MOTIVATIONAL, FOREST MANAGEMENT AND CLIMATIC FACTORS AFFECTING OUTDOOR RECREATION PARTICIPATION AND EXPERIENCE IN FLORIDA By RAMESH PAUDYAL A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA 2017
2017 Ramesh Paudyal
To my beloved wife for her love support and inspiration
4 ACKNOWLEDGMENTS I am sincerely thankful to every person who offered me guidance, support, and encouragement during my PhD program and dissertation. I am grateful to my advisor Dr. Taylor Stein for his committed supervision, support, and encouragement both as a mentor and as a friend throughout my PhD program. Taylor was always available when I needed his support and guidance. He consistently offered me great suggestions, yet allow ing me the flexibility to develop my unique program and dissertation research. I thank Taylor for allowing me the opportunity to get involved in multiple research projects and attend several conferences, which was very helpful to develop thoughts and ideas for my dissertation research. I would like to thank my committee members Dr s Mickie Swisher, Damian Adams, Holly Ober, and Eric Jokela for their input, guidance, and support throughout my PhD program and dis sertation research. Dr. Swisher helped me to understand the concept of research design and research methods. Her input was very helpful in developing dissertation research project and survey instrument. I will always be grateful for her friendly, caring, a nd supportive mentoring. in developing critical thinking regarding implications of research findings. Every meeting and conversation with him always resulted to a new thought about how my research fits in to a border po licy and management perspective. I am thankful to Dr. Ober and Dr. Jokela for helping me to integrate wildlife and forest management component in my dissertation research. I am also thankful to Dr. Brijesh Thapa for his valuable advices and support in my p rogram and resea rch from outside of the committee. I am thankful to School of Forest Resources and Conservation (SFRC) and the Graduate School for providing me a fellowship throughout my PhD program. Without
5 this funding I would not be able to complete my PhD program. My thanks also go to all the SFRC staffs, specially Ms. Amy Abernethy, Ms. Cindy Love, and Ms. Sarah Hernandez, for their admi nistrative supports. My PhD program would not have been complete without the support from numerous individuals and in stitutions. I would like to thank my colleague Dr. Bin Wan for all his guidance and support throughout my PhD program and for assisting in collection and management of trail use data. I also thank Mr. Carson Young for his assistance in onsite data collecti on. I am also thankful to Mr. James Colee for his assistance in statistical analysis. Thanks also go to colleagues Dr. Tracey Ritchie, Ms. Sara Komenda, Ms Sheena Simmons, Dr. Derek Allen and Ms. Cassandra Ward for all their support, feedback, encouragement company, stress sharing, smiles and many other things that helped me succeed in this PhD program I also would like to thank Ms. Elizabeth Ramirez (Ocala NF), Ms. Sarah Lauerman (Aves Consulting) and Ms. Debra Stucki (Osceola NF) for provi ding me maps and assistance in identifying prescribed burning plots and red cockaded woodpecker habitats My sincere gratitude goes to U F Fish and Wildlife Conservation Commission for providing me hunters for online data collection I also thank UF IRB for issuing me a permit to conduct this research. Finally, I would like to thank my family, relativ es, and friends for their love, m otivation, and support. I would not have b een able to complete this journey without them all
6 TABLE OF CONTENTS page ACKNOWLEDGMENTS ................................ ................................ ................................ .. 4 LIST OF TABLES ................................ ................................ ................................ ............ 9 LIST OF FIGURES ................................ ................................ ................................ ........ 10 LIST OF ABBREVIATIONS ................................ ................................ ........................... 12 ABSTRACT ................................ ................................ ................................ ................... 14 CHAPTER 1 INTRODUCTION ................................ ................................ ................................ .... 16 Resear ch Questions ................................ ................................ ............................... 18 Dissertation Format ................................ ................................ ................................ 19 2 MOTIVATIONS AND SITE ATTRIBUTE PREFERENCES OF FLORIDA NATIONAL SCENIC TRAIL VISITORS: MARKET SEGMENTS AND STRUCTURAL RELATIONS ................................ ................................ .................. 21 Background ................................ ................................ ................................ ............. 21 Review of Literatu re ................................ ................................ ................................ 22 Outcome Focused Management and Push/Pull Theory ................................ ... 22 Social Marketing ................................ ................................ ............................... 25 Methods ................................ ................................ ................................ .................. 29 Research Design and Survey Instrument ................................ ......................... 29 Variable Measurement and Analysis ................................ ................................ 30 Results ................................ ................................ ................................ .................... 32 Descriptive Characteristics of Sample ................................ .............................. 32 Motivations and Site Attribute Preferences ................................ ....................... 33 Structural Relations among Push and Pull Factors ................................ .......... 35 Discussion ................................ ................................ ................................ .............. 37 Management Implications ................................ ................................ ................. 38 Limi tations and Future Directions of Research ................................ ................. 43 3 HOW DO ACTIVITY AND KNOWLEDGE OF PRESCRIBED BURNING ION OF SCENIC BEAUTY AND RECREATION SATISFACTION FROM RED COCKADED WOODPECKER HABITAT? ................................ ................................ ................................ ............... 54 Background ................................ ................................ ................................ ............. 54 Conceptual Framework ................................ ................................ ........................... 55 Scenic Beauty and Recreation Satisfaction ................................ ...................... 56
7 Socio cultural and Experiential Factors ................................ ............................ 60 Knowledge and Information ................................ ................................ .............. 62 Methods ................................ ................................ ................................ .................. 66 Research Design and Survey Instrument ................................ ......................... 66 Measurement and Analysis ................................ ................................ .............. 67 Results ................................ ................................ ................................ .................... 72 Descriptive Characteristics of Sample ................................ .............................. 72 Knowledge of Prescribed Burning ................................ ................................ .... 73 Perception of Scenic Beauty and Recreation Satisfaction ................................ 73 Effect of Ecological Information ................................ ................................ ........ 74 Structural Relations ................................ ................................ .......................... 75 Discussi on ................................ ................................ ................................ .............. 76 Management Implications ................................ ................................ ................. 77 Limitations and Future Directions of Research ................................ ................. 79 4 CLIMATIC, TEMPORAL, AND ECONOMIC FACTORS AFFECTI NG RECREATION VISITS TO FLORIDA NATIONAL SCENIC TRAIL: A TIME SERIES MODELING ................................ ................................ ............................... 87 Introduction ................................ ................................ ................................ ............. 87 Review of Literature ................................ ................................ ................................ 90 Effect of Weather and Climate ................................ ................................ .......... 90 Temporal and Economic Factors ................................ ................................ ...... 95 Methods ................................ ................................ ................................ .................. 96 Study Area ................................ ................................ ................................ ........ 96 Data and Variables ................................ ................................ ........................... 97 Model Selection ................................ ................................ ................................ 98 Count data m odel ................................ ................................ ....................... 98 Time series model ................................ ................................ .................... 100 Results ................................ ................................ ................................ .................. 104 Descriptive Characteristics of the Variables ................................ ................... 104 Comparison of Best Fit Time Series and Count Data Model .......................... 106 Regression Results from Final Model ................................ ............................. 107 Discussion ................................ ................................ ................................ ............ 108 Research and Management Implications ................................ ....................... 110 Limitations and Future Research ................................ ................................ .... 112 5 CONCLUSION ................................ ................................ ................................ ...... 130 APPENDIX A QUESTIONNAIRE FOR RECREATION MOTIVATIONS AND SITE ATTRIBUTE PREFERENCES SURVEY ................................ ................................ ................... 13 4 B QUESTIONNAIRE FOR PRESCRIBED BURNING AND AESTHETIC PREFERENCE SURVEY ................................ ................................ ...................... 139
8 LIST OF REFERENCES ................................ ................................ ............................. 145 BIOGRAPHICAL SKETCH ................................ ................................ .......................... 160
9 LIST OF TABLES T able page 2 1 Index mean and measurement reliability of motivations domains of users of the Florida National Scenic Trail. ................................ ................................ ........ 47 2 2 Index mean and measurement reliability of site attribute preferences (pull factor) domains of users of the Florida National Scenic Trail. ............................ 48 2 3 Comparison of cluster segments by motivations and site attribute preferences of users of the Florida National Scenic Trail ................................ .. 49 2 4 Comparison of cluster segments by demographic characteristic s of users of the Florida National Scenic Trail. ................................ ................................ ........ 50 2 5 Comparison of cluster segments by recreation characteristics of users of the Florida National Scenic Trail. ................................ ................................ .............. 51 3 1 Demographic characteristics of adult outdoor recreation users in Ocala National Forest, USA. ................................ ................................ ......................... 82 3 2 Knowledge of prescribed burning of adult outdoor recreation users in Ocala National Forest, USA. ................................ ................................ ......................... 83 3 3 Comparison of index scores of knowledge of prescribed burning, perceived scenic beauty, and recreation satisfaction by recreation activity of adult outd oor recreation users in Ocala National Forest, USA ................................ ... 83 4 1 Descriptive characteristics of independe nt variables in the model examining daily recreation visits (dependent variable) to Ocala National Forest, USA. ..... 115 4 2 Theoretical characteristics of ACF and PACF for stationary data. .................... 115 4 3 AIC and BIC of competing time series models without ex ogenous regressors. 116 4 4 AIC, BIC and errors of best three time series models with exogenous regressors. ................................ ................................ ................................ ........ 116 4 5 Model fit comparison of negative binomial and final time series model. ........... 117 4 6 Regression results from SARIMA (100,101) 7 with regressors model examining daily recreation visits (dependent variable). ................................ .... 117
10 LIST OF FIGURES F igure page 1 1 Diagrammatic structure of dissertation chapters ................................ ................. 20 2 1 Study area and survey sites in Florida, USA. ................................ ..................... 52 2 2 Proportion of respondents by survey sites in Florida, USA. ................................ 52 2 3 Measurement models and structural relations among motivations (push factors) and site attribute preferences (pull factors) of users of the Florida National Scenic Trail ................................ ................................ .......................... 53 3 1 Photographic images used for scenic beauty and recreation satisfaction ratings of adult outdoor recreation users in Ocala National Forest, USA ........... 84 3 2 Comparison of perceived scenic beauty and recreation satisfaction by pictures representing various forest conditions.. ................................ ................. 85 3 3 Comparison of perceived scenic beauty and recreation satisfaction by treatment groups. ................................ ................................ ............................... 85 3 4 Structural relations among knowledge, scenic beauty, and recreation satisfaction, and associations with demographics and activity of participation of adult outdoor recreation users in Ocala National Forest, USA ...................... 86 4 1 Study area with sampling locations at Ocala National Forest Section of F lorida National Scenic Trail, in Florida, USA. ................................ .................. 118 4 2 Frequency distribution of average daily visits to Ocala National Forest Section of Florida National Scenic Trail (dependent variable), USA. ................ 119 4 4 Average daily visits to Ocala National Forest Section of Florida National Scenic Trail by year. ................................ ................................ ......................... 121 4 5 Average daily visits to Ocala National Forest Section of Florida Natio nal Scenic Trail by months. ................................ ................................ .................... 122 4 6 Average daily visits to Ocala National Forest Section of Florida National Scenic Trail by da y of the week. ................................ ................................ ....... 123 4 7 Average daily visits to Ocala National Forest section of Florida National Scenic Trail and daily avg. temperature (C) during the period of January 1, 2010 to May 31, 2015. ................................ ................................ ...................... 124
11 4 8 Average daily visits to Ocala National Forest se ction of Florida National Scenic Trail by various ranges of average daily temperature (C) during the period of January 1, 2010 to May 31, 2015. ................................ ..................... 125 4 9 Occurrence of various heat index values (left axis) and mean recreational visits (right axis) in Ocala National Forest section of Florida National Scenic Trail during the time period of January 1, 2010 to May 31, 2015. ..................... 126 4 10 Average daily visits to Ocala National Forest section of Florida National Scenic Trail and average daily price for regular gas in Florida during the study period of January 1, 2010 to May 31, 2015. ................................ ............ 127 4 11 Autocorrelation function (ACF) and Partial autocorrelation function (PACF) of residuals from final negative binomial regression model and timeseries SARIMAX ((1,0,0) (1,0,1) 7 with regressors) model. ................................ .......... 128 4 12 Histograms of residuals and plots of fitted Vs residuals from final negative binomial regression model and timeseries SARIMAX ((1,0,0) (1,0,1) 7 with regressors) model. ................................ ................................ ............................ 129
12 LIST OF ABBREVIATIONS A CF Autocorrelation Function AFM Activity Focused Management AIC Akaike Information Criterion ANOVA Analysis of Variance AR Autoregressive ARIMA Autoregressive Integrated Moving Average BBM Benefit Based Management BIC Bayesian Information Criterion CFI Confirmatory Fit Index EFM Experience Focused Management FNST Florida National Scenic Trail FTA Florida Trail Association IPCC Intergovernmental Panel on Climate Change MA Moving Average NF National Forest OFM Outcome Focused Management OHV Off Highway Vehicle ONF Ocala National Forest PACF Partial Autocorrelation Function RCW Red Cockaded Woodpecker RMSEA Root Mean Square Error of Approximation ROS Recreation Opportunity Spectrum SARIMA Seasonal Autoregressive Integrated Moving Average SBE Scenic Beauty Estimation
13 SMS Scenery Management System SRMR Standardized Root Mean Square Residual UF University of Florida USDA United States Department of Agriculture USFS United States Forest Service WMA Wildlife Management Area
14 Abstract of Dissertation Presented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy MOTIVATIONAL, FOREST MANAGEMENT, AND CLIMATIC FACTORS AFFECTING OUTDOOR RECREATION PARTICIPATION AND EXPERIENCE IN FLORIDA By Ramesh Paudyal December 2017 Chair: Taylor Stein Major: Forest Resources and Conservation This study identifies various factors affecting outdoor recreation participation and Florida National Scenic Trail ( FNST ) and certain managed areas along the FNST. Analysis of motivations and site attribute preferences and market segments of FNST visito rs reve aled five domains of motivations, four domains of site attribute preference, and four market segments which differed significantly in terms of Complex structural relationships found among motiva tions and site attribute preferences suggest a need to manage forest and recreation areas according to the experiences visitors expect to attain rather than according to the recreation activities they are engaged in. Results also indicated that a t least o ne third of recreation users held some sort of misconception about prescribed burning and its effect on forest health, wildlife, and scenery. recreation satisfaction indicated that the higher perceived scenic beauty of recreation sites, the more likely they will be satisfied with their recreation Findings suggest that
15 desired recreation experience can be enhanced by improving their knowledge about the importance of prescribed burning in forest and wildlife habitat management. Analysis of factors associated with the daily use volume of the FNST indicates that past use levels serve as key explanatory factors of future use. Also, t emperature relative humidity gas price and weekend and public holidays sig nificantly affect daily visits to FNST. Days with average temperature of 4 23 C seems to bring large numbers of visitors, and average temperature range of 10 12 C seems to be most favorable daily temperature to recreate in FNST. Days with heavy rainfall or high heat index are likely to impact recreation use level negatively of not only the same day, but also the next day even if the following day has normal weather condition s Overall, findings indicate that marketing programs should be specific to capture motivations and preferences of each visitor typology to attract visitors with diverse motivations. O utreach programs should target visitors lacking knowledge of prescribed burning and facilitate knowledge transfer among the users. M anagers of facilities that require staffing and other resources management should expect less visitation on days following extreme weather condition s
16 CHAPTER 1 INTRODUCTION Various factors could affect outdoor recreation participation and experience. Most nature based recreation participation research has focused on wildland settings areas relatively far from human development, which include wilderness areas (McFarlane, Boxall, & Watson, 1998) national parks (Warzecha & Lime, 2001) or other recreation areas in remote areas of the western United States (Budruk, Virden, & Waskey, 2009; Stein & Lee, 1995) N ature based recreation studies have also been conducted in the southeastern United States but they are few in comparison to similar studies from the western United States. As a result, many unique attributes associated with the climate, land, forest, and water in the Southeast have not yet been fully explored. For example, many nature based areas are relatively small islands surrounded by private land and urban development. Also, it is common that many of thes e recreation areas are within a few hours driving time for millions of people. Thus, how people in this reg ion, who likely have different relationship s with forest and natural areas, perceive, experience, and prefer nature and nature bas ed activities must be explored. Likewise, this region include s subtropical ecosystems and has a warm and humid climate which potentially ma ke s recreation participation and experience much different in this region than in other parts of the U.S. W eather extremes (e.g., heat index, wind and precipitation extremes) are frequent in this region and specifically in Florida Such extremes significa ntly affect outdoor recreation decisions and can challenge recreation mangers wit h unpredicted shifts in demand and resource use loss of potential profit, and difficulties in administrative scheduling (Manning & Powers, 1984;
17 Yu, Schwartz, & Walsh, 2009) Effect of weather and climate on outdoor recreation is a well researched topic (Aylen, Albertson, & Cavan, 2014; Beaudin & Huang, 2014; Becken & Hay, 2007; de Freitas, 2015; Hewer, Scott, & Gough, 2015; Li & Lin, 2012; Nicholls, Holecek, & No h, 2008) but there are limited studies of this kind from this region. Short winter s, long hot and humid summer s and frequent weather extremes suggest that weather and climate may significantly affect outdoor recreation participation in Florida, but how and to what extent remains unknown Similarly, Florida has one of the 11 national scenic trials in the United States It travers e s several managed areas across the state and offer s a variety of recreation experience s to visitors. However, we have limited understanding about users demographic and recreation profile s experience preference s and choice s of destinations o n this trail. Given the physical settings (topography) and variation of ecosystems (swamps, mangrove s, scrubs, pinelands) in Florida, recreation experience s people attain from the Florida National Scenic Trail (FNST) probably differ from what people experience in other scenic trails like the Appalachian National Scenic Trail. Therefore, findings related to this topic based on studies of other scenic trails are difficult to generalize for this setting. C and preferences is essential for partner agencies like the US Forest Service (USFS) and Florida Trail Associ ation (FTA) to maintain and market this trail. Finally, the southeastern U.S. is rich in the longleaf pine ecosystem, which supports habitat for many endangered and ecologically important keystone species, such as red cockaded woodpecker ( Picoides borealis ) and gopher tortoise ( Gopherus Polyphemus) These longleaf ecosystems also offer attractive destinations for outdoor
18 recreation. However, longleaf forests and some associated wildlife species including the red cockaded woodpecker (R CW ) prefer frequent burning to thrive, w hich may be aesthetically un appealing for recreation users because of temporary negative effect s of fire on scenery (Gobster, 1999; Ryan, 2005; USDA Forest Service, 1995) P eople may consider fire as a n obstacle to attaining the experience they desire because of lack of knowledge and unders tanding about prescribed burning I nformation about recreation users knowledge and perception about prescribed burning and its association with their recreation experience is therefore important for managers to practice multi objective forestry on public lands. The overall goal of m y research is to understand how visitors make decisions based recreation areas and how various factors such as forest management and weather, affect their perceptions and decisions Specifically, it examines and perceptions are related to the specific characterist the environment. To achieve this broad goal, the dissertation addresses thre e major research questions Research Questions 1. What are the types and characteristics of visitors to Florida National Scenic Trail preferences of various site amenities ? 2. How does the management of natural areas using prescribed burning for endangered species habitat (e.g., RCW quality of recreation opportunities in forested settings? 3. How do temporal and climatic factors affect da ily outdoor recreation visits to a long distance scenic trail in Florida?
19 D issertation Format This dissertation is organized into five chapters that include three studies Chapter 1 provides a general background about the problem and research questions tha t are answered in this dissertation. This chapter then explain s the format of the dissertation and the diagrammatic stru cture of chapters (Figure 1 1). Chapter 2 examines salient motivations and site attribute preferences of FNST visitors. It then identifies market segments of visitors and present s the cha racteristics of segments and demographics. This chapter also examines the multi dimension al relations between motivations and site attribute preferences using structural equation modeling. knowledge of prescribed burning and perception of scenic beauty and recreation satisfaction from forest sites represen ting typical RCW habitat s in Ocala NF. This chapter presents multiple approaches to understand associations among activities users chose to participate in knowledge of prescribed burning, perception of scenic beauty and attainment of recreation satisfact ion. Chapter 4 identifies the favorable and preferable range of temperature for trail activities in Florida. It compares two modeling approaches, timeseries and count data, to examine the association of recreation participation with climatic factors like temperature and precipitation and other factors like the day of the week and gas price. Best fit time series model showed that weather extremes not only affect same day recr eation participation but also the next day even if the day has favorable weather co ndition s Finally, Chapter 5 presents a summary of conclusions and implications drawn from th ese three stud ies
20 Figure 1 1 Diagrammatic structure of dissertation chapters
21 CHAPTER 2 MOTIVATIONS AND SITE ATTRIBUTE PREFERENCES OF FLORIDA NATIONAL SCENIC TRAIL VISITORS: MARKET SEGMENTS AND STRUCTURAL RELATIONS B ackground Hiking is one of the most popular outdoor activities in the U S A national survey has shown that about one third of American adults 79 m illion people hiked in 2008 and this number is predicted to increase by 3 10 percent by 2060 (Bowker et al., 2012) Among the various recreation destinations, scenic trails provide diverse recreation opportunities for users. The Florida National Scenic Trail (FNST) is one of eleven national scenic trails in the United States. Stretching 1,000 miles (~1600 km) from Big Cypress National Preserve, in the south, to Gulf Islands National Seashore, in the northwest, the FNST traverses through the most beautiful and unique landscapes in Florida. An assessment of visitation from 2003 to 2014 has shown that FNST receives about 350,500 visits per year (Wan, Paudyal, Huntley, & Stein, 2014) However, we ha ve for visiting the FNST their site preferences associated with attributes of the site like level of development and accessibility. Researchers have argued that the destination serves as a medium through which socio psychological needs or outcomes of recreation users could be attained (Crompton, 1979) and that identifying market segments provides a means to match ds and the management resources available (Kozak, 2002; Smith, Tuffin, Taplin, Moore, & Tonge, 2014) If recreation mange rs have a clear understanding of how different site attributes are preferred by various segments of recreation users, they will be able to maintain recreation areas to meet needs and preference s of diverse recreation visitors. Such information can also hel p to frame effective marketing
22 message s to attract the visitors. Although each recreation destination ha s its unique characteristics in terms of site attractions, recreation users ch o ose the site that they perceive as most likely to match with their motiva tion s. A study conducted earlier with FNST visitors found that almost all visitors were aware that Florida had a national scenic trail, but more than half were unaware that they were recreating on it (Albritton, Denny, & Stein, 2006) Consequently, it is likely that many Floridians are unaware about the exact location of and recreation opportunities offered by the FN ST even if they are using this trail. Hence, developing a characteristics, motivation s and site preferences is important to develop and implement effective marketing program s T his research uses t hree approaches to fill this gap in knowledge : 1) Identify the salient motivations and site attribute preferences of FNST visitors, 2) i dentify the market segments of FNST visitors based on their recreation motivations and site attribute preferences; and 3 ) e xamine the s tructural relations between motivations and site attributes preferences. Review of Literature Outcome Focused Management and Push/Pull Theory Until the early 1970s, the management of recreation resources, services, and facilities in the USA was based on th e supply oriented philosophy or Activity Focused Management (AFM). AFM assume s that providing opportunities for recreation activities like fishing, swimming and camping will bring visitors to recreation sites. Although this management approach was useful f or describing the number of people participating in any given activity, this framework did not
23 answer key important recreation management questions, such as what satisfactions or rewards did participants get from the activity o r how could the quality of th e experience be enhanced Beginning in the early 1970s, Driver and his associates initiated research recreation participation They developed a conceptual foundation and empirical approaches for testing the role of motivations in outdoor recreation participation (Driver, 1983, 2008; Driver & Tocher, 1970; Moore & Driver, 2005) Their work contributed to the change in management approach from ACF to Experience Focused Management (EFM), which focuse s on providing satisfying recreation experience s for visitors. The work of Driver and his associates was based on a psychological perspective, particularly the expectancy vale nce theory of motivation. Expectancy valance theory suggests that people participate in outdoor recreation because of their desire to achieve specific psychological outcomes, such as satisfaction with the recreation experience (Moore & Driver, 2005) E xpectancy valence theory integrates two major concepts expectancy and valence. E xpectancy is the likelihood that an action will lead to a c ertain outcome or goal, whereas valence is the attractiveness of or anticipated satisfaction with an outcome (Moore & Driver, 2005) V al ence and expectancy combine to (Andereck, Gard McGehee, Lee, & Clemmons, 2011) Beginning in the early 1990s, researchers conceptualized the benefits from recreation which led to the development of Benefit Based Management (BBM), now known as Outcome Focused Management (OFM). Th e fundamental concept of benefits include s 1) attainment of a desired condition, 2) attainment of an improved condition,
24 and/or 3) prevention of an unwanted condition (Driver, 2008; Stein & Lee, 1995) Further, the benefits (outcome s ) can pertain to individuals (e.g., attainment of physical and mental health), society at large (e.g., enhance d community pride), the economy (e.g., local economic growth), and the environment (e.g., protection of endangered species). Many researchers have contributed to the development of the OFM approach through understanding of desired outcomes of diverse groups of recreationists and publics (Anderson et al., 2000; Driver, 2008; Kil, Holland, Stein, & Ko, 2012; Moore & Driver, 2005; Stein & Anderson, 2002; Stein & Lee, 1995) For example, Stein and Lee (1995) found connections between the benefits sought by the recreationists and physical, managerial, and social setting e in a variety of activities lead ing to achievement of multiple benefits. Based on the OFM approach this research Elaborating the concept of motivation, Crompton (1979) developed a theoretical framework called push / pull theory that was initially used in the tourism context and later extended to outdoor recreation participation. Research employing this conceptual framework explains motivation s as one of the major determining factor s in choosing a tourism or recreation site/activity (Baloglu & Uysal, 1996; Crompton, 1979; Dann, 1981; Kim, Lee, & Klenosky, 2003) As its name implies, push/pull theory suggests that push and pull factors are the important c onstituents in motivating tourist s and recreationist s (Dann, 1981) Push factors refer to the internal motives of the individual, while pull factors are the destination features and attractions (McCool & Moisey, 2001) According to
25 this notion, push and pull factors are conceptually different factors but they are not independent of each other as travel deci sions take place in a two stage sequential process (Kim, Noh, & Jogaratna m, 2007) Once recreationist s declare their recreation intention based on their intangible or intrinsic desires or motivations like escape from the usual demands of life, rest and relaxation, health and fitness, adventure or social interaction tangibl e destination attributes like aesthetics, accessibility and wildlife influence their decision s in selecting a specific site (Baloglu & Uysal, 1996) Thus, push and pull factors are related to two separate decisions made at two separate points in time ; one focuses on whether to go whereas the other focuses on where to go (Klenosky, 2002) Social Marketing Recreationists are heterogeneous because they exhibit different range s of attitudes, preferences and behavior s (Needham, Vaske, Donnelly, & Manfredo, 2007) Inherent heterogeneity within the user community in terms of needs and preferences can create chal lenge s for managers in providing diverse benefits. A s ocial marketing strategy has been widely used to promote experience opportunities services, and facilities of recreation areas to diverse users. The first conceptual definition of social marketing was programs calculated to influence the acceptability of social ideas and involving considerations of product planning, pricing, communication, distribution, and marketing (Kotler & Zaltman, 1971, p. 12) It is conceptually the same approach as commercial marketing but its focus is on social benefit s rather than commercial profit (Rangun & Karim, 1991) S ocial marketing prima rily seek s to benefit individuals and families or the broader society but not profits for the marketing organization itself.
