AN INTERAC ATTENDAN PARK To evaltx reation areas must be QStimc is to design . of these value actaristics a: PIOblem is to Camper attenda Three staps an The fit: Park in terms Three attractj IQSOuICes' 2) 3) facilities quality 0f the ABSTRACT AN INTERACTION TRAVEL MODEL FOR PROJECTING ATTENDANCE OF CAMPERS AT MICHIGAN STATE PARKS: A STUDY IN RECREATIONAL GEOGRAPHY by Carlton S. Van Doren To evaluate existing and potential outdoor rec- reation areas, a number of economic and social values must be estimated. A primary objective of this work is to design a suitable method for approximating some of these values on the basis of recreation area char- acteristics and the desires of the users. The specific problem is to understand the spatial variation in camper attendance at Michigan State Parks in 1964. Three steps are included in the analysis. The first step involves classifying each state park in terms of its attraction as a trip destination. Three attractive qualities are considered: 1) natural resources, 2) outdoor activity Opportunities, and 3) facilities and services available. The extent and quality of the variables included in these three ele- ments are combined into an attraction index for each state paz ing routi camping a as well a: allurement The a consider; quently uti provided by regression a between atte attraction .1 ity of the a1 camPing Satig A third attraction in recreatimal ‘ model is utilj component or the “Wilhel- o f eightY-eight . time betVeen tions. A She Carlton S. Van Doren state park by the use of factor analysis and a scor- ing routine using the factor loadings. The index of camping attraction provides a classificatory system as well as a measure of the camping opportunities and allurement each park holds for campers. The camping attraction index is derived without a consideration of visitation rates, a measure fre- quently utilized as an estimate of the satisfaction provided by a recreation site. A second step uses regression analysis to compare the relationship between attendance at parks (camper-days) and the attraction indices. This test substantiates the valid- ity of the attraction indices as a measure of potential camping satisfaction. A third and final step includes the use of the attraction indices as one of three components in a recreational travel model. An interaction or gravity model is utilized. In addition to the destination component or attraction indices, other components are the number of camper-days originating from each of eighty-eight origin nodes and the automobile travel- time between the origins and fifty-nine park destina- tions. A successful replication of the system is achieved data is a‘ The the campin dictive toc ing or prop insight int: 0f the state Carlton S. Van Doren achieved for the one year, 1964, for which complete data is available. The travel model and its specialized component, the camping attraction indices, can be used as a pre- dictive tool for estimating the camping use of exist- ing or pr0posed recreational sites as well as providing insight into the site and situational characteristics of the state parks. AN INTERACT I CAMPER. 1n partia .AN INTERACTION TRAVEL MODEL FOR PROJECTING ATTENDANCE OF CAMPERS AT MICHIGAN STATE PARKS: A STUDY IN RECREATIONAL GEOGRAPHY by Carlton S. Van Doren A THESIS Submitted to Michigan State University in partial fulfillment of the requirements for the degree of DOCTOR OF PHILOSOPHY Department of Geography 1967 I wish to so: Donald A. criticism in t 5 Similar eXpr my adViSOl’y co and L. M. Somm- PIOfessor D, 5 and the Encourc a recreational Professors Mil Resource Delec the financial 6 Outdoor Recteat ACKNOWLEDGMENTS I wish to express my thanks to my advisor, Profes- sor Donald A. Blome, for his support, advice, and criticism in the preparation of this dissertation. A similar expression of gratitude is due the members of my advisory committee, Professors D. H. Brunnschweiler and L. M. Sommers. An acknowledgment is also due to Professor D. N. Milstein for his many helpful suggestions and the encouragement he gave me in my attempt to create a recreational travel model. I am deeply grateful to Professors Milstein and L. M. Reid of the Department of Resource DevelOpment, Michigan State University, for the financial assistance extended to me from the Michigan Outdoor Recreation Demand Study. In addition, I am further indebted to D. L. Gibson, director of the Institute for Community Development and Services, Michigan State University, for the financial aid received during my tenure as a graduate student. Without the faith and confidence of these gentlemen, this work would not have been completed. ii For help thanks should fully read my Columbus, Ohi Pbral and fin Professor Edw. As a close €01 ability to ma] the final dra; Preparation 0: “WHY. for her fort“ For help during the latter phases of my writing, my thanks should be extended to Mrs. Ree Sheck, who care- fully read my drafts, and to Mrs. William Neeld of Columbus, Ohio, who so capably typed the dissertation. Moral and financial support was also received from Professor Edward J. Taaffe of The Ohio State University. As a close confidant during the last stages, he had the ability to make me smile while in the throes of preparing the final draft. Mr. Charles E. Trott assisted in the preparation of the illustrations. Finally, I am forever grateful to my wife, Sharon, for her fortitude, patience, and encouragement throughout my graduate program. iii Acszxomsoomn exams . . LIST 0? TABLE: LIST OF mus ; CONTENTS ACKNOWLEDGMENTS. . . . . . . . CONTENTS 0 O O O O O O O O O O LIST OF LIST OF CHAPTER I. II. III. IV. V. VI. TABI‘ES O O O O O O O O ILLUSTRATIONS. . . . . INTRODUCTION . . . . . APPRAISAL OF THE ELEMENTS OF SITE AND SITUATION FOR RECREATIO NAL AREAS . . . COMPONENTS OF AN INTERACTION MODEL FOR MICHIGAN STATE PARKS . THE INDEX OF CAMPING.AT EMPIRICAL VERIFICATION TIONAL TRAVEL MODEL. . CONCLUSIONS. . . . . . APPENDIX A O O O O O O O O O O APPENDIX B O O O O O O O O O O APPENDIX C O O O O O O O O O O APPENDIX D O O O O O O O O O O BIBLIOGRAPHY O O O O O O O O O iv TRACTION. . . . OF THE RECREA- Page ii iv vii 12 31 81 123 166 177 184 210 228 256 Table Per ce: partic; recrea1 and No: Compari outdooz non-can Correla tion in camping N'umeric; land ac1 State Pa Fifty-f1“ fBCtor a Faetor factor Summary ”Odels, Factor 1 factor 5 ForMUla WRAP pro Table 3. 8. 9. LIST OF TABLES Per cent of persons 12 years and over participating in various outdoor recreation activities, United States and North Central Region, Summer, 1960. Comparison of participation in other outdoor activities by campers and non-camper800000000000000 Correlation coefficients of participa- tion in selected activities with camping participation, Summer, 1960 . . Numerical activity weights by water and land activities assigned to Michigan State Parks O O O O O O O O O O O O O O Fifty-five variables used in first factor analysis model . . . . . . . . . Factor loadings for first model four- factor solution (55 variables). . . . . Summary table of factor analysis mdelsO O O O O O O O O O O O O O O O O Factor loadings for Model Five, four- factor solution (43 variables). . . . . Formula and factor loading weights for WRAP ProgramO O O O O O O O O O O O O O Page 36 4O 44 53 84 88 91 97 Table 10. ll. 12. 13. 14. 15. 16. 17. 18, 19, St Ad; Rev rat Table 10. ll. 12. 13. 14. 15. 16. 17. 18. 19. LIST OF TABLES - CONTD. Page Standardized park scores. . . . . . . . . .' 103 'AdjuSted park Scores. 0 o o o o o o o o o o 107 Revised index of attraction with campsite ratio O O O O O O O O O O O O O O O O O O O 112 Error measures for select interact models 139 Predicted camper-days from Model Seven compared to 1964 attendance . . . . . . . . 158 Per cent of activity participation prefer- ences from personal interviews and volun- tary responses - Michigan State Parks - Sumner,l955................179 Most frequent suggestions for park improve- ment, Michigan State Park Users Survey, 1956. O O O O O O O O O O O O O O O O O O O 214 Park inventory tables 0 o o o o o o o o o o 229 Variables inventoried for Michigan State Parks, including possible range of values, number of parks with non-zero value, and sourceOfdataooooooooooooooo234 Estimated origin camper-days, 1962, and actual 1964 camper-days . . . , , , . . . . 249 vi Figure Mic? Camp Camp Park: Intez acti\ servi Trave Michi Campy ratio‘ Cdfllpin ratio, Resid dEViat Michi 1964, Camper 196m 5 LIST OF ILLUSTRATIONS Page Michigan State Parks, 1964. . . . . . . . 5 Camper-days, Michigan State Parks, 1964 . 7 Camper-days by origins, Michigan State Parks, 1964 . . . . . . . . . . . . . . . 8 Interrelationships of natural resources, activity preferences and facilities and 881171083. O O O O O O O O O O O O O O O O 34 Travel times between origin nodes and Michigan State Parks. . . . . . . . . . . 78 Camping attraction index without campsite ratio, MiChigan State Parks, 1964 o o o o 110 Camping attraction index with campsite ratio, Michigan State Parks, 1964 . . . . 115 Residuals from regression - relative deviation of camping attraction indices, Michigan State Parks with camper-days, J'964O O O O O O O O O O O O O O O O O O O 117 Camper-days by hourly increments to Mich- igan State Parks from.Wayne County, mohigan' 1964O O O O O O O O O O O O O O 153 vii The pc activities h and enlargem “Ch as stat. a 599d for m ”‘9 Proposed Praised for t the prOCESS 0 areas, a Vari the thsiCal ‘1 aCtivity dESiz visitors COme numbers? How tI‘dVEI t0 r6a data available CHAPTER I INTRODUCTION The post-war increase in outdoor recreational activities has not only necessitated the improvement and enlargement of many established recreation sites, such as state and national parks, but has also created a need for new recreational sites. With few exceptions, the proposed sites for these new developments are ap- praised for their economic and social feasibility. In the process of evaluating potential outdoor recreation areas, a variety of values must be estimated, such as the physical attributes which would complement the activity desires of potential visitors. Where would visitors come from to utilize the site and in what numbers? How far is the consuming public willing to travel to reach a recreation facility? With the limited data available, these values cannot be satisfactorily estimated v objective c method for In uti. recreational the precise involved. T than a decadl Prosoect. There is c tional stu areas for the patter recreation This ana not °n1Y 'rec estimated without “heroic assumptions."1 The primary objective of this research is to design a suitable method for approximating these values. In utilizing the natural resource base for outdoor recreational purposes, man should identify and classify the precise character of natural and cultural phenomenon involved. The need for such an analysis was stated more than a decade ago in American Geography: Inventory and Prospect. There is great need in the new field of recrea- tional study for careful analyses of recreational areas for the purpose of depicting and explaining the patterns both of recreational structures and recreational activities.2 This analysis should include, as emphasized above, not only ”recreational structures” but ”activities.“ A 1The author was a member of research teams assessing the feasibility of two national lakeshore recreation areas in Michigan. The association with these research teams and a recognition of assumptions necessary for projecting the magnitude and source of potential visitors gave the initial impetus for this dissertation topic. For a dis- cussion of the methods used to predict visitors to these two lakeshore recreation areas, see Institute for Commun- ity Deve10pment and Services, Report on the Economic Feasibility of the Proposed Sleeping Bear National Sea- shore, prepared for the National Park Service (East Lansing: Michigan State University, 1961), and Institute for Community Development and Services, The Proposed Pictured Ricks National Lakeshore, An Economic Study, prepared for the National Park Service (East Lansing: Michigan State University, 1963). 2Preston E. James and Clarence F. Jones, eds., American Geo ra h : Inventor and Pros ect (Syracuse UnIversiEy Press, I953), p. 255. recreationa unman, and should the: sires in mi the motivat area will b Recrea Cal attribut avide selec Fan's use Of the aggregat have bean use derind at ti An anal \ ReCre . satisfy a h a lime . recreational resource must be viewed as a resource useful to man, and public allocation of recreational resources should therefore be made with the potential users' de- sires in mind.3 Opportunity for outdoor activities as the motivation for traveling to and using a recreational area will be a major premise in this work. Recreation sites or phenomena4 have varying physi- cal attributes and activity Opportunities that offer man a wide selection of where to undertake outdoor activities. Man's use of specific recreation sites is measureable in the aggregate by attendance figures, and these figures have been used frequently as a measure of the satisfaction derived at those recreational sites studied. An analysis of the spatial variation of attendance 3Recreational resources are functional in that they satisfy a human want. This notion is in accordance with Zimmermann's definition of a resource. "The word 're- source’ does not refer to a thing or a substance but to a function which a thing or a substance may perform or to an operation in which it may take part. . . .' Eric W. Zimmermann, Werld Resources and Industries (New York: Harper & Brothers Publishers, 1951), p. 7. 4"Recreational phenomena are of three kinds: 1) the peculiar associations of physical and biotic con- ditions that peOple think are conducive to recreation and which, therefore, constitute natural recreational resources: 2) the structures and other facilities that represent capital investment for recreational purposes: and 3) the recreational activities themselves.‘l James and Jones, 0 . cit., p. 252. at recreatio ing and expl resources am are evident-- recreation sj motivations t recreational ically evalua and OUtdoor a: can, hOVeVer' appraisal, in: variation in a tional SYStem. This stud at recreation areas provides an initial level for analyz- ing and explaining man's perceptions of recreational resources and activity opportunities. Behavior patterns are evident--some outdoor activities occur at specific recreation sites and not at others--but the many human motivations that are influential in the selection of one recreational site as Opposed to another cannot be empir- ically evaluated. The combinations of natural resources and outdoor activity opportunities in a recreation area can, however, be subjectively analyzed. From such an appraisal, insights may be obtained to explain the spatial variation in attendance for a specific outdoor recrea- tional system. This study is addressed to the problem of under- Mf/ standing the spatial variation in camper attendance at Michigan State Parks (Figure l). The research is built ‘ h 4 ' we -rhryvr- upon accepted notions of natural and man-made attributes necessary for camping in state parks as well as travel characteristics of visitors to state parks. Generaliza- tions drawn within the framework of these assumptions are synthesized into a behavioral travel model that can be used to analyze the spatial structure of camping in this extensive state park system. A component of the “o ”. .90 as. 87° 86° 0 . 0 FIGURE 1 \ MICHIGAN STATE PARKS‘ \ ‘ 1964 \ \ \ \ \ . \ \\ .1.) “our WILKINS F.J.H‘LA|~\ . RAGA flU’KALLONOE LAKE “(Add . 1/ “’6. q VAN mun p "“°"”‘""°” 32mm tuner/arr: ““5 martyr «mum \ __ Alleluia c Patina! \\ mm»: Ln ._ “ . 9/“! PO I fi -W®I” ‘ \ I} , ‘~-' HEBOYGAN “3‘ II!- , \\v ”7W 3 \ " ‘ “'5" OALOHA , . I \ \ J w WELL: I S \ \ ‘ D NAWAY ‘ \ " ' BURT Lance arm cu _ I ‘ avail I Q / Young. muse rue-1n Imam A J J" MIR/M O 015260 Luge m4. DAY\ [1,! / CHM! OSCWA / 1’! kV R5 Isuzu: “M, ‘ EC”:- can! ucx use“ PKHOEFT l In rats: 141 (ASIA ' I l I OH"! . HAIRSV LL: 1 I new taste i w WIMPW OHIGGINO n I ll EIGTI‘DM A “CW BEACH/ O V041“ CHELL A2. “I” mm museum on son. um am all 0mm “A“ O mm: ma r CRE’CENT mes M . mum Iturunmra lulu“ MIMI NVAON a»! ”L n" “WT" CLOUD mm saw/us in ‘l maul tuner “any . m mom GRAND HAVEN “yr m; -4.‘.-:“. YO yawn m. “M cum “I!“ m "LLB Hound ”OI-L" 1 I "m 1:. Mt: . "Tm n-r. m JAM I am HIGflLAN) W Vaults muss “‘“T°'§ maul. T .muv L59 0 15mm LAKE 1’ an wry-wad mm WATERLDO mature/um an": «.2 W "’ lnwcxu W3". o—1 1 We. unva- cass arm annex “purl: Mm imm- N wuls L t ONLY vaxxs USED m was INTERACTION mount. an: snow" “I”. MIR L‘A‘EVOQT 1 1 l l 1 l 1 68. 87° “. ”O .4. ”O 47" 45° 42° travel m: component tween the late the 1 origins (5 choices of different a 20 2:00 zsgul 1085.30 30.30.: .UOKDOD 0L ”Isl— z...T.........i.-L _ <>> Eon. (aim I .. a .. . - >Km m>dh<§xocnn< (Fowmzz.§ the appeal a parks. The: traction of to an activi Three identified 1 state parks ' automobile t The elements discussant natural resc the appeal a Specific activity holds for visitors to state parks. Therefore, the notion of indexing the retail at- traction of a shapping center is selected for application to an activity-centered recreation travel model. Three components for a recreation travel model are identified in Chapter III: (1) destination areas or state parks, (2) origin areas of the visitors, and (3) automobile travel-time between origin and destination. The elements of a camping attraction are identified and discussed--the activity preferences of campers, the natural resources of the parks, and the facilities and services. The basis for selection of the variables in each group is given in a brief review of literature on planning and development of outdoor recreation areas. Much of the chapter is devoted to a discussion of outdoor activity preferences of campers, accepting preferences as primary motivation for choosing a particular park, and to the relationships between activity preferences and the quality of natural and man-made resources complementing activity preferences. In Chapter IV a method for combining the natural and man-made resource variables into an attraction index for each park is demonstrated by the use of factor analysis. ing the ed and man-mac derived are camping att tunity for Campers as These indic allurement Wtential ‘ u1e satisfé site are f} attractiOn tation rat. a Masure ( conch"ited l 10 analysis. This procedure yields a system for classify- ing the extent and quality of a combination of natural and man-made resources of each state park. The values derived are then combined into an overall index of camping attraction, including a measure of the Oppor- tunity for various outdoor activity preferences of campers as well as the camping capacity of each park. These indices are assumed to be a measure of the camping allurement of each park and also a reflection of the potential enjoyment derived by campers. Estimates of the satisfaction derived by visitors to a recreation site are frequently made by attendance. The camping attraction index, derived without consideration of visi- tation rates, is at this point put to a severe test as a measure of expected camping pleasure. This test is conducted by comparing the relationship between camper- days5 at each park for a season (a calendar year - 1964)6 and the park indices by means of a regression analysis. 5A camper-day is equivalent to one person camping one night and is considered to equal one day of camp use. 6Data availability determined the type of outdoor activity utilized for the attraction index. The only complete count which included the origin and destina- tions of visitors to Michigan State Parks was made in 1964 from camping permits. The in to Michigan Chapter V. with the ex. number of t 130C361 is ba day figures Eight Origin compared "ii root-meaan LE. 50 errO] with the ad: 11 The interaction model for camping activity flows to Michigan State Parks is described and tested in Chapter V. All the parameters of the various models, with the exception of time-distance, are varied in a number of tests. The relative measure of a successful model is based on the best replication of 1964 camper- day figures at each park and from each of the eighty- eight origins. The projected total camper-days are compared with actual camper-days at each park by a root-mean-square (r.m.s.) error. The results of the r.m.s. error support the use of the interaction model with the addition of a specialized component, the camping attraction indices. A summary of the work in the last chapter also suggests the implications of the study for recreational geography and recreational resource planning. The use of the interaction model as a predictive tool for estimating the use of existing or prOpOSed state parks for camping is discussed along with its limitations. The chapter concludes with the view that this investi- gation is one step in an attempt to understand the site and situational characteristics in the use of recreation areas. APPRA SI’I Economi calculating resources anc WIVES, 1» Thf thWeen tWO V activity and CHAPTER II APPRAISAL OF THE ELEMENTS OF SITE AND SITUATION FOR RECREATIONAL AREAS Economists have made a significant contribution in calculating the value of specific outdoor recreation resources and projecting attendance by the use of demand curves.1 These curves have reflected the interaction between two variables-~the total cost of the outdoor activity and the number of visitors originating at varying distances from the recreation site. Such curves provide an estimate of the value of the recreational 1A few examples of these demand studies are Marion Clawson, Methods of Measuringflthe Demand for and Value of Outdoor Recreation (Washington: Resources for the Ffiture,_I§3377_fiT—§770utdoor Recreation Resources Review Commission, Economic Studies of Outdoor Recrea- tion, Study Report 71—7Washington: U.‘§. Government PEIHting Office, 1962); and E. Boyd wennergren, Value 9: water for Boating Recreation, Bulletin 453 (Logan: Agriculture Experiment Station, Utah State University, June, 1965). Using Clawson's method, the author mea- sured the demand for outdoor recreation at a federal reservoir on the Missouri River. See John S. Evans and Calrton S. Van Doren, "A Measurement of the Demand for Recreational Facilities at Lewis and Clark Lake," Sguth Dakota Business Review, a supplement, XVIII, No. 3 (February, 1960). 12 resource ence by used in ‘ ance at ; Thes measureme tion reso assumed t) Participar requiring neglected. emPhaSized recreatiOna Not a] thSiCal re 13 resources er se, thus measuring the value of the experi- ence by total attendance. By varying the assumptions used in these consumption curves, one can predict attend- ance at particular sites. These pioneering ventures have excluded precise \ measurement of human attitudes and reactions to recrea- tion resource characteristics and have unrealistically assumed that an outdoor exPerience is the same for all participants.2 Individual outdoor activity preferences requiring a variety of natural resources have been neglected. Neither have the economists' demand models emphasized importance of the spatial distribution of recreational areas. Not all recreational areas offer the same activities, physical resources, and develOped facilities. Further- more, the spatial distribution of recreational sites is very influential in determining the amount of use that 2Ullman and Volk develOped an operational demand model for predicting attendance at reservoirs in Missouri. Their model, based on a regression analysis, does not include a consideration of site characteristics at the reservoirs. Edward L. Ullman and Donald J. Volk, “An Operational Model for Predicting Reservoir Attend- ance and Benefits: Implications of a Location Approach to Water Recreation," Pa ers of the Michi an Acade of Science, Arts and Letters, XLVII (1962), p. 473. w ' . 1" . each site rec an attempt t: ing intensit- ity Opportun: Travel n a few have 14 each site receives. The model deve10ped in this work is an attempt to investigate separately the spatially vary- ing intensities of use, resource attributes, and activ- ity opportunities. Recreation Travel Models Travel models for recreation have been suggested; a few have been attempted. In discussing the location and use of recreational areas, Perloff and Wingo have stated: Normally, a consumer will use a given facility at a rate related to its distance from his home: the more remote it is, the less he will tend to use it. Or, given several comparable facilities, all other things being equal, he will use the nearer more frequently than--or to the exclusion of--the more remote.3 Perloff and Wingo have suggested an interaction model to project travel to recreation sites. The counterpart of accessibility is "demand poten- tial.” This can be viewed as a so-called “gravity” model in which is incorporated the tendency of demand for a service to vary inversely with its distance from the consumer.4 A similar model of this type has been suggested by Clawson 30.8., Outdoor Recreation Resources Review'Com- mission, Trends in American Living and Outdoor Recrea- tion, Study Report 22 (washington: U. 5. Government Printing Office, 1962), p. 91. 4Ibid., p. 93. in a diagra ters of dif parks.5 In in the Los Case and Le in selectin 15 in a diagram showing the connectivity between urban cen- ters of different sizes and various distances from parks.5 In predicting the demand for outdoor recreation in the Los Angeles MetrOpolitan Area in 1976 and 2000, Case and Levin also suggest the use of a “gravity" model in selecting Optimum locations for recreation sites.6 Interaction Models Considerable work has been done with interaction models for urban land uses, particularly retail activi- ties. Admittedly, there are differences between retail travel behavior and recreation travel behavior, but some of the basic notions used in the retail models are applicable for this study. One purpose of this chapter is to eXplore the similarities between urban travel models and recreational travel models and to show their utility to this particular tapic. Interaction models as applied to human movements have developed from an analogy with Newtonian concepts of gravitation. These concepts view a region as a mass 5Marion Clawson, Land and Water for Recreation (Chicago: Rand McNally and Co., 1963), p. 48. 6U. 8., Outdoor Recreation Resources Review Come mission, The Future of Outdoor Recreation in Metropoli- tan Re ions of the United States, Study Report 21, V01. III TWaShington: U. S. Government Printing Office, 1962), p. 103. structure 1' or initiate tial interL between mas mass. The tr is 16 structure in accordance with principles that constrain or initiate action of the individual particles. Spa- tial inter-connections are viewed as interactions between masses where papulation is used as a measure of mass. The traditional "gravity“ or interaction hypothesis,7 is P1P2 dx (1) where I is directly prOportional to the product of two papulations, P1 and P2, and inversely related to the exponential function of the distance between them, dx. A large number of empirical studies have shown the promise of interaction models for analyzing various spatial inter-connections.8 The measure of I'mass" used has depended on the problem studied, for example 7This is the general form used by Stewart in his initial work. John Q. Stewart, "An Inverse Distance Variation for Certain Social Influences,” Science, XCIII (1941), pp. 89-90. 3A detailed discussion of the deve10pment and application of these models may be found in Gerald A. P. Carrothers, ”An Historical Review of the Gravity g and Potential Concepts of Human Interaction,” Journal of the American Institute of Planners, XXII (Spring, 1956): pp. 94-102. comnodi ty and worker Traff have refin< traffic fl Voorhees, v iS basic tc 'that all t attracted, "if-h Specif gravity mod shOpping ha \ , 9W1111 17 commodity output,9 retail sales,10 floor area in apparel, 'and workers employed.11 Traffic engineers, urban planners, and sociologists have refined ”gravity" or interaction models to predict traffic flow. Prominent among these groups has been Voorhees, whose I'General Theory of Traffic Movement"12 is basic to this research tOpic. Voorhees' concept is "that all trips emanating from a residential area are attracted, or 'pulled' to various land uses in accordance with specific empirical values."l3 Voorhees, using a gravity model,14 demonstrated this idea in research of shapping habits. It was determined that the travel time 9William.Warntz, Toward A Geography_of Price (Philadelphia: University of Pennsylvania Press, 1955). 10Both Chauncy D. Harris and Edgar S. Dunn used retail sales as measuring mass. See Chauncy D. Harris, "The Market as a Factor in the Localization of Industry in the U.S.," Annals of the Association of American Geographers, XLIV (December, 1954), pp. 315-348; Edgar S. Dunn, "The Market Potential Concept and the Analysis of Location," Papers and Proceedings of the Regional Science Association, II (1956), pp. 183-194. 11Alan M. Voorhees, Gordon B. Sharpe, and J. T. Stegmeier, Shopping Habits and Travel Patterns, Special Report II-B (Washington: Highway Research Board, 1955). 12Alan M. Voorhees, ”A General Theory of Traffic Movement,” 1955 Proceedings of the Institute of Traffic En ineers, October, 1955, p. 46. 13Ibid., p. 56. 14Voorhees, et a1., 0 . cit. l between a z the only (:1 another. I relationsh: in terms of this measu: means of re the intel'ac where . A] Willing Do» 18 between a residential area and a shepping center is not the only criterion for selecting one store as Opposed to another. The additional criterion is the competitive relationship among various stores or snapping centers in terms of the amount of retail floor area. By using this measure rather than papulation or attendance as a means of rating the attraction of a shopping center, the interaction model gave more realistic projections. Thus modified, equation one of the model becomes A. a 113 = P1 (2) dijx where Aj is an empirical measure of the attraction or pulling power of a shepping center. It is assumed that a similar equation can be applied in this study: that is, each state park has an empiric- ally derived value, Aj, that will describe certain site characteristics that attract campers. Measuring the mass attractive qualities of parks has been a major problem for park planners. Interaction Models for Recreation Trips The interaction hypothesis was used to project attendance at two parks in southeastern Connecticut by Crevo15 an: Theoretical similar to park was nc He assumed Proportions divided by CaVanaugh E formula cm in questior was in err eXpeCt dif Catton rate 05 int 19 Crevo15 and at Yellowstone National Park by Cavanaugh.16 Theoretical trips were predicted by Crevo in a manner similar to equation one: an attraction value for each park was not used. Cavanaugh's assumptions were similar. He assumed that the cars entering Yellowstone should be prOportional to the population of the state of origin divided by the distance from the national park, or P/D. Cavanaugh postulated that the first P in the Ple/D formula could be drOpped because the mass of the park in question was the same for all visitors entering the park. Catton,17 on the other hand, believes Cavanaugh was in error by making this assumption because one could expect different parks to vary in their attractive mass. Catton used an interaction model to project the rate of interaction between visitors from the forty-eight 1SCharles C. Crevo, "Characteristics of Summer Weekend Recreational Travel,“ Highway Research Record, No. 44, Highway Research Board Publication 1161 (Washington, 1963), p. 51. 16Joseph A. Cavanaugh, “Formulation, Analysis and Testing the Interactance Hypothesis,‘I (unpublished Ph.D. dissertation, Department of Sociology, University of Washington, 1950). 17Cavanaugh's work dealing with national parks is discussed in William R. Catton, Jr., ”The Concept of 'Mass' in the Sociological Version of Gravitation," Mathematical Explorations in Behavioral Science, eds. Fred Massarik and Philburn Ratoosh (Homewood, Illinois: Richard D. Irwin, Inc., 1965), p. 294. states and ferent par To account each park annual vis a measure model; the natural ar Parks that a Closer I rates to Kepler's I tute for I 20 states and selected national parks. He assumed that dif- ferent parks have varying degrees of an attractive mass. To account for variations in the park mass, he indexes each park by a measure of pOpulation--specifically, annual visitor-days. He concludes that visitor-days as a measure of mass are inappr0priate in an interaction model: that it is the 'sacredness” (preservation of a natural area as nearly unchanged as possible) of national parks that constitutes their attractive force. He found a closer relationship between expected and actual visitor rates to western national parks by using a derivation of Kepler's Law and a measure of 'sacredness' as a substi- tute for mass. National and state parks have generally been estab- lished because of unique or scenic natural attributes desirable for preservation in an unaltered state. It is the extent of the resource qualities that help to make a park attractive. The notion that the natural resources of a park do constitute part of its attraction is develOped in this travel model, though it is formu- lated on a probabilistic basis. Huff18 believes 18David L. Huff, ”The Use of Gravity Models in Social Research,” Mathematical ExPlorations in Be- havioral Science, op. cit., p. 317. differenc ments car havior. istic moc of moveme potential interactj In 1 one is 91 Spatial j ChOiCes ‘- Preferem individu, of a Par] IPayoff» aSsighed tYPe Of a model number 0 Huf to recre \ 19E 21 differences among individuals and variations in environ- ments can bring about different forms of Spatial be- havior. He recommends the formulation of a probabil- istic model that ”would seek to determine the likelihood of movement from a given point of origin to various potential sources of destination for any type of spatial interaction."19 In reference to Catton's work, Huff states that if one is given a set of alternative choices for a type of spatial interaction (in this case parks), and if these choices are representative of an individual's tastes and preferences, then the selection of one park by that individual reflects its “utility.” Thus, the prOperties of a park play a part in measuring its utility or “payoff“ to the visitor. Therefore, if each park is assigned an empirical value that reflects a specific type of spatial interaction, this value can be used in a model to project the probability of traveling to a number of parks. Huff recommends the use of park size (acres devoted to recreational purposes) as an empirical measure of 19Ib1d., p. 319. park utili predicting floor area mass or at power of s; fied by co; Voorh. mass consuz with succe; and Psycho. of COUSUme] Primarily ‘ competitiw fEedback. eXpreSSQd j aChiEVed W! PerienCe is lEaVe the x 22 park utility.2o This suggestion stems from his work in predicting flows to shopping centers, where the retail floor area of shopping centers is used as a measure of mass or attraction.21 The parallel between ”pulling" power of shapping centers and state parks can be justi- fied by comparing the objectives of the two trips involved. Voorhees and Huff have incorporated the results of mass consumer preference studies into their travel models with success. Huff22 has analyzed the social, economic, and psychological factors affecting the space preferences of consumers for particular products. His work has been primarily concerned with retail products which have a competitive market structure involving pricing and market feedback. In the case of camping, such a feedback is expressed in terms of individual or group satisfaction achieved while camping at a park. If the camping ex- perience is unsatisfactory, the camper (consumer) will leave the park and probably stay away in the future. 2°Ibid., p. 320. 21David L. Huff, "A Probabilistic Analysis of Shop- ping Center Trade Areas,‘ Land Economics, XXXIX (Feb- ruary, 1963), p. 81. 22David L. Huff, ”A Topographical Model of Consumer Space Preferences," Pa ers and Proceedin s of the Re- gional Science Association, VI (1960), pp. 159-173. The consumer require an 1' taken with a retail goods center accor and type of is assumed t make choices If the Parks with v‘ is coMeet, 1 opportunitie; superi” qua, that do not I diversity Of Customs: by 1 stores. 