Travel Model Blind Spots: The Importance of Understanding Special Markets Related to Visitors

Huntsinger and Ward 1 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 Travel Model Blind Spots: The Importance of Understanding Special Markets Related to Visitors Leta F. Huntsinger, PhD, PE ** Senior Professional Associate Travel Modeling Parsons Brinckerhoff Systems Analysis Group 434 Fayetteville Street, Suite 1500 Raleigh, NC 27601 Phone: 919-836-4086 Email: huntsinger@pbworld.com Kyle Ward Technical Specialist Parsons Brinckerhoff Systems Analysis Group 434 Fayetteville Street, Suite 1500 Raleigh, NC 27601 Phone: 919-836-4048 Email: warddk@pbworld.com ** Corresponding Author Submitted for Consideration for Publication and Presentation at the 95 nd Annual Meeting of the Transportation Research Board, Washington, DC, January 2016 Words = 5434 Tables and Figures = 7 (1750) Total = 7184 35 36 37

Huntsinger and Ward 2 38 39 40 41 42 43 44 45 46 47 48 49 50 51 ABSTRACT In many small and medium-sized communities, the need to reflect the unique characteristics of travel by special markets is an important consideration. Visitor travel is a common special market for many communities that enjoy the benefits of being a destination for visitors from outside of the region. While benefiting the region s economy, visitor travel can also impact the transportation system. Visitor travel is often not modeled explicitly due to the cost associated with collecting behaviorally rich survey data to support model development. This paper presents a low cost option for collecting travel behavior data for special markets through the administration of a small-sample intercept survey of persons staying in recreational vehicle (RV) campgrounds in the French Broad River region of North Carolina. Analysis of the survey data shows that travel by these surveyed RV households is different from both retired and seasonal part-time resident households captured in the regions household travel survey. The paper further discusses the development of a special market model to address travel by RV visitors in the region. 52

Huntsinger and Ward 3 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 INTRODUCTION AND MOTIVATION In many small and medium-sized communities, the need to reflect the unique characteristics of travel by special markets is an important consideration. A special market is different from what travel modelers often refer to as special generators. A special generator refers to a specialized land use such as a regional shopping center or hospital that has trip generation characteristics that cannot be reflected by the standard trip rates. A special market is much more comprehensive in nature and refers to a specialized land use such as a university or airport that has trip generation, distribution, time-of-day and/or mode choice characteristics that cannot be reflected by the models developed for other trip purposes. To better capture the unique travel characteristics of these special markets, the preferred approach is the collection of travel survey data and the development of a separate submodel. Visitor travel is a common special market for many communities that enjoy the benefits of being a destination for visitors from outside of the region. While benefiting the region s economy, visitor travel can also impact the transportation system. The French Broad River Metropolitan Planning Organization (FBRMPO) is one such community. In addition to the more common visitor travel where visitors stay at hotel type lodging facilities during their visit, the FBRMPO region is also home to over 30 recreational vehicle (RV) parks for a total of 3,000 sites were visitors to the region may stay in the region for either short- or long-term stays. Those long-term visitors, defined as those visitors that stay in the region for 31 days or longer, can often seem like an extension of the resident base by those communities that are home to the RV parks where they reside. In fact, during peak seasons certain communities experience a significant increase in demand on the public infrastructure, including the transportation system, as a result of what some might consider part-time or seasonal residents [1]. An important goal for these communities is to better understand those impacts, especially with respect to impacts on the transportation system and the identification and prioritization of transportation projects. This type of visitor travel is often not modeled explicitly in travel demand models due to the cost associated with collecting behaviorally rich survey data to support such model development. This does however limit a community s ability to understand how visitor travel impacts the transportation system and key transportation projects. Developing a better understanding of travel by RV park visitors was a key goal of the FBRMPO during the household survey data collection effort recently undertaken to support travel model development. One of the key assumptions the MPO wanted to test was whether persons visiting the region in RVs have trip rates and travel patterns similar to those of retired households in the region. In addition to developing a better understanding of travel characteristics, the MPO also wanted to explore the development of an enhancement to their travel model to account for RV visitor travel. This paper describes the collection of a small-sample intercept survey of RV visitors, the key findings from that survey effort, and finally describes the development of an RV trip model for the FBRMPO. The approach outlined in this paper offers a low cost alternative for understanding and modeling travel behavior for special markets with a specific focus on RV visitor travel. BACKGROUND The FBRMPO travel model covers the North Carolina counties of Buncombe, Haywood, Henderson, Madison, and Transylvania. The region is located in the mountains of North Carolina, and is home to Biltmore Estate, featuring 8,000 acres of gardens, forest conservation areas, and America s largest home,

