Record for a UF thesis. Title & abstract won't display until thesis is accessible after 2009-12-31.

Permanent Link: http://ufdc.ufl.edu/UFE0021466/00001

Material Information

Title: Record for a UF thesis. Title & abstract won't display until thesis is accessible after 2009-12-31.
Physical Description: Book
Language: english
Creator: Morgan, Kimberly L
Publisher: University of Florida
Place of Publication: Gainesville, Fla.
Publication Date: 2007


Subjects / Keywords: Food and Resource Economics -- Dissertations, Academic -- UF
Genre: Food and Resource Economics thesis, Ph.D.
bibliography   ( marcgt )
theses   ( marcgt )
government publication (state, provincial, terriorial, dependent)   ( marcgt )
born-digital   ( sobekcm )
Electronic Thesis or Dissertation


Statement of Responsibility: by Kimberly L Morgan.
Thesis: Thesis (Ph.D.)--University of Florida, 2007.
Local: Adviser: Larkin, Sherry L.
Electronic Access: INACCESSIBLE UNTIL 2009-12-31

Record Information

Source Institution: UFRGP
Rights Management: Applicable rights reserved.
Classification: lcc - LD1780 2007
System ID: UFE0021466:00001

Permanent Link: http://ufdc.ufl.edu/UFE0021466/00001

Material Information

Title: Record for a UF thesis. Title & abstract won't display until thesis is accessible after 2009-12-31.
Physical Description: Book
Language: english
Creator: Morgan, Kimberly L
Publisher: University of Florida
Place of Publication: Gainesville, Fla.
Publication Date: 2007


Subjects / Keywords: Food and Resource Economics -- Dissertations, Academic -- UF
Genre: Food and Resource Economics thesis, Ph.D.
bibliography   ( marcgt )
theses   ( marcgt )
government publication (state, provincial, terriorial, dependent)   ( marcgt )
born-digital   ( sobekcm )
Electronic Thesis or Dissertation


Statement of Responsibility: by Kimberly L Morgan.
Thesis: Thesis (Ph.D.)--University of Florida, 2007.
Local: Adviser: Larkin, Sherry L.
Electronic Access: INACCESSIBLE UNTIL 2009-12-31

Record Information

Source Institution: UFRGP
Rights Management: Applicable rights reserved.
Classification: lcc - LD1780 2007
System ID: UFE0021466:00001

This item has the following downloads:

Full Text




2007 Kimberly Ludwig Morgan 2


To my father and mother, Peter J. and Susan K. Ludwig, my grandmothers, Violet A. Ludwig and Beverly B. King, and my mentors, Dr. John Holt and Richard G. Miller 3


ACKNOWLEDGMENTS I have only my deepest appreciation and gr atitude to offer to my chairwoman and girlfriend, Dr. Sherry L. Larkin, and the member s of my committee, Dr. Robert L. Degner, Dr. Timothy G. Taylor, Dr. Charles A. Adams, and Dr. Charles A. Jacoby, for their long hours of patient counsel and dedication to my succe ssful completion of this doctoral program. I thank my husband and children for their love and constant support throughout the entire lengthy process of my pursuit of this degree. I thank my brothe rs for simultaneously acting as both my personal coaching staff and my most de voted fan club members. Many more thanks are due to my family and friends for providing unending support during my times of darkness and despair. Finally, there are a multitude of individuals that assisted my re search efforts with priceless tidbits of information at times when I needed them most, and to each of you, I would like to thank you all for being just an email or a phone call away. 4


TABLE OF CONTENTS page ACKNOWLEDGMENTS...............................................................................................................4 LIST OF TABLES................................................................................................................. ..........7 LIST OF FIGURES.........................................................................................................................9 ABSTRACT...................................................................................................................................10 CHAPTER 1 INTRODUCTION................................................................................................................. .12 Purpose of Study.....................................................................................................................12 Floridas Gulf Coast and Red Tide.........................................................................................12 Review of Applied Economic Studies....................................................................................16 Overview of Study Theory and Methods................................................................................18 2 FIRM-LEVEL ECONOMIC LOSSES DU RING RED TIDE BLOOMS: A CASE STUDY OF THREE BEACHF RONT RESTAURANTS......................................................21 Introduction................................................................................................................... ..........21 Methods..................................................................................................................................23 Theoretical and Empirical Model....................................................................................23 Proprietary Sales Data.....................................................................................................24 Environmental Data.........................................................................................................26 Results.....................................................................................................................................27 Model Estimation and Evaluation...................................................................................27 Effect of Red Tides..........................................................................................................28 Effect of Red Tides Relative to Other Environmental Factors (X).................................28 Effect of Time-Related Factors (D).................................................................................29 Discussion...............................................................................................................................30 Conclusions.............................................................................................................................32 3 RED TIDES AND PARTICIPATION IN MARINE-BASED ACTIVITIES: ESTIMATING THE RESPONSE OF SO UTHWEST FLORIDA RESIDENTS..................39 Introduction................................................................................................................... ..........39 Theoretical Model of Behavioral Choice...............................................................................42 An Application to Red Tide Events........................................................................................43 Data..................................................................................................................................43 Empirical Models............................................................................................................45 Hypotheses......................................................................................................................47 Empirical Results.............................................................................................................. ......48 Reaction Across Activities..............................................................................................48 5


Reaction by Activity........................................................................................................49 Type of Reaction by Activity..........................................................................................50 Summary and Discussion of Results......................................................................................51 Conclusions.............................................................................................................................53 4 PUBLIC COSTS OF FLORIDA RED TIDES: A SURVEY OF LOCAL MANAGERS.....64 Introduction................................................................................................................... ..........64 Study Objective......................................................................................................................65 Study Area..............................................................................................................................66 Procedures..................................................................................................................... ..........68 Survey Results................................................................................................................. .......69 Response Rate.................................................................................................................7 0 Agencies..........................................................................................................................70 Funding Sources and Expenditures.................................................................................71 Communication or Activity Protocols.............................................................................73 Summary and Conclusions.....................................................................................................74 5 SUMMARY AND CONCLUSIONS.....................................................................................93 LIST OF REFERENCES.............................................................................................................101 BIOGRAPHICAL SKETCH.......................................................................................................106 6


LIST OF TABLES Table page 2-1 Variable descriptions and definitions.................................................................................34 2-2 Descriptive statistics for continuous variables...................................................................34 2-3 Environmental data by firm reporte d as the number of days observed.............................35 2-4 Estimation results by firm................................................................................................. .36 3-1 Variable names, descriptions and statistics........................................................................55 3-2 Marginal effects for factors hypothesi zed to influence whether a residents participation in a marine-based activity is affected by a red tide......................................56 3-3 Marginal effects and standard errors for factors hypothesized to influence whether a residents participation in each marine-b ased activity is affected by a red tide................57 3-4 Marginal effects and standard errors for factors hypothesized to influence the reaction of a resident for activity whose part icipation has been affected by a red tide ....58 4-1 Population estimates of nine Florida Gulf Coast counties.................................................78 4-2 Tourist tax collections, dollars and per cent of total county taxes, and estimated annual tourist numbers for nine Florida Gulf Coast counties............................................78 4-3 Approximate tourism tax dollars collected per public beach miles for nine Florida Gulf Coast counties............................................................................................................ 79 4-4 Population estimates of Florida Gulf Coast cities..............................................................80 4-5 Disposition of telephone in terviews and response rates....................................................81 4-6 Summary of public departments charged with physical beach/red tide management responsibilities by county..................................................................................................81 4-7 Summary of public departments charged with physical beach/red tide management responsibilities by city....................................................................................................... 82 4-8 Summary of sources of funds for financial beach/red tide management responsibilities and estimated funds and expenditures, by county....................................83 4-9 Sarasota County expenditures for six red tide events by public beach..............................84 4-10 Pinellas County reimbursements for 2005 red tide events by city public beach...............84 7


4-11 Summary of public departments charged with financial beach/red tide management responsibilities and estimated expenditures and labor by city...........................................85 4-12 Summary of communication and activity protocols followed in the case of an active red tide event by county.....................................................................................................88 4-13 Summary of communication and activity protocols followed in the case of an active red tide event by city......................................................................................................... .90 8


LIST OF FIGURES Figure page 2-1 Average daily CPI-adjusted sales for each firm, 1998-2005.............................................37 2-2 Average daily temperature and wi nd speed by month, Southwest Florida........................38 3-1 Reaction of residents to red ti de events by marine-based activity.....................................60 3-2 Predicted probabilities of participation in any activity and beach-going due to a red tide event by the number of activities and the average participation days........................61 3-3 Predicted probability of participation in beach-going, boat fishing, and restaurant patronage due to a red tide event.......................................................................................62 3-4 Change in predicted probability of re sponse by marine-based activity for selected statistically significant variables........................................................................................63 9


Abstract of Dissertation Pres ented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy ECONOMIC ANALYSES OF THE EFFECTS OF RED TIDE EVENTS ON THREE SECTORS OF FLORIDA COASTAL COMMUNI TIES: RESTAURANTS, RESIDENTS AND LOCAL GOVERNMENT By Kimberly Ludwig Morgan December 2007 Chair: Sherry L. Larkin Major: Food and Resource Economics As scientists learn more about the effectiveness of various harmful algal bloom prevention, control, and mitigation strategies, information on local-level economic effects is needed. This study examines red-tide related economic effects across a nine-c ounty region of Floridas Gulf Coast in three related essays, which contribute to the empirical and theoretical literature on HAB-effect measurement and mitigation. Only se lect specific findings are summarized here. The first essay estimated sales reductions due to red tide events for three Southwest Florida beachfront restaurants from 19982005 using time-series linear regression analysis. Locationspecific manager notes were used to indicate red tide, rain, and tropical storm days. Findings revealed that average inflation-adjusted sale s declined by 13.7% and 15.3 % for the two highest grossing restaurants on days when a red tide was present. The second essay estimated the pr obability that a resident in Manatee or Sarasota County that engaged in marine-related activities was affected by a red tide. Using a sample of 569 residents, a series of discrete choice models were estimated to explain whether a resident was affected by red tide events in 2000 and, if so, how they were affected. Findings revealed that variables that can be affected by Extension activities were the only factors that affected the 10


probability that a residents participation in a marine-related activity would change during a red tide. Subsequent analyses by activity and to furt her investigate the type of reaction revealed difference by activity that can be used to design more effective outreach ma terials as a means of mitigation the affects of a red tide. The final essay provided insights into the fina ncial and managerial impacts of a red tide event on nine county and eighteen city governme nts charged with management of public beaches. Two cities had 2006 budgets of $50,000 and $100,000 for red tide beach cleaning. For 2004-2007, six locations reported expenditure s ranging from $11,114 to $250,000 per event and totaling $653,890. Expenditures were directly correlated with public beach length, severity of fish kills, and available beach management budg ets, but should be considered conservative estimates since they did not include in-kind labor or equipment expenses. 11


CHAPTER 1 INTRODUCTION Purpose of Study The overall purpose of this study is to generate information that would provide for a better understanding of the economic cons equences of red tide events on local managers, residents and restaurants located on the Gulf of Mexico in Sout hwest Florida, an area that has historically experienced severe red tide events. This is the first study to explore the economic impacts of red tide events across three inter-related components of Floridas economy within a region that is heavily dependent on marine water quality and related amenities. The combined use of proprietary sales information and primary survey data collected from residents and government managers is expected to provide estimates of business losses, behavi oral choices, and local government expenditures on beach management re sulting from the presence of harmful algal blooms (HABs). There are three objectives of this research. Objective one is to estimate the change in restaurant sales on those days when red tide conditions were no ticeable. Objective two is to predict the change in resident behaviors that ar e specific to their participation in marine-related activities during a red tide. To complete these examinations, other envi ronmental, temporal, spatial, and demographic explanatory variables are incorporated in the models for comparison. Objective three is to ascertain the expenditure s and specific protocols associated with public beach management during red tide events and dete rmine the factors that affect the expenditure levels. The combined analyses are examined in turn in Chapters Two through Four and are expected to provide quantitative and qualitative evidence of the e ffects of a red tide event in the localized study area and contribute to the growing literature on how to measure such effects. 12


Floridas Gulf Coast and Red Tide According to the 2000 National Survey on Recr eation and the Environment, Florida was the first destination of choice for over 22 milli on American travelers that sought out marine recreation on coastal beaches, waterways and wetlands in order to swim, dive, sunbathe, boat, or fish (Leeworthy and Wiley). These travelers spent more than 177 million beach days along Floridas coast in 2000 (Leeworthy and Wiley). Non-market values estimated for Floridas marine-related recreation activities ranged from $81 million for scuba diving up to nearly $18 billion for beach users during 2000 alone (Kildow). High-quality, sustainable marine coastal environments are crucial elements necessary to support and ensure the popularity of Floridas coastlines and a ssociated marine sectors. The increasing movement of the population to coasta l areas has emphasized coastal consumer and business reliance on predictable weather conditions, which in turn, has increased the economic impacts of naturally-occ urring environmental stressors (Changnon) Similar to daily temperature, precipitation, and wind data, s hort-run prediction m odels of uncontrollable environmental stressors, such as harmful algal blooms (HABs) or other water quality measures, are necessary sources of information that imp act decisions made by the public. Harmful algal blooms are defined to have one common unique feature: they cause harm, either due to their produc tion of toxins or to the manne r in which the cells physical structure or accumulated biomass affect co-occu rring organisms and alter food web dynamics (Anderson et al., 2000). When red tide algae ( Karenia brevis ) in particular become concentrated, fish and marine mammals may experience paralysi s and shellfish may become toxic, resulting in massive kills of a variety of fish species a nd endangered manatees among other marine life (Steidinger et al.). For example, the 2003 the de aths of 98 of Floridas endangered manatees were attributed to red tide blooms (Pain), wh ile the 2005 red tide even ts resulted in the 13


documented deaths of 118 endangered sea turtles (Tomalin). Human health impacts resulting from K. brevis toxins released into the air in minute parts per million can include respiratory distress, nasal and eye irritations, and skin ra shes, which can be particularly damaging to individuals with compromised immune systems such as the elderly and asthmatics (Flewelling et al., Backer et al.). Red tide cell count measurements collected by the Florida Fish and Wildlife Research Institute [FWRI] and Mote Marine Laboratory can be used to i ndicate the locati on and intensity of red tide blooms from 1996 through 2005 in Southwest Florida, the main area under consideration in this study. Fo llowing FWRIs key, very low K. brevis cells per liter of seawater range from greater than 1,000 to 10,000, low and medium categories include cell counts ranging from greater than 10,000 to unde r 100,000 and greater or equal to 100,000 to less than a million, respectively; both categories cont ain cell count levels that are considered sufficient to cause human respiratory irritation and probable fish kills. The high category includes measurements of at least one million cells per liter. While an official cell count level that might cause either significant fish kills or respiratory distress in humans has not been established, the Florida Divisi on of Aquaculture has mandated that waters with cell counts exceeding 5,000 per liter will result in shellfish bed closures to prevent NSP (Florida Department of Agriculture and C onsumer Services). Anecdotal evidence of the effects of red tide on human health, sea life, and marine-centric businesses is also abundant. A Google news artic le search for the phrase red tide from 01 January 1996 to 31 December 2005 found more than 1,700 days with articles in just five of the Gulf Coasts print media. Internet web logs such as RedTideAlert.com offer web space to individuals that want to share written or pictoria l descriptions of their personal experiences with 14


the negative aspects of red tide. This site has posts from both fore ign and domestic visitors that indicate their plans to avoid vacationing, or residing, in Fl orida in the future. Unfortunately, testing of local waters can reveal zero or low red tide cell counts even while anecdotal evidence abounds (J. Abbott, personal communication, FWRI Harmful Algal Bloom Group, 5 January 2007). Such contradictions highlight the difficulties associated with absolute predictions of algal bloom impacts on sea life and human health (Landsberg). Red tide event economic losses are suspected to result from a variety of impacts including those on human health (e.g., lost wages, increased sick days and medical treatment costs) and coastal communities (e.g., shellfish fishery closur es, reduced recreational expenditures, and the need for beach cleaning and speci es rehabilitation). The Pinellas County visitor and convention bureau has stated that revenue losses approached $240 million as a result of the nearly year-long series of red tide events in 2005 (Moore). L ee Countys public tax records revealed a $100,000 commitment to red tide-related research. And, in July 2007, Florida Congress members supported a federal authorization of $30 million in each of the next three years for additional red tide research (Tibbetts). The successful management of red tide blooms and associated negative impacts should be examined at a regional scale if the region is the affected by red tides and can contribute to successful prevention, control, and m itigation activities (Fisher et al.). This is especially the case, if red tides and their paths could be predicted, alerted communities might have time to mobilize cleanup crews and establish warning systems before the bloom arrives (FWRI). To that end, this research attempts to not onl y provide quantitative evidence of localized red tide economic impacts (if they exist), but also to contribute to the methodologies that can be used to gather 15


credible information for the purpose of evaluating continued red tide rela ted expenditures by the government. Review of Applied Economic Studies Following several recent prolonged red tide blooms and continued increases in the coastal population, it is not surprising that local, state, national, public and private agencies have been overwhelmed with demands for economic information on HAB events. The earliest estimates found that direct economic impacts of HABs in the United States average $75 million annually, including impacts on public health costs, commercial fishing closures, recreation and tourism losses, and in management and monitoring costs (National Oceanic and Atmospheric Administration [NOAA] ). Similar research efforts have produced estimates of average annual national losse s of $46 million (Hoagland, 2000), $49 million (Anderson, Kaoru, and White, 2000), and $82 milli on (Hoagland and Scatasta, 2006). For the recreation and tourism sector in particular, annual losses were estimated at $4 million (Hoagland and Scatasta, 2006). For commercial fisheries, average annual losses were estimated at $19 million (2000 dollars) (Anderson, Kaoru, and White, 2000). Regional economic analyses of red tide events are faced with lim ited availability of consistent information, and exacerbated by the in herently random and unpr edictable nature of HABs. However, in one of the few region-specific studies, an input -output revealed that 59% of respondents indicated that negative impacts associated with red tide were reflected in their business sales and revenues, resulting in estima ted direct economic impacts of $10 million to Galveston, Texas economy (Evans). Halo effects that persist over time and masking variables (i.e. weather, varied seasonal at tractions, etc.) that coincided w ith red tide events were found to create additional difficulty in the analys is (Jensen; Hoagland and Scatasta, 2006). 16