26 C luster segmentation has been one widely used tool in social marketing to ferences, and behaviors by grouping heterogeneous individuals into homogenous groups (Bright, Manfredo, & Fulton, 2000; Hubert & Gipson, 1996; Kyle, Norman, Jodice, Graefe, & Marsinko, 2007; Needham, 2010; Paudyal et al., 2015) Many researche rs have considered motivation as a basis for market segmentation in various contexts of tourism and outdoor recreation (Beh & Bruyere, 2007; Crompton, 1979; Lee, Lee, & Wicks, 2004; Loker Murphy, 1997; Nyaupane, White, & Budruk, 2006; Park & Yoon, 2009) However, most of these stu dies neglected to include site attribute preferences on market segmentation. Likewise, m any researchers have examined relationships between motivations (push factors) and site preferences (pull factors) with various other experiential, cognitiv, and behav ioral factors. For example researchers have examined how motivation and site attribute preferences relate to leisure involvement ( Kyle, Absher, Hammitt, & Cavin, 2006) satisfaction and destination loyalty (Yoon & Uysal, 2005) support for management practices (Hall, Seekamp, & Cole, 2010) environmental attitude s and behavior s (Kil, Holland, & Stein, 2014) place attachment and future visit intention s (Budruk & Stanis, 2013; Kil, Holland, & Stein, 2015; Kil et al., 2012; Lee, 2009) However, these studies did not examine how push factors and pull factors relate to each other. R esearchers have examined motivations and site preferences of tourists and recreation users to various destinations and have identified various traits of motivation domains and market segments (Beh & Bruyere, 2007; Crompton,
27 1979; Dillard & Bates, 2011; Iso Ahola, 1982; Kim et al., 2007; Kim et al., 2003; Loker Murphy, 1997; Smith et al., 2014; Yoon & Uysal, 2005) For example, t hrough unstructured interview s with a sample from two cities in USA, Crompton (1979) identified nine domains of moti vations, which he classified in to two broad categories socio psychological and cultural. The socio psychological category included seven domains: (1) E scape from a perceived mundane enviro nment (2) exploration and evaluation of self, (3) relaxation (4) prestige (5) regression (6) enhancement of kinship relationships and (7) facilitation of social interaction Likewise, Loker Murphy (1997) examined motivations of backpackers to national parks in Australia and found meeting local people and excitement/adventure the most common motivat ions. Using cluster analysis, Loker Murphy characterized four groups of backpackers as achievers, self developers, social/excitement seekers and escapers/relaxers Similarly Kim et al. (2003) examined push and pull factors among visitors to six national parks in South Korea family togetherness and study, appreciating natural resources and health escaping from everyday routine adventure and building friendship he pull factor domains were key tourist resources information and convenience of facilities and accessibility and transportation Similarly, Yoon and Uysal (2005) examined push and pull motivations of tourists visiting Northern Cyprus and found that safety and fun escape knowledge and education and achievement were the important push factors. leannes s and shopping reliable weather and safety different culture and water activities were perceived as the major pull factors. Beh and Bruyere (2007) e xamined travel motivation s and among visitors to three national reserves in Kenya. M otivation domains identified were escape culture
28 personal growth mega fauna adventure learning nature, and general viewing. escapists, learners, a nd spiritualists. Smith et al. (2014) segmented Australian park visitors based on the purpose of the visit and activity undertaken at the site. Groups identified were ature e xperience s eekers, assive e xperiencers, ature e xplorers, and ocializers As discussed above, both motivation and site attribute preference s are a dynamic and multi dimensional concept that can vary across individuals and groups as well as among destinations (Kim et al., 2007; Kozak, 2002; Stein & Lee, 1995; Uysal & Hagan, 1993) The findings imply that motivations and site attribute preferences, market segments, and structural relation ship s between motivation s and site preference s found in other study location s and population s ma y be less meaningful for marketing to FNST visitors. Few previous studies have examined motivations and/or site attribute preferences of FNST visitors and nearby natural areas (non FNST). However Albritton and Stein (2007) examined and compared FNST visitors to non FNST visitors and found that these two groups significantly differed in terms of their recreation motivations and site preferences. Likewise, Albritton and Stein (2008) segmented non FNST visitors based on their motivations and site preferences in order to develop marketing strateg ies to attract non FNST visitors to FNST. Although these studies offered some understanding about the major motivations and site preferences of FNST visitors and market segments of non FNST visitors, they ignored the heterogeneity that existed among FNST v isitors and the relations hips between
29 their motivations and site preferences. Therefore, identifying the market segments of visitors and examining the relationships between individual dimension s of motivation and site preference is important to better unde rstand the purpose, needs, and preferences of visitors to this FNST Taking a social marketing approach, this study segments a sample of FNST visitors into homogenous subgroups based on their reported recreation motivations (push factors) and importance of site attributes (pull factors) in selecting a recreation site. Then the research examines and contrasts the subgroups based on demographic trait s and trip c haracteristics. Methods Research Design and Survey Instrument The population of interest for this s tudy was recreation trail users in Florida. I first identified m ajor recreation areas along the FNST within an approximate driving distance of two hours from the I selected seven recreation areas randomly and colle cted data onsite using a questionnaire and index. The s tudy sites included two national forests, two wildlife management areas, a state forest, a state park, and a greenway (Figure 2 1). I collected 353 responses during Fall 2011 Spring 2014 with a respo nse rate of 89%. Among the study sites, I collected the high est number of responses (37%) f r o m the Ocala National Forest and the least (5%) from the Green Swamp Wildlife Management Area (Figure 2 2). The study questionnaire and indices included five sections Section A asked respondents about their current and past hiking experience S ection B included the Florida Trail Association S ection C asked about trip expenditur e s and substitutes
30 S ection D included items related to site attribute preference s and motivation s and section E included demographic questions. Variable Measurement and Analysis I assessed r or push factors with 23 items adapted from the Recreation Experience Preference index developed and tested by Driver and colleagues (Driver, 1983; Manfredo, Dr iver, & Tarrant, 1996) I assessed site attribute preferences or pull factors with an index of 14 items related to setting attributes that describe site characteristics and recreation opportunities. Response s for both push and pull items was a five poin t scalar format ranging from 1 ( n ot at all important) to 5 (very important). I used confirmatory factor analysis to identify the domains of push and pull factors and structural equation model ling to examine the relations hips between them in MPlus version 7.4 statistical software. Following Yoon and Uysal (2005) I specified and tested the fit of measurement models (facto rs or latent variables) of both push and pull factors before examining the structural relations and the fit of a full structural equation model. A latent variable like intrinsic motivation refers to an unobserved variable that can be measured through the m ultiple observable variables (Byrne, 2001) (Cronbach & Shavelson, 2004) to assess measurement reliability for each factor. Tests of reliability examine the internal consistency among responses to the variables and provides an indication of whether the items measure a single construct (Vaske 2008) The item total correlation refers to the correlation between a n individual
31 item score across respondents and the summative score for all items across respondents A higher item total correlation indicates higher internal consistency and a value close to zero indicates poor internal consistency (Pett, Lackey, & Sullivan, 2003) measure of a respective domain (Leong & Austin, 2006; Vaske, 2008) I dropped any item that either (a) had item total correlation of <0.4 for th e respective domain or (b) delating that item would significantly coefficient. F or each measurement model, I deleted items having less than 0.4 factor loadings and re r a n the model (Vaske, 2008; Yo on & Uysal, 2005) A factor loading refers to a correlation between indicator (item) and the factor (latent variable). I repeated the process of del e ting weak indicators and re running the model until a meaningful as well as statistically acceptable model fit was achieved. In this process, I dropped eight motivation items and three site attribute items (Table 2 1 and Table 2 2). I used a ratio of Chi Root mean Square Error of Approxima tion (RMSEA), and Standardized Root Mean Square Residual (SRMR) to examine whether each measurement and full structural model had good fit with the data. Value s a model for a given data set (Brown, 2006; Hu & Bentler, 1999; Kline, 2005) Once the reliable indicators were identified for both push and pull factors, I used cluster analysis to identify the market segments of respondents based on their means are the two most commonly us ed clustering algorithms in the tourism and recreation marketing research.
32 However, there is no clear indication in the literature regarding a standard to choose one method over an other (Paudyal et al., 2015) Thus, I compared the result Pe a rson correlation as the is based on a least square criterion, which minimizes the within cluster sum of squares and thus maximizes the wi thin cluster homogeneity (Mooi & Sarstedt, 2010) .The intuitive interpretat ion of results from this method provided the most distinct and meaningful market segments (Kotler, 1994; Paudyal et al., 2015) I compared the market segments u sing Chi square and ANOVA for demographics and trip characteristics in SPSS 24.0 statistical software. I used eta 2 ) to assess effect size on ANOVA test. Effect size should be interpreted as small ( 2 = 0 .01), medium ( 2 = 0 .06), and large ( 2 = 0 .14) as suggested by Cohen (1988) Results Descriptive Characteristics of Sample The average age of the sample p articipant was 43 (S td. D ev. = 14) years old 63% were male, and the majority were White (94%). About two thirds of the respondents (64%) were at least a college graduate, whereas about 10% graduated high school or less The annual household income was les s than $30,000 for about 20% of the respondents, $ 30,000 60,000 for one third of the respondents, and $90,000 or more for one quarter of the respondents. Respondents reported that they drove 55 miles (SD=58.5 miles), on average to reach the recreation site during the trip during which they were interviewed About half of the respondents (48%) were first time visitors. Among
33 the return visitors, about half (54%) had made no or limited (1 6 times) use of the site in the past yea r, whereas about 30% had visited the site more than 20 times in the past 12 months. The m ajority of the respondents (77%) spent one hour to half a day on the trail, whereas about 20% spent more than one day. Likewise, about 60% of the respondents hiked fiv e miles or less, whereas one quarter of the respondents hiked more than ten miles on the trail in the course of the trip during which they were surveyed. Motivations and Site Attribute Preferences A confirmatory factor analysis yielded five domains of moti vations o r push factors : (1) Achievement, (2) N ature and S cenery, (3) F itness and R elaxation, (4) S ocial interaction, and (5) S olitude (Table 2 1 and Figure 2 3). Factor 1, A chievement included four items representing motivations related to using equipmen t, testing skills, sharing skills with other s and meeting new people. The factor loadings for this factor ranged from 0.63 0.74 (Figure 2 3). Factor 2 Nature and S cenery, included three items related to enjoying scenery, experiencing nature, and explor ing the area, with factor loadings ranging from 0.72 0.80. Factor 3 F itness and R elaxation, included three motivation items, getting exercise, getting away from the usual demands of life, and feeling healthier, with factor loadings ranging from 0.53 0.65 (Figure 2 3). Factor 4 S ocial I nteraction, included two motivation items related to being with a group with factor loadings of 0.61 and 0.75 (Figure 2 3). Finally Factor 5, S olitude included three motivation items, such as being away from people and experiencing solitude, with factor loadings ranging from 0.65 0.78 (Figure 2 3). The index mean score of the factor N ature and Scenery was highest ( =4.6 SD=0.6 ) suggesting that experiencing or enjoying nature or scenery or exploring the
34 area was the most important motivation to visit FNST. Fitness and R elaxation ( =4. 4, SD=0.7 ) was the second most important motivation to visit FNST. Respondents rated A chievement ( = 3.1, SD=1.1 ), such as using equipment, testing skills, sharing skills, or meeting new people, as the least important motivation to visit the trail (Table 2 1). A confirmatory factor analysis of site attribute preferences or pull factors yielded four do mains : (1) Quality Nature (2) Consumptive, (3) Convenience, and (4) Arts and C ulture (Table 2 2 and Figure 2 3). Factor 1, Quality Nature included four items related to the environment and nature recreation such as chance to see wildlife/birds and undis turbed nature, with factor loadings ranging from 0.59 0.67. F actor 2 Consumptive, included three site attribute items related to fishing and hunting with factor loadings ranging from 0.67 0.92. Similarly, the third pull factor, Convenience, included t hree items easy access, closeness, and parking, with factor loadings ranging from 0.51 0.71. Finally, Arts & C ulture, included two items, interesting small towns and local crafts or handiwork, with factor loadings of 0.64 and 0.81 respectively. Among t he site attributes, settings that offer the opportunity of experiencing quality nature such as a chance to see wildlife, good environmental quality of air, water, and soil, and wilderness and undisturbed nature, was the most preferred ( =4. 3, SD=0.7 ) site attribute (Table 2 2). On the other hand, the opportunity for c onsumptive recreation, such as fishing and hunting, was the least preferred ( = 1.8, SD=1.0) attribute.
35 Structural Relations among Push and Pull Factors The structural equation model (Fi gure 2 3) indicates that the push factor domain Achievement was positively associated with the two pull factor domains Consumptive (0.47, p<0.01) and Arts and C ulture (0.64, p<0.01). Likewise, the push factor domains Nature and S cenery (0.65, p<0.01) and S olitude (0.43, p<0.01) were positively associated with the pull factor domain Quality Nature Finally, the push factor domain Fitness and R elaxation was positively associated with the pull factor domain Convenience (0.46, p<0.1). On the other hand, the pus h factor domain Social I nteraction was not significantly associated with any of the pull factor domains. Cluster analysis generated four meaningful and distinct segments of respondents based on their motivations and site attribute preferences : (1) Nature F it, (2) Passive, (3) Social, and (4) Opportunist (Table 2 3). The cluster segments did not differ in terms of most demographic characteristics (Table 2 4) H owever, they did differ significantly in terms of recreation characteristics (Table 2 5). The chara cteristics of each cluster segment are described below. Nature F it : The first cluster included 41% of the respondents who were motivated to visit FNST mostly to experience nature and scenery, fitness and relaxation, skills development, and experiencing sol itude (Table 2 3). In comparison to other segments, this segment of respondents placed higher importance on recreation settings that offer the opportunities to experience quality nature and wildlife, for example to see wildlife/birds. This segment of respo ndents had a higher proportion of Whites (Table 2 4) than at least two other segments (Social and Opportunist). Respondents of this segment drove the longest distance to reach the destination, spent the most time, hiked
36 the farthest on the trail, and rated the highest visit experience relative to the other segments (Table 2 5). Passive : The second cluster segment included 28% of the respondents who had no clear motivation and site attribute preference to visit the FNST (Table 2 3). T hey stated low impo rtance for all motivation domains as well as site attribute preference domains. Almost all respondents (99%) in this group were White (Table 2 4) and the majority (57%) were first time visitors (Table 2 4). In addition, respondents in this group spent the least amount of time on the trail and they reported the lowest ratings of the visit experience compared to respondents in other segments (Table 2 4). Social : The third cluster segment included 24% of the respondents who rated the push factor social interac tion being with members of their group as important (Table 2 3). This group of respondents gave greater importance to site attributes related to convenience such as easy access, close to home, and parking when selecting a recreation site on the FNST (Tab le 2 3). Respondents from this segment had lower past use frequency than other segments except Passive. However, this group spent more time on the trail than respondents in other segments except for the N ature F it segment. Opportunist : The fourth cluster i ncluded 7% of the respondents They reported higher importance of almost all motivation domains as well as site attribute domains than respondents in other segments (Table 2 3). Particularly, this segment of respondents reported higher importance of recrea tion settings that offer consumptive recreation experience (e.g., hunting, fishing), convenience
37 (e.g., close to home, parking), and arts and culture (e.g., local crafts, small towns) than recreation users in other groups. T his group expressed higher impor tance of quality nature as well. In terms of motivation, this segment of respondents had higher motivation of achievement (e.g., use equipment and test skills) as well as social interaction (e.g. be with group). However, other motivations, such as experien cing nature/scenery and solitude, were also highly important. In comparison to the other three segments, this segment included the least proportion of Whites (Table 2 4). Likewise, compared to other segments, this segment of respondents drove the least dis tance (28 miles) to reach the destination, were mostly return users (74%) and had a higher proportion of individuals (72%) who visited the trail more than 20 times over the past year (Table 2 5). However, this segment of respondents hiked the least and spent less time on the trail than the other segments. Discussion The primary aim of this research was to identify motivations (push factors) and site attribute preferences (pull factors) of FNST visitors, identify their market segments, and e xa mine the structural relationships between p ush and pull factors. Finding of multiple dimensions of both push and pull factors was consistent with the existing body of literature which suggest that visitors have multiple motivations and setting preferences and there are significant relationships between visitors desired motivations and setting preferences (Kim et al., 2007; Kozak, 2002; Stein & Lee, 1995; Uysal & Hagan, 1993) T he visitor segmentation conducted in this study describes distinctive market segments of visitors that have meaningful and different characteristics in terms of recreation experiences, preferences, and behaviors. The findings of v arious structural
38 relations among motivations and site attribute preferences suggest that recreation settings should be managed in terms of how visitors prefer to experience or benefit from a site rather than the activity they choose (Driver, 2008; Stein & Lee, 1995) I connected to their desired benefits or motivations, and both motivation s and site attribute preference s are multi dimensional concept s (Kim et al., 2007; Kozak, 2002; Stein & Lee, 1995; Uysal & Hagan, 1993) However, m ost past research focused on motivations or site attribute preferences of visitors to wildland settings areas relati vely far from human development such as wilderness areas, national parks, or other recreation sites in remote are as of the western United States. Thus, there is limited knowledge regarding the specific relationships among the push and pull fact ors from visitors to forests and recreation areas that are close to urban settlements. This research contributes to the body of knowledge by further explaining these relationships in the push/pull theory in the context of recreation settings that offer clo se access to nature for millions of people. Management Implications Offering preferred experience opportunities in preferred recreation setting s s for diverse visitors (e.g., personal, soc ial, and economic benefit) and minimiz ing unwanted outcome s (e.g., conflicts associated with recreation and wildlife management). The findings presented in this study are useful for the FTA and partner agencies in managing existing resources in the FNST to satisfy
39 expectations of diverse recreation users. D istinct market segments of visitors and s tructural relations among various domains of motivations and site attribute preferences indicate the importance of manipulating recreation settings keeping in mind specific motivations and preferences of target visitors. This study found that n ature focused and fitness related experiences were the major push factors and easily accessible and less developed areas were the most important pull factors. Although the FNST will never achieve the high degree of pristine and primitive conditions of other national scenic trails (e.g., Pacific Crest Trail and Continental Divide Trail), it does provide easy access to nature for millions of Florida residents and tourists The FNST is intentionally planned to traverse through various natural settings through a fairly urbanized state. It provides all users a convenient and easily accessible place to experience nature and attain fitness and relaxation and this study shows i t is serving those user groups. Past research often does not link primitive nature to experiences like fitness, but this research shows that natural surroundings are an important component of the diverse opportunities natural areas close to urban settings can provide. Since the FNST has a unique role of providing access to nature, further marketing and promotion efforts should be continued to accommodate the diverse audiences who do use and could potentially use the FNST while continuing to maintain the ec ological health of the areas. Finding s show that most visitors highly value opportunities to experience scenic natural landscapes undisturbed and tranquil environment s wilderness areas, and native wildlife and bird habitat s Thus, managers must ensure th ese opportunities are enhanced and sustained In an urbanized state,
40 like Florida, high levels of unmanaged recreation use, encroachment of development, invasive species, and many other human induced changes threaten the qualities of natural environments t hat make the FNST valuable to existing and potentially new users. This study found four market segments of FNST visitors who had various motivations, site attribute preferences, and recreation behaviors. Marketing refers to providing a match between visito visitors can attain their desired outcome. Thus, managers should continue their efforts of marketing various sections of FNST so that current users continue to use the trail and visit other parts of the trail. Marketing efforts based on these segments could also be useful to attract new users, who are perhaps unaware about the existence of this trail, but have similar recreation motivations as the current users have. Among the market segments, the Nature fit se ems to be the most lucrative group because this segment included the largest proportion of visitors, they travelled the longest distance from their residence to the destination, spent more time and hiked farther on the trail, and were most satisfied with t heir experience Satisfied visitors are more likely to revisit the place (Kil et al., 2012) thus establishing a loyal client base which will attract more visitors (Kruger, Viljoen, & Saayman, 2017) Thus, Nature fit segment of visitors is important to the USFS and FTA, as they have the greatest potential to directly benefit the local econom ies along the FNST As the structural analysis shows, visitors wh o would like to enjoy nature/scenery, explore natural area s and experience solitude are
41 likely to prefer large primitive or wilder ness areas and wildlife habitats. Although these are rare in Florida, they do exist, and this study shows users are motivated to visit these areas. Thus, managing and promoting specific sections of the FNST that offer healthy environmental conditions are less disturbed, have more wilderness areas, and provide opportunit ies to see a variety of wildlife and birds will be effecti ve to retain these lucrative visitors and attract new visitors having the similar motivations of experiencing fitness and relaxation in natural environment s The characteristics of the P assive visitor segment was consistent with past research that identif ied a conceptually similar visitor segment (Mehmetoglu, 2007; Palacio, 1997; Park & Yoon, 2009; Smith et al., 2014) This segment was second largest in terms of size and they drove a long distance to the destination like the N ature fit segment did, thus indicating higher potential econ omic benefit to local economies However, the visitors of this segment had no clear motivation and site attribute preference to visit the FNST, they spent the least amount of time on the trail, and rated their experience lower than the other segments Thus further research is needed, preferably qualitative, to better understand what brings these visitors to the FNST, what experience s motivate them to take part in nature based recreation and what attractions will help them to attain their expected experien ce s Marketing programs should approach this kind of current and potential visitor with clear information of all the opportunities each section of the trail provides, so that such users can find the site and opportunity of their interest. The users in the Social segment valued social interaction such as being with members of their group as the most desired outcome and they preferred convenience
42 such as easy access, close to home, and parking when selecting a recreation site on the FNST Thus, developing large group campgrounds or picnic areas on or near the FNST close to roads and settlements and with plenty of parking space will be effective to satisfy current and potential users who prefer to visit in large groups a nd gain more social experience. The Opportunist segment, which was similar to the Want it all segment found by Park and Yoon (2009) had a wide range of motivations and site attribute preferences. However, unlike Park and Yoon (2009) who found 25% of such visitors in a rural tourism village in Korea, this study found only 7% of such vi sitors on the FNST M arketing specific sections of the FNST and associ ated managed areas that offer a wide range of recreation opportunities to people living nearby from the natural areas could attract such users who have multiple motivations. Users in both the Opportunist segment and Nature fit segment placed higher importa nce on achievement. Such users w ho are motivated to use equipment, test skill s and share knowledge, and meet new people, are likely to choose sections of the FNST that offer good fishing/hunting opportunities like Ocala NF and Green Swamp WMA or areas that have interesting small towns or areas where local crafts and handiworks are popular. Thus, promoting such areas along or near the trail in various sections of the FNST could be beneficial to attract new users having similar m otivations for recreation. Multiple dimensions of both push and pull factors and different preferences of different market segments indicate that a one size fits a ll
43 approach in promotion and marketing w ill not be effective to all kind of users Thus, mark eting messages should be specific to capture needs and desires of both current and potential users. While doing so, FNST managers should keep in mind the various user typolog ies ( i.e., and their resp ective motivations and preferences Limitations and Future Directions of Research The push and pull items considered in the instrument of this study do not include all possible dimensions of motivations and site characteristics. A master list of recreation experience preference developed by Driver (1983) includes 300 plus items that represent 39 domains of motivations. However, it was not practical to include all possible domains of motivations in our survey instrument. Keeping this i n mind, this study focused on 23 items representing six domains of motivations that were most commonly found in the literature in this field of study. Likewise, 14 items were included to represent various site attributes, which do not represent all the pos sible attributes of site attraction preferences. Also, motivation is only one of many variables which could (Baloglu & Uysal, 1996) Thus, research (Baloglu & Uysal, 1996) or perceived broader positive outcome (e.g., social, economic) to the community (Stein & Lee, 1995) will be more beneficial for agencies in marketing various sections of FNST. There could be other pull factors, important to specific group of visitors, which this study did not include. Futu re research can use the model developed in this study to further expand our understanding about motivations and site attribute preferences by exploring additional attributes that could attract visitors to FNST.