Huntsinger and Ward 4 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 built by George Vanderbilt and opened to friends and family in 1895 [2]. The Biltmore property is now a National Historic Landmark that welcomes more than 1 million visitors each year [3]. The City of Asheville, located in the center of the MPO region, is well known for its culinary cool, beer scene, and burgeoning artist community [4]. The region is full of history and surrounded by natural beauty. In addition to the Biltmore Estate, other visitor attractions include the Great Smoky Mountains National Park, the Blue Ridge Parkway, Chimney Rock, the River Walk Arts District, and many others. Buncombe County and the City of Asheville alone welcome over 9.1 million visitors a year, with 3.1 million overnight guests [5]. This translates into $1.5 billion spent, $2.3 billion in total business sales, and 23,000 jobs supported [5]. Without tourism, the unemployment rate in Buncombe County would be 15.9% [6]. Statistics on the economic benefits of visitors is easy to find, but what is less clear is how visitor travel impacts the transportation system, and how this information can help identify and prioritize key transportation projects. The interest in understanding visitor travel is well documented, though much of the published research related to transportation impacts has focused on special venues; in particular, improving visitor access and experience at National Parks [7-9]. There is also a heavy focus on long-distance travel in statewide models as it relates to visitors [10-12]. This topic garnered such interest as to support a research project on transferable parameters for long-distance travel in statewide models funded by the National Cooperative Highway Research Program [13]. While these studies are useful, they do not contribute significantly to understanding the travel patterns of visitors to a region, nor do they offer insight into the development of travel models for regions that wish to better represent visitor travel. DATA DESCRIPTION To better understand travel by persons at RV parks, the FBRMPO commissioned a small sample survey of campers staying at RV parks within the region; the survey was conducted by Westat, Inc. during the month of July 2013. The primary goal of the data collection effort was to try and understand the travel impacts of RV parks on the transportation system during peak seasons of the year. The survey design was somewhat limited by available budget as the survey was considered an add-on to the full household survey effort. For this reason, the sample goals were set to try and achieve 50 100 completed surveys with a mix of long- and short-term visitors. Short-term visitors were defined as those staying at the park less than 31 days (one month). Because the MPO was specifically interested in the travel of long-term visitors as it relates to residents, the goals was to have a final sample consisting of 60-70 percent longterm campers. Survey Design and Data Collection Sampling Plan Staff from the FBRMPO provided a shape file of RV parks that was intended to form the basis of the sampling for the RV survey. All of the parks were located within Haywood County, the county with the highest concentration of parks demonstrating a long-term seasonal influx of visitors. With a small sample survey, MPO staff wanted to concentrate the survey effort in these parks in order to increase the probability of collecting data on long-term campers, the demographic they were most interested in, primarily to address the question of how different short- vs. long-term travel patterns are, and whether the long-term campers behave more like retired full time residents.

