232Bg 113 42.62 9.48 107.59 16.77 Pinelands 7 8 232Bg 113 24.05 3.57 49.84 8.54 Hardwood Hammocks and Forest 8 7 232Bg 113 66.50 13.32 49.84 8.54 Mixed Pine Hardwood Forest 9 20 232Bg 113 11.07 2.07 174.04 25.38 Hardwood Swamp 10 19 23 2Bg 113 68.13 15.28 125.20 18.61 Mixed Wetland Forest 11 16 232Bg 113 51.68 11.27 147.63 20.23 Bay Swamp 12 9 232Lb 113 31.90 7.06 121.21 18.59 Pinelands 13 7 232Lb 113 38.91 8.49 49.84 13.76 Mixed Pine Hardwood Forest 14 8 232Lb 113 10.29 1.80 49.84 1 3.76 Hardwood Hammocks and Forest 15 20 232Lb 113 3.23 0.49 174.04 5.06 Hardwood Swamp LC, Land cover; Eco subs, ecological subsection; C, carbon

PAGE 67

67 Table 2 6 Continued Raster value Original LC Ecol subs County C above C below C soil C dead LC name 16 19 232Lb 113 33.02 7.39 147.63 12.23 Mixed Wetland Forest 17 16 232Lb 113 22.64 4.94 147.63 12.23 Bay Swamp 18 5 232Bg 113 15.48 3.23 67.64 16.58 Sandhill 19 5 232Bg 91 14.77 2.88 67.64 14.85 Sandhill 20 9 232Bg 91 32.10 7.03 107.59 16.44 Pinelands 21 7 232Bg 91 50.53 10.80 49.84 12.87 Mixed Pine Hardwood Forest 22 16 232Bg 91 45.55 9.79 147.63 20.23 Bay Swamp 23 20 232Bg 91 26.03 5.38 174.04 15.16 Hardwood Swamp 24 19 232Bg 91 52.90 13.90 125.20 17.05 Mixed Wetland Forest 25 8 232Bg 91 19.27 3.62 49 .84 12.87 Hardwood Hammocks and Forest 26 5 232Lb 113 19.47 4.05 67.64 15.01 Sandhill 27 20 232Bg 33 5.44 0.94 174.04 20.27 Hardwood Swamp 28 19 232Bg 33 48.90 10.80 125.20 17.83 Mixed Wetland Forest 29 7 232Bg 33 116.30 21.48 49.84 16.00 Mixed Pine Ha rdwood Forest 30 9 232Bg 33 45.80 10.12 80.33 16.87 Pinelands 31 8 232Bg 33 62.66 11.78 49.84 16.00 Hardwood Hammocks and Forest 34 16 232Bg 33 10.95 2.07 147.63 20.23 Bay Swamp 36 17 232Bg 33 5.44 0.94 174.04 20.27 Cypress Swamp 37 17 232Lb 113 3.23 0.49 174.04 5.06 Cypress Swamp 40 5 232Lb 91 12.47 2.32 67.64 8.96 Sandhill 41 9 232Lb 91 18.97 4.08 121.21 20.65 Pinelands 42 20 232Lb 91 5.12 0.88 174.04 5.06 Hardwood Swamp 43 19 232Lb 91 40.48 9.12 100.77 22.42 Mixed Wetland Forest 44 8 232Lb 91 1 0.92 1.91 49.84 13.76 Hardwood Hammocks and Forest LC, Land cover; Eco subs, ecological subsection; C, carbon

PAGE 68

68 Table 2 6 Continued Raster value Original LC Ecol subs County C above C below C soil C dead LC name 45 7 232Lb 91 27.51 5.84 49.84 13.76 Mixed Pine Hardwood Forest 46 16 232Lb 91 12.72 2.70 121.21 22.56 Bay Swamp 47 17 232Bg 91 26.03 5.38 174.04 15.16 Cypress Swamp 48 17 232Lb 91 4.72 0.79 174.04 5.06 Cypress Swamp 51 4 232Lb 113 3.59 0.47 80.33 23.65 52 7 232Lb 33 31.68 6.93 49.84 13.76 M ixed Pine Hardwood Forest 53 9 232Lb 33 30.92 6.76 121.21 18.17 Pinelands 54 16 232Lb 33 33.04 7.14 121.21 18.17 Bay Swamp 55 8 232Lb 33 4.25 0.56 49.84 13.76 Hardwood Hammocks and Forest 56 20 232Lb 33 6.21 1.09 174.04 5.06 Hardwood Swamp 57 19 232Lb 33 47.72 10.57 100.77 21.60 Mixed Wetland Forest 58 17 232Lb 33 5.82 1.00 174.04 5.06 Cypress Swamp 59 4 232Lb 91 25.07 5.44 80.33 22.28 Sand Pine Scrub 61 4 232Lb 33 15.76 3.29 80.33 25.02 Sand Pine Scrub LC, Land cover; Eco subs, ecological subsecti on; C, carbon

PAGE 69

69 Table 2 7 Input data for the InVEST water yield model [ Adapted from Tallis, H., Ricketts, T., Guerry, A., Wood, S., Sharp, R., Nelson, E., Ennaanay, D., Wolny, S., Olwero, N., Vigerstol, K., Pennington, D., Mendoza, G., Aukema, J., Foster J., Forrest, J., Cameron, D., Arkema, K., Lonsdorf, E., Kennedy, C., Verutes, G., Kim, C., Guannel, G., Project. Stanford.] Data Description Units Data Source Soil depth GIS Raster Minimum of depth to bedrock and typical water table depth. mm Soil Survey Staff, Natural Resour ces Conservation Service, USDA (2006) Precipitation GIS Raster A non zero value for average annual precipitation for ea ch cell (1998 2008) mm PRISM Climate Group & Oregon State University (2008) Plant Available Water Content GIS Raster Fraction of water that can be stored in the soil profile that is available for the use of plants. Soil Survey Staff, Natural Resources Co nservation Service, USDA (2006) Average Annual Potential Evapotranspiration GIS Raster Potential loss of water from soil by both evaporation from the soil and transpiration by healthy Alfalfa, if sufficient water is available (1998 2008) mm U.S Geologica l Survey Florida Water Science Center (2008) Land cover GIS Raster A LC code for each cell (2003) Florida Fish and Wildlife Conservation Commission (2004) Watershed GIS Shapefile The boundary of one drainage area Natural Resources Conservation Service (2005) Sub watershed GIS Shapefile Subdivision of watersheds Natural Resources Conservation Service (2005) Maximum root depth (per land cover type) Maximum root depth for vegetated land covers mm Canadell et al. (1996) U.S. Department of Agriculture Forest Service

PAGE 70

70 Table 2 7 Continued Data Description Units Data Source Evapotranspiration coefficient (etk; per land cover type) The plant evapotranspiration considering the transpiration characteristics to modify the reference evapotranspiration, which is based on alfalfa Zhang Constant (per watershed) Corresponds to the seasonal distribution of precipitation. Tallis et al. (2011)

PAGE 71

71 Table 2 8 Carbon storage descriptive statistics for the Lower Suwannee (LS) and Pensacola Bay (PB) Florida wate rsheds in Mg/ha Watershed Year Forest area (ha) Min imum Max imum Range Mean Std LS 2003 186,276 72 342 270 196 56 LS 2060 163,692 69 326 257 187 53 P B 2003 54,074 66 229 163 148 40 P B 2060 42,862 64 218 154 137 38 Std, Standard deviation Table 2 9 Ca rbon storage in the four pools ( Mg ) for the Lower Suwannee River (LS) and Pensacola Bay (PB) watersheds, Florida Watershed Aboveground Belowground Soil Dead Total p LS 2003 10,193,638 (27.86%) 2,056,594 (5.62%) 21,136,310 (57.77%) 3,200,208 (8.75%) 36,586 ,750 (100%) 0.0005 LS 2060 8,467,580 (27.64%) 1,697,010 (5.54%) 17,848,700 (58.26%) 2,624,870 (8.57%) 30,638,160 (100%) PB 2003 1,396,430 (17.46%) 280,855 (3.51%) 5,504,360 (68.82%) 816,620 (10.21%) 7,998,265 (100%) 0.0009 PB 2060 1,015,730 (17.31%) 19 1,896 (3.27%) 4,040,830 (68.88%) 618,348 (10.54%) 5,866,804 (100%) p<0.05, significant differences Table 2 10 Timber volume values for the Lower Suwannee (LS) and Pensacola Bay (PB) watersheds, year 2003 2060 Watershed Timber volume Area (ha) Mean* Std Total ** p LS Volnet 2003 186,27 6 109.53 66.22 20,402,395 0.0005 LS Volnet 2060 163,692 110.58 67.92 18,100,416 LS Volsaw 2003 186,27 6 83.09 72.83 15,477,193 LS Volsaw 2060 163,692 82.72 72.33 13,540,937 PB Volnet 2003 54,074 68.60 54.95 3,709,552 0.0044 PB Volnet 2060 42,862 68.34 57.75 2,929,327 PB Volsaw 2003 54,074 54.80 40.96 2,963,514 PB Volsaw 2060 42,86 2 53.87 43.54 2,309,049 in cubic meter per hectare; ** in cubic meter; Std: standard deviation ; Volnet: net volume; volsaw: volume in the sawlog portion ; p<0.05, significant differences Table 2 1 1 Water yield for the Lower Suwannee (LS) and Pensacola Bay (PB) watersheds Watershed Year Mean* Mean Total volume** p LS 2 003 1,293 93,193,344 0.002 LS 2060 1,342 96,694,988 PB 2 003 1,5 20 104,635,694 0.0412 PB 2 060 1,577 108,440,687 in mm; ** in m 3 ; p<0.05, significant differences

PAGE 72

72 Table 2 1 2 Average volume in m3 for the Gopher gage station and the InVEST water yield model Year Volume at gage station Volume water yield model Differ ence 2000 4,052,867,141 4,340,245,073 287,377,932 2001 4,667,723,281 4,591,129,229 76,594,052 2002 3,181,523,048 5,674,599,116 2,493,076,068 2003 10,402,507,944 6,268,311,839 4,134,196,105 2004 9,642,874,632 6,166,135,117 3,476,739,515 20 05 12,636,723,568 5,895,177,138 6,741,546,429 2006 6,300,488,059 4,054,040,486 2,246,447,573 2007 3,637,303,035 4,367,875,584 730,572,549 2008 5,982,335,754 5,303,463,592 678,872,162 Average 6,722,705,162 5,184,553,019 1,538,152,143 Table 2 1 3 Pearson Co rrelation coefficients between c arb on storage, t imber volume, w ater yield and p ercentage (%) of forest change for the Lower Suwannee River and Pensacola Bay watershed s at the sub watershed level % of forest change Carbon Timber Water LS % of forest change 1 LS Carbon 0.981 1 LS Timber 0.986 0.981 1 LS Water 0.856 0.927 0.868 1 PB Forest a rea variation 1 PB Carbon 0.994 1 PB Timber 0.974 0.966 1 PB Water 0.143 0.129 0.201 1

PAGE 73

73 Figure 2 1. Study area: Lower Suwan nee River and Pensacola Bay watersheds Figure 2 2. Approach for the estimation of carbon pools and timber volume using the Forest Inventory and Analysis (FIA) program data Land Cover Forest Species or Forest type description FIA Survey Unit Ecological subsection County Condition status Stand age Map created using ESRI basemap Copyright Esri

PAGE 74

74 Figure 2 3 Reclassification of the original land covers to include the ecologic al subsection and county data.

PAGE 75

75 Figure 2 4 Flow chart for estimating carbon and timber in the Lower Suwannee River and Pensacola Bay Florida watersheds to approximat e the exact locations of the Forest Inventory and Analysis (FIA) plots Carbon and Timber estimates Forest Inventory and Analysis (FIA) data Land cover Map Four carbon pools by forest specie s (tons/ha) Land cover reclassification process Four carbon pools by Land cover (tons/ha) Final reclassified Land cover map Carbon values assigned to Land cover map I nVEST c arbon Model Model Output carbon in Mg/ha Net and sawlog timber volume by forest species Net and sawlog timber volume by Land cover Volume values assigned to Land cover map Output timber volume in m 3 /ha

PAGE 76

76 Figure 2 5 Conceptual framework for the InVEST water yield model [ Adapted from Tallis, H., Ricketts, T., Guerry, A., Wood, S., Sharp, R., Nelson, E., Ennaanay, D., Wolny, S., Olwero, N., Vigerstol, K., Pennington, D., Mendoza, G., Aukema, J., Foster, J. Forrest, J., Cameron, D., Arkema, K., Lonsdorf, E., Kennedy, C., Verutes, G., Kim, C., Guannel, G., Papenfus, M., Toft, J., Marsik, M., and J. Bernhardt. 2011. Invest 2.2.2 (Page 229, Figure 1). The Natural Capital Project, Stanford. ] Precipitation Evaporation Transpi ration Runoff Water availability Water yield = Runoff + Water availability Root depth

PAGE 77

77 A B Figure 2 6 Total Carbon storage in the Lower Suwannee River, Florida. A) Total carbon storage in 2003. B) Range of change in m ean total carbon between 20 0 3 and 2060 by sub watershed.