In a recent study of economic impacts from red tides in Florida, th e changes in business activity that occurred during re d tide events from 1995 through 1999 in the Ft. Walton Beach and Destin areas of Okaloosa County were investig ated (Larkin and Adams). Aggregated monthly gross taxable sales data were used to determ ine red tide-related losses for these two small communities located on Floridas Northwestern Gu lf Coast. A time-series regression analysis was estimated to reveal historic al average losses of 29% and 35% for the restaurant and lodging sectors, respectively, during months when red tide events were measured in nearby offshore waters. While these studies have produced some hist oric measurements of HAB economic costs, they collectively outline comp elling reasons to expend future research efforts on the development of more precise estimates of economi c consequences. First, there appears to be a current trend of increasing harmful algal bloom s in both number of events and duration length (Brand and Compton), which serves to increase the important of estimates that can be used to justify continued expenditures. Second, there exis ts a need for empirical analyses resulting in statistically significant information at the firm le vel necessary to guide owners and operators of beach-dependent businesses in red tide-related loss management. Third, statistical analysis of the effects of red tide on residents that actively engage in marine -related activities may provide estimates of the scope and potential costs of residents decisions to avoid those activities. Finally, there is an unmet need for accurate red-tide induced expenditures incurred by public agencies that are charged with management, and depe ndent upon revenue streams, of beachfront destinations. There is no doubt that an estimation of the economic impacts of red tide events would provide the necessary guidance for prioritization of financial and other types of support to those 17


segments of the economy that are directly aff ected by HAB events. However, the variation of these same economic effects across currently availabl e scientific studies tends to result in more questions, and fewer clear-cut answ ers. Our study provides data an alyses completed at the local level of firms, consumers, and government agenci es that are impacted by red tide events, which add qualitative and quantitative approaches and co ntributions to the exis ting body of literature. Overview of Study Theory and Methods Previous studies that have pr ovided estimates of absolute economic losses as a result of harmful algal blooms in the United States have relied upon readily available secondary data. Three of the most recent studies (Anderson et al., 2000; Hoagland et al., 2002; Hoagland and Scatasta, 2006) used a combination of surveys fr om coastal state experts, literature reviews and individual calculations, to produce a semina l national review of economic implications associated with HAB occurrences from 19871992 on four economic sectors public health, commercial fisheries, recreation/tourism, and monitoring/management. Available red tide loss estimates were aggregated and average annual sh ares were calculated for each sector. While these studies were seminal, the absence of estimates for all hypothesized losses (and measurements at different points in the value chai n) will bias the absolute and relative impacts. Losses to commercial fisheries in particular ha ve been calculated us ing historical harvest levels and dockside prices (Anderson, Kaoru, and White, 2000). Most certainly, however, the magnitude of real losses will depend on such factor s as the number of fishery participants, length of fishery season, and the level of fixed costs as sociated with particip ation in the fishery. Published estimates of economic losses due to HABs have, thus, been based on a variety of case study methods and aggregated to include a wide range of econo mic sectors, including tourism, commercial fisheries, public health, and red tide management and monitoring. The resulting estimates from these data compilations are difficult to compare, given the information 18


was collected for different regions using differe nt sampling frames across such broad sectors, and therefore have limited use for future work. As an alternative approach, comprehensive que stioning of a few individuals across several sectors of a local coastal econom y within a limited geographic area that had recently experienced a red tide event resulted in data that was analyzed with an input -output model (Evans). Respondents were asked to recall pr ecise losses resulting from red tid e events that varied in their duration and intensity. The inconsistency of re spondent types and data recollection, and the reliance of the IMPLAN model on an underlying production function that was based on a broader geographic area and included both coas tal and inland areas suggested limits to the credibility of the methodology and its results. The field of economics provides various me thods for estimation of losses based on historical data. Time series anal ysis with econometrics techniques in particular allow for the use of consistently collected data and bias-free es timation (versus calculati ons based on restrictive assumptions) of red tide impacts. For example, th e empirical analysis of monthly gross taxable sales data from the Florida Department of Reve nue and the FWRI cell count measurements were successfully used to estimate economic losses to the restaurant and hotel sectors in Floridas Panhandle (Larkin and Adams). The availability of primary data analyzed at the finest resolution possible (e.g., that of an indi vidual firm) using econometric te chniques may provide defensible, statistically significant evidence of the effects of red tide days. Consumer behavior theory ba sed on the principles of an individuals goal of selecting goods and services that maximize his or her utilit y has been used to pred ict the probability of various recreational choices. This theory has suc cessfully produced estimation of welfare effects resulting from different management options for recreational salmon fishing (Lin, Adams, and 19


Berrens), measurement of the benefits of water quality improvements to marine recreational fishers in a North Carolina estuary (Kaoru), a nd estimation of average annual access values for recreational fishing in Tampa Bay, Fl orida (Greene, Moss and Spreen). Finally, estimation of economic impacts resulting from environmental extremes such as harmful algal blooms can result from institutional analysis, as It is important for practitioners and researchers to recognize the capacity (e.g., know ledge, power, and resources) to solve complex problems is often widely dispersed across a set of actors located at different levels of government (Imperial). Stakeholders across Flor idas local, regional, and state levels are actively pursuing and funding red tid e-related scientific and economic research, perhaps in large part due to concerns based on future lost proper ty tax revenues, lost values of a Florida beach vacation, and lost tourist interest in Floridas coastal activities. 20


CHAPTER 2 FIRM-LEVEL ECONOMIC LOSSES DURING RED TIDE BLOOMS: A CASE STUDY OF THREE BEACHFRONT RESTAURANTS Introduction Southwest Floridas economy is heavily dependent on its marine amenities. The value derived from marine-related busine ss is ultimately influenced by th e quality of the environment. In addition to the weather, extreme environmen tal conditions such as hurricanes and harmful algal blooms (HABs) can frequently occur in this region of Florida. The main species of HABs in Southwest Florida ( Karenia brevis ) is unique in that brevetoxi ns are produced during blooms. These specific toxins can kill marine life (Flewelling et al.), prevent safe consumption of shellfish, and cause respiratory ir ritation in humans (Backer et al.; Robbins et al.) and thereby cause economic losses to commercial and recreational marine-related businesses (Kusek et al.; Magana et al.; Schneider, Pierce and Rodrick; Casper et al.). Some business sectors have been able to re ceive compensation for HAB-related disasters. For example, the Small Business Association provided each of 36 Florida firms with $4,832 to $81,912 in loans due to red tide events that occurred between 1996 and 2002 (Tester. P.A. Personal Communication. NOAA Nati onal Ocean Service, 13 July 2007). Of the 36 total loans awarded to firms in Florida, only five (13.9%) went to restaurants. Overall, however, the restaurant sector fell behind onl y seafood markets and shellfishing (o f nine total sectors) in the total sum of monies loane d. The restaurant sector is vulnerable to red tide relate d losses and as a service sector contributes to sustainable tour ism in Florida with gr oss taxable sales of $17.3 billion in 1999 (Bureau of Economi c and Business Research [BEBR]). While there is an abundant and growing body of anecdotal information on the detrimental economic effects that HABs have on local econo mies (e.g., Glick; Huettell; Karp; Van Sant; McLaughlin and Spinner; Moore), there is a paucity of rigorous empirical analysis. Most studies 21


have either compared changes in dockside va lues of harvested seafood between seasons (e.g., Tester et al., 1991), calculate d average annual losses by aggreg ating across industries (Anderson et al., 2000; Hoagland and Scatasta, 2006), or esti mating losses using recall data from businesses in a localized area (Evans). One exception is a recent study that used secondary data from the Florida Department of Revenue to estimate hist orical losses of 29% to 35% on average for the restaurant and lodging sectors in two small co mmunities in Northwest Florida, respectively, during months when red tide was present in near shore waters (L arkin and Adams). During 2005, Floridas southwest coastal areas e xperienced a prolonged series of red tide events in nearly every month, raising widesp read concern in the business community (e.g., Glick; Huettel; Moroney; Moore). Since intense, long-lasting and far-reac hing blooms are not an unusual occurrence to this area (Steidinger et al .) and red tides may be becoming more abundant (Brand and Compton), regional eco nomic losses may increase. As a result, the demand for new and alternative prevention, control and m itigation strategies is also increasing. Fortunately, the scientific community is a dvancing several alternative prevention and control strategies for red tides (S chneider, Pierce, and Rodrick; Robbins et al.; Ca sper et al.). This same community also has a long history of advocating the need for exploration of local and regional data to gain accurate estimates of the size and magnitude of business interruptions precipitated by HAB events (Jen sen; Kahn and Rockel; Shumway; Anderson; Boesch et al.; Hoagland et al., 2002). In light of recent technol ogical advances and an increasing number of high profile red tide events, empirical support fo r red tide related business losses is paramount. This need is magnified considering the potenti al to improve forecasting models that could support the specification of ri sk premiums offered by private insurance companies. 22


To provide the statistical evidence of the economic effects of natural disasters and support for potential prevention, control and mitigati on strategies proprietary firm-level data is used to estimate lost sales for several beachf ront restaurants. Our study offers an initial examination of the economic consequences of red tid e events at the firm level. As red tide events are naturally-occurring phenomena, their presence and relative effects on sp ecific restaurants will be compared with other environm ental factors. It is the inte ntion of our study to provide a portion of the information request ed by scientists, resource mana gers, and business leaders. In addition, this research will further empirical analysis. Methods Theoretical and Empirical Model As suggested by Nordhaus, a time-series analysis might be useful for examining the impact of abrupt [climate] changes, for these are similar to extreme weather events. Following this prescription, the theoretical model for our study hypothesizes that, on a daily basis, restaurant sales ( Y ) are a function of exogenous environmental conditions ( X ) and seasonal demands ( D ), such as day of the week, season, and or year. Assuming a linear functional form, which allows for the direct estimation and comp arison of effects, the following empirical model is specified for each restaurant: ,, tj j tk k t jkYXD t (2-1) where t identifies a specific day, j indexes the environmental variables, and k indexes the timerelated variables (Table 2-1). Parameters j, and k will be estimated using a least squares approach for each firm. The random error, is likely to be autocorrelated due to the use of time series data. Thus, empirical equations will be tested for autocorrelation, with subsequent corrections to the estimati on procedure if necessary. 23


In our study, five j variables ( X ) are considered, including temperature, wind speed, rainfall, red tides, and tropical (or stronger) stor m conditions. Temperature is expected to vary directly with daily restaurant sales while the re maining environmental conditions (if present or at higher levels) are expected to vary inversely wi th sales. The time-related variables assumed to affect daily demand for restaurant services ( D ) include holidays, days of the week, months of the year, and years. The time-related variables will be discrete and dichotomous (i.e., 0-1 dummy variables) such that directional impacts on sales will depend on which category is used as the base and included in the intercept (i.e., ). In general, however, sales are expected to increase (coefficients have a positive sign) on holidays, weekends, during the spring (when tourism and the resident population increases), and in the most recent years (due to a gradual increase in the regional population). While estimation would be simplified by defini ng a single model with interactions to capture individual restaurant-lev el differences, it would substan tially increase the number of explanatory variables and thereby complicate the presentation and an alysis of results. In addition, the diversity between restaurants (especially with respect to rest aurant size, type, and unique changes to each during the study period) supports a unique set of explanatory variables that would further complicate the estimation and anal ysis of results from a single model. Thus, separate models will be estimated for each restaurant. Proprietary Sales Data Daily sales were obtained for three beachfront restaurants located directly on the Gulf of Mexico in Southwest Florida. All restaurants were located within fifty feet of the waters edge. Seating capacity ranged from 360 to 500 guest chairs. Outdoor seating accounted for approximately 35 to 50 percent of total seating capacity. The data cover November 1, 1998 24


through December 31, 2005 and include gross sales for each day ( Yt), for a maximum of 2,032 observations. The three restaurants differed in te rms of the average price of menu items; one was considered more up-scale with the highest aver age menu price, one was moderate, and one was relatively casual and had the lowest average prices The restaurants are generically referred to as firm A, B, and C, respectively, to maintain c onfidentiality. All restau rants were open year-round with the exception of Christmas day; however, th ere were a few days of planned closures for maintenance and renovations. The sales data were adjusted for inflation using the Southern regions food-away-fromhome monthly consumer price i ndex (CPI) (Table 2-2). Average daily CPI-adjusted sales (to December 2005 dollars) varied in magnitude and apparent trends between restaurants over the study period (Figure 2-1a). Daily sales for firm A, the smallest restaurant, were relatively unchanged with an average of $2,626 (Figure 2-1a). Firm B, with the highest average daily sales of $24,347, experienced above-inflati on gains in daily sales due to continual updates to the facility, steadily increasing prices and substantial market growth. The average daily sales of firm C were $6,357, although sales increased in 2004 from the addition of a 90-seat banquet area (this change in level of CPI-adjusted sales is less a pparent in Figure 2-1a due to the scale of the vertical axis). The information on these infrastr ucture changes to firms B and C were used to create dummy variables (D ) to allow the model to capture th ese exogenous effects in the model estimation. In general, all three restaurant s received the highest average daily sales in the months of February though May, and the lowest in September (Figure 2-1b). Sales of firm A, the smallest restaurant, ranged from $1,686 to $4,269 depending on the month. For comparison, sales of the largest restaurant, firm B, ranged from $16,527 to $33,895 depending on the month. To capture 25


these seasonal effects, monthly dummy variab les were included in each model. Similarly, variation is present throughout the week with peak sales being on or near weekends and the lowest sales coming early in the week. Thus, du mmy variables were also included to account for these effects in the model. Environmental Data Information on daily environmental conditions that were believed to impact sales was obtained from the manager of each restaurant. Th ese factors included the presence of noticeable red tide effects, whether rainfall occurred, or wh ether a storm of at leas t tropical strength was ongoing (Table 2-3). Using this data source, red ti de events were noted to occur on 52, 55 and 54 days (approximately 2.7% of observations) for fi rms A, B and C, respectively, during the study period. The number of rainy days for all three re staurants averaged between seven to 14% of all operating days. Tropical storms or hurricanes were noted on 15, 14 and 15 days for firms A, B and C, respectively. Temperature and wind speed data were obtaine d from the University of South Floridas (USF) data station. As these data were measured every six to eight minutes, the analysis used the average of measurements from 11 am through midnight to correspond with the operating hours of each restaurant. In general, seasonal variati ons are evident with respect to both temperature and wind speed, which are inversely related from March through October (Figure 2-2). A supplementary source of environmental data was obtained from a monitoring station that is maintained by the National Climatic Data Center (NCDC). For the purposes of this model, the NCDC average daily temperature data were s ubstituted for those days (73 in total or 3.6% of the 2,032 observations) when the primary USF da ta were missing. The absolute variation in temperatures between the two data sets over the study horizon ranged from -1.57 F to -0.58 F. 26


To account for this deviation, the NCDC temperat ure observations where adjusted by the average monthly differences in the cases where NCDC data substituted in for missing USF data. Results Model Estimation and Evaluation Following Greene, condition numbers were calculated as measures of multicollinearity for each model and all were acceptable (i.e., unde r a value of 20), which allows for efficient model estimation. Each model was examined fo r evidence of autocorrelation using DurbinWatson test statistics. The null hypothesis of no autocorrelation was rejected for all firms as autocorrelation of degree one was found. Thus, th e models for all firms were estimated with generalized least squares to correct for the presence of correlated error terms in the first period. The estimated models are shown in Table 2-4. Estimated models for firms A, B, and C had adjusted coefficients of determination of 64.6% 67.4% and 65.9%, respecti vely, indicating that the models appear to fit the data relatively well and consistently across locations. The signs of the parameter estimates corre sponding to the envir onmental variables ( X ) were as expected in all models, i.e. temperat ure was positive and wind speed, red tides, rain, and storm events were negative. The parameter estimat es for all environmental variables, with the exception of red tide in firm A, were statisti cally significant. Of all the time-related dummy variables, only those trying to capture differences in early-week sales (i.e., Tuesday and Wednesday versus Monday) were statistically in significant in each model. For the larger firms (B and C), sales were not found to differ in some fall months (September for firm B and August and October for firm C) from January sales, a lthough these are considered off-peak seasons for this region. Lastly, the annual dummy variables for firm B that were intended to capture continual changes made to the menu and facili ty over the study period indicated that these changes did not begin to affect sales until 2000. 27


To facilitate discussion of model results, a base estimate of daily sales was calculated for each of the three firms using only the average daily temperature and wind speed explanatory variables. These base daily sales estima tes were calculated to be $2,630, $24,361 and $6,367 for firms A, B, and C, respectively. When compared to actual average daily inflation-adjusted sales (Table 2-2), these estimated daily sales values were nearly identical. Therefore, the results discussed in this study are compared to the ac tual average daily inflation-adjusted sales. Effect of Red Tides For two of the three restaurant s, the estimated models reveal ed a statistically significant reduction of daily sales when a red tide was pr esent. Firm A, the lowest-grossing, was the only restaurant where the red tide pa rameter estimate was not statis tically significant during the study period (Table 2-4). Firm B, the highest-earning restaurant with CPI-adjusted average daily sales of $24,347 (approximately 4 to 10 times larger than the other two restaurants), experienced the largest absolute and relative decline due to a red tide event. Firm B incurred a statistically significant decline of $3,734 (15.3%) each day that red tide conditions were noticeable enough for the manager to document. Daily sales for firm C also experienced a d ecline during a red tide event, with an $868 (13.7%) reduction when a bl oom was present. Given the number of days of reported red tide events for each restaurant (Table 2-3), total losses during the seven year time horizon are calculated to total $252,242 for firms B and C, respectively. Effect of Red Tides Relative to Other Environmental Factors (X) All of the other four environmental factors were statistically signi ficant in each model. For the continuous variables (temperature and wind speed), relative changes in inflation adjusted sales were measured assuming an increase equal to one standard deviat ion. Average daily sales were found to increase by 3.0% to 6.3% due to a one standard deviation increase in temperature and decrease by 4.4% to 4.7% from a one standard deviation increase in wind speed. In absolute 28


value, these effects of temperature and wind sp eed are approximately one-quarter to one-third (i.e., the ratio of the effects to the red tide coefficient ranged fr om 0.22 to 0.35) the magnitude of effects on sales from red tides for firms B and C that had a statistically significant red tide parameter estimate. If the restaurant manager recorded ra in, daily sales fell 23.0% to 27.0% across restaurants. The size of this e ffect is larger (with coeffici ent ratios of 1.52 and 1.98 for the rainfall to red tide coefficients, or one and a ha lf to two times larger) than that caused by red tides. Calculation of the effect on an annual aver age basis revealed that the heavy rainfall caused lost revenues of approximately ten times those caused by red tides. This is not surprising, given that a subjective measurement of a rainy day is more liberal as rainfall events are a more common weather occurrence. Tropical storms or hurricanes had relatively la rger effects on daily sales, i.e., a 20.8% to 40.1% decrease in sales. The ratios of the st orm to red tide coefficients were 1.35 and 1.65, indicating storm effects exceeded red tide effects, on average, by approximately one-third to two thirds each day. When calculated on an annual average basis to factor in inci dence, the effect of tropical storms or hurricanes on firms B and C was approximately one-thir d to one-half that of red tides. Given the number of days of reported tr opical storms and hurricanes for each restaurant (Table 2-3), total losses due to storm events dur ing the seven year time horizon were calculated to be $65,728 and $20,034 for firms B and C, respectively. Effect of Time-Related Factors (D) Holidays generated increased sales of 21.4% to 29.9% across firms. An average of six holidays per year resulted in annual re venue increases of $3,378, $33,090, and $11,412 for firms A, B and C, respectively. `The highest average daily sales occurred on Saturdays for each firm, when sales increased by $1,469 to $15,283 or 55.9% to 62.8% for firms A and C, respectively, 29


above early-week sales (i.e., Monday through We dnesday). For comparison, the peak winter month (i.e., March) generated daily sales increases of $1,957 to $17,732, which translates into increases of 71.9% to 74.5% across firms. The seasonal demand has a much larger relative impact on sales than any other factor. Firm B was the only restaurant that experienced a notic eable increase in inflationadjusted sales in the long run, that is, acro ss the seven-year time hor izon. Compared to 1998 and 1999, sales increased from $3,165 in 2000 to $13,753 in 2004. Thus, the menu price increases and infrastructure improvements resulted in s ubstantial increases in real gross revenues. Similarly, the renovations to firm C in 2004 increased average daily sales by $1,103 (17.4%); over the course of a year, the renovations contributed to increased sales of $402,595. Discussion The regression analyses revealed that red tid e events reduced daily receipts at the two highest priced beachfront restaurants in the study. This result was found using environmental data observed by the manager of the restaurant that was on duty during business hours. Like the data on rain and storm days, the designation of the presence of red tide conditions that were sufficient enough to affect sales as perceived by th e manager are essentially subjective data. The benefit and uses of such data are becoming in creasingly common, most notably the use of beach conditions data with respect to water recreational activ ities (Caldwell). In the case of red tides, the observations are likely conservative because the notes were only made when the red tide conditions (e.g., noxious airborne toxins and/or dead fish washed up onto the beach) were indisputable. Moreover, the nega tive effects of red tide blooms on restaurant patrons can vary rapidly (due in part to the in fluence of wind speed and direc tion), suggesting that off-site monitoring by state officials may no t provide relevant data necessa ry to capture economic effects of the restaurant sector. 30