30 dc One Park hen, indiCate a '. : 23 The consumer shOpping trip and the camping trip both require an investment in time and money, both are under- taken with a specific objective in mind. Consumers of retail goods initially order their choice of a shopping center according to their trip objectives and the number and type of products offered at alternative sites. It is assumed that campers have the same opportunity to make choices between state parks which offer varying resources and activity Opportunities. If the assumption that campers do choose between parks with varying resources and activity opportunities is correct, then parks offering a wide range of activity opportunities or resources and activity opportunities of superior quality should be more attractive than parks that do not provide such opportunities. Just as the diversity of stores in a shopping center benefit the customer by reducing time and cost traveling between stores, so does a variety of activity opportunities in one park benefit the camper. Shapping behavior patterns indicate a “pulling" power when sets of retail functions exist.23 ilar linka within a h Claws outdoor re that Campe Situationa 24 exist.23 Parks, like shopping centers, may exhibit sim- ilar linkages by offering a number of outdoor activities within a high quality environment. Attraction Qualities of Recreation Sites Clawson's24 generalized description of the whole outdoor recreation experience substantiates the idea that campers are activity oriented and that site and situational characteristics of a park are important. Such an exPerience consists of five clearly separate phases: (1) anticipation, (2) travel to, (3) on site, (4) travel back, and (S) recollection. About the "travel to“ experience Clawson states: In order to reach the outdoor recreation area of its choice, a family must travel. Considerable expense is involved in such travel, and often as much time is consumed in travel as later on the site.25 23The proposition that retail stores are arranged into shopping centers of various sizes by entrepreneurs to meet the desires of customers to combine shopping trip purposes has been studied by John D. Nystuen, ”Geographical Analyses of Customer Movements and Retail Business Location: (1) Theories, (2) Empirical Patterns in Cedar Rapids, Iowa, and (3) A Simulation Model of Movement” (unpublished Ph.D. dissertation, Department of Geography, University of Washington, 1959). 24Clawson, op. cit., p. 41. 251bid., p. 40. But the 'on outdoor exp this manner- When it site, th Bodies of door recr generally experienc the time expense.2 Perloff and are the Prin SUCCinctly s and Site Che 25 But the ”on site" eXperience is the real payoff for the outdoor eXperience. It is characterized by Clawson in this manner: When it [the family] arrives at the recreation site, the family may engage in many activities. Bodies of water are especially valued for out- door recreation. The activities at the site generally provide the basic purpose of the whole experience, even when they occupy less than half the time and require less than half the total expense.26 Perloff and Wingo also believe that outdoor activities are the primary focus of the outdoor experience and succinctly state the relationship between activities and site characteristics: In the system we are describing, activities are the fulcrum which fixes the overall relationship of the recreation prOpensities of outdoor recrea- tion groups to the array of facilities which are in different degrees available to users. They relate in Specific ways to the behavioral patterns of the outdoor recreation groups and each has certain requirements for the nature of the facil- ities that support it.27 State parks for camping may be viewed as recreation centers that offer a range of outdoor activities for visitors. If campers are truly activity oriented as Clawson believes, then a state park's attractiveness to 261bid., p. 40. 27ORRRC Report 21, o . cit., p. 89. a camper w can be und devoted to park's att called for travel mod]- deve 10pmen major need 26 a camper will partially depend on the activities that can be undertaken. A major portion of this work is devoted to the development of a measure reflecting a park's attractive mass, Aj. Such a measure has been called for but has not been developed for use in a travel model. Clawson and Knetsch have recommended the develOpment of a “rating scale” of attractiveness as a major need for considering recreation supply problems. One line of research on recreation supply is to develop rating scales, or systems, to measure the inherent attractiveness of different outdoor recreation areas. Even the most casual observation shows that some areas are much more attractive than others: often, however, differences are not as clear or lack some kind of specific description or measurement. . . . It seems entirely possible to deve10p specific and rather objective rating scales for different kinds of outdoor recreation areas, for different uses of each, that would have great utility in planning, in research, and in administration.28 This underlines the complexity of planning and de- veloping outdoor recreation sites. Supplying the public with a scenic site, although important, is not enough. Services and facilities must also be supplied, and the entire recreation entity planned for outdoor activities preferred by its users. 28Marion Clawson and Jack L. Knetsch, Outdoor Rec- reation Research: Some Concepts and Suggested Areas of Study (Washington: Resources for the Future, Inc., Reprint Series, 1963), p. 261. Assum of a state campers fo: fied, a ta: Clawson ha: ence pattEJ We know seek ou‘i from it adminis* PhYSicaI USing P1 is tUrn: expects recreat; Consume; This constrUCti: 27 Assuming that outdoor activities are the focal point of a state park for camping, then the preferences of campers for various outdoor activities must be identi- fied, a task which has not as yet been adequately done. Clawson has described the need for user activity-prefer- ence patterns: We know all too little about why different persons seek outdoor recreation, or what they hOpe to gain from it. Too often, we have thought of recreation administration and management in terms of the physical area, and not enough in terms of the using public. Just as modern marketing theory is turning to a study of what the consumer wants, expects, and is willing to pay for, so must modern recreation administration turn to a study Of its consumers.29 This informational gap provides a major obstacle in constructing a camping index. Elements other than out- door activity Opportunities which should also be included in an aggregate measure of attraction are environmental characteristics and facilities and services of the recrea- tion site. The variables included in these two elements can be identified, but rating them for use in an index is more difficult. In the next chapter the activity preferences of 29Leslie M. Reid, Outdoor Recreation Preference--A Nationwide Study of User Desires (Department of Resource Development, Michigan State University, East Lansing, Michigan, by the author, 1963), p. 17. campers ar state repo ferred by 1 After the 4 State Park values are activity p. outdoor 6C] Provides a and genes Elemei i vices Prov. C‘m‘nities. the Variabz fied 0n the 28 campers are identified as determined in national and state reports and surveys. Each outdoor activity pre- ferred by campers is assigned a numerical rating. After the activity Opportunities at each Michigan State Park are recognized, the activity preference values are assigned to each park. Totalling these activity preference values for each park develOps an outdoor activity rating for camping. This rating scale provides a measure of the activity attraction of parks and serves as one element in the attraction index, Aj. Elements number two and three, the natural resource characteristics of each park and the facilities and ser- vices provided, contain many variables Of different qualities. In the next chapter, selection of many of the variables included in these two elements is justi- fied on the basis of recreation planning and research reports. A scoring routine is used to combine all three elements into a camping attraction index. Summary This review of previous attempts to describe out- door recreational systems has concluded with the idea that site and situational characteristics have not been considered elements of tle or no I'z ity desires economists given as ex characteris activity op Prev ic “nation: a mdSUIe < of attenda: interacts) lels betwe and those 29 considered in combination. Most statements have included elements of the location of recreational areas with lit- tle or no regard for site characteristics and the activ- ity desires of participants. The models develOped by economists to predict attendance at recreational sites, given as examples of this type, have ignored the internal characteristics of individual recreation areas and the activity Opportunities available. Previous attempts to use interaction models for recreational trips have indicated the need for deriving a measure of the mass attraction of each park independent of attendance. A discussion of the basic assumptions of interaction hypotheses led to an analysis of the paral- lels between the attractive qualities of shOpping centers and those of state parks. One component of an inter- action model for recreational sites could be a measure of the internal park characteristics and its activity Opportunities. Such a measure or index is not too dif- ferent from the attraction measures used for retail location models. The need for such a measure or rating scale for recreation areas has been recognized but none have been developed. The use of an attraction index for specific recreatior ways. Fir pepulatior activity, physical c existing c index has independer The 1 Component: the elem] of the Va 30 recreational areas in a travel model is valuable in two ways. First, by using an attraction index rather than‘ pOpulation or attendance to demonstrate demand for an activity, in this case camping, both behavioral and physical characteristics can be used to project use at existing or prOposed parks. Second, the attraction index has a basic use in evaluation Of a resource site independent of a travel model. The following chapter identifies and defines the components of an interaction travel model and analyzes the elements of the attraction index and the selection of the variables associated with each element. Inter tional Sy 5 tern resu] People, fa PhYSical _ With the E tional tr: bEtween tx 1. CHAPTER III COMPONENTS OF.AN INTERACTION MODEL FOR MICHIGAN STATE PARKS Interpretation and analysis of a statewide recrea- tional system requires a knowledge of the spatial pat- tern resulting from the complex interaction between peOple, facilities, resources, time, and space. The physical supply of recreational resources must be matched with the spatial demand for the facilities. A recrea- tional travel model that recognizes the "connectivity"1 between two or more unlike points in space involves three components: 1. Supply: The destination areas (state parks), designated A. 2. Demand: Origin areas (campers represented by nodal population centers), designated P. 3. Distance between 1 and 2, designated TD for time-distance. 1Edward A. Ackerman, "Where Is A Research Fron- tier?“ Annals Of the Association of American Geograp- hers, LIII (1963), p. 437. 31 The firsi ically es probably figures f available distance , The index is ; then with attractiox 32 The first problem in develOping such a model is to empir- ically estimate a measure for the supply component, probably the key (or primary) component, as the demand figures for origin areas of campers in the model are available from camping permits, and the third component, distance, is readily Obtained from highway maps. The Supply Component--Camping Attraction Index The initial step in develOping a camping attraction index is selection of the elements of attraction, followed then with selection of elemental variables. The camping attraction index for Michigan State Parks is a composite Of three elements: 1. The outdoor recreation activity preferences of campers. 2. A consideration of the natural resources that enhance the camping experience. 3. Facilities and services that complement the camping eXperience and related outdoor activities. The available natural resources influence the types of acitvities possible: the activity Opportunities are in turn conditioned by facilities and services provided. One method of visualizing the interrelationships of these three elements is construction of a flow chart such as that in Pi and their 1 Michigan 8 inventory more detai and their z 33 that in Figure 4. A complete inventory Of these elements and their associated variables was conducted for each Michigan State Park (see Appendix D, Table 17, for the inventory sheet used). Appendixes A through C provide a more detailed discussion Of each of the three elements and their variables than is included in this chapter. Identifying Outdoor Activity Preferences of Campers The camping attraction index for Michigan State Parks should include specific activities preferred by campers. Unfortunately, no survey or report stands out as providing the Optimal data needed for the index. Several survey reports appear to be representative of the type of data necessary, but omissions of a type or types of activities, the location of interviews, dif- ferences in the phrasing Of questions, and variations in the method of presenting the data make comparison diffi- cult. Some clarification is necessary at this point be- tween two terms--preferences for activities as opposed to participation in activities. Preferences are the desires or aspirations for particular activities, while participation implies actually undertaking a certain 6: LA z I -lzlllli .iriz ilzlll ll- VI mat-slicnu >0 N mu0t>ttm I Q ..~.1\.W\ .hi ,th AJZV. N; >Z.. (- 2h... >. .1. \ .PZLEZRU . MU Id \ mica -Fs>~hnzv< \ minufi-\ ~\,ba‘a\ PHYSICAL ENVIROWENT ACTIVITIES PREFERRED FACILITIES AND L l ’ WATER =E CULTURAL . \ QUALITY J, \\‘I'” ___( STOR CA ——————— HI I L POLLUTIOm TURBIDITYI V EXTENT L I ARCHEOLOOIC it CONTEMPORARY it ENVIRONS I! E ‘11 GREAT WILDERNESS DUNES LAKES ATMOSPHERE vIReIN TIMBER P———-——fit-— -—It-—— — —————————— (D K M Q 2 4 U E SWIMMING LAvspoRTs NATURE [u I STUDY ARCHERY IFLE RANGES SERVICES DI‘ CHILDREN'S 8d PLAY EQUIP. RC HORSESHOE A PITS INTERPREHVE PROGRAMS AND BROCHURES OUTDOOR CENTERS ng activity ai is based or being poss: for the pur park, the \ actiVity, k; exist. The unfulfilled reflected i the Person en‘39-3 for a the attract available, will of nec The E the importa 35 activity at a recreation site. Since the choice of a park is based on what types of activities are perceived as being possible once there, preferences are more important for the purposes of an attraction index. While at the park, the visitor may not choose to partake Of a given activity, but the possibility of participation will still exist. The National Recreation Survey agrees that ”the unfulfilled demand for an outdoor recreation activity is reflected in preferences for the activity, even though the person may not participate."2 Information on prefer- ences for activities are therefore desirable for develOping the attraction index. Since preference data is not always available, however, the selection Of ratings for activities will Of necessity be based on participation rates. The National Recreation Survey provides insight into the importance of camping relative to other outdoor recrea- tion activities (Table l). The most important activities for the North Central Region are picnicking, sightseeing, swimming, playing outdoor games, and fishing. With few exceptions, these are the most pOpular activities on a national basis also. Camping participation, relative to 2U.S., Outdoor Recreation Resources Review Commis- sion, National Recreation Survey, Study Report 19 (Wash- ington: U. S. Government Printing Office, 1962), p. 4. TABLE 10.- pating in ‘ States Activity ‘- Picnics Driving for SViMihg Sightseeing Walking for Playing Ou 1PiSI'ling Attending O threats Other Boatj BicYclin 36 TABLE l.--Percent of persons 12 years and over partici- pating in various outdoor recreation activities, United States and North Central Region, Summer, 1960 Percent Of Persons Participating United States North Central Region Activity % Ranka % Rank Picnics 53 1 58 1 Driving for Pleasure 52 2 58 1 Swimming 45 3 42 3 Sightseeing 42 4 47 2 Walking for Pleasure 33 5 29 6 Playing Outdoor Games 30 6 35 4 Fishing 29 7 33 5 Attending Outdoor Sports 24 8 28 7 Events Other Boating 22 9 27 8 Nature Walks 14 10 15 9 Bicycling 9 11 10 10 .Attending Outdoor Con- 9 ll 11 9 certs, Drama, etc. Camping 8 12 7 ll Hiking 6 13 5 13 Horseback Riding 6 13 5 13 Water Skiing 6 13 6 12 Miscellaneous 5 l4 5 13 Hunting 3 15 2 15 Canoeing 2 16 3 14 Sailing 2 16 2 15 Mountain Climbing l 17 l 16 aRanking was done by the author. Source: U.S., Outdoor Recreation Resources Review Commission, National Recreation Survey, Study Report 19 (Washington: U.S. Govern- ment Printing Office, 1962), Table 1.01. the other developing is on a s. pating in I Using possible t( in Michiga: that are e. in a Comple parks, Us: Ila]. RegiorI “dertaken sightseeing walking fozl much hue: order, 11a tL‘ hOl'szack r so“ 0. 37 the other activities, is very low. This means that in develOping the activity preferences of campers, the focus is on a small prOportion of the total pOpulation partici- pating in outdoor recreation activities. Using the list of activities in Table 1, it is possible to specify activities that can be undertaken in Michigan State Parks and to eliminate many activities that are either impossible in the parks or are not included in a complex of choices desired by campers while in the parks. Using the participation rates for the North Cen- tral Region, the most pOpular activities that can be undertaken in most Michigan State Parks are picnicking, sightseeing, swimmdng, playing outdoor games, fishing, walking for pleasure, and boating. Activities with a much lower percentage Of participation are, in decreasing order, nature walks, bicycling, camping, water skiing, horseback riding, hiking, canoeing, sailing, and hunting. Some of these activities can be eliminated from consideration. Hunting, for example, is an important outdoor activity at some sites: it is allowed in state recreation areas, but only in a few state parks. Since hunting occurs during October to March, after the May to October ca as an acti' Bicyc- majority o: and bicycle exception I considered System.3 Horse: tion areas "in be con For the mos DriVin can be eliml any 0f the to Offer 10 Park 38 October camping season, it will not be considered further as an activity in the index. Bicycling is limited in Michigan State Parks. A majority of the peOple do not bring bicycles with them, and bicycle rentals are not provided in the parks. The exception to this is Mackinac Island, which will not be considered here as a part of the Michigan State Park System.3 Horseback riding is done in a few parks and recrea- tion areas adjacent to the Detroit Metropolitan Area and will be considered here as an attraction in these parks. For the most part, however, this activity will not be pertinent to the attraction index. Driving for pleasure and attending sports events also can be eliminated. Sports events are not organized within any of the parks, and a majority of the parks are too small to offer long scenic drives within the confines of the park. Driving for pleasure may very well be important in considering the distance of parks from.visitors' homes 3Mackinac Island and Fort Michilimackinac are admin- istered by the Mackinac Island State Park Commission and not the Parks Section of the Michigan Department of Con- servation. These two parks have a Special character in that they are primarily historical sites with many come mercial establishments nearby, and, in this respect, are different from a majority of state parks. and the t: ity will its magni nation th Walk” as one ac Separately not be um um and sledd important as elem! occur in takes pl TWC national 39 and the type of landscape to be traversed, but this activ- ity will not be considered further since measurement of its magnitude is difficult without more detailed infor- mation than is now available. Walking for pleasure and nature walks are considered as one activity in this study--hiking is considered separately. Mbuntain climbing is excluded since it can not be undertaken in any state park. Although winter sports such as ice skating, skiing, and sledding are not included in Table 1, they are important activities in Michigan. They are not included as elements of the attraction index because they do not occur in the summer season when the majority of camping takes place. Two national surveys in the ORRRC Reports provide national and regional activity-participation rates for campers. The first compares activity-participation rates between campers and non-campers in a number of outdoor activities. This data is shown in Table 2, reproduced from ORRRC Report 20.4 The information in this table 4U.S., Outdoor Recreation Resources Review Commis- sion, Participation in Outdoor Recreation: Factors Affecting Demand Among American Adults, Study Report 20 (Washington: 0. S. Government Printing Office, 1962), p. 68. IIIIIIIIIIIIIIIIIII onmEnU ucoz on: wmoze mCOE< :OfiuanOwuumm ucou Mom QDHOQEQOIDOE DC! IHOQEDU in 00d0fi>fi006 “000030 H0200 Cfi COfiUUQfiUflOHUQ N0 COQMHHQEOUllaN WQQQH 40 ooa mm m m nuogamolcoz can we aa «a «H anodemo means cyan can ousumz ooa em a m muom8m0|coz can am he mN mm mumaemo muses: ooa no ha ha muomesolcoz cod mm eh mm we muomeno wannabe ooa nu ma m numdfinoucoz OOH me am m~ Hm anodeao mcwoocnu can mcwunom cod em on mm mumdfifioncoz cod an mm mm me muoaemo comma m on $58 no ficgwsm uoovuno fiance and um uoz mofiwa via moawa coumo >Da>wvo< can maybe euH Huuoa mcumsmu Dcoz 0:3 omega madam cowuwmflowuumm ucou mom anGQEdOIcoc can muomaco an noauw>wuom noctuso Honuo cw cowucmwoauunm mo conwuomaooll.~ Manda . .HHH N‘CEU. DfldnmEflU 06.03 023 OIOSH. UGO—Hts COfluflQflOfiukflm UGOOHOE on. “cooll ON Mflqu‘puu .mo .a ..~oma .ooammo ”cowumouoom Hooouso cw mcauefium aeoecuo>oo .m .s «noumahammzv cu phenom henna .uuasea dmowuo5¢_wcofid panama cwuoomud nuoaomm cowuumwowuumm .cOfinnwEEou 3ow>om moousomom coaumouoom uoovuoo ..m.a «ooucom no: one won» unamoon who: popuflao one home .xoccfl cowuumuuum 0:» ca pmvaaocw .oannu Hmcwmwuo ecu cw powaaocw mums coaumxmaou can .mcwoomunmam dawnoeoucm .munomm Hounds Honuo .mcwwxn .mcaucsmn 1. con vm m N nHoQEMOIcoz 4. ooa mm we a m muonsao .mmwpwm Romnmnuom con mm an an muodamoueoz ass a Hm Nm mm muodaao nowcowm noses and mushy gonna mua>nuoa Hobos and um uoz can coumo and Hobos madmauu acox on: muons mcoad coduomwoauuum Dcoouom .O.unOUll.N mnmdfi was obtain pers acti particular fifty-five under $3,0l The data 1 is the cal tion betwe¢ This table Ontdoor am the camper: than one-h; nature Val) camper to I 42 was obtained by a home interview survey and not from camp pers actively undertaking a camping trip. PeOple not particularly interested in outdoor recreation--persons fifty-five years of age and over, and peOple with incomes under $3,000--were excluded from the original table. The data in this table are relevant, however, because it is the only data found that compares activity-participa- tion between those preferring camping and those not. This table indicates that campers are more active in many outdoor activities than non-campers. Three-fourths of the campers go fishing, two-thirds go swimming, and more than one-half go boating. Campers also prefer hiking, nature walks, and horseback riding. The willingness of a camper to “rough it' also appears in the camper's desire for picnics which is higher than for non-campers. By combining the percentages for campers in the first and second columns of Table 2, an estimate of the import- ance of each of the activities to an active camper is read- ily obtained (column three). For instance, picnics are the most pOpular activity of campers, with 91 per cent, followed by fishing with 74 per cent, swimming 68 per cent, boating and canoeing 57 per cent, hiking 47 per cent, nature an 16 per ce combinati sirable f=' preferenc: Simi. Report 20 according 43 nature and bird walks 27 per cent, and horseback riding 16 per cent. These figures provide an estimate of the combination of activities that might be considered de- sirable for the “average camper," ranked in order of preference. Simdlar activity desires of campers are indicated in the National Recreation Survey. This survey, like ORRRC Report 20, attempts to identify participation patterns according to the socio-economic characteristics of per- sons undertaking outdoor activities. Persons are hypo- thesized to have activity preferences that can be predicted. For example, it is expected that an individual who frequently fishes will have an affinity for other water-related activities. Similarly a person who frequently participates in such activities as ‘ sightseeing, driving for pleasure, and attending outdoor events is expected to have an affinity for related urban-centered outdoor activities.5 The correlation coefficients from the National Recreation Survey allow a more detailed comparison of camping with the other activities. Camping is strongly associated with water activities such as fishing, swim- ndng, and boating (Table 3). The strong association of camping with hiking is not surprising since hiking was SORRRC Report 19, op. cit., p. 6. TABLE 3.-- Water Skii Horseback Attending Walking fo I \- Sourc! pack'us SL1 44 TABLE 3.--Correlation coefficients of participation in selected activities with camping participation, Summer, 1960 Activity Correlation Coefficient Hiking 0.28 Boating .22 Fishing .21 Swimming .21 Water Skiing .21 Horseback Riding .09 Attending Outdoor Sports Events .08 Walking for Pleasure .06 Source: U. 8., Outdoor Recreation Resources Review Commission, National Recreation Survey, Study Report 19 (Washington: U. S. Govern- ment Printing Office, 1962), p. 33. defined in the report as being "along a trail with a pack," suggesting an outing requiring an overnight stay. In this survey camping and hiking were treated as “back- woods” activities. For example, campers were asked if they had camped in "developed“ or “in wilderness or remote areas."7 Sixty-five per cent of the campers had patron- ized ”developed“ rather than "wilderness“ campsites. This 6Ibid., p. 108. 7Wilderness areas or undeveloped areas were defined as "an area not accessible by improved roads and without developed campsites.“ Sites of this type would be acces- sible only by ”packing in.“ Ibid., p. 33. would indi to campim tion beth Since Parks are there are to a remot camping 01 ance of h} centage r a is appareJ The 1 45 would indicate that the “backwoods“ connotation attached to camping may not be strong in spite of the high correla- tion between camping and hiking in the national survey. Since most of the campgrounds in the Michigan State Parks are developed rather than undeveloped, and since there are few parks large enough for wilderness hiking to a remote campsite, the relationship between hiking and camping of this type is not significant. With the import- ance of hiking minimized, the similarity between ranking of the correlation coefficients in Table 3 and the per- centage ranking of preferred activities listed in Table 2 is apparent. The two national surveys, ORRRC Reports 19 and 20, provide the best information on the activity preferences of campers. An analysis of several state and local surveys-- two Michigan State Park surveys,8 a Wisconsin State Park 8Thomas L. Dahle, Michigan State Park Users Survey, Bureau of Business Research, College of Business and Public Service, Research Report No. 19 (Michigan State University, East Lansing, Mich., 1956), and Michigan House of Representatives, Report of Committee on State Parks and Public Lands (Lansing, 1962). The last refer- ence is widely known as the Van Til Report after the Committee Chairman, Representative Reimer Van Til. References in the text will be to the Van Til Report. survey,9 19mm, found in obvious : (see App‘ and loca by Dahle eral, Hi The Van is also ities. the Gin: -—_.‘___ 9 H . State p 5 Po {Est 1 m 46 survey,9 and a study of camping parks in Iron County,Mich- igan,1°--is in reasonably close agreement with preferences found in the national surveys.11 However, there are obvious informational gaps in the state and local surveys (see Appendix A for a detailed discussion of the state and local surveys). The first Michigan State Park survey by Dahle may be outdated and, in any event, is too gen- eral, with no breakdown of camper activity preferences. The Van Til Report, the second Michigan State Park survey, is also limited in terms of a discussion of camper activ- ities. Camper activities are discussed in more detail in the Gilbert Study, but it is restricted to a few localized 9H. Clifton Hutchins and Edgar W. Trecker, Jr., 222 State Park Visitor-~A Report of the Wisconsin Park and Forest Travel Study, Wisconsin Conservation Department, Technical Bulletin No. 22 (Madison, Wisc., 1961). 10Alphonse H. Gilbert, 'A Survey of Vacation Camping in Iron County, Michigan" (unpublished Master's thesis, Department of Resource Development, Michigan State Uni- versity, 1963). 11There are many other surveys of the use of state parks available. However, they were not included in the discussion primarily because the states were remote from Michigan. One example was washington State Parks and Recreation Commission, We Come to Camp in Washington State Parks--Overni ht Ca in Surve (Olympia, Wash., I553). An additiona survey has recently been called to the author's attention. This survey is Ohio Depart- ment of Natural Resources, Ohio State Parks--Travel and Use Surve (Columbus, Ohioz-Division of Parks and Recrea- ton, 6). parks. In sp: to the entire to estimate n1 undertaken by To numer activities, 3 facilities ar. necessary. 1 011 the bBSis according to The typ' ”But of the lished in th rating scale (2) asSign . 47 parks. In spite of the lack of data directly applicable to the entire Michigan system, the attempt will be made to estimate numerical preference values for activities undertaken by campers. Rating Scales Based on Activity Preferences To numerically rate the preferences for outdoor activities, attributes of the physical resources, and facilities and services, a judgment rating scale is necessary. The judgment for activity preference is made on the basis of an average preference for each activity according to the studies reviewed. The type of rating scales used throughout the develop- ment of the attraction index are reasonably well-estab- lished in the fields of sociology and psychology.12 The rating scales (1) order or categorize a variable and (2) assign an interval scale to the ordered categories. Both involve judgments. Professionals may not agree, for example, on what is a prOper classification of vege- tative types, or which types are of a higher quality for 12For examples see Warren S. Torgerson, Methods of Scaling (New York: John Wiley 8 Sons, 1958): William 3. Goods and Paul K. Hatt, Methods in Social Research (New York: McGraw-Hill Book Cmmflififia—Wfiite Riley, John W. Riley, Jr., and Jackson Toby, Sociolo ical Studies in Scale Anal sis (New Brunswick, N.J.: Rutgers University Press, 1954). recreatior given set as, say, I done in th each case, informed j The p] one who dis exactly wha insert what. The final t. workability. Four wa “Entioned in boating,13 a1 °°dered part the higheSt p 48 recreation. They may agree still less on whether any given set of classes or activities ought to be scaled as, say, 1-2-3-4 or as 3-7-12-18. The most that can be done in this study is to adapt what appears to be, in each case, a reasonable measurement in the light of informed judgment. The procedure used for scaling is explicit. Any- one who disagrees with any particular scaling can see exactly what has been measured, and precisely where to insert whatever alternative measure may be preferred. The final test of the validity of the procedure is its workability. This is tested in the interaction model. Four water-based activities have repeatedly been mentioned in the literature reviewed: swimming, fishing, boating,13 and water skiing. In ORRRC Report 20, which compared participation rates of campers and non-campers, the highest participation rate was for fishing, followed by swimming and boating. Water skiing was not included (Table 2). The National Recreation Survey listed North Central Region participation rates for these activities as follows: fishing 33 per cent, swimming 42 per cent, 13Boating includes canoeing and sailing. Limited data is available on canoeing and sailing and most reports combine them with boating. boating 27 p' The Dahle 511: indicates the and boating : correlation < M21 (Table the relative camping. The the activity States that a as part of th “WY. swim by fishing, b subject: he Varied f0r 49 boating 27 per cent, and water skiing 6 per cent (Table l). The Dahle Survey, like the National Recreation Survey, indicates that swimming is more pOpular than fishing, and boating is the least pOpular (see Appendix A). The correlation coefficients from the National Recreation Survey (Table 3) are of some help, since they indicate the relative association of these water activities with camping. The Van Til Report states that swimming was the activity most preferred by all park users. It also states that a “majority of campers do not require boating as part of their recreation."14 On the basis of this summary, swimming will be given the most weight, followed by fishing, boating and water skiing. Subjective values assigned to these activities must be varied for each park, depending upon the type of water body adjacent to or within the park. In some parks lo- cated on the Great Lakes and on inland lakes, swimming can be undertaken in both. In these cases a value for swimming as an activity is applied for swimming in the Great Lakes, and an additional value derived for swimming in an inland lake. Since Great Lakes swimming usually 1‘Report of Committee on State Parks and Public Lands, op. cit., p. 11. involves quent la< lake swil In a on the G1 Boating i rivers, w limited. Sites on for boati; each of t] SVimming 1’ishihg Boating "ate! Skii Prefe Walks, hOr 1mm’9-1‘ica1 ‘ water acti‘ “dimes ‘ f” hiking (Tables 2 a hiking, Val? 50 involves relatively cooler water, more currents, and fre- quent lack of diving platforms or rcped areas, inland lake swimming is assigned a higher value. In a similar manner, fishing is given a lower weight on the Great Lakes and on rivers than on inland lakes. Boating is given a higher weight on inland lakes than on rivers, where boating capacity and maneuverability are limited. Michigan State Parks have no boat launching sites on the Great Lakes, therefore no weight is given for boating on the Great Lakes. Weights assigned to each of these four activities are listed below. Inland Lakes Great Lakes Rivers Swimming 20 16 -- Fishing 16 9 9 Boating 8 -..- 4 Water Skiing 6 --- -- Preferences for land activities--hiking and nature walks, horseback riding, and play sports--are assigned numerical values in a manner similar to that used for water activities. An examination of the previous tables indicates that both preference and participation weights for hiking frequently are equal to those for boating (Tables 2 and 3). In a few instances, a combination of hiking, walking for pleasure, and nature and bird walks results in one instan skiing (Ta' Parks are , trails (so: a weight 0 than boati1 Prefe: many State: lar1y impo: Where it i: ’50 increase Becaus appears to Weight Of 3 SeVera These activ 51 results in a higher preference than that for boating. In one instance hiking is as low as the preference for water skiing (Table 1); however, since many Michigan State Parks are of ample size for hiking and have well—marked trails (some with nature brochures), hiking is assigned a weight of 7. This places it above water skiing, lower than boating. Preferences for horseback riding are not listed in many statewide surveys and this activity is not particu- larly important in Michigan State Parks. In the few where it is available, it is given a token weight of 4 to increase the park's attractiveness. Because play equipment--swings, teeters, and slides-- appears to be important to family campers, it is given a weight of 3. Several parks contain archery and rifle ranges. These activities are not included in any of the reports reviewed, but the availability of facilities of this type can be an attraction. Therefore, each of these activities is given a weight of 1.0. A summary of the weights for land activities is listed below. Hiki} HorSc Chil< Rifl1 Seve: nicking, a considerec inherent 1‘ has not be Primary fc highly pre is only pa Table of the fif mitted. 