PAGE 78

78 A B Figure 2 7 Total c arbon storage in the Pensacola Bay w atershed. A) Total carbon storage in 2003. B) Range of change in m ean total carbon between 200 3 and 2060 by sub watershed

PAGE 79

79 Figure 2 8 Differences between 2060 and 2003 in the water yield per hectare for the Lower Suwannee River watershed Figure 2 9 D ifferences between 2060 and 2003 in the water yield (mean) and precipitation (ppt) for the Pensacola Bay watershed 0 20 40 60 80 100 120 140 160 180 1 3 5 7 9 11 13 15 17 19 21 23 25 27 29 31 33 35 37 39 41 43 45 47 49 51 53 55 57 59 61 63 Difference in water yield 2060 2003 (mm) Subwatersheds -100.000 -50.000 0.000 50.000 100.000 150.000 200.000 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 Difference (mm) Subwatersheds Difference in mean Difference in ppt

PAGE 80

80 Figure 2 1 0 Normalized curves showing differences between average water yield volume from the InVEST model and the Gage station data (act ual) for the Lower Suwannee River watershed 0 0.2 0.4 0.6 0.8 1 1.2 2000 2001 2002 2003 2004 2005 2006 2007 2008 Year Actual Model

PAGE 81

81 Figure 2 1 1 Carbon storage, timber volume, and water yield in 2003 (current) and 2060 for the Lower Suwannee (LS) and Pensacola Bay (PS) watersheds

PAGE 82

82 CHAPTER 3 ASSESSING AND MAPPIN G POTENTIAL LOSS OF ABOVEGRO UND CARBON STORAGE AND TIMBER V OLUME FROM HURRICANES Background Hurricanes are considered one of the major threats to forest ecosystems in the southeastern United States (U.S.) since t he frequency of a major hurricane making landfall in the eastern U.S. is every 2 to 3 years (McNulty, 2002) In many cases, the quantification of forest dam age is not taken into account because the damage is low compared to human lives and property losses (McNulty, 2002) However, the damage to forest s can be significant and large areas of forest ecosystems can be altered, as well as their functions and services; consequently affecting human well being (Kupfer et al., 2008; Wear and Greis, 2002b) Predicting the impacts of hurricanes on ecosystems then becomes important in terms of the developmen t of strategies for management, post hurricane response and restoration, salvage logging potential, and especially for long term ecosystem analys es (Kupfer et al., 2008) According to Hanson (2010) hurricanes, t ornadoes and other natural dist u r bances creat e periodic forest clearings One example is H urricane Katrina, a category 5 in the Saffir Simpson Hurricane wind scale that produced an estimated loss of 320 million trees in Louisiana and Mississippi in 2005 (Hanson et al., 2010) Florida is one of the most vulnerable areas to hurricanes (Leatherman and Defraene, 2006) and climate change projections suggest that hurricanes will increase in frequency as well as intensity in the next years (Stanton and Ackerman, 2007) According to Stanturf et al (2007) in th e southern U.S., the probability of having a hurricane event is greater over the next 40 years. This situation generates the need to better understand how a forest As

PAGE 83

83 su ch, h urricanes are considered an ecological driver affecting ecosystems since b y definition drivers are ecological or human factors that can affect ecosystem composition, functioning, and processes, increasing or decreasing the provisioning of ecosystem se rvices (Hanson et al., 2010; Millennium Ecosystem Assessment, 2003) Impacts of Hurricanes on F orests and E cosystem S ervices Hurricanes can affect forests by creating fragmentation, as well as different s tand ages and succession al stages (Foster and Boose, 1992) Myers et al. (1998) mentioned that severe hurricanes (e.g. Hurricanes Camille in 1969 David in 1979 and Gilbert in 1988 ) cause d forest mortality and forest canopy loss of up to 100%. Brokaw and Walker (1991) mentioned defoliation, uprooting, and snapping as the principal types of damage to forest due to hurricanes. Moreover, this effects of forest structure can influence ecosystem services such as the ability of a forest to store and sequester carbon, especially in terms of short term carbon storage (McNulty, 2002; Uriarte and Papaik, 2007; Worley, 2008) F or example, hurricanes can reduce the amount of carbon stored in the aboveground portions of trees; according to McNulty (2002) the 10% of the total annual carbon sequestered in U.S. forests is converted to dead wood due to hurricanes. F or hurricane Katrina 20% of annual forest carbon sequestration was lost, corresponding to 40 mi llion metric tons of forest carbon (McNulty, 2002; Worley, 2008) However, not all the carbon is released into the a tmosphere; carbon is transferred from the living to dead pool, as well as to the soil pool after some period of time (McNulty, 2002; Myers and van Lear, 1998) Frangi and Lugo (1991) found that after Hurricane Hugo impacted a flood plain forest in Luquillo Mountains of Puerto Rico, ten percent of the aboveground biomass was transferred to the forest floor. Tanner et al. (1991) mentioned that after Hurricane Hugo, the total litter generated by the hurricane

PAGE 84

84 was equivalent to two year s of litter fall. Besides the conversion of biomass from live to dead pool, Chambers et al. (2007) also mention ed the shift from larger tree size s to smaller ones thus resulting in lower biomass. To evaluate if hurricanes can have either positive or negative consequences on ecosystem services such as carbon storage several factors need to be analyzed such as the amount of carbon that is converted from living wood to dead wood, as well as how much is released to the atmosphere (McNulty, 2002) T he amount of wood that can be salvaged determine s how much carbon is lost by hurricanes. For example, McNulty (2002) mentioned that after Hurricane Hugo in 1989 only 13% of the timber was salvaged representing an amount of less than 9% of the carbon storage Because a forest requires approximately 15 20 years to recover after a hurricane, a large amount of carbon can be lost (Lugo, 2008; McNulty, 2002) On the other hand, hurricanes can eliminate older trees, thus favoring the establishment of younger tree s, and in doing so, the amount of carbon sequestered can increase because younger trees have more phot osynthetic activity (McNulty, 2002) The loss in timber volume is also important, si nce hurricanes can generate huge financial losses Stanturf et al. (2008) mentioned that the estimated loss in economic terms due to hurricane Katrina was about $1.4 to $2.4 billion and Jacobs (2007) estimated that the potential loss in forest volume due to Charley, Frances, Ivan, and Jeanne hurricanes was about 56,000,000 cubic meter of timb er Tanner et al. (1991) processes such as the reduction in the forest area that will also reduce evapotranspiration after a hurricane impact. They also mentioned that the nutrient

PAGE 85

85 cycling is altered because of all the litter and downed biomass added to the soil. Some authors mentioned that hurricanes can increase the risk of secondary damages. The increase in forest fuel s especially when some dry conditions occur, increases the pr obability of wildfires occurrence with the potential to release carbon into the atmosphere (Dale et al., 1998; Lugo, 2008; McNulty, 2002; Myers and van Lear, 1998; Uriarte and Papaik, 2007) With high levels of rai nfall the soil can be saturated producing a risk of flooding and erosion (Tanner et al., 1991) Other consequences of hurricanes are the increase in disease risk (Lugo, 2008; McNulty, 2002) and the invasion of some species, increasing the risk of the extinction of other species (Lugo, 2008) Horvitz et al. (1998) studied forest areas impacted by Hurricane Andrew in Florida and fo und that in terms of forest species regeneration, after the hurricane 28% of 90 species identified were invasive Factors Controlling Hurricane Damage on F orests The pattern of hurricane damage on forests is the result of the combination of meteorological, physiographic, biotic (Duryea et al., 2007; Foster and Boose, 1992; Oswalt and Oswalt, 2008) historical (e.g. previous natural disturbance) and stochastic processes (Foste r and Boose, 1992; Myers and van Lear, 1998) Myers et al. (2008) mentioned that the cumulative effects of hurricanes alt er forest structure and composition in the long term. The relationship between wind speed and duration with forest damage is not however clearly understood (Boose et al., 1994) S tudies mention that the most important facto r that determine a damage are stand characteristics such as tree diameter, height, age and species (Merry et al., 2010; Xi e t al., 2008) In the particular case of H urricane Katrina, damage to forest s was more related to stand conditions and site characteristics (Kupfer et al.,

PAGE 86

86 2008; Oswalt and Oswalt, 2008; Stanturf et al., 2007) On the other hand, studies also mentioned that the highest damage in forest s was the result of higher wind speeds (Boose et al., 1994; Duryea et al., 2007) and the distance to landfall. For example, Oswalt and Oswalt (2008) concluded that the damage to forest s decreased when the distance to landfall inc reased in other words, when the distance from the coastline (i.e. where hurricane reaches land) to inland areas increase d Additionally McNulty (2002) found no evidence of forest damage from hurricane s with hurricane categories below 3. Boose et al (1994) performed an analysis on the impacts of hurricanes on forests in Puerto Rico and concluded that the damage was strongly related with differences in elevation (below 600 m) and vegetation type (Colorado, palm, and dwarf forest types). Duryea et al (2007) mentioned that the damage to urban trees will depend on several factors such as species, wind speed, soil type, moisture, and depth, tree health, and topography. According to Bromley (1939) soils with hard clay horizons resulted in increased susceptibi lity to uprooting. Regarding stand characteristics, Duryea et al (2007) found that urban trees that grew in groups had 80% of possibility to survive after a hurricane impact, comparing to 70% for the individual, open grown trees. In regards to individual tree characteristics, Foster and Boose (1992) found that damages was influenced by forest height, composition, slope, and aspect In terms of forest type softwoods were most affected by hurricanes resulting in severe damage and hardwoods were less affected (Foster and Boose, 1992) However, Oswalt and Oswalt (2008) concluded that softwoods do not experience more damage than hardwoods. They also mentioned that there was a relationship between forest type and distance to

PAGE 87

87 landfall, as damage in hardwoods increase with the increase in distance to landfall. Further more softwoods experienced more damage in the zone where the distance to landfall was less (generally referred to as zone 1 closer to the coastline ). Kupfer et al (2008) included several variables in the ir analys e s of hurr icanes damage that can also play an important role in the determination of the patterns of damage in forests such as : wind speed, rainfall, topography, and floodplain condit ions (e.g. soil) and found that hurricane damage was a combination of several of t hese factors or variables. Assessing and M apping Hurricane Damage in F orests A combination of field, remote sensing climate (Boose et al., 1994; Kupfer et al., 2008; Oswalt and Oswalt, 2008) and historical data (Boose et al., 1994) are necessary to better estimate forest damage by hurricanes. Kupfer et al (2008) for example used aerial photographs for forest damage analysis, as well as rainfall data, wind speed, duration of hurricane, and wind direction, and the overall accuracy using a classification tree analysis was 71.5%. Jacobs (2007) developed a rapid response method for assessing hurricanes potential damage in fores ts based on the delineation of the damage areas using the hurricane track, wind speed, wind patterns, rainfall accumulation, and tidal surge areas, and forest damage based on U.S Department of Agriculture Forest Inventory and Analysis (FIA) data. Jacobs (2007) mentioned that most of the forest damage is observed immediately adjac ent to the windward side of the track. Further, Oswalt and Oswalt (2008) estimated the forest damage by hurricanes based on FIA data and the damage zones outlined by Jacobs (2007) Wang and Xu (2010) performed a comparison of different vegetation indices techniques (e.g. Normalized Difference Vegetation Inde x NDVI) from pre and post hurricane using Landsat 5 TM imagery and aerial photos to identify damaged forest s

PAGE 88

88 due to hurricane Katrina using the supervised classification method Based on an error matrix to assess accuracy, r esults showed an overall accurac y ranging from 51% to 86%. Chambers et al. (2007) quantified the carbon loss by Hu rricane Katrina using detailed field data (i.e. FIA), remote sensing analyses, and a model to estimate forest damage. Wang and Xu (2009) applied Geographic Inf ormation System (GIS) and logit regression to assess the influence of several factors (i.e. forest characteristics and site conditions) on pattern, severity and probability of damage due to hurricane Katrina. Other studies have used decision tree analys e s to assess hurrican es. Decision tree is an analytic method that can be used to select between different options or choices considering high or low risk, cost or influence using a graphic approach to compare different alternatives (Olivas, 2007) However, the decision comp rises some uncertainty. Li et al. (2009) applied a decision tree analysis and a set of d ecision rules to find the influence of four parameters: sea surface temperature, vertical wind shear, atmospheric water vapor, and zonal stretching deformation on tropical cyclone formation and intensity. Kupfer et al. (2008) also used a classification tree analysis to develop models for Hurricane Katrina based on storm meteorology, stand conditions, and site characteristics to predict forest damage. Howard et al. (1972) applied a decision tree analysis to make the decision of seeding or not a hurricane to reduce its intensity DeLoach and Dicke (2005) presented several questions for landowners that involve timber salvaged an d management decisions and developed a Timber Stand Salvage Decision Model to determin e salvag ing of damaged timber. However, few studies that have assessed the potential role of hurricanes as direct drivers of forest ecosystem services such as carbon stor age and timber. Furthermore,

