LI B RA R Y UfliVCl’Sit)’ r This is to certify that the thesis entitled Simulating Suburban Development in Suffolk County, New York: 1950 through 1990 presented by Ralph J. De Vitto has been accepted towards fulfillment of the requirements for M .A. degree in Geoqraphy SLW Ox WW. Major p oéessor QW k 0-7 639 SIMULATING SUBURBAN DEVELOPMENT FROM 1950-1990 IN SUFFOLK COUNTY, NEW YORK By Ralph Joseph De Vitto A THESIS Submitted to Michigan State University in partial fulfillment of the requirements for the degree of MASTER OF ARTS Department of Geography 1978 ix (XX) ABSTRACT SIMULATING SUBURBAN DEVELOPMENT FROM 1950-1990 IN SUFFOLK COUNTY, NEW YORK By Ralph Joseph De Vitto Suburban sectors within major metropolitan regions of the United States have been growing more rapidly than their urban cores in both population and area since werld War Two. -The population dynamics of the New York Metropolitan Region are analyzed from.l940-l970. Through time urbanization is found to be dependent primarily upon distance from the New York City CBD and the areas of highest population den- sity. Both past and present population trends in Suffolk County, a lowedensity outer suburban county, are examined in detail. Several factors including adjacency to existing development and distance from.urbanized Nassau County are found to be significant. Two Monte Carlo simulation runs are constructed to measure the spatial patterns of develop- ment across Suffolk County from 1949-1969. A third simu- lation is also presented depicting probable development situations in the county's semi-rural eastern townships through 1990. ii To My Parents For Their Encouragement and Support ACKNOWLEDGMENTS I gratefully acknowledge the conscientious advisement received throughout my program from.Professors Stanley Brunn and Deiter Brunnschweiler of the Geography Depart- ‘ment at Michigan State University. Their guidance, assistance and academic stimulation have helped me to realize an academic goal of which I am very proud. I also wish to thank the Department of Geography for the financial assistance awarded during most of my attendance. Gratitude is extended to Sandy Briggs for a typing job well done, and since nothing is possible without the inspiration of friends and loved ones, to my sister and Anne, thank you. iii CHAPTER II. III. IV. V. APPENDIX TABLE OF CONTENTS INTRODUCTION . THE NEW YORK METROPOLITAN REGION . SUBURBANIZATION IN SUFFOLK COUNTY THE SUFFOLK COUNTY SIMULATION MODEL OF SUBURBAN DEVELOPMENT SUMMARY AND RECOMMENDATIONS BIBLIOGRAPHY iv Page 21 46 63 68 70 Table LIST OF TABLES NYMR Population Summary Table, County Population Change: 1940-1970 . Township Land Use and Population Density Township Total Population and Percent Change: 1950-1977 Page 13 34 37 LIST OF FIGURES Figure Page 1. U.S. Urban Population Growth by Census and Urban Regions, 1900-2000 . . . . . . . . . . 3 2. Increase of U.S. Urban Lands, 1920-2000 . . . . 3 3. The New York Metropolitan Region . . . . . . . . ll 4. NYMR: Percentage Population Change, 1940- 1950 . . . . . . . . . . . . . . . . . . 16 5. NYMR: Percentage Population Change, . 1950-1960 . . . . . . . . . . . . . . . . . . . l6 6. NYMR: Percentage Population Change, 1960-1970 . . . . . . . . . . . . . . . . . . . l6 7. NYMR: Relationship between 1950-1960 Population Change and the Distance from the NYC CBD . . . . . . . . . . . . . . . . . . . . l9 8. Suburban Long Island: Population Growth and Projections, 1940-1995 . . . . . . . . . . . . . 24 9. An Exposure of the Ronkonkoma moraine (a) in western Suffolk County partly depleted by intensive sand-gravel mining (b), Long Island's major extractive industry . . . . . . . . . . . 26 10. An Outwash Plain in Brookhaven Township, the Predominant landform in eastern Suffolk County . . . . . . . . . . . . . . . . . 28 ll. Remnant Agricultural Fields in western Suffolk County Fringed by (a) Residential and (b) Industrial Lands . . . . . . . 30 vi Figure 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. Suffolk County: 1970 Population Density and Distance from Nassau County . . Suffolk County. 1966 Land Use and Distance from.Nassau County . . . Suffolk County: Population Growth Rates by Zone, 1940-1977 . . . . . . . . . . . . . Outer Zone: Percentage Annual Population Change, 1960-1970 . . . . Outer Zone: Percentage Annual POpulation Change, 1970-1977 . . . . Suffolk County: Population Gravity Centers, 1960-1977 (by MED) . . . . . . . . . . Suffolk County: Actual Development, 1949 and 1969 . . . . . . . . . . . . . . . Suffolk County: Actual Development, 1949 and 1969 Generalized to Simulation Grid . Suffolk County. Random Simulation of Development, 1949-1969 . . Suffolk County: Controlled Simulation of Development, 1949-1969 . . . . . Outer Zone: Simulation of Development, 1970-1990 . . . . . . . . vii Page 33 33 39 41 41 43 56 56 