Huntsinger and Ward 5 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 A total of seven RV parks were selected from list provided by the MPO, see Table 1. Parks with fewer than 50 hookup sites were excluded from the sample in order to maximize limited resources and time constraints. The target sample size was 50 100 completed surveys. The sample was not a probability sample, and the survey was intended to be informative in nature. Table 1 includes the number of units at the park, expected short- and long-term campers, and the sampling goal. Westat contacted each site before visiting in order to explain the purpose of the survey and to gain permission to approach and survey campers. Information on park capacity during the time of visit was also obtained from the camp host. Westat assumed a fifty percent response rate and a fifty percent contact rate from which they developed a sampling strategy as noted in the table. Table 1 Sampled Parks and Selection Rates [14] Park Name Units Short-Term Long-Term # Sampled RVs Sample Strategy Butch Teague 50 0 50 41 Everyone in park Creekwood Farm 50 50 0 10 Cross Creek 53 27 26 10 Every 5 th occupied RV Pride Resorts 100 70 30 20 Riverside 60 6 54 49 Everyone in the park Stone Bridge 100 60 40 20 Every 5 th occupied RV Windgrey 60 6 54 49 Everyone in the park Total 473 219 254 200 In total, 70 interviews were completed representing 15.1 percent of the occupied lots [14]. Westat noted that the biggest challenge to achieving a higher sample rate was non-contact attributed to people being away from the site during the time of the interview, and campers who use their RV as a vacation home not actively using their RV during the time of the survey. Table 2 provides a breakdown of all respondents surveyed. Table 2 Visitor Type Summary by Campground [14] Park Name Completes Short- % of Long- % of Site Indefinite % of Site Term Site Term Stay Butch Teague 11 6 54.5% 4 36.4% 1 9.1% Creekwood 5 3 60% 2 40% 0 NA Farm Cross Creek 10 0 0.0% 10 100.0% 0 0.0% Pride Resorts 13 6 46.2% 6 46.2% 1 7.7% Riverside NA NA NA NA NA NA NA Stone Bridge 12 8 66.7% 4 33.3% 0 0.0% Windgrey 19 2 10.5% 17 89.5% 0 0.0% Total 70 25 35.7% 43 61.4% 2 2.9%

Huntsinger and Ward 6 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 Survey Instrument The survey was designed to capture key elements of travel similar to those captured in a household travel survey, but also to better understand the purpose, frequency, and duration of travel. Questions related to the primary trip to the region include: Date of arrival Length of stay Frequency of travel to the region Purpose of travel to the region Number of people in the travel party Number of people in the travel party under 18 years of age Demographic variables were also collected for each travel party surveyed, including: Respondent: o Age o Employment status o Student status o Gender Annual household income Trip data included: Trip origin Trip destination Time of departure Time of arrival Mode of transportation (walk, bike, personal/rental vehicle, motorcycle/scooter, public transportation, taxi, sightseeing bus, other) Number in travel party Trip purpose (shopping, recreational, food/drink, visiting friends/relatives, personal business, work, other) Approach All surveys were conducted by trained surveyors using a simple paper questionnaire and face-to-face interviews. The survey was conducted as a 24-hour retrospective survey where respondents were asked the trips they made the previous day. The survey was only administered to one member of the camping party, though data was collected on the number of persons staying in the RV and the number of persons in the trip party. Survey Data Analysis This section provides summary statistics of the RV survey data. One of the key purposes of this data collection effort was to try and develop an understanding of whether travel by RV visitors is similar to the travel of retired households captured in the full household travel survey. To answer that question, comparisons of various travel statistics are made against retired and seasonal (or part-time) households captured in the full household travel survey.

Huntsinger and Ward 7 209 210 211 212 213 RV survey participants were asked about their primary purpose of travel to the region, survey responses are summarized in Figure 1. Seventeen percent of the participants identify their RV as a second home, and another twenty-nine percent identify retired and enjoy the area as the primary reason for their visit. 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 Figure 1 Primary Reason for Travel to the Region Sixty-nine of survey respondents identified themselves as long-term campers, ninety percent of respondents were retired and eighty three percent were over 65 years of age. There was a fifteen percent non-response rate for the question on household income, but of those who did report income, fifty-four percent reported an annual income of less than 60,000 dollars. Four of the RV visitors surveyed had just arrived in the area the day before (e.g. the reported travel day for a retrospective survey), these records were dropped from the analysis. Of the remaining RV visitors, fifteen (twenty two percent) reported no travel on the travel day. These records were included in the analysis to reflect the fact that not all visitors travel every day. Analysis of the trip data shows that RV visitors make an average of 5.05 trips per day, with home-based shopping (HBSHP) and non-home-based (NHB) trips the highest at 1.71 and 1.82 average trips per day, respectively. RV visitors who identified their primary travel purpose to the region as vacation had a slightly higher average daily trip rate (5.63) than those RV visitors in the region for other purposes (4.26). The average trip rate for long-term campers (5.60) is slightly higher than that for short-term campers (4.26), with the biggest difference in the HBSHP trip rates for the long-term campers. Table 3 provides a comparison of average trip rates by RV visitors and retired and part-time resident households from the full household travel survey. The average daily trip rate for RV visitors is lower than that estimated for either retired households or part-time residents. The average daily trip rate for part-time residents is higher than that for both the RV visitors and retired households, though the biggest difference is between the part-time households and the RV visitors. The HBSHP trip rate for RV visitors is much higher than that for both retired and part-time households, perhaps reflecting the fact that many RVs have less room for storing groceries and other household supplies, necessitating more frequent shopping trips. Retired households make the highest average number of HBO trips, perhaps reflecting a greater integration into the community and travel necessary to maintain these community connections.