44 Although recreation destinations often offer multi products, they could also offer different experience opportunities during different seasons (e.g., good hiking during winter and good springs experience during summer). In fact, research that compares winter visitors and other season visitors could g ive better insight about the relations between motivations and site attribute preferences. These differences might be less significant in Florida than in regions where there are stark differences between winter and non winter seasons, but differences likel y exist and should be examined. Galloway (2012) compared motivation and site preference among the recreation users participating in different activities (e.g., wh itewater kayaking, multisport racing, and fishing) and found significant difference s Thus, comparing the structural relations between motivations and site attribute preferences across various recreation activity users will be an additional contribution to the body of knowledge. Research has found the same but different sizes of market segments in two different times (Fix & Taylor, 2011) motivations and site preferences could change over time. Comparing motivations and site preferences, their structural relations, and cluster segments over different time periods could offer more robust knowled ge on this topic. The visitor segments identified here offers a representative profile of recreation users of the FNST, which provides insights for developing marketing programs. However, the sample may not represent the visitors to all sections of the FN ST as only visitors to Central and North central sections were included. Thus, future research should consider taking representative samples from all
45 sections of FNST, which will offer a experiences and s ite preferences. In conclusion, consistent with the previous research, different destination attributes are found to be preferred by users with various motivations to participate in outdoor recreation. However, t his research adds to the body of knowledge by context of urban parks and recreation areas Relationships between motivations and site attribute preferences are different than that found in previous research from rural contexts. For instance, research conducted in rural contexts have found that recreation users motivated for achievement would want primitive and hard to access areas and adventurous activities. However, this study found users motivated for achieveme nt such as testing and sharing knowledge and skills and meeting new people, are likely to choose destinations that offer fishing or hunting opportunities or towns that are small and culturally interesting or are popular for handicrafts Findings presented in this study are unique and could be applicable to urbanized regions where nature based recreation opportunities are within a short distance to millions of people. FNST managers should continue their efforts to enhance and sustain natural environments al ong the trail while maintaining easy access and parking spaces to attract users having various motivations and preferences. As shown in this research, a segmentation of visitors based on motivations and site attribute preferences serves as a useful marketi ng tool as it offers information regarding specific markets to focus in relation to available site attractions. Agencies should maintain and promote FNST keeping in mind various typology of
46 users, as shown in this research, in terms of their motivations, s ite attribute preferences, and recreation behaviors.
47 Table 2 1 Index mean and m easurement reliability of motivations domains of users of the Florida National Scenic Trail Motivation domains and items (push factors) Index mean (SD) Reliability Achievement 3.1 (1.1) 0.79 To use my own equipment To test my skills and abilities To share my skills and knowledge with others To meet new people Nature & Scenery 4.6 (0.6) 0.79 To enjoy the scenery To experience nature To explore the area Fitness & Relaxation 4.4 (0.7) 0.69 To get exercise To get away from usual demands of life To feel healthier Social Interaction 3.7 (1.1) 0.67 To be with members of my group To be with people who enjoy the same things I do Solitude 3.8 (1.0) 0.76 To be on my own To be away from people To experience solitude Values in parenthesis indicate standard deviation. The following items were dropped from the measurement model of push factors because of f actor loading, or item total correlation <0.4. To learn about natural history of the area To learn more about the nature To experience new and different things To have thrills and excitement To be close to nature To enjoy the smells and sounds of nature To do something with my family To relax physically
48 Table 2 2 Index mean and m easurement reliability of site attribute preferences ( pull factor ) domains of users of the Florida National Scenic Trail Site attribute domains and items (pull factor) Index mean (SD) Reliability Quality Nature 4.3 (0.7) 0.67 Chance to see wildlife/birds Good environmental quality of air, water, and soil Wilderness and undisturbed nature Consumptive 1.8 (1.0) 0.84 Good fishing Good big game hunting Good small game hunting Convenience 3.4 (1.0) 0.65 Easy access to the area/being easy to get to Close to home Available parking Arts & Culture 2.4 (1.1) 0.69 Interesting small towns Local crafts or handiwork Values in parenthesis indicate standard deviation. The following items had factor loading or item total correlation <0.4. Thus, were dropped from the measurement model of pull factors. To see the natural water features Good camping Historical, military, or archeological sites
49 Table 2 3 C omparison of cluster segments by motivations and site attribute preferences of users of the Florida National Scenic Trail Domains Index mean Cluster segments post hoc A Nature fit (41%) B Passive (28%) C Social (24%) D Opportunist (7%) Effect size ( 2 ) Motivations (Push F actors ) Achievement 3.1 (1.1) 3.8 (0.8) 2.2 (0.9) 2.8 (0.9) 3.8 (0.9) A,D>C>B 0.41 Nature & scenery 4.6 (0.5) 4.8 (0.4) 4.6 (0.6) 4.5 (0.5) 4.6 (0.6) A>B,C,D 0.05 Fitness & relaxation 4.4 (0.7) 4.7 (0.5) 4.3 (0.8) 4.3 (0.6) 4.4 (0.7) A> B,C,D 0.06 Social interaction 3.7 (1.1) 3.8 (1.0) 2.8 (1.1) 4.2 (0.7) 4.1 (0.8) C,D>A>B 0.23 Solitude 3.8 (1.0) 4.3 (0.6) 3.5 (1.1) 3.0 (0.8) 4.1 (0.9) A,D>B>C 0.30 Site Attribute Preferences (Pull F actors ) Quality Nature 4.3 (0.7) 4.5 (0.7) 4.2 (0.6) 3.9 (0.7) 4.3 (0.8) A,B,D>C 0.07 Consumptive 1.8 (1.0) 1.6 (0.8) 1.6 (0.9) 1.7 (0.8) 4.1 (0.8) D>A,B,C 0.34 Convenience 3.4 (0.9) 3.3 (0.9) 3.1 (1.0) 3.6 (0.8) 4.0 (0.9) D>C>A,B 0.06 Arts & Culture 2.4 (1.1) 2.3 (1.0) 2.4 (1.0) 2.1 (0.9) 4.2 (0.8) D>A,B,C 0.18 Values in parenthesis indicate standard deviation. Effect size should be interpreted as small ( 2 = 0 .01), medium ( 2 = 0 .06), and large ( 2 = 0 .14).
50 Table 2 4 Comparison of cluster segments by demographic characteristics of users of the Florida National Scenic Trail Demographics Sample average Cluster segments Sig. Nature fit Passive Social Opportunist Age (years) 43.1 (14.5) 44.0 (13.9) 44.8 (15.7) 39.4 (13.6) 43.8 (15.5) Gender (%) n.s. Male 63.4 62.4 60.8 68.1 63.2 Female 36.6 37.6 39.2 31.9 36.8 Race/Ethnicity (%) Hispanic 3.2 4.3 5.8 White 94.1 96.4 98.7 87.7 82.4 *** African American 3.0 1.8 1.3 6.2 5.9 n.s. Other # 4.1 1.8 2.6 7.7 11.8 ** Education (%) n.s. High school or below 8.2 9.5 5.2 8.7 10.5 Some college 28.1 31.0 18.2 30.4 42.1 College graduate 37.4 35.3 40.3 39.1 31.6 Some graduate school or above 26.3 24.1 36.4 21.7 15.8 Income (%) n.s. Less than 30,000 18.4 17.1 14.3 26.1 15.8 30,000 59,999 33.0 34.2 31.2 33.3 31.6 60,000 89,999 24.1 23.9 24.7 21.7 31.6 90,000 or more 24.5 24.8 29.9 18.8 21.1 #Other: American Indian, Native Hawaiian, and Asian American; Sig. = significance. ***Significant at 1% level; **Significant at 5% level; n.s. = not significant
51 Table 2 5 Comparison of cluster segments by recreation characteristics of users of the Florida National Scenic Trail Recreation characteristics Sample average Cluster segments Sig. Nature fit Passive Social Opportunist Distance drove (miles) $ 54.9 (58.5) 63.9 (60.8) 61.1 (63.1) 42.3 (52.5) 27.5 (27.4) ** Visit experience # 8.6 (1.4) 8.9 (1.1) 8.3 (1.6) 8.6 (1.3) 8.8 (1.6) Past trail use (%) *** First time visitors 47.9 52.1 57.0 36.2 26.3 Return visitors 52.1 47.9 43.0 63.8 73.7 Past year use (% of return visitors) None 11.7 10.7 11.8 14.6 7.1 Low (1 6) 42.8 42.9 47.1 48.8 14.3 Medium (7 20) 15.9 21.4 17.6 9.8 7.1 High (>20) 29.7 25 23.5 26.8 71.4 Time spent (%) *** 1 hour or less 28.7 24.1 38 19.1 52.6 Few hour half a day 46.8 37.1 51.9 61.8 31.6 One whole day 5.7 9.5 1.3 2.9 10.5 More than one day 18.8 29.3 8.9 16.2 5.3 Miles Hiked (%) **. Less than a mile 9.6 8.6 8.9 10.4 15.8 1 5 miles 49.8 43.1 58.2 50.7 52.6 5 10 miles 15.7 14.7 16.5 19.4 5.3 > 10 miles 24.9 33.6 16.5 19.4 26.3 $ Out of state visitors (9.8%) not included; #visit experience measurement scale ranged from 1 to 10, with 10 being best experience; Sig. = significance ***Significant at 1% level; **Significant at 5% level; *Significant at 10% level.
52 Figure 2 1 Study area and survey sites in Florida, USA Figure 2 2 Proportion of respondents by survey sites in Florida, USA 0 5 10 15 20 25 30 35 40 Ocala NF Goldhead SP Cross Florida Greenway Little Big Econ SF Aucilla WMA Osceola NF Green Swamp WMA % of respondents
53 Figure 2 3 Measurement models and structural relations among motivation s (push factors) and site attribute preference s (pull factors) of users of the Florida National Scenic Trail Note: Bold lines indicate statistically significant structural relations. ***Significant at 1% level; **Significant at 5% level; *Significant at 10% level.
54 CHAPTER 3 HOW DO ACTIVITY AND KNOWLEDGE OF PRESCRIBED BURNING AFFECT D RECREATION SATISFACTION FROM RED COCKADED WOODPECKER HABITAT? B ackground The red cockaded woodpecker ( Picoides borealis ) is a federally designated endangered keystone species of bird in the United States. This bird has a high ecological value as at leas t 27 species of vertebrates have been documented to use red cockaded woodpecker ( RCW ) cavities either for roosting or nesting (US Fish and Wildlife Service, 2015) Even though the species is protected under the Endangered Species Act of 1973, managers face challenges to increasing its abundance to levels high enough for recovery to occur. The primary threat to RCW is lack of suitable habitat. Research has shown that prescribed burning and tree harvesting play important roles in maintaining suitable habitat condition for RCW (Ramirez & Ober, 2014) Prescribed burning and tree harvesting treatments are effective to maintain ecologically healthy ecosystems and habitat for many wildlife species including RCW However, researchers have also found tha t people do not always perceive prescribed burning as aesthetically pleasing and acceptable (Gobster, 1999; Ribe, 1999; Ryan, 2005; USDA Forest Service, 1995; Williams & Cary, 2002) Human dimensions literature in various other cont exts (e.g., climate change) indicates that knowledge is not to know whether knowledge about the ecological importance of prescribed burning would be likely to affect vi from forests that require burning and other management (e.g., thinning) to maintain habitat for wildlife (e.g., RCW).
55 of fire and fire management in the southern United States are significantly fewer (12%) in comparison to the western United States (52%) (Toman, Stidham, McCaffrey, & Shindler, 2013) qual ity and recreation experience, is limited. Knowledge about whether the type of recreational limited; for example, whether hunters perceive prescribed burning differently than hikers. perception of scenic beauty and recreation satisfaction from forest sites maintained as RCW habitat in Florida. This research answers the following questions. 1. Do the activities peop le choose to participate in relate to their knowledge of prescribed burning and perception of scenic beauty and recreation satisfaction from forests sites requiring frequent burning? 2. Does providing ecological information about prescribed burning improve re creation requiring frequent burning? 3. scenic beauty and recreation satisfaction from forest s ites requiring frequent burning? Conceptual Framework Past studies have suggested that public perceptions of forest and landscape aesthetics involve complex relationships between affective perception of visual stimuli and cognitive perception of approval (Ribe, 2002) However, there is not yet a well developed consensus among psychologists regarding activation of human affective perceptions and cognitive perceptions. Some psychol ogists have argued that affective reactions are pre cognitive and they occur without extensive cognitive processing
56 (Zajonc, 1980) Based on this hypothesis, people make affective perception, such as scenic beauty, sooner and with greater confid ence than cognitive judgments, such as acceptability of management treatment or realization of recreation experience (Zajonc, 1980) On the other hand, some scholars have argued that affect can be both pre and post cognitive (Lerner & Keltner, 2000) Thus, when there is absence of conditioning knowledge or information, affective perception of visual stimuli, such as scenic beauty, may not necessar ily be different from more cognitive perception, such as acceptability of landscape management (Ribe, 2002; Zajonc, 1980) When perceptions of acceptability are only partly affective but mainly cognitive, these two perceptions may differ significantly (Ribe, 2002) Using this theoretical framework, this study examines perception of aesthetic and recreation quality of forest sites representing various times after burning. Scenic Beauty and Recreation Satisfaction component of the environment. However, Daniel and Boster (1976) and Daniel and Vining (1983) aning associated with (2001, p. 270) Joh nson, Brunson, and Kimura (1994) operationalized these terms as two conceptually separate constructs. Authors suggested that the relationship between these two constructs are ings in
57 In recreation context, expectancy valence theory suggests that people p articipate in outdoor recreation because of their hope to fulfil preferred needs, motivation s or expected experience s (Manning, 1999; Moore & Driver, 2005) The expectancy valence theory is an outcome of integrating two major concepts expectancy and valence. Expectancy is the likelihoo d that an action will lead to a certain outcome or goal whereas valence is the attractiveness of, or anticipated satisfaction with, an outcome (Lawler, 1973) choice of recreation activity and site (Andereck et al., 2011) In this theoretical perspective, the degree of congruence between aspirations and the perceived reality of experiences is defined as recreation satisfaction (Bultena & Klessig, 1969, p. 349) In this research context, I operationalize recreation satisfaction as the congruence between desired experience and perceived attainment of experience associated with appearance of recreati on sites (Graefe & Burns, 2013) Although scenic beauty of a forest site could always contribute to recreation satisfaction and desire to visit an a rea (Sevenant & Antrop, 2009) scenic beauty may not be the major factor of importance in every case. For example, Tahvanainen, Tyrvinen, Ihalainen, Vuorela, and Kolehmainen (2001) examined perceived scenic beauty and recreation preference of five forest management practices: small clear cutting, thinning, removal of undergrowth, natural state, and traditionally managed cultural landscape. Results s howed no clear association between scenic beauty rating and recreation preference. Likewise, Brunson and Shelby (1992) examined scenic
58 beauty and acceptability of various forest management practices and found that old growth forest setting rated highest an d clear cut rated lowest in terms of both scenic beauty and acceptability for recreation. However, acceptability rating of each forest management practice was higher when the setting was considered a place to hike than when it was considered as a place to camp (Brunson & Shelby, 1992) Thus, it is possible that different recreation users would perceive scenic beauty and recreation value of forest management practices in wildlife habitats di fferently. Studies have shown that the physical characteristics of forest attributes such as tree density and forest type account for a significant proportion of the variance in public perception of scenic beauty (Dan iel, 2001) However, different cultural and social traits of the observers have also been found to affect evaluation of scenic beauty and recreation value of forests (Buhyoff & Leuschner, 1978; Daniel & Boster, 19 76) Based on extensive review of the literature, Zube, Sell, and Taylor (1982) identified four dominant paradigms in landscape perception: the expert parad i gm, the psycho physical paradigm, the cognitive paradigm, and the experiential paradigm. In the expert paradigm of assessing landscape aesthetics, forest and landscape experts assess visual quality of landscape features in terms of color, form, line, and t exture and classify forest and landscape into high, medium, and low visual quality categories (USDA For est Service, 1995) This expert based approach of landscape aesthetics was integrated into the Scenery Management System (SMS), a forest management framework developed by USDA Forest Service. The SMS offered an integration of forest and landscape aesthe tics with other biological, physical, and social/cultural resources in the planning processes. Expert based assessment of
59 landscape aesthetics, however, was heavily criticized for lacking adequate levels of precision, reliability, and validity (Daniel & Vining, 1983) One of the major drawbacks associated with this approach was that the public were not involved in the asse ssment. The psychophysical, cognitive, and experiential paradigms assume that scenic beauty is an output of interactions between the physical appearance of the forest and in volve the untrained public in the assessment of scenic beauty and acceptance of management practices. These paradigms are found to achieve greater reliability and validity than the expert paradigm (Brown & Daniel, 1987; Daniel, 2001) The psychophysical approach to landscape scenic beauty evaluation emphasizes the use of mathematical modeling (scenic beauty estimation [SBE] model) to examine relationships between objectively quantifiable site attributes and subjective ratings of the sit e (Daniel & Boster, 1976) One of the strengths of this approach is that it can relate changes in manipulative site characteristics with their impacts on scenic beauty (Ruddell, Gramann, Rudis, & Westphal, 1989) However, a major criticism of psychophysical models is their limited ability to produce theoretical explanations of predictor variables. In other words, the psychophysical models place emphasis on prediction, with little focus on psychological processes that mediate the relationship between physical attributes of a particular site and its ratings regarding scenic beauty and recreation preferences (Ruddell et al., 1989) Also, the SBE model standardizes scenic beauty ratings of individuals to adjust for potential differences in judgement criteria among the observers (Daniel & Boster, 1976) Therefore, it does not account for the potential variatio n in scenic beauty perception across publics of different socio
60 cultural, cognitive, and experiential characteristics. However, researchers have suggested that forest and landscape preference is a function of both the scene characteristics (forest attribut es and silvicultural treatments) and human characteristics (Edwards et al., 2012; Kearney & Bradley, 2011) Thus, it is important to examine the variation in perceived scenic beauty and recreation preferences across recreationists of different experiential, cognitive, and socio cultural characteristics. The cognitive paradigm involves assessment of forest scenic beauty by human observers takin g the socio cultural factors, human meanings associated with the forest, experience, and expectations into account (Hull & Reveli, 1989; Zube et al. 1982) Likewise, the experiential paradigm involves assessment of forest scenic beauty through examination of human forest interactions (Zube et al., 1982) The concept of human i nteraction with nature refers to what people put into and take from nature. Thus, it is important to have an idea of what people will be doing in their interactions with a landscape to understand the human nature interactions involved. For example, whether visitors hike or view the scenery will significantly affect the nature person interaction (Zube et al., 1982) Thus, researchers have used cognitive and experiential paradigms to examine which forest attributes are preferred more than others with regard to perceived scenic beauty and recreation preference, but also how socio cultural and experiential factors are a ssociated with these perceptions. Socio cultural and Experiential Factors Hull and Reveli (1989) argue that a person's immediate purpose in a n environment influences the type of information sought and the criteria used to evaluate that environment. They suggest that landscape preferences are learned and depend
61 upon a person's culture, past experiences, and current purposes. Based on this concep t, various researchers have examined the role of socio cultural and experiential factors on perception of scenic beauty and recreation preferences. However, there is inconsistency in the literature regarding the effect of experiential and socio cultural fa ctors on the perception of scenic beauty and recreation value of forest sites. Some researchers have found similarities in scenic beauty evaluations of landscapes among persons of different ages, genders, professions, socio economic status, and cultures (Daniel & Boster, 1976; Daniel & Vining, 1983; Frank, Frst, Koschke, Witt, & Makeschin, 2013; Kearney, 2001; Kearney & Bradley, 2011) However, other studies have found significant variations in perception/attitude about scenic beauty and recreation value across observers of different age groups, genders, and residential settings. For example, Tahvanainen et al. (20 01) found that middle aged and older respondents perceived clearcutting as increasing scenic beauty, whereas younger respondents thought clearcutting would decrease scenic beauty. In the same study, rural residents thought that clearcutting would have n egative impact on scenic beauty while urban residents thought clearcutting would improve scenic beauty (Tahvanainen et al., 2001) Similarly, in a national survey of public perception and acceptance of prescribed fire, Lim, Bowker, Johnson, and Cordell (2009) found that African American and Hispanics were significantly more concerned about the impact of prescribed fire on aesthetics and harm to wildlife than Caucasians. Likewise, Fix, Carroll, and Harrington (2013) found that visitors with a greater number of previous visits to a site were more likely to choose to recreate in areas that fit more primitive recreation opportunity spectrum (ROS) classes than visitors with fewer previous visits. This implies that
62 frequent visitors could avoid sites that exhibit a higher proportion of visible human disturbances like prescribed burning. However, it is also possible that visitors who visit a site more frequently are more aware of forest management objectives for that site and have dif ferent standards of acceptance than less frequent visitors. However, none of the previous studies on forest aesthetic perception focused on management practices specific to habitat management for endangered wildlife species like the RCW. Thus, it is import ant to know how the recreation activity of participants and other socio cultural and demographic traits are related to perception of scenic beauty and recreation value of forest sites maintained as RCW habitat. Knowledge and Information Studies have found that Florida residents generally exhibit varying levels of knowledge and support toward prescribed burning (Agrawal & Monroe, 2006; Jacobson, Monroe, & Marynowski, 2001; Monroe, Nelson, & Payton, 2006; T hapa, Holland, & Absher, 2008; Wan, Stein, Paudyal, & Colonna, 2014) For instance, Jacobson et al. (2001) assess the knowledge of fire among rural and suburban Florida residents living in counties that experienced severe wildfire events. They found that over two thirds of respondents accurately answered six technical questions abo ut fire. Over 79 percent of the respondents knew that fire helps improve forest health and 67 percent understood that fire is useful in creating wildlife habitat (Jacobson et al., 2001) In contrast, Thapa et al. (2008) found that two thirds of visitors to natural areas in Florida knew what a prescribed burn is, but only 25% thought that natural areas in Florida should be burned peri odically. Similarly, Wan, Stein, et al. (2014) conducted focus group discu ssions to assess knowledge of prescribed burning among hunter residents and non hunter residents in Northeastern Florida. Both groups of residents had some knowledge about
63 prescribed burning. However, hunter residents demonstrated more sophisticated knowle dge about prescribed burning, such as benefits to specific vegetation and wildlife species, whereas non hunter residents expressed more general knowledge, such as benefits to the ecosystem and reduced risk of wildfire (Wan, Stein, et al., 2014) Despite knowledge and support for prescribed burning, studies have also found that Florida residen ts express high concern about risk of fire to wildlife (DeLorme, Hagen, & Stout, 2005) A study conducte d in Texas has shown that people who believed that prescribed fire creates wildlife habitat and other positive outcomes were more supportive of prescribed burning than others (Rideout, Oswald, & Legg, 2003) This suggests that incomplete knowledge on the part of people who support wildlife management may lead them not to support prescribed burning in public lands because of the perceived negative impacts of fire on wildlife. The well known ecologist Aldo Leopold explained the concept of forest aesthetics based on two perspectives, the scenic perspective and the ecological perspective. Although perception of scenic beauty is based on visual appearance alone, an ecological perspective requires acquired knowledge of the dynamic qualities of forest environments and appreciation of both subtle and dramatic changes during forest lifecycles (Gobster, 1994) This perspective emphasizes the importance of ecological information on the perception of scenic beauty and recreation value or acceptability of a given forest or recreation setting. Som e studies have found a positive effect of ecological information on scenic beauty perception and attitudes toward forest management practices (Anderson, 1981; Buhyoff, Leuschner, & Wellman, 1979; Kearney, 2001; Ribe, 1999) For ex ample,
64 Anderson (1981) found that labels of land use designation significantly affected scenic beauty ratings of forest landscapes. Particularly, observers rated areas labelled received lower ratings of scenic beauty. Likewise, when Kearney (2001) showed pictures of clear cuts from Midwestern forests and included information about the benefits of birch and jack pine regeneration, she found increased acceptance and visual quality ratings. In contra st to the above findings, some researchers have ar gued that humans are resistant to changing aesthetic preferences (Daniel, 2001) and that ecological information only improves the acceptance for forest and ecosystem management, not the perception of scenic beauty (Ribe, 1999) Thus, understanding how recreationists perceive scenic beauty and recreation value of forests managed for e ndangered wildlife species with and without ecological information could expand the body of knowledge on the perception of scenic beauty and recreation experience on lands managed for wildlife species. Research examining the perception of scenic beauty and recreation value from the perspective of recreation users participating in different activities would expand our theoretical knowledge regarding the relationship between aesthetic preference and acceptance behaviors. Such information would also improve th e ability of forest managers to accommodate diverse recreation preferences in managed forests. Specifically, understanding how different recreationists perceive scenic beauty and recreation value of forest management practices intended to maintain habitats for RCW
65 and improve the management of lands for multiple objectives. Past studies examined rt for prescribed burning and found that people with greater knowledge of prescribed burning are more likely to support prescribed burning. However, research is limited regarding ning on recreation satisfaction, and whether and how knowledge of prescribed burning is associated with perception of forest aesthetics and recreation satisfaction. This study follows cognitive and experiential paradigm of landscape aesthetic perception, b orrowing methodological techniques from the psychophysical paradigm in order to examine how the recreation activity people chose affects their perceptions of scenic beauty and recreation satisfaction using RCW habitats as an example. In addition, this stud y examines how knowledge of prescribed burning varies among participants in different recreational activities and how this knowledge and acquired ecological satisfaction from forest sites maintained as habitat for RCW. Based on the review of literature, I developed and tested four hypotheses. Hypothesis 1 (H1): Recreation users who participate in different nature based recreation activities have different levels of knowledge ab out prescribed burning. Hypothesis 2 (H 2 ): Recreation users who participate in different nature based recreation activities have different levels of perceived scenic beauty and recreation satisfaction Hypothesis 3 (H 3 ): Providing ecological information about RCW and its habitat and recreation satisfaction. Hypothesis 4 (H 4 ): Knowledge of prescribed burning affects perceptions of scenic beauty and recreation satisfaction. Pe rceived scenic beauty significantly affects recreation satisfaction and mediates the structural relationship between knowledge and recreation satisfaction.