In the absence of routine, on-site, red tide monitoring stations, direct observations (subjective determinations) were used. To provide comparison, all red tide observations noted by the restaurant managers were found to corres pond with cell counts that averaged 180,853 cells per liter within seven days and six miles of the western edge of the County (as measured by Fish and Wildlife Research Institute [FWRI]). When FWRI recorded cell counts and managers noted a red tide on the same day (13 in total), FWRI cell counts averaged 585,183 cells per liter. These average cell count measurements greatly exceed th e 5,000-threshold level that is used to close commercial shellfish harvesting areas (Florida Department of Agriculture and Consumer Services). Thus, cell count measurements may need to be much higher to impact beachfront restaurant patrons when compared to the shellfish closure threshold levels, pa rticularly in light of corresponding wind speed and wind di rection (Backer et al.). If th ese thresholds are supported in other studies, they could be used to estimate red tide effect thresholds for any beachfront business and, thereby, provide support for the futu re use of subjective red tide data in empirical analysis. Firm-level analysis is, however, essen tial since red tides were not found to affect all restaurants in our study despite thei r close proximity to one another. The economic sustainability of beachfront rest aurants, as with any natural resourcerelated firm, is dependent on the condition of the natural environment, which is largely uncontrollable. Recent scientific advances have, however, suggested a suite of potential prevention, control and mitigation strategies for re d tides (e.g., Casper; Robbins; Pierce et al.; Schneider, Pierce and Rodrick). One result has been improved forecasting models (e.g., Stumpf). The cost of such strategies will need to be compared with potential benefits, which could be proxied with preventable losses to affected busine sses, such as the restaurants measured in this research. These estimated benefits also provide support to the hypothesis that the restaurant 31


sector is affected during red tides along with the traditional commercial fisheries or marinerelated recreation and tourism sectors which seem to have received the most media attention (e.g., Glick; Huettel; Karp; Morone y; Moore). Like the fishery sect or, the restaurant sector could demonstrate its eligibility for disaster assistan ce, such as loans offered by the Small Business Administration, using the empirical approach developed in our study. Conclusions Red tides have occurred in the Southeastern U.S. for more than 150 years of recorded history (FWRI) and have been anecdotally accu sed of causing a suite of negative economic effects. Our study used proprietary data on a more fine geographic, temporal and industry sector resolution to provide empirical evidence of the magnitude and re lative size of economic losses from red tides. Results were generally consistent with the available scie ntific literature (e.g., Larkin and Adams). Statistically significant dail y sales losses due to a re d tide day are relatively close to the minimum Small Business Administra tion (SBA) loan values provided to Florida restaurants in 2006 (Tester. P.A. Personal Communication. NOAA National Ocean Service, 13 July 2007). Each of the restaurant managers observations had notat ions of up to six consecutive red tide days, which suggests that cumulative sales losses based on duration of a red tide event may have driven SBA loan values to the maximum end of the range. These results indicate that onsite record-k eeping of weather even ts by individual firms represented a valuable source of data for determining statisti cally significant support for the absolute and relative effects of extreme envir onmental events. These records are of primary importance should waterfront firm owners consider appealing to state or federal governments, or private industry, for financia l loss reimbursements. For example, the SBA Economic Relief Funds are a potential source of re venue replacement in the case wh ere a firm demonstrates that red tide blooms diminished daily sales. Some Florida counties are reimbursing cities for the 32


beach clean-up costs incurred during a red tide, and evidence of beachfront firm losses may encourage such tourism-dependent county governments to reimburse businesses for their private clean-up expenditures. In addition, privateor publicly-underwritten hazard or business disruption insurance could use forecast and associ ated probability of bloom estimates to include red tide events as they currently do for hurricane or flood policies offered in the market. The methodology and results may also provide a means to extrapolate findings to a regional level for the restaurant sector using the cell count thresholds number of days, and number and type of beachfront restaurants. Such losses can be compared to estimated costs of any proposed red tide prevention, control and mi tigation strategy for other business dependent on the economys tourism sector. This methodology c ould be applied to other businesses sectors (commercial fishing, lodging, beach attendance, et c.) to get a better estim ate of regional costs and benefits associated with proposed red tide control, mitigation and management. 33


34 Table 2-1. Variable descriptions and definitions Variable Definition (units of measure) Y i Inflation-adjusted gross sales for firm i ($) X j = TEMP Average temperature fro m 11am midnight (F) X j = WIND Average wind speed from 11am midnight (meters/second) X j = RTIDE Red tide (1 if yes; 0 if no) X j = RAIN Heavy rainfall (1 if yes; 0 if no) X j = STORM Tropical storm or hurricane condi tions (1 if yes; 0 if no) D k = HOL Holiday, with the exception of Chri stmas Day (1 if yes; 0 if no) D k = DAY1-DAY7 Sunday through Saturday, respectiv ely (1 if yes; 0 if no) D k = MTH1MTH12 January through December, respect ively (1 if yes; 0 if no) D k = YEAR98-YEAR05 Years 1998 through 2005, respectively (1 if yes; 0 if no) D k = EXPAND Expanded seating area for firm C in 2 004 (1 if year 2004 or 2005; 0 if not) Table 2-2. Descriptive statis tics for continuous variables Variable N Mean Standard deviation Minimuma M aximum Y i = Firm A 2,023 $2,626 $1,312 $0 $7,947 Y i = Firm B 2,025 $24,347 $11,677 $0 $80,868 Y i = Firm C 2,023 $6,357 $2,918 $0 $16,589 X j = TEMP 2,032 72.0 F 9.2 F 36.4 F 87.5 F X j = WIND 2,032 4.6 m/sec 2.7 m/sec 0 m/sec 18.4 m/sec a A zero CPI-adjusted daily sales corresponds to clos ures during hurr icanes, which are captured with the X j = STORM variable. The incidence of tropical storms and hurricanes as noted by the manager are shown in Table 2-3.


35 Table 2-3. Environmental data by firm reported as the number of days observed Red tide ( X j = RTIDE) Rain (X j = RAIN) Tropical storm/hurricane ( X j = STORM) Year Firm A Firm B Firm C Firm A Firm B Firm C Firm A Firm B Firm C 1998a 0 0 0 2 2 2 1 1 1 1999 1 1 1 32 35 36 2 2 2 2000 0 0 0 39 40 39 1 1 1 2001 8 11 8 26 25 25 2 2 2 2002 4 5 5 48 51 51 1 0 0 2003 1 1 1 32 35 37 1 1 1 2004 1 1 1 50 49 51 4 4 5 2005 37 36 38 44 44 45 3 3 3 1999-05: Total 52.0 55.0 54.0 271.0 279.0 284.0 14.0 13.0 14.0 Mean 7.4 7.9 7.7 38.7 39.9 40.6 2.0 1.9 2.0 a Only November and December were included in the 1998 data.


Table 2-4. Estimation results by firm Firm A Firm B Firm C Variable Coef. Pr > |t| Coef Pr > |t| Coef. Pr > |t| Intercept 907*** 0.0004 -268 0.9044 2,636*** <.0001 X j = TEMP 18*** <.0001 114** 0.001 21** 0.0182 X j = WIND -44*** <.0001 -397*** <.0001 -111*** <.0001 X j = RTIDE -88 0.5808 -3,734** 0.0086 -868** 0.0118 X j = RAIN -604*** <.0001 -5,693*** <.0001 -1,719*** <.0001 X j = STORM -1,052*** 0.0001 -5,056* 0.0296 -1,431** 0.0166 D k = HOL 563*** <.0001 5,515*** <.0001 1,902*** <.0001 D k = DAY1 923*** <.0001 7,624*** <.0001 2,013*** <.0001 D k = DAY3 -72 0.2719 -401 0.4728 -74 0.6355 D k = DAY4 -33 0.6146 279 0.6186 117 0.4562 D k = DAY5 152* 0.0206 2,184*** <.0001 511** 0.0012 D k = DAY6 844*** <.0001 9,874*** <.0001 2,162*** <.0001 D k = DAY7 1,469*** <.0001 15,283*** <.0001 3,607*** <.0001 D k = MTH2 1,024*** <.0001 9,397*** <.0001 2,703*** <.0001 D k = MTH3 1,957*** <.0001 17,732*** <.0001 4,572*** <.0001 D k = MTH4 1,625*** <.0001 16,639*** <.0001 4,433*** <.0001 D k = MTH5 527*** <.0001 10,690*** <.0001 2,387*** <.0001 D k = MTH6 -329* 0.0117 7,346*** <.0001 1,041*** 0.0003 D k = MTH7 -338* 0.0112 8,643*** <.0001 1,181*** <.0001 D k = MTH8 -798*** <.0001 3,268** 0.0047 -107 0.7148 D k = MTH9 -776*** <.0001 320 0.7853 -811** 0.006 D k = MTH10 -271* 0.0259 4,895*** <.0001 500 0.0591 D k = MTH11 217* 0.0326 5,153*** <.0001 654** 0.0032 D k = MTH12 -315*** 0.0006 1,895* 0.0205 -410* 0.0403 D k = YEAR98 1,400 0.1971 D k = YEAR00 3,165*** <.0001 D k = YEAR01 4,854*** <.0001 D k = YEAR02 6,548*** <.0001 D k = YEAR03 8,194*** <.0001 D k = YEAR04 13,753*** <.0001 D k = YEAR05 13,566*** <.0001 D k = EXPAND 1,103*** <.0001 36


$0 $3,000 $6,000 $9,000 $12,000 $15,000 $18,000 $21,000 $24,000 $27,000 $30,000 $33,000 19981999200020012002200320042005 (1a) Annual $0 $3,000 $6,000 $9,000 $12,000 $15,000 $18,000 $21,000 $24,000 $27,000 $30,000 $33,000 $36,000 JanFebMarAprMayJunJulAugSeptOctNovDec (1b) Monthly Firm A Firm B Firm C Figure 2-1. Average daily CPI-adjusted sales fo r each firm by year (a) and month (b), 19982005 37


55 60 65 70 75 80 85 JanFebMarAprMayJunJulAugSeptOctNovDec 3.0 3.5 4.0 4.5 5.0 5.5 6.0 6.5 7.0 Temperature Wind speed oF m/sec Figure 2-2. Average daily temperature a nd wind speed by month, Southwest Florida. 38


CHAPTER 3 RED TIDES AND PARTICIPATION IN MARINE -BASED ACTIVITIES: ESTIMATING THE RESPONSE OF SOUTHWEST FLORIDA RESIDENTS Introduction With over 22 million people participating in 2000, Florida was the number one destination for marine recreation in the Unite d States (Leeworthy and Wiley). In particular, Florida led the nation as the number one saltwater fishing destination, with 4.7 million individuals spending 56 million days in 2000 (Leeworthy and Wiley). A review of literature provided a range of values from $60 to $100 for a day of recreational saltwat er fishing day in Floridas Gulf Coast, generating potential annual non-market valu es ranging from $3.4 to $5.6 billion in 2000 (Kildow). Florida offers more than 4,500 mari ne fishing access points and more than 2,300 beach and shoreline access sites (Divers et al.). Aside from s upporting a relatively large marinebased tourism industry, 77% of Floridas population resides in coastal coun ties (Kildow). The continued attraction of Flor idas marine recreational area s is predicated on the supply of high quality marine ecosystems that provide healthy beach and recreational experiences. For beaches in particular, poor water quality has been found to reduce the value of a trip to the beach (Freeman). For example, visitors to 23 beach es in the United Kingdom indicated that the scenery, bathing safety, and environmental quali ty (e.g., water quality and absence of sewage debris, litter and unpleasant odors) were the most important factors determining beach choice (Morgan). Ballance, Ryan, and Turpie found th at cleanliness was the primary factor behind beach choice in Cape Peninsula, South Africa, and that up to 97% of the beach value could be lost by a decline in standards of cleanliness. While these results are likely not surprising, Pendleton, Martin and Webster found that residents of Los Angeles, California, predominantly viewed the ocean as a place of pollution, not fo r swimming despite several heavily advertised 39


and successful clean-up campaigns; the authors conc luded that the perceptions of coastal water quality may be influenced more by the media th an with current coastal education campaigns. Using a combination of travel cost surveys a nd public opinion polls, these studies confirm the importance of high-quality water and beach conditions to participants in marine-related activities. They also revealed the potential pow er of perception over proof with respect to residents knowledge of local water conditions. In perhaps the only study of the effects of red tide on recreation, Nunes and van den Burgh assessed the economic value of a program intended to prevent harmful alga l blooms (HABs) at a famous beach resort in Holland. Following a joint travel-cost and contingent valuation approach, the program would be feasible if costs did not exceed 225 million Euro ($302 million in 2007). Their results also indicated that residents living closer to the beach placed a higher value on the HAB prevention program as compared to those th at had relatively highe r travel costs, which highlights the value of nature-based recreation to local residents. While non-market valuation studies are nece ssary to evaluate policy proposals that improve environmental quality and, thus, r ecreational opportunities, information on actual behavioral responses is needed to accurately esti mate any change in use values. For red tides in particular, this information is needed to justify continued expe nditures on preven tion, control and mitigation strategies. Behavioral responses can be evaluated with choice modeling. In particular, random utility modeling (RUM) has been used extensively for th is purpose, as evidenced by a recent summary of several recreational fi shing site choice models that have been published over the last twenty years (Hunt). The majority of th ese models rely solely on the ch aracteristics of the sites to explain the choice decision. Such a model specifi cation entails the estimation of a conditional 40


logit model (e.g., the choices are contingent upon the choice set, which complicates the estimation). If the choices are also hypothesized to depend on characteristics of the individual making the choice, then a mixed logit model is estimated. Alternatively, if the choices are believed to be driven solely by characteristics of the individuals, then it is most appropriate to estimate a multinomial (or generalized) logit model. The underlying RUM methodology can be used to examine the behavioral response to a red tide event, which are either to cut short the activity (i.e., decide the conditions are too unpleasant and go home early), delay the activity (e.g., postpone it from today to tomorrow), or relocate by deciding to go furthe r up or down the coast. Since these alternatives are not correlated in terms of their characteristics, the choice is assumed to solely depend on the preferences and constrai nts of the individual. Thus, multi nomial logit modeling is used to examine behavioral response to red tide events in this paper. This paper examines the behavioral respons e of individuals when faced with red tide conditions. Specifically, we want to know what factors affected the underlying decision of whether an individual reacted to a red tide event and, if they did, what factors determined whether they cut short, delayed, or relocated thei r participation in each activity. Using data from residents of two Southwest Florida counties that have experienced the most red tides, probability based models are used to examine behavior both across all marine-based activities and for four specific activities, namely: beach going, fishing from a boat or pier, and patronage of coastal restaurants. Results are important since resi dent responses are needed to estimate losses in recreational values associated with marine-based activitie s and to guide extension efforts. For red tide research in particular, results will be timely since red tide s may be occurring with more 41


frequency (Brand and Compton), and several potential control and mitigation strategies are currently under considera tion (e.g., Casper et al.; Robbins et al.; Pierce et al.; Schneider, Pierce, and Rodrick). Red tide cell count measurements collated by scientists at the National Oceanic and Atmospheric Administration (NOAA) that were within six miles the shorelines of Manatee and Sarasota Counties revealed positive (greater than zero) cell counts in January, March, August and September of 2000. However, a search of five local newspapers revealed 75 articles which included the terms Florida red tide, in the study area. These news articles appeared in every month of 2000 and ranged from three daily items each in March, July, August and November, up to 12 and 13 daily pieces in January and June, respectively. Theoretical Model of Behavioral Choice The theoretical models developed in our study are based on previous work concerning the rational choice perspectives of consumer be havior (McFadden, 1972). The underlying utility theory assumes individual i is faced with a set of K mutually exclusive choices. Following Gujarati (1995), the decision of the ith respondent to make choice k depends on an unobserved utility index, Uik, which is determined by the set M of explanatory variable s. The larger the value of index Uik, the greater the probability the respondent makes choice k If individual i selects alternative k over alternative l then we know that their utility from the former must have been larger than from the latter (i.e., Uik > Uil). From this, we can assume that they would choose alternative k such that Uik is the maximum utility choice among K choices, and therefore the statistical model is driven by the probability that choice k is made: Pr[Uik > Uil] for all other k l (3-1) To represent actual consumer behavior we compute the probability that Uik > Uil from the cumulative logistic distribution function. Assumption of logistic distribution of the disturbances 42


produces a logit model, which provides parameter estimates and reveals information on the unobservable utility. Logit transformations are lin ear in the explanatory variables, while the probabilities are nonlinear, which allows for dic hotomous response variables that take values ranging from 0 to 1. The multinomial logit model for a consumers choice is, then: ,,, 1 M ikimimik mYX (3-2) where Y* represents the choice k made by the ith respondent. As specified, the utility received by individual i from selecting k is composed of a deterministic and random component ( X and respectively). The random error term is included to capture any unobserved characteristics that might have influenced the probability and is assumed independent of the remaining probabilities. Parameter estimates are computed by maximi zing the log-likelihood function such that the probability of observing the actual choices is ma ximized. For ease of determining the effect of a change in one of the explanatory variables on the likelihood of a choice, the marginal effects were calculated for both continuous and dummy va riables. Following Greene, marginal effects for dummy variables generally produce a reasonable approxima tion to the change in the probability that Y equals 1 at a point such as the regressor means. An Application to Red Tide Events Data A total of 1,006 households in Sarasota and Manatee Counties in Southwest Florida were selected by random digit dialing in early 2001. One adult within each household was then randomly selected by asking to speak with the person over the age of 18 whose birthday was closest to the day of the interview. A call back procedure was employed to speak only with the 43


selected adult. Of those 1,006, 894 we re aware of the term red ti de and, as such, represent the population of residents potentially impacted by such events. Respondents were asked about thei r general awareness of red tide events and their level of participation in four activities during the previo us 12 months, namely: (1) saltwater fishing from a boat, (2) saltwater fishing from a pier or beach, (3) beach-going, and (4) patronage of restaurants located near the beach or bay front areas. Several questions where asked to ascertain their level of knowledge concerning red tides in addition to traditional socio-economic information. Collectively, these variab les will be used to explain how (if at all) a resident reacted to a red tide in terms of alte ring their participation in, or their reaction to, such events. A total of 755 respondents reported participating in at least on e of the four activities during the previous year and, thus, cons titute the sample used in this analysis. Of th is total, 80% reported being full-time residents in the study area and the average length of residency in Florida was 16 years (Table 3-1). Seventy percent of all in terviewees spent at least some time in college, while 18% reported a gross annual household in come of at least $75,000. Respondents averaged 55 years old, but the range of reported ages reveal s significant variation (i .e., 18 to 89 years old). Of the 755 respondents with complete inform ation (with the exception of income), 663 residents or 88% indicated that th ey frequented restaurants located near coastal waters at least once per month. A majority of residents (76%) in dicated that they had spent at least one day engaged in beach-going activities during the prev ious 12 months. Thirty-one percent and 28% enjoyed saltwater fishing from a pier or boat, re spectively, of all residents interviewed in the two counties. Overall, 66% of the sample indi cated that the presence of a red tide event impacted their participation in at least one of the four mari ne-based activities during the previous 12 months. 44