PAGE 89

89 there are few hurricane related studies that identify and map high and low forest damage risk zones that can then be used to predict possible damages or conversely identify areas where pre hurricane management practices can be used to reduce hurricane damage to forests. Dale et al. (2008) mentioned to the possibility of manag ing forests to alter their vulnerability to hurricanes, by the selection of forest species, silvicultural techniques (e.g. thinning ), and genetic enginee ring. Post hurricane forest management includes debris and downed tree clean up to avoid other events such as wildfires and diseases resulting from the accumulation of biomass in the forest (Myers and van Lear, 1998) However, this kind of information is not available for Florida. Therefore, this study will develop a framework based on a decision tree analysis that can be used for assessing forest damages due to hurric anes with available field data. This information can also be used to help land managers understand the consequences of hurricanes on their forests in terms of loss of ecosystem services, as well as to predict the responses of ecosystems and forest/emergen cy management activities, to this kind of disturbances (Dale et al., 1998; Kupfer et al., 2008) Decision tree analysis can also be useful for determining areas with high probability of timber salvaged after a hurr icane. By knowing this information managers can be prepared for such events taking into account in their management plans, and consequently they can reduce losses to forests, property, and ecosystem services (Kupfer et al., 2008) In this chapter, the risk of forest damage by hurricane was assessed and mapped. Specifically, t potential los s of ecosystem services in terms of aboveground carbon storage and timber volume that can be affected by a direct driver hurricanes Another objective wa s to develop a decision tree analysis

PAGE 90

90 framework to identify zones of high and low forest damage risk by hurricane s using existing literature and a vailable plot level spatial data such as: tree diameter and height, type of species (e.g. softwoods and hardwoods), basal area, and distance to landfall (damage zone). Methods Study A rea s The highly forested Low er Suwannee River watershed in west Florida and the more urbanized Pensacola Bay located in Northwestern Florida were selected for this study (Figure 2 1; Table 2 1). The Lower Suwannee watershed consists of an area of 408,828 ha, and is one of five watersh eds that comprise the Suwannee River watershed (Katz et al., 1997) Ei ght counties are within the watershed boundary: Columbia, Dixie, Gilchrist, Lafayette, Levy, Madison, Suwannee, and Taylor. The Suwannee river is the second largest river in Florida in terms of average discharge ( Light et al., 2002) Several studies were done in the Lower Suwannee watershed due to the threat of the transition from natural ecosystems to more intense land use practices (Martin, 2007) and the area has been defined by Obreza and Means (2006) as one with unique biota and water resou rces that need to be protected. The Suwannee watershed is also characterized by the presence of karst featur es, wetlands, lowland topography, and a small number of tributary streams (Katz et al., 1997) Interactions between groundwater and surface water occur within the Suwannee River watershed due to the presence of karst features. The land cover type in the watershed is predominantly forest, agriculture, and wetlands (Katz et al., 1997) and has 63 sub watersheds. The watershed has subtropical weat her with warm summers and moderate winters (Katz et al., 1997) The av erage precipitation in the watershed is

PAGE 91

91 1,270 mm per year (Obreza and Means, 2006) The Lower Suwannee River watershed was affected by several hurricanes. A ccording to Light et al. (2002) the hurricanes impacting the watershed were: Alma (1966), Agnes (1972), Josephine (1996), Earl (1998), Opal (1995), and Georges (1998). The Pensacola B ay watershed comprises an area of 140,825 ha and is part of the Pensacola Bay system that is composed by five watershed s (Schwenning et al., 2007) The counties of Escambia, Okaloosa, and Santa Rosa are within the watershed boundaries. This watershed includes the Pensacola B ay estuary and parts of four major rivers: Escambia, Blackwater, Shoal and Yellow Rivers, and some small tributaries (Florida Department of Environmental Protection, 2007; Thorpe et al., 1997) The watershed was chosen because of its differen t land cover characteristics compared to the Lower Suwannee River watershed. Th e Pensacola B ay watershed is more urbanized than the Lower Suwannee with around 30% of its area covered by urban areas. The Pensacola B ay system has unique features such as its morphology and the hydrodynamics that are considered key points for the natural ecosystems. However, over the last decades it was impacted by human activities such as habitat damage, urban development, and point and non point source pollution such as industrial and domestic waste and non point source pollutants from urban, suburban, and agricultural areas (Thorpe et al., 1997; U.S. Environmental Protection Agency, 2004) Natural ecosystems such as wetlands and aquatic habitats have been affected by urban related deposition and sedimentation (Thorpe et al., 1997) as well as through conversion and fragmentation (Northwest Florida Water Management District, 2006) The Pensacola

PAGE 92

92 B ay watershed has a humid subtropical weather and annual precipitation ranges from 730 to 2,280 mm (Hays, 2009) with an average of 1,625 mm (Th orpe et al., 1997) Pensacola Bay is affected by hurricanes approximately every five to ten years (NOAA National Hurricane Center, 2012 a ) The last major hurricanes in the area were Ivan in 2004 and Dennis in 2005 (Hays, 2009) According to NOAA National Hurricane Center (2012a ) the hurricanes that affected Pensacola B ay watershed were: Eloise(1975), Fredrick (19 79), Elena (1985), and Opal (1995). Quantification of Aboveground Carbon S torage and T imber V olume Aboveground c arbon storage and timber volume were estimated using the USDA Forest Inventory and Analysis (FIA) data and the Florida Vegetation a nd Land Cover 2003 (Florida Fish and W ildlife Conservation Commission, 2004) Both data sets were input into the InVEST model, developed for the Natural Capital Project to analy zi ng the amount of aboveground carbon storage (Natural Capital Project, 2011) Forest Inventory and Analysis Program D ata The USDA Forest Service Inventory and Analy sis (FIA) program data was used for analyzing forest carbon storage. Several studies (Jenkins et al., 2003; Smith et al., 2004; Woodbury et al., 2007) have used this database for calculating forest carbon stocks. T he FIA data set contains database for the forest lands in all the U.S including forest characteristics, such as extent, condition, and volume. The FIA Program started in 1929 with periodic forest inventories of nation wide network of plots in all the U.S. and since 1999, the se inventories are conducted annually (Jenkins et al., 2003; Woudenberg et al., 2010) The plot level data is based on individual tree measurement. One limitation with the FIA data is that exact geographic coordinates (latitude and

PAGE 93

93 Congress in the Food Security Act of 1985 to ensure private landown plot integrity. The location of the plots is approximately with a buffer of 1 mile or mile from the exact position. Moreover, 20% of the plots are swapped, which means that some private plots are exchanged with other similar private plot in the same county (Woudenberg et al., 2010) This situation creates an uncertainty while plots need to be overlaid with other GIS data. Sabor et al. (2007) found that the potential error of using the altered location of the plots in combination with other layers, for example for a 30 meters by 30 meter raster layers plot misclassification ranged from 32% to 66%. However, different approaches, based on other FIA attrib utes, can be applied to better use the data and approximate the location. For example, only the 1 10 % of plots were misclassified using ecological subsection data (Sabor et al., 2007) To deal with the problem of the precise FIA plot location, a different approach can be developed to approximate the aboveground carbon and timber volume for the different forest types. Some attributes available at the FIA database used for this the FIA Survey Unit, ecological subsection, county, condition status, and stand age (Figur e 2 2). To extract the aboveground c arbon and timber volume from the FIA data, the Tree table was used (Woudenberg et al., 2010) which contains individual tree data A tree expansion factor, which can be found in the Tpa_unadj attri bute in the Tree table was used to scale each tree on a plot to a per acre basis, considering the number of trees per acre (TPA) (Woudenberg et al., 2010) Per tree values for aboveground carbon and timber volume were multiplied by the corresponding Tpa_unadj to get the plot level da t a

PAGE 94

94 Using the Pivot table function in Microsoft Excel (Microsoft, 2010) these values were summed by plot (PLT_CN) considering specie code (SPCD), FIA survey unit (UNITCD), ecological subsection (ECOSUBCD), county (COUNTYCD), condition status (COND_STATUS_CD), and stand age (STDAGE). The aboveground carbon and timber volume for each plot was calculate d Then, the average was calculated using these plot data to get the value for each species, considering the ecological subsection and county. To assign the aboveground carbon and timber volume to a raster file (i.e. reclassified land cover) from unidentif ied FIA plot locations, the forest species were identified in both data, so the carbon values in the FIA data for one specific forest specie, one ecological subsection, and one county were assigned to the each forest ecosystem in the raster file that corre sponded to the same attributes (i.e. ecological subsection and county). Florida Vegetation and Land C over 2003 The 2003 Florida Vegetation and Land cover map was developed based on Landsat Enhanced Thematic Mapper satellite imagery with a 30 meter spatial resolution and had detailed distinction between different ecosystems. It contain ed 43 classes of vegetation and land cover, 26 being natural and semi natural vegetation types, 16 types of disturbed lands (e.g., agriculture and urban), and a single water cl ass (Stys et al., 2004) For the development of the hurricane fram ework 11 forest cover classes were selected and these classes were reclassified in to two main tree types for analyses: softwoods and hardwoods (Table 2 1). Urban forests (i.e. all trees and other associated resources within urban areas), we re not consider ed in the analysis.

PAGE 95

95 Decision tree F ramework F orest damage risk by hurricanes in this study considered the following as the main factors behind forest damage: tree diameter, height, basal area, tree type (softwoods and hardwoods), and distance to landfall (Merry et al., 2010; Oswalt and Oswalt, 2008; Xi et al., 2008) The reclassified land cover and main factors were combined to develop a forest ecosystem map to identified forest damage risk zones. The tree diameter and height, as well as the basal area, and type of species (e.g. softwoods and hardwoods) were extracted from the FIA database. Mean values for each ecological subsection and county were calculated for diameter, height, and basal area considering softwoods or hardwoods. The mean tree diameter and the mean tree height were extracted from the Tree table that contains per tree level data; however the mean basal area and the tree type (e.g. softwoods and hardwoods) were extracted from the Condition table that c ontain per plot level data. The FIA data corresponded to cycle 8 (2002 2007) were used in the analysis because the land cover corresponds to 2003. To deal with the problem of the precise plot location, a different approach was used to extract the mean tree diameter and height, as well as the basal area from the FIA database. The FIA attributes considered to identify the plots that are within the boundaries of the study area were: FIA Survey Unit, ecological subsection (U.S.D.A Forest Service ECOMAP Team, 2007) county, condition status, and stand age The FIA survey, the ecological subsection, and the county were used to approximate the location of the plots. However, the condition status and stand age were used to define the characteristics of the forests. Only plots in the study area with the condition status considered in the analysis. The description of each attribute is shown in Ta ble 3 1

PAGE 96

96 The FORTYPCD attribute from the Condition table in the FIA database was used to classify the tree type in the database and to better extract the information and assign to the previously created raster file The basal area of live trees on the plo t (BALIVE) was also extracted from the condition table and information was added to the Tree table that contains tree diameter (DIA) and total height (HT) (Woudenbe rg et al., 2010) The average values for DIA, HT, and BALIVE according to tree type, plus ecological subsection and county were then calculated. The DIA, HT and BALIVE were classified in two categories: small and large short and tall and low and high, r espectively (Table 3 2 ) This classification was done using the ArcGIS 10 so ftware and the Natural breaks ( i.e. Je nks) method. The natural breaks are based on natural groupings inherent in the data. Class breaks are identified that best group similar valu es and that maximize the differences between classes (ESRI, 2011b) The distance to landfall, defined as the intersection of the surface center of hurricane with a coastline (NOAA National Hurricane Center, 2012 b ) was used for determining the two forest hurricane damage zone as : high and moderate low. Specifically using 7 98) the wind damage zone s were de termined according to distance to hurricane landfall (Peacock et al., 2005) The ASCE 7 98 shows a Florida map of wind risk assessment associated with hurricanes where the maximum wind speed for a given area is expected to affect building construction standards (Figure 3 2). The ASCE 7 98 was used because it presents the average worst case situation, in terms of the maximum wind speed that can affect a specific are a in Florida T his map can the refore

PAGE 97

97 be used to delineate different hurricane damage zones by taking into acc ount these worst case wind speeds. For the study area, two areas of damage were identified based on the ASCE 7 98. One area had a shorter distance to landfall and a high wind speed: 13 miles of distance and 120 miles per hour for the Lower Suwannee and 2 miles of distance and 140 miles per hour for Pensacola B ay ; this area was identified as h igh hurricane damage zone (zone 1) T he farther area and l ower wind speed was identified as Moderate Low hurricane damage zone (zone 2) and had a distance ranging from 13 to 65 miles and a wind speed of 100 110 miles per hour for the Lower Suwannee and for Pensacola B ay with a distance from 2 to 23 miles with 130 miles per hour of wind speed. The two class vegetation map (softwoods and hardwoods) was then combined with the Ecological subsection and county raster files, a s well as with the hurricane damage zone s based on the ASCE 7 98. The result of this analysis was one raster file showing the unique combination of classes for each type of forest ecosystem: hardwoods or softwoods, ecological subsection, county and hurricane damage zone. The final raster used to develop the risk of forest damage due to hurricanes i n the decision tree analysis contained : 1) T ree type: Hardwood or Softwood, 2) H urricane damage zone: high (zone 1) and moderate or low (zone 2), 3) E cological subsection, 4) C ounty, 5) FIA Survey Unit, 6) M ean diameter: small (code = 2) and large (code = 1), 7) M ean total height: short (code = 2) and tall (code = 1), and 8) B asal area: high (code = 1) and low (code = 2). A decision tree scenario was developed to identify areas with high and low risk of forest damage from hurricanes (Figure 3 1) based on i nformation presented in Oswalt