66 Methods Research Design and Survey Instrument The target population of interest in this study includ ed outdoor recreation users of managed natural areas in Florida. The accessible or study population was adult outdoor recreation users to Ocala National Forest I conducted onsite data collection with recreationists, 18 years or older encountered at trailh eads or in parking lots of various recreation areas in Ocala National Forest (ONF) during the period April 2016 February 2017. If party size was two or more, the individual with the most recent birthday was invited to complete the questionnaire and index (USDA Forest Service, 2013) Among the 242 recrea tion users approached onsite, 209 agreed to t ake the survey prod ucing a response rate of 8 6 .0%. The questionnaire included five sections. Section A included questio n s related to included test of knowledge about prescribed burning and indices of w ildlife value orientation s and attitudes about and intentions to support prescribed burning. Section C asked respondents to rate scenic beauty and recreation satisfaction from forest sites depic ted in 12 photographic images that showed sites at various times after prescribed burning and evidence of other active management practices, such as thinning and clearcutting. Section D included index measuring lace attachment. Finally, section E included demographic questions. A two stage test of the questionnaire occurred prior to data collection. First, seven graduate and undergraduate students from forestry and natural resource programs at U niversity of Flor ida provided feedback regarding clarity of the questions, such as wording, e ase of comprehension, and meaning of questions. After revising the
67 questionnaire based on this test, I tested the questionnaire with 42 recreation visitors to Ocala National Forest With the feedback from respondents regarding ease of comprehension and ten items intended to measure knowledge and attitude about prescribed burning and intentions to support prescribed burning in public forest wer e either re worded or replaced. Measurement and A nalysis The objective of this study was to understand how participation in outdoor recreation, knowledge about prescribed burning, and receiving ecological information (treatment) affect perception of sceni c beauty and recreation satisfaction. The first hypothesis was r ecreation users who participate in different activities have different levels of knowledge about prescribed burning (H 1 ) Grouping respondents by recreation activity was critical for testing t h is hypothe sis as almost all respondents reported that they participated in more than one activity. Therefore, I asked respondents to report their first, second, and third major activities and grouped them based on the reported first activity Now, to compare knowledge of prescribed burning by activity (H 1 ), I assessed following seven items. 1. Do you think leaving forests in a natural state is better for wildlife than using active forest management? 2. Have you heard or read about the use of prescribed burning or controlled fire for management of wildlife habitat? (Loomis, Bair, & Gonzlez Cabn, 2001) 3. Do you think prescr ibed burning or controlled fire a. would reduce the chance of high intensity wildfire? (Loomis et al., 2001) b. usually results in the death of the majority of animals in the area? (Loomis et al., 2001)
68 c. would create long term negative impa ct on the scenery in the forest ? d. would improve health of forest ? e. would improve habitat for many wildlife species? To minimize potential variation in respondents interpretation of what the terms active forest management and prescribed burning mean, I provided this definition thinning, fertilization, weed control, timber harvesting, and replanting; and prescribed burning refers to any controlled fire intentionally ignited to meet specif ic land Among the seven items, I adapted three from Loomis et al. (2001) and developed the other four items fo following (Loomis et al., 2001) Now, t o compare among the groups (H1) I c alculated a summative knowledge score based on 1 point for correct answers and 0 I tested the reliability of score by using Test of reliability examine s the internal c onsistency among responses to the variables and provides an indication of whether the items measure a single construct ( Vaske, 2008) The item total correlation refers to the correlation between a n individual item score across respondents and the summative score for all items across respondents iable measure of a respective domain (Leong & Austin, 2006; Vaske, 2008) I dropped one it prescr ibed burning or controlled fire would reduce the chance of high intensity wildfire than 0.4. Once this item was dropped, item total correlation of rest of the items ranged from 0.43 to 0.63 an 0.79, thus indicating reliability of the score. Then, I obtained composite scores of knowledge for
69 each respondent by adding scores on each item. Thus, the score for test of knowl edge ranged from zero to six. The second hypothesis was recreation users who participate in different activities perceive scenic beauty and recreation satisfaction from forest sites differently (H 2 ) There are many challenges for researchers to immediate perce ption of scenic beauty and recreation satisfaction from various management conditions (e.g., recent burning, thinning) in forest sites. Using photographs has the advantage of exposing people to various site conditions in less amount of time and it minimize s the logistical burden of transporting people to sites (Brunson & Shelby, 1992; Daniel & Boster, 1976) Most of the past studies used photographs to assess scenic beauty of forest sites from college students or the public in the indoor setting, assuming (Kreimer, 1977) However, those research ers did not capture the potential effect of onsite experience on perception of scenic beauty and recreation satisfaction of various forest conditions. Considering this gap in research method, I used photographs but obtained the ass essment of scenic beauty and recreation satisfaction from the on site recreationists. I took 48 pictures from various forest sites in ONF by using a digital crop sensor SLR camera. The pictures represented typical RCW habitat and surrounding areas with div erse tree sizes and densities, evidences of clearcutting, residuals from thinning, and various times after prescribed burning. To minimize effect of camera exposure on the visual quality of pictures I maintained the same setting ( 1/250 sec. f/8 35mm auto ISO, flash off) for all the pictures. F rom the 48 pictures, I then manually selected 24 pictures
70 making sure to capture variation in 1) time after burning, 2) tree size and stand density, 3) ground cover, and 4) mechanical treatments (thinning / harvesting) Lastly, from the 24 pictures, I selected 12 pictures (Figure 3 1 ) randomly and included them for the assessment I prepared a photobook with one color picture (10.5 inch by 7.25 inch) on each page, printed o n copy paper and laminated with 3 M il thickness. Respondents were asked to rate scenic beauty and t heir likely attainment of recreation satisfaction if they had to recreate on forest sites where surrounding view was as depicted on each picture. Scenic beauty ratings scale was 1 (not at all sce nic) 10 (very scenic) and recreation satisfaction rating scale was 1 (not at all satisfied) 10 (very satisfied). I adapted a scenic beauty rating scale from Daniel and Meitner (2001) who used a 10 point rating scale as 1 (very low scenic beauty) 10 (very high scenic beauty). Likewise, adaption of recreation satisfaction rating scale wa s from Yoon and Uysal (2005) who used 4 point rating scores as 1 (Not at all satisfied) to 4 (very satisfied). To make the rating scores consistent with the scenic beauty rating and to avoid confusion among the respondents, I modified the rating score range to a 10 point scale. I used item total correlation and co efficient to assess reliability of index scores. were 0.93 for scenic beauty index and 0.91 for recreation satisfaction index and item total correlation of pictures ranged from 0.49 to 0.80 for scenic beauty and 0.40 to 0.79 for recreation satisfaction thus suggesting reliability of the scores Once t he scores were tested for reliability, I obtained composite scores of both scenic beauty and recreation satisfaction for each respondent by taki ng averages of his/her ratings o f all pictures.
71 To test the third hypothesis (H 3 ), the effect of ecological info rmation on perception of scenic beauty and recreation satisfaction, I randomly divided respondents into two groups control and treatment. The respondents on the treatment group received the following information, which the control group did not receive. it is managed to maintain habitat for a federally endangered bird, the red cockaded woodpecker [a picture of RCW was shown]. The red cockaded woodpecker is a very important sp ecies because their activities provide habitat for hundreds of other species. However, foresters need to apply different management treatments, such as frequent The only information the control group received was that the pictures were taken from different sections of the Ocala National Forest. The fourth hypothesis was knowledge of prescribed burning affects perceptions of scenic beauty and recreation satisfaction; and pe rceived scenic beauty significantly affects recreation satisfaction as well as mediates the structural relationship between knowledge and recreation satisfaction. For this I fitted a measurement model (confirmatory factor analysis) to obtain a latent scor e of knowledge. The composite scores of both scenic beauty and recreation satisfaction were included in the structural model as the observed va riable. In addition, to control the effect of activity and demographic variables, I also included activities and demographic variables in the model as binary variables (1,0) I used analysis of variance and equivalent Welch test (for unequal variances) in SPSS 24.0 statistical software to compare knowledge of prescribed burning, perception of scenic beauty, and recre ation satisfaction among activity user groups (H 1 H 2 ). Likewise, I examined the effect of treatment information on perception of scenic beauty and recreation satisfaction (H 3 ) by comparing control group and treatment group using
72 independent sample t test For all group comparison tests, I used eta 2 ) to assess effect size. Following Cohen (1988) effect size is interpreted as small ( 2 = 0 .01), medium ( 2 = 0 .06), and large ( 2 = 0 .14) I used structural equation modeling in MPlus version 7.4 statistical software to assess the relationships among knowledge of prescribed burning, perception of scenic beauty, and recreation satisfaction (H 4 ). In this structural model I used confirmatory factor analysis and item total correlations to assess measure knowledge. Finally, I used a rat io of Chi 2 /df), Confirmatory Fit Index (CFI), Root mean Square Error of Approximation (RMSEA), and Standardized Root Mean Square Residual (SRMR) to examine whether the structural 2 / a good fit of model for the given data (Brown, 2006; Hu & Bentler, 1999; Kline, 2005) Results Descriptive Characteristics of S ample Among the 209 responses, categorization of respondents by activity included hiking or backpacking ( 30 %), OHV riding ( 1 6 %), camping ( 18 %), hunting (12%), and canoeing or kayaking (24 %). The average age of the respondents was 42 (Std. Dev. = 14) ye ars old with 30% of the respondents being less than 30 years old and 14 % being more than 60 years old Proportion of male respo ndents (71%) was much higher than the female respondents (31%). About half of the respondents (47%) had an education 2 0% had education attainment of h igh school or below About 60% of the respondents were fullt ime employed, 14% were retired, and about 10 % each were self employed and students (Table 3 1 ).
73 In terms of race and ethnicity, almost all respondents reported their race as White ( 96 %) and 8% of the respondents reported their ethnicity as Spanish, Hispani c or Latino (Table 3 1 ). One third of the respondents had an annual household income of less than $50,000, whereas one quarter of respondents had an annual income of $100,000 or more (Table 3 1 ). Knowledge of Prescribed Burning In general, respondents were knowledgeable about forest management and prescribed burning (Table 3 2 ). However, only about one third of the respondents (36 %) gave correct answers to all questions asked to assess knowledge. This indicates that almost two th ird of the respondents had some sort of misconception about active forest management and prescribed burning. For example, 8 9 % of the respondents said they had heard or read about the use of prescribed burning or control fire for management of wi ldlife habi tat; however, only 52 % knew that using active forest management is better for wildlife than leaving forests in a natural state (Table 3 2 ). Comparison of knowledge index score by recreation activity (Table 3 3 ) shows h unters were more knowledgeable about p rescribed burning than other recreation users thus supporting my first hypothesis (H 1 ) Perception of Scenic Beauty and Recreation Satisfaction Among the 12 photographic images (Figure 3 1) used to assess respondents' perception of scenic beauty and recre ation satisfaction, three pictures received the highest rating of both scenic beauty and recreation satisfaction (Figure 3 2): 1) forest sites with mixed pine trees and hardwood trees (picture 8), 2) a mature pine forest site with saw palmetto of about 2 f eet tall and more than two years after burning (picture 6), and 3) a mature pine forest site with evidence of regeneration (picture 11). Likewise, a
74 recent large clear cut (picture 5) and a mature pine forest burnt less than a week earlier and with a large burnt open area (picture 4) received the lowest rating of both scenic beauty and recreation satisfaction (Figure 3 2). In general, respondents gave higher ratings of both scenic beauty and recreation satisfaction to forest sites that had mixed species of trees, mature and thinned trees with large depth of view, no recent evidence of burning, sites with lush palmetto, and evidence of regeneration. On the other hand, forest sites with immature and dense trees, recent burning, large clear cuts, and thinning r esidues received the lower ratings of both scenic beauty and recreation satisfaction. Comparison of scenic beauty and recreation satisfaction index scores among activities (Table 3 3 ) shows significance of group differences for both scores (p<0.05) thus s upporting my second hypothesis (H 2 ) Consistent with knowledge of prescribed burning, hunters rated significantly higher scenic beauty than other recreation users In terms of recreation satisfaction, hunters rated highest recreation satisfaction, followed by hikers/backpackers and OHV users C ampers and those participating in canoeing/kayaking rated the least recreation satisfaction (Table 3 3 ). Effect of Ecological Information Results show no significant effect of treatment information on perception of sc enic beauty and recreation satisfaction (Figure 3 3 ) thus failing to support my third hypothesis (H 3 ) Index s core (1 10) of scenic beauty perception from treatment group (6.5) was not statistically different than that from the control group (6. 3 ). Like wise, index score (1 10) of recreation satisfaction from treatment group (6. 4 ) also did not differ significantly from the control group ( 6.0 ). Thus, results indicated that providing
75 ecological information about RCW and its habitat requirement was unlikel y to affect respondents' perception of scenic beauty and recreation satisfaction. Structural Relations The ratio of chi square and degrees of freedom ( 1.53 ), CFI (0.96), TLI (0.9 5 ), RMSEA (0.0 5 ), and SRMR (0.0 5 ), indicated good fit of the full structural m odel (Brown, 2006; Hu & Bentler, 1999; Kline, 2005) In the measurement model of the latent f actor loading s ranged from 0.4 2 to 0.75, thus suggesting a validity of a single factor (Vaske, 2008) The variance explained (square of factor loadings) by the knowledge variable ranged from 18 % for read about the use of prescribed burning or controlled fire for management of wildlife 56 % for the i rescribed burning or contr olled would improve ha bitat for many wildlife species (Figure 3 4). Results of the structural models show that knowledge of prescribed burning significantly affected both perception of scenic beauty and recreation satisfaction (Figure 3 4). With a one unit increase in knowledge, perceived scenic beauty increased by 0.3 3 (p<0.01) and recreation satisfaction increased by 0. 11 (p<0.0 5 ). Likewise, perceived scenic beauty significantly affected recreation satisfaction (Figure 3 4). With one unit increase in perc eived scenic beauty, recreation satisfaction increased by 0. 79 (p<0.01). In addition to a small direct effect, knowledge affected recreation satisfaction also indirectly through the mediation effect of scenic beauty. In fact, the indirect effect of knowled ge on recreation satisfaction through the mediation effect of scenic beauty (0.2 7 p<0.01) was stronger than its direct effect (Figure 3 4). Thus, results from structural model supported my fourth hypothesis (H 4 ) that k nowledge of prescribed burning affect s perceptions of scenic beauty and recreation satisfaction and p erceived
76 scenic beauty significantly affects recreation satisfaction and mediates the structural relationship between knowledge and recreation satisfaction. Among the control variables, activ ity and demographics, hunters were more knowledgeable about prescribed burning (0. 58 p<0.05) and rated higher scenic beauty (0.38, p<0.1) than other recreation users. Hispanics ( 0.58, p<0.05) and females ( 0.36, p<0.05) were likely be less knowledgeable than non Hispanics and males, respectively. Participation in hunting did not associate with recreation satisfaction directly. However, it did associate with recreation satisfaction i ndirectly through knowledge and scenic beauty perception (0.15, p<0.05), t hus suggesting that hunters were more likely to attain recreation satisfaction because of knowledge about prescribed burning and higher rating of scenic beauty than other activity users. On the other hand, campers were likely to attain lower recreation sat isfaction ( 0.21 p<0.05) than other activity users. Discussion The aim of this research was to better understand how chosen activity, ecological recreation satisfaction a nd perception of scenic beauty. This research contributes to the body of knowledge in several ways. First, this study is the first to examine the effect of both information intervention and held knowledge about forest management practices ption of forest aesthetics and recreation satisfaction. Index scores of scenic beauty and recreation satisfaction did not differ and the effect of treatment information was not significant. Although there is controversy in the literature regarding the rela tionship between affective perception and cognitive perception, this study supports the claim that when conditioning information is absent or not effective cognitive perception, such as acceptability of landscape management may not be
77 different from affec tive perception of visual stimuli, such as scenic beauty (Ribe, 2002; Zajonc, 1980) This finding indicates that the perception of aesthetic beauty of forest and recreation areas is a very important factor for recreation users to a ttain satisfying recreation experience. Also, the finding that existing knowledge is related to higher perceptions of scenic beauty and recreation satisfaction suggests that improving as prescribed burning, would improve their aesthetic perception and acceptance of forest and recreation areas. On the other hand, information provided at the time of the recreation experience (e.g., information boards or on site brochures) are likely less Second, this study offers an important methodological implication regarding using photographs for assessment of forest and landscape aesthetic perceptions and preferences. I used photographs to capt ure various forest conditions, but unlike most past research, I obtained ratings of both scenic beauty and recreation satisfaction from onsite users. This approach allowed respondents an opportunity to better reflect their perspective to the actual sites w hile rating the forest sites depicted in the pictures. In such as hiking and camping, while rating forest sites shown on the pictures. Past research often relies on respon dents to remember or imagine a recreation visit. Thus, I believe this approach offers more valid and reliable assessment related to perception and preference of forest aesthetics. Management I mplication s This study offer s important implications useful for the management of public forests to attain the dual goals of recreation and wildlife habitat management. Hunters,
78 hikers/backpackers, and OHV users were likely to attain better recreation experience s from forest sites undergoing frequent burning than camp ers Thus, promoting hiking /backpacking hunting, and OHV riding opportunities in RCW habitats (or areas requiring active management) and keeping camping sites away from these areas could nt of expected experience will be less impacted by prescribed burning. This will be beneficial to minimize conflict with land management objectives and achieve the multiple objectives of forest management. This study shows that most recreation visitors und erstand that healthy forests require burning, but there continues to be some misconceptions about the use of fire as a management tool for the improvement of wildlife habitat. Forest visitors are an easy to access audience that managers can target with edu cation and outreach on this complex ecological topic. Differences in knowledge of prescribed burning across different activities and demographics suggest the need for different approaches in reaching various recreation users. For example, h unters were the most knowledgeable about prescribed burning relative to recreationists engaged in other activities ; therefore, and should focus on transferring these knowledge to other r ecreation users. This study shows that t he extent to which people accept fire in natural areas relates to their understanding of how that fire affects or benefits a forest However, simply providing beneficial purposes of prescribed burning will likely no t resonate with recreation visitors to public lands. Thus, outreach and education programs should be careful in contextualizing the purpose of fire with the hope of providing long term
79 educational benefits What information to include in those information programs is also important. Most users of FNST, which pass through ONF, were either visiting to enjoy nature and scenery, or visiting to obtain solitude. Recreation settings having quality nature, wilderness areas, and wildlife habitats were the most impo rtant site attraction attributes for these visitors. Thus, prescribed burning education message should focus on how fire improves forest scenery (in long term), habitat quality for wildlife, and health of the wilderness. In addition, linking prescribed fi re to the protection of endangered species being used in the message should be selected carefully Past re search has shown that Floridians had higher values and were m ore familiar with gopher tortoise s than RCW (Johnson, 2015) with gopher tortoise as a focal species in framing information re lated to prescribed burning could be more effective in educational programs. Limitations and Future Directions of Research I classified and compared respondents by activity of participation. While doing so, I classified respondents based on their first ma jor activity (at the time of the survey). However, activities people chose to recreate are not always mutually exclusive as users could participate in more than one equally important activity. Although asked to report three activities, in order of importance, certain respondents (about 10%), such as those who hiked, as well as camped, struggled to report one activity as more important than other. Thus, there could be a certain bias in classification of respondents by activity. Future research should consider adopting more effective classification criteria to offer robust conclusion s regarding differences associated with the a ctivity people chose to recreate.
80 I introduced the information treatment during the assessment of perceived scenic beauty and recreation satisfaction, and the treatment was not significant. However, research has shown that providing ecological information knowledge and attitude about prescribed burning (DeLorme et al., 2005; Loomis et al., 2001; Toman & Shindler, 2006) In this research, I found knowledge of prescribed burning strongly affects recreation satisfaction through the mediation effect of perceived scenic beauty. Thus, future research using treatment i nformation to manipulate knowledge and assess ing how that change in knowledge affects perception of scenic beauty and recreation satisfaction would be a noble contribution to the body of knowledge. In addition, I used photographic images representing diffe rent times after burning; however, the forest sites were different. Thus, tracking same forest sites before and after burning for multiple years (time lapse photos) knowledge, scenic beauty perception, and recreation satisfa ction in longitudinal research setting will offer more robust evidence to confirm findings presented in this study. In conclusion, this study confirms that forest aesthetic perception and preference is something learned, and it has more to do with a person 's immediate purpose in using change with immediate information intervention, such as information boards; however, held knowledge about forest management practices, such as prescribed burning, could change these perceptions and preferences. To allow users to attain more positive outcome, camping sites should be kept away from RCW habitats (or areas requiring active management) Instead, recreation opportunities can be provide d in such sites for
81 hunters, hikers/backpackers, and OHV users. Hunters were the most k nowledgeable about prescribed burning, and they also rated higher scenic beauty and recreation satisfaction of forest sites representing various forms of active manageme nt. Thus, agencies should develop hands on programs to facilitate knowledge transfer from hun ters to other recreation users.
82 Table 3 1 Demographic characteristics of a dult outdoor recreation users in Ocala National Fores t, USA Demographic characteristics % of respondents Age 30 31 45 years 25 46 60 years 31 > 60 years 14 Gender Male 71 Female 29 Education Below High School 1 High School graduate 19 Some College no degree 24 9 31 Some Graduate School 3 11 Doctoral Degree 2 Occupation Fulltime employed 59 Retired 14 Student 8 Self employed 10 Race/Ethnicity White 96 Black, African American 1 Amer. Indian or Alaska Native 3 Asian 2 Spanish/Hispanic/Latino 8 Household Income Under $25,000 13 $25,000 $49,999 20 $50,000 74,999 23 $75,000 99,999 19 $100,000 149,999 12 $150,000 199,999 7 $200,000 or more 3 Not disclosed 4
83 Table 3 2 K nowledge of prescribed burning of a dult outdoor recreation users in Ocala National Fores t, USA Knowledge questions $ % of respondents (N=209) Yes No Don't know Leaving forests in a natural state is better for wildlife than using active forest management 34 52 15 Heard or read about the use of prescribed burning or controlled fire for management of wildlife habitat 89 11 Prescribed burning would reduce the chance of high intensity wildfire 90 2 8 Prescribed burning usually results in the death of the majority of animals in the area 16 76 8 Prescribed burning would create long term negative impact on the scenery in the forest 21 73 6 Prescribed burning would improve health of forest 84 6 10 Prescribed burning would improve habitat for many wildlife species 72 15 13 $ ance of high further analysis because of item total correlation value <0.4. Table 3 3 Comparison of index scores of knowledge of prescribed burning, perceived scenic beauty, and recreation satisfaction by recreation activity of a dult outdoor recreation users in Ocala National Fores t, USA Index scores Recreation activities Sample mean F or Welch Statistic $ p value Effect size HB OHV CAM HUNT CK Knowledge of prescribed burning 4.3 a 4.1 a 4.5 a 5.4 b 4.0 a 5.3 3.9 <0.001 0.06 Scenic beauty 6.3 a 6.5 a 6.2 a 7.4 b 5.9 a 6.4 2.6 0.017 0.06 Recreation satisfaction 6.4 a 6.4 a 5.7 b 7.3 c 5.5 b 6.1 3.4 0.004 0.08 Knowledge score ranged from 0 6 and it indicates summative scores of six items coded as 1, if correct answer, 0 otherwise. Scenic beauty and recreation satisfaction scores represent average scores of ten pictures and values ranged from 1 10. Effect size should be interpreted 2 2 2 =0.14 ). abc: different letters indicate significance of group difference in post comparison test at 5% level. $ Welch test was used for knowledge index score because of unequal variances. HB:
84 Hiking or Backpacking, OHV: Off Highway Vehicle, CAM: Camping, HUNT: H unting, CK: Canoeing or Kayaking. Figure 3 1. Photographic images used for scenic beauty and recreation satisfaction ratings of a dult outdoor recreation users in Ocala National Fores t, USA 6 1 2 3 7 8 9 10 11 12 5 4
85 Figure 3 2. Comparison of perceived scenic beauty and recreation satisfaction by pictures representing various forest conditions. Note: Measurement score ranged from 1 10. Figure 3 3. Comparison of perceived scenic beauty and recreation satisfacti on by treatment groups. Note: Difference between treatment group and control group was not statistically significant (p>1.0) for both scenic beauty and recreation satisfaction. 0 1 2 3 4 5 6 7 8 9 10 Pic. 1 Pic. 2 Pic. 3 Pic. 4 Pic. 5 Pic. 6 Pic. 7 Pic. 8 Pic. 9 Pic. 10 Pic. 11 Pic. 12 Index score Scenic Beauty Recreation Satisfaction 5.6 5.8 6 6.2 6.4 6.6 Scenic beauty Recreation satisfaction Index score Control group Treatment group
86 Figure 3 4. Structural relations among knowledge, scenic beauty, and recreation satisfaction, and associations with demographics and activity of participation of a dult outdoor recreation users in Ocala National Fores t, USA Note: Only significant variables were retained in the model. Hispanic, female, camping, and hunting w ere binary variables. Scenic beauty and recreation satisfaction scores ranged from 1 10. Items measuring latent variable, knowledge, were coded as 1, if correct answer; 0, otherwise. One variable wildfire was dropped because of the factor loading being l ess than 0.4. Values in measurement model of latent variable indicate standardized factor loadings, whereas values in structural models indicate standardized linear regression coefficients. *indicates significance of coefficient in 10% level **indicates si gnificance of coefficient in 5% level and *** indicates significance of coefficient in 1% level. Active : leaving forests in a natural state is better for wildlife than using active forest management. Heard : heard or read about the use of prescribed burning or controlled fire for management of wildlife habitat. Death : prescribed burning or controlled fire usually results in the death of the majority of animals in the area. Scene : prescribed burning or controlled would create long term negative impact on the scenery in the forest. Health : prescribed burning or controlled would improve health of forest. Habitat : prescribed burning or controlled would improve habitat for many wildlife species.