Huntsinger and Ward 8 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 Table 3 Average Trip Rates by Household Type Household Type HBW HBSHP HBO NHB Total RV - 1.71 1.52 1.82 5.05 Retired 0.25 0.95 3.23 2.49 6.93 Part-Time Residents 0.49 0.96 2.93 2.86 7.23 The biggest difference in travel characteristics between the different household types is seen in the average trip length in miles. Table 4 shows the average trip length in miles by household type and trip purpose. The average trip length in miles is lowest for retired households at 6.89 miles compared to 8.79 for part time residents and 8.70 for RVs. The average trip length for HBSHP and HBO is highest for RVs likely reflecting the remote nature of the majority of the RV campgrounds as compared to retired and part time resident households that are scattered throughout the region. Table 4 Average Trip Length in Miles by Household Type and Trip Purpose Household Type HBW HBSHP HBO NHB Total RV (n=67) NA 10.90 11.18 4.71 8.70 Retired (n=707) 7.33 6.67 7.00 6.79 6.89 Part-Time Residents (n=69) 8.12 8.91 10.89 6.72 8.79 The small sample size for the RV survey does limit the disaggregation of the data for fully understanding travel behavior or for estimating models, but it also offers some interesting insights and comparisons against retired and part-time households in the general population. The observed variations in trip rates and trip lengths support the development of a separate model component to better capture travel for RV visitors to the region. RV TRIP MODEL This section describes the development of the RV trip model for the FBRMPO region using the small sample intercept RV survey described in the previous section. Model Development The RV trip model follows a traditional 3-step process of trip generation, trip distribution, and highway assignment by time of day. There is no mode choice as all trips from the RV parks are assumed to be vehicle trips given the remote nature of these parks and the lack of transit access. This is supported by the survey data that indicated that all trips off site were made by a personal vehicle. The vehicle trips are split by time of day and assigned with other trip types during the final assignment. There are 32 RV parks across the FBRMPO region for a total of 3,000 sites. The RV survey data revealed that the majority of the visitors of the parks are seasonal, staying for a month or longer, rather than the multiple-day or weekly average common for most visitors staying in hotels or other rental lodgings. For survey purposes, each RV was treated as a household. Survey respondents reported a total of 182 trips, equating to 338 unexpanded trips when accounting for party size. Expansion factors were developed for the survey records using the total number of RV sites and the occupancy rate for the survey period. Using the address information provided for each place visited, the survey records were geocoded