PAGE 98

98 and Oswalt (2008) Stanturf et al. (2007) Merry et al. (2010) Xi et al. (2008) and Kupfer et al. (200 8) Forest damage, for the purpose of this study, is some degree of alteration of forests (i.e. plot level, individual tree level). According to Oswalt and Oswalt (2008) the hardwoods are more affected by hurricanes when the distance to landfall increase (i.e. zone 2), and softwoods are more affected when they are closer to the coast (i.e zone 1 ). For tree diameter, Oswalt and Oswalt (2008) report that when the diameter is smaller in softwoods and larger in hardwoods they were more affected due to hurricanes In terms of total height when both softwoods and hardwoods are taller they are most affected by hurricanes For basal area, both softwoods and hardwoods are more affec ted when the basal area is lower (i.e. low density) (Duryea et al., 2007; Escobedo et al., 2009b; Oswalt and Oswalt, 2008) Specific size, height, and density criteria and respective units are provided in Table 3 2 Based on this information and using the final raster containing all the decision tree factors previously described the high and low forest damage risk zones were selected using the C onditional F unction in ArcGIS Spatial Analyst. The conditional functio n allows the selection of severa l attributes from a raster file. By doing so, the areas with high and low forest damage risk zones can be identified based on the five factors analyzed. The expression used in the conditional function can be found in T able 3 3 In zone 1 (high damage) the most affected type of tree species was s oftwoods, with : smaller diameters (code = 2), great heights (code = 1), and lower densities (code = 2). However, in zone 2 (moderate or low damage) the most affected were hardwoods wit h larger diameter (1), greater height s (1), and lower densit ies (Oswalt and Oswalt, 2 008) A map showing the high and low forest damage risk zones due to hurricanes was

PAGE 99

99 created for each watershed and t he flow chart for the decision tree analysis is shown in Figure 3 3. A sub watershed analysis was also done to identify the sub watersheds with high percentage of high forest damage risk zones Therefore, based on the total forest area of each sub watershed, the percentage of area in both high and low forest damage risk zones was calculated. Potential L oss of A boveground C arbon S torage and T i mber V olume Using the aboveground carbon storage, timber net volume and timber in the sawlog portion results based on the 2003 Florida Vegetation map and the USDA FIA data ( C hapter 2) aboveground carbon storage and timber volume for each zone (i.e. high a nd low forest damage risk ) were assessed and mapped The high low forest damage risk map resulting from the decision tree analysis was used with the 2003 aboveground carbon storage and timber volume map to calculate the potential amount of loss of these tw o ecosystem services in each of these zones. Later, using the Zonal statistics as a Table in ArcGIS Spatial Analyst the carbon storage and timber volume raster values were summarize within each forest damage risk zone (e.g. high and low) to get total abo veground carbon storage and timber volume estimates in each zone (ESRI, 2012b) Results Potential Loss of A boveground Carbon S torage D ue to H urricanes In 2003 the aboveground carbon stor age was 10,193,638 Mg for the Lower Suwannee River watershed. Considering the worst case scenario w here hurricanes potentially impact all the forest areas in the watershed the approximate amount of potential loss of carbon aboveground can be up to 10,193,638 Mg. Highest average potential losses were located in the northeastern part of the watershed in t he Suwannee

PAGE 100

100 County (Figure 3 4). At the sub watershed level, the current carbon aboveground ranged from 22 to 115 Mg/ha, corresponding the highest average value to Rocky Hill Tower sub watershed that is composed by 54% of forests distributed in 38% of hard woods and 16% of softwoods. Table 3 4 shows the potential aboveground carbon storage for the identified high and low forest damage risk zones due to hurricanes for the Lower Suwannee River watershed. Most of the areas with high risk were located in the No rtheastern part of the watershed (Figure 3 5). The area corresponding to the high risk was 23,840 ha (13%) and the low risk area was 162,435 ha (87%). The aboveground carbon in the high forest damage risk zone included 31% of the total estimated for the wa tershed. Considering each zone the high risk zone (zone 1) and the moderate or low risk zone (zone 2), the first one had 79,260 Mg (1%) and the second one had 3,037,890 Mg (30%). At the sub watershed level, 22 out of 63 sub watersheds had a high risk of ex periencing some damage in their forests due to hurricanes (Figure 3 6). Rocky Hill Tower sub watershed registered the greater high risk considering the total forest area within its boundarie s that is equal to 326 hectares; or 70% of the forest has a high r isk of hurricane damage and 30% of the sub under low risk. For the Pensacola B ay watershed, the aboveground carbon stor age in 2003 was 1,396,430 Mg and ranging from 3 to 116 Mg/ha, thus, the carbon that could be potentially lost in the worst scenario wa s considered. High average values were located in the western part of the watershed (Figure 3 7). The sub watershed with the highest average aboveground carbon was Bayou Texar Bayou Garcon Frontal, with 64 Mg/ha,

PAGE 101

101 where 27% of the area is composed by forests, 15% being hardwoods and 12% softwoods. About 0.3% of the existing forest area in the watershed is under high risk of damage due to hurricanes; but the rest, 99.7% are under low risk of damage. The most affected areas were located i n the coastal areas of the watershed (Figure 3 8). Table 3 5 shows the aboveground carbon values in the high and low forest damage risk zones for the Pensacola Bay watershed. The aboveground carbon in the high risk zone was 0.5% of the total aboveground ca rbon registered in the watershed. At the sub watershed level only 2 out of 15 sub watersheds registered a high risk of damage based on the forest vegetation characteristics. The highest value corresponded to Fundy Bayou Williams Creek Frontal sub watershed with 59% of the areas defined as high risk (Figure 3 9). This sub watershed registered only 3% of forest area. Urban areas were not assessed in this analysis. Potential L oss of T imber Volume D ue to H urricanes The average timber net volume (volnet) for the Lower Suwannee in 2003 was 109 m 3 /ha over an area of 186,275 ha of forest, resulting in a total timber net volume of 20,402,395 m 3 approximately. Considering the worst scenario where hurricanes can damage all the forest areas the approximate amount of vo lnet that can be potentially lost can be up to 20,402,395 m 3 At the sub watershed level, the timber net volume ranged from 33 to 194 m 3 /ha, with the highest average value for Gopher River sub watershed. In this sub watershed the 73% of the area is covered by forest where 71% corresponded to hardwoods and 2% to softwoods. The timber volume in the sawlog portion (volsaw), in 2003 was 15,477,193 m 3 with an average of 83 m 3 /ha for the Lower Suwannee River watershed. In the worst scenario

PAGE 102

102 where the hurricane a ffects all the forest areas the approximate amount of volsaw that can be potentially lost can be around 15,477,193 m 3 Table 3 2 shows the timber net volume and the timber in the sawlog portion for the identified high and low forest damage risk zones due to hurricanes for the Lower Suwannee River watershed. The timber net volume in the high risk zones was 15% of the total observed in the watershed, and the timber sawlog portion was 32%. For Pensacola B ay, the average timber net volume in 2003 was 69 m 3 /ha in an area of 54,381 ha, the total timber net volume was approximately 3,730,642 m 3 This is the potential timber volume that can be lost in the worst hurricane scenario. At the sub watershed level the average timber net volume ranged from 45 to 157 m 3 /ha. The highest average value corresponded to Bayou Texar Bayou Garcon Frontal where 27% of the area is covered by forests; being 15% hardwoods and 12% softwoods. The timber volume in the sawlog portion, in 2003 was 2,980,362.85 m 3 approximately with an avera ge of 68 m 3 /ha. Table 3 4 shows the values for timber net volume and timber in the sawlog portion in the Pensacola B ay watershed. The amounts of timber net volume and in the sawlog portion were low in the high risk zones : 0.7% and 0.5% of the total registe red in the watershed, respectively. Discussion A decision tree analysis based on previous studies can help to determine the zones with high and low forest damage risk zones due to hurricanes considering the structural characteristics of forests such as tre e diameter and height, as well as basal area. Results show that this method can be used to determine the effects of direct drivers on ecosystem services, specifically how aboveground carbon storage and timber

PAGE 103

103 volume can be potentially affected due to hurri canes in the Lower Suwannee River and Pensacola B ay watersheds. Potential Loss of A boveground Carbon S torage D ue to H urricanes Hurricanes can forests However, not all the carbon will be lost different changes will occur in the short term as well as in the middle and long term In the short term carbon pool changes occur as aboveground carbon in live materials shift s to carbon dead and carbon can also be stored in the salvaged timber (McNulty, 2002) According to McNulty (2002) only 15% of the total carbon can be salvaged after a major hurricane, the rest is decomposed and returned to the atmosphere in a long term. For the Suwannee River watershed, the area with high risk of forest damage was 13% of the total forest s in the watershed, 11% being in zone 2 (moderate or low damage) where most of the forest are hardwood ( 87% ) Oswalt and Oswalt (2008) calculated the mean percent of plot damages by hurricane Katrina and they found that 8% were damaged considering all the zones (closest and furthest from landfall). In terms of forest area damage by Hurricane Katrina, Wang and Xu (20 09) found that 60% of the total forest land in the region was damaged by the Hurricane. The aboveground carbon in the high forest damage risk zone was 31% of the total estimated for the watershed, with a 30% of the amount in the zone 2 because most of the forest is located within this zone (93%). In the Pensacola B ay watershed, the area with high forest damage risk due to hurricanes was only 0.3% of the total forest registered in the watershed The aboveground carbon in the high risk zone was only 0.5% of the total aboveground carbon estimated for the watershed ; most forest area ( 93% ) being located within this

PAGE 104

104 zone. Previous studies mention that the amount of aboveground carbon that can be transferred from the living to dead pool is about 10% for a single hurricanes (McNulty, 2002) in terms of carbon emissions about 9 18% of the carbon storage can be released (Zeng et al., 2009) and for H urricane Katrina the carbon loss was 77.6 kg per ha (Zeng et al., 2009) Apparently and according to the results of this study the loss of aboveground carbon storage in the Pensacola B ay watershed is potentially low compar ed to documented loss es from Hurricane Katrina ; in spite the area being more a ffected by hurricanes because of its geographic location (i.e. ASCE 7 98 maximum wind speed for this region is 130 140 miles per hour ) Potential L oss of T imber Volume D ue to H urricanes The timber net volume in the high forest damage risk zone was 15% and 0.7% of the total observed in the Lower Suwannee and Pensacola B ay watershed respectively. The timber in the sawlog portion was 32% for the Lower Suwannee and 0.5% for Pensacola B ay watershed. Jacobs (2007) found that 10% of the total timber was destroyed after hurricane Camille (category 5) that impacted Southern Mississippi. S tanturf et al. (2007) mention that approximately the 20% of the standing timber volume was de stroyed as consequence of Hurricane Katrina. However, McNulty (2002) mentions that across all the damage zones, 37% of the standing tree was damaged from major hurricanes such as Katrina. Jacobs (2007) mentioned that with hurricane Andrew, there was a loss in standing volume of 66% in high and moderate damage zones with hardwoods receiving more damage. In the case of Hurricane Camille in 1969 (category 5 ), Jacobs (2007) produced a decrease in timber volume of 21 37%. The percen tage of damage can vary significantly because of the factors that control forest damage, such as forest characteristics (Oswalt and Oswalt, 2008) region al

PAGE 105

105 characteristics (climate, topography) (Boose et al., 1994) and because of the uncertainty in hurricane prediction (Oswalt and Oswalt, 2008) Some of the limitations of this study include the decision tree analysis that was based on exi sting literature that identif ied factors controlling forest damages due to hurricanes. Thus, t o get better results, this method should be used with existing pre and post hurricane field data to be better able to determine effects based on the percentage of forest damage and to find relevant factors. Another limitation is hurricane complexity, since many factors interact to determine the degree of forest damage by hurricane, such as wind speed, rainfall, mechanical and chemical effects of storm surge s stres s from hurricane induced damage, and mortality from subsequent post hurricane disturbances (i.e. fires), as well as the variability of forest structural characteristics and conditions (Kupfer et al., 2008) Conclu ding Remarks The main objective of this study was to determine the potential loss of aboveground carbon storage and timber volume due to drivers such as hurricanes and to identify high forest damage risk zones due to hurricanes based on the most important factors influencing forest damage. According to existing literature these factors were the tree diameter and height, the distance to landfall and the type of species t hat composed the forest (hardwood or softwood). Using decision tree analyses, a scenario was developed and combined with available USDA FIA and Florida vegetation and land use spatial data to better identify spatially explicit high and low forest damage ri sk zones in the study area. However, the uncertainty associated with hurricanes can generate highly variab i l ity resulting in uncertainty when predict ing damages to forests. This

PAGE 106

106 uncertainty is the product of the random nature of hurricane occurrence in ter ms of space and time as h urricanes are natural forces that cannot be replicated Further research could explore the damage to forests from hurricanes through an analysis that includ ed more factors that can exert and influence in the degree of damage on for ests, as well as different analysis for different categories of hurricanes and different climatic situations. Other ecosystem services can be included and analys e s at different scales should also be done, because factors can vary across different scales T he identification of high forest damage risk zones is important for pre and post hurricane forest management. S ome forest management at the stand level can be done to minimize the risk of carbon and timber loss due to hurricanes, the management practices c an be done before or after the hurricane impact. Overall, based on the literature review, decision tree analysis is not commonly used in this kind of study; however, it is a good alternative for assessing hurricane damage to forests. T his method could also be applied at a state level to identify high forest damage risk zones across the State of Florida The potential for timber salvag ing activities can be also included in the decision tree analysis to identify areas after a hurricane T his study decision tree could also be used to develop an online portal for landowners to know how vulnerable their forests are to hurricane impacts. The inclusion of urban areas in the analysis can also be done in future studies to have a complete idea about ecosystem servic es loss es across all ecosystems in the watersheds Knowledge on the potential effects of hurricanes on forest ecosystems and their services is growing because of the predicted increase in hurricane intensity and occurrence. map and identify the most vulnerable

PAGE 107

107 zones, as well to evaluate how hurricanes can affect aboveground carbon storage, and consequently carbon sequestration in a long term, and timber volume to intensify the management of the areas with high probability of ecosystem services loss. Land managers should include the potential occurrence of disturbances, such as hurricanes in their management plans to be aware of the consequences in terms of loss of ecosystem services especially when they have in mind some cons ervation strategies such as payment for ecosystem services.