87 CHAPTER 4 CLIMATIC, TEMPORAL AND ECONOMIC FACTORS AFFECTING RECREATION VISITS TO FLORIDA NATIONAL SCENIC TRAIL: A TIME SERIES MODELING I ntroduction Environmental (w eather and climate 1 ) temporal ( seasons, months, holidays) and economic factors, such as gasoline price, are among the key factors that affect outdoo r recreation participation and experience W utility from their recreation experience. For example, changing temperature, why), the ir choice of tourism/recreation site and activities, and their eventual experience/satisfaction and willingness to revisit the place (Becken & Hay, 2007; De Urioste Stone, Scaccia, & Howe Poteet, 2015; Gssling, Scott, Hall, Ceron, & Dubois, 2012) In addition, c hanges in climate patterns may lead to changes in the populations and the composition of wildlife and vegetation in nature based recreation areas, and these changes might ultimately affect visitation choice and behavior (Paudyal et al., 2015; Richardson & Loomis, 2004) Also, people tend to have different expectations and levels of tolerance for weather and climatic conditions as they relate to different activities and recreation settings (Hewer et al., 2015; Li & Lin, 2012) In addition to weather and climate, economic factors (e.g., tr avelling costs ) and temporal factors (e.g., seasons, holidays ) significantly affect ou tdoor recreation participation. Peo ple tend to travel more during holidays and when expected travel expenses are low. 1 Weather refers to an atmospheric condition at any give n time or place, whereas, in terms of the mean and variability of relevant quantities over a period of time ranging (IPCC, 2007)
88 Given the dependence of the outdoor recreation and tourism industry on weather, climate, and market factors, fluctuations in weather factors (temperature and precipitation) and economic factors such as gasoline price, could create unforeseen shifts of demand and supply. Uneven distribution of use over time is one of the most significant problems in outdoor recreation and tourism, which may cause inefficient use of resources, loss of potential profit, difficulties in administrative scheduling, and result in negative impacts to the recreation experience and ecological condition (Manning & Powers, 1 984) Thus, unexpected changes in weather patterns climate and economy could affect local and regional economi es and management of resources. Weather and climate are among the major economic assets for the tourism and outdoor recreation industry, whic h, at various times and locations, could be classified along a spectrum of favorable to unfavorable (de Freitas, 2003 ) In general visitors prefer weather range that does not require specific behavioral adjustment (de Freitas, 2015) Thus, it is important to know how weather and clim ate favor or negatively affect tourism and outdoor recreation at various times and locations. More specifically, information related to what people consider ideal, suitable, acceptable, or unacceptable conditions, and what specific weather and climate rela ted criteria people use to make recreation choices (de Freitas, 2003) are important. Most of the previous weather and climate literature which have integrated tourism and outdoor recreation components are superficial, in a way that relationships between weather and outdoor recreation is more assumed than observed or objectively tested (de Freitas, 2003) In addition most of the existing outdoor recreation and tourism literature that integrated weather or climate are either based on short term data of
89 recreation participation and weather, or based on the stated preference to hypothetical scenario of weather and clima te (De Urioste Stone, Le, Scaccia, & Wilkins, 2016; Paudyal et al., 2015; Richardson & Loomis, 2004) One of the challenges to both l ines of literature is the lack of availability of daily visits data for an extended time period from diverse park and recreation areas (Aylen et al., 2014) Activi ty and destination specific climate information is important for potential visitors with diverse leisure activity preferences (Yu et al., 2009) P revious research suggests that participation in tourism and outdoor recreation rates are generally higher in summer than winter (McGinn, Evenson, Herring, & Huston, 2007) I n locations where the winter season is fairly long and summers are not extremely hot, an increase in temperature in comparison to the long term average is p ositively associated with activity participation (Aikoh, Abe, Kohsaka, Iwata, & Shoji, 2012) However, these findings cannot be generalized for all regions including places like Florida, where summers are extremely hot and humid and winters are mild. The Florida tourism and outdoor recreation season differs from that of most of the continental United States. For example, summ er is the best season for hiking in the northern United States. However, summer in Florida is hot and humid and participation in activities like hiking, mountain biking, camping, kayaking/canoeing, and OHV riding are typically low. On the other hand, unlik e northern regions of the nation, winter in Florida is the best season for many activities associated with terrestrial settings. Also, an analysis of ideal climatic conditions for tourism and outdoor recreation over the past 50 years has shown a declining trend in Central Florida, mostly due to the occurrence of more frequent heat indices above 3 5 C (Yu et al., 2009) Therefore, although it is
90 obvio us that climate and weather affect outdoor recreation participation, we do not know how and to what extent weather factors are associated with outdoor recreation participation in warm weather climates, like Florida. In addition, literature has thus far neglected to examine the consequences of extreme weather condition s on recreation participation on days following the extreme weather. To fill this gap in knowledge, this study uses t ime series and count data modeling to examine how weather factors, inclu ding temporal factor and gasoline price affect daily outdoor recreation visits Specifically, this research aims to answer the following specific research questions. 1. How do weather factors (e.g., temperature, precipitation, humidity, heat index) affect dai ly outdoor recreation participation ? 2. Does an unfavorable weather condition of a day affect recreation participation the following day? 3. How are temporal factors (weekend and public holidays) and the market price of fuel associated with recreation participa tion ? I answered these questions by using a regression modeling approach. For this, I developed a conceptual model as: (4 1) Where, Y = daily recreation participation; T= temperature; PPT= precipitation; RH= relative humidity; HI= heat index; WPH = weekend and public holidays; M= months; and GP = gasoline price Review of L iterature Effect of Weather and Climate Interaction between tourism/outdoor recreation demand and weather/climate is complex because of difficulty to both predict and manage consequences of changes
91 (B ecken & Hay, 2007; Denstadli, Jacobsen, & Lohmann, 2011; Dubois, Ceron, Gssling, & Hall, 2016) Nonetheless researchers have examined the impact of weather and climate on outdoor recreation participation around the world. This body of l iterature examini ng the effect s of weather and climate on tourism and outdoor recreation participation includes in terms of context and activity hiking (Li & Lin, 2012) urban trail traffic (Lindsey, Wilson, Rubchinskaya, Yang, & Han, 2007) golfing (Nicholls et al., 2008) skiing (Beaudin & Huang, 2014; Hamilton, Brown, & Keim, 2007; Shih, Nicholls, & Holecek, 2009) physical activity (Chan, Ryan, & Tudor Locke, 2006; McGinn et al., 2007) beach and public pool visits (Finger & L ehmann, 2012) and visitation to zoo s park s and natural areas (Aylen et al., 2014; Becken, 2013; Richardson & Loomis, 2004) The following review of the literature provides an overview of the existing knowledge about climatic, temporal, and fuel price factors on tourism and outdoor recreation. Several effects of weather and clima te on tourism and outdoor recreation have destination selection (Hamilton & Lau, 2005; Yu et al., 2009) More specifically, weather and climate can impact tourism and outdoor recreation industry by affecting travel and outdoor recreation decisions and tour p lans (Steiger, Abegg, & Jnicke, 2016) causing trip cancellation (Tervo, 2008) and affecting duration of stay (Coghlan & Prideaux, 200 9) when to go (Lohm ann & Kaim, 1999) Thus, in the tourism and outdoor recreation industry, weather and climate could serve as both an attraction and limitation (Steiger et
92 al., 2016) or push factor and pull factor (Agnew & Palutikof, 2006) For example, warm temperatures in Florida during the winter could serve as pull factors to attract visito rs from other northern states, whereas extreme hot weather during the summer could push Florida residents to other states. Second, weather and climatic factors affect participation and experience in tourism and outdoor recreation (Denstadli et al., 2011; Yu et al., 2009) Depending upon location, activity, and time, such impacts could be both positi ve and negative. For example, an unexpected early snowfall could reduce participation in some outdoor activities such as hiking and camping while increasing participation in other activities like skiing, ice skating, and sledding. Also, poor weather condit ion s could negatively with their recreation activity by negatively affecting their attainment of expected experience and perception of value for the money (Coghlan & Prideaux, 2009) ultimately affecting likelihood of future visits. Third, weather and climate determine seasonality in tourism and outdoor recreation (Butler, 2001; Yu et al., 2009) For example, summer in Florida is extreme ly warm and humid while winter is dry and clement. As a result, participation in hiking and other land based activities are high during the winter and much lower during the summer. On the other hand, participations in fresh water based activities like swimming and tubing in springs are significantly high er during the summer than in the winter, even though spring water temperature remains constant year round. A set of climatic factors (temperature, precipitation, humidity) describes the variability in t he actual daily weather, whereas the count of number of visitors to a specific site indicates the demand for tourism and outdoor recreation participation
93 (Agnew & Paluti kof, 2006) Researchers have examined impact of weather and climatic factors by modeling the number of recreation visits (daily, weekly, monthly, or yearly to one or multiple places) with various climatic factors. The major significant climatic factors that are identified as affecting outdoor activity participation are temperature, precipitation, humidity, cloud cover, wind speed, and sunlight hours (Aylen et al., 2014; Chan et al., 2006; Hamilton et al., 2007; Li & Lin, 2012; Scott, Gs sling, & de Freitas, 2008; Steiger et al., 2016) In the tourism and outdoor recreation context, de Freitas (2003) cl assified these climatic factors into three categories: 1) thermal (temperature, humidity, and solar radiation), 2) physical (rain and wind), and 3) aesthetic (sunshine and cloud condition). Temperature has a non linear effect on participation in land based recreation such as hiking with increase in recreation participation with rise in temperature up to a threshold, and then a decrease with the increase in temperature (Aylen et al., 2014) Researchers have found that tourists and outdoor recreation visitors consider temperature of 20 26 C as ideal for various land based recreation activities (Li se & Tol, 2002; Steiger et al., 2016; Wilson, Nicol, Nanayakkara, & Ueberjahn Tritta, 2008) The tolerable threshold level of temperature may be activity specific (Li & Lin, 2012) When temperature increases beyond th at threshold, even a small increase in temperature could significantly reduce recreation participation s (Maddison, 2001) Precipitation has been found as another important factor to affect outdoor recreation participation (Steiger et al., 2016) Although light and short period rains are not generally considered as detrimental as temperature, precipitation in the form of snow, showers, and thunderstorm has the potential to greatly affect travel and outdoor
94 recreation decisions (Hamilton et al., 2007; Morgan & Williams, 1999; Yu et al., 2009) Amount and duration of rain as well as time of the day at which rainfall events occur can pose variable e ffects on outdoor activity participation. For example, a study conducted in Prince Edward Island, Canada found a rainy day (14 mm of total rain) negatively affected physical activity participation (Chan et al., 2006) Likewise, Scott and Jones (2007) found that morning rains affect golf participation more negatively than afternoon rains. Clouds and sunshine are less likely to determine whether outd oor recreation activity will take place or not; however, they might negatively affect aesthetics of the destination and lessen recreation satisfaction for certain activity participants (Morgan & Williams, 1999; Yu et al., 2009) For instance, Morgan and Williams (1999) asked a panel of coastal managers and environmental science graduate students to rate aesthetics of various beaches based on a panoramic video. Results showed that cloud cs of the beaches. Thus, in recreation sites, where depth of view is important (e.g., mountain and beach), cloud cover could negatively affect recreation. Even though mild wind gusts could be favorable for outdoor recreation, strong winds could affect outd oor activity participation more severely than the temperature (Hewer et al., 2 015; Yu et al., 2009) For instance, a study found strong winds and heavy rain were the most influential factors affecting weather related decision making and onsite behavioral adjustment among campers. As, wind speed interacts with body mass index to a ffect physical activity (Chan et al., 2006) lean individuals are more likely to be affected by wind speed while participating in outdoor activit ies
95 Although each climatic factor affects tourism and outdoor recreation participation to a different extent, what people experience when outdoor s is the combined effect of multiple factors. Research has also shown that tourists and outdoor recreationists consider more than one climatic factor when planning a trip. For instance, Hamilton and Lau (2005) found that 91% of tourists consider more than one climate attribute when they collect climate information about the destination (Hamilton & Lau, 2005) In line with this finding, researchers have suggested including combined experiential climatic factors when developing predictive models (Spagnolo & De Dear, 2003; Yu et al., 2009) For example Yu et al. (2009) examined tourism related climatic variations and trends over the past 50 years in central Florida by using the heat index as one of the predictors. A heat index is a measure of th e combined effect s of temperature and relative humidity on human physiological responses to and perceptions of temperature (Rothfusz, 1990) Temporal and Economic Factors Temporal variables (e.g., seasons, months, day of the week, public holidays) affect tourism and outdoo r recreation industry by determi ning short term seasonality of demands For instance, people are more likely to plan their trips during the weekends and public holidays than on weekdays. In recreation demand models, it is common to include temporal factors as control variables to explain additional variance in recreation participation (Aylen et al., 2014; Finger & Lehmann, 2012; Shih et al., 2009) For example, Finger and Lehmann (2012) found temperature and precipitation as the significant determinants of daily visits to public pools in Zurich, Switzerland, by controlling the effect of non weather factors (e.g., weekends and holidays) in the regression model.
96 C hange in market price of goods and service can also affect tourism and outdoor recreation participation. For example, when gasoline price goes high, it can indirectly affect price of other merchandise, and people could minimize their travel expenses to com pensate for the additional expenses. R esearchers have found a negative effect of gas price on outdoor recreation participation (Napier & Bryant, 1980; Pergams & Zaradic, 2006) For example, Napier and Bryant (1980) examined the potential impact of increasing fuel costs on future outdoor recreation participation in Ohio and found that residents would reduce future participation in outdoor recreation activity if fuel prices continued to increase. Likewise, Pergams and Zaradic (2006) found increasing oil price as one of the factors (others being hours spent on electronic entertainment media) responsible for the decline in per capita national park visitation in the USA during 1993 to 2003. In the last five years, average daily gasol ine price in Florida has fluctuated significantly However, we have limited information about whether and how this changes in gas price is associated with ou tdoor recreation participation. Methods Study Area I conducted this study in the Ocala National For est (ONF) section of Florida National Scenic Trail (FNST). As one of 11 national scenic trails 2 in the United States, the FNST is a federally designated non motorized recreation trail, which extends over 2 National scenic trails are 100 miles (~160 km) or longer, continuous, primarily non motorized routes of outstanding recreation opportunity (National Park Service, 2016) established under the National Trails System Act of 196 8. Although either N ational Park Service o r US Forest Service are responsible for administrating each National Scenic Trail, various government and private entities own or manage lands along each National Scenic Trail.
97 more than 1600 km through the most beautiful, unique landscapes in Florida (Fig ure 4 1 ). Out of the total length of the trail, the ONF alone encompasses about 160 km and ONF is one of three and the second largest national forest in Florida. Data and Variables I collected data from January 1, 2010 to May 31, 2015. The counts of daily recreation visits were obtained by installing and operating infrared mechanical counters at three trailheads (Lake Delancy, Juniper Spring, and SR 19) in ONF (Fig ure 4 1). The dependent variable represents the sum of the counts f rom these three locations. The regressor variables included in the model were various climatic, temporal, and gasoline price factors (Table 4 1). I obtained daily weather data of the nearest weather station (about 55 km away) from National Climatic Data Ce nter. The climatic regressor variables included in the model were temperature, precipitation, relative humidity, and cold snap. The cold snap in the context of weather in Florida was defined as a dummy variabl e, if daily average temperature during the wint er season was >6C below the long term normal. In addition, I included two dummy variables representing suitability of a day for recreation 3 5 C 2.54 cm ; and 0 otherwise. an unf avorable day had normal weather condition s, 0 otherwise to test whether unfavorable weather condition s on a da y would affect recreation participation on the next day. Likewise, I calculated heat index from daily maximum temperature and relative humidity by following the equations developed by Rothfusz (1990) and used by National Weather Service.
98 In addition to the cl imatic variables, I included temporal factors (e.g., weekend and U.S. public holidays) and fuel price which improved the variance explained by the model without causing any multi collinearity issue. A recent survey has shown that 70% of drivers in the U.S own gasoline operated vehicles (American Automobile Association, 2016) Thus, state average price /liter for regular gasoline was included in the model. Daily average price of regular gas in the state of Florida was ob tained from U.S. Energy Information Administration. Model Selection The dependent variable (daily recreation visits) was highly skewed and it included many zero scores (Fig ure 4 2 ). Thus, the data did not meet the assumption of the ordinary least squares (OLS) model, which assumes a normal probability distribution. Log transformation was not appropriate because of a large proportion of zero counts in the data. In addition, the depe ndent variable was time series count data, i.e., the data had properties of both count data (non negative integer) and time series (temporal dependence). The most commonly used regression models for this kind of data and analysis are either count data or t ime series models. However, there is no clear standard in the literature regarding the choice between these two competing models as both have their own strengths and weakness. Thus, I used both models and compared the results to identify the best model fi t Count data model Researchers have used count data models in analysis of number of trips taken by recreationists to one or multiple sites over a specific period of time (Creel & Loomis, 1990; Scrogin & Berrens, 1999) Creel and Loomis (1990) state that when a dependent variable represent a count data, Poisson and negative binomial regression model s with
99 count probability distribution can perform better than the least square regression models with normal probability distribution. The choice between Poisson and negative binomial model depends upon the dispersion on the data (Yen & Adamowicz, 1993) The Poisson model assumes that the conditional mean and the conditional variance are equal. Many recreation participation data, however, common ly exhibit over dispersion, i.e., conditional variance exceeds the conditional mean. In that case, if the Poisson model is used, the standard errors of parameter estimates are biased because the Poisson model cannot capture the unobserved heterogeneity (Cameron & Trivedi, 2013; Hilbe, 2011; Yen & Adamowicz, 1993) As in other recreation demand literature, the dependent variable in this dataset also represents the count data, and the variance is considerably greater than the mean. Use of negative binomial regression has been common in recreation participation literature, where the dependent variable was over dispersed (Bowker et al., 2007; Finger & Lehmann, 2012; Scrogin & Berrens, 1999; Zawacki, Marsinko, & Bowker, 2000) Following Yen and Adamowicz (1993) and Englin, Holmes, and Sills (2003) the negative binomial probability distribution function can be represented as: (4 2) Where, is a non negative integer, is the over dispersion parameter, is gamma distribution, is conditional mean )), and = is greater variance than the m ean, thus justifying the appropriateness of the negative and Var are equal to
100 ; which indicates Poisson model as the more appropriate choice over the negative binomial regression model In time series data, an observatio n at a certain time usually correlates with the observation(s) from a previous time. As a result, the count data models exhibit serial dependence, thus resulting inefficient or biased estimates (Cameron & Trivedi, 2013) One potential option to work with this kind of data would be to include lag(s) of dependent variable as the explanatory variable(s) in the model. Including a lag(s) of the dependent variable in the model infers that the observations are dependent across time. However, the count data model assumes that the observations are independent (King, 1989) Brandt and colleagues (Brandt & Williams, 2001; Brandt, Williams, Fordham, & Pollins, 2000) have argued that if a count regression model includes a lagged dependent variable, the exponentiated coefficient of the lag represents the linear exponential growth rate not the autocorrelati on component. Thus, lagged dependent Poisson or negative binomial models are not suitable for stationary time series data. Time series model One of the most common approaches to take account of the temporal dependence in the data is to use a time series mo del. Use of time series model is common in the recreation demand literature (Aylen et al., 2014; Hamilton et al., 2007) Autoregressive (AR) and Moving Average (MA) are two fundamental roots of time series models. In the AR model, a value at a certain time is regressed on the value(s) from the previous time(s). A general AR model is give n as: (4 3)
101 Where, t is residual error term. In MA model, a value at a certain time is a function of average of values at certain previous times. (4 4) i terms are the weights applied to prior values in the time series. Although some timeseries data are best explained by either AR or MA models, some data require combined AR and MA model, e.g., Autoregressive Integrated Mo ving Average (ARIMA). Inspection of the plots of autocorrelation function (ACF) and partial autocorrelation function (PACF) provide a clue for the selection of appropriate timeseries model. ACF is defined as the correlation between a time series value at t ime and its value at time (Montgomery, Jennings, & Kulahci, 2008) Thus, ACF is given as: (4 5) For stationary timeseries data, the standard deviation is the same at all time on the series. Thus, ACF for stationa ry timeseries data is: (4 6) In general, a partial correlation is defined as the correlation between two variables after adjusting for the common factor(s) that might be affecting them. Thus, PACF is the autoco rrelation between a time series value at time and its value at time ) after adjusting for (Montgomery et al., 2008) A general equation of PACF is given as:
102 (4 7) The plots of ACF and PACF indicate the possible number of lags. ACF cuts off after lag q for a MA(q) model and PACF cuts off after lag p for an AR(p) model thus suggesting t he possible lags for each AR and MA (Table 4 2). One of the assumptions of the time series model is stationarity of the data. In a stationary time series, the joint probability distribution of the observations is equal to the probability distribution of the observations (Montgomery et al., 2008) Stationarity in the time series data is important because if the mean or variance were changing systematically, the prediction model should t ake this into account (Aylen et al., 2014) The common approach to achieve stationarity in the non stationary data are transformation, trend adjustment, and differ encing (Montgomery et al., 2008) The most practiced statistical test to check the stationary in the time series data is Augmented Dickey Fuller unit root test (Dickey & Fuller, 1979) The Dickey Fuller Test for 12 lags (ADF = 4.80, p= 0.01) show ed that the null hypothesis of unit root test can be rejected, thus suggesting that the dependent variable in this data does not have a unit root (i.e., the data is stationary). The ACF plot shows exponential decay after second lag, and the PACF plot shows damped sinusoid after first lag, thus suggesting the possibility of first or second order ARIMA model (Figure 4 3 and Table 4 2) In addition, both the ACF and PACF plots show consistent higher values at the seventh cycle, thus suggesting for a Seasonal (weekly) ARIMA (SARIMA 7 ) model (Brockwell & Davis, 2002; Hamilton et al., 2007) Following Montgomery et al. (2008, p. 283) a seasonal ARIMA model of orders (p, d, q) x (P, D, Q) with a seasonal cycle s is given as:
103 (4 8) Based on the ACF and PACF plots of dependent variable (Figure 4 3), I tested and compared 15 possible SARIMA 7 models with various combinations of AR(p) and MA(q) compared. While there are various model parameters available to compare the competing models, I considered Akaike information criterion (AIC) and Bayesian informa tion criterion (BIC). Among the competing models, the model that has the lowest values of AIC and BIC indicates the best fit model ( Cameron & Trivedi, 2013; Montgomery et al., 2008) The SARIMA models (2,0,1) x (1,0,1) 7 (1,0,1) x (1,0,1) 7 (1,0,0) x (1,0,1) 7 were the best three models based on their lowest values of both AIC and BIC (Table 4 3 ). Once I identified the best three SAR IMA7 models, I introduced the exogenous regressors ( Table 4 1) in each model and compared the results. This time the comparison of the models was based on the errors (mean error [ME]), root mean square error [RMSE], and mean absolute error [MAE]) in additi on to AIC and BIC. Similar to AIC and BIC, the lowest value of each error term indicates the best fit model (Nau, 2017) Among the three competing SARIMA 7 X models, information criteria and errors from SARIMA ((1, 0, 0) (1, 0, 1) 7 ) X) were lowest, thus suggesting this model as the best fitting time series model (Table 4 4 ). Once the best fit SARIMA 7 X mode l was identified, I used the same regressors to run a negative binomial regression model and compared the results with the time series model. Between the time series and negative binomial regression, I selected the best model based on information criteria, errors, Ljung Box test, and squared correlation between actual and predicted values.
104 Ljung Box is a goodness of fit test to examine whether the time series observations are white noise, i.e., uncorrelated and random with constant variance (Montgomery et al., 200 8) In terms of comparison of competing models, the best fit model should resemble white noise. In other words, the autocorrelations of residuals from a best fit model should be no different than zero up to a specified lag. In Ljung Box test, a higher p value ( >.05) indicates presence of no serial autocorrelations (Montgomery et al., 2008) The squared correlation between actual and fitted values ( R 2 CORR ) which is equivalent to the multiple R 2 in linear models, indicates percentage of variance in the depend ent variable that is explained by the model (Fogarty & Monogan, 2014) Despite the drawbacks that the R 2 CORR value can decrease with the addition of regressors, this m easure is recommended as a fit index to compare disparate models (Cameron and Trivedi, 2013, 193). Thus, a model that has substantially higher squared correlation of actual and fitted values than the other model more accurately explains the data. I combine d the data for daily visits, daily observations for the climatic factor, and daily gasoline price and stored in an Excel Spreadsheet (version 2016). I computed the analyses and graphs in the R statistical software (version 3.3.2) and R studio using tseries (Trapletti, Hornik, LeBaron, & Hornik, 2017) ggplot2 (Wickham & Chang, 2015) forecast (Hyndman et al., 2017) MASS (Ripley et al ., 2013) and AER (Kleiber & Zeileis, 2009) packages. Results Descriptive Characteristics of the Variables During the study period, cumulative daily visits to the study area ranged from zero to 120, with an average of 21 ( 2 0 ) visits per day. The average daily visits in 2011
105 was highe r ( = 24 19.5 ) than in other years (Figure 4 4). Among months, March received the highest number of average daily visits ( =3924 ), followed by February ( =3525 ), and Januar y ( =3326), whereas t ime between June and September received 10 or fewer visits/day (Figure 4 5). Among the days of the week, the average daily visits were highest on Saturday ( =4029 ), followed by Sunday ( =3125) whereas the average visits during the weekdays ranged from 13 18 visits/day (Figure 4 6). T here was a significant negative correlation between daily visits and average daily temperature ( r= 0.452, p<0.05), indicating a decrease d recreation visits with increase d temperature ( Figure 4 7 ) Analysis showed an increase of average daily visits with increase in average temperature up to 12 C (Figure 4 8) When average temperature went above 12C visit number decreased gradually with significant drops for temperature above 24 C (Figure 4 8). D ays with average temperature range of 4 23 C received at least the mean number of visits (21 visits/day) and 1 0 12 C received the highest daily visits (40 visits/day), thus indicating the former as favorable and later as the most preferable range of a verage daily temperature for scenic trail use in Florida (Figure 4 8). Likewise, days with an average temperature of more than 27 C received 10 or fewer visits/day indicating this range of temperature as less favorable for hiking. A trend of mean visitatio ns with heat index (feels like temperature) (Figure 4 9) also showed that recreation participation on the scenic trail significantly declined below aver age when heat index went above 35 C thus suggesting a significant effect of heat index on recre ation participation in Florida.