Residents reactions to a red tide bloom were relatively minimal among restaurant patrons and much stronger for beach-goers, ranging fr om 37% to 70%, respectively (Figure 3-1). Respondents who reacted were asked to identify how red tide altered their participation in each activity. For those engaged in beach-going, boat or pier fishing, respondents were asked if they cut short, delayed, or relocated when a re d tide event occurred, while restaurant patrons were asked only to specify whether they delayed or relocated th eir dining experiences. Overall, across the beach going and fishing activ ities, the majority of respondents (56% to 58%) delayed their participation. An additional 18% to 26% stated that they relocated to another area when faced with a red tide occurrence at their original de stination. Another 17% to 24% indicated that a red tide bl oom forced them to cut short their pa rticipation. For restaurant patrons, 64% stated they relocated and 36% chose to delay their dining experience. Empirical Models The empirical analysis began with the speci fication of the underlying dichotomous choice model to explain the probability th at an individuals participation in any one of the four marinebased activities will change during a red tide. Sinc e we are using stated be havior data, the model will explain the probability that indi vidual i reacted to a red tide (REACTi) by either cutting short, delaying, or re-locating any of the four activities in the pa st 12 months. To further refine the results, the model was re-estimated for each of the four activities to test whether specific explanatory factors had a unique a ffect on an individuals particip ation in a give n activity during a red tide (REACTi,j). Lastly, a multinomial logit mode l was estimated for each activity to explain the choice of reaction, whic h could be either to cut shor t the trip, dela y the trip, or relocate (REACTi,j,k). We hypothesize that certain indivi dual characteristics constrain the behavioral choice (i.e., reaction) and the relative importance of such characteristics is likely to differ by activity. 45


A total of 15 explanatory variables were hypot hesized to affect th e reaction (X) and are categorized into three types of variables: thos e designed to capture diffe rences in demographic characteristics (D), levels of participation in the various activ ities (A) and levels of knowledge with respect to red tides (K). The seven demogr aphic variables included whether the individual was a full-time resident of the study area (D_FULL), the number of years they have lived in Florida (D_YRS), whether they were male (D_G END), whether they had any college education (D_COLL), how old they were (D_AGE), whethe r they were Hispanic (D_HISP), and whether their household income was at least $75,000 last year (D_INC). Given that information on income was only available from a sub-set of th e sample (569 of the 755 total), the first model that explains whether the responde nt reacted (regardles s of the activity) was estimated with and without income. Whether an individual reacted (or not) to a red tide event, especially for each activity ( j ), is also hypothesized to depend on how often they pa rticipated in each activity. Thus, the initial model to explain a reaction acros s all activities is assumed to depend on the number of activities ( J = 4) that they participated in during the previous year (A_NUM). In the subsequent models for each activity, this total was repl aced with the average monthly le vel of participation during the previous year (A_j activity). Finally, three variables were created to captu re each individuals knowledge level with respect to red tide. The first, K_INFO, is the number of distinct red tide information sources; it is essentially the modes of information (e.g., televi sion, newspaper, workshops Internet, etc.) that the respondent has used for red tides. The second, K_DESC, is the number of red tide effects that the respondent could list. It wa s an open-ended question and all answers were unprompted. It is assumed to be a proxy for what the individual thi nks s/he knows about red tides. In contrast, the 46


third and final variable, K_REDT, is the number of correct true-false questions on the biology and human health impacts of red tide events. Th is variable tests for actual knowledge about red tide events. All models were analyzed using the LIMDEP statistical software package. LIMDEP develops starting values from the average of an initial ordinary least squares regression. The Newton-Raphson iterative algorithm was then used to calculate maximum likelihood estimators that were unbiased, consistent, and asymptoti cally normal (Amemiya). Parameter estimates are presented in terms of the marginal effects, where the partial derivatives of the cumulative distribution function with respect to the vector of characteristics were computed at the means of the explanatory variables. Hypotheses Since red tide blooms vary grea tly in their extent, location, duration, and unpleasant side effects, full-time residents or those that have liv ed in Florida longer were expected to have a higher probability of a reaction (e.g., decide not to participate in a marine-related activity). Having college experience or being in the highe st income category could identify full-time employees with the ability to pay for a change in plans (i.e., react) but not the time (do not react since cannot reschedule), so the expected signs of the coefficients are ambiguous. As individuals age, increased health concerns are expected to increase the probability of a reaction since this demographic is likely to have a more sensi tive respiratory system. Those individuals that participate in more activities ar e expected to have an increased likelihood of reacting to the presence of a red tide since they are more likel y to encounter red tides. Higher values for all three of the knowledge variables we re expected to increase the probability that a resident would change their behavior and deviate from their normal participation levels in an activity during a 47


red tide. This is because what has been written a bout red tide, especially in the popular press, has traditionally focused on the negative economic consequences. Empirical Results A general review of overall mode l test statistics revealed satisfactory levels of statistical significance. For all models, (reactions across activit ies, reactions within each activity, and type of reactions within each activity), convergen ce was achieved within five iterations. The coefficient estimates for each model maximized the value of the log-likelihood functions, ln L which ranged from -91.80 to -307.55 across all models Overall, the estima ted models correctly predicted respondent reaction choices of 0 a nd 1 between 56% and 75% of the time (where predicted values equal to one when the probabi lity exceeded a threshold level of 0.500 and zero otherwise) indicating reasonable performance acro ss models. Marginal effects expressions for the parameters were computed at the samp le means of the explanatory variables. Reaction Across Activities The dependent variable REACTi represented whether respondent i's participation in at least one of the four marine-based activities in the study had been affected by a red tide event. To determine whether missing income data was an em pirical problem, two models were estimated: one with and one without income (D_INC). The results indicated that the coefficients were robust to the inclusion of income as a regres sor (Table 3-2). Nevertheless, as income is theoretically expected to affect response, and since there are a sufficient number of observations remaining, we retain the use of the D_INC variable in lieu of increased observations. One of the most notable results was the lack of statistical signif icance of any of the demographic variables, including income, in th is model. The only statistically significant variables were those that indicated either the num ber of activities the resp ondent participated in, the number of red tide effects they could list, or the number of accurate red tide statements the 48


respondent identified. These results highlight the importance of a respondents participation in multiple activities and his/her ability to describe the effects of a red tide event, both unprompted and the correct responses to a set of questions. Th ere was a 13.5% increase in the probability (in absolute value) that a resident was affected by the presence of red tide for each increase in the total number of marine-rel ated activities s/he participated in during the previous year. For a resident that was able to accura tely recall an additiona l red tide effect, that individual was 10.1% more likely (in absolute value) to have their participation impact ed by a red tide bloom (i.e., they decide to either cut short, delay or relocate). Correct identification of an additional true-false statement resulted in a 2.5 percen tage point increase in the likelihood that th is resident altered their marine-related activity involvement during a red tide. Reaction by Activity Replacing the total number of activities (A_NUM) with the level of participation for a given activity (A_j activity), allows us to explai n differences in choice regarding whether to react (i.e., change participation) for each activity durin g a red tide. Estimation results are in Table 3-3. Only three of the demographic variables were st atistically significant in any one of the four models: the number of years re siding in Florida (D_YRS), wh ether the respondent was a fulltime resident of Florida (D_FULL), and whethe r gross household income was at least $75,000 in the previous year (D_INC) (Table 3-3). Overall, each additional year of residence increased the probability of a reaction by 0.4 percentage points (or 4 percentage points for each 10 years of residence). Residents in the higher income gr oup had a 28.6 percentage point lower probability of having their pier fishi ng activities affected by a re d tide event. Restaurant patrons that lived in the area year-round were 13.4 percentage points more likely to respond that red tide events affected their patronage of b each or bayside restaurants. 49


As hypothesized, the level of participation in each activity affected the probability that respondents were affected by red tide events, with the exception of boat fishing. This is, perhaps, obvious as boats can easily be moved to better cond itions. In contrast, each additional day of pier fishing per month increased the probability of a red tide affecting their pi er fishing activities by 4.1 percentage points, or 9.0, at th e average participation level. As was evident in the overall model, those individuals that were able to accurately describe red tide effects without prompting had a highe r probability of having their beach-going, boat fishing and restaurant patronage affected by a red tide; probabilitie s increased by 7.1 to 9.7 percentage points for each effect they provi ded. The more knowledgeable the respondent, the higher probability their beach-going activities woul d be affected by a red tide. At the average number of correct responses (compared to some one that knew nothing), the probability of an effect on beach-going increased 37.4 percentage poi nts. Finally, restaurant patrons that reported having more red tide information sources had a 5.1% higher probability (in absolute terms) of having their participation altered dur ing a red tide for each source. Type of Reaction by Activity The dependent choice variable in these models represented the multiple options for how respondents reacted to red tide ev ents for each marine-related activity. The choice of reactions included delay, cut short, or relocate fo r the beach-going, boat fishing, and pier fishing activities. The restaurant patrons were limited to two choices, either delay or relocate, as it was assumed that most consumers would not leave a restaurant once the dining location had been selected and food or drinks ordered. Results are presented in Table 3-4. Beach-going residents who were able to accurately recall red tide effects without interviewer prompting were 5.0 per centage points more likely to cut short their day at the beach for each effect described, while male residents we re slightly more likely (8.4 percentage points) 50


to relocate to another beach that was unaffected by a red tide. Saltwater boat fishers that lived in the study area year-round were 35.8 percentage points less likely to cut short thei r excursion, as compared to part-time residents. Those resident s with more knowledge of red tide effects were 11.2 percentage points more likely to delay their boat-fishing trips. Pier fishing residents with some college had a 21.7 percentage point increase in probability that they would cut short their fishing pursuits relative to those without colleg e experiences. Restaurant customers with more red tide information sources were 5.7 percentage points more likely to select another location as compared to the option of delaying their dining experience when a red tide bloom was evident. Summary and Discussion of Results Using a model to explain a respondents overall reac tion (or lack thereof) to a red tide event in terms of their participation in any one of four marine-based activities, revealed that the probability of a reaction (i.e., that an individual would react negatively by cutting short, delaying or relocating) increases with the number of act ivities they participate in and their knowledge level regarding red tides (measured either one of two ways). None of the demographic variables were found to have a statis tically significan t affect on this overall probability. Better refinement of the analysis was acco mplished through estimation of the probability of any reaction for each activity by replacing the number of activities with the average monthly level of participation in each activity. These re sults supported the conclusi on that the probability of a reaction (and the associated factors) varies by activity. The probability of a reaction for boat fishing was only affected by how many red tide effects the individual could describe: the more e ffects listed (such as fish kills and offensive odor) the higher the probability an individual will react (i .e., cut short, delay, or relocate). For beach or pier fishing, on the other hand, the more years of residence, lower income levels, and higher participation levels were statistically significant in explaining whether an individual 51


would react. Higher prob abilities of losses to restaurant patronage were supported for year-round residents, those that frequented restaurants more often, those that could describe more effects, and for those that had more information sources. Lastly, for the beach going activity, the residents who have lived in Florida the longest, spent more days at the beach on average, were able to recall accurate red tide effects, and were more knowledgeab le of red tide effects were more likely to react. In terms of what can be done to help mitig ate potential losses from red tide events, the knowledge-related variable results are further evaluated using figur es to facilitate comparisons and highlight the magnitude of the results. The nu mber of red tide effects that respondents could describe without being prompted (K_DESC) serves as a proxy for what the they believe they know about red tides. Figure 3-3 shows that the probability of an effect on beach-going and the probability of boat fishing being a ffected (Figures 3-3a and 3-3b, respectively) are similar; the probabilities increase from about 50% to over 90% when individuals can list more than five red tide effects. The effect on probability for restau rant patronage reveals the same magnitude of increase, but at a lower overall probability (Figure 3-3c), which is to be expected given that restaurant patronage is lower overall. These results (Figures 3-3a-c) can be contrasted with those of the variable that best describes what the individuals act ually know as ascertained from thei r ability to co rrectly answer 17 true and false questions about the nature and health effect s of a red tide bloom (K_REDT). Although the overall probability of a reaction is less sensitive (Fi gure 3-2c), the probability of beach-going being affected during a red tide (by individuals choosing to cut short, delay, or relocate) was only approximately 30% with one correct answer, however, this probability increased to over 60% with six correct answer s and exceeded 90% with 15 correct answers 52


(Figure 3-2f). Continued comparison of the ove rall and beach-going mode ls revealed similar sensitivity with respect to a residents menti on of red tide effects, where individuals that described one to five red tide e ffects resulted in higher probabilit ies of reacting of 50% to 90%, respectively (Figure 3-2b and 32e). Respondents engaged in at least one activity had a 50% probability of reacting to a red tide that increas ed to over 90% with participation in all four activities (Figure 3-2a), and t hose that spent a single day on average at the beach had a 70% probability of reacting to a red tide (Figure 3-2d). The final empirical analyses estimated the probabilities of specifi c reactions, namely, whether an individual would cut sh ort, delay or relocate their pa rticipation in response to a red tide. Notably, some additional demographic variable s were statistically signi ficant at this finer level of resolution (Figure 3-4); males had a high er absolute probability of relocating their beachgoing activities and a higher probability of re location away from a waterfront restaurant. Individuals that were over age 55 had a slightly lower probability of relocating to another restaurant. Those individuals that had attended some college were more likely to cut short fishing from a pier. Full time residents were, perhaps the least impacted; the probabilities of cutting short a boat fishing trip or of relocating patronage of a restaura nt were less. Figure 3-4 also shows the impact of red tide information source s and how much residents think they know; the former increased the probabilities of delaying a boat fishing trip or relocating restaurant patronage, the latter increased the proba bility of cutting short a beach trip. Conclusions This paper examined how red tide effects have influenced Sarasota and Manatee county residents participation in four marine-related activities. The findings suggest potential areas where public management prescriptions may be us ed to influence how a resident engaged in 53


marine-related activities chooses to react to the presence of a red tide. This is especially relevant since such information could serve to mitigate losses caused by red tides. The share of residents surveyed that were aff ected by red tide events ranged from a low of 37% for restaurant patronage to a high of 70% for beach going; this is not surprising as beach going entails more interaction with affected wate rs and aerosolized toxins. The more information sources an individual had, the mo re likely they were to delay their patronage or go somewhere else during a red tide. A respondent s existing knowledge of red tide effects (whether true or not) was associated with a higher probability of being affected across activities, and especially for beach-going. What a respondent actually knew, however, had no affect on the probability of a reaction for restaurant patronage The boat fishers that live in the area year-round were less likely to cut short their trip, perhaps due to their fa miliarity with these naturally-occurring red tides, and/or their ability to drive th e boat away from affected waters into better fishing territory. These findings offer strong support for provision of extension materials targeted towards residents engaged in marine-rel ated activities. Public managers may affect awareness and, therefore, participation deci sions by providing information wh ere conditions are optimal for beach-going as opposed to information on where conditions may be unpleasant or even harmful. Such reports are analogous to daily alpine ski reports, although providing them for red tides is more complicated in that bloom effects are not limited to a single, isolated beach location. That said, coastal communities offer a variety of nature -related activities that could be included such reports. These types of education activities may fo ster an atmosphere where coastal populations are protected from the negative aspects of a re d tide bloom, while simultaneously encouraging residents to frequent alternativ e, unaffected, local marine-related businesses and therefore retain consumer spending in the area. 54


Table 3-1. Variable names, descri ptions and statistics (n = 755) Variable Description Min Max Mean Std deva REACT Participation altered in any marine activity due to the presence of a red tide? (1 if yes, 0 if no) 0 1 0.66 N/A Demographic: D_FULL Full time resident (1 if 12 mo./year, 0 if not) 0 1 0.80 N/A D_YRS Duration of residency (years) 1 76 16.24 13.20 D_GEND Respondents gender (1 if male, 0 if female) 0 1 0.41 N/A D_COLL Formal education includes at least some college (1 if yes, 0 if no) 0 1 0.70 N/A D_AGE Respondents age (years) 18 89 55.00 16.58 D_HISP Respondent is Hispanic (1 if yes, 0 if no) 0 1 0.04 N/A D_INCb Annual income $75,000 (1 if yes, 0 if no) 0 1 0.18 N/A Activities:c A_NUM Participation in marine-related activities (number) 1 4 2.24 0.98 A_j activity Participation level in activity j (number per mo.) A_BCH: j = beach-going (avg. days/mo.) 0.1 30.4 3.45 5.52 A_BFSH: j = boat fishing (avg. days/mo.) 0.1 30.4 2.06 4.44 A_PFSH: j = pier fishing (avg. days/mo.) 0.1 30.4 2.19 4.32 A_REST: j = restaurant patronage (meals/mo.) 1 52 3.44 5.16 Knowledge: K_INFO Information sources about red tide (number) 0 7 2.41 1.27 K_DESC Description of red tide (number of effects listed) 0 6 2.12 1.08 K_REDT Knowledge of red tide (number correct of 17) 0 16 9.35 2.59 a N/A indicates that the statistic is not applicable to this variable. b As a result of non-responses or refusals, n = 569 for this variable. c Beach going, boat fishing, pier fishin g, and restaurant patronage were enjoyed by n = 576 (76%), 215 (28%), 236 (31%), and 663 (88%) of the respondents, respectively. 55