61 61 61 CHAPTER I INTRODUCTION Urbanization is the dominant theme of twentieth century population dynamics in both the United States and the entire world. Geographer Clyde Browning (1974:1) has stated that, "urban growth is one of the central facts of the twentieth century,‘ and considering recent urban population growth figures recorded in the United States, this statement is well qualified. The nation's urban population, for instance, increased by 66 million persons from.1900-l950 (Gottmann, 1961:21) a value roughly comparable to three times the estimated 1976 population of Canada. Rugg, in Spatial Foundations of Urbanism (1972), estimates that one-third of the entire world's population lives in cities. In describing the urban population growth characteristics of the United States, Rugg notes that the rural population was first exceeded by the urban population as early as 1920, and "in the half-century since then the proportion of urban papulation to total has become three- fourths, having grown from.two-thirds in only one decade, the 1960's" (Rugg, l972:xii). In a later chapter Rugg conclusively asserts, "we truly live in an age of urban 1 2 explosion" (Rugs. 1972:72). Yeates and Garner have also acknowledged the vola- tility of contemporary urban growth in their text, The North American City (1971). Concerning the topic of in- creasing urbanization they state: Our first assumption, therefore, with respect to North America is that the urban explosion over the next few decades will be of dra- matic proportions (Yeates and Garner, 1971:472). These viewpoints of Browning, Rugg, Yeates and Garner suggest that urbanization has been and will continue to be a dominant feature of contemporary papulation dynamics. More specifically, the increase of urban pepulations, and the subsequent expansion of built-up areas in the United States, identify a significantly critical problem worthy of scientific research (Figures 1 and 2). Urbanization in the United States By far the most significant changes in the urban and rural proportions of the United States population have taken place since 1950. Approximately 74 percent of the nation's total population was classified as urban by the Bureau of the Census in 1970. When campared to the urban population percentage recorded for 1950 and 1960, the overall 15 percent increase between 1950-1970 indicates Population Total of °/o FIGURE 1 '00, us. URBAN POPULATION GROWTH . BY CENSUS AND URBAN REGION .- ’2‘ — ’ ISOO- 2000 a) O U h- 01 O ’ ‘h -‘ " ’ I 409”’ Urban population a by: . -- census region d 20 —— urban region P 0 l l 1 J I L l l i I900 I920 I940 I9 60 I980 2000 ‘ Source: Yeates a Garner, I97I , p. 30. Relily,l973, p.84. FIGURE 2 INCREASE OF U.S. URBAN LANDS ISZO-ZOOO AIDA/URBAN .. -—1920 ,, “PF—I940 -—|960 ISBO 2000 URBAN LAND Shown as percentage of fatal U.S'. land area/2.9m/l/ianmi‘} Source= Reilly, I973, p.84. 4 an unprecedented shift in the nation's basic population structure. Simple percentage increase figures, however, fail to identify the various dimensions of recent urbanization processes. Suburban areas, for instance, have been grow- ing more rapidly than urban areas since world war Two in both papulation and area. Since that time, the percentage of the nation's total population residing outside of these urban areas has been increasing. Correspondingly, the percentage of the total population located in nonmetro- politan areas (areas away frmm cities and their associated suburban regions) decreased from 38 percent in 1950 to 31 percent in 1970 (Reilly, 1973:85). And during the same two decades, a 10 percent increase in persons residing outside of central cities, from 27 percent to 37 percent, was matched by a 6 percent decrease in persons residing outside of nonmetropolitan areas. Increases in suburban populations have therefore accounted for the major portion of urban growth between 1950-1970. During the 19703, a new demographic shift is evident as nonmetropolitan populations have increased both abso- lutely and relatively. Roseman recently noted, "Now there is a tendency toward movements down the size hierarchy of urban centers and movements from urban areas to rural areas..." (Roseman, l977:iv). These more recent patterns and the reasons behind them, however, are beyond the scope of this study. Numerous factors have been identified which account for the rapid increase in suburbanization since World War Two. These factors affecting the growth of suburban pap- ulations and the areal expansion of suburban lands to meet residential, commercial and industrial needs are varied and complex. The major stimuli have been cited in the report compiled by the Task FOrce on Land Use and Urban Growth (1973). More people, with more money, following their expressed preference for low density living in a metropolitan area can mean only one thing--bigger