'I each Park and a Stan tr”Y and baSed on t that they majority o Eleve activity J: recreation 52 Hiking - Nature Walks Horseback Riding Children's Play Equipment Rifle - Archery Range HNNQ Several land activities have been eliminated. Pic- nicking, although an important activity, has not been considered directly. It is primarily a day-use activity inherent in any camping experience. Camping, likewise, has not been given a weight, since this activity is the primary focus of the attraction index. Sightseeing, a highly preferred activity, is difficult to measure and is only partially included elsewhere as part of the index. Table 4 lists the activity weights assigned to each of the fifty-nine parks in which family camping is per- mitted. These activity weights are summed by rows for each park and then standardized with a mean of 0.0 and a standard deviation of 1.0. The weights are arbi- trary and admittedly allow for improvement, but they are based on the best surveys available, and it is believed that they do eXpress_general activity preferences for a majority of the camping pOpulace. Eleven of the fifty-nine Michigan State Parks have activity rating scores of 1.00 or higher. Five of these recreation sites are state recreation areas in southern saga W NE JL Michigan State Park or Recreation Area ' inc H Great Lakes - Swimm / ALGOINLAC ALOHA mUNTA I N 53 TABLE 4 NUMERICAL ACTIVITY WEIGHTS BY WATER AND LAND ACTIVITIES ASSIGNED TO MICHIGAN STATE PARKS 14 ouoom oumccmum ouoom Hobos oceans somnomuoe someone I wanna ucofimwsvm Moan ocexem 8 9 101112 13 mnemocmu I mcwunom I muo>wm 7 anaemwm I muo>wm mcwwxmuouoz I moxmq pcmacH ,6 mcwuoom I moxmq pcmacH .5 mcwgmwm I moxmq unmacH .4 mcHEEflsm I moxmq caanH .3 mcwnmwm I moxoq unouw .2 mcflfiaflam I magma Donna .1 n o m .1 r.t 9 0a .1,e e .n.t.k r a c.u r c e m_ a e r cop.R_A .98 23 - ALGONAC 52 + .49 6 8 20 16 ALOHA BALD .83 25 16 MOUNTAIN 44 + .10 16 BARAGA .78 26 - BAY CITY, 16 .98 22 - BENZIE 66 +1.18 6 20 16 BRIGHTON .83 25 - 16 BRIMLEY 65 +1.13 6 8 20 16 BURT LAKE .88 24 CHEBOYGAN 16 TABLE 4 . --Contd. 54 1 2 3 4 6 7 8 9 1o 11 12 13 14 D. H. DAY 16 7 23 - .93 DoDGE BROS. 20 16 6 so + .39 No. 4a EAST TAwAs 16 9 33 - .44 FAYETTE 16 24 - .88 PORT cus- 1 1 -2.00 TERI“-l FORT waers 16 7 31 - .54 GLADWIN 9 2 11 -1.52 GOGEBIC 20 16 6 7 2 59 + .83 LAKE GRAND 16 9 9 34 - .39 HAVEN HARRISVILLE 16 7 31 - .54 HARTWICK 9 7 2 26 - .78 PINES WALTER J. 20 16 6 2 52 + .49 HAYES HIGGINS 20 16 6 7 2 59 + .83 LAKE HIGHLAND 20 16 6 7 2 2 61 + .93 p. H. HOEFT 16 9 7 2 42 .oo HOLLAND 16 9 16 2 43 + .05 TABLE 4.-—Contd. 55 1 2 3 4 6 10 11 12 13 14 HOLLY 20 16 6 2 68 +1.27 INDIAN LAKE 20 16 6 66 +1.17 INTERLOCHEN 20 16 6 2 52 + .49 ISLAND LAKE 20 16 6 70 +1.37 LAKEPORT 16 24 - .88 LUDINGTON 16 20 16 6 2 84 +2.06 F. J. 16 2 18 -1.18 MC LAIN CHARLES 16 2 27 - .73 MEARS METAMORA- 20 16 44 + .10 HEADLEY WM. MITCHELL 20 16 6 2 52 + .49 MUSKALLONGE 20 16 6 50 + .39 LAKE MUSKEGON 16 16 6 2 64 +1.08 ONAWAY 20 16 6 2 59 + .83 ORCHARD 16 2 18 -1.18 BEACH ORTONVILLE 20 16 6 2 2 61 + .93 OTSEGO LAKE 20 16 6 2 52 + .49 PINCKNEY 20 16 6 2 2 70 +1.37 TABLE 4.--Contd. 56 1 3 4 5 6 10 11 12 13 14 PONTIAC LAKE 20 16 8 6 2 1 2 71 +1.42 PORCUPINE 16 16 8 69 +1.32 MTS. PORT 16 8 2 35 - .34 CRESCENT PROUD LAKE 20 16 8 6 2 72 +1.47 ROCHESTER- 1 17 -1.22 UTICA SILVER LAKE 20 16 8 6 2 52 + .49 SLEEPER 16 8 2 33 - .44 STERLING- 8 2 10 -1.57 MONROEa STRAITS 8 2 10 -1.57 TAHQUAMENON 16 36 - .29 TRAVERSE 16 8 2 26 - .78 CITY VAN RIPER 20 16 8 6 2 72 +1.47 WARREN 16 2 27 - .63 DUNES WATERLOO 20 16 8 6 2 68 +1.27 JoWo WELLS 16 2 1 39 " .15 WHITE CLOUD 9 -1.62 57 1 2 3 4 5 6 7 8 9 10 11 12 13 14 WILDERNESS 16 8 9 7 40 .10 WILSON 20 16 8 6 2 52 .49 YANKEE 20 16 8 6 7 57 .73 SPRINGS YOUNG 20 16 8 6 50 .39 NUMBER OF PARKS RE- CEIVING WEIGHT 25 7 28 33 46 28 27 9 31 41 5 4 aCamping not permitted. 'i' = 42 7 a .00 0 B 20.40 0 8 1.00 Michigan--Br 19 Waterloo. Wit easily accessi politan Area. west central M Ludington and 1 have lands acce Great Lakes. 2 itY'rating SOD] and Indian Lake the Upper Peniz Six of the less than‘Loo. Beach, Roche Ste Of these Park 3 58 Michigan--Brighton, Holly, Island Lake, Proud Lake, and Waterloo. With the exception of Waterloo, all are easily accessible for visitors from the Detroit Metro- politan Area. Two of the eleven parks are located in west central Michigan on Lake Michigan. These are Ludington and Muskegon State Parks. Both of these parks have lands accessible to inland lakes as well as the Great Lakes. The remaining four parks with high activ- ity-rating scores are Burt Lake in the lower Peninsula and Indian Lake, Porcupine Mountains and Van Riper in the Upper Peninsula. Six of the fifty-nine parks have activity scores of less than'l.00. These are Gladwin, McLain, Orchard Beach, Rochester-Utica, Straits, and White Cloud. All of these parks are adjacent to a water resource, either the Great Lakes, inland lake, or river, but in all cases the natural resources are not of sufficient quality or have not been develOped to permit active water-based activities. A more detailed appraisal of the natural resources is included in the next section of this chapter. Identifying Natural Recreation Resources A second major factor to be considered for the index of camping attraction is the natural environmental factor. 59 Natural-cultural features are of principal significance to campers in Michigan State Parks. There are problems, however, in measuring the way the physical environment in a park is perceived and valued, rationally or irra- tionally, for outdoor recreation purposes. The ultimate objective is to relate the outdoor activities preferred by campers to the physical environmental attributes necessary for each activity. Experience Approaches Evaluation of a landscape for a specific outdoor recreational activity such as swimming can involve some tangible attributes. The length of a beach or the qual- ity of a beach surface can be measured and expressed in swimming Opportunities that could be provided. However, appraisal of a park's total landscape in terms of its scenic qualities and aesthetic appeal involves individual human emotions and elusive intangible values. One method for measuring qualities of the physical environment is by user satisfactions or dissatisfactions of the outdoor experience. It is apparent from the results of this study that users' remarks about area quality are strictly relative and reflect personal judgments. Expres- sions of user satisfaction do not necessarily 60 measure the “quality” of an area, but instead the enjoyment of the visitor, based on a multi- plicity of factors impinging on his specific visit. This satisfaction is based on personal desires and preconceived ideas about the area, the activities available, and the facilities provided to assist in the enjoyment of these activities, as well as the natural setting and the kind of maintenance the area receives.15 A survey of the type undertaken in ORRRC Report 5 could not be done for this study. However, a future survey of user Opinions could use the results of the index develOp- ed here for comparison and analysis. Without a user survey, one can approach the tangible and intangible values or qualities of a recreational landscape by considering the many facets of recreational experience. The National Advisory Council of Regional Recreation Planning (NACRRP) in A User - Resource Recrea- tion Planning_Method suggests six values that may be used individually or in combination to measure the out- door recreational experience: physical, emotional, aesthetic, educational, social, and intellectual exper- iences.16 150. 8., Outdoor Recreation Resources Review COMP mission, The Quality of Outdoor Recreation: As Evi- denced by User Satisfaction, Study Report7§ (Washington: U. S. Government Printing Office, 1962), p. 43. 16National Advisory Council on Regional Recreation Planning, A User - Resource Recreation Planning Method (Hidden Valley, Loomis, Calif., 1959), p. 27. 61 The physical experience is probably the easiest of the six recreational experiences to recognize and mea- sure since it entails actual participation and commit- ment by the individual. This value is tangible; the recreational landscape requirements for a variety of outdoor activities can be identified and quality and resulting use of these resources measured or approxi- mated. The values of the remaining five recreational ex- periences are much more difficult to identify17 because they constitute the intangibles of the recreational experience. Depending to some extent on the quality of the physical environment, the five, particularly the 17The emotional experience, in a sense, is directly related to the physical experience and can be identified by physical reaction in activity situations such as the exaltation experienced after scaling a mountain peak. Aesthetic experiences are the most intangible and much more difficult to isolate. A mental response is neces- sary to appreciate the flatness of the plains, the depth of a mountain canyon, or the solitude of an alpine lake. Each person views these landscapes differently. Educa- tional and intellectual experiences in relation to recreation are just beginning to be recognized in our 'cultural revolution.“ Social experiences are more easily measured by planning the use of natural resources to provide a variety of situations to encourage the soc- ial eXperience; however, this involves major problems of design and land-use planning. Some campers relish the social contact with neighbors in a campground, while others desire the solitude provided by a wilderness tract. 62 emotional and aesthetic experience, involve the complex- ities of human personality. Investigations into these two experiences by social scientists are scarce;18 however, the need for studies of this type is recognized19 and ORRRC Report 520 is close to being an investigation of this type. The results of one continuing study on wilderness perception and use recently published21 deals with the Boundary waters Canoe Area in Minnesota. It is not directly applicable to this analysis but the research method is a good one and is preferable to the approach that has been feasible here. The method involved an investigation of the users' perceptions of the wilderness environment and the outdoor activity uses taking place. In the 18As this study was completed, another study worthy of mention was concluded. Among other analyses, it in- volves a discussion of the recOgnition of users' images of a satisfying experience within a recreation landscape. Unfortunately it was completed too late to be incorpor- ated into the discussion above. The study is Clare A. Gunn, A Concept for the Design of A Tourism -- Recrea- tion Region (Mason, Mich.: The B. J. Press, 1965). 19Ibia., p. 53. 20ORRRC Report 5, 0p. cit., p. 43. 21Robert C. Lucas, ”Wilderness Perception and Use: The Example of the Boundary Waters Canoe Area,” Natural Resources Journal, III (January, 1964), p. 394. 63 present analysis, the outdoor recreation activity experi- ences desired by campers are matched with the estimated ability of the natural environment to provide the re- sources necessary for these activities. Resource Approaches What physical attributes of a park influence indi- vidual preference for one kind of landscape over another for various outdoor recreation activities? One of the earliest professional discussions of park attributes was a 1937 article22 in which six types of park I'appea1" were listed and briefly discussed. They are (l) aesthe- tic appeal to sight, sound, touch, and smell; (2) the curious and unusual, such as mineral springs and caves; (3) scientific interest of a geological or ecological nature; (4) primitive appeal-wilderness concept; (5) historic, and (6) utilitarian appeal, or outdoor- activity centered. These attributes of a park are still applicable today, but several more recent listings com- bine the attributes differently. Basically the attri- butes are sub-divided in terms of natural and cultural features of the landscape. 22S. F. Brewster, “Park Appeal,“ Planningand Civic Comment (October-December, 1937), p. 38; fa Just! 27$ '90 /2: 9.: rt- 31‘ 64 Baker lists the following as major environmental factors for recreational land use:23 I. II. Cultural environmental factors A) Location in relation to population B) Historic association Natural environmental factors A) Physiographic pattern (1) Water access and shoreline conditions (2) Topographic configuration (3) Drainage conditions B) Vegetative pattern (1) Composition or type (2) Form - nature, crown cover, etc. C) Biologic pattern (1) Bird and animal pOpulation (2) Fishing conditions D) Climatic pattern (1) Temperature (2) Wind (3) Precipitation (4) Amount of sunshine, fog, etc. 23W. M. Baker, "Assessing and Allocating Renewable Resources for Recreation,“ in Resources for Tomorrow: A Re ort to the De artment of or ern airs an National Resources (Montreal, October 23-29, 1961), p. 998. “ or S fica Comb to t and acti ifie. thre. acti' int reso: EVal: hilh Wate and; of RI 31?. 1949] Land: in 1: W1 nois‘ 65 The literature contains many examples of a general or specific nature that are similar to the above classi- fications.24 Combined Approaches Several persons have gone a step further in relation to the physical environment by attempting to classify and rate the quality of natural features relative to activity preferences of the users. ORRRC Report 5 class- ified twenty-four recreation sites in the nation in three ways: (1) by administrative agency, (2) by major activity attraction (this was done for Michigan Parks in the previous section), and (3) by type of physical resources.25 The bio-physical character of sites were evaluated according to tapographic relief--flat, rolling, hilly, or mountainous; water resources--contiguous to water bodies, rivers or streams, or absence of water, and by vegetation--whether barren, pastoral, or forested. 24Some examples are Robert D. Campbell, The Geogpaphy of Recreation in the United States (unpublished Ph.D. dissertation, Department of'Geography, Clark University, 1949), p. 30; C. Frank Brockman, Recreational Use of Wild Lands (New York: McGraw-Hill Book Co., 1959), pp. 113- I27——and Phillip H. Lewis, Jr., Recreation and Open Space in Illinois (Urbana: Division of Landscape Architecture and the Bureau of Community Planning, University of Illi- 3018' 1961), Po 9. 25ORRRC Report 5, op. cit., p. 13. 66 In the Meramac Basin Research Project, Ullman, Boyce, and Volk discuss the physical quality of reser- voirs relative to their use for specific recreational activities.26 Their six factors are sub-divided for purposes of judging a site or area. Major Factors Sub-Components I. Appearance and Landform--310pe and relief character of the Soil and bedrock--beach type shoreline Vegetation--type of trees and extent of cover Cultural features II. Physical charac- Size teristics of the Shape lake Fluctuation in water level Depth of water III. Water quality Turbidity Pollution Temperature IV. Fishing quality V. Climate VI. Accessibility Included in their study is a discussion of the physical factors on outdoor recreation activities at reservoirs. 26Edward L. Ullman, Ronald R. Boyce, and Donald J. Volk, The Meramac Basin - Water and Economic DevelOpment, Vol. III: Water Needs and Problems (St. Louis, Mo.: Meramac Basin Research Project, Washington University, December, 1961), Chapter V, p. 42. Portion for app Parks. rate I.‘ These rate 0 order index scales ments. if a 1 Prese: until binar ontdo refle udtio 67 Portions of the ensuing analysis draw upon this report for appraisal of physical qualities in Michigan State Parks. Rating the Qualities of the Recreation Environment Both of the before mentioned studies attempted to rate the physical qualities of the recreation landscape. These ratings were of a subjective, generalized nature. None of the studies reviewed have attempted to rate or scale the qualities of physical environment in order to combine thses scores into an overall rating or index of attractiveness. In this study actual interval scales will be used to measure some environmental ele- ments. Other variables have binary values, i.e., 0.00 if a physical variable is absent, and 1.00 if it is present in a park. A measure of this type is crude, but until other means are devised for scaling these variables, binary values are necessary. To analyze each physical attribute relative to the outdoor activities and to select a numerical value that reflects its quality for activities requires a step by step inventory and evaluation. Several inventory eval- uation schedules from national agencies and commissions 1‘13] have be (Append. numeric sheet 1' park. ing of numeric Not all were us tried j causatj 68 have been studied to construct the inventory sheet (Appendix D, Table 17) and to obtain insights into numerical scales that have been used.27 The inventory sheet records as many variables as possible for each park. Appendix D, Table 18, contains a complete list- ing of the variables inventoried, type of scale used, numerical range of variables, and source of information. Not all the variables inventoried and initially measured were used in the final index of attraction. Some, when tried in the index, proved not to be particularly strong causative variables. Below is an outline of the variables in the natural environment considered important in the camping experi- ence and in outdoor activities desired by campers. Each variable was analyzed for its significance in undertaking outdoor recreational activities and then classified for use in the aggregate index. See Appendix B for a detailed 27For examples see ORRRC Report No. 5, op. cit., pp. 52-583 U. 8., Bureau of Outdoor Recreation, Inventory, Classificatipn, and Evaluation of Existing_0utdoor Recrea- tiSn Areas and Facilities, Form B.O.R. 8-73 TWashington, D.C.: Department of the Interior, 1965); U. S.,Bureau of Land Management, Recreation Site Inventopy_and Evaluation, Form 4-1644 (washington, D.C.: Department of the Interior, June, 1963): U. 8., Forest Service, Work Plan for the National Forest Recreation Survey - A Review of the Qutdoor Recreation Resources of the National Forests (Washington, D.C.: Department of Agriculture, August, 1959), mimeographed. 69 discussion of each element and the judgment rating se- lected. I. Terrain - SlOpe, local relief, landform types - falls, cliffs, springs, sand dunes II. Size of park III. Vegetation - Evergreen, mixed evergreen and deciduous, deciduous, barren, extent of virgin timber and wilderness characteristics, amount of shade in the campground IV. Wildlife V. Climate - Average daily air temperature in July VI. Water resources - Size of water body (including interlake connections), aver- age daily water temperature in July, frontage of inland and Great Lakes shorelines, front- age on rivers and streams, water quality - amount of pollution, fishing quality and success VII. VIII. 70 VII. Beach characteristics - Length and width of dry beach, composition of dry and wet beach, extent of dropoff for wet beach VIII. Cultural features - Historic and archaeologic, contemporary (Straits of Mackinac Bridge, shipping on Great Lakes) Identifying Facilities and Services Outdoor activities undertaken by campers are also influenced by the facilities and services available at a park. A discriminating camping pOpulace will seek out parks that offer facilities, services, and environments which will hopefully fulfill their camping expectations. Certain basic facilities and services have been demanded and are currently expected to be available in any Michigan State Park. Some of these demands for services and facilities are necessary for reasons of sanitation. Others are necessary in order that certain outdoor activ- ity desires can be met. A select few such as showers are obviously assets in the camping experience. The basic facilities and services must be included as an 635801 those more I in ter may nc simila ment c tions which includ resour well a daily I not de and pi Can he "ate: as can Can be 25 71 essential part of the attraction index, while a few of those that could be considered assets also are included. Facilities and services have been interpreted in more than one way. ORRRC Report 20 discusses facilities in terms of recreational sites or entities which may or may not include an entire park or forest area.28 A similar definition of facilities is implied in the state- ment of Perloff and Wingo.29 Essentially, these defini- tions refer to the availability of a physical area within which outdoor activities can be undertaken. This also includes man-made structures or modifications of natural resources which are necessary for some activities as well as services that may or may not be amenities for daily living. For purposes of this study, however, facilities are not defined as the entire park, but as the structures and physical improvements within the park. Services can be defined in terms of utilities such as toilets or water supply and in terms Of concession Operations such as camp stores and boat rentals. In addition, services can be recognized as the collection of refuse or the 28ORRRC Report 20, op. cit., p. 7. 29Perloff and Wingo, Op. cit., p. 89. 72 availability Of a naturalist. Viewed in this manner, it is obvious that the quality of facilities and ser- vices is an important aspect Of the camping eXperience. To a limited number of campers, those who like to ”rough it," the presence of too many facilities and services are undesirable. The demands of these campers tend to be satisfied better in the less-develOped state and national forests Of Michigan. It should be re- emphasized that the index of camping attraction developed for this study is intended to reflect the present condi- tions applicable to a majority of campers in Michigan State Parks. What facilities and services are preferred by cam- pers in Michigan State Parks and should be included in a measure Of a park's camping attraction? ORRRC Report 5 contains a list of such facilities and services, including a discussion of user Opinions of these facilities and services at twenty-four recreation areas in the nation.30 The respondents in this report were asked to indicate satisfaction or dissatisfaction with the following: 1. Water Supply 4. Marked Nature Trails 2. Campgrounds 5. Concession Stands 3. Parking 6. Signs and Information Trails 3oORRRC Report 5, Op. cit., p. 38. 73 7. Rental Facilities 10. Roads 8. Toilets 11. Tours and Organized 9. Boat Docks and Ramps Groups 12. Other (Write in Comments) This survey was not limited to the camping pOpulation, but included day users, consequently the preferences of campers are difficult to isolate. The listing, however, provides an excellent starting point for evaluation of facilities and services according to user Opinions. All eleven Of the above facilities and services were care- fully evaluated for use in the camping attraction index, but only campground characteristics, marked nature trails, concessions, rental facilities, and boat dock and ramps are actually included in the index. Explanations of the omissions are noted in Appendix C. The Demand Component-~Origip Areas Of Campers For the second component, origin areas of campers, or P, two sets of data are used. The first is an empirical estimate Of the prOpensity Of the pOpulace in each origin area to go camping. It is based on the participation rates for camping as listed in the National Recreation Survey.31 This estimate of P for each county 31ORRRC Report 19, Op. cit., p. 126. IL 74 is explained in Chapter V where it is utilized in the model. In final tests Of the model in Chapter V, the estimates are replaced by a second set Of data, the actual 1964 values of P for each origin area. This set of data is obtained from camping permits and is listed in Appendix D, Table 19. The Distance Component--Trave1 Time between Parks and.Origins In addition to the internal attractiveness Of parks, a camper's selection is dependent upon the spatial arrange- ment Of parks. State parks, like shOpping centers or individual retail establishments compete with each other for campers. Clawson has characterized the relationships between parks very well when he states: . . . all recreation areas or resources are in varying degrees substitutes for one another, and attendance characteristics at one area are con- ditioned by the existence and characteristics at others. If the different resources or areas are highly similar, each area is then almost completely competitive with each other area. If one had water recreation and another did not, it might be argued that peOple would go to each park inde- pendently Of the others. In practice the situation is almost always somewhere between these extremes. That is, it is probable that all areas accessible to a given pOpulation are to some extent competitors or rivals, but also to some extent are independent of one another. The degree of substitutability or 75 competition between areas will in large part depend on the inherent attraction Of the area and its location.32 In the travel model it is assumed that the magnitude of the camping attraction index for each park will, when combined with the distance component, provide the competitive and/or substitutability criterion for the interaction model. The distance between the origin base of campers and parks can be a factor that determines a camper's choice Of one park as Opposed to another. Distance, like the attraction of a park then, is a function that describes the prOpensity of the camper to travel tO various parks. Such a function can be eXpressed in physical units, monetary units, or temporal units. The third component used as a function Of distance is "over the road' travel time, or time-distance.33 Research results Of similar 32Clawson and Knetsch, Op. cit., p. 258. 33'The concept of time-distance is defined as the time required to travel a specified distance. When a certain location is said to be '. . . ten minutes from downtown', the emphasis is on time rather than on dis- tance. TO many peOple, especially in larger urban areas, a point to point orientation in terms Of time is more meaningful than distance orientation, and the question 'How much time will it take to reach my destination' is Of greater concern than 'How many miles is it to my destination?‘ In fact, the spatial arrangement of shOpping patterns and visits tO alternative locations 76 models Of this type tend to verify driving times as the best distance measures.34 Intuitively, travel time appears to be the best measure. For many vacation campers, time is at a premium because vacation time is limited. Campers can be expected to select parks not only on the basis of their attractive qualities but on the basis Of travel time to a selected park. A map showing major highways and mileages between Michigan State Parks and county pOpulation centers was available from the Michigan Outdoor Recreation Demand Study (MORDS).35 The pOpulation centers and parks are are usually decisions which are all based on an attempt to minimize ones total travel time. However, a minimiza- tion Of the total travel time between two points does not necessarily minimize the total distance traveled. In order to determine a minimum travel time, one must consider variations in traffic congestion, quality of the highway or street surface, legal speed limits, and other impediments to traffic movement.“ Donald A. Blome, A Map Transformation Of the Time-Distance Rela- tionships in the Lansing Tri-County Area (Institute for Community DevelOpment and Services, Michigan State University), 1963, p. l. 34D. S., Department of Commerce, Calibrating and Testinga Gravity MOdel with a Small Computer, Bureau of Public Roads, Office Of Planning (Washington: U. S. Government Printing Office), 1963, p. II-7. 35This map was compiled by Jack Ellis and C. E. Tiedemann Of the MORDS staff. Major routes were se- lected by inspection Of the 1963 Highway Traffic Flow Map published by the Michigan Department of Highways. Highways with fewer than 300 vehicles were omitted, unless necessary for a park-origin connection. connected by map and esti nodes and st author calcu for DDRDS (s origins--sev counties,37 Politan Stat A major function of derivation o 77 connected by the most direct routes. Using the mileage map and estimated driving times for each link (county nodes and state parks or highway intersections), the author calculated the travel time for this study and for MORDS (see Figure 5).36 There are eighty-eight origins--seventy-one Michigan counties or groups Of counties,37 and seventeen out-of-state Standard Metro- politan Statistical Areas (Appendix D, Table 19). A major problem in utilizing travel time as a function of distance in the interaction model is the derivation Of a weight for this component. Probably one reason why travel models Of this type have not been develOped is the lack of data leading to information on the |'range" of travel of campers. This “range” or 'distance decay function“ for camping in Michigan State 35The average speed in miles per hour for each link is an estimate based on the type of highway, two- or four-lane or limited access interstate, the tOpography traversed, and the urban or rural land use adjacent to each link. Major highway links without limited access which traversed urban areas were assigned speeds as low as 25 miles per hour in a few cases. Once the speed of traversing each link was estimated, it was divided into the highway mileage to obtain travel times. 37Several counties in Michigan with populations of less than 10,000, primarily in northern Michigan, were combined into one node to reduce the number of links. This was necessary for the systems model used for MORDS. To facilitate comparison of the two travel models at a later date, these counties were not separated for use in this study. ‘5' «I q t ORIGIN NODES (z) ORIGIN NUMBER FROM TABLE 0 STATE PARK -.5 - TRAVEL TIME N HOURS m OF --.-_ COMBINED COUNTY ORIGINS 78 FIGURE 5 TRAVEL TIMES BETWEEN ORIGIN NODEs AND MICHIGAN STATE PARKS 47‘ SOURCE: ADOPTED FROM MICHIGAN OUTDOOR RECREATION DEMAND STUDY - SVSTEM LINEAR GRAPH O 20 4O 60 80 MILES 46° 45' 42° Parks has n mate of thi basis of a probability Three supply. den defined. '. Threfi elem in this ca campiBQ‘-a facilitie, planning . assgciate no one he 79 Parks has never been derived. As a consequence, an esti- mate of this parameter is Obtained in Chapter V on the basis Of a sample of Michigan camping permits, using a probability method and a regression analysis. Summary Three components of a recreational travel model-- supply, demand, and distance--have been identified and defined. The primary component of the three is supply. Three elements that influence the supply component or, in this case, the attractiveness of a state park for camping--activity Opportunities, natural resources, and facilities and services--are standard considerations in planning state parks. Although these elements and their associated variables appear frequently in the literature, no one has attempted to combine them into a comprehensive measure Of park attractiveness. A review of outdoor recreational survey and planning studies served to identify the most important variables identified with each element. Recent state and national surveys were analyzed to recognize the most important outdoor activities preferred by campers. The same literature allowed for an estimation of the importance of each scale. outdoor a The supply co Resource . a numeric.- Chapter I\ a manner t attraction Final m“181 were consists 0] Camping par and Primary of Congerva an estimate individual 1 Ponents are in Chapt er ‘ 80 of each outdoor activity and the application Of weighting scale. Each Michigan State Park was scaled as to total outdoor activity Opportunities. The variables of the remaining two elements of the supply component were then identified and discussed. Resource and facility and service variables were assigned a numerical value for use in a factor analysis. In Chapter IV these two elements are grouped and scaled in a manner that is useful for constructing a camping attraction index for each park. Finally the remaining two components of a travel model were Operationally defined. The demand component consists of two separate sets of data--estimates Of camping participation Obtained from a national survey and primary data Obtained from the Michigan Department Of Conservation. The distance component is based on an estimate Of travel time between origin nodes and individual parks. The manner in which these two com- ponents are used in the travel model will be discussed in Chapter V. CHAPTER IV THE INDEX OF CAMPING ATTRACTION Given the many variables inventoried for each Michigan State Park, the next step is to combine them into a meaningful index that will reflect the attraction Of each park for camping. In order to reduce the vari- ables measured into a smaller and more manageable num- ber of dimensions and to analyze the magnitude of asso- ciation between them, multiple-factor analysis is used. This method does not require selection of independent or dependent variables. It is, rather, a technique for finding the similarities or interrelationships among a large number of items which are not clear by inspec- tion. The assumption is that the intercorrelations of several variables may indicate distinct groupings called factors. Thus factors are representative of the 81 82 combined traits or characteristics of the variables.1 In this study, factor analysis is used to group a large number of variables related to camping into a relatively few explanatory factors. Judgment is present since results of various models depend on the original scaling of the variables.2 In a given scale, however, the criteria are relatively clear and objective for eval- uating any one of the models. The selection of the number and type of factors is determined by which (a) is most realistic in terms of acceptable theory and (b) eXplains the highest prOportion of the total variance in the data. Once a particular model is accepted, the results are used to construct a numerical index of park quality. 1For a more concise explanation of factor analysis, the reader is referred to Hubert M. Blalock, Social Statistics (New York: McGrawbfiill Book Co., 1960), pp. 383-389. Blalock describes factor analysis, on p. 383, as “a technique which can be used to take a large number of operational indices and reduce these to a smaller number of conceptual variables." A more de- tailed and sOphisticated discussion of factor analysis can be found in H. H. Harmon, Modern Factor Analysis (Chicago: University of Chicago Press, 1962). 2For example, in a study using the same variables but different scaling, somewhat different numbers and types of factors were identified. See Jack B. Ellis, “The Description and Analysis of Socio-Economic Systems by Physical Systems Techniques" (unpublished Ph.D. dis- sertation, Dept. of Electrical Engineering, Michigan State University, 1965), p. 16. Sevent and activit numerically D, Table 18. initially 51 tion.3 It was would colla; With the thz 1“ Chapter I aha1Y8is4 Wi not, howeVer 83 The Factor Analysis Seventy-two natural-cultural, facility and service, and activity variables have been inventoried and scaled numerically for each Michigan State Park (see Appendix D, Table 18). Fifty-five of the seventy-two items are initially selected as representative of a park's attrac- tion.3 It was hypothesized that the fifty-five variables would collapse into three factors (Table 5) identical with the three elements of a recreation area selected in Chapter III. Results of the initial varimax factor analysis4 with the fifty-five selected variables did not, however, strictly confirm the hypothesis of the three factor grouping--natural resource environment, outdoor activities, and facilities and services. The solution on Model One indicates a strong inland lakes 3These fifty-five variables were selected with the assistance of L. M. Reid, Professor of Parks and Recrea- tion, Department of Parks and Recreation, Texas A. 8 M. €01.19” 0 4The Fanod 4 Library program on the CDC 3600 Com- puter was used. This program produces the means and standard deviations of the variables, a correlation matrix, eigenvalues, principal axis factor loadings, quartimax or varimax factor loadings, prOportions of total variance of each rotated factor, and the communal- ity of each variable. For further details, see Factor Analysis Programs: Fanod 3 and Famin 3, Computer Insti- tute for Social Science Research, Technical Report 2 (rev.: East Lansing, Mich.: Michigan State University, September 22, 1964). 84 Ado. 0xua pedacwucowuwmomaoo comma #03 “omv masblmusumummaou Hmum3 mxma pcmacH “Hey moxmq ummuwlsowuwmomeoo comma #03 «one hannlmusumummamu noun: dogma umouu 1mm. mend mandamuammmn mo emmcmq 1mm. mummy sumaoum Anne mmxmq umouounummn mo cumsmq A~mv munch mamwm lame mama mandamnammas comma lame masmum manage mo eumamq .mn. moxmq ummnulguvwz comma «Ame madman Doom «0 nvusmq ism. mend unmeasuaoamsmomeoo momma sun imae mmsmsasomm m>smmmmumumman gov. mmxmq umouwncowuwmomfioo comma who .Hav Amvmusummm Hmouuoumflm Ammanmwum> hwy moow>umm can mowuaawocm consamm can mowuw>wuo¢ Hoopuso HH uouomm mwmonuommm AFNC masoummsmo as manna .mae amass mama Ammv mmxma occacfi mo mmmmuod Aoav mowumfiumuomumno mumsumoawz .ms. mammaoum um>mm Ame umnswm mamua> Rome mmmucoum mmxmq ummuw “my coaumummu> Ava. madam any mommuam «one and. mxooaum>o can mumwau “we oomuuam xumm and. mmemumm .m. mmsamu Homes Acme musumuomEmu mush mmcum>4 made dawn xumm Ammanuaum> adv nauseouw>cm Amusuuz H Houomm mamonuommm 0H0§ afimxndflfi HOHOflM HmHflM Gd. “flag aflHnflflHfl> 0>HMI~nUMflhllom WEB 85 .ma magma .n mammmmmm cw cowumcmwmoc Hmowumesc name» on ummou manmwum> sumo noumm muonEszm Ahoy oxma osmasfluamucmu umom AmNV auwoauuooao saws nonwomamo Ave. mumwm Away muossmq 1mm. mxma mamammumummmmmsq .mNC non no mmmamummmaaoa Amvv noxmq umoHUIUHmsmoMflq Aewv wxwd vcddawlwmaosnumm .mm. mumsomm «Hwy ououm pcaoummamu .vv. magma uoouuuomsonnumm A~v mouammfimo no Honasz immammaum> Nae mmom>umm mam moanedmomm HHH uouomm mammmmommm “any Ho>wulmaasoc5oa umom ion. mammxmummms Rom. um>muuamamsouommo mammmsm ism. mama mamamaumummms mammma>mm mo ommmxm .