Table 3-2. Marginal effects for factors hypothe sized to influence whether a residents participation in a marine-based activ ity is affected by a red tide (REACTi) Marginal effects (n = 569) Marginal effects (n = 755) Variable Coefficienta Standard error Coefficienta Standard error Constant -0.503* 0.153 -0.484* 0.131 Demographic: D_FULL 0.063 0.060 0.050 0.050 D_YRS 0.0001 0.002 0.001 0.002 D_GEND -0.045 0.043 -0.037 0.038 D_COLL -0.001 0.045 -0.000 0.041 D_AGE -0.002 0.002 -0.002 0.001 D_HISP 0.033 0.102 0.007 0.095 D_INC -0.087 0.053 N/A N/A Activities: A_NUM 0.135* 0.024 0.147* 0.022 Knowledge: K_INFO 0.011 0.017 0.015 0.015 K_DESC 0.101* 0.022 0.085* 0.019 K_REDT 0.025* 0.009 0.021* 0.008 a A single asterisk (*) indicates statis tical significance at the 5% level. 56


Table 3-3. Marginal effects and standard errors (in parentheses)a for factors hypothesized to influence whether a residents participation in each marine -based activity is affected by a red tide (REACTij) Activity j Variable Beach-going (n = 438) Boat fishing (n = 173) Pier fishing (n = 187) Restaurant patronage (n = 495) Constant -0.366 (0.149)* -0.086 (0.281) -0.574 (0.287)* -0.595 (0.156)* Demographic: D_FULL 0.008 (0.062) 0.199 (0.141) 0.226 (0.130) 0.134 (0.059)* D_YRS 0.004 (0.002)* 0.002 (0.003) 0.008 (0.003)* 0.002 (0.002) D_GEND -0.051 (0.046) -0.070 (0.078) -0.094 (0.083) -0.004 (0.046) D_COLL -0.021 (0.050) -0.017 (0.084) 0.083 (0.099) 0.028 (0.052) D_AGE -0.002 (0.002) 0.002 (0.003) 0.003 (0.003) 0.001 (0.002) D_HISP -0.034 (0.118) 0.017 ( 0.146) -0.155 (0.179) 0.152 (0.118) D_INC -0.076 (0.055) -0.134 ( 0.101) -0.286 (0.087)* -0.054 (0.052) Activities: A_j activity 0.019 (0.006)* 0.018 (0.012) 0.041 (0.016)* 0.0003 (0.0001)* Knowledge: K_INFO -0.002 (0.018) -0.002 (0.033) 0.019 (0.032) 0.051 (0.019)* K_DESC 0.097 (0.022)* 0.076 (0.039)* 0.066 (0.040) 0.071 (0.023)* K_REDT 0.040 (0.009)* -0.019 (0.016) -0.005 (0.017) 0.003 (0.010) a A single asterisk (*) indicates statis tical significance at the 5% level. 57


Table 3-4. Marginal effects and standard errors (in parentheses)a for factors hypothesized to influence the reaction of a resident ( k ) for activity j whose participation has been affected by a red tide (REACTijk) Activity j Reaction k Variable Cut-short Delay Relocate Beach-going (n = 312): Constant -0.243 (0.179) 0.450 (0.205)* -0.207 (0.154) D_FULL -0.057 (0.074) -0.054 (0.089) 0.112 (0.075) D_YRS -0.001 (0.002) 0.001 (0.002) -0.0003 (0.002) D_GEND -0.036 (0.052) -0.048 (0.059) 0.084 (0.043)* D_COLL -0.044 (0.058) 0.055 (0.067) -0.010 (0.051) D_AGE 0.0004 (0.002) -0.001 (0.002) 0.001 (0.002) D_HISP 0.030 (0.133) -0.138 (0.153) 0.108 (0.094) D_INC -0.005 (0.060) -0.042 (0.068) 0.047 (0.050) A_BCH -0.006 (0.006) 0.003 (0.006) 0.003 (0.004) K_INFO -0.023 (0.020) 0.032 (0.023) -0.009 (0.017) K_DESC 0.050 (0.024)* -0.035 (0.028) -0.014 (0.021) K_REDT 0.019 (0.011) -0.012 (0.013) -0.007 (0.010) Boat fishing (n = 109): Constant 0.374 (0.325) 0.002 (0.400) -0.375 (0.332) D_FULL -0.358 (0.156)* 0.310 (0.215) 0.048 (0.181) D_YRS 0.003 (0.003) -0.001 (0.004) -0.002 (0.004) D_GEND 0.013 (0.086) -0.170 (0.107) 0.167 (0.094) D_COLL 0.024 (0.088) -0.106 (0.113) 0.082 (0.099) D_AGE -0.004 (0.003) 0.003 (0.004) 0.001 (0.003) D_HISP 0.085 (0.169) -0.214 (0.217) 0.129 (0.165) D_INC -0.224 (0.140) 0.095 (0.149) 0.130 (0.113) A_BFSH -0.016 (0.016) 0.018 (0.015) -0.002 (0.011) K_INFO -0.069 (0.037) 0.112 (0.045)* -0.043 (0.037) K_DESC -0.018 (0.041) -0.028 (0.049) 0.045 (0.040) K_REDT 0.022 (0.018) -0.029 (0.023) 0.007 (0.019) 58


Activity j Reaction k Variable Cut-short Delay Relocate Pier fishing (n = 101): Constant -0.373 (0.253) 0.630 (0.393) -0.257 (0.373) D_FULL 0.075 (0.144) -0.392 (0.226) 0.316 (0.238) D_YRS 0.001 (0.002) 0.0003 (0.004) -0.001 (0.003) D_GEND 0.005 (0.071) -0.079 (0.107) 0.074 (0.100) D_COLL 0.217 (0.098)* -0.150 (0.138) -0.067 (0.125) D_AGE -0.0004 (0.003) -0.001 (0.004) 0.001 (0.004) D_HISP 0.104 (0.172) -0.357 (0.294) 0.253 (0.234) D_INC -0.161 (0.098) 0.115 (0.134) 0.047 (0.123) A_PFSH -0.001 (0.009) -0.004 (0.011) 0.005 (0.010) K_INFO -0.034 (0.026) 0.016 (0.037) 0.018 (0.034) K_DESC 0.008 (0.031) -0.008 (0.048) -0.001 (0.044) K_REDT 0.011 (0.017) 0.006 (0.024) -0.018 (0.022) Restaurant patronage (n = 190):b Constant N/A N/E 0.421 (0.278) D_FULL N/A N/E -0.184 (0.094)* D_YRS N/A N/E 0.005 (0.003) D_GEND N/A N/E 0.226 (0.074)* D_COLL N/A N/E 0.101 (0.094) D_AGE N/A N/E -0.008 (0.003)* D_HISP N/A N/E -0.201 (0.171) D_INC N/A N/E 0.118 (0.086) A_REST N/A N/E 0.0002 (0.0002) K_INFO N/A N/E 0.057 (0.028)* K_DESC N/A N/E 0.021 (0.037) K_REDT N/A N/E -0.016 (0.017) Table 3 4. Continued. a A single asterisk (*) indicates statistical significance at the 5% level. b N/A indicates that the reaction was not applicable to this activity,; N/E indicates that the reaction was not explicitly estimated for this activity; delay was the base category. 59


60 Figure 3-1. Reaction of reside nts (n = 755) to red tide events by marine-based activity ( j ) Pier Saltwater Fishing (n=236)Relocate 26% Delay 57% Cut Short 17% Effect 53% No Effect 47% Beach Going (n=576)Relocate 18% Delay 58% Cut Short 24% Effect 70% No Effect 30% Boat Saltwater Fishing (n=215)Relocate 23% Delay 56% Cut Short 21% Effect 63% No Effect 37% Restaurant Patronage (n=663) No Effect 63% Effect 37% Delay 36% Relocate 64%


Model REACTi (n=569) (a) A_NUM (b) K_DESC (c) K_REDT Model REACTij=Beach-going (n=438) 61 (d) A_j=Beach-going (e) K_DESC (f) K_REDT Figure 3-2. Predicted probabilities of participation in any activity (REACTi) and beach-going (REACTij=Beach-going ) due to a red tide event by the number of activiti es (A_NUM, fig. a) and the av erage participation days (A_j Beach-going; fig. d), respectively, the red tide effects the respondent descri bed (K_DESC; figs. b and e), and the num ber they know correctly (K_REDT; figs. c and f)


62 (a) REACTij=Beachgoing Model (n = 438) (b) REACTij=Boatfishing Model (n = 173) (c) REACTij=Restaurant Patronage Model (n = 495) Number of red tide effects the resident described (K_DESC) Figure 3-3. Predicted probability of participation in (a) beach-going, (b) boat fishing, and (c) restaurant patronage due to a red tide event by the number of red tide effects the resident described (K_DESC)


8.4% 5.0% -0.8% -35.8% 11.2% -18.4% 22.6% 5.7% 21.7% -40% -30% -20% -10% 0% 10% 20% 30% Relocate Relocate Relocate Cut short Relocate Cut short D_GEND D_FULL D_FULL K_INFO D_COLL D_GEND D_AGE K_INFO Delay K_DESC Relocate Cut short Figure 3-4. Change in predicted probability of response by marine-based activity for selected statistically significant variab les Beach-going Boat fishing Pier fishing Restauran t patronage Percent change in probability 63


CHAPTER 4 PUBLIC COSTS OF FLORIDA RED TIDE S: A SURVEY OF LOCAL MANAGERS Introduction Floridas attractive climate has served as a motivating force behind its past, present and predicted population and tourism growth. Be tween 1990 and 2000, the states population grew by more than three million people, a 23.5% increase, and this trend is expected to continue with the influx of migration from ot her states (Perry and Mackun). Overall, 77% of Floridas population resides in coastal count ies and these areas accounted for an estimated $402 billion, or 77%, of the states overall economy in 2003 (Kildow). As a popular tourist destination, Florid a hosted 77.2 and 6.4 million domestic and international visitors (respectively) in 2005 (VISIT FLA). In 2000, Florida was the number one U.S. destination for at least one of 19 types of marine recreation in cluding beach visitation, swimming, snorkeling, and scuba di ving, and its beaches alone attracted more than 15 million U.S. tourists (Leeworthy and Wiley). Changnon noted that the increasing movement of the population to coastal areas has increased coastal consumer and business exposur e, which in turn, has increased the economic impacts of naturally-occurring environmental st ressors. Harmful algal blooms (HABs) are one such stressor and the coas tal waters of Florida host many indigenous species of marine algae, including Karenia brevis which produces blooms known as red tides. Florida has long been familiar with these natu rally-occurring red tides, with the earliest recorded event noted over 150 years ago (Fish and Wildlife Research In stitute [FWRI]). The causative agent, Karenia brevis is known to produce a suite of as many as nine toxins, which are released in the water column and can kill marine life (Baden et al.). Red tide blooms contaminate shellfish beds and present public health concer ns due to decomposing marine life on beaches, 64


and the possibility of neurotoxin shellfish poisoning (NSP) if contam inated shellfish are ingested (Steidinger et al.; Flewelling et al.). Coastal wind conditions may exacerbate an algal bloom, and result in the release of airborne toxins that have been measured up to three miles inland on populated beach areas (Kirkpatrick, B., personal communication, Mote Marine Laboratory, 22 August 2006), where humans may expe rience eye, nose, and respirator y irritation (Backer et al.; Kirkpatrick et al., 2004). A 2001 report submitted by the National Ocean ic and Atmospheric Administrations National Sea Grant College Program to the Co mmittees on Appropriations suggested national and local socio-economic impacts of red tide events that need to be quantified and addressed in order to develop efficient, timely management re sponses. The report stated that research efforts are needed to determine the nature and extent of private and public sector interactions in the case of a HAB event. Once these affected areas and corresponding issues are identified, accurate and efficient management decisions and aid estimates could be calculated and implemented at both the local and national levels. Study Objective To our knowledge, no studies have attempte d to quantify costs incurred by resource managers to address the impacts caused by red ti des, despite the need for resource allocation decisions. The overall objective of our study wa s to solicit specific information on costs associated with red tide blooms from municipal and county-level managers located on Floridas Gulf Coast. This study was intended to quantify public ex penditures and procedur es resulting from red tide-related management and mitigation i ssues which have affected publicly managed beaches. Respondents were queried to determin e which county or municipal departments are responsible for beach and red tide manageme nt, their budget sources and allocations, the 65


existence and types of public relations efforts, and actual red tide-related beach cleaning protocols. This study sought to improve on previous analysis of direct HABs costs to the public by targeting a larger geographic re gion and using a standardized surv ey instrument. It is expected that the results of the survey effort will provide estimates of red tide-related expenditures incurred by local governments that can be used to guide financial planning for other public agencies. Study Area The total population for each county ranged from approximately 11,000 in Franklin to 905,000 in Pinellas, with an average of 307,000 pe rsons (Table 4-1). L ee and Collier counties each experienced population growth rates exceed ing 20% from 2002 to 2005. Manatee, Sarasota, and Charlotte also witnessed an influx of residents over this same time period, with populations increasing by 13.9%, 10.4% and 9.5%, respectively. Pinellas was the only county that experienced a negative growth rate, le ss than two percent, from 2002 to 2005. Okaloosa County hosted nearly three million vis itors in 2005 (Table 4-2), with slightly fewer visiting Pinellas and Lee counties, wh ich received 2.4 and 2.3 million domestic visitors, respectively (VISIT FLA). Sarasota and Collier e ach attracted more than one and half million tourists in 2005, while Manatee saw 770,000 visitors to its area. Due to sa mple size constraints faced by the marketing agency, tourist estimates for Franklin, Gulf and Charlotte counties were not available. With the exception of Franklin County, each of the remaining eight counties in the sample had some number of publicly managed beachfront, ranging from a maximum of 35 miles in Pinellas to a minimum of seven miles in Sara sota, for an average of nearly 17 miles (Table 43). While Sarasota had the fewest overall mile s, it had nearly as many public access point, or parks, as Pinellas, with each having 30 and 31, respectively. Access to public beaches may be 66


free or fee-based, and can include boardwalks, pier s, jetties, parking lots state parks, and in some cases are accessible by motorized vehicle. Each countys beaches exhibit a range of characteristics, from the sugar-white sands of the Panhandle beaches, the flat expanses in Pinellas, the barrier islands of Sarasota and Manatee, and the interwoven marshes of Collier. To support the tourism industries in these re gions, all of the c ounties collect tourist development taxes which are administered thro ugh the related Tourism Development Councils (Table 4-3). Tax rates ranged from two percent in Franklin to five percent in Pinellas, Charlotte, and Lee. The remaining five counties collected four percent in taxes on tourism-dependent business revenues. In total, FY 2005-06 tourist development collections ranged from $304,000 in Gulf, up to nearly $22 million in Pinellas, with an average of over $8 million across all nine counties. Lee and Collier ranked ju st behind Pinellas, with total to urist tax collections of $17 and $13 million, respectively. Okaloosa and Saraso ta each collected just over $7 million and Manatee slightly less than $5 million annually. Four of the counties indicated that at least 25% of annual tourism tax collections were earmarked for public beach management, maintenan ce, and improvements. In order to generate a scale of beach-related expenditures for each county, one-fourth of annual tourism tax collections were divided by miles of public beach managed by each county (Table 4-3). Tax collection records revealed that Sarasota a nd Lee Counties could spend $265,429 and $236,528, respectively, per mile of public beach. Pinell as and Collier Counties were each capable of allocating $164,023 and $148,364 per public beach mile. Census population estimates ranged from th e small community of North Redington Beach, with its 1,482 inhabitants, to the sprawling metropolis of St. Pete rsburg, with nearly a quarter-million individuals, both of which are locate d in Pinellas. Overall, 8 of the 18 cities, or 67


44%, were located in Pinellas County, while Mana tee had three cities, A nna Maria, Bradenton Beach, and Holmes Beach, with an average of 2,754 residents. Sarasota, Lee, and Collier counties were each represented by two cities with average populations of 35,240, 6,313, and 17,928, respectively. Longboat Key claimed county affiliations with both Manatee and Sarasota, and is home to an estimated 7,603 people. Procedures Nine Florida counties and 28 coas tal cities located within these regions were selected due to their proximity to Gulf waters, historical patterns of exposure to red tide blooms, and popularity as tourist destinations. The counties se lected (from northwest to southeast) were Okaloosa, Franklin, Gulf, Pinellas, Manatee, Saraso ta, Charlotte, Lee and Collier. In an effort to estimate the fiscal costs of red tide events at a local level, 28 municipali ties within each of the nine sample counties were targeted, based primar ily on their location to Gulf waters. A total of eighteen successful interviews were comple ted during January through March of 2007. The 18 municipalities are located within the boundaries of five of the nine counties Pinellas, Manatee, Sarasota, Lee and Collier (Table 4-4). Top level administrators within these locati ons were identified as the target sample. A database of names and contact information was compiled using the 2006 Membership Directory published by the Florida Association of Counties and the Florida League of Cities, Inc. The interviews were conducted between January a nd March 2007, and final contacts completed by March 2007 by a single person. The interviewer initially cont acted the top-level county or city administrator to obtain permission and recommendations with respect to id entifying appropriate i ndividuals within their organizations that would most likely have access to the information associated with red tide events. Due to the sporadic nature of red tide blooms and the complexity of government titles 68


and responsibilities across locales characterized by large ranges of population numbers, tourism dollars, and public beach areas, it was necessary to broaden the traditional scope of potential respondents. To that end, an effort was made to canvas all individua ls that were actively involved with beach management issues or funding and employee management responsibilities within each public agency. Respondents were first asked to discuss general beach management programs, and then queried about costs and activities specifically asso ciated with red tide events. Respondents were encouraged to describe general types of beach management or maintenance programs, and to provide data concerning fiscal year expenditures on both labor a nd equipment used in support of these programs. Questions pertaining to red tide events were designed to elicit detailed information for each responding county or city agency. The red tide-sp ecific section included act ual or estimated labor and equipment costs, evidence of communication protocols related to either clean-up activities or public relations, types of activities undertaken or sponsored by the agency, funds allocated to red tide mitigation or management, historical respons es to red tide events, and identification of agency departments charged with red tide-related responsibilities. Survey Results Responses were grouped into three informati on categories. The first category includes a list of all departments (public or private, local or state) char ged with physical responsibilities related to general beach maintenance and/or red tide bloom management or mitigation. The second category includes public funding sources for both beach and red tide management and actual or estimated fiscal a nd labor expenditures provided by these funding sources. The third category includes descriptions of any type of communication or activity protocols followed by 69


each county or city in the event of an active red tide bloom. Information will be summarized across counties and cities for each category following a description of the response rate. Response Rate In sum, 37 telephone interviews were attemp ted by the Florida Survey Research Center (Table 4-5). Six municipalities were either unr eachable, or unwilling to respond to the survey questionnaire. Of the total completed interviews, four responses were deemed ineligible, due to their distance from Gulf waters, or their lack of publicly managed Gulf-facing beaches. A total of 27 interviews were completed during the survey time period, which began in January and ended in March of 2007, for a response rate of 87.1%. These 27 respondents included all nine counties (Okaloosa, Gulf, Franklin, Pinellas, Manatee, Sarasota, Charlotte, Lee and Collier) and 18 cites located within these counties. Agencies Six counties involved at least two or more of thei r departments in the physical management of beach/red tide management res ponsibilities (Table 4-6). Parks and Recreation and Public Works/Utility departments were men tioned by the majority of county respondents (4 or more). Natural Resources or Environment/P ollution Control departments were mentioned by three of the counties. Sarasota a nd Gulf mentioned the i nvolvement of their local branches of the Florida Department of Healt h, while Gulf also included the Florida Department of Environmental Protection. Franklin and Pinellas both mentioned th e role of their Public Waste departments in fish kill cleanups. Sarasota and Lee referred to the roles of outside private contractors in water monitoring and beach clean ing responsibilities, respectively. Okaloosa was the only county to refer to the role of their Tourism Development Council. Finally, Pinellas, Manatee and Sarasota stat ed the inclusion of the Management and Budget Office, the Division of 70