metro- politan areas. If past trends continue, 36 million of the 54 million population increase expected by the year 2000 will live in suburbs, more than moved there between 1950 and 1970 (Reilly, 1973:82). In essence, many of the metrpolitan regions in the United States have experienced significant population increases and associated expansion in outer-lying suburban areas. These areas will also continue to accommodate a major portion of future population growth. The processes of suburbanization will transform.existing suburban lands into urban areas and existing rural lands into suburban areas. Geographers and other social scientists have identi- fied this "explosive" nature of post World war Two 6 suburbanization as urban sprawl. The term.itself as de- fined by Northam indicates more precisely the pace at which the conversion of land from.nonurban to urban uses is presently occurring rather than the mere physical expansion of urban areas, a process that has been taking place for several thousand years (Northam, 1975:378). As urban sprawl occurs, it gives rise to a number of problems that suggest the use of geographic research for their solutions. One such problem is thatcontemporary suburbia, an out- growth of urban sprawl, is unique in terms of its rate of development and internal patterns and processes. Only recently, however, have geographers and others attempted to investigate suburbia as a subsystem functioning both independently and interdependently within a larger urban system. Muller‘s recent publication, The Outer City (1976), is perhaps the most current and rigorous investigation into the modern suburban spatial system» Although the paper is a general survey of "only the major dimensions of suburban social and economic spatial organization," it does provide some valuable information concerning the’ diversities and dynamics of the suburban environment (Muller, l976:3). Muller begins his thesis by identifying the various suburbanization processes which have trans- formed "the tightly focused single-core urban region of the past to the widely dispersed multinodal metropolis 7 of today" (Muller, l976zl). He goes on to suggest that, "suburbia in the late 19703 is emerging as the outer city," a direct result of various intrametr0politan socioeconomic decentralization and deconcentration trends occurring since the early 19503. Indicating that both the causes and con- sequences of the "urbanization of the suburbs are multi- faceted and far reaching," Muller asserts that there is a clear "need for geographers to conduct more research on contemporary suburbia, particularly on regional variations in suburbanization processes and patterns" CMuller, l976:3). Present Objectives In an attempt to elucidate some of the major factors influencing the internal and external spatial morphology of modern suburbia, this study has three main objectives. The first, found in Chapter II, is a general descriptive analysis of the population growth and development patterns which have taken place since the end of world war Two in the largest metropolitan region in the United States: the New York-Northeastern New Jersey urbanized area. The second objective is a detailed analysis of the changing population and land use patterns occurring within one suburban county of the New York Metropolitan Region: Suffolk County. Chapter III therefore provides such an investigation into the recent trends of suburban growth in this county. The 8 third objective is found in Chapter IV and involves the formulation and application of a simple Mbnte Carlo simu- lation model to measure recent and future population and development patterns within rapidly urbanizing Suffolk County. Chapter V provides a summary of this study's findings along with some recommendations for further in- vestigations.‘ The completion of the above stated object- ives will provide a comprehensive understanding of some of the major dimensions of suburbanization in the New York area and yield greater insight into the dynamic spatial aspects of contemporary suburban processes. CHAPTER II THE NEW YORK METROPOLITAN REGION In 1959 Hoover and Vernon formally delineated the New York Metropolitan Region (NYMR) in their classic research monograph, Anatomy of a Metropolis (1959). Their region included twelve counties from New York, nine from New Jersey, and one from Connecticut. They asserted that, "No other metropolitan region in the United States approaches it in population, employment, or wealth" (Hoover and Vernon, l959:3). For instance, in 1956 an estimated 15 million peOple resided in the NYMR, account- ing for 48 percent of the total 1950 population of Gottman's Megalopolis1 in only 13 percent of its total area. The NYMR was divided into three zones based upon the level of land development found within each of its twenty-two counties in the 