~se mama mamamanmmmmmmuommo mmsmmsm 1mm. mama mamammummsmommma mmom .mm. momma mmmoouaommsmuommo mmammmm imme oxms mamammu.m om mommmmsm momma mos 1mm. mama mamammnmmsasmzm .mve momma ommuon.m om mommmmam momma mm: “mm. mmxmq ommuoumaassasm .pucou I HH Houonm mammnuomwm mflflflOUllom mqmdfi 86 factor combining all natural features, water activites, and facilities and services associated with inland lakes. However, this factor also includes high loadings on Great Lakes variables, suggesting a total water resources factor. The second strongest factor is the natural environment, excluding variables on inland lakes and Great Lakes but including river and stream characteris- tics. The third combines all natural features and facility and service variables related to the camping experience, such as shade in the campground, showers, laundry, etc. This grouping also includes Great Lakes variables with loadings equal to those appearing on the first factor. Solutions of the first model reveal a fault in the hypothesis. Analysis of the three factor model indicates that the natural resource variables do not group as one factor but are distributed among all the factors--a distribution which, upon inspection, is logical. For the first factor, while focusing on inland lakes, also includes activities and facilities related to the water resource. The second factor is essentially a scenic one of land features, but includes river features not included elsewhere. The third, camping amenities, is an 87 acceptable activity factor consisting of natural resources and facilities related to the camping experience. Analysis of the four-factor solution of the same model is more encouraging (Table 6). Essentially the same factors are evident that occurred for the three- factor solution, but the four-factor solution removes the Great Lakes characteristics from the third factor and groups them in the fourth. Great Lakes variables are still loading on the first and fourth factors however. If four factors explain more variance than do three, might not five or six do better? To answer this question, five- and six-factor solutions of Model One were analyzed. A five-factor solution is not satisfac- tory conceptually. Many of the Great Lakes variables loaded equally on the fourth and fifth factors. The six-factor solution is similarly unsatisfactory in that its two dominant variables are rifle and archery ranges. These facilities are available in few parks but are not particularly important activities on a camping trip. Using the criterion of prOportion of total variance explained, the four-factor solution is preferable to the three factor hypothesis. The three-factor results 88 pawn. vnmw. vav. Hmmw. chwv. moon. mmmm. comm. ONMF. «map. mmam. Homo. mmmm. hmmm. mama. 0>w#d#0umu0#sH oak0#mam H0>wuiuh#wdmav mcwnmwh mmmwao H0>wullmcw#mom mmcwnmm amass» mamua> scammmmmm> mm0cu0paw3 madam 0on#coum mmxmq #m0uu m0w~0u Hoooq mewxam 0mm#soum H0>wm xuom mo 0wm0uo¢ mcuomoq u0#omm manmnum> ”ma n mmmwmamxm 0osmfium> mo sow#uomoum .#n0asouw>cm Hous#mz. HH u0#omm vohv.l ou0#03 0anmmw>mz mmNm.| 0#wm maanocsua #mom momm.l H0#:0H #0om mvno.l n#ow3 noo0m mmpm.| .m 0# 00cM#nwo nom0n #03 moms.3 0mnonn#0m mocm.l cum500MMq somm.u momma mamammnumumammm mmnmmnm oamo.l cow##oomsoo som0n #03 mNNm.I .mfi0# H0#03 mash 0mou0>4 oemm.| cow#wmomaoo som0n mun humm.u mcwwxm u0#03 mmvm.l mcdaawzm mcwvmoq u0#omm 0H3 m> mom a mmmnmamxm 00smauo> mo sow#uomoum .mmxmq mamHmH. H homomm soa#m#ou Xmawum> A00Hnmwuo> mm“ soa#nHOm u0#omulusom H0008 #ouwm Mom mmswpmoa H0#omhll.m mummy 89 mom a ©0Gw0amxm 0o:0wu0> H0#OH HNNM. £0003 mo £#mc0q «ham. n#pw3 nom0m meow. m0qsc ocmm mmvm. cow#wmomeoo som0n #03 nape. 0mso:n#mm omen. m0an #00lem#wamsm mcwnmwm Namh. oumam0qu Hume. H0wm wwwomoq u0#omm 0Hnmwum> mod n omcsmamxm 00cmwum> mo co«#uomoum Ammxmq mmmuo. >H uomomm «vom.l 0H0#m name. 0mcou mu0nuum «who. mawmu# 0vaum Nome. 0mcmn 0HMflm onam. 0Hs#mu0ma0# Mash 0mmu0>¢ hmam.l ocsoumgemo cu 0vonm mhoo.| m0#wmmsmo mo H0nssz Naoh.n a#wowu#o0a0 c##3 u0#Hm mmow.| mundane ~mom.l m#0awoa avmm.l mumzonm mswcmog H0#omm 0Hnmwum> mad I ommamamxm 00cmwum> mo sow#uomoum Am0w#«c0sanucwmemuv HHH u0#ouh .Q#GOUII.w mqmdfi 9O explain 48 per cent, while the four factor solution ex- plains 54 per cent of the total variance (Table 7). The five-factor solution, accounting for 57 per cent of total variance, and the six-factor solution, for 62 per cent, explain a higher prOportion of variance, but in terms of the grouping of variables they are unacceptable on intuitive grounds. Consequently, the four-factor solution is judged to be the most desirable to use. Several additional models are conducted to confirm whether the four-factor solution is best and to increase the amount of total variance eXplained for four factors. These models are summarized in Table 7. Twenty-four variables are deleted in Model Two, including all activity variables.5 This model proves to be the best solution in terms of the total variance ex- plained (64%) by the four factors (Table 7). But in spite of the large amount of the total variance explained by the four factors on Mbdel Two, this solution is not used in constructing the index. By using only thirty-one variables, too many important facility and service 5These variables in Model One are entered as binary values. Henceforth, activity opportunities at each park are scored separately in the index on the basis of the empirical weights established in Chapter III. 91 .H.w|0om I#coum com0n ..H.«Iomm0uom ..A.OIcom0n mo summon .manmom oases: ..QE0# new hash .0>w#0#0um Iu0#cw .2500 use .mmsmo #mom#so .m0#wmmemo #0 u0nesz mm mm mm vH om mm an N mm «m mm «H «a Na mm H @0064 o0#0H0Q mH0#omm mu0#omm mu0#omm HHH HH H m0Hnm«H0> H0nasz m m m #0 H0Uoz H0naaz com moofi>0um H0#oa u0#omm Bonn moanmwnm> he codenamxm 00cmwum> mo sow#uomoum cow#0#ou xMEAH0> I na0ooE nwuhamcm u0#omm mo 0Hn0# humeesmll.h mamaa 2 9 .H0#c0u #mon .cocama #mon ..q.fil000#souu momma ..a.a 0mm0nom .mu0>wu ..a.au»#aamnq mmaamam ..a.o can .q.wl0on0# undo nom0n #03 ..A.w 0mM#coum ..mflsw0 hmam .maamom mmaaaa ..mE0# wane .>m .850056 .mmEmo #mom#ao .m0#wn umsmo mo amasmz om mm mm ma ma ma mm m momma mommama muomoma maoaomm maooomm HHH an H mmaamaum> amasmz m m n no ammo: Honasz cam msow>0um Eoum moanmwuo> Ho#oa H0#omm an o0samamxm 00cmwum> mo so«#uomoum .6#:OUI!.5 mamafi 93 O‘HOU ”gm .a.wl0os0# umam aomma mm: .sammaz am mm mm aa ma ma mm m xumm «0 mmmmuom mm mm mm ea ma om ma « momma mommama muooomm mammomm muoaomm aaa aa a mmaamaam> umasoz m m m «o ammo: HonEsz cam u50a>0um HM#OB HO#00m Baum m0anmaum> an o0cwmamxm 0osmaum> mo s0a#uomoum .UUGOUII.5 mqmdfi 94 variables are eliminated from.the analysis--perhaps some of these should be included in an attraction index. Many of these facility and service variables--such as fishing quality, hiking trails, boat launching ramps, and boat rental--are believed to be necessary for suc- cessfully undertaking many of the activities preferred by campers. Another weakness in Model Two is the per- sistence of high factor loadings for the Great Lakes variables in the inland lakes as well as the Great Lakes factor. These two water resources provide somewhat different activity Opportunities. For purposes of the index, a stronger contrast is desirable. An effort to achieve the highest explained variance via four factors, using variables thought to be directly associated with outdoor activities preferred by campers, is continued on a trial-and-error basis. Model Three includes many of the variables previously deleted from the analysis and does not include any of the activity variables. The total amount of variance explained by the four factors on Model Three is 55 per cent, a reduc- tion from the 64 per cent achieved on Model Two (Table 7). On Model Pour, the acreage of each park is elimi- nated as a variable without substantially lowering the 95 total eXplained variance. This is done on the basis that some of the most popular camping parks are very small in acreage, suggesting that campers may find amenities desired for camping and preferred outdoor activities regardless of a park's size. The last factor analysis, Model Five, includes forty-three variables. The major difference between this and the original model is that all activity variables are excluded. Also excluded are selected natural-cultural and facility and service variables with very low communalities or factor loadings with signs Opposite to those of the majority of variables on a particular factor. The vari- ables used in Model Five produce the most acceptable combination of loading of Great Lakes variables on the Great Lakes factor and depresses these loadings on the inland lakes factor.6 The distribution of forty-one of 6The more acceptable factor loading of the Great Lakes variables on Model Five is one of the major differ- ences between Models Five and Two. If the goal is to identify the fewest variables possible for a model with the highest explained variance, then Model Two should be used for compiling the index. However, by a trial-and- error method, Medel Five proved to be the best analysis that includes the variables intuitively plausible and more acceptably loaded in the index. By using Model Five, the explained variance is reduced from 64 per cent to 56 per cent. Pragmatically, the difference between Model Five and Model Two for purposes of index construc- tion is not important. This is verified by constructing 96 the forty-three variables on each factor and their load- ings are listed in Table 8. Two variables--outpost camps and average July air temperatures at parks-~are drOpped from the analysis at this point. These two variables load negatively on the third factor and can not be used in the factor-scoring routine to be used in calculating the composite index. Given that the various models tend to explain from 49 per cent to 69 per cent of the total variance in the data, how reliable are the results for managerial deci- sions? In the light of the rudimentary state of research in this field, these percentages are acceptable. The 56 per cent explanation in Model Five, the one actually used for index construction, is without activity scores, which would presumably account for a reasonable percent- age of the remaining variance. Improvements in data reliability might further reduce the unexplained “noise“ or variance in the system by a significant amount. an index with the factor loadings from Model Two. This index, compared with the index listed in Table 11, is basically the same. Thirty-eight parks on the Model Two index went down, while twenty-one Parks went up in value. However, only thirteen parks went up or down by more than 5 index points, and only two parks--Gogebic, -ll.93 and Ludington, -lS.9--differed by more than 10 index points. 97 «NNv. H0>wu >#wamau mcwcmwm mace. acooau0>o I mmmwao oomv. OHHO#mwm mace. H0cEa# camua> name. muswumm name. c0fi#M#0u0> «mow. mm0cu0vaw3 our». madam ova». m0wHOH Hmooq vmmm. 0om#coum m0xma #m0uu mmhm. 0mm#soam u0>wm mama. maamu# mcaxam weapmoq u0#omm 0acmwum> mma u mmaamamxm 00cmaum> mo cow#uomonm “#cussoua>cm HmowmhchIpcoqv HH u0#omm .ammaamaam> as. -w¢.I m0cua panama mo 0mm0uuc mome.I 0mm#soau coc0m oomm.l mcwcocsma #mom owow.I Ao#c0u #nom «mhw.l c#oa3 coo0m oamh.I 0msocc#mm cmNm.I %#HHQ:U mcwcmwh doom.I pumam0mwn ~Nh¢.l coa#auomaoo canon #03 mmmm.l .mE0# H0#m3 mash 0mmu0>¢ boom.l co«#amomaoo c000c mun weapmoa H0#omm 0Hcmaum> mma I mmmamaaxm 0osmaum> mo coa#uomoum .m0xma pcmacuIIu0#m3v H HO#omm c0a#M#ou unfiaum> I coa#:aom u0#OMMIH90m .0>Hm Hope: you umcavdoa u0#00MII.m mecca 98 .0cw#sou mcauoom m¢m3 0c# cw p0m: 0c #0: canon 0cm .HHH moaucm co aa0>w#mm0c p0emoa m0acmaum> 0m0ca .0Hs#ma0ms0# Ham Mann 0mmu0>m 0cm mason #mom#ao «00mmouo 0H03 m0acmwum> 03# .csa mwc# Bonn cow#omu##m mo x0pca 0c# mcw#smaoo ch mmm n mmmamamxm mommaum> ammoa ov~M.I com0c mo c#mcoa mmwm.| amuse pcmm mmmm.l .mE0# H0#03 mash 0mmu0>4 mmao.: c#o«3 com0m ammm.u cow#amomeoo com0c who meos.I »#aamsv mcacmam moon.l cow#amomfioo com0c #03 ooNh.I 0msocc#mm Nth.I H0wm Hamh.l cumsm0mwa wcapmoq u0#omm 0acmaum> mma n mmaamamxm 00cmaum> mo cofl#uomoum «m0coq #m0uUIIH0#o3V >H u0#omm «mam. #c0smwsv0 ocsoummmam mean. 0HO#m were. ossonmmsmo ca 0omcm mo #c0ou0m comm. m0#aumfioo mo H0casz once. m#aowu#o0a0 c#a3 m0#«mmsmu omom. hnpcsmq ommm. m#0awoa mmnm. mu03ocm Aucaomoa H0#omm 0acmanm> maa n mmcamamxa 0oc0aum> mo co«#uomoum .mmauammemmmmamsmo. aaa uomomm .v#cOUII.m mgmdfi 99 Calculation of the Campigg.Attraction Index The camping attraction index for each park is de- rived directly from factor loadings of the forty-one variables. For example, if a park is located on an inland lake and contains all of the variables listed under the first factor, all of these variables are used in computing a single inland lake factor score for this park. It is assumed that all variables are unweighted. Then, to obtain the single factor score for the inland lake characteristic of this park, weights are assigned according to the factor loadings of each variable under the first factor. Thus a variable factor loading of .91 would result in a weight of 05.294, while a factor loading of .66 would have a weight of 01.169. These weights are determined by the formula: r W a l-r2 where r is a factor loading for one variable and W is a factor loading weight (see Table 9 for weights). The remaining three factors for the park are simi- larly assigned weights. If the park is not located on the Great Lakes, then no weights are assigned and it can be expected to receive a lower overall index. After the TABLE 9.--Formula and factor loading weights for Loading 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 100 WRAP program Formula: r r weight a :7 Weight 00.208 00.220 00.231 00.243 00.255 00.267 00.279 00.291 00.304 00.317 00.330 00.343 00.357 00.370 00.384 00.399 00.414 00.429 00.444 00.460 00.476 00.493 00.510 00.528 00.546 00.564 Loading 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 Weight 00.583 00.603 00.624 00.645 00.667 00.689 00.713 00.737 00.762 00.789 00.816 00.844 00.874 00.905 00.938 00.971 01.007 01.045 01.084 01.126 01.169 01.216 01.265 01.317 01.373 factor loading Loading 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 Weight 01.800 01.891 01.992 02.102 02.222 02.355 02.503 02.668 02.853 03.063 03.303 03.579 03.901 04.281 04.737 05.294 05.990 06.884 08.076 09.744 12.245 16.413 24.747 49.749 101 Loading Weight Loading Weight Loading Weight 71 01.432 72 01.495 73 01.563 74 01.636 75 01.714 Source: Weighting Ranking and Printing (WRAP) was programmed by James Clark, Department of Psychology, Michigan State University. 102 weights for individual variables under each of the four factors are assigned, the single factor score for each park is obtained.7 Four scores are thus derived for each park--one for inland lakes, natural resources, camping amenities, and the Great Lakes. If a park is not endowed with one of the factors, then one or more of the single factor scores will be 0.00 (see, for example, the inland lakes score for Algonac State Park on Table 10). The four derived scores are standardized with a mean of 0.00 and a standard deviation of 1.00. The results are listed in Table 10 and Table 11 with each park's activity score from Table 4. Since the attraction index is more uniform numerically if standardized around a mean of 100, with a standard deviation of 50, these additional transformations are made. Thus Aloha, with a score of 1.26 on the inland lake factor, is adjusted to 163 with the base 100 and a standard deviation of 50. A park like Bald Mountain, with a negative score on inland lakes of -.39, is 7The WRAP computer program develOped by James Clark of the Department of Psychology at Michigan State Univer- sity was used. The program computes an estimate of the factor scores for a given factor analysis by selecting sub-sets of variables which are representative of each factor (Table 9), and computes the regression between these factors and factor loadings. 103 TABLE 10.--Standardized park scoresa (;'= 0, 0 = 1)b Factors I II III IV V Inland Physical Camping Great Activity Park Lakes Environ. Amenities Lakes Scores Algonac 0.000 -.17 -.57 0.00 -.93 Aloha 1.26 .34 .93 0.00 .49 Bald Mountain -.39 0.07 -l.63 0.00 -.83 Baraga 0.00 -.27 .53 .39 .10 Bay City -.87 -.31 1.04 .59 -.78 Benzie 0.00 -.11 -1.32 -l.29 -.98 Brighton 1.40 -.27 -1.40 0.00 1.18 Brimley 0.00 -.08 .78 -.03 .83 Burt Lake 1.18 -.09 .91 0.00 1.13 Cheboygan 0.00 -.26 -l.67 -.61 —.88 D.H. Day -.75 .45 -1.46 -.50 -.93 East Tawas 0.00 -.32 .56 1.76 -.44 Fayette 0.00 .09 -l.53 -.82 -.88 Fort Wilkins -.62 .00 .53 -1.36 -.54 Gladwin 0.00 -.20 .45 0.00 -l.52 Gogebic 1.09 -.07 .73 0.00 .83 Grand Haven 0.00 -.16 .61 2.10 -.39 Harrisville 0.00 -.18 .76 -.77 -.54 Hartwick Pines -.87 .89 .54 0.00 -.78 W.J. Hayes 1.29 -.12 .81 0.00 .49 Higgins Lake 1.44 -.13 1.49 0.00 .83 Highland -.57 -.13 -1.46 0.00 .93 P.H. Hoeft 0.00 -.08 .10 -.35 .00 Holland -.85 -.33 .92 1.75 .05 Holly 1.20 .17 -1.38 0.00 1.27 Indian Lake .96 -.02 .67 0.00 1.17 Interlochen 1.32 -.23 .80 0.00 .49 Island Lake 1.24 .07 -1.48 0.00 1.37 Lakeport 0.00 -.28 .90 .13 -.88 Ludington 1.18 1.57 1.01 .53 2.06 McLain 0.00 -.10 -.65 -1.13 -l.18 Mears 0.00 -.24 .49 2.11 -.73 TABLE 10.‘ Metamora- Hadley Mitchell Muskallong Lake Muskegon Onaway Orchard Be 0rmrwille 0tsego Lak Pianney Pontiac La: Porcupine M("lnténlns Port Cresce PrOUd Lake Silver Lake TABLE 10.--Contd. 104 Factors I II III IV V Metamora- .64 -.25 .58 0.00 .10 Hadley Mitchell 1.24 -.34 .61 0.00 .49 Muskallonge .55 -.07 -1.44 -1.26 .39 Lake Muskegon -.68 .49 .86 2.34 1.08 Onaway .24 -.15 .51 0.00 .83 Orchard Beach 0.00 -.23 .61 -.44 -1.18 Ortonville .69 -.17 -1.45 0.00 .93 Otsego Lake 1.30 -.23 .85 0.00 .49 Pinckney 1.21 .25 -.35 0.00 .93 Pontiac Lake 1.48 .08 -l.60 0.00 .98 Porcupine -.70 6.65 -.72 -.31 1.32 Mountains Port Crescent 0.00 0.24 .25 -.24 -.34 Proud Lake .60 .44 -l.37 0:00 1.47 Silver Lake 1.45 .50 .78 -.01 .49 Sleeper -.74 -.26 .95 .60 -.44 Straits 0.00 .02 .58 -.20 -1.57 Tahquamenon 0.00 2.42 .77 -1.11 -.29 Falls Traverse City 0.00 -.23 1.04 -.46 -.78 Van Riper 1.41 .05 -.41 -.32 1.47 Warren Dunes 0.00 .22 .73 0.00 -.63 Waterloo 1.02 .13 -.44 .72 1.27 J.W. Wells 0.00 .59 .61 0.00 -.15 White Cloud 0.00 .01 .35 .03 -1.62 Wilderness 0.00 1.74 .71 0.00 -.10 Wilson 1.08 -.22 .76 -.32 .49 Yankee 1.38 .14 .74 0.00 .73 Springs Young 1.26 -.31 .64 0.00 .39 105 TABLE 10.--Contd. aSources: Factor scores from.WRAP program: activity score from Table 4. bOn Factors I and IV the signs have been reversed since all factor loadings were negative. c0.00 indicates the park did not include inland lakes or was not located on the Great Lakes. 106 multiplied by 50; this product is subtracted from 100, to give an adjusted score of 80.5 on this factor. The set of adjusted park scores for each of the four factors and the activity score for the park are listed in Table 11. To obtain the attraction index, each of the five adjusted scores for an individual park are summed and the total divided by five to preserve the standardization. In the absence of strong evidence suggesting any other weighting, this provides an equal weight for each score. The indexes derived for each of the fifty-nine Michigan State Parks where camping is allowed are listed in Table 11 and mapped in Figure 6. Revision of the Attraction Index Use of the camping attraction index in the inter- action model may inadequately represent the pulling power for each park. An apparent weakness of the index is that it considers only internal (site) characteristics of each park; it does not actually take into account attractions outside the confines of a park (situation). Thus, attendance at a few parks, particularly in the Upper Peninsula, may be underpredicted by the travel model. The location of parks situated within or near recreational attractions, or a complex of attractions, 107 n0sam m.pso o.mmm o.am oo.c o.p~a m...a m.mm aoasooma e.oso m.mmm c.mm m.am o.mma °.am oo.o maaaomaummm m.aoa o.mom m.om o.mo~ m.oma o.~m oo.o mm>mm mamao a.maa m.osm m.ama oo.c m.mna m.ma m.mma oaammoo ~.pmo o.mm~ o.m~ ao.o m.-a o.cm oc.o masmmao a.ms° m.omn o.mm o.~m m.m~a c.coa o.mmmmaaaae mama m.mmo c.m.~ o.mm o.mm m.m~ m.moa oo.o mummsmm m.pmo o.mmm o.mm c.mma o.m~a m..m oo.o mmama mama a.mm° m.omm m.mm o.ms c.3m m.-a m.~m mma .a.a m.m.o o.m- c.mm m.mm m.ma o.sm oo.o mmmaoamao m.aaa m.mmm m.mma oo.o m.mva m.mm o.ama mama mama «.4sc o.~sm m.mm m.mm o.am o.mm oo.c amasaum o.mmo o.mam o.mma oo.o o.om m.mm o.oaa moaamaua o.mmc o.ma~ o.am m.mm o.mn m.ma oo.o mammma s.mao m.mm. o.am m.a~a o.~ma m..m m.mm moao sma n.5mo m.smm c.mca m.aaa m.m~a m.mm oo.o momuma ca0#csoz m.omc o.mm~ m.mm oo.c m.ma m.ma m.om mama «.moa o.mam m.m~a oo.o c.mma o.mm o.mma maoam ~.mq° c.ma~ m.mm oc.o m.as m.am co.o ommomam maznvaH. > OH H mwhoom MOVE...” mwfiurwfimfid .QOHH>GH nflxflfl u—Hdm momma .maoo ama>amom ommao mmamemo amoamaam mamaaa 25m > >a aaa aa a .cm a o .coa u.w. mmuoom aamm mmommhmmru.aa mamas 108 o.omo c.oom m.m¢a oo.o m.s~ m.am m..ma maaa>momao m.hmo o.mmm m.am o.ma m.oma m.am oc.o momma mammoao n.4mo m.asm m.a¢a om.c m.m~a m.~m c.~aa smsmmo m.caa n.4os c.4ma o.sa~ o.mma m.m~a o.mm mommammz 0xma p.am° m.mo. m.maa c.>m o.m~ m.ma m.m~a mmmoaamammz o.ooa o.c°m m..~a oo.c m.oma c.mm o.~ma aammomas M0Homm s.omo m.mmm o.mca oc.o m.m~a n.5m c.~ma umaosmmmz «.mmo m.mpm m.mm m.mo~ m.a~a o.~m oo.o mamma m.a.o o.a.~ o.am m.m. m.em o.mm oo.o mamaoz «.mma o.saa o.mo~ m.m~a m.oma m.msa o.ama mommmamma o.mno o.omn o.mm m.mm o.maa o.mm oo.o muoamama o.~mc o.oma m.mma oo.o °.m~ m.moa c.~ma mama mamama m.moa o.aam m..~a oo.o o.oma m.am o.mma mmmooaammma m.poa a.mmm m.mma oo.o m.mna °.am o.m.a mama mmamma m.~mo °.mm. m.mma oo.m o.am m.moa o.oma maaoa ~.maa o.msm m.~oa c.sma o.mma m.mm m.nm mamaaoa m.mso o.mmm c.ooa m.~m o.moa o.mm co.o mamoa .a.a a.mso m.mam m.mma oc.o o.s~ m.ma m.m~a mamammaa m.maa m.amm m.a.a oo.o m.msa m.ma o.~>a mama mmammaa n.4oa a.m~m m.e~a oc.o m.cma o..m m..ma momma .m.z m *ku09 > On. H amp—00m amu—QA mmwuafiflfid .GOHHbfim MTV—dd u—Hflm momma .maoo ama>aooa ommuo mmaasmo amoamama mamama 25m > >a aaa aa a .6#GOUII.HH Manda 109 m.aao o.aa¢ m.maa oo.o c.~ma m.mm o.mma ammo» m.aoa o.ama m.mna oo.o c.5ma c.soa o.amammmaaam mmamma a.aoa m.mom m.m~a oo.o o.mma o.mm c.4ma momaaz o.mmo o.omm c.4m o.ma m.mna o.nma oo.o mmmmummaaz ..emo o.sm~ o.ma oc.c a.~aa a.cca oo.o mmoao mmamz o.mao c.mam m.~a m.aoa m.cma a.cs oc.o maamz .z.m m.mmo o.mam m.mma oo.o o.mp m.moa o.ama ooaamamz v.°mo c.~a4 m.mm o.mma a.mma c.aaa oo.o momma mmaamz a.mca m.m~m m.msa cc.o m.m~ a.~oa o.msa amaaa mm> a.sso m.mmm o.am o..m o.~ma m.ma oo.m moan mmam>maa maama 4.4oa o.-m m.mm o.ps m.mma o.a- oo.o mommsmammma ~.mmo o.ma~ m.a~ m.mm o.m~a o.aoa co.o mmamaoa a.aoa m.acm o.mn o.oma m.ema c.5m o.nm amammam o.~ma o.omm m.e~a m.ma o.mma e.m~a m.~aa mama am>aam «.amo o.smm m.msa oo.o m.am o.-a o.oma mama mmoaa n.45o m.amm o.ma c.ma m.~aa c.mm oc.o ammommao mama aca0#cnoz n.~ma a.aam o.mma m.am c.vm c.~m. o.mm mmaamoaoa «.mmo o.mv. c.a¢a oo.o o.o~ o.mca o.mna mama omammoa m.ooa o.~cm m.mma oo.o m.~m m.~aa m.oma ammaomaa o.moa o.mmm m.¢~a oo.o m.~ea m.mm o.mma mama ommmmo m¢HfluOB > 3 H awn—00m mwxmq aflfluflcmfid .GOHHbflm 0030A u—me x0ocH .aaoo h#a>a#oc #m0uo mcwmamo Hmoamhcm ocmacH saw > >a aaa aa a .O#GOUII.HH mammfi 110 M as as" 87° 35° 35° 84' 93 . I I I T I I I I 48 I— .a 48 \‘ FIGURE 6 CAMPING ATTRACTION INDEX WITHOUT CAMPSITE RATIO \ MICHIGAN STATE PARKS Iss4 47‘ .1 47° \ — 46° -— 45° “'I— —-I 44° 9 151 AND ABOVE 0 0 150 TO ‘0‘ O 43‘ I- 100 TO O .1 so AND O BELOW SOURCE: CARLTON 8. VANDOREN. MICHIGAN OUTDOOR RECREATION _, 42. 42° » ocmuo smov. June. was ~ 0 110 «1) “1) so MILES i 1 1 I I I I 111 may constitute an additional pulling power for these parks. Examples are Fort Wilkins in the Keweenaw Peninsula and Brimley near the Soo Locks. This omission cannot be fully evaluated at this point. One modification, however, recognized as essential in improving the index, is the use of a campground capacity weight for each park. When the index for each park is multiplied by the ratio of the number of camp- sites in that park to the mean number of campsites in the Michigan State Park System (181 campsites), the attraction index should be a much more successful com- ponent in the travel model.8 This is analogous to using a location quotient. The index weighted by the campsite ratio is listed in Table 12. Notice that parks with high attraction indices, when weighted by this capacity figure, generally are elevated in attractiveness. Examples are Ludington, Holland, and Higgins Lake. Some parks, such as Gogebic and Otsego Lake, remain the same. Others, such as Indian Lake and Porcupine Mountains, are lowered. 8FromTable 8 it can be seen that the number of campsites in each park was included under Factor III-- Camping Amenities. The camping capacity of each park proved to be a stronger component within the system than anticipated, and the relatively low factor loading on Factor III did not sufficiently influence the total index. 112 TABLE 12.--Revised index of attraction with campsite ratio Orig- Number inal of Camp- Capacity Revised Park No. Index sites Ratio Index Algonac 01 43.2 218 1.20 51.8 Aloha 02 103.2 306 1.69 174.4 Bald Mountain 03 50.8 40 .22 11.2 Baraga 04 87.5 97 .54 47.2 Bay City 05 96.7 300, 1.66 160.5 Benzie 06 43.0 125 .69 29.7 Brighton 07 89.0 219 1.21 107.7 Brimley 08 74.4 205 1.13 84.1 Burt Lake 09 111.3 232 1.28 142.5 Cheboygan 10 45.8 35 .19 8.7 D.H. Day 11 68.1 125 .69 47.0 East Tawas 12 97.6 185 1.02 99.6 Fayette 13 48.6 40 .22 10.7 Fort Wilkins '14 78.1 80 .44 34.4 Gladwin 15 47.2 65 .36 17.0 Gogebic 16 114.1 180 .99 113.0 Grand Haven 17 101.6 237 1.31 133.1 Harrisville 18 70.7 227 1.25 88.4 Hartwick Pines 19 71.6 45 .25 17.9 W.J. Hayes 20 104.7 202 1.12 117.3 Higgins Lake 21 116.3 582 3.22 374.5 Highland 22 79.1 40 .22 17.4 P.H. Hoeft 23 76.6 120 .66 50.6 Holland 24 115.2 376 2.08 239.6 Holly 25 92.6 195 1.08 100.0 Indian Lake 26 107.9 165 .91 98.2 Interlochen 27 108.8 487 2.69 279.2 Island Lake 28 92.0 103 .57 52.4 Lake Port 29 76.0 252 1.39 105.6 Ludington 30 163.4 299 1.65 269.6 McLain 31 49.4 94 .52 25.7 Mears 32 95.2 90 .50 47.6 Metamora- 33 90.7 145 .80 72.5 Hadley 113 Table 12.--Contd. Orig- Number inal of Camp- Capacity Revised Park NO. Index sites Ratio Index Mitchell 34 100.0 180 1.00 100.0 Muskallonge 35 81.7 150 .83 67.8 Lake Muskegon 36 140.9 214 1.18 166.3 Onaway 37 94.3 91 .50 47.2 Orchard Beach 38 67.6 180 1.00 67.6 Ortonville 39 80.0 40 .22 17.6 Otsego Lake 40 109.0 192 1.06 115.5 Pinckney 41 100.4 301 1.66 166.7 Pontiac Lake 42 89.4 30 .17 15.2 Porcupine 43 162.3 131 .72 116.9 MOuntains Port Crescent 44 74.3 180 1.00 74.3 Proud Lake 45 91.4 124 .69 63.1 Silver Lake 46 132.0 202 1.12 147.8 Sleeper 47 101.1 292 1.61 162.8 Straits 48 59.2 148 .82 48.5 Tahquamenon 49 104.4 226 1.25 130.5 Falls Traverse City 50 77.1 326 1.80 138.8 Van Riper 51 105.1 222 1.23 129.3 warren Dunes 52 90.4 137 .76 68.7 Waterloo 53 99.8 288 1.59 158.7 J.W. Wells 54 79.0 150 .83 65.6 White Cloud 55 47.4 45 .25 11.8 Wilderness 56 98.0 205 1.13 110.7 Wilson 57 101.1 150 .83 83.9 Yankee Springs58 109.8 252 1.39 152.6 Young 59 99.8 127 .70 69.9 TOTAL 10,694 114 A majority of parks with indices of less than 100 are reduced by the capacity weighting: Cheboygan and White Cloud, with indices in the 40's, are lowered to less than 10 index points. The revised index for each park is illustrated in Figure 7. The Attraction Index as a Measure of Attendance Now that a camping attraction index has been devel- Oped for each Michigan State Park, the utility of the index in measuring the mass attractive power of each park remains to be tested. The relationship between the attraction index and camper-days at a park can be examined by a regression analysis. The attraction index is treated as the independent or causal variable and camper-days as the dependent or resultant variable. The regression equation Y = 4814.38 + 418.07x eXpresses the relationship between the two variables, with a coeffi- cient of determination (r2) of +0.78.9 The “explained variation," which exceeded 75 per cent, is evidence of a relatively high degree of correspondance between the 9This figure was significant at the 5 per cent confidence level with TB = 14.33 and 57 degrees of freedom. 115 a. 201 AND ABOVE — zoo To 151 150 TO 101 100 TO 51 50 AND BELOW 42°- SOURCE: CARLTON 3. VANDOREN. MICHIGAN OUTDOOR RECREATION DEMAND STLDY. JUNEJMB A_ L— .— \cAMPINe \ WITH CAMPSITE RATIO MICHIGAN STATE PARKS O 20 4O 60 so MILES L A l 1 FIGURE 7 ATTRACTION INDEX :7! 1964 \ 87 ° 85“ 48 47° “0 42° 116 1964 camper-days at Michigan State Parks and the camping attraction index for each park.lo Since the attraction indices are to be used in an interaction model as a measure of camper-days, an analy— sis of the spatial pattern of the relative differences between the eXpected and actual values is useful. This analysis can be accomplished by mapping the residuals from the regression equation and thus pinpointing which parks have attraction indices that may not be reasonably representative of their true camping attraction. The utility of mapping residuals from regression for this purpose has been discussed by Thomas.11 The value of the residual indicates the magnitude of the difference between the computed and observed values relative to the standard error of the estimate.12 The residual (Y—YC)/SyC lS mapped on Figure 8. On the basis of the map of residuals, it is obvious 10It should be pointed out that a regression equa- tion between camper-days and the attraction indices without the adjustment for the number of campsites in each park yielded an r2 of +0.31. llEdwin N. Thomas, Maps of Residuals from Regres- sion: Their Characteristics and Uses in Geographic Research (Iowa City: Department of Geography, State University of Iowa, 1960). 12 Syc is equal to 16,073 camper—days. ... 48° ——1 47° _. 450 q '1 44° 83° 840 1964 117 HOUREB RESIDUALS FROM REGRESSION RELATIVE OEVIATION OF CAMPING ATTRACTION INDICES 87" MICHIGAN STATE PARKS WITH CAMP—DAYS \ 89 E m L N D S S E R G E R M O R F S N m T m V E D 830 840 I.” j' o o. o 00. 0 85° M? I E T A w T S E F O R O R R E D R A D N A T 6 5. 1 + N A H T R E T A E R 6 49° 45° — 44° - 86° 87"“ MILES 89 r‘ 5. 1 + O T O. 1 + C] -1.o TO +1.0 -15 T0 —1.0 - SMALLER THAN -1.5 42° *— 118 that the attraction indices of eight parks with a SYc of +1.0 or more can be expected to underpredict camper-days. Three of these parks in southwestern and southcentral Michigan--Warren Dunes, Yankee Springs, and Waterloo-- have standardized error estimates in excess of +1.5. Algonac, East Tawas, Traverse City, and Wilderness, located in the lower Peninsula, and Straits, located in the Upper Peninsula, have estimates from +1.0 SYc to +1.5 S The fact that these eight parks have under- Yc' predicting camper-days on the basis of their attraction indices can be explained by factors external to the parks. For example, Warren Dunes, Algonac, Wilderness, and Straits are located on major highways carrying a large volume of summer traffic. Many Of their campers may be overnight campers enroute to other areas. Warren Dunes is the closest Michigan park to the Chicago MetrOpolitan Region: Algonac is within a few miles of a major inter- national gateway between Michigan and Ontario, and Wilderness and Straits are located on both sides of the Straits of Mackinac. Traverse City State Park is located within the corporate limits of Traverse City and within a thriving resort region. East Tawas, like Traverse City, is a small park within the city limits. 119 Waterloo and Yankee Springs are both state recreation areas near large metropolitan areas and are also located in a part of the state where state parks are not numer- ous13 (in the case of Waterloo this statement is true only if the areas north and west are considered). Since the attraction indices include internal characteristics, it is probable that the underprediction for these eight parks is due to a number of external factors. An analysis of parks with negative deviations from the regression line is difficult. The nine parks with a deviation of -1.0 Syc or less have overpredicted camper-days. Porcupine MOuntains, Gogebic, Bay City, and Brighton have deviations of -1.5 Syc or less. It is evident that the camping attraction index as derived is much higher for some of these parks than actual attendance of campers--for example, Pinckney, Brighton, Bay City, and Aloha. Attraction indices in these cases may indicate camping parks that are as yet “undiscovered” 13The unit areas indicated on Figure 8 were derived by 1) connecting lines between each park and the nearest adjacent parks, 2) bisecting these connecting lines at 90°, and 3) connecting the bisecting lines to form a theoretical service area for each park. Unit areas of this type provide a base map for plotting residuals and also allow a visual analysis of the spacing of parks. The larger the unit area, the more scattered the parks. 120 by the camping pOpulace. Only in two cases were camper “turn aways" evident at these parks in 1964.14 For four of the Upper Peninsula parks the attraction indices relative to camper-days may truly indicate "undiscovered" horizons for camping. Porcupine Mountains and Muskallonge, fairly remote from major highways, have a wilderness character and their isolation is probably partially influential in the overprediction, particularly for Porcupine Mountains with a relatively high attraction index of 117. Gogebic and Van Riper are relatively near major highways, both are attractive parks with an index of 113 for Gogebic and 129 for Van Riper. There is no apparent reasonable explanation for the overprediction of Grand Haven. This park is one of the smallest (forty-three acres) but consists of an excellent sand beach on Lake Michigan. The attraction index, 133, reflects the qualities of the park. If the overpredic- tion of camper-days is considered an indication of true demand for camping at this park, that indication can be substantiated by the fact that camper "turn aways" at Grand Haven in 1964 totalled 16,120 camping parties. 14Brighton had 580 camping parties refused admit- tance and Aloha 6,470. This data is recorded annually by the Parks Section of the Michigan Department of Con- servation. 121 Summary Factor analysis has been used for grouping park characteristics into four eXplanatory factors reflecting the recreational resources available in the Michigan State Park System. Each of the four factors--inland lakes, scenic or land resources, camping amenities, and Great Lakes--inc1ude variables of natural resources and facilities and services, two of the three elements con- sidered paramount for a good recreation area. A camping attraction index for each Michigan State Park is develOped by combining weighted factor loadings of the variables into one overall factor score for each park. These are combined with apark's activity oppor- tunity score, the third element essential to a recrea- tion area. After standardization, this combination constitutes the camping attraction index and is assumed to be a measure of each park's camping attraction as well as a measure of the value of the camping experience. The appropriateness of the index as a substitute for camping attendance is suggested by a regression analysis. By mapping the residuals from the regression, it is evident that the attraction index for a few parks is not coincident with camping attendance. 122 The attraction index in itself does not include a measure of the spatial situation of the parks. The index includes specific on-site characteristics, but does not include a value reflecting the relative location of parks with respect to other outdoor recreational and tourism attractions. Further analysis would be neces- sary to include a measure of this kind. The degree to which this shortcoming detracts from the index cannot be known until the index is utilized as a component within a travel model. Results at this point, however, show the attraction indices as sufficiently valid for use as a component in the interaction model. CHAPTER V EMPIRICAL VERIFICATION OF THE RECREATIONAL TRAVEL MODEL Although the components necessary for an interaction model have already been identified, detenmining the para- meters of these components is difficult primarily because of the absence of necessary data on recreational travel flows to Michigan State Parks. Using available data sources, the initial parameters of the interaction com- ponents are derived according to the probabilistic method discussed by Isard.1 The following portion of this chapter is devoted to the empirical verification of the recreational travel model. Estimatingthe Parameters - Camper Population An estimate of Michigan's annual camping population can be obtained from the National Recreation Surve ,2 1Walter Isard, Methods of Regional Analysis, An Introduction to Regional Science (New York: John Wiley 5 Sons, Inc., 1960), p. 494. 2ORRRC Report 19, op. cit., p. 126. 