Huntsinger and Ward 9 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 to traffic analysis zones (TAZs) and processed into trip records with data fields for production zone, attraction zone, trip purpose, and number of people in the travel party. The trip purposes considered include home-based shopping (HBSHP), home-based other (HBO), and non-home-based (NHB). Trip Generation The average trip production rate by trip purpose is summarized below: HBSHP: 1.7 person trips per RV household HBO: 1.5 person trips per RV household NHB: 1.8 person trips per RV household Trip attraction rates were estimated using the expanded trip end data from the RV survey and the employment data in each TAZ. Trip attraction rates by trip purpose are summarized below: HBSHP: Retail = 0.318 HBO: Office and Service = 0.264 NHB: Office, Service, and Retail = 0.157 Given the limitations in the survey data, no attempt was made to stratify the trip rates other than by trip purpose. To estimate trip productions by TAZ, the trip rate by trip purpose is applied to each occupied RV household in a given TAZ. See Equation 1, where HBOP i the HBO productions in zone i, RVTT i is the number of recreation vehicle households in zone i, and HBOP rate is the HBO trip production rate. The data on occupancy rates was obtained from the RV parks surveyed. HBOP i = RVHH i (HBOP rate ) (1) Trip attractions by TAZ are estimated by applying the trip attraction rates by trip purpose and employment type to the employment data within each TAZ in the region. See Equation 2 where HBOA j is the estimated HBO attractions in zone j, EMP xj is the total number of employees of type x in zone j, and HBOA ratex is the HBO attraction rate for employment type x. HBOA j = EMP xj (HBOA ratex ) (2) Total attractions are balanced to productions prior to trip distribution. Trip Distribution The trip distribution model "connects" the independent productions and attractions estimated by the trip generation model based upon a quantitative description of the relative difficulty in reaching each potential destination zone from an origin zone and an understanding of the underlying functional relationship between these variables (i.e., productions, attractions, and impedance). In other words, RV households are influenced by the attractiveness of the destination (as measured by the estimated quantity of trip attractions), but also tend toward selecting the first destination which satisfies the purpose of the trip. Due to this behavior, the gravity model tends to over-estimate visitor trip lengths given the intrinsic limits associated with the functional form of the F ij factors (friction factors between zone i and zone j). The

Huntsinger and Ward 10 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 intervening opportunities model, however, possesses this first-satisfied attribute in its functional form. The premise governing the intervening opportunities model states that total impedance is minimized subject to the condition that every destination has a finite probability of being accepted, subject to the desire that each trip be as short as possible. The intervening opportunities model generally has a steeper decay slope to its distribution function, which leads to shorter trips. The mathematical form of the model is as follows [15]: where: T ij = T i (e L R j 1 e L R j) (3) T ij = number of trips from zone i to zone j T i = number of trips produced in zone i R j = rank of destination zone j R j-1 = rank minus 1 of destination zone j L = probability of accepting a destination if it is considered. Earlier applications of this model for the distribution of visitor trips showed promising results against observed visitor travel survey data, leading to the selection of this model to distribute RV household trips [16]. The RV survey data was processed to create observed trip tables by trip purpose, and these trip tables paired with network skims were used to develop observed trip-length frequency distributions. The intervening opportunities model was applied with an initial probability parameter, λ, of -0.225 for each trip purpose. The model estimated trip length distributions were compared to the observed distributions as discussed in more detail in the following section. Time of Day The last piece of observed behavior that was assembled from RV survey to support model development was the directional split of trips by time of day. These factors are needed to convert the daily productionattraction person trip tables to origin-destination vehicle trip tables by time of day. According to the survey, the average persons per RV trip are 1.88 persons per trip. The directional split by time of day is provided in Table 5. In comparison to the time of day factors for residents in household survey, RV households are much more likely to make trips during the middle of the day, and much less likely to make trips during the AM or at night. Table 5 RV Time of Day Directional Split Factors Purpose Direction AM MD PM NT Daily HBS From Home 0 0.423 0.058 0.019 0.50 To Home 0 0.206 0.237 0.057 0.50 HBO From Home 0.097 0.323 0.048 0.032 0.50 To Home 0.022 0.304 0.087 0.087 0.50 NHB From Home 0 0.39 0.095 0.015 0.50 To Home 0 0.39 0.095 0.015 0.50