19508. Counties grouped into the Core, for example, demonstrated the highest percentages of developed land to total developable land, while the 1Megalopolis encompassed 54,000 square miles stretch- ing 600 miles along the northeastern seaboard from southern New Hampshire to northern Virginia. Its population grew iggm.3§)million in 1950 to 37 million in 1960 (Gottmann, 9 10 Inner Ring and Outer Ring Counties exhibited successively lower percentages (Figure 3). During the last thirty years the NYMR has experienced the most dramatic population changes nationwide. Between 1950-1960, the region increased its population by 16 per- cent, adding 2.2 million persons to its base population of nearly 14 million (Nassau-Suffolk Regional Planning Board, l968:l). The outlying suburban counties of New York City, that is, those located in the Outer Ring account- ed for the major portion of this increase. In contrast, all but one county located in the Core (Queens County) recorded population decreases during the 19503. As New Yerk City preper experienced a population decline (losing over 100,000 residents), both the Inner and Outer Rings grew by 48 and 56 percent respectively. If taken individually, the population changes evidenced by each county since 1940, identify an intraregional trend towards a decreasing urban and increasing suburban popu- lation which parallels the national trends previously dis- cussed. These changes are summarized in Table 1 and illustrated in Figures 4-6 to provide a temporal perspect- ive of the NYMR's overall papulation dynamics. From an examination of these population changes occurring between 1940-1970, several generalizations hold. Prior to 1950 the regionis population growth took place primarily to the east and west of New York City, specifically in the 11 .mnm. .:o=.o>a :26: .2; 25.5233. E .. .. .. .w. . ii ow. or om em 0.. V mun... 383$ But-\‘I L ”.55.. . 033.5. .\. Z .\ .1; .e wm_._.ZDOo ozE mambo “NJ _ _ mm:zzoo 02E «$22.: mw_._.ZDOo mmoo Q 20.0mm z<._.30n_om._.m_2 xmo» >>mz MI... m $50”: 12 suburban counties on Long Island and in northeastern New Jersey. The northernmost counties in the Lower Hudson Valley and eastern Connecticut at the same time remained relatively stable, experiencing very minimal population increases. Immediately after WOrld War Two and continuing into the 19503, however, explosive increases occurred throughout the NYMR; the most dramatic changes took place in the previously stable counties of the Lower Hudson Valley. In the ten years from.l960-l970, three of these counties (Rockland, Putnam and Dutchess) along with Suffolk County and all of New Jersey's Outer Ring counties main- tained relatively high growth rates while the majority of Inner Ring counties declined sharply. Based upon these changes, it is hypotheSized that through time intraregional growth will vary locally in response to the three following factors: 1) the size of the base populations; 2) the existing leVel of urban de- ve10pment; and 3) the distance from the region's urban center. It follows from.the first hypothesis that higher growth rates are characteristically exhibited by those' areas with small base populations since even relatively low increases in the number of inhabitants represent high percentage increases. Hence, no further discussion of this relationship is necessary. The second and third hypotheses concerning growth rates in relation to the amount of existing development and distance from.the La. 13 mmaa m.a coma m.ma mam a.mm Boa samaaz mam a.aa oma a.ea ama m.am oae camaam mma a. «ma a.a moa ~.m amm xaaam mam a.mm mam a.ma maa a.a aha meoaaoam emme ~-aa comm a.aa mama a.m~ mama meam Hosea aom ~.- aam a.m- Nam a.- Nam coamse amaa m.a mama a.ma amma m.aa mama aeamso mesa m.m mmaa m.a- amea a.a mama neoam mama o.a- mama a.a- mmam m.a mama aamcaev amaaooam amma e.a- mama «.ma- omaa a.m mama aaaom gaze amummaeaz mama m-a mama a.a- meaa m.m «cam once one maaa amewau moaa ameano maaa amamao oaaa N a a Ammammaonh hmv oaaa-oeaa Homemao soauaaaaoa menace canny anofiESm coaumaaeom mzwz .a magma l4 omm a.mm ama m.ma am «.mm ea memaxuom am a.ma mm «.ma mm m.m~ ea ameusm mam a.o~ ama m.o~ mma a.m oaa amsaao «mm m.ma mmm a.am «ma m.om mma manna: mma a.am «mm «.ma mam .m.am ama nonoaeoz «mm a.am «ma m.mm mmm m.