106 During the study period, average daily price for regular gas in Florida wa s significantly lower in 2010 ( = $0.73 0.03 / liter ) in comparison to other years ( p Although annual averages for the rest of the years were almost equ al, average daily price fluctuated among the months for each year (Figure 4 10). For instance, the price ranged from $ 0.70/lit during summer of 2010 to $ 1.03 / lit in April 2012, and dropped to $ 0.57 / lit in January 2015. As data show, most of the peaks in ga s price were during the summer time in each year when recreation visits were low thus indicating a negative correlation between gas price and recreation participation (Figure 4 10). Comparis on of Best Fit Time Series and Count Data Model The significant p value of the likelihood ratio (LR) test indicates that the negative binomial model was a better fit to the data over Poisson model (Table 4 5 ). However, when compared to best fit time series model [SARIMA ((1, 0, 0) (1, 0, 1) 7 ) X], the overall performanc e of the time series model was better than that of negative binomial model (Table 4 5 Figure 4 11, and Figure 4 12). First, despite that the negative binomial model had lower values of both AIC and BIC than in the time series model, values of all the error terms (RMSE, ME, and MAE) were lower in the time series model than in the negative binomial model (Table 4 5 ). Secon d, the squared correlation between actual and fitted values was significantly higher from the time series model (R 2 = 0.59) than from the negative binomial model (R 2 = 0.5 2 ), thus suggesting a better fit of data with the time series model (Table 4 5 ). Thir d, as indicated by Ljung Box test (Table 4 5 ), there was presence of a serial correlation among residuals from the negative binomial model (p<0.001), whereas no serial correlation was found among residuals from the time series model (p=0.56). Fourth, ACF a nd PACF values of residuals from the time series model were almost equal to zero up to 30th lag, whereas many ACF and PACF
107 values from the negative binomial model were significantly bigger than zero (Figure 4 11). Fifth, the histogram of residuals from the time series model resembled closeness to normality better than those from the negative binomial model (Figure 4 12). In addition, plots of residuals versus fitted indicated a better fit of data with time series model than negative binomial model (Figure 4 11). Thus, SARIMA (1, 0, 0) (1, 0, 1) 7 ) X was selected as the best fit model for this data. Regression Results from Final Model All the time series lag factors, first order auto regression (AR (1)), seasonal first order auto regression (AR (1) 7 ), and seas onal first order moving average (MA (1) 7 ) were statistically significant (p<0.01), thus indicating the suitability of this time series model to fit the data (Table 4 6 ). Among the climatic factors, temperature, relative humidity, and cold snap had signific ant negative effects on visitation numbers. Other factors remaining constant, 1C increase in average temperature decreased daily recreation visits to the study area by 0.9 4 (p<0.001). Likewise, a one percent increase in relative humidity resulted in a dec rease in recreation visits by 0.1 6 (p<0.001). A sudden fall in temperature (cold snap) also had a negative effect. Specifically, a day in w inter season received about 11 fewer visits than other days (p<0.001) if average temperature was below the long term daily average by 6C or more Rainfall, as a continuous variable, was not a significant predictor of recreation visits, but effect of rainfall extreme was negatively significant. Specifically, a n unfavorable day with 2.54 cm or higher rainfall (24 hour) or maximum heat index of 35 C or higher received about three fewer visits than other days (p<0.0 5 ), keeping everything else constant (Table 4 6 ). A day with unfavorable weather conditions negatively affected recreation visits of not only the same day, but also the visits of the
108 n ext day, even if next day had normal weather condition s In other words, a day with normal weather following an unfavorable weather day, also received about f ive fewer visi ts than other days (p<0.001), all other factors remaining constant. An analysis of frequency distribution of heat index during the study period has shown that about 40% of the total days in a year have heat index above 3 5 C This indicates that annual seasonality in land based recreati on in Florida is mostly driven by heat index. Among the temporal variables and gasoline price, weekends, public holidays, and the month of March (a month with highest average daily visits) were positively associated and gasoline price was negatively associ ated with the recreation visits. Particularly, everything remained constant, weekend or public holidays received about eight more visitors and a day in March received about 10 more visitors than other days (p<0.001). Likewise, number of recreation visits w as likely to decrease by about 27 visits per day (p<0.001) for every one dollar (per liter) increase in gas price keeping all other factors constant (Table 4 6 ) Discussion The aim of this research was to examine how climatic, temporal, and economic factors affect outdoor recreation participation in Florida. Past research conducted elsewhere shows that tourists and outdoor recreation visitors consider temperature s between 20 27C as favorable for various land based recreation activities (Lise & Tol, 2002; Steiger et al., 2016; Wilson et al., 2008) The favorable range of temperature identified in this study w as slightly lower than that found in past studies from other locations, which might be due to several reasons First, the range of favorable and preferable temperature identified in this study was average daily temperature, not the temperature at the time of visits. Second, the perceived (felt) actual temperature might
109 have been much higher than the measured temperature because of high humidity in Florida temperature might have been different than the objective measure of temperature as o utdoor physical activity could more depend on perceived temperature than the objective measures (McGinn et al., 2007) The finding regarding significance of heat index and heavy rainfall was as expected and is consistent with the finding from other studies (Li, 2008; Yu et al., 2009) Likely of heat related sickness and physical danger associated with stronger storms and possibili ty of thunder and lightning during the time of heavy rainfall could be the major reason for the significant negative effect of high heat index and heavy rainfall on hiking participation in forest land based recreation settings (Nicholls et al., 2008) One of the important findings of this study is that an undesirable weather condition tends to negatively affect recreation visitation not only on the day when it occurs but it also affects visitation on the following day, even if the weather condition on the following day is normal. Such lag effect of weather on recreation participation was also found in studies conducted elsewhere in different recrea tion contexts. For example, Hamilton et al. (2007) found that the amount of snow seen in town was significantly associated with the number of skiers to the nearby ski areas the next day. However, rainfall is more transitory than snow and temperature is something that people can feel rather than observe. Perhaps, certain cognitive process lead people to rely on the weather condition on one day to make their recreation decision for the next day This hypothesis is still unclear and further research is needed. Results also indicated that trail use volume is likely to decline with the increase in gas pri ce. There is evidence in the literature that when gas prices increases, people will
110 prefer to drive fewer miles and less travel will occur for recreation participation (Napier & Bryant, 1980; Pergams & Zaradic, 2006) This finding was expected and consitent with the existing literature. Previous research has shown that recreation users to this area drove a one way distance of about 156 km and only 7% of the vi 35 km (Paudyal & Stein, 2014) ; therefore, a gas price increase has the logical affect of decreasing trail use volume. Research and Management Implications This research adds to the body of knowledge in two ways. First, it identifie s the best approach of model ing long term data of daily recreation use with weather factors and relevant other temporal (e.g., weekend/public holidays) and market factors (e.g., gas price). Second, it improves our knowledge about the dependence of recreati on on weather. Specifically, it moved beyond the obvious result that when it is hot, and/or rainy people do not hike as much, but specific temperature conditions are specified for Florida outdoor recreation, which were previously not examined. Managers sho uld expect a higher use volume to hiking trails when average daily temperature is 6 23 C with most use in days wit h average daily temperature of 10 12 C The model developed in this study would be useful to the agencies in predicting daily use volume with the information of weather condition and past use volume. Information of near future recreation use volume in relation to weather condition will be useful for staffing and other resources management (Manning & Powers, 1984) which will be beneficial to operate and maintain economically viable recreation facility and maintaini ng healthy natural resources. In addition, the finding of delayed response to bad weather conditions has important implications for management and future research. Specifically, if the weather
111 condition of a day is not favorabl e, recreation visitation num bers on the following day, will be significantly fewer, even if the weather condition is within the favorable range. With this understanding, managers will more likely predict use on a day, that might seem favorable, but experience less use because of pre vious bad weather conditions. Research has shown that recreationists and day visitors usually plan their trip at short notice and adjust their plan according to short term weather predictions (McEvoy et al., 2006) Thus, more research is needed t o better understand why use drops on normal days following d ays with unfavorable weather conditions Perhaps certain psycholo gical or cognitive process makes recreation users rely on weather experienced on previous day over forecast for the same day. This study finds significant negative effect of temperature and p recipitation extremes on recreation participation. Change in climate is predicted to increase in frequency of higher maximum temperatures, heat waves, and precipitation extremes in the near future (IPCC, 2001) A n analysis of the past 50 years of weather data of central Florida has shown an increasing trend in frequency of heat indices above 3 5 C (Yu et al., 2009) Research has shown that mean temperature in Florida has increased between 0.25 and 1.4 C between 1961 and 2000 (Soule, 2005) and is expected to increase 3 C by 2050 and 5 C by the end of the century (Stanton & Ackerman, 2007) I n this scenario, model developed in this study suggests that daily recreational visitation in FNST will decrease by 13.43% by 2050 and by 22.38% by end of century only because of increase in temperature. Th eses number could even go higher because of delayed response effect identified in this study, as unfavorable weather condition (e.g.,
112 extreme heat) is likely to decrease recreation participation of not only the same day but also of the following day with n ormal weather condition. Research has found a significant decrease in forest and agricultural lands and increase in urban areas (26.6% ) in Central Florida during the period of 1995 2006 (Hernndez, Hwang, Escobedo, Davis, & Jones, 2012) With the inc rease in human population and associated increase in urban areas, concern has been raised regarding nature deficit disorder among the new generation because of their low outdoor recreation participation (Louv, 2008) Some of the constraints identified for reduced recreation participations were lack of recreation areas nearby, shortage of family time for outdoor activities, increasing amount of time spent with electronic devices (e.g., computer, TV, video games), and d ecline in park budget and associated increase in park entrances fees (Louv, 2008; Nyaupane & Andereck, 2008; Pergams & Zar adic, 2006) As recreation manag ers already have challenge s to overcome these constraints, negative effect of weather extremes on outdoor recreation participation found in this study could create additional challenge s through sole effect or interaction effect with previously identified c onstraints. Limitations and Future Research The dependent variable in this study was the total count of daily visits from three trail heads of a section of FNST in ONF collected over a period of five and half years. T he data do not represent the entire use volume of ONF or the entire section of FNST within ONF. There are more than ten official trailheads of the FNST within the ONF, of which three trailheads were selected to install and operate infrared counter to obtain be taken to interpret the overall use volume of both the ONF and FNS T.
113 Many past studies examined how recreation visitors would adjust their recreation pursuits with change in climate using different hypothetical scenarios (Paudyal et al., 2015; Richardson & Loomis, 2004) However, little is known about how recreation users would adapt to fluctuating weather conditions. Thus, research regarding how recreation users process weather inform ation in their recreation decision and how unforeseen weather would affect recreation decisions could be beneficial to develop theory based As found in this study, scenic trail use in Florida is significantly low in summer and fall, and is high during the winter and spring. Florida offers many destinations with natural springs, which are popular during the summer and fall. Thus, it is likely that recreation users go to spring water b ased activities (e.g., swimming, tubing) when weather conditions are unfavorable for hiking and similar land based activities. Thus, future research that focuses on how recreation users substitute activity and destinations based on objective or perceived w eather condition would be more beneficial to understand net effect of weather on recreation in Florida. The best fit time series model explain ed about 60% of the variance in the dependent variable. Within a same day, weather condition could fluctuate from unfavorable to favorable as the heat index goes down at the evening and heavy rainfall could last just a few hours. Thus, m odeling with hourly data, as did by Yu et al. (2009) of both weather and visitation, and considering heat index over temperature could improve the estimation. To conclude, this study examined how climatic (temperature, precipitation), temporal (weekends, holidays), and economic (gasoline price) factors affect daily
114 recreation participation in Florida National Scenic Trail. Finding shows that most people recreate in FNST when daily average temperature is 4 23 C with daily average temperature of 10 12 C being mos t preferable. Unfavorable weather condition (extreme heat or heavy rainfall) is found to affect recreation participation of not only the same day but also of the next day, even if weather condition is within the favorable range. Thus, managers of recreatio n facilities should expect low recreation participation on days following extreme weather conditions and maintain staffing accordingly.
115 Table 4 1 Descriptive characteristics of independent variables in the model examining daily recreation visits (dependent variable) to Ocala National Forest, USA Variables Definition Mean Std. Dev. Climatic Temperature Average temperature ( C ) 21.81 5.80 Precipitation Daily precipitation ( cm ) 0.33 1.05 Relative humidity Relative humidity in % 71.56 9.5 3 C old s nap 6C below the normal in winter; 0 otherwise 0.04 0 .19 Unfavorable day 1 if a day had heat index >35C or precipitation >2.54 cm 0 otherwise 0.38 0 .47 Day following an unfavorable day 1, if next day of an unfavorable day had normal weather; 0 otherwise 0.0 5 0 .23 Economic Gas price Avg. price for regular gas in Florida ($/ liter ) 0.85 0 11 Temporal Weekend & public holidays 1, if weekend or US public holidays; 0 otherwise 0.31 0 .46 March 1, if month is March; 0 otherwise 0.09 0 .29 Table 4 2 Theoretical characteristics of ACF and PACF for stationary data Model ACF PACF AR (P) Exponential decay and/or damped sinusoid Cuts off after lag p MA (q) Cuts off after lag q Exponential decay and/or damped sinusoid ARMA (p,q) Exponential decay and/or damped sinusoid Exponential decay and/or damped sinusoid Note: ACF: Autocorrelation Function; PACF Partial Autocorrelation Function; AR: Autoregressive; MA: Moving Average; Table s ource: Montgomery et al., (2015)
116 Table 4 3 AIC and BIC of competing time series models without exogenous regressors SN Model AIC BIC 1 ARIMA (1, 0, 0) x (0, 0, 1)7 16692.74 16715.09 2 ARIMA (1, 0, 0) x (1, 0, 0)7 16452.83 16475.19 3 ARIMA (1, 0, 0) x (1, 0, 1)7 16078.79 16106.74 4 ARIMA (0, 0, 1) x (0, 0, 1)7 16755.34 16777.69 5 ARIMA (0, 0, 1) x (1, 0, 0)7 16492.95 16515.31 6 ARIMA (0, 0, 1) x (1, 0, 1)7 16117.41 16145.36 7 ARIMA (1, 0, 1) x (0, 0, 1)7 16694.58 16722.53 8 ARIMA (1, 0, 1) x (1, 0, 0)7 16453.21 16481.16 9 ARIMA (1, 0, 1) x (1, 0, 1)7 16072.40 16105.93 10 ARIMA (2, 0, 0) x (0, 0, 1)7 16694.61 16722.56 11 ARIMA (2, 0, 0) x (1, 0, 0)7 16453.40 16481.35 12 ARIMA (2, 0, 0) x (1, 0, 1)7 16074.93 16108.46 13 ARIMA (2, 0, 1) x (0, 0, 1)7 16576.49 16610.03 14 ARIMA (2, 0, 0) x (1, 0, 0)7 16404.05 16437.59 15 ARIMA (2, 0, 1) x (1, 0, 1)7 16029.40 16068.53 Note: AIC: A kaike i nformation c riterion ; BIC: Bayesian i nformation c riterion ; *Lowest value indicates the best fit model Table 4 4 AIC, BIC and errors of best three time series models with exogenous regressors SN Model Information criteria* Errors* AIC BIC ME RMSE MAE 1 SARIMA ((1, 0, 0) (1, 0, 1) 7 ) X) 15 894.17 159 72.42 0.0 8 13.3 6 9.4 5 2 SARIMA ((1, 0, 1) (1, 0, 1) 7 ) X) 15 895.29 15979.13 0.08 13.3 6 9.46 3 SARIMA ((2, 0, 1) (1, 0, 1) 7 ) X) 15 895.91 159 85.33 0.0 9 13.3 6 9.45 Note: AIC: Akaike i nformation c riterion ; BIC: Bayesian i nformation c riterion ; ME: Mean error; RMSEA: Root mean square error; MAE: Mean absolute error; Regressors(X) are as listed in Table 4 1. *Lowest value indicates the best fit model
117 Table 4 5 Model fit comparison of negative binomial and final time series model Fit indices Models Negative Binomial SARIMA (100,101) 7 with regressors AIC 14 859.62 15 894.17 BIC 149 20.10 159 72.42 ME 0.0 8 0.08 RMSE 14.35 13. 36 MAE 10. 18 9.4 5 R 2 (original and fitted) 0.52 0.59 Ljung Box test 2 42 .6 6 (p<0.001) 10.10 (p=0.43 ) LR test 10 3 75 52 (p<0.001) n/a Regressors(X) on both models were as listed in Table 4 1. Table 4 6 Regression results from S ARIMA (100,101) 7 with regressors model examining daily recreation visits (dependent variable) Variables Coefficient Std. error Z value P>|Z| Climatic Average t emperature (C) 0.9 4 0.12 8.08 <0.001 Precipitation (cm ) 0.45 0.29 1.53 0.1 20 Relative h umidity 0.16 0.04 4.13 <0.001 Cold snap (avg. temperature >6C below long term normal in winter) 11. 14 2.14 5.20 <0.001 Unfavorable d ay 2.7 9 1.17 2.39 0.02 0 Normal day following an unfavorable day 4.7 2 1.35 3.49 <0.001 Economic 2.79 Gas p rice ($/ liter ) 26.58 9.53 3.48 <0.0 0 1 Temporal 6.65 Weekend & p ublic h olidays 7.5 8 2.18 11.61 0.001 March 10.4 8 1.58 66.01 <0.001 AR(1) 0.26 0.02 31.96 < 0.00 1 AR(1) 7 0.94 0.01 8.41 <0.001 MA(1) 7 0.76 0.02 8.08 <0.001 Intercept 73.63 8.7 6 1.53 <0.001 AR(1): 1st order auto regression, AR(1) 7 : Seasonal (weekly) 1st order auto regression; MA(1) 7 : Seasonal (weekly) 1st order moving average; Unfavorable day: (heat index >35C or P recipitation >2.54 inch)
118 Figure 4 1 Study area with sampling locations at Ocala National Forest Section of Florida National Scenic Trail in Florida, USA
119 Figure 4 2 Frequency distribution of average daily visits to Ocala National Forest Section of Florida National Scenic Trail (dependent variable) USA
120 Figure 4 3 Autocorrelation fu nction (ACF) and Partial autocorrelation function (PACF) of dependent variable
121 Figure 4 4 Average daily visits to Ocala National Forest Section of Florida National Scenic Trail by year.
122 Figure 4 5 Average daily visits to Ocala National Forest Se ction of Florida National Scenic Trail by months
1 23 Figure 4 6 Average daily visits to Ocala National Forest Section of Florida National Scenic Trail by day of the week
124 Figure 4 7 Average daily visits to Ocala National Forest s ection of Florida National Scenic Trail and daily avg. temperature ( C ) during the period of January 1, 2010 to May 31, 2015
125 Figure 4 8 Average daily visits to Ocala National Forest s ection of Florida National Scenic Trail by various ranges of average daily temperature ( C ) during the period of January 1, 2010 to May 31, 2015
126 Figure 4 9 Occurrence of various heat index values (left axis) and mean recreational visits (right axis) in Ocala National Forest s ection of Florida National Scenic Trail du ring the time period of January 1, 2010 to May 31, 2015
127 Figure 4 10 Average daily visits to Ocala National Forest s ection of Florida National Scenic Trail and average daily price for regular gas in Florida during the study period of January 1, 2010 to May 31, 2015.
128 Figure 4 11 Autocorrelation function (ACF) and Partial autocorrelation function (PACF) of residuals from final negative binomial regression model and timeseries SARIMAX ((1,0,0) (1,0,1) 7 with regressors) model
129 Figure 4 12 Histograms of residuals and plots of fitted Vs residuals from final negative binomial regression model and timeseries SARIMAX ((1,0,0) (1,0,1) 7 with regressors) model
130 CHAPTER 5 CONCLUSION K nowledge about various factors affecting outdoor recreation participation helps management agencies conserve and maintain natural resources and recreation service and facilities to satisfy needs and preference s of diverse visitor s. I examined various factors affecting outdoor recreation participation and experience of visitors to the FNST and vari ous managed areas in Florida. For this I divided these factors into three sets, 1) recreation motivations and site attribute preferences 2) climatic (temperature, humidity), temporal (weekends, holidays), and economic factors (gasoline price), and 3) forest management (prescribed burning). To examine the effects of these three sets of factors, I conducted three separate studies, which I p resented in three chapters. The second chapter of this dissertation identified motivations and site attribute preferences examined structural relations between them, and identified market segments of FNST visitors. In third chapter, I used the cognitive and experiential paradigm approach to examine how activity of recreation participation, knowledge of prescribed burning, ecological information treatment, and demographic factors affect perception of scenic beauty and recreation satisfaction, by taking RCW habitat as an example case. Finally, research discussed in the fourth chapter attempted to understand effects of climatic factors (temperature, precipitation, humidity), temporal factors (day of the week, month), and economic factors ( gasoline price ) on r ecreation participation in FNST. Findings presented in this dissertation offer better understanding about how people living in the urbanized regions, such as Florida, perceive,
131 experience, and prefer nature and nature bas ed activities. Unlike in the West a nd other rural states, most nature based recreation areas in the southeastern region are relatively small islands surrounded by private lan d and urban development, and are close to millions of people. Most nature based recreation research, however, focused in wildland settings areas relatively far from human development, such as wilde rness areas national parks and recreation areas in remote areas of the western U .S. As a result, there are limited number of research of this kind from urban park and recre ation area settings. In this context, this dissertation provides various important implications applicable to Florida and other developed areas, to attain dual goals of wildlife management and providing satisfying recreation opportunities for people living close from park and recreation areas. First, t his research confirms the previous findings that recreation motivations and site attribute preferences are both multidimensional concepts and various domains of motivations relate with various domains of site attractions (Kim et al., 2007; Kozak, 2002; Stein & Lee, 1 995; Uysal & Hagan, 1993) In other words, various recreation users prefer various setting ch aracteristics depending upon their recreation motivation. The research adds to the body of knowledge by identifying certain relationships between motivations and site attribute preferences that are unique and applicable to urban parks and recreation areas. For instance, unlike p reviously known, this study shows that recreation users being motivated for achievement (test skills, share skills, meet new peoples) do not necessarily prefer rural and secluded wilderness s ettings or areas that offer opportunities of challenging experiences. I n urban parks and recreation areas, users are likely to attain such experience from natural areas that offer hunting
132 and fishing opportunities and sections of trails that pass through interesting small towns and places popular for arts and handiworks. In addition, managers can provide satisfying experience to large proportion of visitors by managing wildlife and bird viewing opportunities, maintaining natural environment, and tranquility in parks, trails, and recreation areas, w here visitors can experience nature and attain fitness and relaxation. However, managers should also keep in mind of various typology of recreation users, in terms of their motivations and site attribute prefere nces, and maintain wide variety of opportunit ies. For instance experiencing solitude was most importance for some users and attaining social interaction was most important for some other users Second, this research finds that perceived scenic beauty of forest and recreation areas significantly aff ect To maintain healthy longleaf pine ecosystem and manage habitat for keystone species (RCW and gopher tortoise) foresters prescribe controlled burning every two to three years of rotation. This research shows that most recreation visitors understand that healthy forests require burning, but there were also some misconceptions about the use of a fire as a management tool for the improvement of wildlife habitat and scenery of forest in long term. Thus, t he tempo rary loss of scenery because of prescribed burning is likely to negatively affect recreation satisfaction of many users, especially those who place higher importance to experiencing natural environment and scenery I mproving e about forest management practices, such as
133 acceptance of forest and recreation areas. Thus, agencies should develop outreach and nding about prescribed burning. Outreach and extension message that focus on benefit of prescribed burning on improvement of wildlife habitat, and post fire positive effect on forest aesthetics, with demonstration of exa mples, could be effective to improve species, such as RCW and gopher tortoise, and their dependence on fire, could be beneficial in offering satisfying recreation experience t o visitors. Third, this research improves our understanding about dependence of outdoor recreation participation on weather conditions, specifically in Florida. Finding shows that most trail uses in Florida are during the days when average daily temperature is 4 23 C and days with temperature of 10 12 C were most preferred. Best fit time series model showed that daily recreation use of FNST is negatively affe cted by temperature and precipitation extremes and gasoline price. In addition, a bad weather condition (extreme heat or heavy rainfall) is likely to negatively affect not only the same day recreation use volume, but also of the next day. Specifically, if the weather condition of a day is not favorable, recreation visitation number, on the following day, will also be significantly fewer, even if the weather condition is within the favorable range. This finding is a noble contribution to the body of knowledg e. Although, socio psychological factors associated with this phenomenon are not yet clear, and more research is needed, mangers should expect fewer recreation users on days following extreme weather conditions and should plan and manage staffing and othe r resources accordingly.
134 APPENDIX A QUESTIONNAIRE FOR RECREATION MOTIVATIONS AND SITE ATTRIBUTE PREFERENCES SURVEY Florida Outdoor Recreation Visitor Study To be completed by surveyor if interview given on site: Surveyor: ___________________ Date: ___________________ Site: ________________________ Time: ___________________ Access Point: ________________ 1. Was this your first time on this particular trail? ____Yes (Go to question 4) ____ No (Go to question 2) 2. In what year did you make your first visit? __________ 3. Over the past year, how many times have you used this trail? ___None ___13 20 times ___1 6 times ___21 30 times ___7 12 times ___ more then 30 (#___) 4. About how long did you spend on the trail? ____1 hour or less ____Half a day ____More than 1 day (_____number of days) ____A few hours ____One whol e day 5. If you spent more than one day in the area, where did you stay overnight?  At a nearby hotel/condo  At a campground off the trail  In an established campground along the trail  In a nearby residence of friends or family  I live in the area 6. Approximately how many miles did you travel on the trail during this visit?  Less than a mile  3 5 miles  More than 10 miles (# of miles __________)  1 2 miles  6 10 miles 7. Hand the participant the activity card, Ask: From this list of activities, please rank the 3 activities that best describe the reason you visited the trail today? 1 st ____________ 2 nd _____________ 3 rd _____________ 8. Including yourself, how many people were you with? _______Total number of people (___#males, ___#females) Number of people under 18 9. What type of group are you traveling with? _____________________________________ 10. How did you learn about this trail? (Check all that apply)  Friends or Family  Roadside Signs  Magazine, please specify___  I live nearby & saw t he trail  Guidebook  Website  Brochure  Newspaper Article  Other, please specify _____________ SECTION A: CURRENT AND PAST HIKING EXPERIENCE
135 11. On a scale of 1 to 10, with 10 being the perfect experience, how would you rate your experience on this trail? ________ 12. If you did not rate your trail experience as a 10, why not? ________________________________________________________________________________ _________ 13. Are there any other improvements you would like to see on the trail? ________________________________ _________ SECTION B TRAIL AND TRAIL ORGANIZATIONS IN FLORIDA For this section we would like to understand what you know about trails and trail organizations in Florida. 14. What is the name of the trail you are now hiking on? ______________________________________________ Go to question 15 If yes and say FNST, go to question 15 If no or incorrect, go to question 16 Go to question 16 15. Other than this trail, have you hiked any other sections of the Florida National Scenic Trail?  No 16. Are you familiar with the Florida Tr ail Association?  Friends or family  Newspaper article  Website, please specify ________________________  Guidebook  Brochure  Travel agent  Magazine, please specify _______________________  Roadside signs  Other, please specify __________________________  No 17. Are you a member of the Florida Trail Association? ou been a member?  1 year or less  6 10 years  2 5 years  More than 10 years  No, but have been a member in the past for about _______years  No, not at all SECTION C RECREATION EXPENDITURE AND SUBSTITUTE Now we would like to learn about your recreation expenditures and preferred alternative activities. 18. Was visiting the FNST a sole or major purpose of your trip from home?  Yes  No, but it was one of many equally important reasons  No, it was just an incidental stop or spur of the moment decision 19. Approximately how long did you drive from your home (or hotel) to this trail?