Huntsinger and Ward 11 359 360 361 362 363 364 365 366 367 368 369 370 371 372 373 374 375 376 377 378 379 380 381 382 383 384 385 386 387 388 389 390 391 392 393 394 Analysis and Results In addition to the rates and parameters discussed in previous sections, the RV trip model requires as input the number of RV park hookups and campsites by TAZ and a seasonal occupancy factor for spring, summer, fall, and winter. The base year model was applied using the spring occupancy factor, which was determined to be 0.59 from the survey data collection effort. This occupancy factor reflects the percent of the sites that are occupied during the specified time of year. Changing this value will have the effect of either increasing or decreasing the number of trips produced by the RV trip model as only occupied sites generate trips. Trip distribution for the RV trip model was estimated using the intervening opportunities model and estimated average trip lengths and trip length distributions were compared to observed values. The NHB trip model performs very well in comparison to the survey data, 5.4 minutes estimated and 6.4 minutes observed. However, the HBS and HBO modeled average trip length trip, 30.9 and 29.6 minutes respectively, is much longer than the observed values of 15.3 and 15.9 minutes, respectively. Multiple model iterations were performed with a range of values on the probability parameter, λ, in an effort to improve model fit. The resulting change in the model average trip length was negligible. This pointed to a number of possibilities that required testing: either the intervening opportunities model was not a good model for these two trip purposes, or the distribution of available activities at a regional level did not support the shorter trip lengths reported by survey participants. The investigation into the locations of the RV parks, both surveyed and non-surveyed, in comparison to the location and magnitude of available activities provided a clear explanation of the discrepancy between the survey and the model application. Figure 2 shows a plot of the modeled HBS RV productions (blue) and modeled HBO RV productions (green). In addition, red and orange colors show RV attractions by trip purpose, which are proportional to the employment in each zone. Finally, the triangle in Figure 3 shows the location of the TAZs where RV parks were surveyed. Recall from the earlier discussion that the surveys were concentrated in the Haywood County parks to increase the probability of collecting data on long-term campers and to minimize the cost of data collection. The surveyed parks are located much closer to potential activities than are many of the RV parks not included in the survey. It is therefore logical that longer travel times would be required to participate in HBS and HBO activities from many of these more-remote parks. Based on this analysis, the modeled trip length distributions for HBS and HBO were allowed to exceed the observed survey data.

Huntsinger and Ward 12 395 396 397 398 399 400 401 402 403 404 405 406 407 408 409 410 411 412 413 414 415 416 417 418 419 420 421 422 FIGURE 2 Modeled RV Trip Productions and Attractions The RV model estimates approximately 6,000 vehicle trips per day in the FBRMPO region, though this number will be much higher during peak seasons. The analysis shows that the RV households in the region are similar to seasonal and retired households with respect to trip purposes, trip rates, and locations visited, though there are a few notable differences. RV households make fewer trips on average than the other two household types and the average trip rate for shopping trips is almost twice that of retired and part-time resident households. Based on the collected survey data, they are less likely to visit key visitor attractions in the region, which runs counter to what would be expected of visitors staying in traditional lodging establishment such as hotels or motels. The travel patterns of RV households also differ from that of retired and part-time resident households with respect to average trip length and travel by time of day. The analysis summarized in this paper has shown the benefits of collecting a small sample intercept survey for understanding the travel behavior of special populations. The analysis has also shown the benefit of modeling the RV households as a special market. Not including these households in the model will have the effect of underestimating travel demand. Modeling RV households as a component of resident travel would have the effect of overestimating travel by specific trip purposes, and by extension to particular locations, as well as overestimating travel during the AM and PM peak periods. SUMMARY AND CONCLUSIONS The work described in this paper demonstrates the use of a small sample intercept survey of long- and short-term campers for understanding travel behavior for recreational vehicle (RV) visitors in the French Broad River Metropolitan Planning Organization (FBRMPO) region of North Carolina. This data collection effort provided insights into the unique travel characteristics of RV households in comparison to travel characteristics observed from household survey data for retired households and seasonal part-