123 124 which gives the percentage of persons twelve years or over who participate in camping in the North Central Region. These participation rates are listed by various socio- economic characteristics and by place of residence, both within and outside of Standard MetrOpolitan Statistical Areas (S.M.S.A.).3 The percentages by place of residence are applied to the 1960 Census of Population4 for each Michigan county to obtain an estimate of the annual camping pOpulation within Michigan. Using this method, P, the annual camping population within Michigan, is equal to 523,815. County camping estimates for Michigan are listed in Appendix D, Table 19. Given an estimate of the annual camping pOpulation in Michigan, the number of trips, T, generated by the camping population is estimated by dividing P by the average number of persons per camping party in the park system in 1962.5 3The percentage figures for the North Central Region are as follows: Residence in an S.M.S.A.--Urban (over 1,000,000-Wayne County only), 6%, Urban (under 1,000,000), 5%, Rural, 14%: Residence in Non-S.M.S.A.--Urban, 12%, Rural Farm, 4%, Rural Non-farm, 3%. ORRRC Report 19, ibid., p. 126. 4U.S.,Bureau of the Census, Eighteenth Census of the United States: 1960. Po ulation, I (Part XXIV, Michigan), p. 14. 5Michigan, Department of Conservation, Parks and Recreation Division, Summary Camping Information (Lansing, Mich., 1962), p. 3. (Mimeographed.) 125 T a 23%f§%§.= 123,250,6 where T is the number of trips. Since no information is currently available on the total number of annual camping trips in Michigan, this method for estimating T appears to be the most satisfactory alternative. Conceivably, the number of camping permits in the Michigan State Park system could have been used as an estimate of T, but extensions of camping permdts, averaging 33 per cent annually, would have resulted in an overestimation of actual camping trips in the state. No information was available at the time T was estimated on camping trips to state or national forests within Michigan. Following the probabilistic method of Isard, the next step in developing the interaction model is to establish the friction of distance at zero. This proced- ure requires an estimate of the number of camping trips originating in one county of Michigan and terminating in a state park. Hence it is necessary to utilize the 6The total number of camping permits issued in Michigan State Parks in 1962 was 188,276. Sixty-seven per cent of these, or 126,955, were to Michigan resi- dents. This would indicate that the difference between estimated camping trips by Michigan residents is fairly close to the actual in 1962, particularly if it is assumed that 33 per cent of the total permits issued each year are for extended stays (extensions) in the same park. 126 attraction indices for state parks. For example, suppose an individual in a Michigan county, Pi: were to go camp- ing. The probability that he would go to Wilson State Park, represented here by its attraction index (101.1) and designated Aj, is equal to the ratio Aj/P, the empirically derived attraction index for Wilson State Park divided by the total camping pOpulation of the state. Thus, the probability of a camper traveling to Wilson State Park is Aj/P = $2Li£2_{= .019 or 1.9 per 523,815 cent. (The attraction index has been multiplied by 100 for convenience in the example.) Since any camper in the state, under the hypothe- tical assumptions here, is identical with any other camper in Michigan and the friction of distance is zero, it appears permissible to estimate the number of camping trips any individual makes as the average number of trips per capita for Michigan. Such an average is T/P, or 23 250 £-‘--’- .24, and is designated k. By using k, one can 523,815 now estimate the absolute number of camping trips a camper in Michigan can be expected to make to Wilson State Park, k(Aj/P) = .24 (19111.9... = .24 (.019) = .00456 (Prfifibfld‘j> 523,815 or .452 This indicates that if 1.9 per cent of all camping trips in Michigan are made to Wilson State Park 127 and the average number of camping trips per capita are .24, then each camper in Michigan could be expected to make less than one visit to Wilson State Park. This applies, however, to any camper in Michigan. It does not tell how campers in specific Michigan counties might be inclined to travel to Wilson State Park. To estimate the number of campers from an individual county--for instance Wayne County--to Wilson State Park, the estimated number of campers in Wayne County, designated Pw, must be known. Pw here is equal to 165,337 (Table 19, Appendix D). Now, A: P T- = k 1 W a .24 (10,110) (165:337) = 7 6 3" p 523,815 6 ' (1) where Tjw identifies the total number of camping trips by Wayne County campers to Wilson State Park.7 Such hypothetical camping trip volumes can similarly be de- rived for all counties and parks in Michigan. Before 7This figure is very close to the actual number of camping permits issued to campers originating in Wayne County. In 1962 there were 694 permits issued in Wilson State Park to camping parties from Wayne County and in 1964 there were 664 permits. Both figures include extensions. The 1962 figure was obtained from Michigan State Highway Department, Origin Survey of Campers at Michigan State Parks, 1962 (Lansing, Mich., 1963), p. 28. (Mimeographed.) The 1964 figure was obtained from the Michigan Outdoor Recreation Demand Study. 128 this is done on a sample basis, a new variable, the dis- tance function, is added in the model. EstimatinggG - Constant and Distance Exponent Until now, the friction of distance has been as- sumed to be zero. In reality this is not so. Estimated time-distances now make it possible to measure the effect of time-distance between origin counties and parks in terms of the number of hypothetical trips originating and terminating. Since actual data on the number of trips between counties and parks are not available, the number of camping permits is used to estimate traffic volume. A recent survey includes the number of permits issued to campers from Michigan coun- ties in the twenty-five parks having the highest number of camping permits.8 A sample of actual trips by Michigan campers is drawn from this survey. By utiliz- ing the camping permit figures for counties listed at ‘Origin Survey of Campggg, op. cit., p. 4. At least three counties were listed for each of the twenty- five parks. The following counties were among those listed: Wayne, Oakland, Genesee, Kent, Ingham, Macomb, Saginaw, Bay, Muskegon, Midland, Kalamazoo, Montcalm, and Ottawa. 129 each of the twenty-five parks, 107 origin and destination links for Michigan are obtained.9 The number of camp permits between an originating county and a state park, used here as actual trip volume, are designated Ii Actual trips are divided by the j' expected or theoretical trip volume (Tij) to obtain a ratio of actual to theoretical trips, Iij/Tij' To com- pute Tij for each of the 107 sample links, equation (1) was programmed.10 There are 766 theoretical trips, Tij' to Wilson State Park from‘Wayne County, while the 1962 figure was 694. The ratio of actual to theoretical trips for this link is Iij/Tij or 694/766 = .9061. Following Isard's method, the ratio of expected trips to theoretical trips is plotted on a graph with a logarithmic11 scale on the Y axis (dependent variable), 9There are several limitations to the use of this data as a sample, but it was the only data available. The number of camping trips are preferable to camping permits. In addition, none of the twenty-five parks in the survey were state recreation areas located near metropolitan areas in the southern one-half of the state. Since state recreation areas where camping is allowed are used in this model, this deletion in the state survey is unfortunate. loPARKATTR was written for the author by Charles Hart of Michigan State University. 11A logarithmic transformation is used to reduce the problem of dealing with extreme cases and to improve linearity. 130 and the log of the time-distance variable tdij on the X axis (independent variable). A least squares regression analysis of these variables is calculated.12 The equa- tion for the line is I. log -%; = a-b log tdij. (2) T1] In this equation g’is a constant which is the intercept of the straight line with the Y axis. The constant 2. defines the slape of the line. If the logs are removed from equation (2) and g.is used for the antilog of g, I. c the intercept is .11. = b or Tij tdij CT- Iij = __£%; . (3) tdij By replacing the value of Tij in equation (3) with the value of Tiw in equation (1) and setting the constant G equation to ck/P, a basic gravity or interaction model is obtained. 12The DAP-2 program was used. This program was written by James Clark of the Department of Psychology, Michigan State University. DAP-2 forms scatter diagrams of pairs of variables plotted on standard score axes in 1/4 standard deviation categories. In addition, it prints out the mean, standard deviation, and variance of each variable, covariance, product moment correlation, regression coefficients, slope By,x and intercept, Ay,x. In the equation 2'= 5.67 and 2.= -.97. 131 Iij=G 551-. (4) tdijb Using the regression coefficients and the previously estimated values for k and P, the following values are obtained for the model: 2, the antilog of'g in equa- tion (2), equals 4.9, k = .24, and P = 523,815; therefore G = ck/P = 4.9 x .24/523,815 = .000002245. The slope b.is equal to the exponent for td, the time- distance between parks and counties. In this case b is equal to -.97 or -l.0. Therefore, the following formula is used to predict the actual trip volumes between seventy-one origin counties and fifty-nine Michigan State Parks where camping is allowed: I-- = (.000002245) A_1__'Pi . (5) 13 1 o tdij ' However, since it is necessary to predict the movement of campers to a park from all seventy-one counties, the model will have to predict movements from county P1, plus county P2, to . . . county P71 or 113 + 12j + . . . I71j. Therefore, equation (4) becomes 132 71 ( 71 Pi]. 0 (5) 2 I . = (.000002245)(A°) E . ' 0 71 td When the time-distance variable was estimated, seventeen out-of-state origins were also measured. Obviously, not all campers in Michigan State Parks are from Michigan; in 1962, 67 per cent of the campers were. The percentages of campers in Michigan parks from the surrounding states13 are as follows: Ohio 7.6% Illinois 6.2 Indiana 5.4 Wisconsin 3.1 Minnesota 1.8. Selection of Standard Metropolitan Statistical Areas (S.M.S.A.) as origins in the above states is based on the premise that most of the campers will come from the largest population centers in these states14 and from those S.M.S.A.'s that are closest to Michigan. A majority of these origins are within 150 miles of the Michigan boundary. 13Summary of Camping Information, op. cit., p. 3. 14This premise is true for Michigan campers. Thirty- four per cent of the campers at the twenty-five parks used in the Origin Survey of Campers had come from the three largest neErOpolitEn counTfiEE-of Wayne, Oakland, and Macomb. These counties contained 48 per cent of Michigan's pOpulation in 1960. Origin Survey, op. cit., p. iii. 133 The 1962 survey15 did not contain the number of camp permits by individual counties or S.M.S.A.'s within the surrounding five states, but listed total permits from each state. To estimate G and the exponent by the same probability method used for Michigan origins and including out-of-state origins, a weighted time-distance figure was devised for each of the five states. The ‘ weights are the estimated camper pOpulation within each S.M.S.A. Using these weights for each of the sample parks, the number of sample origins for estimating G and the exponent are increased from 107 to 186. The DAP-2 Program yields a regression equation with a_= 14.8 and b_= -l.4. Using this data, equations (2), (3), and (4) are computed with G = ck/P = 14.8 x .24/1,366,933 = .000002598 and the exponent equal to -l.40. Hence the interaction model including seventeen out-of-state S.M.S.A.'s becomes 88 88 14 Z 1883. = (.000002598)(Aj) Z tdij ° . (7) isl i=1 134 Verification of Time-Distance Exponent One major problem in the application of an inter- action model to a specific problem is the selection of the empirically derived exponential function for distance. Two exponents, -1.00 and -1.40, have been derived at this point: the latter one, including the seventeen out-of-state origins, should be used.16 How do these exponents compare with empirical studies of other trip types?17 The eXponent for travel-time 16In the Crevo model, mentioned in Chapter II, theoretical trips were predicted in a manner similar to equation (1); however, an attraction index was not used. The ratio of theoretical trips to actual trips was obtained and plotted against travel time in minutes, with both variables converted to logarithms. The results of this analysis produced exponents of -0.81 and -l.3, an indication that the estimates computed for this model are probably valid. Crevo, op. cit., p. 51. In the Connecticut study, however, the longest trip was ninety-five minutes, while in this study some trips are more than fourteen hours. A major shortcoming of the gravity or interaction model is that it frequently loses its validity over long or unlimited distances. It is successfully used for predicting intracity move- ment, but it has only been reasonably successful for intercity travel predictions. John T. Lynch, Glenn E. Brokke, Alan M. Voorhees, and Morton Schneider, ”Panel Discussion on Inter—Area Travel Formulas,“ Traffic Origin-and-Destination Studies, Bulletin 253 Elashington: Highway Research Board, 1960), p. 128. 17For intracity trips, where travel time is frequent- ly used as a function of distance, an exponent of 1.5 for total trips is common. Exponents for total travel between cities of 2.5 are numerous; unfortunately, the parameters for intracity and intercity travel are dif- ferent. The latter is usually based on airline distance. U.S., Department of Commerce, 0 . cit., pp. ii-3. 135 varies by trip purpose--work trips have a lower eXponent than social-recreation trips.18 Examples of exponents for specific trip types in urban areas are school trips, 2.0+; shOpping trips, 2.0: social trips, 1.1: and work trips, 0.9.19 This suggests that distance becomes a less restrictive factor as trips become more important. A camping trip, unlike a work trip, is infrequent, but the very infrequency of such a trip increases its importance. Intuitively it seems that camping parties are willing to travel long distances or periods of time to reach an attractive park for camping. Therefore, the relatively low eXponents that have been derived appear valid for application in the model. Results and Analysis of Model Solutions The parameters used in the first interaction model20 are those discussed in the preceding chapters and listed below: 1. The attraction index without the adjustment for capacity, i.e., the campsite ratio. lalbido' p. 11-7. 19Walter G. Hansen, 'How Accessibility Shapes Land Use," Journal of Ehe American Institute of Planners, Vol. xxv (May, 1959), p. 74. 20The program INTERACT 1, an interaction model, was ‘written for the author by Charles Hart, Michigan State University. ER: ,4 136 2. The estimated number of camper-days21 from each origin area. 3. The time-distance links between the fifty-nine parks and eighty-eight origin nodes. 4. The exponent of -l.40 derived by the least squares method. All of these parameters, with the exception of time- distance, are varied in successive applications of the model in order to achieve the best simulation of 1964 camper-day attendance at each park. A second goal is to match the 1964 camper-day figures from each of the eighty-eight origins. The successful simulation of these figures, however, does not insure that the model correctly predicts flows from each of the origins to one particular park, or from one origin to each of the fifty-nine parks. A separate statistical analysis of one county is done for a number of solutions to measure accuracy of the model in this reSpect. The basic measure used to compare various model 21A camper-day is one person camping one night. An estimate of camper-days at each origin was computed con- currently with estimates of the number of campers. This estimate was also derived from information in the National Recreation Survey. For the North Central Region, the number of camping activity days per participant is listed as 3.7 for campers in Non-S.M.S.A. counties and 6.1 for campers in an S.M.S.A. These figures were multiplied by the estimated number of campers in each origin for an approximation of camper-days (see Appendix D, Table 19). 137 solutions is the root-mean-square error22 (r.m.s.) of the predicted total camper-days at parks and at origins. The root-mean-square error for parks is 59 Parks r.m.s. error =,£_, 2: (per cent of error 59 3:1 of park j)2 and for origins 1 88 ( er cent error 88 p Origins r.m.s. error = i=1 of origin i)2 . The per cent error is defined as follows: predicted actual 1964 camper-days _ camper-days Per cent error of at park (or at park (or park (or origin) 3 origin) origin) x 100. camper-days actual 1964 camper-days at park (or origin) The 1964 origins and destinations of Michigan State Park campers are available from the Michigan Outdoor Recrea- tion Demand Study. For parks, two additional measures are arbitrarily used to quickly judge poor results. 1. The number of parks with a camper-day percent- age error equal or less than 20. 2. The number of parks with a camper-day percent- age error equal or greater than 50. The ultimate success of this model is dependent upon its ability to simulate the system. Thegoal, 22The root-mean-square error is defined as the square root of the averaged sum of the squared percent- age difference between predicted and actual camper-days. 138 therefore, is to achieve the lowest r.m.s. predictive measure feasible. In a behavioral system such as this, an r.m.s. measure of plus or minus 20 per cent for parks and origins is chosen as an acceptable range. Analysis of Basic Model One Description G = .000002598 Attraction Index - Without Campsite Ratio Exponent = 1.4 Origin POpulations - Estimated Camper-days The basic model did not produce a successful simu- lation of actual attendances. Test measures on this run resulted in an r.m.s. error of 403.5 for parks and an r.m.s. for origins of 318.9 (Table 13). Fifty-four per cent of the fifty-nine parks had over or underpre- dicted camper-days by more than 50 per cent, while twelve parks were equal to or less than 20 per cent of the 1964 attendance. A detailed examination of results of this model revealed areas for improvement. Since the model had obviously under or overpredicted camper-days for some parks, consistency, in terms of actual 1964 attendance, was checked. Did parks with low attendance have low 139 predictions, or was the reverse true? Finally, the at- traction indices were studied to discover if the model had predicted according to the magnitude of the indices. TABLE 13.--Error measures for select interact models Number of R.M.S. Error Parks Model Exponent Number Parks Origins 520% 250% Used 1 403.5 318.9 12 32 1.4 2 59.7 456.1 16 18 1.4 3 77.6 110.1 14 24 1.4 4 56.4 125.8 21 15 1.0 5 75.0 51.6 12 25 1.4 6 55.4 43.0 19 16 1.0 7 42.4 37.2 26 12 .4 All but three of the twelve parks having a prediction of 20 per cent or less of attendance, had very low 1964 actual attendance. Three of the parks, Grand Haven, waterloo, and Yankee Springs, had very high 1964 attend- ance (Figure 2). An inspection of the attraction indices of the twelve parks indicated that none of the parks had 140 either very high or very low indices, with a range from 43 for Algonac to 109 for Yankee Springs. All twelve parks were scattered over the state. Twenty-two parks overpredicted by more than 20 per cent and eleven of these are state recreation areas located in the Detroit MetrOpolitan Region. The attrac- tion indices of these twenty-two parks ranged from 46 for Cheboygan to 140 for Muskegon. Camper-day attendance at twenty-five parks that were underpredicted included eight parks in the Upper Peninsula. The seventeen lower peninsula parks are concentrated in the northern one-half of the peninsula, with the exception of East Tawas and Lakeport, while a majority of the parks are located on Lake Michigan--from the Straits of Mackinac south to Silver Lake State Park in Oceana County. Several parks are located within fifty miles of the Great Lakes. Two parks, Higgins Lake and Otsego, are further inland and located adjacent to Interstate 75. Ten of the twenty-five parks had relatively high camper-day totals in 1964, which may have influenced their being underpredicted. But seven of these ten parks have attraction indices of 95 or higher, and two, Ludington and Silver Lake, have very high indices of 163 and 132 respectively. 141 Several generalizations appear from this analysis. Parks that are overpredicted are those located very close to centers of pOpulation, and frequently are clustered, as for example the state recreation areas in southeastern Michigan. There appears to be a con- sistent overprediction of camper-days for the first and sometimes the second park on the route campers travel over out of a metrOpolitan area. For instance, attend- ance at Muskegon State Park is overpredicted, while that of the next park north from Muskegon, Silver Lake, is underpredicted. Another example of an overpredicted park attendance is Warren Dunes, the first Michigan park to be reached by residents of Gary, Hammond, South Bend, and Chicago. That the model is overpredicting to the first or second park and not beyond is substantiated by overpredicted camper-days in such relatively unattractive parks as Gladwin and White Cloud by more than 190 per cent. In the case of Gladwin, the nearest population centers are Midland and Bay City; campers at Bay City State Park, four miles north of Bay City, also are over- predicted. The next park north of Gladwin, Wilson, is within the 20 per cent acceptable range, while Higgins Lake, 30 miles north of Wilson, with a very high attrac- tion index, is underpredicted by 76 per cent. Since 142 most of the parks that are underpredicted are in the northern one-half of the state, the first conclusion is that the attraction indices as used in the model are not adequately weighted. A second conclusion is that the exponential value does not adequately reflect long distance accessibility. Revision of Attraction Index A re-evaluation of the attraction index resulted in the decision to include a camping-capacity weight as a part of the index. This revision was discussed in Chapter IV. Each park's attraction index is multiplied by the ratio of the number of campsites in the park to the average number of campsites (181) in all fifty-nine parks. For example, Bald Mountain State Recreation Area has an index of 50.8. In Model One camper-days for this park are overpredicted by more than 999.9 per cent. By multiplying the original index for Bald Mountain by the campsite ratio (40/181 = .22), the index is lowered to 11.2. Before the attraction index with the campsite ratio is utilized, two other modifications are necessary. Addition of Terminal Time Since the parks closest to the centers of the origin pOpulations are overpredicting in Model One, the distance 143 component is also seemingly in error. Previous research with gravity models has shown that for intercity travel, where trips are long, the omission of a terminal time was not important. For intracity travel, however, where travel times are short, neglecting this factor had a con- siderable affect.23 Due to the range in travel times within the model, from less than one hour to fourteen hours, it is concluded that the addition of a terminal time of one hour to each time-distance link might decrease the number of short trips but have little affect on trips of long duration. In particular, such a refinement may reduce short trips from metropolitan centers to nearby parks such as state recreation areas.24 The addition of a terminal time in a model such as this appears justified no matter how long the trip, a camping trip of any length requires a terminal time to set up camp at the site. 23U.S., Dept. of Commerce, Calibrating and Testing a Gravity Model, op. cit., p. II-4. 24The model was solved with the addition of terminal time but without the attraction index with the campsite ratio. This run produced an r.m.s. error for parks of 143.0 and for origins of 232.3, a definite improvement from Model One. Predictions at individual parks were substantially improved. Bald Mountain, which has over- predicted on Model One by more than 999.9 per cent, was overpredicted by 360 per cent. Brighton drOpped from +257 per cent on Model One to +44 per cent. Both state recreation areas are located near the Detroit Metropolitan Area. 144 Adjustment Ratio Total annual camper-day predictions on Model One were 3,256,226, whereas actual camper-days in 1964 were 3,136,977. The difference between these two totals is negligible.25 But to assure more accurate predictions of total camper-days for Michigan and for each park, a new ratio is incorporated within the model. Camper-day predictions are made and are adjusted up or down to bring them in line with the actual 1964 totals according to the ratio Actual 1964 Campgr-Days (3,136,977) Total Predicted Camper-Days Analysis of Medel Two Description G = .0002598 Attraction Index - With Campsite Ratio Exponent = 1.40 Origin Populations - Estimated Camper-days One Hour Terminal Time Adjustment Ratio Used - Total = 3,136,977 Medel Two introduces three modifications: the campsite ratio as a part of the attraction index, the 25The difference between the actual 1964 total and the predicted total on Model One is negligible, but when the model was run with the one hour terminal time, the difference was substantial, 3,136,977 to 1,709,334. 145 one hour terminal time to each time-distance link, and the total camper-day adjustment ratio. The results, after including these modifications, were gratifying. Now the r.m.s. error for parks was 59.7; for origins it was 456.1. Camper-day attendance predictions at indi- vidual parks are substantially improved. Sixteen parks are now predicted equal to or less than 20 per cent of their actual 1964 attendance while eight parks are off by more than 50 per cent (Table 13). Improving the prediction of camper-days from origins began with examination of the origin output on Model Two, which indicated that predicted out-of-state origins were especially high.26 Origin predictions for Michigan counties were also inaccurate, but less so than out-of- state origin predictions. As an example, Model Two predicted 751,974 camper-days from.Chicago while the esti- mated camper-days from the National Recreation Survey27 26The r.m.s. error for origins was enough indication that the model was not predicting origins accurately; however, some other measures were also available. The program computed the mean and standard deviations of total predictions for parks and origins. The mean number of camper-days for the eighty-eight estimated origins was 26,587 while the predicted mean was 35,647. The standard deviation of estimated camper-days was 68,175 and the predicted standard deviation was 103,482. 270RRRC Report 19, op. cit., p. 126. 146 were 35,665. Actually there were only 135,269 camper-days from the entire State of Illinois in 1964.28 The neces- sity to scale down the estimates of camper-days for out-of-state origins is apparent. Revised estimates of annual camper-days for all out-of-state origins are calculated by allocating a prOportion of the actual 1962 percentage of camper-days for each state among selected origins for that state. For example, there were 2,723,766 camper-days in the entire Michigan State Park system in 1962, and 5.1 per cent of these were from Illinois, or 138,912 camper-days. Originally the total number of camper-days estimated for Chicago was 2,428,605 (Table 19, Appendix D). Assuming that only 5.1 per cent of these camper-days take place in Michigan, then an estimated 123,615 camper-days from Chicago (.0509 x 2,428,605) could be eXpected in Michigan. Camper-days are estimated in a similar manner for the three remaining Illinois origins. This reappor- tionment results in a total of 138,639 camper-days from Illinois. Camper-days from S.M.S.A. origins in Indiana, 28The Illinois total of camper-days was obtained from the Parks Section of the Michigan Department of Conservation. State totals of this type are tabulated annually but not by specific origin within a state. 147 Wisconsin, Ohio, and Minnesota are estimated by the same method. These revised figures are listed in Table 19 of Appendix D. By using this method the total estimated camper-days from all in-state and out-of-state origins are reduced from the original estimate of 8,013,794 to 3,113,765. The latter is obviously closer to 1964 actual camper-days in the Michigan State Park system. The revised out-of- state origin estimates are used in Models Three and Four. Analysis of Models Three and Four - Revised Out-of-State Origin Estimates Description G = .000002598 Attraction Index - With Campsite Ratio One Hour Terminal Time Origin Populations - Out-of- State Revised Adjustment Ratio Used - Total 3,136,977 Model Three - Model Four - Exponent = 1.40 Exponent = 1.00 A revision of estimated camper-days for out-of-state origins substantially improved the accuracy in predicting camper-day origins. Whereas Model Two had an r.m.s. error of 447.8 for origins, Model Three resulted in an r.m.s. of 110.1 for origins. With the new origin data, the predicted r.m.s. value for parks on Model Three, 77.6, was worse than the 59.7 on Model Two. 148 A fourth model utilizing the lower of the two derived exponents was then attempted. The lower exponent was derived by using a sample of Michigan origins. Using the 1.00 exponent on Model Pour, the r.m.s. error for parks of 56.4 is the best achieved to this point. The origin r.m.s. value is 125.8, worse than on Mbdel Three (110.1), but still much better than that obtained on Models One and Two. In addition to the relatively low r.m.s. for parks, there were more parks, twenty-one, with predictions of 20 per cent or less than there had been on previous models. Only fifteen parks were off by more than 50 per cent. The improved results attained on Model Four by using a lower exponent cannot be fully explained. Perhaps time-distance is not prohibitive to the camper when he selects a park, and conceivably the park itself, if sufficiently attractive, can pull campers from long distances. It is apparent that more knowledge is needed about travel preferences and the behavior of campers. Although Model Four is reasonably successful and justifies the utility of an interaction model to predict attendance at each Michigan State Park, it is not 149 yielding a simulation of the system satisfactory for planning purposes. Both r.m.s. measures for parks and origins are too high, particularly that for origins. To improve the model, the known origin components are used as input. Estimates of camper-days at each of the eighty-eight origins have been used in the model to this point. There are several reasons for this. If the model is to be of practical value for planning purposes, it should require a minimum of actual data. For this reason estimates of camper-days by origins were calcu- lated with data from the National Recreation Survey?9 in an attempt not only to bypass the need for actual figures, but also to test the validity of utilizing component estimates based on this source. An analysis of predicted camper-days at each origin on Model Four indicates that Wayne County predictions of 1,013,382 were more than double (98 per cent) the actual 1964 attendance of 512,000. Chicago predictions of 84,876 were also more than double the estimated 1964 attendance of 35,665. These results suggest that the estimated camper-days at each origin are not realistic, assuming the attraction index was correct. In the 290RRRC Report 19, 02. cit. 150 remaining models, the actual 1964 camper-days from each of the eighty-eight origins are used in lieu of the estimated camper-days (see Appendix D, Table 19). Analysis of Models Five and Six Origins - 1964 Camper-Days Description G = .000002598 Attraction Index - With Camp- site Ratio One Hour Terminal Time Origin Populations - 1964 Data Adjustment Ratio Used - Total 3,136,977 Model Five - Model Six Exponent = 1.40 EXponent = 1.00 By using a known component in the model, the predic- tions of camper-days from origins are substantially improved. The r.m.s. error for origins on Model Five is 51.6. Predictions are also improved from specific origins--for example, total camper-days from Wayne County are overpredicted by 28 per cent while Chicago is underpredicted by 31 per cent. Improvement in the ability of the model to predict camper-day origins is not reflected in the capability of the model for predicting attendance at parks. On Model Five, the r.m.s. error for parks is 75.0 with only twelve parks within 20 per cent of the actual 1964 attendance, and twenty-five parks are off by more than 151 50 per cent (Table 13). While the origin predictions are approaching acceptability, the attendance predic- tions at parks are becoming worse. Medel Six is identical to Mbdel Five with the exception of the distance exponent. The solutions on Model Four, with an exponent of 1.00, prompted experimentation on Model Six with the same eXponent. On Model Six an r.m.s. error of 55.4 is the best achieved to this point. Nineteen parks are within 20 per cent of actual camper-days and sixteen parks are more than 50 per cent from the actual value. The r.m.s. error measure of origin predictions is 43.0, also the best obtained. Although Model Six produces attendance predictions with the lowest r.m.s. error at origins and destinations, analysis shows that the distribution of camper-days from an individual county to each of the fifty-nine parks is not close to reality. The model is inadequate for predictions of this detail. For example, when predictions for Wayne County camper-days at each of the fifty-nine parks are compared with the 1964 figures and the r.m.s. error is computed, the error for all fifty-nine parks is 489. The model overpredicted 152 camper-days from Wayne County for a majority of state recreation areas and for popular parks such as Grand Haven and Holland. The histogram in Figure 9 provides a visual analysis of the actual 1964 camper-days from wayne County by time-distance zones. Wayne County was selected above because it has the largest camper-day populace of any origin node utilized in the model. The 512,000 camper-days from this county amounted to one-sixth of all camper-days in Michigan State Parks in 1964. A regression analysis of 1964 camper-days from the county was computed using the BAP-2 Program. When the logarithms of camper-days at each park are placed on the Y axis and the log of the time-distance between Wayne County and each park on the X axis, the result is a regression line with a sloPe of -.40. Thus a new exponent is available for testing the model. Theoretically the new exponent is valid for use in the model. The two previously used exponents were derived from a sample of 1962 camper permits, whereas this ex- ponent is based on 1964 camper-days from the primary origin node in the state's park system. Foregoing runs of the model indicate that the lower of the two hhhhhhhhhhhh S K R m a m a: C N A Sm em . vH- .0; // gAmw mm 72 ED.MN mo / Rmem mm UESCWv TN mPTE m mm yflx/%////%///m m %////////////////////fl//a.W~ mummmmwmwmm1 1 m><0.mm¢5<0 l0 kWh—26301;. 154 exponents produce the best solutions. The new value provides an even lower exponential function for improv- ing the model. Analysis of Model Seven Adjustment Ratio - 3,136,977 Exponent = .40 Description G = .000002598 Attraction Index - With Campsite Ratio EXponent = .40 Origin Populations - 1964 Data One Hour Terminal Time Adjustment Ratio Used Total 3,136,977 By utilizing the empirically derived exponent based on Wayne County camper-day travel patterns, the predic- tive ability of the model is improved. The r.m.s. error measure for parks on this run is 42.4, the best obtained.30 Twenty-six out of the fifty-nine parks are predicting attendance within 20 per cent of the actual, and eleven are at the other extreme with camper-day predictions of 50 per cent or more. The modifications made in the model for this run are justified in terms of the increased accuracy attained in predicting origin camper-days. An r.m.s. error of 37.2 was obtained for origins. 3oA regression analysis of the relationship between projected camper-days and actual camper-days on Model Seven produced an r2 of 0.82. 