136 One way distance ________ (miles) One way travelling time ___/ ____ (hours/min) 20. What type of vehicle did you use to travel to this trail? (Check one)  Full size Pick up/SUV  Small Pick up/SUV  Small Car  Medium sized Car  Large Car/Van 21. Was this vehicle a hybrid?  Yes  No 22. How many people travelled in the same vehicle with you? _____________# of people (including yourself) 23. Can you tell us how much money you spent on the following items in this trip? If you are in a multiple day trip, please provide the amounts you expect to spend on each items. Items  Expenditure so far OR  Expected expenditure Fees/Stamps/Entrance $ Transportation (gas, etc.) $ Restaurants/bars $ Groceries $ Hiking equipment and supplies $ Lodging (hotel, campground) $ Guide service $ Souvenirs/gifts $ Miscellaneous $ 24. If you could not hike on this section of trail on a typical day, what would you do instead? (Check one)  Go to another trail in Florida for hiking. If so, how many miles (one way) from your residence? (Go to question 25)  Go out of state for hiking. If so, how many miles (one way) from your residence? (Go to question 25)  Go somewhere else in Florida for another activity. If so, how many miles (one way) from your residence? (Go to question 28)  Stay home (Go to NEXT SECTION)  Go to wor k (Go to NEXT SECTION) 25. If you decided to hike somewhere else, Which trail would you prefer to use? Name or location__________________ How long would you spend hiking there? _____/_____ (days/hrs) How much out of pocket money would you expect to spend? ($)_______ 26. In comparison to your alternate hiking site, how would you rate the expenses in this section of FNST?  More exensive by approximately $______  Less expensive by approximately $______ 27. In comparison to your alternate hiking site, how would y ou rate the following for this section of the FNST? Proximity 1 (Very far) 2 3 4 5 (Very close) Environmental quality/site attraction 1(Poor) 2 3 4 5 (Excellent) Facilities/services 1(Poor) 2 3 4 5 (Excellent) (Go to NEXT SECTION) 28. If you decided to go somewhere else in Florida for another activity, other than hiking, What would be your preferred alternative activity? ___________________ How much time would you spend doing that activity? ______/______(days/hours)
137 How mu ch out of pocket money would you expect to spend? ($)_______ (Please hand the second set of pages to the visitor to fill out on their own.) SECTION D RECREATION EXPERIENCE PREFERENCE AND BENEFITS 29. Please indicate how important each item below was in choosing your leisure destination for this trip. Reason for Visit Not at all important Neutral Very Important Historical, military, or archeological sites 1 2 3 4 5 To see the natural water features 1 2 3 4 5 Wilderness and undisturbed nature 1 2 3 4 5 Good fishing 1 2 3 4 5 Good big game hunting 1 2 3 4 5 Easy access to the area/being easy to get to 1 2 3 4 5 Good environmental quality of air, water, and soil 1 2 3 4 5 Close to home 1 2 3 4 5 Interesting small towns 1 2 3 4 5 Good small game hunting 1 2 3 4 5 Chance to see wildlife/birds 1 2 3 4 5 Good camping 1 2 3 4 5 Local crafts or handiwork 1 2 3 4 5 Available parking 1 2 3 4 5 30. People go to particular areas and participate in recreation activities for any number of reasons. Please indicate how important each experience was for you during your visit to this area today. Experiences Not at all important Neutral Very Important To enjoy the scenery 1 2 3 4 5 To relax physically 1 2 3 4 5 To do something with my family 1 2 3 4 5 To get exercise 1 2 3 4 5 To explore the area 1 2 3 4 5 To experience nature 1 2 3 4 5 To be on my own 1 2 3 4 5 To use my own equipment 1 2 3 4 5 To learn about natural history of the area 1 2 3 4 5 To be away from people 1 2 3 4 5 To have thrills and excitement 1 2 3 4 5 To learn more about the nature 1 2 3 4 5 To meet new people 1 2 3 4 5 To test my skills and abilities 1 2 3 4 5 To enjoy the smells and sounds of nature 1 2 3 4 5 To get away from usual demands of life 1 2 3 4 5 To share my skills and knowledge with others 1 2 3 4 5 To be with members of my group 1 2 3 4 5 To be close to nature 1 2 3 4 5 To be with people who enjoy the same things I do 1 2 3 4 5 To experience new and different things 1 2 3 4 5 To experience solitude 1 2 3 4 5 To feel healthier 1 2 3 4 5
138 31. People see different benefits from their recreation experience. Please indicate your agreement with Benefits Strongly disagree Neutral Strongly agree Reduce health maintenance costs 1 2 3 4 5 Improve outdoor oriented lifestyle 1 2 3 4 5 Increase family or friend bonding 1 2 3 4 5 Provide positive contributions to local regional economy 1 2 3 4 5 Improve functioning of individuals in family or friends 1 2 3 4 5 Help increase local tourism revenue 1 2 3 4 5 Reduce social isolation 1 2 3 4 5 Improve economic benefits through increased work productivity 1 2 3 4 5 Improve parenting skills 1 2 3 4 5 Help improve local economic stability 1 2 3 4 5 SECTION E DEMOGRAPHICS We would like to ask a few questions about you, your background, and your past experiences. This information will be used for statistical analysis only, and all information will remain strictly confidential. 32. I am  Male  Female 33. Which of the following best describes your status?  Married  Divorced  Single  Widowed 34. How many children currently reside in your household? ________ 35. What is the highest level of education you have completed? (please mark one)  Eighth grade or less  Some College  Graduate Degree or beyond  Some High School  College Graduate  High School Graduate or GED  Some Graduate School 36. (Please mark all that apply)  Employed Full Time  Unemployed  Retired  Part Time Student  Employed Part Time  Full Time Homemaker  Full Time Student 37. If employed part time, how many hours a week do you work? __________hours/week 38. What is your profession or occupation? ______________________ 39. What year were you born? _______________________ 40. What race or ethnic group(s) would you place yourself in? (please mark all that apply)  African American  Hispanic or Latino  Asian American  Native Hawaiia n or Pacific Islander  American Indian or Alaskan Native  White 41. What was your approximate total household income, before taxes this past year?  Less than $10,000  $10,001 to $19,999  $20,000 to $29,999  $30,000 to $39,999  $40,000 to $49,999  $50,000 to $59,999  $60,000 to $69,999  $70,000 to $79,999  $80,000 to $89,999  $90,000 to $99,999  $100,000 or more 42. Zip Code: ___________________ 43. Do you have any comments you would like us to know?
139 APPENDIX B QUESTIONNAIRE FOR PRESCRIBED BURNING AND AESTHETIC PREFERENCE SURVEY Informed Consent Hello, The purpose of th is study is to examine recreation experience and preference of recreation visitors to public lands in Florida and understand how recreationists perceive scenic beauty and recreation value of different forest management practices. It takes about 10 minute to complete this survey. Your participation in this study is completely voluntary, you have the right not to answer any specific questions and you may withdraw your consent at any tim e. There are no ue feelings. There is no ant icipated risk or direct benefit to participants. However, indirect b benefits for visitors participating in this survey include the opportunity to have their input on improving recreation planning and management of recreation ser vices and facilities in public forests T he survey is anonymous; we would like to assure that your answers will not be connected with you in the analysis and reporting of results Will you participate in this study? Are you 18 years old or older? IF YES : Thank you! [Then continue with survey] IF NO: Thank you for your time! [Then randomly select next respondent] If you have any questions concerning this study, please contact: Whom to contact about your rights as a research participant in the study : UFIRB Office Box 112250 University of Florida, Gainesville, FL 32611 2250 Phone: (352) 392 0433 Taylor Stein, Ph.D., Professor School of Forest Resources and Conservation 345 Newins Ziegler Hall P.O. Box 110410 University of Florida Gainesville, FL 32611 0410 Phone: (352) 846 0860 E mail: firstname.lastname@example.org Ramesh Paudyal, Ph.D. Candidate School of Forest Resources and Conservation 227 Newins Ziegler Hall P.O. Box 110410 University of Florida Gainesville, FL 32611 0410 Phone: (352) 846 0874 E mail: email@example.com
140 Outdoor Recreation Experience & Forest Management Perception Survey 2016 Surveyor: ______________________________ Date: ___________________ Site: _____________________ Time: ___________________ 44. Was this your first visit to this forest site? __Yes ( Go to question 3 ) __ No ( Go to question 2 ) 45. Over the past year, how many times have you visited this forest site?  None  1 5 times  6 10 times  11 15 times  16 20 times  More than 20 times 46. About how long did you spend in this forest site during your current trip?  1 hour or less  A few hours  Half a day  One whole day  More than 1 day (# of days____) 47. Hand the participant the activity card, Ask: From this list of activities, which activity best describes the reason you visited this forest today? _________ If other(s) specify____________________________________________ 48. Including yourself, how many people are you with? Total______ Adult males _____ Adult females_____ Boys (under 18) ____ Girls (under 18) ____ 49. What type of group are you traveling with? ___Alone ___Friends ___Family ___Friends & Family ___Organized Group ___Other 50. Approximately how many miles (or hours) did you drive to reach this destination? ____miles ____hours 51. On a scale of 1 (Not at all satisfied) to 10 (Very satisfied), how would you rate your recreation experience on this forest site today? _______ 52. Based on your recreation experience today, how would you rate the overall scenic beauty of this forest site on a scale of 1 (not at all scenic) to 10 (very scenic) _____ 53. On a scale of 1 (very unlikely) to 5 (very likely), how likely are you to revisit this forest site within next 12 months? _______ SECTION B ECOLOGICAL KNOWLEDGE, VALUE ORIENTATION & ATTITUDE In this section we would like to know your opinion about management of natural forest areas and wildlife habitat in public lands, like the Ocala National Forest. The inform ation you provided will be useful for land management agencies to prioritize visitors' need and preferences. Please note, active forest management may include prescribed burning, thinning, fertilization, weed control, timber harvesting, and replanting. Pr escribed burning refers to any controlled fire intentionally ignited to meet specific land management objectives. SECTION A: CURRENT & PAST RECREATION EXPERIENCE
141 B 1 Statements Yes No Do you think leaving forests in a natural state is better for wildlife than using active forest management? Have you heard or read about the use of prescribed burning or controlled fire for management of wildlife habitat? Do you think Yes No would reduce the chance of high intensity wildfire? usually results in the death of the majority of animals in the area? would create long term negative impact on the scenery in the forest? would improve health of forest? would improve habitat for many wildlife species? B 2 Please answer the following questions on a scale of 1 (Absolutely Not) to 5 (Absolutely Yes) Statements Absolutely Not Neutral Absolutely Yes all living things are part of one big family? 1 2 3 4 5 fish and wildlife are on earth primarily for people to use? 1 2 3 4 5 wildlife should have rights similar to the rights of humans? 1 2 3 4 5 humans should manage fish and wildlife populations so that humans benefit? 1 2 3 4 5 we should strive for a world where there is an abundance of fish and wildlife for hunting and fishing? 1 2 3 4 5 hunting is cruel and inhumane to the animals? 1 2 3 4 5 feel a strong emotional bond with animals? 1 2 3 4 5 care about wildlife as much as you care about people? 1 2 3 4 5 B 3 To what extent do you disagree or agree with the following statement? Statements Strongly disagree Neutral Strongly agree Forest managers should periodically burn public pine forests to maintain habitat for wildlife 1 2 3 4 5 Public land managers should prioritize maintaining habitat for wildlife over providing recreation opportunities 1 2 3 4 5 All fires, regardless of origin, should be put out as soon as possible 1 2 3 4 5 Forest managers should avoid using prescribed fire because of the potential health problems to people from smoke 1 2 3 4 5 Prescribed burning should be allowed on public lands even if scenery will be negatively affected 1 2 3 4 5 Prescribed fire is too dangerous to be used 1 2 3 4 5 Prescribed burning is a waste of money 1 2 3 4 5 B 4 If you get an opportunity to participate in a planning meeting about management of a public forest in your area, how likely are you to SUPPORT each of the following management options?
142 Statements Very unlikely Neutral Very likely Create habitat for endangered birds and animals 1 2 3 4 5 Burn the forest in a controlled way every two three years of rotation 1 2 3 4 5 Exclude recreation activities in certain areas to promote habitat for wildlife 1 2 3 4 5 Use thinning (reduce the number of trees) to minimize competition among the plants 1 2 3 4 5 Clear cut (remove all the trees) in certain areas to establish longleaf pine plantation 1 2 3 4 5 SECTION C SCENIC BEAUTY & RECREATION SATISFACTION In this section, we would like to know your perception about different forest scenes. (Hand the picture album to the respondent) These pictures are taken from different sections of this forest where it is managed to maintain habitat for a federally endangered bird, the red cockaded woodpecker. (Show the RCW picture to the respondent). The red cockaded woodpecker is a very importan t species because their activities provide habitat for hundreds of other species. Foresters need to apply different management treatments, such as frequent prescribed burning and thinning to maintain suitable habitat for this species. For each of these pic tures, we ask you two different questions. The first question asks you to rate the scenic beauty (visual aesthetic quality) of the forest site. The second question asks you how satisfied do you think you would be doing this recreation activity (primary act ivity you are here for today) if the surrounding view is as shown on the picture. Picture # C1. How do you rate the scenic beauty of this forest site? 1 (not at all scenic) 10 (very scenic). C2. How satisfied do you think you would be doing this recreati on activity if the surrounding view is as shown on this picture? 1 (not at all satisfied) 10 (very satisfied). 1 2 3 4 5 6 7 8 9 10 11 12  Pre scanned
143 SECTION D RECREATION EXPERIENCE PREFERENCE AND PLACE ATTACHMENT D 1 People go to particular areas and participate in recreation activities for any number of reasons. Please indicate how important each experience was for you during your visit to this area today. Experiences Not at all important Neutral Very Important To enjoy the scenery 1 2 3 4 5 To do something with my family 1 2 3 4 5 To get exercise 1 2 3 4 5 To explore the area 1 2 3 4 5 To be on my own 1 2 3 4 5 To learn about natural history of the area 1 2 3 4 5 To learn more about the nature 1 2 3 4 5 To test my skills and abilities 1 2 3 4 5 To get away from usual demands of life 1 2 3 4 5 To share my skills and knowledge with others 1 2 3 4 5 To be with members of my group 1 2 3 4 5 To be with people who enjoy the same things I do 1 2 3 4 5 To experience solitude 1 2 3 4 5 To feel healthier 1 2 3 4 5 D 2 Please indicate your level of agreement with each of the following statements Statements Strongly Disagree Neutral Strongly Agree I am very attached to Ocala National Forest 1 2 3 4 5 I get more satisfaction out of visiting Ocala NF than from visiting any other forest 1 2 3 4 5 Ocala NF means a lot to me 1 2 3 4 5 I enjoy visiting Ocala NF more than any other forest 1 2 3 4 5 I have a special connection to Ocala NF and the people who use it for recreation 1 2 3 4 5 I would not substitute any other forests for the type of recreation I do in Ocala NF 1 2 3 4 5 I identify strongly with Ocala NF 1 2 3 4 5 Recreation in Ocala NF is more important than recreation in any other forest 1 2 3 4 5 SECTION E DEMOGRAPHICS Now, we would like to ask a few questions about you. This information will be used for statistical analysis only, and all information will remain strictly confidential. E 1.  Male  Female  Other E 2. Which of the following best describes your marital status?  Married  Separated  Divorced  Widowed  Never Married E 3. How many children currently reside in your household? ________ E 4. What is the highest level of education you have completed? (please mark one)
144  Eighth Grade or Less  Some High School  High School Graduate  Some College no degree  Some Graduate School  Professional Degree  Doctoral Degree E 5. What was your area of study (e.g., major or program) for your highest degree? (Only if highest level of education is Associate Degree or higher) _______________________________________ E 6.  Employed Full Time  Employed Part Tim  Self employed  Unemployed  Homemaker  Retired  Full Time Student  Part Time Student E 7. What is your profession or occupation? ______________________ E 8. What year were you born? _______________________ E 9. Are you Spanish / Hispanic / Latino?  No, not Spanish / Hispan ic / Latino  Yes, Mexican, Mexican American, Chicano  Yes, Puerto Rican  Yes, Cuban  Yes, other Spanish / Hispanic / Latino (Specify) _____ E 10. With which racial group(s) do you most closely identify? (Please check ALL that apply)  Whi te  Black, African American  Amer. Indian or Alaska Native  Asian Indian  Chinese  Filipino  Japanese  Korean  Vietnamese  Other Asian (Specify) _______  Native Hawaiian  Guamanian or Chamorro  Samoan  Other Pacific Islander (Specify) ___ E 11. What was your approximate total household income, before taxes this past year?  Under $25,000  $25,000 $49,999  $50,000 74,999  $75,000 99,999  $100,000 149,999  $150,000 199,999  $200,00 0 or more E 12 Zip Code: ___________________ E 13. Do you have any comments you would like to share?
145 LIST OF REFERENCES Agnew, M., & Palutikof, J. (2006). Impacts of short term climate variability in the UK on demand for domestic and international tourism. Climate Research, 31 (1), 109 120. Agrawal, S., & Monroe, M. C. (2006). Using and improving social capital to increase community preparedness for wildfire. The public and wildland fire management: social science findings for managers. Gen. Tech. Rep. NRS 1. (pp. 163 167). Newtown Square, PA: US Department of Agriculture, Forest Service, Northern Research Station. Aikoh, T. Abe, R., Kohsaka, R., Iwata, M., & Shoji, Y. (2012). Factors influencing visitors to suburban open space areas near a northern Japanese city. Forests, 3 (2), 155 165. Albritton, R., Denny, C., & Stein, T. V. (2006). Florida National Scenic Trail visitor assessment. Annual Report 2005 2006 A report presented to USDA National Forests in Florida and Florida Trail Association: University of Florida, School of Forest Resources and Conservation. Albritton, R., & Stein, T. V. (2007). Exploring marketing opportu nities for the Florida National Scenic Trail. A report presented to USDA National Forests in Florida and Florida Trail Association: University of Florida, School of Forest Resources and Conservation. Albritton, R., & Stein, T. V. (2008). Market segmentatio n of trail visitors: Marketing the Florida national scenic trail to Non FNST visitors. A report presented to USDA National Forests in Florida and Florida Trail Association: University of Florida, School of Forest Resources and Conservation. American Automo bile Association. (2016). Premium fuel research: Proprietary research into the use of premium octane gasoline when not required by the manufacturer. Heathrow, FL: American Automobile Association, Inc. Andereck, K., Gard McGehee, N., Lee, S., & Clemmons, D. (2011). Experience expectations of prospective volunteer tourists. Journal of travel Research 0047287511400610. Anderson, D. H., Nickerson, R., Stein, T. V., Lee, M. E., Gartner, W., & Lime, D. (2000). Planning to provide community and visitor benefit s from public lands. Trends in outdoor recreation, leisure and tourism. 197 211. Anderson, L. M. (1981). Land use designations affect perception of scenic beauty in forest landscapes. Forest Science, 27 (2), 392 400. Aylen, J., Albertson, K., & Cavan, G. (2014). The impact of weather and climate on tourist demand: the case of Chester Zoo. Climatic Change, 127 (2), 183 197.
146 Baloglu, S., & Uysal, M. (1996). Market segments of push and pull motivations: a canonical correlation approach. International Journal of Contemporary Hospitality Management, 8 (3), 32 38. Beaudin, L., & Huang, J. C. (2014). Weather conditions and outdoor recreation: A study of New England ski areas. Ecological Economics, 106 56 68. Becken, S. (2013). Measuring the effect of weather on tourism: a destination and activity based analysis. Journal of travel Research, 52 (2), 156 167. Becken, S., & Hay, J. E. (2007). Tourism and climate change: risks and opportunities Victoria Road, Clevedon: Channel View Publication. Beh, A., & Bruyere, B L. (2007). Segmentation by visitor motivation in three Kenyan national reserves. Tourism management, 28 (6), 1464 1471. Bowker, J. M., Askew, A. E., Cordell, H. K., Betz, C. J., Zarnoch, S. J., & Seymour, L. (2012). Outdoor recreation participation in th e United States: -Projections to 2060: A technical document supporting the Forest Service 2010 RPA Assessment : US Department of Agriculture, Forest Service, Southern Research Station. Bowker, J. M., Murphy, D., Cordell, H., English, D., Bergstrom, J., Star buck, C., . Reed, P. (2007). Wilderness recreation participation: Projections for the next half century. In A. Watson, J. Sproull, & L. Dean (Eds.), Science and Stewardship to Protect and Sustain Wilderness Values: Eighth World Wilderness Congress Symp osium; September 30 October 6, 2005 Anchorage, Alaska: USDA Forest Service Proceedings RMRS P 49. Brandt, P. T., & Williams, J. T. (2001). A linear Poisson autoregressive model: The Poisson AR (p) model. Political Analysis 164 184. Brandt, P. T., Willia ms, J. T., Fordham, B. O., & Pollins, B. (2000). Dynamic modeling for persistent event count time series. American Journal of Political Science 823 843. Bright, A. D., Manfredo, M. J., & Fulton, D. C. (2000). Segmenting the public: An application of valu e orientations to wildlife planning in Colorado. Wildlife Society Bulletin, 28 (1), 218 226. Brockwell, P. J., & Davis, R. A. (2002). Introduction to time series and forecasting New York: Springer Brown, T. A. (2006). Confirmatory factor analysis for app lied research : The Guilford Press.
147 Brown, T. C., & Daniel, T. C. (1987). Context effects in perceived environmental quality assessment: scene selection and landscape quality ratings. Journal of Environmental Psychology, 7 (3), 233 250. Brunson, M. W., & Shelby, B. (1992). Assessing recreational and scenic quality: How Does "New Forestry" Rate? Journal of Forestry, 90 (7), 37 41. Budruk, M., & Stanis, S. A. W. (2013). Place attachment and recreation experience preference: A further explor ation of the relationship. Journal of Outdoor Recreation and Tourism, 1 51 61. Budruk, M., Virden, R., & Waskey, M. T. (2009). Community conversations around public lands: Exploring beneficial outcomes in the Wood River Valley. Journal of Park & Recreati on Administration, 27 (1). Buhyoff, G. J., Leuschner, W., & Wellman, J. (1979). Southern pine beetle infestation affects esthetic values of forest landscapes. Southern Journal of Applied Forestry, 3 (2), 48 49. Buhyoff, G. J., & Leuschner, W. A. (1978). Es timating psychological disutility from damaged forest stands. Forest Science, 24 (3), 424 432. Bultena, G. L., & Klessig, L. L. (1969). Satisfaction in camping: A conceptualization and guide to social research. Journal of Leisure Research, 1 (4), 348. Butl er, R. (2001). Seasonality in tourism: Issues and implications. Seasonality in tourism 5 21. Byrne, B. M. (2001). Structural equation modeling with AMOS, EQS, and LISREL: Comparative approaches to testing for the factorial validity of a measuring instrum ent. International Journal of Testing, 1 (1), 55 86. Cameron, A. C., & Trivedi, P. K. (2013). Regression analysis of count data New York, NY: Cambridge University Press. Chan, C. B., Ryan, D. A., & Tudor Locke, C. (2006). Relationship between objective me asures of physical activity and weather: a longitudinal study. International Journal of Behavioral Nutrition and Physical Activity, 3 (1), 21. Coghlan, A., & Prideaux, B. (2009). Welcome to the wet tropics: the importance of weather in reef tourism resilie nce. Current issues in tourism, 12 (2), 89 104. Cohen, J. (1988). Statistical power for the behavioral sciences (2 ed.). Hillsdale, NJ: Lawrence Erlbaum Associates. Creel, M. D., & Loomis, J. B. (1990). Theoretical and empirical advantages of truncated cou nt data estimators for analysis of deer hunting in California. American Journal of Agricultural Economics, 72 (2), 434 441.
148 Crompton, J. L. (1979). Motivations for pleasure vacation. Annals of Tourism Research, 6 (4), 408 424. Cronbach, L. J., & Shavelson, R. J. (2004). My current thoughts on coefficient alpha and successor procedures. Educational and Psychological Measurement, 64 (3), 391 418. doi:10.1177/0013164404266386 Daniel, T. C. (2001). Whither scenic beauty? Visual landscape quality assessment in th e 21st century. Landscape and Urban Planning, 54 (1), 267 281. Daniel, T. C., & Boster, R. S. (1976). Measuring landscape esthetics: the scenic beauty estimation method USDA Forest Service Research Paper RM 167 Fort Collins, CO: Rocky Mountain Forest and Range Experiment Stattion. Daniel, T. C., & Meitner, M. M. (2001). Representational validity of landscape visualizations: the effects of graphical realism on perceived scenic beauty of forest vistas. Journal of Environmental Psychology, 21 (1), 61 72. Dani el, T. C., & Vining, J. (1983). Methodological issues in the assessment of landscape quality. In I. Altman & J. Wohlwill (Eds.), Human Behavior and Environment (Vol. VI, pp. 39 84). New York: Plenum Press. Dann, G. (1981). Tourist motivation an appraisal. Annals of Tourism Research, 8 (2), 187 219. de Freitas, C. R. (2003). Tourism climatology: evaluating environmental information for decision making and business planning in the recreation and tourism sector. International journal of biometeorology, 48 (1), 45 54. de Freitas, C. R. (2015). Weather and place based human behavior: recreational preferences and sensitivity. International journal of biometeorology, 59 (1), 55 63. De Urioste Stone, S. M., Le, L., Scaccia, M. D., & Wilkins, E. (2016). Nature based Maine. Journal of Outdoor Recreation and Tourism, 13 57 65. De Urioste Stone, S. M., Scaccia, M. D., & Howe Poteet, D. (2015). Exploring visitor perceptions of the influence o f climate change on tourism at Acadia National Park, Maine. Journal of Outdoor Recreation and Tourism, 11 34 43. DeLorme, D. E., Hagen, S. C., & Stout, I. J. (2005). Perspectives on prescribed burning: Issues and directions for developing campaign messag es. The Environmental Communication Yearbook, 2 99. Denstadli, J. M., Jacobsen, J. K. S., & Lohmann, M. (2011). Tourist perceptions of summer weather in Scandinavia. Annals of Tourism Research, 38 (3), 920 940.