Marine Rescue, and Emergency Services, respectively, as holding responsibilities for beachrelated physical management tasks. The majority of cities interviewed, 12 out of 18, or 67%, assigned physical beach or red tide tasks to their Public Works de partment, while more than half of all cities (10) hired private contractors, contract labor, consulting firms, commercial fi shers, marine inspectors, or equipment and boat rental suppliers to handle be ach cleaning work (Table 4-7). Five cities mentioned the top administrator, i.e. Town Clerk, City Manager, City Council, or Mayor, as having primary responsibilities for managing beaches and red tide events. Three cities, Sarasota, Venice, and Marco Island, noted that their beach es were physically cared for by their counties. Parks and Recreation and Natural Resource depart ments were mentioned by three cities each, four of which are located in Pinellas. Anna Maria was the only one to mention the Garbage Collection department, although this could be considered equivale nt to larger cities Public Works departments. Madeira Beac h involved their Finance Direct or in physical management tasks, primarily in the role of assigni ng funds to beach-related responsibilities. Not surprisingly, three-quarters of the counties claimed their Tourism Tax Funds as the source of dollars used in both b each maintenance and red tide management chores (Table 4-8). Five of the counties also mentioned their own county government regular, emergency or contingency reserves funds, with Pinellas and Charlotte relying solely on their own budgets for funding (no mention of tourism tax funding). While Franklin County used its own funds to clean its beaches, the respondent claimed that it has no cities on the Gulf and is not greatly bothered by, nor concerned with, red tide or other HABs. Funding Sources and Expenditures Overall, six counties provided estimated and hi storical financial information with respect to overall beach maintenance efforts (Table 4-8). General annual beach management and 71


maintenance costs ranged from nearly $1.5 milli on in Sarasota, down to $76,000 in Gulf. While some counties did not provide red tide-related co sts, several respondents noted allocations of large portions of tourism tax dollars towards annual emergency beach cleaning accounts (which would be used in the event of a red tide), which ranged from $25,000 in Okaloosa up to $400,000 in Sarasota. Four counties had kept precise records of red tide related beach cleaning expenditures, and included Pinellas, Sarasota, Lee and Collier. Sarasota respondents provided current red tide cleaning expenditures of $51,148 for six separate events in FY 2006-07, which included labor, equipment and vendor costs (Table 4-9). Pinellas offers a reimbursement program to its cities that had incurred costs related to red tide cl eaning in 2005, and seven ci ties received $78,090 in total (Table 4-10). Pinellas Office of Mana gement and Budget has limited reimbursement parameters to include actual overtime, temporary labor, and equipment costs related directly to red tides that occurred during a specific time frame. Lee recorded costs of $250,000 for a single 2004 red tide event in Fort Myers, and Collier spent $250,000 in 2005 in red tide-related cleaning expenditures. Seven cities are reimbursed by their host count ies for at least some, but not all, of the labor or dollar expenditures on red tide cleaning efforts, six of th ese in Pinellas, and one in Lee (Table 4-11). Four cities (Holme s Beach, Sarasota, Venice, and Marc o Island) indicated that they would notify their host counties in th e event of a fish kill and, therefore, were not responsible for fiscal or labor expenditures. Three municipa lities (Bradenton Beach, Fort Myers Beach, and Naples) indicated their own budgets were the only source of funds used to clean beaches after a red tide event. Two cities, Clearwater and St. Pe tersburg, used beach parking fee collections to maintain their beaches, and this would include cleaning up during a red tide event. 72


A total of 11 of the 18 cities or 61%, provided red tide-re lated financial and/or labor costs (Table 4-11). These numbers ranged from $1,420 received by Belleair Beach from Pinellas County for its 2005 red tide clean-up efforts, up to Long Boat Keys annual red tide line item budget allocation of $100,000 for cleaning its 10.5 miles of public beaches. The seven cities receiving reimbursement funds from their c ounties collected between $1,420 up to $45,310 as a result of 2005 red tide events. Naples was the sole self-funded city to have a historical red tide annual cleaning allocation of $50,000 in its bud get, although Long Boat Key established its $100,000 red tide budget in 2006. The majority of labor and equipment used to clean red tiderelated fish kills is provided by regular city staff and machinery, and most counties waived the dumping fees associated with dead fish dispos al. However, several respondents mentioned the need for overtime, contract labor, and prisoner tr ustees required to expedite the cleaning process, depending on volume and location of dead marine creatures. Overall, five counties shifted existing pe rsonnel and equipment for red tide cleaning efforts, including Pinellas, Manatee, Sarasota Charlotte and Collier (Table 4-12). These same five, plus Lee, hired additional te mporary labor or private contract ors and utilized prison trustees to achieve the timely removal of dead fish. Communication or Activity Protocols Five of the counties followed some program of public relations in the case of a red tide event. Sarasota and Manatee have equipped their lifeguards with Blackberries, which are used to send twice-daily reports of red tid e conditions for their beaches that are staffed for 8-10 hours per day, year-round. These two are joined by Charlotte and Collier in placement of red tide warning signs on their public beaches. Gulf, Sarasota, Char lotte and Collier also issue press releases and emails to media, hotels, Tourism Developmen t Council, Chamber of Commerce, health care agencies, and county websites. Ma natee sends their Chief Lifeguard out into the community to 73


educate beach users, schools and other organizations. Sarasota was the only county with a written, red tide-specific protocol designed to provide stringent guidelines as to policies and procedures for beach cleaning and public safety notifications. Two counties do not manage their beaches (Okaloosa) or do not ha ve municipalities exposed to the Gulf of Mexico (Franklin). Lastly, Pinellas and Lee Counties did not engage in a ny type of public notification efforts. A total of 13 of the 14 cities that were directly responsible for red tide beach cleaning followed similar action plans desc ribed as follows in the event of a red tide (Table 4-13). Typically when a complaint (odor or dead fish) wa s received by the city it was then investigated by natural or marine reso urce personnel. Following their recommendations and any environmental or health guidelines establishe d by state or federal agencies (e.g., Florida Department of Health, Environmental Protection Agency), the Public Works and/or Parks and Recreation departments combined existing personnel and/or temporary labor and equipment to begin the cleaning process. St. Petersburg, with its small beach length of approximately 650 feet, had their usual private contract or remove any dead fish resulting from red tide blooms, and provided no further elaborations The remaining four cities notified their host counties as previously mentioned, although Holmes Beach wa s willing to assist the county on a where needed and as manpower is available basis. Holm es Beach is unique in that it possesses several blind canals where fish kills build up, and it has hired commercial fishers to collect these with nets and haul them back out into the Gulf. Only one city, Indian Rocks, provided the public with red tide information, including red tide fact sheets that were provided by Pinellas County. Summary and Conclusions The overall objective of this survey effort was to provide an analysis of the direct costs associated with harmful algal bloom events incu rred by city and county agencies located on or near coastal waters with histor ical exposure to red tide blooms. Information was gathered via in74


depth telephone interviews of individuals charged with red ti de-related responsibilities ranging from physical public beach cleaning, to funding allo cation decisions, to pub lic relations efforts. Respondents were successfully interviewed for a to tal of nine Florida Gulf Coast counties and 18 cities within these counties, each of whic h had experienced red tide events in 2005. The majority of funds for red tide-related cleanups were ge nerated by tourism tax dollars, with only two counties relying stri ctly on their regular county dolla rs, perhaps due to the lack of public beaches in these areas (e.g., none were reported in Franklin and only one in Charlotte). In all, four counties and two cities were able to provide actual dollar am ounts specific to red tide events that occurred on their public beaches. Thes e six locations provided red tide-specific costs totaling $653,890 over the 2004-07 time period, with total expenditures per event (including labor, equipment, supplies and vendor fees) ranging from $11,114 to $250,000. Only two cities, Longboat Key and Naples, have placed red tide cleaning costs as a line-item in the annual budget, in the amounts of $100,000 and $50,000, respectively. Although Sarasota County provided the only official written protocol outlining specific policies and procedures in the cas e of a red tide event, each of the other counties and cities appeared to follow a similar pattern of activit y. Initially, a complaint of odor from a red tiderelated fish kill was received by th e agency, either from a member of the public or from beach or park personnel. An agency member, or private consultant, with some level of resource management experience, was sent to the area to investigate the claim and establish a cleaning protocol that would meet any human welfare, environmental and access re strictions (e.g., human health hazard, turtle nesting site, protected dunes, etc.). At this point, cleaning personnel were assigned from existing staff, outsi de labor agencies, or prison trustees, while machinery was also either diverted from usual uses or rented from local suppliers. Once the debris was collected in 75


either trucks or garbage bags, it was hauled to local waste disposal si tes following prescribed regulatory procedures (e.g., dead fish might be bagged, buried, or incinerated in designated locations). In addition to data concerning red tide fiscal costs, respondents provided insight into the difficulties associated with cleaning public beaches in the event of a fish kill. For example, many of the Gulf County beaches harbor protected nesting areas for tu rtles and seabirds, as well as native flora that have low tolerance levels for invasive mechanized equipment. Several beaches have strict environment protoc ols in place to limit or prevent removal of washed up marine materials for a set period of time in an effort to preserve the natural state of coastal ecosystems. Such policies include criteria such as no-rak e areas, cleaning only when there are significant numbers of dead fish, or they require one la rge fish per foot of s horeline or substantial portion of the beach be covered by fish for 24-48 hours, or to a depth of six inches before cleaning can occur. Adherence to environmental policies must be enforced by public officials on private businesses, and in some cases excepti ons have been granted for resorts that are grandfathered and allowed to follow their own cleani ng policies. On at least one occasion, the state health department steppe d in and required a county to cl ean private homeowners beaches as the fish kill was deemed a human health hazard. Five of the counties, and only one city, me ntioned public notification of an ongoing red tide event, typically by placing warning signs on the beach and sending alerts to tourism-related businesses. However, a few counties and cities mentioned financial suppo rt of the grassroots organization START, or Solutions To Avoid Red Tide, which has active membership in most of the responding regions and works to educate the public and busi nesses about red tide. Manatee 76


and Sarasota counties have equipped their lifeguards with Blackbe rries, which are used to send twice-daily messages concerning red tide and other beach conditions. Our study has provided some insight into ac tual public management strategies and funding sources linked to a red tide event. An im portant finding is the estimated costs of a red tide event per linear foot of beach, which can be extracted from the data provided by Sarasota and Pinellas counties. Sarasota spent an averag e of $4.87 per linear foot of beach to provide the labor and equipment necessary to re move the dead fish resulting fr om a single red tide event that occurred in October 2006 through February 2007. In Pinellas, seven cities were reimbursed an average of $14.27 per linear foot of beach for red tide-related cleaning required throughout 2005; however, incidence and duration of the events were not mentioned, and city expenditures may have exceeded county reimbursements due to in-k ind labor and equipment reallocations. This information may provide a useful baseline for estimation of red tide-related budget needs for other cities and counties that are responsible for public beach management. However, it should be noted that public government pr otocols associated with red tid e events are strongly dependent on all of the following factors: the timing, durati on and severity of an event; size of budget and labor force; overall importance of tourism (evidenced by tourism tax collections); quantity and accessibility of public beaches; and the environmen tal regulations that are specific to each locality. 77


Table 4-1. Population estimates of nine Florida Gulf Coast counties County County populationa Population change, 2002-05 (number) (%) Okaloosa 177,284 + 4.0 Franklin 11,057b N/A Gulf 13,332b N/A Pinellas 905,158 1.8 Manatee 300,828 + 13.9 Sarasota 359,783 + 10.4 Charlotte 154,716 + 9.5 Lee 539,097 + 22.3 Collier 302,514 + 20.3 a U.S. Census Bureau, 2005. b U.S. Census Bureau, 2002 (2005 not available). Table 4-2. Tourist tax collections, dollars and pe rcent of total county taxes, and estimated annual tourist numbers for nine Florida Gulf Coast counties County Estimated annual domestic tourists, 2005a Tourist development tax rate, FY 2005-06b Tourist development tax collections, FY 2005-06b (1,000) (%) ($1,000) Okaloosa 2,934 4.0 7,364 Franklin N/A 2.0 669 Gulf N/A 4.0 304 Pinellas 2,393 5.0 21,651 Manatee 772 4.0 4,760 Sarasota 1,621 4.0 7,432 Charlotte N/A 5.0 1,625 Lee 2,316 5.0 17,030 Collier 1,544 4.0 13,056 a Estimated tourist numbers from VIIST FLA -Domestic Visito rs to Florida, Florida Visitors Study, 2005 (N/A refers to unavailable data). b Validated tax receipts data for July 2005, through June 2006, Florida Department of Revenue, Office of Tax Research. 78


Table 4-3. Approximate tourism tax dollars coll ected per public beach miles for nine Florida Gulf Coast counties County Public beachfront lengtha Tourist tax dollars per mile of public beach b (miles) ($/beach mile) Okaloosa 24 76,708 Franklin 0 c N/A Gulf 17 4,471 Pinellas 35 164,023 Manatee 14 85,000 Sarasota 7 265,429 Charlotte 12 33,854 Lee 18 236,528 Collier 22 148,364 a Public beachfront access miles retrieved from various online county government sources Sarasota: http://apoxsee.co.sarasota.fl.us/ ; Charlotte: http://www.charlotte-florida.com/ ; Okaloosa: http://www.co.okaloosa.fl.us/ ; Lee: http://www.lee-county .com/ ; Pinellas: http://www.pinellascounty.org/ ; Gulf: http://www.visitgulf.com/ ; Collier: http://www.collierg ov.net ; Manatee http://www.flagulfislands.com/ b Calculated as 25% of annual FY2005-06 Tourist Development Tax collections (see Table 4-2). c Franklin County reported no Gulf-front public beaches. 79


Table 4-4. Population estimates of Florida Gulf Coast cities City 2002 populationa County (number) Belleair Beach 1,751 Pinellas Clearwater 108,787 Pinellas Indian Rocks Beach 5,072 Pinellas Indian Shores 1,705 Pinellas Madeira Beach 4,511 Pinellas North Redington Beach 1,474 Pinellas Saint Petersburg 248,232 Pinellas Treasure Island 7,450 Pinellas Anna Maria 1,814 Manatee Bradenton Beach 1,482 Manatee Holmes Beach 4,966 Manatee Longboat Key 7,603 Manatee/Sarasota Sarasota 52,715 Sarasota Venice 17,764 Sarasota Fort Myers Beach 6,561 Lee Sanibel 6,064 Lee Marco Island 14,879 Collier Naples 20,976 Collier a U.S. Census Bureau, 2002. 80


Table 4-5. Disposition of telephone interviews and response rates Description Number Percent (number) (%) Total Interviews Attempted 37 100.0 Unsuccessful Attempts 6 16.2 Successful Interviews 31 83.8 Total Interviews 31 100.0 Ineligible Respondents a 4 12.9 Total Completed Interviews 27 87.1 a Ineligible respondents included the cities of Apalachico la, Carrabelle, Northport, and Punta Gorda, due to their proximity from Gulf waters and/or l ack of public Gu lf-front beaches. Table 4-6. Summary of public departments char ged with physical beach/red tide management responsibilities by county County Departments Okaloosa Okaloosa Tourism Development Council Beach Manager Franklin Waste Disposal Department Gulf Public Work Department; Florida St ate Health Department; Department of Environmental Health and Protection Pinellas County Solid Wast e; Technical Management; Office of Management and Budget; Department of Environmental Management Manatee Parks and Recreation Department; Ut ility Operations; Division of Marine Rescue, Department of Public Safety Sarasota Sarasota County Health Department (branch of Florida State Department of Health); Mote Marine Laboratory (w ater testing); Parks and Recreation Department; Emergency Services; Public Works Charlotte Parks and Recreation Department Lee Parks and Recreation Department; Na tural Resources; Division of Public Works; Private Contractors Collier Parks and Recreation Department; Pollution Control 81


Table 4-7. Summary of public departments char ged with physical beach/red tide management responsibilities by city City Departments County Clearwater Landscape Manager (Parks and Recreation); Clearwater Marine Aquarium Inspector Pinellas Treasure Island Public Works, Assistant Director; Beach Stewardship Committee; Consulting Firm Pinellas St. Petersburg Parks and Recreation, Manager of Athletic Operations; Public Contractors Pinellas Indian Rocks Public Service Pinellas Belleair Beach Communication Services IT Director Public Works; Private Contractors Pinellas Indian Shores Public Services; Workforce Contract Labor Pinellas Madeira Beach Parks and Recreation; Fi nance Director; Utilities Pinellas North Redington Beach Town Clerk; Public Work s; Private Contractor Pinellas Bradenton Beach Public Works Manatee Holmes Beach Public Works; Mayor; Commercial fishers Manatee Anna Maria Public Works; Garbage Collection Manatee Long Boat Key Public Works; Private Contractor Manatee/Sarasota Sarasota Sarasota County Sarasota Venice Sarasota County Sarasota Fort Myers Beach City Council; City Personnel Lee Sanibel City Manager; Private Contractor; Public Works; Natural Resources Manager; Lee Naples Natural Resource; Publ ic Works; Contract Labor; Hauling Contractors; Equipment Rental Suppliers; Boat Suppliers Collier Marco Island Collier County Collier 82