155 In addition to standard tests of the model, predicted attendance at each park from Wayne County are examined. The r.m.s. error of these predictions is totally un- acceptable at 369.0. The model is now approaching acceptable solutions for prediction at each of the fifty- nine parks from eighty-eight origins, but for detailed predictions from individual counties to each park, it is inadequate. Summary of Mbdel Runs In successive runs of the model all components have been adjusted or revised with the exception of the time-distance links. Three exponents have been utilized with the one of lowest magnitude yielding the best solution. The most complete camper-day data at origins is included. A logical revision of the attraction indices, which included a camping capacity measure at each park, was inserted at a beginning stage. With these adjustment and revisions in the model, it is assumed that a great deal of 'noise' has been eliminated. In a behavioral system of the type simulated in this model, r.m.s. measures of 20 per cent or less were arbitrarily considered as a successful simulation of the system. Such a measure has not been attained for 156 park predictions here and it is doubtful that such an arbitrary limit can be achieved without completely re- vising the attraction indices. .Assuming that better predictions at parks could be obtained by revising the attraction indices, then the model can be of some use for planning purposes. The functional capability of the model, demonstrated by Model Seven, can be evaluated further by assessment of annual attendance fluctuations at individual parks. Between 1963 and 1964, overall camper-day attendance at fifty-nine Michigan State Parks increased by 8 per cent, but changes at specific parks vary. Some parks had camper-day increases of more than 100 per cent while others decreased by more than 40 per cent. A root-mean- square measure of the percentage change in attendance at the fifty-five parks between 1963 and 1964 is 33.4 per cent. From 1962 to 1963, the r.m.s. error is 28.7 per cent.31 It is apparent that from year-to-year camper- days at individual parks fluctuate considerably. This model has been developed to simulate attendances for one year. Recognizing the actual year to year fluctua- tions at parks, the 42.4 per cent r.m.s. on Model Seven, 31Ellis, Op. cit., p. 32. 157 although not as desirable as a lower figure, may not be excessive. Presuming that the attraction indices are at fault in the model, a supplementary analysis of some of the worst solutions (50 per cent or more) on Model Seven should assist in establishing which park indices are in error as well as the type of revision necessary. Pre- dicted attendances from Model Seven are listed on Table 14. Four of the most poorly predicted attendances--of the eleven parks with attendance projections of 50 per cent or more--are, in order of decreasing per cent error, Muskallonge (+162.5%), Bay City (+107.3%), Brighton (+98.9%), and Pontiac Lake (+74.l%). For all four parks the project- ed camper-days according to the model are more than would be expected. Three of these parks are the same that indicated a high negative standard error of estimate from the regression analysis in Chapter IV. These three are Muskallonge,32 Bay City, and Brighton. The fourth park, Pontiac Lake, has one of the lowest attraction indices, 15.2, and a low 1964 camper-day attendance of 6,018. 32It should be noted that attendance projections at Muskallonge State Park on all runs of the model were well above the actual 1964 attendance figures. TABLE l4.--Predicted camper-day from Mbdel Seven Compared 158 to 1964 attendance Predicted Attendance Park (Camper-days) Error % Error Algonac 1 31,962 6,538 -17.0 Aloha 2 86,622 31,773 57.9 Bald Mountain 3 7,611 1,183 18.4 Baraga 4 18,668 4,452 31.3 Bay City 5 97,297 50,366 107.3 Benzie 6 14,936 10,601 -4l.5 Brighton 7 74,608 14,392 98.9 Brimley 8 39,141 31,046 -44.2 Burt Lake 9 73,104 10,631 ~12.7 Cheboygan 10 4,234 388 -8.4 D.H. Day 11 23,031 5,904 -20.4 East Tawas 12 53,844 19,540 -26.6 Fayette 13 4,614 577 -ll.l Fort Wilkins 14 12,754 12,067 -48.6 Gladwin 15 9,589 2,283 31.2 Gogebic Lake 16 42,550 13,009 44.0 Grand Haven 17 78,672 1,009 1.3 Harrisville 18 45,195 14,098 -23.8 Hartwick Pines 19 10,472 11,194 -51.7 W.J. Hayes 20 74,687 5,137 -6.4 Higgins Lake 21 204,730 23,625 13.0 Highland 22 11,785 185 -1.5 P.H. Hoeft 23 23,757 2,564 12.1 Holland 24 144,450 12,162 9.2 Holly 25 67,437 26,531 64.9 Indian Lake 26 43,351 11,727 -21.3 Interlochen 27 145,776 12,294 9.2 Island Lake 28 36,736 13,838 60.4 Lakeport 29 64,631 16,294 -20.1 Ludington 30 141,533 15,494 12.3 F.J. McLain 31 9,824 12,649 -56.3 Charles Mears 32 25,770 3,067 -10.6 159 TABLE l4.--Contd. Predicted Attendance Park (Camper-days) Error % Error Metamora- 33 49,898 35,234 -4l.6 Headley Wm. Mitchell 34 55,125 7,278 15.2 Muskallonge Lake 35 29,525 18,278 162.5 Muskegon 36 95,587 16,392 20.7 Onaway 37 22,630 2,364 11.7 Orchard Beach 38 35,077 8,248 -l9.0 Ortonville 39 11,375 3,901 52.2 Otsego Lake 40 61,847 4,984 -7.5 Pinckney 41 113,325 33,237 41.5 Pontiac Lake 42 10,479 4,461 74.1 Porcupine Mts. 43 44,149 13,026 41.9 Port Crescent 44 42,588 3,377 8.6 Proud Lake 45 44,231 27,739 -38.5 Silver Lake 46 80,030 6,177 -7.2 Sleeper 47 95,630 29,119 43.8 Straits 48 23,748 27,235 -53.4 Tahquamenon 49 58,604 6,545 12.6 Traverse City 50 69,802 14,125 -16.8 Van Riper 51 52,729 12,294 30.4 Warren Dunes 52 39,715 25,766 -39.3 Waterloo 53 107,639 16,893 -13.6 J.W. Wells 54 30,499 120 0.4 White Cloud 55 6,658 1,326 24.9 Wilderness 56 53,499 20,530 -27.7 Wilson 57 48,310 6,617 -12.0 Yankee Springs 58 96,113 15,936 -14.2 Young 59 34,791 18,283 -34.4 ....... 160 Unlike the other three parks, its standard error of esti- mate on the regression analysis was between -.5 and -l.0 SYC. As a state recreation area nine miles west of the city of Pontiac, this relatively unattractive camping park is probably rated too high when considered in combination with the other model components. Four additional parks out of the eleven also indi- cate overprojections of camper-days in excess of 50 per cent. These are Aloha (+58%), Holly (+65%), Island Lake (+60%), and Ortonville (+52%). Of these four only Aloha had an unusually high standard error of estimate from the regression analysis. Holly, Island Lake, and Ortonville, like Pontiac Lake, are state recreation areas in southeastern Michigan. Ortonville, like Pontiac Lake, had less than 10,000 camper-days in 1964, while Holly had 41,000 and Island Lake had 23,000. The model is obviously inadequate for projecting camper-days at the nine state recreation areas closest to the metrOpolitan centers of southeast Michigan. In addition to overpredictions for Brighton, Pontiac Lake, Holly, Island Lake, and Ortonville, Metamora and Proud Lake had underprojected camper-days by -42 per cent and -38 per cent respectively. Only Bald Mountain (+18%) 161 and Highland (+1.5%) were projecting within the critical range stated as acceptable (both have attraction indices of 17 or less). It is believed that much of the difficulty in projecting attendance at all nine areas is due to inadequacies in the attraction indices. However, the closeness of these areas to the centers of population also suggests that the population and distance parameters may also be in error. Perhaps some camping behavioral element has been overlooked. It may be that these parks are too close for a majority of campers in southeast Michigan. It could be assumed a person in Detroit planning a camping trip would desire a park farther away than one of the state recreation areas, preferring a park that might offer more environmental contrast than that near his home. The remaining three of the eleven parks with projec- tions of 50 per cent or more are Hartwick Pines (-52%), McLain (-56%), and Straits (-53%). All three parks have underpredicted camper-days according to the model. Finding Straits State Park within this group is not unexpected, since it was one of the parks in the regres- sion analysis that was in error by more than +1.0 SYc' 162 In Chapter IV, eight parks with Syc of +1.0 or more were listed. Straits is the only one of the eight that retained an underprojection of more than -50 per cent when its attraction index was used in the model. Results from the model for the remaining eight parks are Warren Dunes (-39%), Yankee Springs (-14%), Waterloo (-l4%), Algonac (-17%), East Tawas (-27%), Traverse City (-l7%), and Wilderness (-28%). All the above attendances were underprojected, but it is apparent that the addition of other components in the model bring the attendance estimate closer to real- ity. There are attractions adjacent to Straits State Park not included in its index.33 For McLain State Park the underprojection of attendance can be partially attributed to the attraction index, since its Syc was between 0 to +.5 in the regression analysis. Hartwick Pines' underprediction is the reverse of what could be expected from the regression analysis where it indicated 33The importance of attractions external to Straits State Park and its strategic position at the Straits of Mackinac are emphasized even more by the large number (34,440) of camping parties turned away from the park in 1964. According to the Department of Conservation, one other park in the system had more turnaways. This park was Warren Dunes with a total of 45,000 parties. Warren Dunes was underpredicted by -39 per cent. Like that of Straits State Park, the location of Warren Dunes is probably the primary factor disrupting the model's projections. 163 an overprojection of camper-days. With origin pOpulation and distance components added to the model, its projection would suggest an attraction index higher than that com- puted. Located in Crawford County adjacent to Interstate 75, it is easily accessible to the population centers in southern Michigan. Hartwick Pines is also the first park north of Higgins Lake State Park, which in 1964 had a very high number of camper turnaways.34 Campers may have driven that short distance to find available camping sites at Hartwick Pines, thus increasing camping at the less desirable park. Refinements of the Index In the case of Bay City State Park, near large metrOpolitan areas around Saginaw Bay, and the state recreation areas, overpredictions of attendance indicate the need for addition of one or more behavioral elements in the attraction indices. The missing elements, essential for prOper evaluation of a park's pulling power, can only be determined by further study-~prefer- ably by field interviews to determine directly from 34Turnaways at Higgins Lake for the 1964 season, according to the records of the Department of Conser- vation, totaled 17,430 camping parties. 164 participants how they evaluate a particular park for camping and how far they prefer to travel. Such an approach also will define the type of persons using the parks. Differing socio-economic characteristics of campers in the state recreation areas adjacent to the Detroit Metropolitan Area and those of campers to northern Michigan parks may account for inaccurate predictions attained for these parks in the travel model. By personal interviews it might also be found well-established parks act as a deterrent to camping in relatively new parks located nearby. A major deficiency in the attraction index is the lack of attention to features external to each park. In addition to park site, characteristics necessary in the index must include the situation of a park, that is, nearness to other parks or recreation attractions' used by campers. The deletion of this factor is an apparent failing for several Upper Michigan parks-- Straits, MCLain, Fort Wilkins, Brimley, Gogebic, and Porcupine Mountains--and may be a factor for other parks, particularly the state recreation areas. In addition, a more detailed analysis of origin-destination data for many of the parks might prove useful. From 165 such data it might be determined that campers in south- eastern Michigan do prefer to drive beyond the nearer state recreation areas to camp. Although a revision of the attraction indices is not done for this study, solutions to the model strongly suggest that such a revision would substantially improve the model. Revisions should not be undertaken without personal interviewing to obtain the users' perceptions of the qualities of the parks. CHAPTER VI CONCLUSION Summary This study was initiated on the premise that a need exists for classifying and analyzing the natural resource base for outdoor recreational purposes. It was empha- sized that an investigation of this type would be particularly useful in assessing feasibility of prOposed recreational sites as well as in obtaining insight into the present character and value of existing sites. The first step in the investigation was to review methods that have been utilized to measure the satisfac- tions derived by participants in outdoor recreational activities, recognizing that previous models (1) did not take cognizance of the resources at the site, and (2) did not specifically relate to the interaction factors that spatially connect a recreational system. The second step in the investigation was construction of an index to replicate the attractive qualities of a 166 167 state park for a particular outdoor activity--camping. This index included 1) the activity preferences of campers, 2) an assessment of the natural resources, 3) the man-made services and facilities, and 4) the capacity of each park for camping. Camping attraction as measured by the index was then compared with the attendance at parks by a regression analysis. Finding that there was a statistically significant relationship between the index values and attendance, the attraction indices were used in an interaction travel model for the purpose of replicating the movement of campers to fifty-nine Michigan State Parks in 1964. The results of the travel model provide evidence that an empirical value of the attractive qualities of parks for one outdoor activity was successfully exhibited.1 This success lends support to the assumptions made and methods used in constructing the index. Variations in camping attendance are dependent upon and can be partially explained by site characteristics. 1For additional verification of the merits of the attraction index and its use in a systems travel model, see Jack B. Ellis and Carlton S. Van Doren, "A Compara- tive Evaluation of Gravity and Systems Theory Models for Statewide Recreational Traffic Flow," Journal of Regional Science, Vol. VI (Winter, 1966), p. 57. 168 Implications The attraction indices used independently or as a component of a travel model have many merits. A park- by-park analysis of the recreational attributes of many points in space can be obtained by a comparison of the indices. Such an analysis has utility for obtaining a spatial overview to assist in recreation resource management studies. Sub-components of the index could be utilized in the same manner; individual attributes could be located and related Spatially. For example, parks with qualities attractive for inland lake boating could be readily identified. Future research may take this course. The primary utility of the attraction indices, however, is in travel models. Such use provides the Opportunity to analyze interconnections in the existing system and to provide solutions to problems encountered when considering the location of recreational phenomena. A travel model utilizing an attraction index as a sub- stitute for attendance allows new insights into problems of accessibility, attractiveness, competing recreational opportunities, and estimates of attendance saturation or overcrowdedness. In addition, by developing an 169 attraction index for a prOposed park based on the master plans of a state or national planning agency and the anticipated activity participation rates, the park's attendance could be simulated with an established park system. The flexibility of the attraction index and the interaction model means that necessary adjustments can be made for making demand projections of future attendance, using reliable demographic predictions and anticipated outdoor recreation participation rates. Not only can the use of new parks be anticipated but measuring benefits of planned refurbishing or enlarge- ment of existing parks is possible. Recalling the overprediction of camping attendance at Muskallonge Park, (Chapters III and IV) the model results may be anticipating a future increase in use as the camping public discovers the park. The model could be adapted for predictions of uses other than camping. It is conceivable that indices for predicting attendance for wilderness camping, or for other specific activities such as boating or fishing could be derived. If Factor III (camping amenities) is removed from the present index before the scoring 170 routine is attempted and parks are weighted empirically for day-use activity preferences, then an index for day-use visitors might be constructed. At present the problem is complicated by a lack of day-use origin and destination data to verify such an index. One of the merits of the attraction index and its use in an interaction model is that it is basically simple to construct and make operational. The simplicity of the model means that it can be duplicated for applied purposes by persons with limited analytical training. Problems in operating the model can be attributed pri- marily to difficulties in obtaining the necessary data inputs not only for prediction but for verification. Even at this stage in the develOpment of recreation travel models, when modest inputs are required, obtaining this information in the prOper amount and form is diffi- cult. Types of information difficult to obtain for this model are (l) origin and destination patterns of campers including the duration of visits, and (2) detailed characteristics of the physical environment and facil- ities within parks. In Michigan, origin and destination data have been potentially available from camping permits for a long time. But it was not until a specific request 171 was made by the Michigan Outdoor Recreation Demand Study for this information2 that its utility in two types of travel models could be demonstrated.3 Many of the park characteristics for the attraction index were readily available, while others could be gathered from the planners and administrative staff of the Michigan Department of Conservation. Information concerning activity preferences of campers is more difficult to obtain; however, the Bureau of Outdoor Recreation has recently (September, 1965) undertaken a large nation- wide survey to amass such data. The distance component used in the model is time consuming to construct but once it is completed changes resulting from highway relocations or improvements can be easily made. Limitations of the Model and Future Investigations One of the limitations inherent in the model is its ability to predict only single purpose, one-way trips. 2A full determination of the origins and destina- tions of Michigan state park camping permits for 1964 was made under the supervision of Jack B. Ellis for the Michigan Outdoor Recreation Demand Study, Department of Resource DevelOpment, Michigan State University. 3Ellis and Van Doren, ibid. 172 Many campers undoubtedly camp at more than one park and multiple stOps for camping are not included in the model. No attempt was made in constructing the model to account for park overflows at any time during a camping season, or how a camping party may react when this occurs. To the author's knowledge, this constraint has not been applied to interaction models. It was not attempted in this investigation, since the major focus was that of testing the attraction indices. The entire question concerning intervening oppor- tunities in an interaction model has not been explored. The variable magnitude of the attraction indices may be minimizing the intervening effects to some extent, but this needs verification. In addition, analysis of why the model does not simulate the flows from individual counties to parks is necessary. Aggregative attendance figures are replicated but not flows from separate counties. Possibly by using several empirically derived distance eXponents for selected counties, this limitation could be corrected. The model has been verified on the basis of success- ful predictions for one year, but even with accurate results questions would still remain as to why the model 173 is apparently viable in this reSpect. Travel behavior implications associated with the model, particularly the attraction indices, should be investigated further. Campsite interviews would reveal why campers select one park over another. Such insights into the perceptions4 of campers will not only assist in improving models such as this one, but will eventually lead to a body of empirical data necessary for the development of hypo- theses, concepts, and theories on park location. The goal in this study has not been to develop theory but to demonstrate the applicability of a recreation resource classification system in a recreation travel model. Conclusion This investigation has centered on a method for evaluating recreational land use at selected points in space. The development of an index of camping attraction quality demonstrates a method of area differentiation, and as a classificatory scheme not previously attempted, this attraction index provides an empirical method for 4Several geographers have concerned themselves with the perceptions of the landscape. See Lucas, 0 . cit , p. 394, and Robert Kates,"The Pursuit of Beauty,“ Paper prepared for delivery to a Symposium of the Natural Resources Institute, The Ohio State University, Columbus, Ohio, May 24, 1966. 174 rating the quality of Michigan State Parks and comparing their spatial relationships to other recreational and non-recreational phenomena. The attraction index includes an analysis of a combination of natural-cultural resources and relates these features to the behavioral characteristics and desires of man in relation to camping. Such a complex resource appraisal for specific points in space is considered to be one of many major contributions of the geographer.5 The general theme of the investigation was expanded beyond the descriptive phase useful for area differen- tiation to include a much wider scope in terms of Space and content. An attempt was made to utilize the classi- ficatory scheme in a travel model, under the assumption that campers have the ability to rationalize their choice of a park. In short, the study is an attempt to encom- pass the major forces which form a system of space relationships. The model, then, includes structural components of two dissimilar landscapes as well as the "connectivity"6 or information flow between them. By assuming that the distance and population components 5James and Jones, 0p. cit., p. 230. 6Ackerman, op. cit., p. 437. 175 within the model were correct, then the addition Of these components and the resulting spatial connections illus— trated provided for an analysis of the validity Of the classification scheme, that is, the attraction indices. The classification of phenomena is the major task Of any science. A basic notion as to what constitutes the attractive qualities of a park for any activity is a necessary prerequisite for moving ahead to more diffi— cult problems concerning the spatial content of recrea- tional land use. The attraction index provides a basic classification and rating scheme and in itself provides clues for the spatial interconnections that are Operative at one point in time. Continued research along these lines is commonly recognized as a contribution to the discipline Of geography and also to applied problems in outdoor recreation research.7 Further develOpment of models such as this one, tested in a variety Of spatial situations, will assist in building a framework for planning and locating outdoor 7For a general discussion of what geographers have done and are qualified by their training to do for out- door recreation research see Richard E. Murphy, “Geo- graphy and Outdoor Recreation: An Opportunity and An Obligation," The Professional Geographer, Vol. XV (September, 1963). p. 33, and R. I. Wolfe, "Perspective on Outdoor Recreation — A Bibliographical Survey," Geographical Review, Vol. LIV (April, 1964), p. 203. 176 recreation sites. As the use of outdoor recreation sites increases and outdoor recreation demands grow, the adage that ”Parks are like gold, where you find them,"8 is no longer applicable. Meeting these recreational demands is highly dependent on the prOper location of recreation sites. It is clearly within the sphere of the geographer to find solutions to these locational problems. 8From a presentation by Newton B. Drury, California Chief of Beaches and Parks and former Director of the National Park Service, to the National Conference on State Parks, November, 1957. APPENDIX A 177 APPENDIX A STATE AND LOCAL STUDIES OF ACTIVITY PREFERENCE PATTERNS Attempts to associate other outdoor activities with camping are based on national surveys found in a number of the ORRRC Reports. This national and regional infor- f mation must be compared with somewhat different data which is available for Michigan. Two statewide surveys of visitors to Michigan State Parks conducted in the last decade included questions on visitor activities while in parks. The first Of these, the Michigan State Park Users 1 Survey, consisted of two parts--one done by 1,452 personal interviews, the other by distribution Of a questionnaire to be filled out voluntarily by park visi- tors. (Statistical sampling theory was not applied to the personal interviews.) Results Of the activity preferences trom.both are listed on Table 15. For the visitors' voluntary responses activity preferences are arranged in decreasing order. Neither the personal 1Dahle, Op. cit. 178 179 TABLE 15.--Per cent of activity participation preferences from personal interviews and voluntary responses Michigan State Parks - Summer, 1956 Personal VOluntary Activity Interviews (Total 2 1,452) Responses (Total = 4,147) Swimming 20.0% 21.0% Relaxation .2% 21.0% T Boating 5 .0: 18. o: 5‘ Camping 39.0% 15.0% ’1 Picnicking 17.0% 15.0% Fishing 14.0% 10.0% Hiking 2.0% 9.0% Nature Study ****a 4.0% Touring 3.0% 1.0% All or Most of Above 11.0% 2.0% aNot a category used in personal survey. Source: Thomas L. Dahle, Michi an State Pgrk Users Surve , 1956, Bureau O Business Research, CO ege O usiness and Public Service, Research Report Number 19 (East Lansing, Mich.: Michigan State University, 1956), p. 8. r33 ,5 180 interviews nor the voluntary response questionnaires identifies campers' activity preferences separately, but it is possible to generalize that water activities were very important in Michigan State Parks in 1955. The study indicates that swimming, picnicking, fishing, relaxing, and camping were popular activities, while boating, hiking and nature study were less pOpular. A second statewide study on state parks and public lands conducted in 1962, commonly known as the Van Til 2 Report, included a park users survey at thirty-five : state parks. Nineteen questions were used, the major- ity Of them pertinent to this discussion. Even more important, a preponderance of the 2,000 visitors sur- veyed were campers.3 Findings of this report are related to the present analysis; apparently, however, no measures were taken during the survey to allow for statistical testing of the results, so reliability of the data is Open to question and only_generalizations can be drawn from the survey. One survey question requested respondents to list their primary purpose in visiting one of the Michigan 2Repgrt 2; Committee pp_State Parks and Public Lands, Op. cit. 3Ibid., p. 8. The report does not say how many per- ,sons in EHe survey were campers. ”There were 2,000 questionnaires distributed--mostly to Park Campers in 35 State Parks.” 181 State Parks. They were asked to order their prefer- ences in terms of recreation, scenic, or historic purposes. Answers to the questions indicated that 86% Of the respondents considered recreation as their primary purpose, while 12% listed scenic and 1% listed historic. Recreation is difficult to define, but as . “Z’U‘CA 1.1 . the report states, the implication is that facilities for recreational activities may be more important than the natural environment of the park. This statement is essentially in agreement with the previous quotation from Perloff and Wingo:4 that is, that activities are the center of attraction and are directly dependent upon activity facilities available. The report also states that the majority of campers apparently did not consider camping as a recreational activity; the respondents believed camping was a means to prolong a visit and participate in recreational activities within the parks. The most conclusive result of this survey is the desire and need of the camper for adequate swimming facilities. Suitable water for swimming is necess- ary for the campers as indicated by the fact that larger prOportion of users do not stay at parks where swimming is unavailable than those who do not stay at parks without fishing or boating. (Sixty percent to 30% to 25 percent respectively).5 4Perloff and Wingo, Op. cit., p. 89. 5Report of Committee on State Parks and Public Lands, op. cit., p. 9. 182 A majority Of campers interviewed in this survey did not require boating as a part Of their recreational activities. When questioned about the most enjoyable activity while on vacation at a Michigan State Park, 51% preferred camping; 16% swimming: 11% sightseeing; 3% visiting; and 2% boating.6 Although campers were in .:.".T’.;--.E L. i asked to state their preferences for specific groups of recreational activities, the total responses were insufficient for proper analysis. It was obvious, res-a“ mhrb‘h’} h.“ 25' however, that swimming was preferred when it was included as one of the activity choices. In 1958, a survey similar to the Van Til Report was undertaken in twenty-seven state parks in Wisconsin.7 Activity preferences were included in a question on the recreational purpose of a visit to a park. Since day users as well as overnight visitors were interviewed, a comparison of the results with the Michigan camping data is hampered. Sightseeing was stated as the primary purpose by 34% of the visitors: picnicking, by 19%; camping, 17%; and swimming, 14%. Boating, nature study, and fishing attracted a small percentage Of the visitors to Wisconsin parks.8 6Ibid., p. 14. The respondents answers to this question seems to contradict the previous statement that campers did not consider camping as recreation. When asked the most enjoyable activity, they overwhelms ingly stated camping. 7Hutchins and Trecker, Op. cit. 8Ibid., p. 19. 183 During the summer of 1959, the Department Of Resource Development of Michigan State university con- ducted comprehensive interviews with vacation campers at seven selected campgrounds in Iron County, Michigan.9 None Of these campgrounds were in a state park, but the results Of the survey are still meaningful if the camp- ing population is assumed to be similar in origins and socio-economic characteristics to campers visiting Michigan State Parks.10 Imm‘qr’l' 'mermlm . 9 A!» itlaxnl The survey asked, ”What would you say were the one or two things you enjoyed most?" Forty per cent of the men and 20% of the women interviewed considered water-oriented activities (fishing, swimming, boating) the most desirable.11 For example, fishing was pre- ferred by 48% of the men: swimming, by 15%; relaxing, 11%; and hiking, 2%. Similar preferences were indicated by women campers, with the exception of a slight pref- erence for swimming over fishing. Hiking was more popular with women than with men by a margin of 7%.12 9Gilbert, Op. cit. 10This assumption has been made after consultation with L. M. Reid, director Of the survey. The county parks in Iron County are considered to be on a par with most state parks. The 1959 survey indicated that the county parks surveyed were attracting visitors from.long distances, comparable to the attraction areas of state parks. This county, incidentally, contains no state parks. ' llGilbert, Op. cit., p. 118. 121bid., p. 131. APPENDIX B 184 APPENDIX B ELEMENTS OF THE NATURAL ENVIRONMENT ATTRACTIVE FOR CAMPING Each natural resource variable is analyzed in the discussion that follows in terms Of its significance in outdoor recreational activities. The item is then classified and scaled and its usefulness in the final index is explained. The fact that some variables did not ultimately contribute much to the index does not mean that they might not be useful in another type of attraction index, or useful if scaled by other methods. Terrain The forms and features Of the landscape have a considerable affect on the type Of outdoor recrea- tional activities pursued in a given area. A level terrain is desirable for camping, assuming well-drained soils, while for picnicking an undulating terrain may be favored.1 A level area with a sharp rise in terrain is ideal for archery and rifle ranges. For hiking, the most important of the land activities with the exception Of camping, trails are Often more 10. 8., Forest Service, Op. cit., p. 80. 185 186 inUflguing if they traverse a rough or undulating terrain. A landscape of this type will frequently afford vistas. The same general statements could be made concerning bridle trails. The application Of scales to measure the terrain and to isolate unique landforms is formidable. Terrain I measurements are less vexing conceptually than the problem Of identifying classes of terrain types, in that lepes and local relief as differentiating char- acteristics can be measured. Landforms for outdoor recreation, although inherently a part Of and included in such measures as slope and local relief for terrain, are more difficult to rate numerically. In this analysis a simplified measure is adopted for landforms. Reid in ORRRC Report 5 measured terrain by sub- jective classifications such as flat, rolling, hilly, or mountainous,2 and the Bureau of Outdoor Recreation has a.similar rating on their inventory forms. A rating of this type, however, requires a visual inspec- tion of the site and some knowledge Of the local relief and amount of slope. If the same individual or survey team judges the terrain for every recreational site evaluated by this method, as was the case in Reid's study, then consistency Of judgment could be expected. The problem in constructing this type of scale is 2ORRRC Report 5, Op. cit., p. 13. 187 that no clear-cut numerical values can be assigned. A different team might classify many Of the same recre- athmxsites quite differently, depending upon their individual backgrounds and academic training. Since the Bureau Of Outdoor Recreation inventory forms had been completed for Michigan State Parks and these forms were made available to the author, the terrain characteristics-~hilly, rolling, and so forth --were c0pied on the inventory sheets for later use if These data were not used in the final necessary. analysis, however. Instead, more exact measures Of terrain in each park were adOpted. The local relief of each Michigan State Park was measured from 0.8. Geological Survey tOpographic quad- rangles; when these were not available, it was taken from Park Master Plans. For Michigan it is assumed that the greater the local relief, the more attractive the park, although this might not hold true for state parks in more mountainous states. .An additional measure of terrain by per cent of slope was prOposed in the Forest Service Work Plan.3 This survey recommended classifying and rating per cent Of lepe in the following manner: a. Abrupt--Slopes of 30% or more with a 3U. 8., Forest Service, Op. cit., p. 80. 188 value of 4. b. Irregular--Slopes Of 20-30% with a value of 3. c. Rolling--Slopes of 10-20% with a value of 2. d. Regular--Slopes of 0-10% with a value of 1. The Work Plan states that slopes Of 30% or more were not suitable for occupancy in terms of campground development or for most land activities. Directly measuring lepes in each of the Michigan State Parks was considered extremely time consuming and was not attempted, particularly after another source of lepe and terrain information was recognized. The additional source is the work of the geog- rapher Hammond, who has developed a map with "Classes of Land-Surface Form in the Forty-Eight States."4 Three measures of terrain are included: per cent of slope, designated by ampital letter A, B, C, or D: local relief, expressed by a number 1 to 6; and profile type, shown by a lower case letter. These three measures are listed below. 4E. H. Hammond, “Classes of Land-Surface Form in the Forty-Eight States,“ Annals of the Association of American Geographers, Map Supplement No. 4, L V (ihufifla, I964T. 189 SLOPE (lst letter) A. 80% of area gently sloping. B. 50-80% of area gently sloping. C. 20-50% Of area gently sloping. D. 20% of area gently leping. LOCAL RELIEF (2nd letter) 1. 0-100 feet. 2. 100-300 feet. 3. 300-500 feet. 4. 500-1000 feet- 5. 1000-3000 feet. 6. 