~m mam xaaoammaz mam a.am mam m.a~ «om m.m~ maa maaaaaamm mam m.m~ maa a.mm ama m.ma ama maaaouam «awe, a.aa. .mmom elem mmaa a.am mama meam amuse aam nroa aom a.am .mmm a.a mam Hoaaaaauaaz mam a.a mom a.m~ mam a.am mam scan: ama a.aa ace m.mm amm a.a aom camaaea maaa «mango mmaa .omemao oaaa amcaso maaa N N N a.a.aeoav .a magma 15 .50 .m .mmma .Hnom one amamamm "mouaom .Nmmha H.HH .mmaoH urna mmmma a.aa mHmNH Houoa mNHH N.mo new n.H¢H emu a.am maa xaommnm mma .m.mm «ea ¢.m¢ mm H.mm as uomamaom oaaa «madam omaa amaaeo oaaa amemno maaa a. a. .a a.a.uaooV .a mamaa I PERCENTAGE POPULATION CHANGE , I940 - l950 PERCENT -zs-—o 0-25 26—50 5: -IOO IOI—ISO PERCENTAGE POPULATION CHANGE , l950- I960 PERCENT '25-0 0-25 26-50 Sl-IOO IOI -l50 NYMR PERCENTAGE POPULATION CHANGE, I960- l970 PERCENT '25-0 0-25 26-50 5| - IOO IOI-ISO l7 region's center, however, need to be more fully investi- gated. Newling's Mbdel Newling (1966) developed a mathematical model of intraurban growth suggesting that a systematic relation- ship exists between a population's rate of growth and its distance from the center of the urban area. Newling's hypothesized relationship indicates that the two variables (distance and growth rate) are positively related: as distance from the urban center increases, papulation growth rates also increase. The majority of his research that tested this model was carried out on the internal spatial structures of several cities. In his Pittsburgh study, population growth rates were analyzed for each of the city's thirty-two wards and showed a significant positive relationship outward from.the central business district (CBD). Newling's model probably has applicability in a larger dimension. To illustrate, when the conceptual basis of the model is applied to the NYMR's population growth patterns for 1950-1960 (the decade exhibiting the most dynamic changes), the hypothesized relationship can- forms with the empirical situation. Population growth rates were highest in the Outer Ring and relatively lower in the Inner Ring; the Core recorded population declines 18 (Figure 7) . Newling has also determined that although a relationship exists between a population's distance from.the urban center and its rate of growth, a third variable, papulation density (persons per square mile) must also be considered. He suggests that in cities where "topographic controls and the development of the city through the coalescence of several primary settlement nodes have produced a highly complicated spatial structure,‘ simple linear distance from.the urban center does not adequately explain the resultant patterns (Newling, 1966:220). Assuming, therefore, that the spatial structure of the NYMR is highly complex as a result of various topographic and developmental factors, county growth rates also require analysis with respect to their papulation densities. Employing Spearman's rank-order correlation technique to measure the relationship between the 1950-1960 percentage population change and the 1956 population density for each of the region's twenty-two counties, a coefficient of -.44 is obtained indicating that a moderate inverse relationship exists between the two variables; it was significant when tested at the .05 level. It is concluded that the papulation changes experienced by the counties of the NYMR have been dependent upon two primary factors: 1) distance from the New York City CBD; I960 I950— CHANGE, POPULATION °lo FIGURE 7 I50 NYMRI RELATIONSHIP I30 BETWEEN "0 l950-I960 G POPULATION m° (t CHANGE AND DISTANCE FROM THE 80 N YC CBD 60 Date from Bergman a Pot", 4O l975, p.67. 20 - 2 O 1 L r l 1 l 1 l 1 I l l L l l I 0 20 4O 60 80 MILES FROM NYC CBD 20 and 2) existing population density.2 Within the region, the counties located the greatest distances away from New York City and having the lowest population densities will experience the most significant changes in both population and development in the coming years. The gross regional patterns discussed above are the result of local variations in suburbanization processes not able to be identified by such a large-scale analysis. Therefore, in Chapter III a more detailed investigation ' of one of the NYMR's low-density, outer suburban counties, Suffolk County in particular, is presented; it provides greater insight into such intraregional processes of suburban growth. 2Population density is used here as a measure of the level of urban development from the second hypothesis concerning intraregional growth trends. CHAPTER III SUBURBANIZATION IN SUFFOLK COUNTY The Nassau-Suffolk Standard Metropolitan Statistical Area Nassau and Suffolk Counties, located on Long Island, have been among the fastest growing counties in the United States since the late 19403 according to the Nassau-Suffolk Regional Planning Board (NSRPB, l968:2). This bi-county region was the scene of: ...one of the nation's earliest, most powerful surges to suburbia...as tens of thousands of young families poured out of New York City after Werld war Two (Reilly, 1973:39). By 1970 their combined populations equaled 2.6 million persons, representing over 14 percent of the NYMR's total population and more than 1 percent of the entire country‘s. The federal government in 1972 designated Nassau and Suffolk Counties as the nation's first "all suburban" Standard Metro- politan Statistical Area (Muller, 1976:2). The Nassau- Suffolk SMSA today ranks ninth among