Table 4-8. Summary of sources of funds for financial beach/red tide management responsibilities and estimated f unds and expenditures, by county County Source of funds Estimated funds and/or expenditures Okaloosa Okaloosa Tourism Development Council Emergency Funds $25,000 annually for emergency beach clean-up Franklin Franklin County Government Budget Not applicable Gulf Gulf County Government Budget Gulf Tourism Development Council Funds Tourism Tax Funds 25% of annual tourism tax funds (~$76,000 based on Table 4-2)] Pinellas Pinellas County Government Contingency Reserves $78,000 total paid to reimburse eight cities for 2005 red tide beach cleaning costs (see Table 4-10) Manatee Manatee County Tourism Development Council Tourism Tax Funds $1.25 $1.275 million annually for beach maintenance Sarasota Sarasota County Tourism Development Council Tourism Tax Funds $1.4 $1.5 million annually for beach maintenance $400,000 annually for emergency beach cleaning $51,148 paid out for six red tides in FY 2006-07 (see Table 4-9) Charlotte Charlotte County Government Beaches cleaned by hand (MayOctober ~ $10,000; Nov-April ~ $40,500) Lee Lee County Tourism Development Council Tourism Tax Funds; Lee County Government $300,000 $340,000 annually for emergency beach cleaning $250,000 paid out for single 2004 red tide in Fort Myers Collier Collier County Tourism Development Council Tourism Tax Funds $550,000 annually for beach maintenance $250,000 paid out for 2005 red tide beach cleaning 83


Table 4-9. Sarasota County expenditures fo r six red tide events by public beach Cost per event Public beach/ Event numbera Red tide days Labor Equipment Vendor Total Cost per beach area (days) ($) ($) ($) ($/event) ($/ft) Siesta Beach#1 37 10,201.50 5,166.60 1,165.00 16,533.10 6.89 Siesta Beach#2 2 327.12 327.12 0.00 654.24 0.27 Siesta Beach#3 25 2,813.00 1,237.73 1,865.00 5,915.73 2.46 Siesta Beach#4 20 10,147.00 5,776.43 720.00 16,643.43 6.93 North Jetty#1 7 5,522.00 3,712.86 1,890.00 11,154.86 12.39 North Jetty#2 1 137.00 109.23 0.00 246.23 0.27 AVERAGE 15 4,862.94 2,721.66 1,410.00 8,524.60 4.87 a Siesta Beach #1: October 2 Novemb er, 8, 2006; Siesta Beach #2: Nove mber 9-10, 2006; Siesta Beach #3: December 4-29, 2006; Siesta Beach #4: January 8-28 2007; North Jetty #1: February 1-7, 2007; North Jetty #2: February 22, 2007. Table 4-10. Pinellas County reimbursements fo r 2005 red tide events by city public beach City Labor Equipment/ supplies Total Costs per beach area ($) ($) ($) ($/ft) Belleair Beach 985.14 126.80 1,111.94 10.14 Indian Rocks Beach 9,214.96 5,095.75 14,310.71 5.04 Indian Shores (1) 9,972.49 304.02 10,249.51 22.32 Indian Shores(2) 8,160.00 20,878.00 29,038.00 38.09 Madeira Beach 10,868.00 35,998.00 46,866.00 7.01 North Redington Beach 1,198.86 842.80 2,041.66 14.64 Treasure Island 7,851.03 12,633.80 20,484.83 2.61 AVERAGE 4,862.94 5,064.29 8,524.60 14.27 84


Table 4-11. Summary of public departments char ged with financial beach/red tide management responsibilities and estimated expenditures and labor by city City Source of funds Estimat ed labor and/or expenditures Clearwater Beach parking fees $200,000 standard beach maintenance for 1.4 miles public beach; staff of 12 clean daily from 5am to 1:30 pm routinely Treasure Island Reimbursed by Pinellas County for some red tide cleaning costs After 2005 red tides, city requested $20,485, received $9,660 from county for 3 miles public beach; regular beach cleaning biweekly with tractor and rake; 18 employees wanted to work overtime during 2005 red tide St. Petersburg Beach parking fees Pay private contractor $200/month for standard beach cleaning for 600-700 ft public beach Indian Rocks Reimbursed by Pinellas County for some red tide cleaning costs During 2005 red tides, 10 full-time personnel cleaned beaches, and county reimbursed $11,000 to city; standard beach raking during first week of month; county waived landfill dumping fees Belleair Beach Reimbursed by Pinellas County for some red tide cleaning costs if Work cannot be done by regular [city] personnelreluctant reimbursement program After 2005 red tides, county reimbursed $1,420 to city; pay dead fish dumping fees of $37/ton Indian Shores Reimbursed by Pinellas County for some red tide cleaning costs After 2005 red tides (June 13 through August 19), received $45,310 from county after submitting two re quests (labor and equipment); dump fees waived 85


City Source of funds Estimat ed labor and/or expenditures Madeira Beach Reimbursed by Pinellas County for some red tide cleaning costs During 2005 red tides (June through midSeptember), city personnel logged 770 hours cleanup, 8-22 man hours per day; front-end loader, pickup truck, dump fees not waived (fish mixed with garbage), paid overtime of $10,800; city requested $46,866, received $8,310 from county North Redington Beach Reimbursed by Pinellas County for some red tide cleaning costs After 2005 red tides, requested and received reimbursement of $2,050 from county Bradenton Beach General city budget for overtime and part-time; no red tide line item None given Holmes Beach Manatee County Budget Follow Manatee County red tide protocols Anna Maria Place annual requests for red tide clean-up funds, not granted; however, clean-up funds appear in the overtime and part-time budget possibly to avoid negative images Average 3 red tides per year, from early July through mid-September; if major event, county assists beach cleaning with rakes, 4-12 laborers, 4-6 prison trustees, who handpick dead fish Long Boat Key Prior to 2005, red tide funds from City General Revenues; from 2006 to present, City Red Tide Budget $100,000 annual red tide cleaning funds for 10.5 miles public beaches Sarasota Sarasota County budget Follows Sarasota County red tide protocols Venice Sarasota County budget Follows Sarasota County red tide protocols Fort Myers Beach City budget During 2006 event, dead fish hand-picked for ~ 2 months, approx. costs of $15,000 (beach raking prohibited) Table 4-11. Continued. 86


City Source of funds Estimat ed labor and/or expenditures Sanibel City presents county Tourism Development Council with estimated red tide cleaning costs which are paid by county Parks and Recreation funds February 2004 city received $11,114 from county; January 2007 city received $9,742 for Nov. red tide for 11 miles public beach; average 4-5 fish kills per year; 8-10 man hours per week beach inspection during red tide event Naples City Red Tide Budget $50,000 annual red tide cleaning funds; no severe outbreaks in last 4-5 years Marco Island Lee County budget Follow Lee County red tide protocols Table 4 11. Continued. 87


Table 4-12. Summary of communica tion and activity protocols followed in the case of an active red tide event by county County Communication and activities Okaloosa None cities manage their own beaches Franklin None (respondent indicated that no m unicipalities are exposed to red tide) Gulf Gulf County Tourism Development Council issues Public Service Announcements to county and touris m-related businesses as needed Pinellas No public notification efforts Reimburses municipalities following 2005 FEMA standards regarding extreme environmental events by providing contingency reserves funds to those municipalities that incur be ach clean-up expenditures and submit requested reimbursement amount Shifts existing beach cleaning personnel and equipment, approve overtime, hire temporary workers Manatee Participates in twice-daily lifegua rd red tide monitoring program via Blackberries, who also place beach si gnage on all public beaches for 8-10 hours per day, year-round Participate in PR efforts to ove rcome surveys done by universities Chief Lifeguard presents red tide information to beach-goers, schools and other organizations Shifts existing beach cleaning personnel and equipment, and add prisoner workers, extra staff if necessary Sarasota Participates in lifeguard red tide monitoring program via Blackberries for six county beaches Follows Florida Department of Health protocol with respect to water hazard threats to human health Follows an extensive beach cleaning policy developed by county in 1995 and updated in 2006 County cleans up municipality beaches 88


County Communication and activities Charlotte Charlotte County Parks and Recreation Department issues press releases to media and hotels and posts red tide warning signs at the beaches Shifts existing county beach cleaning personnel and tractor and golf cart, no overtime Lee Reimburses Fort Myers and Sanibel for HAB clean-up costs Cleans unincorporated and regional beaches within cities and Bonita Springs Uses temporary labor and hire private contractors Collier Posts red tide warning signs on beaches Monitors fish kill reports, NOAA and HAB bulletins Sends emails to Tourism Development Council, Chamber of Commerce, health care agencies, media, and county website Shifts existing beach maintenan ce crews (2 beach cleaners and 1 supervisor) and rake and drag equipment In the case of limited access beaches and canals, uses community service individuals to hand pick d ead fish if heavy deposits Table 4-12. Continued. 89


Table 4-13. Summary of communica tion and activity protocols followed in the case of an active red tide event by city City Communication and activities Clearwater Clearwater Marine Aquarium inspect or checks turtle nesting spots Shifts existing beach cleaning personnel to clean up fish kills at public beaches In 2005, FL Department of Health declared dead fish a heal th risk, so city hired extra help to clean private beaches as well Treasure Island Beach Stewardship Committee determines no-rake areas Shift existing labor, approve overtime File reimbursement request with Pinellas County St. Petersburg Private contractor paid to do standa rd beach cleaning removes dead fish Indian Rocks Public Service Department monitors beaches and provides information and fliers, as well as red tide f act sheets provided by Pinellas County Clean up dead fish when significant numbers Department of Environmental Protec tion permit required to dispose of dead fish File reimbursement request with Pinellas County Belleair Beach Parks management work done by private contractors Dead fish are bagged, put into dumps ters, taken to county incinerator File reimbursement request with Pinellas County Indian Shores Shifts existing Public Works personnel, pays overtime Hire contract labor Workforce File reimbursement request with Pinellas County Madeira Beach Shift existing Parks and Recreation and Utility Department labor and pays overtime Hire temporary workers File reimbursement request with Pinellas County 90


City Communication and activities North Redington Beach Rent skid loader, hire day laborers to hand pick debris, place in piles and city dumpster File reimbursement request with Pinellas County Bradenton Beach City receives odor complaints from fi sh kills washed into bay (Gulf beach owned by Manatee County) Public Works director determines clean-up need Shift existing personnel County provides free county stoc kade trustees, if needed. City garbage truck hauls fish away Holmes Beach Manatee County cleans up without cost to city City assists county on where needed and as manpower is available basis by providing shovels, forks to load fish into front-end loaders City has several blind canals wher e fish kills build up, cleared by 2-3 commercial fishers using nets to haul fish back out into Gulf Anna Maria City notified of foul odor Public Works Director investigates, employees alerted to possible fish kill (smell without dead fish ofte n occurs with onshore winds) Shift existing Public Works and Garbage Collection employees to collect and haul away dead fish Long Boat Key City receives complaint Mote Marine Laboratory scientists walk beach and tell Public Works when safe to clean up dead fish (avoid nesting turtles) Private contractor uses w eed harvesters to clean Sarasota Follows Sarasota County protocols Venice Follows Sarasota County protocols Table 4 13. Continued. 91


City Communication and activities Fort Myers Beach City has development code which forb ids beach raking (certain areas are grandfathered and allowed to rake because they are resorts, etc.) Must get a permit from FL Departme nt of Environmental Protection to rake or clean the beach Shift existing personnel to hand-pick dead fish, pay overtime Use jail inmates, at no cost, to hand-pick dead fish Sanibel Public Works Director actively manages clean-up, determines where and when based on environmental regulations i.e. one large fish per foot of shoreline, avoid wildlife nesting sites Natural Resources biologists inspect and contribute to clean-up decisions Estimates cleaning costs and presents to Lee County Tourism Development Council with requested reimbursement Shift existing Public Works staff to hand-pick dead fish Private contractors already under beach cleaning contracts also participate in beach cleaning Naples Daily city beach cleaners call in a fish kill Public Works Director and Natural Resources Manager inspect beaches Clean-up criteria: Substan tial portion of the beach be covered by fish for 24-48 hours or to a depth of six inches Most of clean-up done with contracted labor With existing red tide line item in city budget, purchase orders with hauling contractors, equipment and boat suppliers are already set up to facilitate rapid response Marco Island Collier County is notified and count y provides equipment, labor and funding Table 4 13. Continued. 92


CHAPTER 5 SUMMARY AND CONCLUSIONS Red tide blooms have occurred on Floridas co astlines for decades prior to the states rapidly growing population. Historically, as co astal areas faced incr easing pressure from development, targeted environmental management efforts (e.g., Florida Beach Erosion Control Program (1964), Statewide Coastal Monitoring Pr ogram (2000), Marine Turtle Protection Act (1979), Florida Healthy Beaches (1998), The BE ACH Act (2000)) were developed to protect native biota and simultaneously ensure continued public access to safe, clean, and healthy marine environments. The origin, movement, and expected duration and intensity of naturally-occurring Karenia brevis blooms (i.e., red tides) have proven diffi cult to predict, further increasing the difficulty of quantifying economic impacts. While anecdotal evidence of the negative effects resulting from red tide events on water and shoreline conditions have a ppeared in local popular press for more than 50 years, the prolonged series of events in 2005 served to hi ghlight the need for more precise information. This research provides estimates of the localized economic impacts of re d tide events, which can help to justify continued or proposed prevention, mitigation, and control efforts. In particular, this study focused on impacts to specific to restau rants, local residents, and coastal governments, all of which are located within nine counties along Floridas Gulf Coast. The analysis of daily restau rant sales over a seven year period revealed many unique insights and contributions regard ing how to measure and model the short-run fiscal impacts on individual firms. First, it is the only case study (to our knowledge) that quantifies a reduction in sales due to red tide events by using proprietary data that has been consistently collected over time. Average inflation-adjusted daily sales were found to decline by 13.7% and 15.3% for the two highest-grossing restaurants, ceteris paribus ; however, these declines would be relatively 93


larger if noticeable red tide conditions were pres ent on low volume sales days that tend to occur on Mondays and Tuesdays, or in the month of September. Second, the empirical findings were based on subjective notations of environmental conditions (i.e., red tides, rainfall and tropical storms or hurricanes ) at the location of sales, not by secondary data collected at distant monito ring stations. A compar ison of the red tide observations noted by the restaurant managers corresponded with FWRI recorded cell counts measured within seven days (recall that cell count monitoring is infrequent). More importantly, the subset of observations that matched (i.e., subjective observations were made on the same days that cell counts were recorded), cell count s averaged 585,183 cells per liter. Thus, FWRI cell count measurements may need to be much hi gher than average levels (or levels used to regulate shellfish harvests) to impact beachfront restaurant patrons. If these thresholds are supported in other studies, they could be used to estimate red tide effect thresholds for any beachfront business and, thereby, allow for the us e of nearby cell count data in subsequent empirical analysis. Third, these models were specific to the unique waterfront location and physical characteristics of each restaurant and, thus, th e results cannot be used to predict or project aggregate sales losses to the restaurant sector. However, this approach highlights how proprietary data can be used with nearby cell count data to estimate sales reducti ons due to red tide events in order to pursue opportunities for economic disaster relief. For example, the U.S. Small Business Administration (SBA) has provided loans ranging from $4,832 to $81,912 per event to individual Florida restaurants from 1996-2002. By comparison, this study found daily sales reductions of $3,734 and $868 when a red tide was present, rang ing from $868 to $22,404 per red tide event. 94


However, given the relatively small total losses of $252,242 for two of the restaurants over a seven-year time period, development of privat e business disruption insurance such as that available for hurricane or flood disasters with re latively low-cost annual premiums may be more appropriate solutions as compared to public a ssistance programs. The relevance of insurance instruments provided by the market is further supported when sales fo r the two high-grossing restaurants in 2005 are closely examined, as this year was noted for its intense, long-lasting red tide bloom activity. Using estimated percentages of daily sales losses generated by the model, inflation-adjusted average 2005 sales, average daily 2005 sales, and the number of 2005 red tide days noted by the manager, restaurants B a nd C experienced sales losses of 2.4% and 2.2%, respectively, as a direct result of the presence of a red tide in 2005. Lastly, the firm-level investigation of sales reductions during red tide events revealed that aggregate losses to each firm (in absolute values) in this study were relatively small compared to reported gross sales. Thus, it seem s appropriate that the SBA con tinue to provide loans, which could perhaps be used to fund im mediate clean-up activit ies or replacement staff (i.e., to solve a cash flow problem) as opposed to offering grants. Mo reover, the characteristi cs of red tides (i.e., as an exogenous environmental ev ent with localized impacts) c ould support the development of private business disruption insurance such as that available for hurricanes or floods. While such an industry would need fairly ac curate prediction mode ls, the state of the science is progressing and could support relatively low-co st annual premiums given that red tide conditions can vary across short geographic distances. Such a mitigation approach may be the most appropriate in addressing the impacts of red tides as oppos ed to other forms of public assistance. The second analysis, which provided empirical models of the probabi lity of a residents decision to alter their participation in marine-rela ted activities during a red tide, offers additional 95


insights and contributions regarding economic impact s. Overall, the survey revealed that 70% of residents in Sarasota and Manatee counties had th eir participation in at least one marine-related activity disrupted due to red tides in 2000. A ch ange in participation (i.e., disruption) was defined as whether the individual cut short, delayed, or relocated during a red tide. The model to explain this overall effect found that none of the demogr aphic characteristics (such as age, education, income, or residency st atus) mattered; only the factors that could be affected by Extension activities had a measurable impact on the probability that an individuals participation in at least one marine-related activity would change. These Extension-related variables included red tide knowledge levels and the number of information sources that they had used to obtain information on red tide events; high er levels of each were associated with higher probabilities of disrupted participation, which implies economic losses associated with reduced recreational expenditures in the immediate region. When a change in the probability of a disrup tion was examined for ea ch of four distinct marine-related activities, the models revealed differences that would support unique education and outreach campaigns to mitigate economic lo sses during red tides. For example, beach-going residents that have lived in Florida longer, vi sit the beach more often, or have more red tide knowledge or more accurate recollections of red tid e effects, they have a higher probability of being affected. In particular, be ach users that know more about the environmental characteristics and effects of a red tide had a higher probability of cutting short a beach visit. Restaurant patrons that resided year-round in Florida, dined out more often, accessed more sources of red tide information, and recalled more red tide effects also had a higher probability of reacting during a red tide event; in general by dining out less. Speci fically, the restaurant patrons who reacted (i.e., had a higher probability of having their patronage affected) were less likely to relocate away 96