PAGE 108

108 Table 3 1 Description of the Forest Inventory and Analysis (FIA) attributes used in the hurricane framework [ Adapted from Woudenberg, S. W., Conkling, B. L., O'Connell, B. M., LaPoint, E. B., Turner, J. A., and K. L. Waddell. 2010. The Manual Version 4.0 for Phase 2. U.S. Department of Agriculture, Forest Service, Rocky Mountain Research Station, Fort Collins, CO.] Attribut e code Attribute FIA Table name Attribute description UNITCD FIA Survey unit Condition, Tree, Plot Survey units are groups of counties within each State ECOSUBCD Ecological Subsection Plot An area of similar surficial geology, lithology, geomorphic proce ss, soil groups, sub regional climate, and potential natural communities COUNTYCD County Condition, Tree, Plot The identification number for a county, a governmental unit in a State COND_STATUS_CD Condition status Condition The condition status indicates the basic land cover (e.g. forest land) STDAGE Stand age Condition The average age of the stand (years) FORTYPCD Forest description Condition Forest type code DIA Current diameter Tree Current diameter of the tree in inches. HT Total height Tree The t otal length of the tree (in feet) from the ground to the tip of the apical meristem. BALIVE Basal area Condition Basal area of live trees over 1 inch of diameter at breast height (in square feet per acre)

PAGE 109

109 Table 3 2 Range of values for the two defined classes for tree diameter, tree height and basal area for the Lower Suwannee River and Pensacola Bay watersheds Attribute Value in Lower Suwannee River Value in P ensacola Bay Tree diameter DIA 1 >=8.4 inch >=7.5 inch Tree diameter DIA 2 < 8.4 inch < 7.5 inch Tree height HT 1 >=45.8 feet >=50 feet Tree height HT 2 < 45.8 feet < 50 feet Basal area BA 1 >=80 square feet/acre >=68 square feet/acre Basal area BA 2 < 80 square feet/acre < 68 square feet/acre Table 3 3. Expressions used in the Conditional func tion in ArcGIS Spatial Analyst to define the risk of forest damage by hurricanes Area of damage Expression Risk of forest damage High forest risk damage zone : 1 Land cover = softwoods and Area of damage = 1 and mean height = 1 and mean diameter= 2 and ba sal area = 2 High Moderate or Low forest risk damage zone : 2 Land cover = hardwoods and Area of damage = 2 and mean height = 1 and mean diameter = 1 and basal area = 2 High Table 3 4. Potential a boveground carbon, timber net volume, and timber in the sa wlog portion in the forest damage risk zones f or the Lower Suwannee watershed Risk zone Area (ha) Aboveground c arbon (Mg) % Timber net volume (m 3 ) % Timber sawlog portion (m 3 ) % Low 162,435 7,076,490 69 17,337,987 85 10,563,310 68 High 23,841 3,117,150 3 1 3,064,328 15 4,913,881 32 Total 186,276 10,193,640 100 20,402,314 100 15,477,192 100 Table 3 5. Potential a boveground carbon, timber net volume, timber in the sawlog portion in the forest damage risk zones for the Pensacola B ay watershed Risk zone Are a (ha) Aboveground c arbon (Mg) % Timber net volume (m 3 ) % Timber sawlog portion (m 3 ) % Low 53,904 1,384,500 99.5 3,681,811 99.3 2,950,522 99.5 High 170 7,547 0.5 27,391 0.7 13,661 0.5 Total 54,074 1,392,047 100 3,709,202 100 2,964,183 100

PAGE 110

110 Figure 3 1. Decision tree scenario for determining forest damage risk due to hurricanes Tree type Risk Tree height Tree diameter Basal area Hardwoods Zone 1 Sof twoods Zone 1 Softwoods Zone 2 Hardwoods Zone 2 High High Height 1 (greater heights) Height 2 Height 1 (greater heights) Height 2 Diameter 1 (larger trees) Diameter 2 (smaller trees) Diameter 1 (larger trees) Diameter 2 (smaller trees) Basal a rea 1 (high) Basal area 1 (low) Basal area 1 (high) Basal area 1 (low) Forest damage risk

PAGE 111

111 Figure 3 2. American So 7 98 map with wind risks for the two watersheds: Lower Su wannee River and Pensac ola B ay [ Adapted from Peacock, W. G., Brody, S. D., and W. Highfield. 2005. Hurricane Risk Perceptions among Florida's Single Family Homeowners. (Page 127, Figure 1). Landscape and Urban Planning ]

PAGE 112

112 Figure 3 3. Flow chart for the decision tree analysis for determining forest damage risk zones Hurricane framework F orest I nventory A nalysis (FIA) Data L and cover (LC) Map FIA forest reclassification: hardwoods and softwoods LC reclassification proces s : hardwoods and softwoods Mean diameter (DIA), total height (HT), and Basal area (BALIVE) Final reclassified LC DIA, HT and BALIVE by LC Decision tree analysis High and Low forest damage risk zones

PAGE 113

113 Figure 3 4. Aboveground c arbon storage in the Lower S uwannee River watershed in 2003

PAGE 114

114 Figure 3 5. Risk of forest damage by hurricanes in the Lower Suwannee River watershed

PAGE 115

115 Figure 3 6. High risk of forest da mage by hurricanes in the Lower Suwannee River watershed by sub watersheds

PAGE 116

116 Figure 3 7. Aboveground c arbon storage in the Pensacola B ay watershed in 2003

PAGE 117

117 Figure 3 8. Risk of forest damage by hurricanes in the Pensacola B ay watershed

PAGE 118

118 Figure 3 9. Hig h risk of forest damage by hurricanes in the Pensacola B ay watershed by sub watersheds

PAGE 119

119 CHAPTER 4 CONCLUSION S Summary The concept of e cosystem services is becoming more relevant to environmental planning and management (Boyd and Banzhaf, 2007; Brown et al., 2007; de Groot et al., 2010; Farley and Costanza, 2010) The Millennium Ecosystem Assessment (2003) defined ecosystem services as those services provided by ecosystems with a direct influence on human well being. The southern US fores ts are one of the most biologically diverse temperate forests in the world and consequently they can produce a bundle of ecosystem services (i.e. a set of multiple ecosystem services) (Raudsepp Hearne et al., 2010) Although people can benefit from multiple ecosystem services; they often have to decide which services are more relevant to them depend ing o n what benefit they want from these ecosystems. So, the inclusion of the population in the decision making process is key to know what ecosystem services are important for a region. Change in land cover can affect the provision of ecosystem services b y reducing them. All ecosystems constantly change over time, and many of these changes are due to human impacts (e.g. urbanization) and natural processes (e.g. wildfire) (ICSU et al., 2008; Millennium Ecosystem Assessment, 2003) Understanding the causes of these changes, or drivers as done in this study is fundamental to minimize their impact on the provisioning of ecosystem services as well as on the human well being (Millennium Ecosystem Assessment, 2003) Effects of U rbanizati on on Carbon Storage, T imber Volume, and Water Yield Ecosystem can be impacted from different drivers that can alter their functionality as well as their processes and services The first step in understanding the magnitude

PAGE 120

120 of the impacts to ES provision is to quantify ES as we ll as the drivers affecting them. Urbanization defined as the development of new urban areas in non urban lands is considered one of the major drivers that can affect ecosystems (Grimm et al., 2008; Xu et al., 200 7) As urbanization areas increase, ecosystems functions, processes and services can be impacted. This study estimated the effects of urbanization as a driver of carbon storage, timber volume, and water yield in the Lower Suwannee River and Pensacola Bay watersheds in Florida. The study combined Geographic Information System (GIS) and Forest Inventory and Analysis (FIA) data to estimate ES using a 2003 current scenario and a projected 2060 urbanized scenario in these two areas. The Integrated Valuation an d Ecosystem Services Tradeoffs (InVEST) GIS model was used to quantify the ecosystem services. Results show that there were significant changes in the two study areas between the current and urbanized 2060 scenarios in terms of provision of the three ecosy stem services. Thus, urbanization directly affects carbon storage, timber volume, and water yield ; c arbon storage and timber volume we re negatively correlated to urbanization and water yield was positively correlated. Results showed that urbanization will conver t forest areas in the year 2060, and that will be result in the decrease of the provision of carbon storage and timber volume and i nto an increase in water yield. Potential Loss of Aboveground Carbon Storage and T im ber Volume from H urricanes Hurrica nes are also a driver affecting ecosystems, as well as its processes and services. This study identified forest damage risk zones due to hurricanes that can be useful for landowners and policy makers and is crucial for management purposes. This study also estimated the potential loss of ecosystem services in terms of aboveground

PAGE 121

121 carbon storage and timber volume by hurricanes in the Lower Suwannee (LS) and Pensacola B ay (PB) watershed s in Florida The identification of the high and low probabilities zones w ere based on five factors identified in previous studies as important for determining the degree of forest damage: tree diameter and height, distance to landfall, type of species (e.g. hardwoods or softwoods), and basal area. A decision tree framework was developed based on five factors identified in previous studies as important for determining the potential damage in forest due to hurricanes. Those factors were: tree diameter and height, distance to landfall, type of species (e.g. hardwoods or softwoods), and basal area. Results showed that this method can be used to determine potential losses of aboveground carbon storage and timber volume due to hurricanes in the Lower Suwannee River and Pensacola B ay watersheds. Integra ting the Concept of Ecosystem Serv ices in D ecisions One of the major challenge that scientist are facing is the translation of ecosystem services in to action. The objective of this is to make better decisions, resulting in better use of the natural resources (Daily et al., 2009) Several studies have been done for evaluating ecosystem services in different aspects (Liu and Costanza, 2010; Millennium Ecosystem Assessment, 2003; Nelson et al., 2009) ; however n o t many of them have been put in practice. So, the integration of ecosystem services in decisions becomes very important to improve human life. Decisions in terms of land use change have been done base only in economic aspects putting aside ecological aspe cts (McKenzie et al., 2011) As a result of the lack of information on the quantity and value of ecosystem services some land covers that are multifunctional (and consequently multiple ES provi sion ) are conver ted often to single functional land covers (e.g. croplands)

PAGE 122

122 Therefore, to produce the information about ecosystem services and how these are affected by different drivers is the first step to include them in the decision making process. The next step is t o put that information available to the public, land owners, policy makers, urban planners, and others decision makers but in a very clear, simple and understandable way. Other effective strategies are education and awareness programs for all who lack info rmation (McKenzie et al., 2011) Daily et al. (2009) for example, presented a framework to integrate ecosystem service into decisions, they started describing the how d ecisions are linked with ecosystems, they proposed the use of scenarios to see the effects of decisions on ecosystems. Next, they mention the importance of knowing about ecosystem services and their values to provide this information to the institutions th at have the power to create, for example conservation incentives. Integrating ecosystem services into policies is not impossible. There are some examples, ecosystem services are included in the National Strategy for Growth and Reduction of Poverty of Tanza nia for the year 2005 (McKenzie et al., 2011) Basically they include the protection and enhancing of ecosystem services to improve the welfare of people. Information about ecosystem services can also influence decisions, this is the case of Borneo, there was an interest in change a forest area to an oil palm plantation; however, an assessment on the values of ecosystem services favor the forest area over the oil palm plantation (McKenzie et al., 2011) The Millennium Ecosystem Assessment (2005) presents several other examples. Even though several studies have been done to show the consequences of ecosystem service changes on the population erception abo ut ecosystems

PAGE 123

123 does not change, no possible study can have any impact on them (McKenzie et al., 2011) Therefore, i f people do not value nature and they are more interested in short term economic growth; the int egration of ecosystem services in decisions will be very difficult (Daily et al., 2009; Nelson et al., 2009) In conclusion, w hen people begin understand ing the importance of ecosystem services to th e i r well being incorporati ng ecosystem serv i ces in to policies will become easier.