149 Dickey, D. A., & Fuller, W. A. (1979). Dist ribution of the estimators for autoregressive time series with a unit root. Journal of the American statistical association, 74 (366a), 427 431. Dillard, J. E., & Bates, D. L. (2011). Leisure motivation revisited: why people recreate. Managing Leisure, 16 ( 4), 253 268. Driver, B. (1983). Master list of items for Recreation Experience Preference scales and domains Unpublished document. USDA Forest Service. Rocky Mountain Forest and Range Experiment Station. Fort Collins, CO. Driver, B. (2008). Managing to optimize the beneficial outcomes of recreation : Venture Publishing. Driver, B., & Tocher, S. R. (1970). Toward a behavioral interpretation of recreational engagements, with implications for planning. Elements of outdoor recreation planning, 8 9 31. Duboi s, G., Ceron, J. P., Gssling, S., & Hall, C. M. (2016). Weather preferences of French tourists: lessons for climate change impact assessment. Climatic Change, 136 (2), 339 351. Edwards, D., Jay, M., Jensen, F. S., Lucas, B., Marzano, M., Montagn, C., . Weiss, G. (2012). Public preferences for structural attributes of forests: Towards a pan European perspective. Forest Policy and Economics, 19 12 19. Englin, J. E., Holmes, T. P., & Sills, E. O. (2003). Estimating forest recreation demand using count data models Forests in a market economy (pp. 341 359): Springer. Finger, R., & Lehmann, N. (2012). Modeling the sensitivity of outdoor recreation activities to climate change. Climate Research, 51 (3), 229 236. Fix, P. J., Carroll, J., & Harrington, A. M. (2013). Visitor experiences across recreation settings: A management or measurement issue? Journal of Outdoor Recreation and Tourism, 3 28 35. Fix, P. J., & Taylor, S. C. (2011). Generalizability of a cluster solution over time: results from a visitor st udy at the Kennecott National Historic Landmark. Journal of Leisure Research, 43 (2), 246. Fogarty, B. J., & Monogan, J. E. (2014). Modeling time series count data: the unique challenges facing political communication studies. Social science research, 45 73 88. Frank, S., Frst, C., Koschke, L., Witt, A., & Makeschin, F. (2013). Assessment of landscape aesthetics validation of a landscape metrics based assessment by visual estimation of the scenic beauty. Ecological Indicators, 32 222 231.
150 Galloway, S. (2012). Recreation specialization among New Zealand river recreation users: a multiactivity study of motivation and site preference. Leisure Sciences, 34 (3), 256 271. Gobster, P. H. (1994). The aesthetic experience of sustainable forest ecosystems. Sustai nable Ecological Systems: Implementing an Ecological Approach to Land Management. General Technical Report RM 247, USDA Forest Service, Rocky Mountain Forest and Range Experiment Station, Fort Collins, CO 246 255. Gobster, P. H. (1999). An ecological aes thetic for forest landscape management. Landscape Journal, 18 (1), 54 64. Gssling, S., Scott, D., Hall, C. M., Ceron, J., & Dubois, G. (2012). Consumer behaviour and demand response of tourists to climate change. Annals of Tourism Research, 39 (1), 36 58. Graefe, A. R., & Burns, R. C. (2013). Testing a mediation model of customer service and satisfaction in outdoor recreation. Journal of Outdoor Recreation and Tourism, 3 36 46. Hall, T. E., Seekamp, E., & Cole, D. (2010). Do recreation motivations and wi lderness involvement relate to support for wilderness management? A segmentation analysis. Leisure Sciences, 32 (2), 109 124. Hamilton, J. M., & Lau, M. A. (2005). The role of climate information in tourist destination choice decision making. In S. Gsslin g & C. M. Hall (Eds.), Tourism and global environmental change: Ecological, economic, social and political interrelationships (pp. 229). London: Routledge. Hamilton, L. C., Brown, C., & Keim, B. D. (2007). Ski areas, weather and climate: time series models for New England case studies. International Journal of Climatology, 27 (15), 2113 2124. Hernndez, J. L., Hwang, S., Escobedo, F., Davis, A. H., & Jones, J. W. (2012). Land use change in central Florida and sensitivity analysis based on agriculture to urb an extreme conversion. Weather, Climate, and Society, 4 (3), 200 211. Hewer, M. J., Scott, D., & Gough, W. A. (2015). Tourism climatology for camping: a case study of two Ontario parks (Canada). Theoretical and Applied Climatology, 121 (3 4), 401 411. Hilb e, J. M. (2011). Negative binomial regression : Cambridge University Press. Hu, L. T., & Bentler, P. M. (1999). Cutoff criteria for fit indexes in covariance structure analysis: Conventional criteria versus new alternatives. Structural Equation Modeling: A Multidisciplinary Journal, 6 (1), 1 55. doi:10.1080/10705519909540118
151 Hubert, W. A., & Gipson, R. D. (1996). Angler survey contributes to socially acceptable modification of harvest regulations to preserve cutthroat trout fishery in Snake River, Wyoming, US A. Environmental Management, 20 (5), 707 713. doi:10.1007/BF01204141 Hull, R. B., & Reveli, G. R. (1989). Cross cultural comparison of landscape scenic beauty evaluations: A case study in Bali. Journal of Environmental Psychology, 9 (3), 177 191. Hyndman, R J., O'Hara Wild, M., Bergmeir, C., Razbash, S., Wang, E., & Hyndman, M. https://cran.r project.org/web/packages/forecast/forecast.pdf IPCC. (20 01). Impacts, adaptation and vulnerability. Contribution of Working Group II to the Third Assessment Report of the Intergovernmental Panel on Climate Change. Cambridge University Press. IPCC. (2007). Climate Change 2007: The Physical Science Basis. Working Group I Contribution to the Fourth Assessment of the Intergovernmental Panel on Climate Change Retrieved from New York, NY: Iso Ahola, S. E. (1982). Toward a social psychological theory of tourism motivation: A rejoinder. Annals of Tourism Research, 9 (2), 256 262. Jacobson, S. K., Monroe, M. C., & Marynowski, S. (2001). Fire at the wildland interface: the influence of experience and mass media on public knowledge, attitudes, and behavioral intentions. Wildlife Society Bulletin 929 937. Johnson, R. L ., Brunson, M. W., & Kimura, T. (1994). Using image capture technology to assess scenic value at the urban/forest interface: a case study. Journal of Environmental Management, 40 (2), 183 195. Johnson, S. A. (2015). Tradeoffs among wildlife diversity, soci al values, and ecosystem services: considerations for wildlife habitat management. (PhD Dissertation), University of Florida, Gainesville, FL. Kearney, A. R. (2001). Effects of an informational intervention on public reactions to clear cutting. Society &Natural Resources, 14 (9), 777 790. Kearney, A. R., & Bradley, G. A. (2011). The effects of viewer attributes on preference for forest scenes: Contributions of attitudes, knowledge, demographic factors, and stakeholder group membership. Environment and Be havior, 43 (2), 147 181. Kil, N., Holland, S. M., & Stein, T. V. (2014). Structural relationships between environmental attitudes, recreation motivations, and environmentally responsible behaviors. Journal of Outdoor Recreation and Tourism, 7 16 25.
152 Kil, N., Holland, S. M., & Stein, T. V. (2015). Experiential benefits, place meanings, and environmental setting preferences between proximate and distant visitors to a national scenic trail. Environmental Management, 55 (5), 1109 1123. Kil, N., Holland, S M., Stein, T. V., & Ko, Y. J. (2012). Place attachment as a mediator of the relationship between nature based recreation benefits and future visit intentions. Journal of Sustainable Tourism, 20 (4), 603 626. Kim, K., Noh, J., & Jogaratnam, G. (2007). Mul ti destination segmentation based on push and pull motives: Pleasure trips of students at a US university. Journal of Travel & Tourism Marketing, 21 (2 3), 19 32. Kim, S. S., Lee, C. K., & Klenosky, D. B. (2003). The influence of push and pull factors at K orean national parks. Tourism management, 24 (2), 169 180. King, G. (1989). Event count models for international relations: Generalizations and applications. International Studies Quarterly, 33 (2), 123 147. Kleiber, C., & Zeileis, A. (2009). AER: Applied Econometrics with R. R package version 1.1 New York: Springer end investigation. Journal of travel Research, 40 (4), 396 403. Kline, R. B. (2005). Principles and practice of structural e quation modelling (2nd ed.). New York, NY: The Guilford Press. Kotler, P. (1994). Marketing management, analysis, planning, implementation, and control New Jersey: Prentice Hall. Kotler, P., & Zaltman, G. (1971). Social marketing: an approach to planned s ocial change. The Journal of Marketing 3 12. Kozak, M. (2002). Comparative analysis of tourist motivations by nationality and destinations. Tourism management, 23 (3), 221 232. Kreimer, A. (1977). Environmental preferences: A critical analysis of some re search methodologies. Journal of Leisure Research, 9 (2). Kruger, M., Viljoen, A., & Saayman, M. (2017). Who visits the Kruger National Park, and why? Identifying target markets. Journal of Travel & Tourism Marketing, 34 (3), 312 340. Kyle, G., Absher, J. D., Hammitt, W. E., & Cavin, J. (2006). An examination of the motivation involvement relationship. Leisure Sciences, 28 (5), 467 485.
153 Kyle, G., Norman, W., Jodice, L., Graefe, A., & Marsinko, A. (2007). Segmenting anglers using their consumptive orientatio n profiles. Human Dimensions of Wildlife, 12 (2), 115 132. doi:10.1080/10871200701196066 Lawler, E. E. (1973). Motivation in work organizations Monterey, CA: Brooks/Cole Publishing Co. Lee, C. K., Lee, Y. K., & Wicks, B. E. (2004). Segmentation of festival motivation by nationality and satisfaction. Tourism management, 25 (1), 61 70. Lee, T. H. (2009). A structural model to examine how destination image, attitude, and motivation affect the future behavior of tourists. Leisure Sciences, 31 (3), 215 236. Leon g, F. T., & Austin, J. T. (2006). The psychology research handbook: a guide for graduate students and research assistants California: SAGE Publications, Inc. Lerner, J. S., & Keltner, D. (2000). Beyond valence: Toward a model of emotion specific influence s on judgement and choice. Cognition & Emotion, 14 (4), 473 493. Li, C. (2008). The acceptable level of weather condition among the different types of hikers. J Phys Education High Education, 10 (2), 43 54. Li, C., & Lin, S. H. (2012). Influence of weather conditions on hiking behavior. International journal of biometeorology, 56 (4), 777 781. Lim, S. H., Bowker, J. M., Johnson, C. Y., & Cordell, H. K. (2009). Perspectives on prescribed fire in the south: doe s ethnicity matter? Southern Journal of Applied Forestry, 33 (1), 17 24. Lindsey, G., Wilson, J., Rubchinskaya, E., Yang, J., & Han, Y. (2007). Estimating urban trail traffic: Methods for existing and proposed trails. Landscape and Urban Planning, 81 (4), 2 99 315. Lise, W., & Tol, R. S. (2002). Impact of climate on tourist demand. Climatic Change, 55 (4), 429 449. Lohmann, M., & Kaim, E. (1999). Weather and holiday destination preferences image, attitude and experience. The Tourist Review, 54 (2), 54 64. Lo ker Murphy, L. (1997). Backpackers in Australia: A motivation based segmentation study. Journal of Travel & Tourism Marketing, 5 (4), 23 45. Loomis, J. B., Bair, L. S., & Gonzlez Cabn, A. (2001). Prescribed fire and public support: Knowledge gained, atti tudes changed in Florida. Journal of Forestry, 99 (11), 18 22.
154 Louv, R. (2008). Last child in the woods: Saving our children from nature deficit disorder Chapel Hill, NC: Algonquin Books. Maddison, D. (2001). In search of warmer climates? The impact of cl imate change on flows of British tourists. Climatic Change, 49 (1), 193 208. Manfredo, M. J., Driver, B., & Tarrant, M. A. (1996). Measuring leisure motivation: A meta analysis of the recreation experience preference scales. Journal of Leisure Research, 28 (3), 188. Manning, R. E. (1999). Studies in outdoor recreation : search and research for satisfaction (Vol. 2nd ed). Corvallis, Or: Oregon State University Press. Manning, R. E., & Powers, L. A. (1984). Peak and off peak use: redistributing the outdoor re creation/tourism load. Journal of travel Research, 23 (2), 25 31. McCool, S. F., & Moisey, R. N. (2001). Tourism, recreation, and sustainability: Linking culture and the environment : CABI Publishing. McEvoy, D., Handley, J., Cavan, G., Aylen, J., Lindley, S., McMorrow, J., & Glynn, S. (2006). Climate change and the visitor economy: the challenges and Sustainability Northwest (Manchester) and UKCIP (Oxford) McFarlane, B. L., Boxall, P. C., & Watson, D. O. (1998). Past experience and behavioral choice among wilderness users. Journal of Leisure Research, 30 (2), 195. McGinn, A. P., Evenson, K. R., Herring, A. H., & Huston, S. L. (2007). The relationship between leisure, walking, and transportation activity with the natur al environment. Health & place, 13 (3), 588 602. Mehmetoglu, M. (2007). Typologising nature based tourists by activity Theoretical and practical implications. Tourism management, 28 (3), 651 660. Monroe, M. C., Nelson, K. C., & Payton, M. (2006). Communica ting with homeowners in the interface about defensible space. The public and wildland fire management: social science findings for managers. Gen. Tech. Rep. NRS 1. Newtown Square, PA: US Department of Agriculture, Forest Service, Northern Research Station 99 109. Montgomery, D. C., Jennings, C. L., & Kulahci, M. (2008). Introduction to time series analysis and forecasting Hoboken. New Jersey: John Wiley & Sons. Mooi, E., & Sarstedt, M. (2010). Cluster analysis A concise guide to market research (pp. 237 284): Springer.
155 Moore, R. L., & Driver, B. (2005). Introduction to outdoor recreation: Providing and managing natural resource based opportunities. Journal of Park and Recreation Administration, 25 (3), 113 115. Morgan, R., & Williams, A. (1999). Video pan orama assessment of beach landscape aesthetics on the coast of Wales. Journal of Coastal Conservation, 5 (1), 13 22. Napier, T. L., & Bryant, E. G. (1980). The Impact of Increasing Fuel Costs on Future Outdoor Recreation Participation Ohio State Universit y. National Park Service. (2016, 11/01/2016). National trail system: general Information. Nau, R. F. (2017). What's the bottom line? How to compare models. Retrieved from https://people.duke. edu/~rnau/compare.htm Needham, M. D. (2010). Value orientations toward coral reefs in recreation and tourism settings: a conceptual and measurement approach. Journal of Sustainable Tourism, 18 (6), 757 772. doi:10.1080/09669581003690486 Needham, M. D., Vaske, J. J., Donnelly, M. P., & Manfredo, M. J. (2007). Hunting specialization and its relationship to participation in response to chronic wasting disease. Journal of Leisure Research, 39 (3), 413. Nicholls, S., Holecek, D. F., & Noh, J. (2008). Impact of weather variability on golfing activity and implications of climate change. Tourism Analysis, 13 (2), 117 130. Nyaupane, G. P., & Andereck, K. L. (2008). Understanding travel constraints: Application and extension of a leisure constraints model. Journal of travel Research, 46 (4), 433 439. Nyaupane, G. P., White, D. D., & Budruk, M. (2006). Motive based tourist mar ket segmentation: An application to native American cultural heritage sites in Arizona, USA. Journal of Heritage Tourism, 1 (2), 81 99. Palacio, V. (1997). Identifying ecotourists in Belize through benefit segmentation: A preliminary analysis. Journal of S ustainable Tourism, 5 (3), 234 243. Park, D. B., & Yoon, Y. S. (2009). Segmentation by motivation in rural tourism: A Korean case study. Tourism management, 30 (1), 99 108. Paudyal, R., Poudyal, N. C., Bowker, J. M., Dorison, A. M., Zarnoch, S. J., & Green G. T. (2015). A value orientation approach to assess and compare climate change risk perception among trout anglers in Georgia, USA. Journal of Outdoor Recreation and Tourism, 11 22 33. Paudyal, R., & Stein, T. V. (2014). Segmenting FNST visitors based on factors important in choosing leisure destination Paper presented at the SAF National Convention, October 8 11, 2014., Salt Lake City, Utah.
156 Pergams, O. R., & Zaradic, P. A. (2006). Is love of nature in the US becoming love of electronic media? 16 ye ar downtrend in national park visits explained by watching movies, playing video games, internet use, and oil prices. Journal of Environmental Management, 80 (4), 387 393. Pett, M. A., Lackey, N. R., & Sullivan, J. J. (2003). Making sense of factor analysi s: The use of factor analysis for instrument development in health care research Thousand Oaks, CA: Sage Publications. Ramirez, E. M., & Ober, H. K. (2014). Nest site selection and reproductive success of red cockaded woodpeckers in Ocala National Forest. The American Midland Naturalist, 171 (2), 258 270. Rangun, V., & Karim, S. (1991). Teaching note: Focusing the concept of social marketing. Cambridge, MA: Harvard Business School Ribe, R. (1999). Regeneration harvests versus clearcuts: public views of t he acceptability and aesthetics of Northwest Forest Plan harvests. Ribe, R. (2002). Is scenic beauty a proxy for acceptable management? The influence of environmental attitudes on landscape perceptions. Environment and Behavior, 34 (6), 757 780. Richardso n, R. B., & Loomis, J. B. (2004). Adaptive recreation planning and climate change: a contingent visitation approach. Ecological Economics, 50 (1 2), 83 99. doi:10.1016/j.ecolecon.2004.02.010 Rideout, S., Oswald, B. P., & Legg, M. H. (2003). Ecological, poli tical and social challenges of prescribed fire restoration in east Texas pineywoods ecosystems: a case study. Forestry, 76 (2), 261 269. Ripley, B., Venables, B., Bates, D. M., Hornik, K., Gebhardt, A., Firth, D., & Ripley, M. https://cran.r project.org/web/packages/MASS/MASS.pdf Rothfusz, L. P. (1990). The heat index equation (or, more than you ever wanted to know about heat index). Fort Worth, TX: Scientific Services Division, National Weather Service Southern Region Headquarters. Ruddell, E. J., Gramann, J. H., Rudis, V. A., & Westphal, J. M. (1989). The psychological utility of visual penetration in near view forest scenic beauty models. Environment and B ehavior, 21 (4), 393 412. Ryan, R. L. (2005). Social science to improve fuels management: A synthesis of research on aesthetics and fuels management Gen. Tech. Rep. NC 261 (pp. 58). St. Paul, MN: U.S. Department of Agriculture, Forest Service, North Centra l Research Station.
157 Scott, D., Gssling, S., & de Freitas, C. R. (2008). Preferred climates for tourism: case studies from Canada, New Zealand and Sweden. Climate Research, 38 (1), 61 73. Scott, D., & Jones, B. (2007). A regional comparison of the implicat ions of climate change for the golf industry in Canada. The Canadian Geographer/Le Gographe canadien, 51 (2), 219 232. Scrogin, D. O., & Berrens, R. P. (1999). Determinants of Lottery participation for big game hunting privileges: resident versus nonresid ent Elk hunters in New Mexico. Human Dimensions of Wildlife, 4 (1), 36 49. Sevenant, M., & Antrop, M. (2009). Cognitive attributes and aesthetic preferences in assessment and differentiation of landscapes. Journal of Environmental Management, 90 (9), 2889 2 899. Shih, C., Nicholls, S., & Holecek, D. F. (2009). Impact of weather on downhill ski lift ticket sales. Journal of travel Research, 47 (3), 359 372. Smith, A., Tuffin, M., Taplin, R., Moore, S., & Tonge, J. (2014). Visitor segmentation for a park syste m using research and managerial judgement. Journal of Ecotourism, 13 (2 3), 93 109. Soule, P. T. (2005). A comparison of 30 yr climatic temperature normals for the southeastern United States. southeastern geographer, 45 (1), 16 24. Spagnolo, J., & De Dear, R. (2003). A human thermal climatology of subtropical Sydney. International Journal of Climatology, 23 (11), 1383 1395. Stanton, E. A., & Ackerman, F. (2007). Florida and climate change: the costs of inaction. Florida and climate change: the costs of inac tion Steiger, R., Abegg, B., & Jnicke, L. (2016). Rain, rain, go away, come again another day. Weather preferences of summer tourists in mountain environments. Atmosphere, 7 (5), 63. Stein, T. V., & Anderson, D. H. (2002). Combining benefits based management with ecosystem management for landscape planning: Leech Lake watershed, Minnesota. Landscape and Urban Planning, 60 (3), 151 161. Stein, T. V., & Lee, M. E. (1995). Managing recrea tion resources for positive outcomes: an application of benefits based management. Journal of Park and Recreation Administration, 13 (3), 52 70. Tahvanainen, L., Tyrvinen, L., Ihalainen, M., Vuorela, N., & Kolehmainen, O. (2001). Forest management and pub lic perceptions visual versus verbal information. Landscape and Urban Planning, 53 (1), 53 70.
158 Tervo, K. (2008). The operational and regional vulnerability of winter tourism to climate variability and change: The case of the Finnish nature based tourism en trepreneurs. Scandinavian Journal of Hospitality and Tourism, 8 (4), 317 332. Thapa, B., Holland, S. M., & Absher, J. D. (2008). Perceived risk, attitude, knowledge, and reactionary behaviors toward wildfires among Florida tourists Retrieved from Albany, CA: Toman, E., & Shindler, B. (2006). Communicating the wildland fire message: influences on knowledge and attitude change in two case studies. In P. L. Andrews & B. W. Butler (Eds.), Fuels Management How to Measure Success: Conference Proceeding Portlan d, OR: USDA Forest Service, Rocky Mountain Research Station. Toman, E., Stidham, M., McCaffrey, S., & Shindler, B. (2013). Social science at the wildland urban interface: A compendium of research results to create fire adapted communities Gen. Tech. Rep. N RS 111 Newtown Square, PA, USA.: USDA Forest Service, Northern Research Station. http://cran .utstat.utoronto.ca/web/packages/tseries/tseries.pdf US Fish and Wildlife Service. (2015). Red cockaded Woodpecker Recovery. Retrieved from http://www.fws.gov/rcwrecovery/rcw.html USDA Forest Service. (1995). Landscape Aesthetics: A Handbook for Scenery Management Agriculture Handbook Number 701 USDA Forest Service. (2013). National Forest Visitor Use Monitoring Training Manual National Visitor Use Monitoring Program. Uysal, M., & Hagan, L. A. R. (1993). Motivation of pleasure travel and tourism. In M. Khan, M. Olsen, & T. Var (Eds.), Encyclopedia of hospitality and tourism (pp. 798 810). Vaske, J. J. (2008). Survey research and analysis: Applications in parks, recreation and huma n dimensions. State College, Pennsylvania: Venture Publishing Wan, B., Paudyal, R., Huntley, C., & Stein, T. (2014). Florida National Scenic Trail Visitor Assessment. University of Florida: School of Forest Resources and Conservation, University of Flori da. Wan, B., Stein, T., Paudyal, R., & Colonna, A. (2014). Public views of prescribed fire: perceptions, concerns, and communication preferences Retrieved from University of Florida:
159 Warzecha, C. A., & Lime, D. W. (2001). Place attachment in Canyonlands National Park: Visitors' assessment of setting attributes on the Colorado and Green Rivers. Journal of Park & Recreation Administration, 19 (1). Wickham, H., & Chang, W. (2015). ggplot2: An Implementation of the Grammar of Graphics. 2015. URL http://CRAN R project. org/package= ggplot2. R package version, 1 (1). Williams, K., & Cary, J. (2002). Landscape preferences, ecological quality, and biodiversity protection. Environment and Behavior, 34 (2), 257 274. Wilson, E., Nicol, F. Nanayakkara, L., & Ueberjahn Tritta, A. (2008). Public urban open space and human thermal comfort: the implications of alternative climate change and socio economic scenarios. Journal of Environmental Policy and Planning, 10 (1), 31 45. Yen, S. T., & Ada mowicz, W. L. (1993). Statistical properties of welfare measures from count data models of recreation demand. Review of Agricultural Economics, 15 (2), 203 215. Yoon, Y., & Uysal, M. (2005). An examination of the effects of motivation and satisfaction on d estination loyalty: a structural model. Tourism management, 26 (1), 45 56. Yu, G., Schwartz, Z., & Walsh, J. E. (2009). A weather resolving index for assessing the impact of climate change on tourism related climate resources. Climatic Change, 95 (3), 551 5 73. Zajonc, R. B. (1980). Feeling and thinking: Preferences need no inferences. American Psychologist, 35 (2), 151. Zawacki, W. T., Marsinko, A., & Bowker, J. M. (2000). A travel cost analysis of nonconsumptive wildlife associated recreation in the United States. Forest Science, 46 (4), 496 506. Zube, E. H., Sell, J. L., & Taylor, J. G. (1982). Landscape perception: research, application and theory. Landscape planning, 9 (1), 1 33.
160 BIOGRAPHICAL SKETCH R amesh is a natural resources social scientist and a nature photographer from Nepal. His research interest includes a broad spectrum of human dimensions of natural resources, recreation, and climate change. Over the past 12 years, Ramesh worked as a freelancer and researcher in projects that involved a large variety of stakeholders, including buffer zone forest users and farmers in Nepal to hunters and anglers in the United States, and diverse contexts, including climate change risk perception to recreation experience and ecotourism. Ramesh holds one ma forest r esources from University of Georgia and one other in environmental s cience from Nepal. During his last four years as a PhD student, Ramesh led and assisted in several projects related to outdoor recreation in Florida and ecotourism in Nepal. He also served as a co instructor and teaching assistant for variety of courses. After accomplishment of PhD, Ramesh aims for a career in academia and dreams to be a research University faculty in the future. Other than conducting research and t eaching, Ramesh has been involved in various service, leadership, and professional development positions. For instance, Ramesh served as treasurer of Forestry Graduate Student Organization (FGSO) from 2016 2017, coordinated Southeaster Recreation Researc h Conference (SERR) in 2016, and assisted to organiz e Forestry Graduate Student Symposium in 2017. Considering, his notable contribution to research, teaching, and service, Ramesh was awarded with 7. Nature and passion together with his career goal in academia.