3000-5000 feet. PROFILE TYPE (3rd letter) a. 75% of gentle slope is in lowland. b. 50-75% of gentle lepe is in lowland. c. 50-75% of gentle slope is on upland- d. 75% Of gentle lepe is on upland. Since this classification embodies many of the terrain qualities inherent in selecting and classifying the qualities Of recreation sites, Hammond's map was used to classify the terrain characteristics Of the Michigan State Parks. Parks were found to be located within six differ- ent classes, from A1 to B4b. These classes were recorded as A1 = l, A2b I 2, . . . . B4b = 6. The assumption is made that the more slope, the greater 190 the local relief, and the more land in upland, the more attractive the park is. As an example, Porcupine Mountains State Park is located within a B4b classifi- cation, and therefore received a numerical score Of 6; Bay City State Park, located in an Al area, received a score Of 1. This classification and scoring system 1| for terrain seems ideal and should not be overlooked in future studies. In the technique used for combining the numerical variables into an index, however, it was found to be inadequate as a measure when compared with Llocal relief. 'Consequently, it was not used in comput- ing the final index in this study. Measurement of various landform types is more difficult than measurement of terrain. As with terrain, unusual landforms add to the scenic qualities of an area. Michigan parks do have atypical landforms such as falls, high cliffs, overlooks, massive sand dunes, Features Of this type were recorded on A suitable way Of and springs . the inventory sheets for each park. measuring these features could not be devised. To include them within the index, they are treated as binary data. This avoids the problem Of estimating the intangible value derived from features of this type, but at the same time, acknowledges the presence or absence of the feature in a park. The parks receiving a score of 1 for having one or more of these features can be identified in Table 18, Appendix D. 191 Size of Park The acreage of a park is an additional variable directly associated with terrain. This variable, included as a component within the index, is excluded in the final model. It is evident in the discussion that follows that the assignment of numerical values to physical environmental variables inherently recog- nizes the size of the park. Some additional reasons for its exclusion are given in Chapter IV. Vegetation The biotic character of a recreation site, when considered with terrain, may add untold value to all recreational experiences. Hiking, including nature walks, is the outdoor activity that benefits directly. Nature study is enhanced if a wide variety of biotic types are available, or if one or two atypical types are abundant. The vegetative character of each park is analyzed, classified, and scaled into one Of four classes: 4. Evergreen. 3. Mixed Evergreen and Deciduous. 2. Deciduous. l. Barren. This scale Of vegetation desirability for a park is adapted and condensed from the Meramac Basin Study.s 5Ullman, Boyce, and Volk, Op. cit., p. 44. 192 This scale is certainly Open to questions of judgment. Most campers in Michigan State Parks are from the more populated southern counties, so parks with predomin- antly evergreen vegetation Offer the majority Of campers a marked change in vegetative character from that of their home area. Such a change is assumed to be a positive quality factor. Some additional measures of quality Of vegetation Parks with large stands Of virgin are also attempted. Virgin timber timber are rated on a binary basis. stands are a rarity in this region, found only in three parks, and are assumed to be desirable for enhancing the total recreational experience. Another measure, the wilderness quality of a park, also is scaled on a binary basis. If a park is 5,000 acres or larger and has few unnatural environmental detractions, it is rated as a "wilderness type" park. This definition is similar to that established by the U. S. Forest Service for virgin areas.6 The three parks that fit this classification have unique physical qualities, includ- ing a vast area Of virgin and second growth forests, that could not be overlooked. Certainly campers who desire develOped campgrounds along with the Opportunity for day-long hikes into the semi-wilderness, would find these parks desirable. 6U. 8., Forest Service, Op. cit., p. 11. -_ '4.- I: 193 Shade Camping enjoyment is usually increased if camp- sites are well shaded, although not always. In northern Michigan direct sunlight may at times be pre- ferred by campers, not only for warmth but to hasten the evaporation of moisture from.tents. The Gilbert Study included a question on shade in the campgrounds. In this survey it was found that 95% of the men and women campers preferred well-shaded campsites.7 With this information in mind the amount of shade in each state park campground was estimated. One Of four percentage figures--75%, 50%, 25%, or 10% of the area in shade-~is used as a measure of this quality. Wildlife The wildlife Of a park can be one of the most gratifying amenities. Empirical measurement Of this amenity for an index of attraction, however, is quite burdensome. NO data on the wildlife of the parks were collected for use in this analysis. Climate The climate of parks can affect activity use more than any other factor. Campbell discusses three gen- eral ways in which climate can encourage or discourage 7Gilbert, Op. cit., p. 122. 194 8 The first of these is the use of a recreation area. the effect on human comfort, such as the sensible temperature. This has been measured by the U. S. Weather Bureau in a comfort index, in which extreme discomfort is felt when the temperature and humidity are high. A lack of available data precluded the use of a comfort index for each park in the attraction index. A second category discussed by Campbell is the psychological effect of climate. This intangible effect occurs when cloud overcast shuts out direct sunlight for long periods of time, or when precipitation occurs for prolonged periods. For hay fever sufferers, cer- tainly the psychological boost received by visiting relatively pollen-free areas of northern Michigan is large. Rainy periods or a lack of sunshine are not necessarily detrimental to the relative attractions among Michigan State Parks, therefore, these elements are not measured. Campbell's last category is the effect of climate on outdoor activities per se. Climate primarily affects water activities: land activities to a lesser extent. A high comfort index in urban areas may encourage the pOpulace to travel to and camp in a state park. The comfort index may still be high at the park, but heat and humidity are often tolerated better in an outdoor 8Campbell, Op. cit., p. 36. 195 setting, particularly if swimming is a possibility. Climate also directly affects the temperature of the water. Other physical variables, however, also affect water temperatures and the activity of swimr ming. If the water temperature is sufficiently warm --generally above 68’F.--swimming frequently will be attempted.9 Even if the water temperature is suitable, a combination of low air temperature and high wind may dampen enthusiasm for swimming. Sunbathing, considered as a part Of the swimming experience, is directly related to the availability of direct sun- light.10 The Forest Service suggests that the temperature differential between a recreational site and the user pOpulation centers be used as a measure of attractive- ness.11 A similar measure was used in ORRRC Report 5.12 Recreation sites were scaled in this study in terms of ”climatic relief.” 9U. S. Forest Service, Op. cit., p. 84. 10In one survey, 5.9% Of all groups taking part in sunbathing as an activity were dissatisfied with the activity. Thirty per cent of the reasons cited for dis- satisfaction could be attributed to bad weather, refer- ring tO clouds, cold, and wind. (ORRRC Report 5, 9p; cit., p. 34.) 11U. S. Forest Service, Op. cit., p. 79. 12ORRRC Report 5, Op. cit., p. 217. 196 Excellent--High relief from.discomfort-- 20° Differential Good ------- Moderate relief from discomfort --10° Differential Fair ------- Slight relief from discomfort-- 0-10° Differential Poor ------- NO apparent relief--Negative Differential In addition to temperature differential, scaling was done on the basis of other atmospheric conditions. For one recreation site near Los Angeles, smog was recognized in rating the recreation site.13 The idea Of a temperature differential between parks and the origins Of a majority of users has considerable merit, but is not undertaken here. The major reason is the complexity of the measure involved for fifty-nine parks and the fact that the origins of users were not clearly established at the time the index was con- structed. Measuring the climatic effect on outdoor activi- ties in Michigan State Parks between June and September is not as easy as it might seem; lack of suitable climatic data is a major problem. As a result, air temperature is the only measure used. A majority of parks keep a daily temperature record; however, no 13Personal conversation with L. M. Reid, May 18, 1965. 197 indication is given as to the time the temperature is recorded. When records of this type were available, air temperatures at each park for July were averaged and recorded. If air temperature was not available at the park, the average July temperature at the nearest weather station was used. This measure gives parks located in northern Michigan a higher attraction rating than parks in southern Michigan. Water Resources The importance Of water-related activities to campers has already been emphasized. NOt only are water-based activities preferred, but other outdoor activities are more attractive when undertaken near water sites. Most of the variables selected and measured are related to inland lakes and Great Lakes characteris- tics, with a limited amount of variables devoted to the characteristics of rivers and streams. Rivers and streams are more difficult to develOp and utilize for water activities, particularly for swimming, and river or stream frontage adjacent to campgrounds is less desirable than lake frontage. However, the former is preferred to no water frontage at all.14 14Gilbert, Op. cit., p. 123. 198 Size of the Water Body The importance Of the size Of a body Of water varies when considering water activities. For swim- ming, which utilizes the margins of a water body, size is important for its relation to temperature. A very large body Of water, if of sufficient depth, can be a very cold one. As mentioned under climate, swimming is not a popular activity if the water temp- erature is below 68°F.15 Lake Superior, as a result of its size and great depth, frequently is not util- ized for swimming. The waters of Lakes Huron and Michigan attain somewhat warmer temperatures, and as a result, parks located on these lakes are popular for bathing. In contrast, a small, shallow inland lake can heat to temperatures Of 80’ or more. Lakes of this type, although intensively used, may be too warm for the enjoyment of some bathers. The same records that included data on air temp- eratures for parks include space for recording water temperatures. Unfortunately, Conservation Department records for July,l963 included water temperatures for only twenty-two Of the thirty-eight parks where it is possible tO swim. For these twenty-two parks, water temperatures for the month were averaged. July water temperatures for the remaining eleven parks were 15U. 8. Forest Service, op. cit., p. 84. "r: g. pun. 199 estimated with the assistance of planners in the Parks Section of the Conservation Department. These esti- mates and the twenty-two average temperatures are used as one quantitative measure of the swimming enjoyment in parks where swimming is permitted. A park with relatively high water temperature is considered to be ._fl more attractive for swimming than parks with lower water temperatures. L'HAE _‘-.-‘ ‘ The size of the water body also is extremely a... important for activities requiring the use of water- craft. Congestion Of motor boats on inland lakes in Michigan has become a major problem in recent years.16 ”Rule Of thumb" capacity standards for this activity as a partial planning guide for reducing water conges- tion have been scarce.17 The acreage Of inland lakes with public access for boating is used as one measure of the quality of the water resource for boating and 16C. S. Van Doren, ”Recreational Boating in Michi- gan,“ Unpublished paper presented at Michigan Academy Of Science, Arts and Letters, Wayne State University, Detroit, Michigan, March, 1961. 17One of the few "rule of thumb" standards of this type was located. This study stated: '. . . it has been assumed on the basis of interviews and Obser- vation that there is a minimum requirement Of 20 acres of a desirable degree of skiing satisfaction.“ . This study also recommends that inland lakes of less than 50 acres should have water skiing prohibited. Wisconsin, Department of Resource DevelOpment, Recreation in Wisconsin (Madison, Wisconsin, November, pp. 0 200 as a measure Of the total scenic qualities of the park. A measurement of this type means that the largest lakes are the most attractive,as large lakes can accommodate more boaters, fishermen, and water skiers with less congestion and possibly less conflict between water activities. if An additional measure of size relative to water activities is also attempted. Some boaters prefer to pleasure cruise, and navigable connections between W'X': ': lakes add to this activity experience. Since there i- are state parks located on inland lakes that have navigable access to adjacent lakes, this Opportunity is included as a variable. A different analysis is necessary for the Great Lakes. The dangers Of boating on the Great Lakes in small boats except during periods of assured calm are Obvious. There are few develOped boat launching sites in Michigan State Parks with access to Great Lakes. Parks on the Great Lakes are used intensively, however, for swimming, shoreline hiking, and sunbathing. For this reason the foot frontage of a park on the Great Lakes is used as a measure of size for this variable, complementing acreage measure of size for inland lakes. Shapes Of Water Bodies The shape of a water body, while not Of importance for swimming, is very important for water skiing 201 and boating. A large, round lake is much safer for water skiing than an elongated, narrow one. In con- trast, a narrow lake with irregular, penetrating spurs may be very popular for pleasure cruising. Sailing requires a large Open space for unobstructed winds, while canoeing may be hindered by these characteris- tics. The shape Of water bodies is briefly discussed here to take cognizance of its importance, but no measures of this variable are attempted for the index.18 Rivers and Streams For river waters size and shape are less impor- tant, since few water activities are undertaken unless the river is of sufficient width for boating. In Michigan State Parks there are few points Of access for boaters, and there are no develOped swimmdng sites on river banks, but the rivers are important for such activities as bank fishing, hiking along the banks, and canoeing. The velocity of a river or stream and Obstructions such as falls and rapids do affect water-related activities. Streams and rivers with swift currents are sometimes enjoyable for “white water“ canoeists, while others prefer calm 18For a method Of measuring shapes that might be useful in classifying the shapes of inland lakes for planning purposes, see Chapter III in William Bunge, Theoretical Geography, Lund Studies in Geography, Ser. 5. General and Mathematical Geography NO. l (Lund, Sweden: C. W. K. Gleerup, Publishers, 1962). 202 water. In many cases rivers containing "white waters“ were recorded, but this variable was not used in the index since few of the parks are large enough for long canoe trips. The length and width of the river and stream frontage in each park was determined, and these values were used as measures of the quality of this 1.... can] . ,4 resource . Water Quality The water quality of a lake or river in relation to water activities can be measured by its turbidity, temperature, and the amount of pollution. Turbidity is probably the most difficult to measure and the one for which information is lacking. Obviously a clear lake or river is aesthetically preferable to a cloudy or murky one. When information on turbidity was available, it was recorded on the inventory sheets but it was not used in compiling the final index because a measure of this variable was impossible to obtain for all parks. The temperature of a water body has been dis- cussed previously. Water temperature is directly related to the depth Of a lake and its circulatory system. A large, shallow lake will have higher water temperatures than a deep lake and, depending upon its size and water circulation, varying degrees of pollu- tion. 203 The effect of polluted water is felt more directly by park visitors. Polluted waters do not necessarily preclude boating as an activity, but could temporarily hamper and even curtail swimming and water skiing if infectious organisms or organic chemicals were of sufficient content to endanger those in contact with the water. Pollution problems are more acute on the Great Lakes and some Michigan rivers than on inland lakes. One state park, Sterling-Monroe, received a lower activity rating because pollution of Lake Erie waters has required the Conservation Department to post no-swimming signs at this park. A measure of pollution was recorded but was not used in the index because Of incomplete data. Pollution has been indirectly measured, however, in terms of a fishing quality measure. Pollution, water temperature, and depth of a water body all influence the fish habitate and therefore have a direct bearing on the success and quality Of this outdoor activity. Fishing Quality and Success Fishing ranks second among the activities preferred by campers. For this reason the fishing quality and fishing success on water bodies located within or ad- jacent to Michigan State Parks was recorded for the attraction index. Quality as used here implies the type of Sport fish, that can be caught, such as pike 204 and trout, as opposed to less desirable species. Success implies a reasonable probability of catching a desirable species while on a camping trip. Each of the inland lakes and rivers having access points within Michigan State Parks was rated for fishing quality and success by Dr. W. H. Tody of the Fish Division Of the Michigan Department Of Conservation according to the following numerical scale. 4 -- Excellent. 3 -- Good. 2 —- Medium or Average. 1 -- Poor. 0 -- None. Similar ratings were applied to Michigan State Parks located on the Great Lakes, providing a dock or pier was accessible to those wishing to fish. These structures did not have to be located within the park proper but within a reasonable walking distance (one mile) from the campground. A rating Of this type is an empirical estimate on Dr. Tody's part, but with his knowledge Of creel census data, fish habitat characteristics, and when and where lakes or rivers have been stocked, his judgments are con- sidered the best available. 205 Beach Characteristics Since swimming was the most popular of all out- door activities preferred by campers, considerable attention was given to the quality of beaches on the Great Lakes and inland lakes. Two surveys already mentioned point out the importance of beaches in state parks. The 1956 Dahle Survey states that when given an Opportunity to suggest improvements in parks, the public expressed a desire for extending and enlarging beaches, although desire for more beach area ranked twelfth out Of nineteen suggestions on the list (see Table 16).19 The 1961 state park visitors in Wisconsin were much more positive in their desire for beach improvements. Of eleven major improvements suggested by visitors, the need for beach improvements was ranked third.20 Five beach characteristics assumed to effect the attractiveness of a park for swimming as well as wading, sunning, and shoreline hiking are length and width of the dry beach, composition of the dry and wet beaches, and distance of the wet beach to a five-foot depth.21 19Dahle, Op. cit., p. 10. 20Hutchens and Trecker, Op. cit., p. 37. 21The previously mentioned WOrk Plan for the National Forest Recreation Survey recommended these factors as attributes for utilitarian use of beaches. Many of these same attributes were recommended by a geographer and 206 The length and width Of useable beach frontage was estimated from.Park Master Plans and from dis- cussion with employees of the Parks and Recreation Section of the Conservation Department. In some cases the length of beach corresponded with the total water frontage in a park, but this was not always true. The footage Of beach length was used as a measure Of attractiveness-~the longer the beach, the more attrac- tive the park. The width of beaches was judged in the same manner. Beaches with widths of fifty feet or more, for example, have been recommended as best for swimming and boat launching purposes.22 Rating composition Of the dry and wet beaches might have been done according to the system.recommen- ded by the U. S. Forest Service.23 -- Sand. -- Gravel. Timbered. N w :5 01 I I 1 -- ROCkO Without on-site evaluation of the quality Of beaches a rating of this type was impossible. If a personal measured in another area in 1948. See Alfred W. Booth, The Lakes of Ehe Northeastern Inlanthmpire, A Study Of Recreational Sites, Bulletin NO. 5 (Bureau of Economics and Research, State College of Washington, Pullman, April, 1948). 22U. 8. Forest Service, op. cit., p. 85. 23Ibid. 207 inspection of the beaches had been possible, then both dry and wet beaches could have been evaluated in detail. For example, a fine sand beach extending from.the high water mark for more than fifty feet could have been scored higher than a sand beach with gravel at the high water mark. Dry and wet beaches are scored on a binary basis here, 1.0 for sand beach and 0.0 for gravel or rock. These scores were allocated with the assistance of staff members of the Parks Section, Conservation Department. The wet beach score was applied to the composition Of the wet beach to a depth of five to six feet. This evaluation was made for swimming activities since wet and dry beach composi- tion is not as crucial for boating activities. However, good characteristics for swimming are also good for boating. The lepe of the Off-shore bottom also was estimated. A distance of 100 to 300 feet to a five- foot depth was considered the most desirable. In a few parks the five-foot depth is not reached before. 300 feet. These wet beaches are thought to be less desirable from one safety standpoint. At these sites, children could advance a long distance from.shore, and if for some reason their life were endangered, a would-be rescuer would have a lengthy span of water to cross before reaching them. However, from another 208 viewpoint, a gentle lepe, regardless of the length from shore to a five-foot depth, provides a less dan- gerous situation in that there is less Opportunity for a child to step into deep water by moving only a few feet. Taking this viewpoint, it.was decided that the estimated length Of the beach to a five-foot depth would be used as a measure. This means that parks with wet beaches of gentle lepes are more attractive than those with steep slopes. Such a judgment is biased in favor of swimming and is not necessarily good for boat launching. For boating, slopes should be greater, dropping rapidly at least to a two-or three-foot depth, and reaching a five-foot depth within 24 Since swimming is the more impor- twenty-five feet. tant of these two activities, the qualities ideal for boating have been placed secondary to swimming qualities. Cultural Features The term ”cultural features" is defined as evi- dence Of past or present human occupance or activity in an area. Several Michigan State Parks contain historical structures that constitute attractions in their own right. A good example is Fort Wilkins State Park, with the reconstruction Of a nineteenth century settlement and fort.' Archeological attractions also 24Ibid. 209 fall into this category as do unique features Of a contemporary nature in sight of a park--such as the bridge across the Straits of Mackinac or the view of shipping on the Great Lakes.25 The measurement of cultural or historic features, like natural features, is difficult. As a result, cultural features at a park are also treated as binary variables. 25The Van Til Report found that campers in Michi- gan State Parks were not interested in the historic or natural aspects of a park or its environs. The report elaborates by saying that this is probably a truthful statement, since there were only two parks with his- toric features of significance in the survey and the state's naturalist services were limited to a few parks. The report suggests that more public interest might be demonstrated if these aspects Of a park's attraction received more emphasis. Report of Committee on State Parks and Public Lands, op. cit., p. 10. APPENDIX C 210 APPENDIX C RATING THE QUALITY OF FACILITIES AND SERVICES Water Supply The water supply in campgrounds in Michigan State Parks was not measured for inclusion in the attraction index. The assumption is made that a sanitary water supply is provided at convenient locations in all parks.1 When some campgrounds are at maximum capacity for prolonged time periods, there is evidence of minor dissatisfaction with the water supply. Campers also have expressed a desire for tap water as Opposed to pumps.2 Such dissatisfactions, however, are minor relative to other facilities and services. 1Minor dissatisfaction was expressed by respon- dents surveyed in ORRRC Report 5 as to the location of the drinking water supply. Overall, the public expressed satisfaction with the drinking water supply at the sample locations, in spite of the fact that one location had no water and visitors had to bring their own. The report does not elaborate on this point, but it can be assumed that the attractiveness of this particular area was reduced to a large number of potential campers. This means that its attractive- ness to those wishing to be isolated was probably increased. ORRRC Report 5, pp; cit., p. 39. 2Dahle, op. cit., p. 10. 211 212 Cappgpounds Campground, as used in ORRRC Report 5, is a broad term. It encompasses many facilities included at each campsite and utilities available at selected locations within a campground. Facilities to be considered for each campsite include electricity, fireplaces, tables, and the amount of shade or cover in the campground. (The latter feature already has been discussed.) Features frequently present within a campground in a few selected locations are a water supply, toilets, laundry facilities, showers, a wood supply, and refuse containers. In addition, a nearby campground store is frequently considered essential. With the exception of a store or concession, these features of a campground were not explicitly listed in ORRRC Report 5, but complaints expressed by users indicated that they were included in a user's evalua- tion of the site. A majority of campgrounds at the twenty-four interview locations used in ORRRC Report 5 were satisfying to the public. Specific complaints were directed towards overcrowded facilities, lack Of showers, and dirty grounds. Many of these same complaints have been directed to facilities and services in Michigan State Parks as evidenced by responses to two questions in the 213 1956 Dahle Survey related to facilities and services.3 Although the camper population was not interviewed as a distinct group, most Of the respondents in the sur- vey directed a majority Of their comments towards campground and toilet facilities (Table 16). The Van Til Report indicates that when asked about facility expansions, visitors--primarily campers--made similar responses. Fifty-four per cent thought toilets and shower buildings needed expansion and 35% and 33% respectively expressed a need for more campsites and campstoves.4 A direct comparison of these two surveys with ORRRC Report 5 is impossible. The ORRRC Report did not discuss each campground variable that is to be consid- ered here. However, general complaints in Michigan State Parks which indicate crowded conditions, such as the requests for more restrooms and expanded camp- grounds, can also be identified in the ORRRC Report.5 The Gilbert Study6 provides the best insights into the facilities and services desired by campers 3Users were asked, “What, if any, facilities appeared to be inadequate or lacking in the park?“ and "What new improvements or expansion of present facil- ities (other than those previously mentioned) should be made?” Ibid., p. 9. 4Report of Committee on State Park and Public Lands, op. cit., p. 39. 5ORRRC Report 5, Op. cit., p. 38. 6Gilbert, Op. cit., p. 121. 214 TABLE 16.-~Most frequent suggestions for park improvement from Michigan State Parks Users Survey, 1956 I ‘- Personala Voluntaryb Suggestions Survey Response More restrooms I 55 142 More electric outlets 43 39 Cleaner restrooms 40 34 Better laundry facilities 35 57 Fire places and stoves 34 38 TOO crowded (expand park) 33 4 Better parking facilities 30 63 Extend beach 24 16 Boat launching facilities 23 30 Hot water service 23 43 Expand camp area 23 58 More water service 20 -- Mosquito control 17 8 More tables 16 61 Running water 16 42 Showers 16 68 More recreational facilities 13 -- Clean it up 13 14 Diving platform. 11 20 Miscellaneous 17 -- aThe suggestions listed in the original table have been reordered for the personal survey by decreasing number. bNot all voluntary returns were tabulated on this question. Source: Thomas L. Dahle, Michi an State Pgrk Users Surve , 1956, Bureau o? Business Research, CO ege 0 Business and Public Service, Research Report Number 19 (East Lansing, Mich.: Michigan State University, 1956). p. 10. 215 in or adjacent to campgrounds. This study grouped facilities and services into categories that would demonstrate desires for developed or underdeveloped campgrounds. In this study, 120 men and women campers were asked to rank the desirability of twenty-two facilities and services in the three campground types listed below. 1. Full facility campgrounds - which included tap water, flush toilets, electricity, tables, and stoves. 2. Minimum facility campgrounds - with pumped well-water, pit toilets, tables, and stoves. 3. Isolated campgrounds - with none of the above facilities-and with limited accessibility. . A full facility campground was preferred by 58% of the men and 64% of the women: minimum facility grounds, by 37% of the men and 32% of the women. Only 3% of the men and 1% of the women desired an isolated site.7 These preferences, when compared with those implied from the Dahle Survey and the Van Til Report, indicate the demand for developed or semideveloped campground facilities. Since the Iron County campgrounds surveyed 7Ibid., p. 121. 216 in the Gilbert Study are believed to be representative of Michigan State Park campgrounds,8 many Of those facilities and services included in a full facility campground as defined in the Gilbert Study were inven- toried and measured for inclusion in the index of attraction. The twenty-two facilities and services included in the Gilbert Study are listed below. The list has been divided into four groups, according to the cate- gorical ranking of importance as shown by the survey questionnaires.9 very Important Safe water Toilet facilities Garbage disposal Tables Preserving area in natural state NO crowding Important Life guard Well-marked trails Electricity Launching ramps Brush clearance Not Too Important Laundry facilities Campground showers First aid stations Fireplaces Smooth trails Boat docks Not at All Important Boat rental Cafe or restaurant 8This assumption was made after consultation with L. M. Reid, Department Of Research Development, Michi- gan State University. 9Gilbert, 9p. cit., p. 190. 217 Library facilities Planned recreation Child-care facilities The facilities and services liSted in each category Of importance include variables that are not necessar- ily a part of the campground prOper, but are related to certain outdoor activities. These are discussed under their appropriate subtitles. At this point the above list is useful for selecting and measuring important campground facili- ties and services. Facilities and services directly related to the camping experience within a campground are garbage disposal, tables, electricity, laundry facilities, showers, first aid stations, and fire- places. If water supply (which has already been discussed) and toilets (to be discussed separately) are added, then a rather complete list of facilities and services for the camping experience can be compiled. If these variables are analyzed according to the pref- erences in the above list, then it becomes apparent that campers are concerned first with health and sani- tation measures and then with facilities and services that make camp houskeeping duties easier. It is assumed, as was the case with water supply, that facilities directly related to health and sani- tation are adequately provided and maintained in Michigan State Parks. This would eliminate garbage disposal as a variable as well as first aid stations. 218 It is also assumed that a table is a ubiquitous item provided at each campsite. Electricity, showers, laundry facilities, and fireplaces are amenities to a camping experience. Since they are provided in many Michigan State Parks, and are not available in less-developed campgrounds such as state or national forest campgrounds, they may constitute a part Of the camping attraction for some state parks. The demand for electricity at campsites is evi- dent in the Dahle Survey of 1956 (Table 16). The Van Til Report indicated a need for more trailer sites, which implies sites with electrical outlets.10 Trailer campers, and many tent campers as well, prefer a source of electricity.11 The number of campsites with electric outlets in each park is used as a measure Of this campground service. Better laundry facilities were requested by large numbers of respondents in the Dahle Survey, while showers did not receive a high preference rating (Table 16). In the Van Til Report, however, showers were indicated as a feature that needed expansion, and 10Report Of Committee on State Parks and Public Lands, Op. cit., p. 14. 11It is now the policy Of the Michigan Department of Conservation to provide electricity at every camp- site. (Personal consultation with Harold Guillaume, Parks Section, May 14, 1965.) 219 no mention was made of laundry facilities.12 These two variables were listed as “not too important" in the Gilbert Study. Since they are present in Michigan State Parks, as opposed to other public campgrounds in the state, and there is some evidence that they are preferred by campers, their inclusion in the index seemed warranted. These two variables are measured on a binary basis. Fireplaces for each campsite, like showers and laundry facilities, were ranked as ”not too important" in the Gilbert Study. The Dahle Survey (Table 16) and the Van Til Report13 indicated that fireplaces or campstoves were desired by campers. An inventory of the number of campsites in each park with fireplaces theoretically was possible, but was not felt to be justified. Even when fireplaces are provided, build- ing fires is frequently discouraged in parks due to fire danger. One remaining feature relative to campgrounds is included in the index. In all of the surveys and re- ports discussed, an indication of public dissatisfaction has been expressed by such phrases as "no crowding,"M 12Report of Committee on State Parks and Public Lands, op. cit., p. 14. 13Ibid., p. 14. 14Gilbert, op. cit., p. 190. 220 "too crowded,”15 or ”expand campsites."16 Obviously the capacity of a campground has an effect on the attractiveness of a park. For this reason, the number of campsites in each park is used as a measure of capacity and, indirectly, attractiveness. It is pointed out in a more detailed discussion in Chapter V that this measure of a park's attractiveness for camping is a primary variable in the index. Parking Little dissatisfaction was expressed in ORRRC Report 5 with parking facilities, but there have been complaints of this type in Michigan State Parks, as noted in the Dahle Survey (Table 16) and in the Van Til Report.17 The size of each campsite is designed to include parking space for one vehicle. Parking problems generally are acute in day-use activity areas, for instance boat-launching ramps. If these activity areas are of sufficient distance from camping areas to require automobile travel to reach them, then parking problems at these areas would detract from.part of the camping experience. Time was not allocated for 15Dah1e, op. cit., p. 14. 16Report of Committee on State Park and Public Lands, op. cit., p. 14. 