PAGE 124

124 LIST OF REFERENCES Abd Elrahman, A., Adams, D., Borisova, T., Delphin, S., Escobedo, F., Frank, A., Kil, N., Kreye, M., Kroeger, T., Stein, T., and N. Timilsina, 2012, "Stewardship Ecosys tem Services Survey Project: Final Report", Escobedo, F. and N. Timilsina (Eds.), Unpublished manuscript, last modified May 16 2012. Abt, R. C., Cubbage, F. W., and G. Pacheco, 2000, "Southern Forest Resource Assessment Using the Subregional Timber Supply (SRTS) Model," Forest Products Journal 50(4):25 33. Alig, R. J., Kline, J. D. and M. Lichtenstein, 2004, "Urbanization on the US Landscape: Looking Ahead in the 21st Century," Landscape and Urban Planning 69(2 3):219 234. Alvarez, R., Cropper, W., Har well, M., Jagtap, S., Landsea, C., Letson, D., Parker, C., Shivlani, M., Wanless, H., and J. Winchester, 2001, Feeling the Heat in Florida: Global Warming on the Local Level Fiedler, J., F. Mays and J. Siry (Eds.), New York, NY: Natural Resources Defense Council. Arnold, J. G., Bircket, M. D., Williams, J. R., Smith, W. F., and H. N. McGill, 1987, "Modeling the Effects of Urbanization on Basin Water Yield and Reservoir Sedimentation," Journal of the American Water Resources Association 23(6):1101 1107. Bengston, D. N., Potts, R. S., Fan, D. P., and E. G. Goetz, 2005, "An Analysis of the Public Discourse about Urban Sprawl in the United States: Monitoring Concern about a Major Threat to Forests," Forest Policy and Economics 7(5):745 756. Birch, E. L. an d S. M. Wachter (Eds.), 2008, Growing Greener Cities: Urban Sustainability in the Twenty First Century The City in the Twenty First Century Philadelphia: University of Pennsylvania Press. Birdsey, R., Pregitzer, K., and A. Lucier, 2006, "Forest Carbon M anagement in the United States: 1600 2100," Journal of Environmental Quality 35(4):1461 1469. Boggess, D. H., 1968, Water Supply Problems in Southwest Florida Tallahassee, FL: U S Geological Survey Open File Report FL 68003. Boggs, J. L. and G. Sun, 2 011, "Urbanization Alters Watershed Hydrology in the Piedmont of North Carolina," Ecohydrology 4(2):256 264. Boose, E. R., Foster, D. R., and M. Fluet, 1994, "Hurricane Impacts to Tropical and Temperate Forest Landscapes," Ecological Monographs 64(4):36 9 400.

PAGE 125

125 Bosch, J. M. and J. D. Hewlett, 1982, "A Review of Catchment Experiments to Determine the Effect of Vegetation Changes on Water Yield and Evapo Transpiration," Journal of Hydrology 55(1 4):3 23. Boyd, J. and S. Banzhaf, 2007, "What Are Ecosystem Services? The Need for Standardized Environmental Accounting Units," Ecological Economics 63(2 3):616 626. Braatz, S. M. (Eds.), 2001, Rome, Italy: Food and Agriculture Organization of the United Nations. Brokaw, N. V. L. and L. R. Walker, 1991, "Summary of the Effects of Caribbean Hurricanes on Vegetation," Biotropica 23(4):442 447. Bromley, S. W., 1939, "Factors Influencing Tree Destruction during the New England Hurricane," Science 90:15 16. Brown, T. C., Bergstr om, J. C., and J. B. Loomis, 2007, "Defining, Valuing, and Providing Ecosystem Goods and Services," Natural Resources Journal 47(2):329 376. Bystriakova, N., Brown, S., Gonzalez, P., Irland, L. C., Kauppi, P., Sedjo, R., and I. D. Thompson, 2005, "Timber Fuel, and Fiber," in Ecosystems and Human Well Being: Current State and Trends : Findings of the Condition and Trends Working Group of the Millennium Ecosystem Assessment Hassan, R. M., Scholes, R. and N. Ash (Eds.), Washington, D.C.: Island Press, 243 269. Canadell, J., Jackson, R., Ehleringer, J., Mooney, H. A., Sala, O., and E. D. Schulze, 1996, "Maximum Rooting Depth of Vegetation Types at the Global Scale," Oecologia 108(4):583 595. Carr, M. H. and P. D. Zwick, 2007, Smart Land Use Analysis: The LUCIS Model Land Use Conflict Identification Strategy Redlands, CA: ESRI Press. Chambers, J. Q., Fisher, J. I., Zeng, H., Chapman, E. L., Baker, D. B., and G. C. Hurtt, 2007, "Hurricane Katrina's Carbon Footprint on U. S. Gulf Coast Forests," Science 31 8(5853):1107 1107. Chapin III, F. S., Matson, P. A., and H. A. Mooney, 2002, Principles of Terrestrial Ecosystem Ecology New York, NY: Springer. Chen, N., Li, H., and, L. Wang, 2009, "A GIS Based Approach for Mapping Direct Use Value of Ecosystem Servic es at a County Scale: Management Implications," Ecological Economics 68(11):2768 2776.

PAGE 126

126 Conner, R. and A. Hartsell, 2002, "Forest Area and Conditions," in Southern Forest Resource Assessment. Wear, D. N. and J. G. Greis (Eds.), Asheville, NC: U.S. Departm ent of Agriculture, Forest Service, Southern Research Station, 357 402. Conner, R. C. and R. M. Sheffield, 2006, Analysis of the Timber Situation in Florida, 1995 to 2025 Asheville, NC: U.S. Department of Agriculture, Forest Service, Southern Research St ation. Conte, M., Ennaanay, D., Mendoza, G., Todd Walter, M., Wolny, S., Freyberg, D., Nelson, E., and L. Solorzano, 2011, "Retention of Nutrients and Sediment by Vegetation," in Natural Capital: Theory and Practice of Mapping Ecosystem Services Kareiva, P., H. Tallis, T. H. Ricketts, G. Daily and S. Polasky (Eds.), New York, NY: Oxford University Press, 89 110. Cooley, T. N. and D. F. Martin, 1979, "Cadmium in Naturally Occurring water Hyacinths," Chemosphere 8(2):75 78. Costanza, R., d'Arge, R., de G root, R., Farber, S., Grasso, M., Hannon, B., Limburg, K., Naeem, S., O'Neill, R. V., Paruelo, J., Raskin, R. G., Sutton, P., and M. van den Belt, 1997, "The Value of the World's Ecosystem Services and Natural Capital," Nature 387(6630):253 260. Coulson, D. P., Joyce, L. A., Price, D. T., McKenney, D. W., Siltanen, R. M., Papadopol, P., and L. Kevin., 2010, Climate Scenarios for the Conterminous United States at the County Spatial Scale Using SRES Scenarios A1B and A2 and Prism Climatology U.S. Departmen t of Agriculture, Forest Service, Rocky Mountain Research Station (March 19 2012) [http://www.fs.fed.us/rm/data_archive/dataaccess/US_ClimateScenarios_county_ A1B_A2_PRISM.shtml]. Cubbage, F. W., Harris Jr, T. G., Wear, D. N., Abt, R. C., and G. Pacheco, 1 995, "Timber Supply in the South: Where Is All the Wood?," Journal of Forestry 93(7):16 20. Daily, G. C. (Eds.), 1997, Nature's Services: Societal Dependence on Natural Ecosystems Washington, DC: Island Press. Daily, G. C., Polasky, S., Goldstein, J., Kareiva, P. M., Mooney, H. A., Pejchar, L., Ricketts, T. H., Salzman, J., and R. Shallenberger, 2009, "Ecosystem Services in Decision Making: Time to Deliver," Frontiers in Ecology and the Environment 7(1):21 28. Dale, V. H., Lugo, A. E., MacMahon, J. A. and S. T. A. Pickett, 1998, "Ecosystem Management in the Context of Large, Infrequent Disturbances," Ecosystems 1(6):546 557.

PAGE 127

127 Davies, Z. G., Edmondson, J. L., Heinemeyer, A., Leake, J. R., and K. J. Gaston, 2011, "Mapping an Urban Ecosystem Service: Qua ntifying above Ground Carbon Storage at a City Wide Scale," Journal of Applied Ecology 48(5):1125 1134. de Groot, R. S., Alkemade, R., Braat, L., Hein, L., and L. Willemen, 2010, "Challenges in Integrating the Concept of Ecosystem Services and Values in Landscape Planning, Management and Decision Making," Ecological Complexity 7(3):260 272. DeFries, R. S., Foley, J. A., and G. P. Asner, 2004, "Land Use Choices: Balancing Human Needs and Ecosystem Function," Frontiers in Ecology and the Environment 2(5) :249 257. DeLoach, T. and S. Dicke, 2005. General Forestry Information: Salvaging Timber, Frequently Asked Questions Raymond, MS: Mississippi State Extension Service. Duryea, M. L., Kampf, E., and R. C. Littell, 2007, "Hurricanes and the Urban Forest: I Effects on Southeastern United States Coastal Plain Tree Species," Arboriculture & Urban Forestry 33(2):83 97. Dymond, J. R., Ausseil, A. G. E., Ekanayake, J. C., and M. U. F. Kirschbaum, 2012, "Tradeoffs between Soil, Water, and Carbon: A National Sca le Analysis from New Zealand," Journal of Environmental Management 95(1):124 131. Egoh, B., Reyers, B., Rouget, M., Richardson, D. M., Le Maitre, D. C., and A. S. van Jaarsveld, 2008, "Mapping Ecosystem Services for Planning and Management," Agriculture Ecosystems & Environment 127(1 2):135 140. Escobedo, F., Varela, S., Staudhammer, C., Thompson, B., and J. Jarratt, 2009a, Southern Escambia County, Florida's Urban Forests Gainesville, FL: University of Florida IFAS, EDIS. FOR 293. Escobedo, F. J., Lu ley, C. J., Bond, J., Staudhammer, C., and C. Bartel, 2009b, "Hurricane Debris and Damage Assessment for Florida Urban Forests," Arboriculture & Urban Forestry 35(2):100 106. ESRI, 2011a, ArcGIS Desktop: Release 10 Redlands, CA: Environmental Systems R esearch Institute. ESRI, 2011b, Classifying Numerical Fields for Graduated Symbology (May 18 2012) [http://help.arcgis.com/en/arcgisdesktop/10.0/help/index.html#//00s50000001r000 000.htm]. ESRI, 2012a, How IDW Works (May 22 2012) [http://help.arcgis.com/e n/arcgisdesktop/ 10.0/help/index.html#/How_IDW_works/009z00000075000000/].

PAGE 128

128 ESRI, 2012b, Zonal Statistics as Table (Spatial Analyst) (May 20 2012) [http://help.arcgis.com/en/arcgisdesktop/10.0/help/index.html#//009z000000w800 0000.htm]. European Commission, 2009, Ecosystem Goods and Services Luxembourg: Publications Office. Farley, J. and R. Costanza, 2010, "Payments for Ecosystem Services: From Local to Global," Ecological Economics 69(11):2060 2068. Florida Department of Environmental Protection, 2007, Water Quality Assessment Report: Pensacola Bay Lord, L. (Eds.), Tallahassee, FL. Florida Fish and Wildlife Conservation Commission, 2004. Florida Vegetation and Land Cover 2003 Tallahassee, FL. Florida Oceans and Coastal Council, 2009, The Effects of Ocean and Coastal Resources: a Special Report to the Florida Energy and Climate Commission and the People of Florida Tallahassee, FL. Foley, J. A., DeFries, R., Asner, G. P., Barford, C., Bonan, G., Carpenter, S. R., Chapin, F S., Coe, M. T., Daily, G. C., Gibbs, H. K., Helkowski, J. H., Holloway, T., Howard, E. A., Kucharik, C. J., Monfreda, C., Patz, J. A., Prentice, I. C., Ramankutty, N., and P. K. Snyder, 2005, "Global Consequences of Land Use," Science 309(5734):570 574. Foster, D. R. and E. R. Boose, 1992, "Patterns of Forest Damage Resulting from Catastrophic Wind in Central New England, USA," Journal of Ecology 80(1):79 98. Frangi, J. L. and A. E. Lugo, 1991, "Hurricane Damage to a Flood Plain Forest in the Luquillo Mountains of Puerto Rico," Biotropica 23(4):324 335. Frick, E. A., 1998, Water Quality in the Apalachicola Chattahoochee Flint River Basin, Georgia, Alabama, and Florida, 1992 95 : U.S. Geological Survey [http://pubs.water.usgs.gov/circ1164]. Furniss, M J., Staab, B. P., Hazelhurst, S., Clifton, C. F., Roby, K. B., Ilhadrt, B. L., Larry, E. B., Todd, A. H., Reid, L. M., Hines, S. J., Bennett, K. A., Luce, C. H., and P. J. Edwards, 2010, Water, Climate Change, and Forests: Watershed Stewardship for a Cha nging Climate Portland, OR: U.S. Department of Agriculture, Forest Service, Pacific Northwest Research Station. Gilbert, T. and B. Stys, 2004, Descriptions of Vegetation and Land Cover Types Mapped Using Landsat Imagery Tallahassee, FL: Florida Fish and Wildlife Conservation Commission.