Huntsinger and Ward 13 423 424 425 426 427 428 429 430 431 432 433 434 435 436 437 438 439 440 441 442 443 444 445 446 447 448 449 450 451 452 453 454 455 456 457 458 459 460 461 462 463 464 465 466 time residents. In general the total trip rate per household is lower for RVs as compared to retired households and part time resident households. The trip rates by trip purpose also varies between household types, as does the average trip length by trip purpose and the percentage of trips made by time of day. The survey was administered with very little startup cost and followed an approach that could easily be transferred to other regions with a desire to understand travel by special markets. The simple design and administration of the survey lends itself well to implementation by agency staff or trained interns in order to keep survey costs low. The work documented further describes the development of a special market model for estimating travel for RV households. The model includes submodels for trip generation, trip distribution, and time of day all developed from the observed survey data. Application of the model results in approximately 6,000 vehicle trips in the region assuming an average RV park occupancy rate of 59 percent. The model allows the FBRMPO to evaluate different occupancy rates in order to evaluate changes in travel demand by season. The RV trip model accounts for approximately 0.4 percent of the daily trips in the case study region, though the impact is higher in the subareas where the RV parks exist. During peak capacity, this value could climb as high as 1 percent. As a point of comparison, the transit share in the region is 0.3 percent, though much higher for the parts of the region served by transit. Even for large urban areas, the regional transit share is often less than 2 percent, and yet the investment in mode choice models is rarely disputed. Special market models are not needed for all communities, just as mode choice models are not needed for all communities. However, for communities where special markets are important, collecting a small sample survey and adding this model component provides local planners with a better tool for understanding the complete picture of traffic flows in and around the region while also providing a tool to measure the seasonal impacts of visitor-related travel. ACKNOWLEDGEMENTS The authors would like to acknowledge Pam Cook with the North Carolina Department of Transportation, Paul Black, Lyuba Zuyeva and Vicki Eastland at the French Broad River Metropolitan Planning Organization and Westat, Inc. for the use of the recreation vehicle survey and all supporting data. REFERENCES 1. Black, P., Director FBRMPO. 2012: Asheville, NC. 2. Biltmore. Estate History. 2015 [cited 2015 July 2]; Available from: http://www.biltmore.com/. 3. Biltmore. Biltmore Thrives as a National Brand. 2015 [cited 2015 July 3]; Available from: http://www.biltmore.com/media/newsarticle/biltmore-thrives-as-a-national-brand. 4. AshevilleAreaChamberofCommerce. Asheville, Discovery Inside and Out. 2015 [cited 2015 June 16]; Available from: http://www.exploreasheville.com/. 5. TourismEconomics. The Economic Impact of Tourism in Asheville, North Carolina. 2012 [cited 2015 June 16]; Available from: http://www.ashevillecvb.com/. 6. AshevilleConventionandVisitorsBureau. Economic Impact. 2015 [cited 2016 June 16]; Available from: http://www.ashevillecvb.com/economic-impact/. 7. Lawson, S., et al., Modeling the Effects of Shuttle Service on Transportation System Performance and Quality of Visitor Experience in Rocky Mountain National Park.

Huntsinger and Ward 14 467 468 469 470 471 472 473 474 475 476 477 478 479 480 481 482 483 484 485 486 487 488 489 490 491 492 Transportation Research Record: Journal of the Transportation Research Board, 2011. 2244: p. 97-106. 8. Lawson, S., et al., Integrated Transportation and User Capacity Research in Yosemite National Park. Transportation Research Record: Journal of the Transportation Research Board, 2009. 2119: p. 83-91. 9. Nelson, B., V. Taylor, and J. Nabti, Shifting Modes of Travel to National Parks: Pilot Study at Muir Woods National Monument, California. Transportation Research Record: Journal of the Transportation Research Board, 2008. 2077: p. 166-174. 10. LaMondia, J., L. Aultman-Hall, and E. Greene, Long-Distance Work and Leisure Travel Frequencies. Transportation Research Record: Journal of the Transportation Research Board, 2014. 2413: p. 1-12. 11. Moeckel, R., R. Fussell, and R. Donnelly, Mode Choice Modeling for Long-Distance Travel, in 92nd Annual Meeting of the Transportation Research Board, TRB, Editor. 2013: Washington, DC. 12. Xiao, S., R. Yu, and Y. Wang, A Modeling Framework for Long Distance Pleasure Travel Supply Analysis, in 92nd Annual Meeting of the Transportation Research Board, TRB, Editor. 2013: Washington, DC. 13. TRB, Long-Distance and Rural Travel Transferable Parameters for Statewide Travel Forecasting Models, TRB, Editor. 2012, National Academy of Sciences: Washington, D.C. 14. Westat, Western North Carolina Household Travel Survey Final Report. 2013, French Broad River Metropolitan Planning Organization. 15. Ortuzar, J.d.D. and L.G. Willumsen, Modeling Transport. 4th ed. 2011: John Wiley and Sons, Ltd. 16. ParsonsBrinckerhoff, Non-Resident Visitor Model Estimation & Calibration. 2000, Regional Transportation Commission of Clark County Nevada. 493 494 495