from beachfront dining locations if they were older and lived full time in Florida, and more likely to select another restaurant if they were male and used additional sources of information about red tide. Although this survey effort was conducted prio r to the widespread re d tide events in 2005, it serves to preview the potentia l for amplification of negative economic impacts in cases where accurate public education efforts are not forthc oming. Results should spur private and public sector beach-related concerns to develop appr opriate educational materials and make them available in both time and location-sensitive manner in an effort to mini mize both short and long run probable losses. Given the general finding that each additional piece of red tide information and source resulted in increased likelihood that a resident would alter his or her participation in a marine-related activity (and, thus, incur lost reve nues), extension materials could focus on a few direct educational messages provide d at a single venue to targeted recreationists, rather than an abundance of literature and messages scattered ac ross many sources and aimed at all residents. Lastly, extrapolation of study results to all residents within the study coun ties generates absolute numbers of affected residents based on participation in each activity. For example, survey results revealed that 76% of reside nts engaged in beach-goi ng, which suggests that 503,989 people visit Manatee and Sarasota beaches at least once per year Of these, 70%, or 352,793 people have indicated that red tide affected their beach participation. Fifty-eight percent, or 204,620 residents, delayed their day at the beach, while 84,670 cut short a beach day and 63,503 relocated away from red-tide affected beache s. These changes in participation levels revealed in this study can be combined with estimated values of a day at the beach, or on a boat fishing trip, or beach and pier fishing, or rest aurants average daily sales to estimate potential losses to local businesses and communities reliant on marine-related activities. Future resident 97


surveys could focus on their willingness-to-pay for red tide prevention, mitigation, and control measures to avoid disruption in recreational activities that ar e affected by red tide events, particularly in the case of beach users. The final study provides perspective on the pub lic management protoc ols, and labor and equipment expenditures resulting from red tide ev ents on public beaches that were incurred from 2004 through 2007 in Florida. Encompassing Franklin, Gulf, Okaloosa, Pinellas, Manatee, Sarasota, Charlotte, Lee and Collier counties and eighteen coastal municipa lities located within these nine counties, it presents a seminal colle ction of recorded red tide expenditures.. Six localities recorded red tide expenditures tota ling $653,890 during the three-year time period. Two cities that privately manage their beaches had annual line-item budget allocations of $50,000 and $100,000 for red tide-related clean up and research. It is important to emphasize that the hidde n costs of in-kind expenditures incurred by public agencies due to lost or redirected labor or equipment hour s used to assist in red tide cleaning efforts are not captured in these numbers Additional labor is also provided by prison trustees that represent an implicit cost to taxpa yers that was not direc tly included in red tide expenditure calculations recorded by managers. These costs are not explicitly recorded and, thus, are excluded from the calculations in this study, ye t they represent explic it costs to taxpayers. Using the precise dollars expended by Pinell as and Sarasota coun ties on red tide cleaning efforts and approximate linear feet of public county beaches, it is possible to generate conservative cost estimates for all beaches. Sa rasota County spent an average $0.47 per linear beach foot per day of a red tide event in 200607, and with its seven miles of beaches, this agency could spend an average $17,371 per red tide day were its entire beachfront exposed to a red tide bloom. Pinellas County spent an annua l average $14.27 per linear beach foot in 2005, 98


and its 35 miles of public beaches could cost taxpayers $2,637,096 in annual red tide cleaning costs. These estimates could be used to determine a maximum amount that beach users may be willing to pay via tax dollars to ensure timely red tide beach cleaning, although the numbers are conservative due to in-kind matches. These survey re sults suggest that it is vital to recognize the benefits of quantifying both dir ect and indirect red tide-related costs, and required personnel and financial efforts necessary for timely red tide event management, for planning and budgeting purposes. In sum, the results indicate that the restaurant s, residents, and public agencies are aware of, and have experienced negative econo mic impacts due to, red tide ev ents, in spite of limited (or no) public education efforts and erratic and localized red tide ev ents. This study showed that beachfront restaurants have incurred statistically significant reductions in daily sales when red tide events leave behind dead fish, or aerosolized toxins prevent diners from enjoying water-side dining. As revealed in this study, economic losses are likely given reside nts decisions to cut short, delay or relocate their ma rine-related activities due to red tide events. While such impacts may be short run, such behaviors may translate into future decreases in demand for both tourism and property values should tourists choose an alte rnate red tide-free destin ation, current residents abandon beach properties, or potenti al residents select away from locations susceptible to red tide events. The results revealed by these efforts suggest ma ny fruitful directions for future research efforts. Analysis of other firms that are dependent on beach-related activities following a timeseries regression approach may reveal statistic ally and economically si gnificant ne gative losses during a red tide event. Improvement of curre nt red tide cell count measurements through implementation of a statistically valid sampli ng process would provide much-needed data for 99

PAGE 100

firm or local analyses. Finally, there is a st rong need for empirical analysis of various educational messages related to red tide, which could serve to guide the timing, placement, and exposure levels necessary to ensure public aw areness of the nature of red tide events. 100

PAGE 101

LIST OF REFERENCES Amemiya, T. Qualitative Response Models: A Survey. Journal of Economic Literature XIX(December, 1981):1483-1536. Anderson, D. M. ECOHAB: The Ecology and O ceanography of Harmful Algal Blooms A National Research Agenda. Woods Hole Oceanographic Institution, Woods Hole, MA, 1995. Internet site: h ttp://www.whoi.edu/redtide/nationp lan/ECOHAB/ECOHABhtml.html (Accessed January 2006). Anderson, D.M., P. Hoagland, Y. Kauro, and A.W. White. Estimated Annual Economic Impacts from Harmful Algal Blooms (HABs) in the United States. Technical Report #WHOI-2000-11, Woods Hole Oceanographic In stitution, Woods Hole, MA, September 2000. Backer, L.C., Fleming, L.E., Rowan, A., Cheng, Y.S ., Benson, J., Pierce, R.H., Saias, J., Bean, J., Bossart, G.D., Johnson, D., Quimbo, R., Baden, D.G. Recreational Exposure to Aerosolized Brevetoxins During Florida Red Tide Events. Harmful Algae 2(2003):19-28. Baden, D.G., A.J. Bourdelais, H. Jacocks, S. Mi chelliza, and J. Naar. Natural and Derivative Brevetoxins: Historical Backgr ound, Multiplicity, and Effects. Environmental Perspectives 113,5(May 2005): 621-625. Ballance, A., P.G. Ryan, and J.K. Turpie How Much is a Clean Beach Worth? The Impact of Litter on Beach Users in the Cape Peninsula, South Africa. South African Journal of Science 96(May 2000):210-214. Boesch, D. F. D. M. Anderson, R.A. Horner, S. E. Shumway, P.A. Tester, and T.E. Whitledge. Harmful Algal Blooms in Coastal Waters: Options for Prevention, Control and Management. Decision Analysis Series No. 10, National Oceanic and Atmospheric Administration Coastal Ocean Program, 1997. Brand, L.E., and A. Compton. Long-term Increase in Karenia brevis abundance along the Southwest Florida. Harmful Algae 6(2007):232-252. Bureau of Economic and Business Research [BEBR] Florida Statistical Abstract. University of Florida, Gainesville, FL, 2000. Caldwell, P.C. Validity of North Shore, Oahu, Hawaiian Islands Surf Observations. Journal of Coastal Research 21,6(November 2005):1127-1138. Casper, E.T., S.S. Patterson, P. Bhanushali, A. Farmer, M. Smith, D.P. Fries, and J.H. Paul. A Handheld NASBA Analyzer for the Fiel d Detection and Quantification of Karenia brevis Harmful Algae 6(2007):112-118. Changnon, S.A. Shifting Economic Impacts from Weather Extremes in the United States: A Results of Societal Changes, Not Global Warming. Natural Hazards 29(2003):273-290. 101

PAGE 102

Divers, T, J. Farr, C. Hall, J, Huntington, K. Mikata, and H. Recksiek. FACT: Florida Assessment of Coastal Trends. D. Bagley, M. Harrison, and C. McCay, eds. Florida Coastal Management Program, Florida Departme nt of Community Affairs, Tallahassee, FL (December 2000). Evans, G.K. Three Essays in Regional Economic Impact Analysis. Ph.D. dissertation, Texas A&M University, College Station, August 2002. Fisher, W.S., T.C. Malone, and J.D. Giattina. A pilot project to detect and forecast harmful algal blooms in the northern Gulf of Mexico. Envi ronmental Monitoring and Assessment, Vol. 81. Kluwer Academic Publishers, The Netherlands, 2003. Flewelling, L.J., Naar, J.P., Abbott, J.P., Bade n, D.G., Barros, N.B., Bossart, G.D., Bottein, M.D., Hammond, D.G., Haubold, E.M., Heil, C.A., Henry, M.S., Jacocks, H.M., Leighfield, T.A., Pierce, R.H., Pitchford, T. D., Rommel, S.A., Scorr, P.S., Steidinger, K.A., Truby, W.W., VanColah, F.M., Landsberg, J.H. Red Tides and Marine Mammal Mortalities. Nature 435(2005): 755-756. Florida Department of Agriculture and Consumer Services Division of Aquaculture. Internet site: http://www.floridaaqu aculture.com/seas/seas_mngm t.htm (Accessed July 2007). Florida Department of Revenue Offi ce of Tax Research. Internet site: http://dor.myflorida.com/dor/ (Accessed July 14, 2007). Florida Fish and Wildlife Cons ervation Commission Fish an d Wildlife Research Institute (FWRI). Internet site: http://resear ch.myfwc.com (Accessed January 2006). Freeman, A.M. The Benefits of Water Quality Improvements for Marine Recreation: A Review of the Empirical Evidence. Marine Resource Economics 10,4(Winter 1995):385-406. Glick, J. The Long Hot Smelly Summer. The Sarasota Herald Tribune, The New York Times Co., New York, NY, September 5, 2005. Greene, G. C.B. Moss, and T.H. Spreen. The Demand for Recreational Fishing in Tampa Bay, Florida: A Random Utility Approach. Marine Resource Economics 12 (1997): 293-305. Greene, W. H. Econometric Anal ysis, 3rd ed. Upper Saddle River, NJ: Prentice-Hall, Inc., 1997. Hoagland, P., and S. Scatasta. The Economic E ffect of Harmful Algal Blooms. In Ecology Study Series, E. Graneli and J. Turner, eds., Ecology of Harmful Algae 189,30(2006): 391402. Hoagland, P., D.M. Anderson, Y. Kaoru, and A.W. White. The Economic Effect of Harmful Algal Blooms in the United States: Estimates, Assessment Issues, and Information Needs. Estuaries 25,4b(August, 2002):819-837. Huettel, S. Truth in Tourism? Mums the Word on Red Tide. St. Petersburg Times, The Times Publishing Company, St. Petersbur g, Florida, September 1, 2005. 102

PAGE 103

Hunt, L.M. Recreational Fishing Site Choice M odels: Insights and Fu ture Opportunities. Human Dimensions of Wildl ife: An International Journal 10,3(October 2005):153-172. Imperial, M.T. Institutional Analysis and Ecosystem-Based Management: The Institutional Analysis and Development Framework. Environmental Management 24,4(1999):449-465. Jensen, A.C. The Economic Halo of a HAB. In Proceedings of the 1st International Conference on Toxic Dinoflage llate Blooms, V.R. LoCicer o, ed., The Massachusetts Science and Technology Foundation, Boston, MA, 1975. Kahn, J. and M. Rockel. Measuring th e Economic Effects of Brown Tides. Journal of Shellfish Research 7(1988):677-682. Kaoru, Y. Measuring Marine Recreation Bene fits of Water Quality Improvements by the Nested Random Utility Model. Resource and Energy Economics 17,2(August, 1995): 119-136. Karp, H. Red Alert. Wall Stre et Journal, Dow Jones & Comp any, Inc., New York, New York, August 12, 2005. Kildow, J. Phase I Facts and Figures Florida s Ocean and Coastal Economies. National Oceans Economics Program, June 2006. Internet site: http://noep.mbari.org/download/ (Accessed September 23, 2006). Kirkpatrick, B., L.E. Fleming, D. Squicciarini, L.C. Backer, R. Clark, W. Abraham, J. Benson, Y.S. Cheng, D. Johnson, R. Pierce, J. Zaias, G.D. Bossart, and D.G. Baden. Literature Review of Florida Red Tide: Impli cations for Human Health Effects. Harmful Algae 3,2(April 2004): 99-115. Kusek, K.M., G. Vargo, and K. Steidinger. Gymnodi nium breve in the Field, in the Lab, and in the Newspaper A Scientific and Journali stic Analysis of Florida Red Tides. Contributions in Marine Science, 34(1999):1-228. Landsberg, J.H. The Effects of Harmfu l Algal Blooms on Aquatic Organisms. Reviews in Fisheries Science 10,2(2002):113-390. Larkin, S.L., C.A. Adams. Red Tides and Coastal Businesses: Measuring Economic Consequences in Florida. Society and Natural Resources 2007. Leeworthy, V.R. and P. C. Wiley. Current Pa rticipation Patterns in Marine Recreation National Survey on Recreation and the Envi ronment 2000. Silver Springs, MD: U.S. Department of Commerce, National Oceanic and Atmospheric Administration, National Ocean Service, November 2001. Lin, P.C., R.M. Adams, and R.P. Berrens. Welfare Effects of Fishery Policies: Native American Treaty Rights and Recreational Salmon Fishing. Journal of Agricultural and Resource Economics 21,2(December, 1996): 263-276. 103

PAGE 104

Magana, H.A., C. Contreras, and T.A. Villareal. A Historical Assessment of Karenia brevis in the Western Gulf of Mexico. Harmful Algae 2(2003):163-171. McFadden, D. Conditional Logit Analysis of Qual itative Choice Behavior. P. Zarembka, ed. Frontiers in Econometrics Academic Press, New York (1973). McLaughlin, K. and K. Spinner. Red Tide Brings Red Ink to Coastal Businesses. The Sarasota Herald Tribune, The New York Times Co., New York, New York, October 9, 2006. Moore, C. Fear of Red Tide Remains. St. Petersburg Times, The Times Publishing Company, St. Petersburg, Florida, July 26, 2006. Morgan, R. Preferences and Priorities of Recreational Beach Users in Wales, UK. Journal of Coastal Research 15,5(1999):653-677. Moroney, J. County Says Tourism Drowning. The News-Press, Fort Myers, FL, December 5, 2005. National Oceanic and Atmospheric Administration [NOAA] National Marine Fisheries Service. Prevention, Control, and Mitigation of Harmful Algal Blooms: A Research Plan. September 2001 Internet Site: http://www.whoi.edu/science/B/redtide/pertinentinfo/PCM_HAB_Research_Plan (Accessed March 2007). Nordhaus. W.D. The Economic Impacts of Abrupt Climatic Change. Prepared for Meeting on Abrupt Climate Change: The Role of O ceans, Atmosphere, and the Polar Regions, National Research Council, January 22, 1999. Nunes, P. and J. van den Bergh. Can People Va lue Protection against In vasive Marine Species? Evidence from a Joint TCCV Survey in the Netherlands. Environmental and Resource Economics 28(August 2004):517-532. Pain, J. Manatees Red Tide Victims. The Asso ciated Press, New York, New York, January 3, 2004. Pendleton, L., N. Martin, and D.G. Webster. Pub lic Perceptions of Environmental Quality: A Survey Study of Beach Use and Perceptions in Los Angeles County. Los Angeles, CA: University of Southern California, June 2000. Perry, M.J. and P.J. Mackun. Census 2000 Brief Population Change and Distribution, 19902000. Washington, DC: U.S. Census Burea u, Pub. No. C2KBR/01-2 (April, 2001). Pierce, R.H., M.S. Henry, C.J. Higham, P. Blum, M.S. Sengco, D.M. Anderson. Removal of Harmful Algal Cells ( Karenia brevis ) and Toxins from Seawater Culture by Clay Flocculation. Harmful Algae, 3(2004):141-148. 104

PAGE 105

Robbins, I.C., G.J. Kirkpatrick, S.M. Blackwell, J. Hillier, C.A. Knight, and M.A. Moline. Improved Monitoring of HABs Using Au tonomous Underwater Vehicles (AUV). Harmful Algae 5(2003):749-761. Schneider, K.R., R.H. Pierce, and G. E. Rodrick. The degradation of Karenia brevis Toxins Utilizing Ozonated Seawater. Harmful Algae 2(2003):101-107. Shumway, S.E. A Review of the Effects of Algal Blooms on Shellfish and Aquaculture. Journal of the World Aquaculture Society 21(1990): 65-104. Steidinger, K.A., Landsberg, J.H., Tomas, C.R., Burns, J.W. Har mful Algal Blooms in Florida. Harmful Algal Bloom Task Force Technica l Advisory Group Report #1. Submitted to Floridas Harmful Algal Bloom Task Force, Florida Department of Environmental Protection, Tallahassee, FL (1999). Stumpf, R. P. Application of Satellite Ocean Color Sensors for Monitoring and Predicting Harmful Algal Blooms. Human and Ecological Risk Assessment 7,5(2001):1363-1368. Tester, P.A., R.P. Stumpf, F.M. Vukovich, P.K. Fowler, and J.T. Turner. An Expatriate Red Tide Bloom: Transport, Dist ribution, and Persistence. Limnology and Oceanography 36(1991):1053-1061. Tibbets, S. Mack, Buchanan, and Castor Figh ting for More Red Tide Research Funding. Congressman Vern Buchanan Pr ess Release. Internet site: http://buchanan.house.gov/apps/list/press/fl13_buchanan/tide.html (Accessed August 1, 2007). Tomalin, T. Red Tide Takes Toll on Sea Turtles. St. Petersburg Times, The Times Publishing Company, St. Petersburg, Florida, September 23, 2005. Turner, R.K., J.C.J.M van den Bergh, T. Soderqvi st, A. Barendregt, J. van der Straaten, E. Maltby, E.C. van Ierland. Ecological-ec onomic Analysis of Wetlands: Scientific Integration for Management and Policy. Ecological Economics 35(2000):7-23. U.S. Census Bureau. Internet site: http://www.census.gov/ (Accessed July 2007). Van Sant, W. Red Tide Toll: 732 Tons of Decay. St. Petersburg Times, The Times Publishing Company, St. Petersburg, FL, September 10, 2005. Visit Florida [VISIT FLA] Domestic Visitors to Florida, Florida Vis itors Study, 2005. Internet site: http://media.visitf lorida.org/about/research/ (Accessed 25 July 2007). 105

PAGE 106

BIOGRAPHICAL SKETCH Kimberly Ludwig Morgan graduated fr om Mendham High School in Mendham, New Jersey, in 1988. She received her Bachelor of Sc ience in Animal Science in December 1993, and her Master of Science in Food and Resource Economics was awarded in December 1997, both from the University of Florida in Gainesvi lle. Since 2000, she has worked as an economic analyst for the Florida Agricultural Market Research Center, participating in many agricultureoriented survey research efforts for a va riety of producer, extension, and government organizations. For the past five years she has serv ed as an adjunct faculty member for Southern New Hampshire University, facilitating distance education courses such as Principles of Macroeconomics and Microeconomics, Manageri al Economics, and Economics for Business. Kimberly lives in High Springs, Florida, with her husband Michael and th ree beautiful children, on a mini-ranch where they cater to eight equine friends, three cats, a standard poodle, and an African grey parrot. 106