PAGE 129

129 Gimona, A. and D. van der Horst, 2007, "Mapping Hotspots of Multiple Landscape Functions: A Case Study on Farmland Afforestation in Scotland," Landscape Ecology 22(8):1255 1264. Gordon, H. B., Rotstayn, L. D., McGregor, S. P., Waterman, L. J., Hirst, A. C., Wilson, S. G., Collier, M. A., Watterson, I. G., and T. I. Elliott, 2002, The CSIRO Mk3 Climate System Model Aspendale, Australia: CSIRO Atmospheric Research. Gorte, R W., 2009, Carbon Sequestration in Forests Washington, D.C.: Congressional Research Service. Grimm, N. B., Foster, D., Groffman, P., Grove, J. M., Hopkinson, C. S., Nadelhoffer, K. J., Pataki, D. E., and D. P. C. Peters, 2008, "The Changing Landscape: E cosystem Responses to Urbanization and Pollution across Climatic and Societal Gradients," Frontiers in Ecology and the Environment 6(5):264 272. Hancock, J., 2010, "The Case for an Ecosystem Service Approach to Decision Making: An Overview," Bioscience H orizons 3(2):188 196. Hanson, C., Yonavjak, L., Clarke, C., Minnemeyer, S., Boisrobert, L., Leach, A., and K. Schleeweis, 2010, Southern Forests for the Future Washington, DC: World Resources Institute. Harmon, M. E., 2001, "Carbon Sequestration in For ests: Addressing the Scale Question," Journal of Forestry 99(4):24 29. Hays, B., 2009, "Environmental Setting," in Ecology of Pensacola Bay Caffrey, J. (Eds.), 1 38. Heath, L. S. and J. E. Smith, 2000, "An Assessment of Uncertainty in Forest Carbon Bud get Projections," Environmental Science & Policy 3(2 3):73 82. Heath, L. S., Smith, J. E., Woodall, C. W., Azuma, D. L., and K. L. Waddell, 2011, "Carbon Stocks on Forestland of the United States, with Emphasis on Usda Forest Service Ownership," Ecospher e 2(1):1 21. Hibbert, A. R., 1967, "Forest Treatment Effects on Water Yield," in International Symposium on Forest Hydrology Sopper, W. E. and H. W. Lull (Eds.), Oxford, England: Pergamon Press, 527 543. Hodges, D. G., Gunter, J. T., Swalm, C. M., and J. L. Regens, 1998, Impact of Urbanization and Environmental Protection Policies on Timber Availability: A GIS Application, Paper read at Proceedings of the First International Conference Geospatial Information in Agriculture and Forestry, 1 3 June 1998, a t Lake Buena Vista, FL.

PAGE 130

130 Hollis, G., 1977, "Water Yield Changes after the Urbanization of the Canon's Brook Catchment, Harlow, England," Hydrological Sciences Journal 22(1):61 75. Horvitz, C. C., Pascarella, J. B., McMann, S., Freedman, A., and R. H. Hofs tetter, 1998, "Functional Roles of Invasive Non Indigenous Plants in Hurricane Affected Subtropical Hardwood Forests," Ecological Applications 8(4):947 974. Howard, R. A., Matheson, J. E., and D. W. North, 1972, "Decision to Seed Hurricanes," Science 17 6(4040):1191 1202. Huggett, R., Wear, D. N., Li, R., Coulston, J., and S. Liu, 2002, "Forecasts of Forest Conditions," in Southern Forest Resource Assessment Wear, D. N. and J. G. Greis (Eds.), Asheville, NC: U.S. Department of Agriculture, Forest Servic e, Southern Research Station, 1 71. ICSU, UNESCO, and UNU, 2008, Ecosystem Change and Human Well Being: Research and Monitoring Priorities based on the Millennium Ecosystem Assessment Paris, France: International Council for Science. IPCC, 2001, Climate Change 2001: Impacts, Adaptation and Vulnerability, Contribution of Working Group II to the Third Assessment Report of the Intergovernmental Panel on Climate Change Cambridge, UK, and New York, USA: Cambridge University Press. Jacobs, D. M., 2007, Fores t Inventory, Catastrophic Events and Historic Geospatial Assessment in the South, Paper read at ASPRS 2007 Annual Conference, May 7 11, at Tampa, FL. Jandl, R., Vesterdal, L., Olsson, M., Bens, O., Badeck, F., and J. Roc, 2007, "Carbon Sequestration and F orest Management," CAB Reviews: Perspectives in Agriculture, Veterinary Science, Nutrition and Natural Resources 2(17):1 16. Jenkins, J. C., Chojnacky, D. C., Heath, L. S., and R. A. Birdsey, 2003, "National Scale Biomass Estimators for United States Tre e Species," Forest Science 49(1):12 35. Jones, R. C. and C. C. Clark, 1987, "Impact of Watershed Urbanization on Stream Insect Communities," Water Resources Bulletin 23(6):1047 1055. Katz, B. G., DeHan, R. S., Hirten, J. J., and J. S. Catches, 1997, "I nteractions between Groundwater and Surface Water in the Suwannee River Basin, Florida," Journal of the American Water Resources Association 33(6):1237 1254. Kupfer, J. A., Myers, A. T., McLane, S. E., and G. N. Melton, 2008, "Patterns of Forest Damage i n a Southern Mississippi Landscape Caused by Hurricane Katrina," Ecosystems 11(1):45 60.

PAGE 131

131 Lal, R., 2005, "Forest Soils and Carbon Sequestration," Forest ecology and management 220(1):242 258. Leatherman, S. and N. Defraene, 2006, 10 Most Hurricane Vulner able Areas International Hurricane Research Center (02/20 2012) [http://www.fema.gov/emergency/reports/2006/nat112006.shtm]. Li, W., Yang, C., and D. Sun, 2009, "Mining Geophysical Parameters through Decision Tree Analysis to Determine Correlation with T ropical Cyclone Development," Computers & Geosciences 35(2):309 316. Light, H. M., Darst, M. R., Lewis, L. J., and D. Howell, 2002, Hydrology, Vegetation, and Soils of Riverine and Tidal Floodplain Forests of the Lower Suwannee River, Florida, and Potent ial Impacts of Flow Reductions : U.S. Geological Survey. Liu, S. and R. Costanza, 2010, "Ecosystem Services Valuation in China," Ecological Economics 69(7):1387 1388. Lugo, A. E., 2008, "Visible and Invisible Effects of Hurricanes on Forest Ecosystems: A n International Review," Austral Ecology 33(4):368 398. Martin, J. B., 2007, The Suwannee River: A Coastal Plain Watershed in Transition Gainesville, FL: University of Florida, Florida State University, University of South Florida, University of Central Florida, University of Georgia, USGS, USDA, and SRWMD. McKenzie, E., Irwin, F., Ranganathan, J., Hanson, C., Kousky, C., Bennett, K., Ruffo, S., Conte, M., Salzman, J., and J. Paavola, 2011, "Incorporating Ecosystem Services in Decisions," in Natural Cap ital: Theory and Practice of Mapping Ecosystem Services Kareiva, P., Tallis, H., Ricketts, T. H., Daily, G. and S. Polasky (Eds.), New York, NY: Oxford University Press, 339 355. McNab, W. H., Cleland, D. T., Freeouf, J. A., Keys, J., J.E., Nowacki, G. J ., and C. A. Carpenter, 2007, Description of Ecological Subregions: Sections of the Conterminous United States Washington, DC: U.S. Department of Agriculture, Forest Service. McNulty, S. G., 2002, "Hurricane Impacts on Us Forest Carbon Sequestration," En vironmental Pollution 116:S17 S24. Mendoza, G., Ennaanay, D., Conte, M., Todd Walter, M., Freyberg, D., Wolny, S., Hay, L., White, S., Nelson, E., and L. Solorzano, 2011, "Water Supply as an Ecosystem Service for Hydropower and Irrigation," in Natural Ca pital: Theory and Practice of Mapping Ecosystem Services Kareiva, P., Tallis, H., Ricketts, T. H., Daily, G. and S. Polasky (Eds.), New York, NY: Oxford University Press, 53 72.

PAGE 132

132 Merry, K. L., Bettinger, P., Grebner, D. L., and J. Hepinstall Cymerman, 2010 "Perceptions of Foresters of Wind Damage in Mississippi Forests," Southern Journal of Applied Forestry 34(3):124 130. Metzger, M. J., Rounsevell, M. D. A., Acosta Michlik, L., Leemans, R., and D. Schrter, 2006, "The Vulnerability of Ecosystem Services to Land Use Change," Agriculture, Ecosystems & Environment 114(1):69 85. Microsoft, 2010, Microsoft Excel Redmont, Washington: Microsoft. Millennium Ecosystem Assessment, 2003. Ecosystems and Human Well Being: A Framework for Assessment Washington, D .C.: Island Press. Millennium Ecosystem Assessment, 2005. Ecosystems and Human Well Being: Synthesis Washington, D.C.: Island Press. Moore, J., Bartlett, J., Boggs, J. L., Gavazzi, M. J., Heath, L. S., and S. G. McNulty, 2002, "Abiotic Factors," in Sout hern Forest Resource Assessment Wear, D. N. and J. G. Greis (Eds.), Asheville, NC: U.S. Department of Agriculture, Forest Service, Southern Research Station, 429 452. Munn, I. A., Barlow, S. A., Evans, D. L., and D. Cleaves, 2002, "Urbanization's Impact on Timber Harvesting in the South Central United States," Journal of Environmental Management 64(1):65 76. Myers, R. K. and D. H. van Lear, 1998, "Hurricane Fire Interactions in Coastal Forests of the South: A Review and Hypothesis," Forest Ecology and M anagement 103(2 3):265 276. Nasi, R., Wunder, S., and J. J. Campos A., 2002. Forest Ecosystem Services: Can They Pay our Way out of Deforestation? Bogor, Indonesia: CIFOR for the Global Environmental Facility (GEF). Natural Capital Project, 2011, Integ rated Valuation of Ecosystem Services and Tradeoffs 2.2.2 (InVEST), Stanford, CA. Natural Resources Conservation Service, 2005, Florida Hydrologic Unit Area Map USDA Natural Resources Conservation Service [http://water.usgs.gov/GIS/huc.html]. Nelson, E. Mendoza, G., Regetz, J., Polasky, S., Tallis, H., Cameron, D. R., Chan, K. M. A., Daily, G. C., Goldstein, J., Kareiva, P. M., Lonsdorf, E., Naidoo, R., Ricketts, T. H., and M. R. Shaw, 2009, "Modeling Multiple Ecosystem Services, Biodiversity Conservati on, Commodity Production, and Tradeoffs at Landscape Scales," Frontiers in Ecology and the Environment 7(1):4 11.

PAGE 133

133 Nelson, G. C., 2005, "Drivers of Ecosystem Change: Summary Chapter," in Ecosystems and Human Well Being: current State and Trends Hassan, R ., R. Scholes and N. Ash (Eds.), Washington, DC: Island Press, 74 76. Newell, R. G. and R. N. Stavins, 2000, "Climate Change and Forest Sinks: Factors Affecting the Costs of Carbon Sequestration," Journal of Environmental Economics and Management 40(3):2 11 235. Nielsen, J. P. and J. M. Norris, 2007. From the River to You: USGS Real Time Streamflow Information Pembroke, NH: U S Geological Survey. NOAA National Hurricane Center, 2012a, Glossary of NHC Terms (May 18 2012) [http://www.nhc.noaa.gov/aboutgl oss.shtml#l]. NOAA National Hurricane Center, 2012b, NHC Data Archive (March 20 2012) [http://www.nhc.noaa.gov/pastall.shtml]. Northwest Florida Water Management District, 2006, Surface Water Improvement and Management Program: Priority List for the Nort hwest Florida Water Management District Havana, FL. Nowak, D. J., Noble, M. H., Sisinni, S. M., and J. F. Dwyer, 2001, "People & Trees Assessing the Us Urban Forest Resource," Journal of Forestry 99(3):37 42. Nowak, D. J., Walton, J. T., Dwyer, J. F. Kaya, L. G., and S. Myeong, 2005, "The Increasing Influence of Urban Environments on Us Forest Management," Journal of Forestry 103(8):377 382. Obreza, T. and G. Means, 2006, Characterizing Agriculture in Florida's Lower Suwannee River Basin Area : Univ ersity of Florida IFAS, EDIS. SS460. Olivas, R., 2007, Decision Tree: A Premier for Decision Making Professionals (April 28 2012) [http://www.stylusandslate.com/decision_trees/index.html]. Oswalt, S. N. and C. M. Oswalt, 2008, "Relationships between Comm on Forest Metrics and Realized Impacts of Hurricane Katrina on Forest Resources in Mississippi," Forest Ecology and Management 255(5 6):1692 1700. Peacock, W. G., Brody, S. D., and W. Highfield, 2005, "Hurricane Risk Perceptions among Florida's Single Fa mily Homeowners," Landscape and Urban Planning 73(2 3):120 135. Prestemon, J. P. and R. C. Abt, 2002a, "The Southern Timber Market to 2040," Journal of Forestry 100(7):16 22.