17Ibid. 221 collecting data on the capacities of the campground parking lots, but data was gathered-~as discussed below-~on parking space at boat ramps. Toilets If the order of facilities and services in ORRRC Report 5 is to be followed, a discussion of this var- iable is not warranted at this point. Any discussion of campground facilities, however, must include an analysis of facilities as important as toilets, and for this reason, it is included here. Toilets were found to be adequate by users of the national sites in ORRRC Report 5, although one-fourth of the respondents were dissatisfied with "dirty” toilets.18 A similar complaint is listed in the Dahle Survey (Table 16). The public also voiced dissatisfaction with toilet facilities in the Van Til Report, in that more restrooms were needed.19 Nothing was said in ORRRC Report 5 on the desir- ability of flush toilets as opposed to box or pit toilets. In heavily used areas, flush toilets are preferable to box toilets in terms of sanitation and for reducing objectionable odor. Since a majority of campgrounds are used to capacity, flush toilets are 18ORRRC Report 5, op. cit., p. 40. 19Report of Committee on State Parks and Public Lands, op. cit., p.’II. 222 obviously better. Consequently, campgrounds with flush toilets are given a value of 1.0, while parks with box toilets are given a value of 0.0. The Gilbert Study offers some justification for this assumption, in that flush toilets were included as a preferred variable for full facility campgrounds.20 Marked Nature Trails Next to the camping experience itself, hiking is the most important land activity. Only a few Michigan State Parks have marked nature trails. These parks are measured by binary scale under interpretative facilities (Table 18) which includes nature trails with interpretative brochures, outdoor centers, infor- mational displays, and guide services for special groups. This variable essentially includes the ”signs and information trails“ and the I't:ours and organized group“ categories included in ORRRC Report 5. In gathering information on hiking and nature trails for the attraction index, a more direct measure was used in addition to the interpretative facilities binary measure. Two features related to trails are considered to reflect the quality of this facility-- the length of the trails and the extent of trail marking. They are scaled by measuring the length in 20Gilbert, op. cit., p. 189. 223 feet of all marked trails within each park. For the thirty-two parks with marked trails, these values range from 4,000 to 370,000 feet. This method of measurement means that parks with lengthy marked trails are more attractive than parks with unmarked trails or than parks with no trails. Bridle trails are measured by the same method. Concessions and Rental Facilities Since a concession operator in Michigan State Parks in many cases will also rent articles for out- door activities, these two categories from.CRRRC Report 5 are combined. The chief complaint listed in ORRRC Report 5 against concessions and rented facilities "21 No information is available was “too expensive. for Michigan parks on complaints of this type. These twotnategories have not been evaluated in any of the surveys and studies discussed so far. Concession and rental Operations are amenities to a camping experience. The presence or absence of such services affects the activity participation at parks. Therefore, in parks where they are provided they are considered as assets to the attractiveness of the park. Based on judgment as to what is conven- ient distance, parks are given a value of 1.0 if there 21ORRRC Report 5, 0p. cit., p. 40. 224 is a store less than a mile from the campground. Parks with no stores or stores more than a mile from the campground receive a value of zero on this var- iable. Parks with boat, canoe, or horse rentals also are scored in this manner. Boat Docks and Ramps Users generally were satisfied with boat docks and ramps according to ORRRC Report 5.22 The primary complaint among the few dissatisfied was that condi- tions were crowded, particularly at boat ramps. The presence or absence of boat-launching facilities is noted for each park. There are some parks where congestion problems are acute. Congestion on the launching ramp and inadequate parking space adjacent to the ramp is recorded on a binary basis. This data, however, proved not significant for the index. The availability of a breakwater or pier within or adjacent to a park located on the Great Lakes is also recorded. These were considered meaningful for strolling and for fishing. Roads Road characteristics and conditions in the vicin- ity of parks are treated superficially. The completed Bureau of Outdoor Recreation Inventory Sheets for each park contained a judgment of access roads, principal 225 road conditions, and accessibility to each park.23 This data was recorded on the inventory sheet but in the final analysis was not included in the attraction index. The travel model takes this variable partially into account in that time distances are computed on the basis of road conditions and minimum distances. Other Facilities and Services The National Survey (ORRRC Report 5) included a category for write-in comments. The comments received in this category were directed toward specific facil- ities such as fireplaces, tables, firewood, insect 24 Tables have been dis- control, and playgrounds. cussed previously. With the exception of playground equipment, none of these minor facilities and services were inventoried or evaluated for the attraction index. An effort was made to determine the availability of playground equipment in parks. There are a number of reasons for this. The camping attraction index is designed to reflect the desires of the average party of family vacation campers. Such parties fre- quently include one or two small children who derive 230. 8., Bureau of Outdoor Recreation, Inventory, Classification and Evalpation of Existing Outdoor RecreationallAreas and Facilities (Form B.O.R. -73), Department of the Interior, Washington, 1965. 24ORRRC Report 5, op. cit., p. 42. 226 considerable pleasure from playground equipment.25 These facilities serve as a social gathering place for the young and they enhance the camping experience for adults as well. Respondents in the Van Til Report heavily endorsed swing-teeter-slides as apparatus that would be used if available. The report also states that in many cases it is the children who influence where a family takes a camping vacation.26 Parks which included campgrounds with play equipment were given a score of 1.0, while those without play equip- ment were assigned 0.0. Several other facilities and services not men- tioned in ORRRC Report 5 which were inventoried and scored for possible inclusion in the index are museums, outpost camps, rifle and archery ranges, dune ride concessions, lifeguards, and bath or beach houses. Congestion In all surveys reviewed overcrowded parks were mentioned as a complaint. The attraction index developed for this study does not include this factor. However, data on congestion at boat launching sites on inland lakes and on the per cent of the time each 25"The typical camping family totals 4.6 persons with 2.2 children under 16; . . .' Report of Committee on State Parks and Public Lands, opicit., p. 9. 261bid., p. 11. 227 park was at 100% campsite occupancy during the 1963 season was collected.27 In a sense, information of this type constitutes a detraction rather than an attraction. Campers may learn that some attractive parks will continuously be full on holidays and weekends and select another less attractive park for camping where they will not have to queue to gain admittance. As the attraction index is refined, it may be possible to utilize congestion figures of this type in such a way that they will negate a portion of a park's attractive qualities. 27One of the purposes for collecting the data on congestion was to use it to test the attraction index after it had been used in a travel model. For example, the attraction indices could be compared to measures of congestion by a correlation analysis. APPENDIX D 228 TABLE l7.--Park Inventory Table. REGION ACREAGE LAND WETLANDS WATER TOTAL CAMPSITES--LOTTED UNLOTTED TOTAL TENT TRAILER OUTPOST (shelters) PARK NAME Terrain--LoR.H. L. Rank-Rolling Flat Hilly MtSo Landform Surface Vegetation Evergreen Deciduous Mixed Barren Virgin Cutover Cover-Shade loo-50% 10-25% 25-50% 10% or Less Unusual Vegetation Wilderness No Detractions Serious Minor Unacceptable Substantial Wildlife Habitat Excellent Poor Normal Historical Archeology Contemporary Museum Springs Falls Cliffs Overlook Dunes Interpretive Facilities 229 230 Capping Store In park Out park Showers Toilet-Box Flush Laundry Electricity Occupancy--% 90% 50-90% Below 50% Play Sports Equip. Horseback RidipggRental Miles-No. Trails Hiking-No. Trails Miles-Trails Other Great Lakes-- Total Beach 1 Mile + l/2 Mile to 1 Mile Less than 1/2 Mile Water Tegpgrature 73°F+ 68°-72’ 60°-67° 59‘ Less Beach-Width 100'+ 100'-50' 50'-25‘ 25' Less Beach-Composition Sand Gravel-Rock Offshore-Bottom.to 5' Depth Sand Pebbles Sand + Pebbles Distance to 5' Depth 300'+ 100'-300' 100'-50' 50' Less 231 TABLE l7.--Contd. Swimming Class A B C D Bath House Life Guard Fishing Qualipy None Medium Poor Good Excellent Pier Access Distance Paved Highway Access Roads-Adequate Princ. Roads-Adequate Rec. Class-(Area)-ORRRC I- III- V- II- IV- VI- River-Stream Length Width Flow-Still hite Fishing Canoeing Inland Lakes Size 1000 Acres+ 50-1000 A. 0-50 A. T Marsh-Bog Marsh-Lake WarmpWater Cold-Water Pollution None or Minor Polluted-Acceptable Unacceptable 2 O o Z O O Z O O 232 Turbidity 52;. ES: Clear Mnrky Cloudy Water Tepperature 73°F+ 68°-72° 60°-67° 59° Legs—II.— Water Level Little Change Moderate Major Change Beach-Lengph 1 Mile 1/2 Mile to 1 Mile Less 1/2 Mile Beach-Width 100'+ 100'-50' 50'-25' 25' Less Beach-Composition Sand Gravel Offshore-Bottom to 5' Depth Sand Pebbles Sand + Pebbles Offshore Distance to 5' Depth 300'+ 100'-200' 100'-50' 50' Less Swimming Class A B C D Bath House Life Chard ROped Area Boating Launching Rental Obstructions Navigable Waters Limited Expanded 233 TABLE 17 o "COtho Water Skiing Area-Limited Extensive Water Congestion Major Normal None Fishing Quality None Medium Poor Good Excellent Pier Attendance Camper Days 1962 1962 ‘ ' 1963 1963 1964 1964 Average Temperature Diff. 234 new: oflcmmumomoa hm>usm Hmowmoaoou .m.o paw cowus> nummcoo mo useguummmo savanna: as ummm omm.a on m moaamm swoon .m usommum u H ca Hammad n o mmemo umomuso .¢ Hflmm ”H“ u H om Demand n o mwuwm Hmafimna .m mouwmmEmu mm hmv on o m0 nonesz .m cowum>ummsou mo uameuuwama sameness so mmuom sws.mm o» es «Nam xuum .H sumo mo muusom o 1 .um> mamom mannaum> cuss mxumm Ho mosam> no mo .02 mmcmm manwmmom seas .uonouoo cmuoossoo mama .mump mo mousOm can .msam> ououlcoc cuss mxumm mo Hones: .mmsan> mo omens manwmmom mcwpsaocfl .mxumm mumum savanna: Mom pmwuoucm>sw moansfius>ul.ma mammfi 235 .Asmma .noumzc >Hq .mumcmwum Andean none can; unamsmc ham n .maaam nuns «chasm. amm .uawmsm umapmouuHc n~m u .mpcdams so macaw no «msucmuo I Q'mno loom snowumE¢ mo cowumfloomnm ncwmam cuoosm. Omd u m on» no massed .e .oz ucmsmammsm Ampssasoa cw macaw an: ...¢.m.o .mmuaum unmam «0 mmnnomun Imuuom on» :w Euom oommusmuocuq ncflmam cuooamv and n N oosmusm mason mo mmmmmso. .ucoeeum .m sheen ms .mcausm passe H4 n H unama «mad .5 .mwms ..o.a .coumcficmma .uOAumucH can no unusuumdmo .xnnnm .m.o.m such. mwauwawomm can nomad sowusmuomm uooousoxmwwumwxm msocwuussoz I v mo cowumsam>m can «acaumoaua Sedan u n unusao .muouso>cu .cowumouoom mcwaaom I ~ :«nuuoa no uoouuao no smousm ..m.= «m pass a H moumusm swam .o sumo mo mousom o x .um> madam manswuc> cues exude no mosau> no mo .02 meson manwmmom .UHQOUIlomH HAMuomcoo usomoum I H mo unmeuummoo cmmHnOHz m ucomnd I o uwnEHa chuH> .m .mmmH .mchcmn ummm .asHmum>HsD somumum>m n v mumum cmchOHz .usosm0H0>mQ cmmnmum>m can 00H50mmm mo acmeuummmo msospwomn pmxwz I m a.smmHEOHz cH ummuom acmEmHu uncapwowo I N uuommm no man:- .covsm> .o.b em couumm I H GOHuuuomo> .m sumo mo mousom o x .um> mHmom OHQMHMM> cuwz mxumm no mosHm> no mo .02 mmcsm mHnHmmom .CUGOUII.QH MHM¢H 237 c ”In I... II. II I: mHHsuu Doom an mHnHm Imooom I ucommum I H h ucomn¢ I c mouse pssm .mH mason» noon an macaw Imooom I ucomoum I H meOHum>o «H scoped I o I mMMHHU .hH usamup poop an manna m Imooow I ucommum I H Hammad I o mHHMh .mH «Hanna uoow an mans» Inooom I ucmmoum I H m acmmnd I o mmcHumm .mH anemone I H m Demand I o scams: .vH coHum>H0mcoo uconoum I H «Hanson mo usmEuummmo csmHEOHz h Hammad I o mumuomamucou .MH sumo mo oousom o \ .Hu> oHcom mHnnHum> cqu nxumm no nosHs> no mo .02 omcdm mHnHmnom ovuflOUll.wH WAM¢H 238 h. Ugmmem an H hv ucsnna I o musHHoa cmsHo .mN HgmflHm I H Hv ucsmnd I o nuozocm .NN coHusbusssou «HHS sso v.I H mo unsauusooo :smHEOHz mN sHHE sso.n I o snoum .HN .msmH .sHau .s .oz .mh .Ho> .souseaoo mo usuauusoso .mmmH .mHso .sstcoH: .suso Hso IHmOHousEHHU .sssusm usaussz ..m.: can .msaH .coHus>umm Icon «0 ussEuusoso :schon .sOHuosm sowussuosm pas mmmH .susu sxuso ou musspssucHusosm xusm Isusoasa uHs an smuuuensu museums asxmmz as .m .vs o» .mo sHsu muuuoss .o~ coHus>usmcou ussmsum I H nsHuHHHuso no ucssuusoso ssmHEOHz om ucsmnd I o stususuousucH .mH suso mo sousom o x .Hs> OHsom ansHHs> cqu mxuso Ho nosHs> no mo .02 smcsm sHmemom .quOUllomH mqm¢a 239 hi I... Hflmm ”Hm " H «nouns mm ussmos I o uoom poxnsz .om mHHsnB m umom ccs.a~ on ovm.mn mnsunm omens: .mu cOHus>nsmsou ucsmono I H xnso GH «0 ucsEunsoso cschqu v ucsmns I o ussm now sunom .mN conus>nmm mmmn no nos u on Icoo mo ussfiunsooo :smHEOHz no>o HQHImm H mm pcsonmosso sH .coHuosm cOHussnosm was mxnso nm>oo Rmmuom u,om Aspsnmv ns>oo unsum mchssHo an mousEHumm mm no>oo womIocH I me «0 ucso nsm .hm psHosooo usuHmoasU usmo om nna sea on «H nan soasnsooo .mm as mounm has on v nunonnuomnm .mm coHus>nsmcoo ucsmsno I H so uaosunmamo nauseous mm saunas u o museums .«N suso mo sonsom o \ .ns> sHsom sHoans> :qu mxnso no nssHs> no mo .02 omcsm oHonnoo .UHGOUII.QH mummfi 240 cOHus>nsnsoo mo ucueunsoso cschon .soHuosm GOHussnosm was mxnso .mmsum mcHsssHo sou an ssoH>ono mxnso moxsq ussnw so manna sauna: eons suuuanumn Nm some csn.snn on can may no omuuconm .sm sons mcHofisU ucsmsno I H cH ucsEoHsvm nu unmmns n o sasonmmunm .mm ucsmsnm I H cOHsmsocou N ucsmns I o ssHm ssso .vm unmm ”Hm n- H m humans a a «mass unmanns .mm ucsmono I H m ucsmnm I o mmcsm OHHHM .~m cOHus>nsscoo usso mHHsna ucHme mo unmannuaoa cumnnonz mm ooo.cnm on ooo.¢ no saunas .Hm suso mo sonsom o ‘ .ns> sHsom anans> ouHs nxnso no mssHs> no mo .02 amass anHsnom .OHGOUIl.mH wands 241 w. . _ at .E 11'... anti.- ”gigg— .mmmn .m .02 conumunnnsm .cmmn Icmmz mo hummn0>mso «noons cosy csmmmOHz sxsq mo nouns: nsusz , pas mucsnnsu .susuHumsH conssnsm msxsq ussno pas .AvmmH .hHso .H .02 nooso HsoHccosa .csmH uses: no sunmnm>ncs «scans ans. mdHmwlumsmlwdlausfllummsmlmmm mmmmmmmm .susuHuscH conssssm msxsq usmnw unsunsom mcH3oHH0m ssu Bonn psusEHums onus monsu Isnsoeou .nxnso mH mchHss Ion scu nom .momH cH psuuHE Inns mucssssucHnsosm xnso mo munoosn mess3 Eonu oscHsuno anus .msxsq usono IIsnsusnsoEsa snm3 mxnso HH nom msnsusnsoesa on .m ems ou com nous: mmsnm>¢ .mm :oHus>nsmcoo mo ucsfiunsoso csoHcOHz .cOHuosm coHussnosm saw unnmm .mnuum mancqmnm scu an pssH>ono sxnso mo moxsq ussno so ncsHo nsumsz Eonm sousEHumm Hm ussm omo.mm ou omH cumssq cossm .hm suso no sonsom o \ .ns> sHsom anans> cqu mxnso no nssHs> no mo .02 smssm sHoHsnoo .pucooll.mH Hausa 242 nummo .m ou socsusHo m~ ussm woo ou cm cosso us: .Nv mxuom 3mm nqu scsm I H ssxsq usonw xoom IsOHuHmooEoo en no\s:u Hm>mno I o zooms use .nq mxoom 3mm cuHs pssm I H msxsq usmnw xoom IsOHanooEoo s~ noxscu Ho>sno u o scams sun .oe .coHus>nsmsoo mo ucsaunsoso csmHEOHz .coHuoom :OHusonosm sum menus .uuuum mannamnm an mmumsnumn sac .mmmn .munmnm>nca susum csmHQOHz .usseo0H0>so sonsonsm no ussEunsoso «mchcsq Imam .Amunn manumnnsmc «an» Icsou ssmHEOHz mo sGOHusoHMHm ssxsq ussnu ImsHU sowusnonm .nhncofism .m.U Hm usso om ou oH InucH3 noswm .mm suso mo sonoom o x .ns> sHsom anans> cuHs mxnso no sssHs> no mo .02 smcsm sHonmoo .quOUllomH quda 243 aoHus>nsmaoo uasmsno I H msxsq ussnw mo uamEunsoso asmHaoHZ m uasmnd I o a0 nsHo .nv uasHHsoxm I v aOHus>nsmaoo @000 I n no .uoso asoHaOHz .aOHu EsHosz I N moxsq nomm guns as» so soon .m.3 noon I H usmnoumun an saoo maHusn Hsuasfimpah h saoz I o IHsao oaHamHo .mv uasmsno I H msxsq ussnu m acmmna u o usumsmmmnq .ms msxsq uasssno I H ussnwlsmsom ¢H uaomoa I o aossm no suso .vv aoHus>nmmaou sssoHHm I H msxsq so ucmaunuamo cmmnnonz mm smsonns uoz u o unassumanasnzm .ms suso mo sonsom o x .ns> sHsom OHAans> auH3 mxnso no msaHs> no «0 .oz smasm sHonmom .GUGOUII.QH mqmdfi 244 ~saH .muHmnw>Hco «scum asmHaon .uasEoOHs>so sonsOmsm unsusom mo uasaunsoso «maHmasH ussm ou anHs .Amuaaoo mm .mmIH .mozv naHusH Inoooa .xnsm IHam anouas>aH mxsq asmHaOH: aH ssxsq pasH .ammno .m .m can manages: .m.o mm a on H nan so nwnssz .mm nm>Hm aoHus>nsmaoo sHonmsooa I H ImaHsoaso mo uassunsoso asmHaOHz m mmsooa 02 I 0 was maHusom .Hm nmsHmumuH sN m>ons so mHnanm> mam uHsao mcHamHa .om mH ussm mm ou oH aust ns>Hm .mv aOHus>nsmaoo mo uasEunsoso asmHaOHz .aoHuosm aOHussnosm can «name .ummum mchcmHa sau an psUH>ono mxnso mo mamHa nmummz eons smumeHumm - Hams ooo.maH on emu omsucona no>Hm .ma suso mo sonsom o \ .ns> sHsom anans> aqu nxnso no msaHs> no mo .02 smasm anHmsoo .Uuaooll.mH mamas 245 aoHus> Insmaoo mo uasEunsoso asmH anon: .noHuumm menus .umsum maHaasHm mau mo mnsQEsE mau mo soasumHmms sau auH3 sousEHuss snss mxnso mH maHaHsEsn sau now msnsu Isnsossa .momH .nuasoasu IaHnsoam xnso mo munoosn thoos Bonn osaHsuoo onus mxnso HH nan nsnausnsofisa nuHmnm>Hco ousum asmHEUHz . uamaongso sunaossm mo uamEunsoso .manaoeam .m.o mo monoosm ow .m can ou can AN no H s hpSOHU I m aqu menus m. snug: I H vs nssHU I o aOHussnosm nan sHosuosoosao was psuaHHoo I N smsns>a ann .mmnuq sasHaHIonsu Isnsofisa nousz .om nsxsq oasHaH «0 nuHannse .mm manna scuch IcoHusHHoa .vm nsxsq oasHaH mo omssnud .mm MHMQ HO OUHHHOW AN no H s chsuosooa nqu mnnma m. nan smusHHom I H vs noaHz no oaoz I c an «onus omo.sH on an o m .nm> mHmom aqu mxnso no usaHs> no mo .02 smasm anHmmoo oHannm> ovuaOU|l.mH mnm¢9 246 suasumHo nw ussm com ou mm aossm usz .No nsxsq oasHaH ~s>ons IaOHuHsooEou ow ov sHoans> mom. aosom us3 .He .aOHus>nsmaou mo uasaunsomo msxsq asmHEOHz .aOHuosm aOHussnosm Hs>oos oasHaHIaOHuHs can mnnma .msmum ocnccmHm mm as mHnanu> mom. uoasoo scams .om nsxsq pasHaH mm ussm cow ou mH Iaust aosom .mm msxsq sasHaH ao .m>onm mm mHnaHnu> ammo an noon oom.¢ on ocH mmsuconm nouns .mm ..s>oos om sHoans> mom. .mxnso no nos: asHm nouns: .aOHus>nsm msxsq oasHaH Iaoo no uassunsoso asmHaOHz mm uwso ooo.mH ou om¢.H ao smsuaono .hm suso mo sonaom o x .ns> sHsom anans> auHs mxnso no ussHs> no mo .02 smasm anHmmoo ovuaOUll.mH mam¢9 247 aOHus>nsmaou mo uasEunsoso asmHaon .aOHu uaonsooa Iosm aoHusonosm was mxnso mo aOHummmaou I H aoHummmaou mmsum maHaasHo oau mo uasEmpao m aoHumsmaOU oz I o maHaoaasH usom .mw msxsq uasmsno I H pasHaHIHsuasm mH uaswna I o soaso no usom .hm msxsq uasmsno I H pasHaHImoEsm he uasmoa I o aoaasq usom .mm Hs>oo< msxsq pasHaH on ms oHomHns> mom. usnssoouHH .mo soxsq uasmsno I H sasHaHIomsom mH uasmna I o aossm no ausm .vw aOHus>nsmaou maHEEHSm I H wsxsq so unassumomo cmmHnon mm ocHseHzm oz I o sauHaHumoHasHsm .ms suso mo sonsom o x .ns> sHsom sHaans> nuns mxnso no msaHs> no mo .02 smasm anHmsoo .fluaoollomH mamca 248 F... Hui i. I; . IHIN ill“ Esau—fl u Ho>oos we «HomHnm> some aOHus> Inssaou mo uaseunso loo asmHaOHz .aOHu loom aoHussnosm pas manna mo mason man: -asHo ago so nauseous noxsq Ho>oo< scmHanunuH an we oHnanm> some uHuno mcHsuHm .ms muoHHu Iaoo pas aoHu Issmaou nansz aoaeou uoa MHHsnsaso msHuH>HuoaVnsusz ou sao mEanono mH was aoHummmaou maHme nouns mm ocHme nouns oz moxsq sasHaH noauo ou sHoaasaU no soscsoxm aoHusmH>sz saoo ss3 maHaoaasH snags sxsq cu m souHsHH maHusom msxsq pasHaH IaOHunmmaou o unoom nouns .Hs o mannnm nouns .os msxsq scsHaHumnousz o maHusom sHosmesz .mo suso mo sonaom o \ .ns> sHsom anans> auHs mxnso no usaHs> no mo .02 mmasm anHmnoo .QHGOUIlowH mqmdfi 249 fin-III 200m OM aoH.m mmm.H pom 5H s snomzano www.mH ssm.s~ ohm.v oH .amzm. couaHHo s-.4 Hmn.H mom mH oano con.~ mmm.m ~ma «H oumnooono smm.~ HHv.m «mm «H xHo>mHnsso «so.» aes.o «Hm.H NH ammo mom.mm mms.m¢ son.~H HH noonHuu moo.» mHH.s mNa.H oH gunman OHm.m~ eHa.HH mnm.HH m :oHnnom ohs.~ mmo.~ Hmm m smouHmoH a Hucom mam.m¢ soH.Hm has.» a .amzm. sum was.HH ~om.m mam.H s nouns H-.H maH.H mam m canons ssH.m nmm.~ mas o commuo a anuna aoasnofi co¢.m Hns.m sum.~ m Iago: a scooHa som.m~ use.HH maH.n m camoHHa ems.H ssH.H «Hm H usoomo a moooHs «smH announoosmo «mmH nsmH nooeoz guano nocsoo mhsolnsosso mo aOHmH>sm mhsolnoofisu mnsossu opoz no ousumuuso nuasoo Hugues mooumsHomm .mwsplnooaso vomH Hsauos pas .NmmH .nmsplnsoaso aHmHno pousSHusMII.mH mamas 250 Le mamawH a a mmo.mH MMM.MH mmm.H mm aoumoaH>HH «pH.hH . Hoomm mm osssaoq cmm.m~m mom.noH mam. mm sHooomo a ost Hem.Hm mm . scm.m~ mm .amzm. anon msm.~m «.me cmc.mH Hm Hausa. oousesHsn mmm.mH mmm.oa mms.HH om .mmzmc cannons mmH.e mmH m mum m an sHHonsmH mmo.m mm . mum mm anemoo a canon m mH mmH.oH . «Hm.mOH . smh.m mm nHaoH vmm.c mmH.mm mmm.MH om HaMva asamaH mmo.o m~.m m~¢.H mm aonam mmm. mmm m Hmm H mm onsmHHHm «H Hmo.m . MOW N MN UOHUQHU Hem.OH mxmmnHmn vhm.m mm¢u0H mmm.~ «N a smns>sna sasnw ohm. Hmm H mmm HN ansssHo OHm mmm.mmH . mmm.H . msH.m~ om Assam. consume mH~.mm ~mm.m «mo.H mH ocean smm Hm mHm m mH Hausa. canon mmmH unmounmosmo mom H mwsolnsofiso mo aOHmH>om masolnsoEso mnwmmH nonesz osonu auaaou o Eso spoz no usumluao huanoo Hsauoa pousEHusu .suaoouu.mH MHmas 251 DUI-I mom.m «HH.m ~HH.H mm omHHcmm mmo.m Hm~.oH mms.m mm common .nm mmm.om mmm.mm mmo.m mm nHmHo .um mmm.mm mmm.om smm.mH Hm .Hmzmc zmchum mmm.m omH.m mmm om onH summons mmm.mm omm.mm cmH.s mm msmuuo mos.m mac.~ mam mm mouooo OHo.mmm mHH.Hmm 4mm.mm mm Assam. scanmo Hmm.m ooo.m Hmo.H um omnmsoz mmm.mm mmm.ms omm.HH mm Assam. comoxmsz mmm.cH mph.» «mm.H mm sHuooaoz mso.mm mmm.°m omm.m mm mango: UHOHUHOQ— mom.m mmm.m mom.H mm a monsmmmmz mmm.mm mmH.mH mmo.m Hm oamHst Hmm.m ~om.m omm.H om mumoooz mm..m mHm.m mmm.H mm comm: Hso.m mmm.m omm.H mm moumHasz mmm.Hom mmm.~sH mmm.mm mm Assam. oaoomz mmmH mnmounooemo mmmH mmmH noosaz moons nucsoo unsounsoesu mo aonH>sm masoInsoesU mnsoeso oooz no ousumuuso nuosoo Hsauod ususfiHumm .quOUlI.mH Manda 252 nHoaHHHH OHm.m Hoam.mc mm~.ma Hmm.OH so .aauanauamHmasaao Hamzmc mmm.m Hamm.ac mom.mmH amm.mm as mHaaHHHH .aHnaam Assam. mHm.nm HmHm.m~Hc mam.m~m.~ mmH.mmm ma mHaaHHHH .aaaaHno Hamzmc mmH.m Hams.mc mHm.mm omm.mH «a aHaaHHHH .snannaam Hamzm. has.H .Hmm c mmm.om mmm.m mo .amHz .mamaaam Assamc mmm.m HemH.H. Hmm.oe ch.HH mm .amnz .aaHaam . Hamzmc nms.oH Imam.m. oso.mam mso.as Ha .amHz .aanaasHH: Ham.m .mmm.H. mmm.mOH mam.sH as .amHz .aamnsaz aHm.ama amm.mao.H mmm.mmH mm .amzmc mans: vmm.mm mam.om mo~.mH mm .smzm. anaauammz mam.m mmm.m mmm.m mm aanamaa> mmm.mH mmm.m mmm.H am aHaaaan amm.mm mam.mH msm.m mm aaamazaHam mamH unmaanaoEIo mamH mamH naaaaz aaano noaaao mmsolnsoeso mo aonH>sm mmsolnsoesu onsoEso 0002 no oumumnuao nuaaao Hssuoa osusEHumm .puaOUII.mH wands 253 omm.m mmm.m mam.m om naaaao :aaaHxaHa smm.m mmm.OH «Hm.m ms nuaaao auHaa Oflflwxuflz ama.m mmo.qH mam.m me a aaaH .azooanao ems mmm mmm em nuaaao saunas mom mam.m has as nuaaao namHa Hamzmc mmm.eH Hmmm.mmc www.msm mam.OHH amp aHna .saaHa>aHo mmm.m .Hmo.mc mms.mm mem.m ms .amzmc aHaa .aeHH 3%sz mmH.sm .Hmm.mH. mma.mmH mOH.om mm aHaa .aaaHaa 29$ emm.m .mmm.mc mHm.mm Hmm.m He .san .aHocaz Hamzmc .san Hmm.om .mmm.m. ama.aHH mmH.mH om .aanae anon Hanan. .san smH.mH .mmm.mc mom.aoH mmm.mH ma .aaam suaam .san mam.m Hmmm.OHc mom.mm~ mmm.mm mm .nnmousaosaam HamH unmannaasmo mamH mmmH naasaz aaano naaaao masonnsofisu mo aonH>sm mwsolnsoaso unsoaso sooz no ausumuuao nuaaao Hsauud UsusEHumm .vuaOUIl.mH Hausa 254 . lb" ..,\ for a. .a.«...." Inn‘s.:u 31......) EN mmo.Hmm.m man.omm.m mm».omm.m mHm.m~m aaHaaaao aaoHnaHz Savanna—am Emma. maH.a .moo.H. mom.mm mme.m aom .aaHz .mam guano Haozm. auoa oaH.m .HHm . omm.mHH mho.mH mm naaanz .auaHaa Hmm mmo.m mHm mm muaaao unanaHaanam omo.H ~m~.H mmm mm huaaou aomsaouao mmo.m moo.o oms.H mm suaaao aacHaaaaz smm.m omm.mH omm.m mm naaaao auaaaonaz moo.H mmH.m mom mm naaaao aanH .s.oo 3saoo3ox mmm.~ H~H.m mmH.~ mm a aauaoaas soo.H mHm.m mam.m Hm maaaao anaaoao momH unmaunaaeao momH momH naasaz oaano muaaao masolnsoeso no aOHmH>sm mhsoInsoEsu snsofiso 0002 no ousumuuao nuaaao Hssuoa psusEHumm ovunooliomH Handy 255 .oomasao uoa onos mnonaaa aHmHno one .onmHsas oau Eonm commons nousH uoa pooaHoaH hHHsaHmHno onot ..ooH3 .noHnoosm pas .oHao .aanoq .o..a.m.z.m ousunIMOIuao o3so .ao>nam aOHusonoom HsaOHusz oau Bonn sousn aOHusoHOHunso pas aOHusHaooo oomH ao soosn ons sousEHumms smm.omH.m mos.mHH.m mmm.mHa.m mmm.oom.H Hanan mmm.mHm Hmm.mmm amo.mmH.m mHH.mmm m..«.m.z.m aaauouoaa momH unmannoosmo momH momH noaeaz oaano nuance mmsoInooEsU mo aOHmH>om masolnooaso snooaso opoz no ousuonuao nuasoo Hsauoa oousEHuom owuaOUII.mH mamas BIBLIOGRAPHY 256 IlIII‘lllrlnlll'III-lll )1! ll Ilcll I II II I lil Allll II ll .’ I i | .ll BIBLIOGRAPHY Public Documents Michigan. Department of Conservation, Parks and Recreation Division. Summa of Ca in Infor- mation, 1962. Lansing, I969. (MimeograpEeH.) Michigan. House of Representatives. Report of Committee on State Parks and Public Lan s. Lansing, 1962. Michigan. State Highway Department. Ori in Surve of Cam. rs at Michigan State Parks, I962. Lansing, I969. (Mimeographed.) Ohio. Department of Natural Resources. Division of Parks and Recreation. Ohio State Parks--Travel and Use Survey. Columbus, I969. U. S. Department of Agriculture, Forest Service. Wbrk Plan for the National Forest Recreation Survey-A Review 0? the Outdoor Recreation?- sources of the National Forests.Ithhington, U. S. Bureau of the Census. Eighteenth Census of the United States: 1960. PopuIation, Vol. I, Part 24 (Michigan). Washington: U. 8. Government Printing Office, 1961. U. S. Department of Commerce, Bureau of Public Roads, Office of Planning. Calibrating and Testing A Gravit Model with a SmaII Com uter. WEShington: U. . Government Printing Office, October, 1963. U. S. Department of Commerce, Weather Bureau. Cli- matological Data, Michi an, vol. LXXIIIV, No .7. waShington, JuIy, I963. U. S. Department of the Interior, Bureau of Land Management. Recreation Site Inventor and Eval- uation. (Form 4- -1644), IWashington, June, I963. 257 258 U. S. Department of the Interior, Bureau of Outdoor Recreation. Inventori, Clagsification and Eval- uation of Existinc Outdoor Recreational Areas and Facilities. (Form.B.O.R. - . as ing- ton, . ’ U. S. Outdoor Recreation Resources Review Commission. The Qualit of Outdoor Recreation: As Evidenced by User Satis action. (StudyReportINo. 5.Y’ Washington: U. 5. Government Printing Office, 1962. . National Recreationi§urvey. (Study Report No. 19.) washington: U. S. Government Printing Office, 1962. . Partici ation in Outdoor Recreation: Factors Affecting Demand Among American Aafilts. (Stfidy Report No. 26.) Washington:U. S} GOvern- ment Printing Office, 1962. . The Future of Outdoor Recreation in Metro- pglitan Region§_pf {He UnitEd States. (Stfiay Re- port 1, Vol. III.) Washington: U. 8. Government Printing Office, 1962. . Trends in American Livin and Outdoor Recreation. {Stuay ReporE No. 2;.) WasfiingEon: U. 9. Government Printing Office, 1962. . Economic Studies_in Outdoor Recreation. (9Eu3y Report No. 24.) washington: U. S. Gavern- ment Printing Office, 1962. washington. State Parks and Recreation Commission. We Come to Cam. in Washinggon State Parks--Overnigh€‘ Campigggurvey, 1956. Iympia, I956. Wennergren, E. Boyd. Value of water for BoatingRecrea- tion. (Bulletin 453.) Logan: Agricfilture Experi- ment Station, Utah State University, 1965. Wisconsin. Department of Resource DevelOpment. Recrea- tion in Wisconsin. Madison, 1962. Books Blalock, Hubert M. Social Statistics. New York: McGraw-Hill Boo Co., . Brockman, C. Frank. Recreational Use of Wild Lands. New York: McGrawzfiIII_§66k Co;T”I9§9. 259 Bunge, William. Theoretical Geo a h . (Lund Studies in Geography, Ser. C. GeneraI anfi Mathematical Geography No. 1) Lund: Sweden, C. W. K. Gleerup, Publisher, 1962. Clawson, Marion. Land and Water for Recreation. Chicago: Rand McNally Co., 1963. Garrison, William L., Berry, Brian J. L., Marble, Duane F., Nystuen, John D., and Morrill, Richard L. Studies in Highway DevelOpment and Geographic Ohénge. Seattle: University ofWashihgton Press, Goode, William.J., and Hatt, Paul K. Methods of Social Research. New York: McGraw-Hill BooE 50., 1957} Gunn, Clare A. AConcept for the Design of a Tourismr Recreation Region. Mason, Michigan: Tfie E. 3. Press,1965. Hagood, Margaret J., and Price, Daniel 0. Statistics for Sociologists. New York: Henry Holt and Co., I952. Harman, H. H. Modern Factor Analysis. Chicago: Univer- sity of Ch cago Press, . Isard, Walter. Methods ofRegional Anal sis, An Intro- duction to RegionaIScience. New YorE: JOKE WiIey and Sons, Inc., 1960. James, Preston 3., and Jones, Clarence F. American Geogpaphy: Inventory_and Pros ct. Syracuse: Syracuse University Press, I . Massarik, Fred and Ratoosh, Philburn (eds.). Mathe- matical E lorations in Behavioral Science. Home- woad, IIIinois: RicHard D. Irwin, Inc., I965. Riley, Matilda White, Riley, John W., and Toby, Jackson. §gciological Studieg_in Scale Analysis. New BrunswicE} Rutgers University Press, I954. Torgerson, Warren S. Theory and Methods of Scaling, New YOrk: John WiIéy and Sons, I958. Warntz, William. Toward a Geo aphy of Pgice. Phila- delphia: University of Pennsylvania Press, 1955. Zimmermann, Erich W. WOrld Resourges and Industries. New York: Harper 8 Brothers, Publishers, 195I. l ‘I i l. 260 Articles Ackerman, Edward A. "Where is a Research Frontier?“ Annals of the Association of American Geographers, VbIo LII 11563,, 129-1160 I Brewster, S. F. “Park Appeal,” Planning and Civic Comment (October-December, I937). Carrothers, Gerald A. P. 'An Historical Review of the Gravity and Potential Concepts of Human Interaction,” Journal of the American Institute of Planners, XXII (Spring, 1* , - . Clawson, Marion, and Knetsch, Jack L. Outdoor Recrea- tion Research: Some Cgpcepts and Su estediAreas find , Washington: Resources for %'ie Future, Inc., Reprint Series, No. 43 (Nevember, 1963). Crevo, Charles C. "Characteristics of Summer Weekend Recreational Travel,” Hi hwa Research Record, No. 44, Highway Researcfi Boa¥a Publicatibn I161 (1963), 51-60. Dunn, Edgar S. ”The Market Potential Concept and the Analysis of Location," Papers and Proceedin s of the Regional Science Associatibn, II (I956), IFI=I94. Ellis, Jack B. and Van Doren, Carlton S. 'A Comparative Evaluation of Gravity and System Theory Models for Statewide Recreational Traffic Flows,“ Journal of Regional Science, XI, No. 2 (Winter, 1966), 57-79. Evans, John S., and Van Doren, Carlton S. "A Measure- ment of the Demand for Recreational Facilities at Lewis and Clark Lake," South Dakota Business Review Su lement. Vermillionzischool ofiBusiness, State UnIversiEy of South Dakota (February, 1960). Hansen, walter G. "How Accessibility Shapes Land Use,” Journal of the American Institute of Planners, XXV, “a; 2 ("hy'Igsgi' 67-66. Harris, Chauncy D. "The Market as a Factor in the Local- ization of Industry in the U. S.,' Annals of the Association of American Geographers, XCIV (December, ' O Huff, David L. 'A Probabilistic Analysis of Shapping Center Trade Areas," Land Economics, Vol. XXXIX (February, 1963), sl-yua 261 Huff, David L. 'A.Topographical Model of Consumer Space Preferences," Papers and Proceedin s of the Re ional Science Association, VI (I960), I59-I73. Lucas, Robert C. "Wilderness Perception and Use: the Example of the Boundary Waters Canoe Area," Natural Resources Journal, Vbl. III, No. 3 (Jan- uary, I963), 391-411. Lynch, John T., Brokke, Glenn E., Voorhees, Alan M., and Schneider, Morton. "Panel Discussion on Inter- Area Travel Formulas,“ Traffic Ori in-and-Destina- tion Studies, Highway ResearchfiaifaffiuIIetin 253 p -1380 Murphy, Richard E. “Geography and Outdoor Recreation: An Opportunity and An Obligation,” The Professional Geographer, val. XV, No. 5 (1963), 37:53} Stewart, John Q. “An Inverse Distance variation for Certain Social Influences,“ Science, Vol. XCII (1941), 89-90. Ullman, Edward L. and volk, Donald J. “An Operational Model for Predicting Reservoir Attendance and Benefits: Implications of a Location Approach to Water Recreation,” Pa ers of the Michi an Acade of Science, Arts, a tiers, VbI. fiaII (I962), 175:154. voorhees, Alan M., Sharpe, Gordon B., and Stegmeier, J. T. Sho in Habits andTEavel Patterns, Special Report II-fi, Washington: Highway Research Board (1955). Wblfe, R. I. “Perspective on Outdoor Recreation--A Bibliographical Survey,“ a hical Review, Vbl. LIV, No. 2 (April, 196;), 593-235. Reports and Proceedings Baker, W. M. ”Assessing and Allocating Renewable Re- sources for Recreation,” Resources for Tomorrow: A Report to the Department of Northern Affairs and NationaI Resources, Monfreal, OctOber 59-29, I96I. Booth, Alfred W. The Lakes of the NOrtheastern Inland Empire, A Stuiy_of Recreational Sites. (Eireau of EEonomics and Research, BulIetin No. 5) Pullman: State College of Washington, 1948. 262 Dahle, Thomas L. Michigan State Park Users Survey, 1956. (Bureau of Business Research, College of‘fiusiness and Public Service, Research Report No. 19) East Lansing: Michigan State University, 1956. Great Lakes Research Institute. Currents and Water Masses of Lake Huron. (TechnicaI Paper No. 1) Ann Arbor: University of Michigan, July, 1954. . Currentsj and Water Masses of Lake Michigm (P EBIication No. )3 Ann Arbor: University of Michigan, 1958. Humphrys, C. R. Shoretype Classifications of Michigan Counties. (Bulletins 1-29) East Lansing: Depart- ment of Resource Development, Michigan State Univ- ersity, 1958. Humphrys, C. R. and Green, R. F. Michigan Lake Inventory. (Bulletins 1-83, By County) East Lansing: Department of Resource Development, Michigan State University, 1962. Hutchins, H. Clifton and Trecker, Edgar W., Jr. Th__e_ State Park Visitor--A Report of the Wisconsin Park and Forest Travel Study. (Technical Bulletin No. 22)’ Madison: Wisconsin Conservation Department, 1961. Institute for Community Development and Services. The PrOposed Pictured Rocks National Lakeshore,L An Economic Study. Prepared for the National Park Service. East Lansing: Michigan State University, 1963. . Report on the Economic Feasibility of the Proposed Sleeping Bear National Seashore. Prepared fSr the National ParEEService. East Lansing: Michigan State University, 1961. Lewis, Philip H., Jr. Recreation and Open Space in Illinois. (Division of Landscape Architecture and Bureau of Community Planning) Urbana: University of Illinois, 1961. National Advisory Council on Regional Recreational Planning. A User--Resource Recreation Planning. Method. Hiaaen vaIIey, Laamis, 5aIifornia,l 9. Reid, Leslie M. Outdoor Recreation Preferences-HA Nationwide S tidy of User Desires. East Lansing: Michigan State University,I9697' 263 Thomas, Edwin N. Maps of Residuals from Regression: Their Characterisiics and Uses in Geogrephicfie- search. Iowa City: Department of_§eography,_§fate University of Iowa, 1956. Ullman, Edward L., Boyce, Ronald R., and VOlk, Donald J. Recreation;fThe Meramec Basin. (Vol. III, Water Needs and Prohlemg) St. Louis, Missouri: WasHing- ton University, 1961. Voorhees, Alan M. ”A General Theory of Traffic Move- ment,” 1955 Proceedin s of the Institute of Traffic Engineers.(0ctober, 1955), 36-50. Unpublished Materials Campbell, Robert D. ”The Geography of Recreation in the United States.“ Unpublished Ph.D. dissertation, Department of Geography, Clark University, 1949. Cavanaugh, Joseph A. "Formulation, Analysis and Testing the Interactance Hypothesis.” Unpublished Ph.D. dissertation, Department of Sociology, University of Washington, 1950. Ellis, Jack B. ”The Description and Analysis of Socio- Economic Systems by Physical Systems Techniques.“ Unpublished Ph.D. dissertation, Department of Electrical Engineering, Michigan State University, 1965. Gilbert, Alphonse H. "A Survey of vacation Camping in Iron County, Michigan.“ Unpublished Master's thesis, Department of Resource Development, Michigan State University, 1963. Nystuen, John D. ”Geographical Analyses of Customer Movements and Retail Business Locations: (1) Theor- ies (2) Empirical Patterns in Cedar Rapids, Iowa, and (3) A Simulation Model of Mbvement." Unpub- lished Ph.D. dissertation, Department of Geography, university of Washington, 1959. Van Doren, C. S. ”Recreational Boating in Michigan." A paper presented at the Michigan Academy of Science, Arts and Letters, Wayne State University, Detroit, Michigan, March, 1961. 264 Other Sources Computer Institute for Social Science Research. Factor Anal sis Programs: Pep d gand Famin 3. TecEfiicaI Report 2 (Revised), East Lansing: Michigan State University, September, 1964. Drury, Newton 8. California Chief of Beaches and Parks and Former Director of the National Park Service. Speech delivered at the National Con- ference on State Parks, NOvember, 1957. Rates, R. "The Pursuit of Beauty." A paper prepared for delivery to a symposium of the National Resources Institute at The Ohio State University, Columbus, May 24, 1966. Maps Hammond, E. H. ”Classes of Land-Surface Form in the Forty-Eight States.“ Annals of the Associa- tion of Americgngeogra hers, Map SuppIement Numberl Vol. LIV, No. 7March, 1964). Veatch, J. 0. Map of Resettlement Forestin Michigan. East Lansing: Department of Resource DevelOpment, Michigan State University, 1959. HIGRN STQTE UNIV.