PAGE 134

134 Prestemon, J. P. and R. C. Abt, 2002b, "Timber Products Supply and Demand," i n Southern Forest Resource Assessment Wear, D. N. and J. G. Greis (Eds.), Asheville, NC: U.S. Department of Agriculture Forest Service, Southern Research Station, 299 325. PRISM Climate Group and Oregon State University, 2008, Total Precipitation (August 10 2011) [http://prism.oregonstate.edu]. Raudsepp Hearne, C., Peterson, G. D., and E. M. Bennett, 2010, "Ecosystem Service Bundles for Analyzing Tradeoffs in Diverse Landscapes," Proceedings of the National Academy of Sciences of the United States of Ame rica 107(11):5242 5247. Reyers, B., O'Farrell, P. J., Cowling, R. M., Egoh, B. N., Le Maitre, D. C., and J. H. J. Vlok, 2009, "Ecosystem Services, Land Cover Change, and Stakeholders: Finding a Sustainable Foothold for a Semiarid Biodiversity Hotspot," E cology and Society 14(1):38. Rodriguez, J. P., Beard, T. D., Jr., Bennett, E. M., Cumming, G. S., Cork, S. J., Agard, J., Dobson, A. P., and G. D. Peterson, 2006, "Trade Offs across Space, Time, and Ecosystem Services," Ecology and Society 11(1):28. Sa bor, A. A., Radeloff, V. C., McRoberts, R. E., Clayton, M., and S. I. Stewart, 2007, "Adding Uncertainty to Forest Inventory Plot Locations: Effects on Analyses Using Geospatial Data," Canadian Journal of Forest Research Revue Canadienne De Recherche Fores tiere 37(11):2313 2325. Schwenning, L., Traci Bruce, T., and L. R. Handley, 2007, "Pensacola Bay," in Seagrass Status and Trends in the Northern Gulf of Mexico: 1940 2002 Handley, L., D. Altsman and R. DeMay (Eds.): U.S. Geological Survey Scientific Inv estigations Report, 129 142. Seidl, R., Rammer, W., Jaeger, D., Currie, W. S., and M. J. Lexer, 2007, "Assessing Trade Offs between Carbon Sequestration and Timber Production within a Framework of Multi Purpose Forestry in Austria," Forest Ecology and Man agement 248(1 2):64 79. Smail, R. A. and D. J. Lewis, 2009, "Forest Land Conversion, Ecosystem Services, and Economic Issues for Policy a Review," In. Portland, OR: U.S. Dept. of Agriculture, Forest Service, Pacific Northwest Research Station [http://pur l.access.gpo.gov/GPO/LPS119022]. Smith, J. E., Heath, L. S., and P. B. Woodbury, 2004, "How to Estimate Forest Carbon for Large Areas from Inventory Data," Journal of Forestry 102(5):25 31.

PAGE 135

135 Soil Survey Staff, Natural Resources Conservation Service, and A griculture, U. S. D. A ., 2006, U.S. General Soil Map (Statsgo2) (August 10 2011) [http://soildatamart.nrcs.usda.gov]. Stanton, E. A. and F. Ackerman, 2007, Florida and Climate Change: The Costs of Inaction Medford, MA: Global Development and Environment Institute, Tufts University and Stockholm Environment Institute US Center, Tufts University. Stanturf, J., 2008, Lessons from Katrina and Rita, Compass 1 2. Stanturf, J. A., Goodrick, S. L., and K. W. Outcalt, 2007, "Disturbance and Coastal Forests: A S trategic Approach to Forest Management in Hurricane Impact Zones," Forest Ecology and Management 250(1 2):119 135. Stavins, R. N., 1999, "The Costs of Carbon Sequestration: A Revealed Preference Approach," American Economic Review 89(4):994 1009. Stys, B., Kautz, R., Reed, D., Kertis, M., Kawula, R., Keller, C., and A. Davis, 2004. Florida Vegetation and Land Cover Data Derived from 2003 Landsat Etm+ Imagery Tallahassee, FL: Florida Fish and Wildlife Conservation Commission. Su, S., Xiao, R., Jiang, Z ., and Y. Zhang, 2012, "Characterizing Landscape Pattern and Ecosystem Service Value Changes for Urbanization Impacts at an Eco Regional Scale," Applied Geography 34(2012):295 305. Sun, G., McNulty, S. G., Lu, J., Amatya, D. M., Liang, Y., and R. K. Kolk a, 2005, "Regional Annual Water Yield from Forest Lands and its Response to Potential Deforestation across the Southeastern United States," Journal of Hydrology 308(1 4):258 268. Sun, G., McNulty, S. G., Myers, J. A. M., and E. C. Cohen, 2008, "Impacts o f Multiple Stresses on Water Demand and Supply across the Southeastern United States," Journal of the American Water Resources Association 44(6):1441 1457. Swank, W. T., Vose, J. M., and K. J. Elliott, 2001, "Long Term Hydrologic and Water Quality Respon ses Following Commercial Clearcutting of Mixed Hardwoods on a Southern Appalachian Catchment," Forest Ecology and Management 143(1 3):163 178. Tallis, H., Ricketts, T., Guerry, A., Wood, S., Sharp, R., Nelson, E., Ennaanay, D., Wolny, S., Olwero, N., Vig erstol, K., Pennington, D., Mendoza, G., Aukema, J., Foster, J., Forrest, J., Cameron, D., Arkema, K., Lonsdorf, E., Kennedy, C., Verutes, G., Kim, C., Guannel, G., Papenfus, M., Toft, J., Marsik, M., and J. Bernhardt, 2011, Stan ford: The Natural Capital Project.

PAGE 136

136 Tanner, E. V. J., Kapos, V., and J. R. Healey, 1991, "Hurricane Effects on Forest Ecosystems in the Caribbean," Biotropica 23(4):513 521. Tawil, N., 2007, The Potential for Carbon Sequestration in the United States Wa shington, D.C.: Congress of the U.S., Congressional Budget Office. Thorpe, P., Bartel, R., Ryan, P., Albertson, K., Pratt, T., and D. Cairns, 1997, The Pensacola Bay System Surface Water Improvement and Management Plan: A Comprehensive Plan for the Restor ation and Preservation of the Pensacola Bay System Cairns, D., Macmillan, T., Bartel, R., Culbertson, M., Ryan, P., and P. Thorpe (Eds.). Havana, FL: Hank Montford. Timilsina, N., Escobedo, F., Cropper, W. P. J., Abd Elrahman, A., Brandeis, T., Delphin, S., and S. Lambert, 2011, "A Framework for Identifying Carbon Hotspots and Forest Management Drivers," Unpublished manuscript, last modified March 29 2012. Troy, A. and M. A. Wilson, 2006, "Mapping Ecosystem Services: Practical Challenges and Opportunitie s in Linking GIS and Value Transfer," Ecological Economics 60(2):435 449. U nited S tates, Department of Commerce, Census Bureau, 2002, International Data Base: World Population U.S. Census Bureau, U.S. Department of Commerce (7 March 2012) [http://www.ce nsus.gov/population/international/data/idb/informationGateway. php]. U nited S tates, Environmental Protection Agency, 2004, The Ecological Condition of the Pensacola Bay System, Northwest Florida (1994 2001) Gulf Breeze, Florida U.S. Environmental Protect ion Agency, Office of Research and Development, National Health and Ecological Effects Research Laboratory, Gulf Ecology Division. U nited S tates, D epartment of A griculture, Forest Service ECOMAP Team, 2007, Ecological Subregions: Sections and Subsections for the Conterminous United States Washington, DC: U.S.D.A Forest Service. U nited S tates, Geolog ical Survey, Florida Water Science Center, 2008, Florida Potential and Reference Evapotranspiration 1998 2008 (August 10 2011) [http://hdwp.er.usgs.gov/]. U riarte, M. and M. Papaik, 2007, "Hurricane Impacts on Dynamics, Structure and Carbon Sequestration Potential of Forest Ecosystems in Southern New England, USA," Tellus Series a Dynamic Meteorology and Oceanography 59(4):519 528. Vitousek, P. M., Mooney, H. A., Lubchenco, J., and J. M. Melillo, 1997, "Human Domination of Earth's Ecosystems," Science 277(5325):494 499.

PAGE 137

137 Waage, S., Stewart, E., and K. Armstrong, 2008, Measuring Corporate Impact on Ecosystems: A Comprehensive Review of New Tools: Synthesis Re port San Francisco, CA: Business for Social Responsibility. Wang, F. and Y. J. Xu, 2009, "Hurricane Katrina Induced Forest Damage in Relation to Ecological Factors at Landscape Scale," Environmental Monitoring and Assessment 156(1 4):491 507. Wang, F. and Y. J. Xu, 2010, "Comparison of Remote Sensing Change Detection Techniques for Assessing Hurricane Damage to Forests," Environmental Monitoring and Assessment 162(1 4):311 326. Wear, D., 2002a, Forecasts of Land Uses Wear, D. N. and J. G. Greis (Eds. ), Southern Forest Resource Assessment: Technical Report Asheville, NC: U.S. Department of Agriculture, Forest Service, Southern Research Station, 1 71. Wear, D., 2002b, "Land Use," in Southern Forest Resource Assessment: General Report Wear, D. N. and J. G. Greis (Eds.), Asheville, NC: U.S. Department of Agriculture, Forest Service, Southern Research Station, 153 173. Wear, D. N. and J. G. Greis (Eds.), 2002a, Southern Forest Resource Assessment: Technical Report Ashville, NC: US Department of Agricul ture, Forest Service, Southern Research Station. Wear, D. N. and J. G. Greis (Eds.), 2002b, Southern Forest Resource Assessment: Summary Report Asheville, NC: U.S. Department of Agriculture, Forest Service, Southern Research Station. Winter, T. C., 1999 Ground Water and Surface Water: A Single Resource Denver, CO: U.S. Geological Survey. Woodbury, P. B., Smith, J. E., and L. S. Heath, 2007, "Carbon Sequestration in the Us Forest Sector from 1990 to 2010," Forest Ecology and Management 241(1 3):14 27. Worley, S., 2008, How Much Forest Carbon Is Lost after a Major Hurricane?, Compass 21 22. Woudenberg, S. W., Conkling, B. L., O'Connell, B. M., LaPoint, E. B., Turner, J. A., and K. L. Waddell, 2010, The Forest Inventory and Analysis Database: Database Fort Collins, CO: U.S. Department of Agriculture, Forest Service, Rocky Mountain Research Station. Xi, W., Peet, R. K., Decoster, J. K., and D. L. Urban, 2008, "Tree Damage Risk Factors Associated wi th Large, Infrequent Wind Disturbances of Carolina Forests," Forestry 81(3):317 334.

PAGE 138

138 Xu, C., Liu, M., An, S., Chen, J. M., and P. Yan, 2007, "Assessing the Impact of Urbanization on Regional Net Primary Productivity in Jiangyin County, China," Journal of Environmental Management 85(3):597 606. Yonavjak, L., Hanson, C., Talberth, J., and T. Gartner, 2011. Keeping Forest as Forest: Incentives for the U.S. South Washington, DC: World Resources Institute. Zang, S., Wu, C., Liu, H., and X. Na, 2011, "Impact of Urbanization on Natural Ecosystem Service Values: A Comparative Study," Environmental Monitoring and Assessment 179(1 4):575 588. Zektser, I. S. and H. A. Loaiciga, 1993, "Groundwater Fluxes in the Global Hydrologic Cycle Past, Present and Future," Journal of Hydrology 144(1 4):405 427. Zeng, H., Chambers, J. Q., Negron Juarez, R. I., Hurtt, G. C., Baker, D. B., and M. D. Powell, 2009, "Impacts of Tropical Cyclones on Us Forest Tree Mortality and Carbon Flux from 1851 to 2000," Proceedings of the National Academy of Sciences of the United States of America 106(19):7888 7892. Zhang, C., Tian, H., Chen, G., Chappelka, A., Xu, X., Ren, W., Hui, D., Liu, M., Lu, C., and S. Pan, 2012, "Impacts of Urbanization on Carbon Balance in Terrestrial Ecosystem s of the Southern United States," Environmental Pollution 164:89 101. Zhang, D. W. and R. V. Nagubadi, 2005, "The Influence of Urbanization on Timberland Use by Forest Type in the Southern United States," Forest Policy and Economics 7(5):721 731. Zhang L., Dawes, W., and G. Walker, 1999, Predicting the Effect of Vegetation Changes on Catchment Average Water Balance Canberra: Cooperative Research Centre for Catchment Hydrology. Zhang, L., Dawes, W., and G. Walker, 2001, "Response of Mean Annual Evapot ranspiration to Vegetation Changes at Catchment Scale," Water resources research 37(3):701 708. Zheng, D., Heath, L. S., and M. J. Ducey, 2012, "Potential Overestimation of Carbon Sequestration in the Forested Wildland Urban Interface in Northern New Eng land," Journal of Forestry 110(2):105 111. Zwick, P. and M. Carr, 2006, Florida 2060: A Population Distribution Scenario for the State of Florida Gainesvi lle, FL: GeoPlan Center Universi ty of Florida.

PAGE 139

139 BIOGRAPHICAL SKETCH Sonia Elsa Rocio Delphin Perez was born in Asuncion, Paraguay in 1980. She got a bachelor degree in forest e ngineering from the National University of Asuncion in 2005 Her thesis analyzed f orests fires based on remote sensing and geographic information s ystems (GIS) methods. After gra duation she joined an environmental consulting firm focusing her work in the use of GIS to analyze forest ecosystems. She work ed there for 2 and half years. Later, she started working for World Wildlife F und (WWF) in Paraguay as a GIS a nalyst. Her work wa s focused on the monitoring of the deforestation in the Upper Parana Atlantic Forest in the Eastern Region of Paraguay bas ed on Landsat satellite imagery. Her job also included different GIS analysis of forests and GIS training to environmental employees f rom the Paraguayan Government She has been working there for 2 and half years. During that time she continued her education in GIS and remote s ensing at the Kansas State University and the Polytechnic University of Madrid. She entered at the School of For est Resources and Conservation at the University of Florida in the f all 2010 to get and she was graduated in August 2012