the nation's 272 SMSAs with a population larger than that of Philadelphia, the 21 22 country's fourth largest city. In addition, this region is expected to accommodate a major portion of the NYMR's sub- urban population growth through the 19803. In a recent publication it has been projected that by 1985 over 3 million of the region's total 22 million residents will live in the bi-county area (NSRPB, 1968:1, 5). Over the last twenty years, suburbanization has con- sumed almost all of the developable lands in Nassau County which has resulted in the stabilization of growth and associated development. In fact, with the county's current pOpulation density at almost five thousand people per square mile, "growth is at a virtual standstill" (Morris, 1977:17). But suburban development continues to be vigorous in Suffolk County where the population density is only 29 percent that of Nassau's and the population growth rates have been con- sistently higher since the latter part of the 19503. According to population projections made by the NSRPB, this trend will continue throughout the remainder of this century as Suffolk County is expected to accommodate 93 percent of Long Island's suburban population increases (NSRPB, l976:6). The latest available census data published by the Long Island Lighting Company (LILCO) reveal that between January 1, 1976 and January 1, 1977, Suffolk County gained over eighteen thousand residents while its western neighbor (Nassau County) 23 gained only two thousand (LILCO, l977:2).3 Future develop- ment in this area will also be affected by the relocation of light industry fram New York City and Nassau County, a local trend now in evidence. Overall, Suffolk County is experiencing dynamic growth and thus provides a prime area for the study of contemporary suburban development (Figure 8) . Land Use Changes in Suffolk County Suffolk County occupies the eastern two-thirds of Long Island, encompassing approximately 900 square miles (2,331 square kilometers) of land as it extends 85 miles (136 kilometers) from its western boundary into the Atlantic Ocean. The county is elongated in shape with its maximum width of 20 miles (32 kilometers) occurring along the Nassau-Suffolk boundary line from.Long Island Sound to Great South Bay. In the east the county terminates in two peninsulas or forks that represent the topographic highs of the Harbor Hill moraine on the north flank and the Ronkonkoma moraine in the south. These morainal ridges are merged at the western end of the county, but diverge towards the east as the Ronkonkoma moraine crosses from the north to the south. Moderate elevations of 100-200 feet (30-60 meters) are characteristic of both moraines 3An explanation of LILCO's population estimation methods is found in the Appendix. 24 »hz:oo Dhznoo xaomuam Om m. Om m. 059 00m. Cum. ova. zozbwwomm mmzmz mnmzwo .mj \ I \ magioea. .mzofiommoma. oz< 5305 222.53.”... .52de 0204 24mm3m3m w .55on CON 00¢ 000 com coo. CON. 00¢. Goo. Goo. NI NOIlV‘II‘IdOd SONVSDOHJ. 25 (Figure 9). The remaining physiographic features of Suffolk County are the result of erosional and depositional processes occurring from.the Cretaceous period to the present. Exten- sive outwash plains and channels are common throughout, along with several isolated pre-Pleistocene plateau remnants (Fuller, 1914:23). Physiographically, the county is basic- ally homogeneous and is characterized predominantly by sand- gravel plains and hummocky ridges of glacial origin. The subtlely diverse landscapes of Suffolk County-- hills and plain3--have influenced suburban growth both directly and indirectly. The economic costs required to develop the morainal areas, for example, are restrictive and subsequently limit their potential as building sites. As Eschman and Marcus have noted: ...even though in these days man's technology allows him to move mountains, the economic costs of overcoming geologic and geomorphologic factors continue to impose directional and aerial constraints on urbanization. Thus...the basic landscape on which cities are situated continues to play an important role in modern urbani- zation (Detwyler and Marcus, 1972:27). Therefore, even though the moraines are prominent landforms only in the county's northwestern and northcentral parts, the roughness of the local terrain is a direct control affecting suburbanization. The plains exert a positive impact; they attract high-density deve10pment since obstacles (a) An exposure of the Ronkonkoma moraine(a) in western Suffolk County partly depleted by in- tensive sand-gravel mining(b) , Long Island's major extractive industry. 27 such as irregular depressions and slopes are absent (Figure 10). On these sandy plains, the construction of single family detached dwellings (the predominant housing type) and low-rise, sprawling industrial buildings is easier and less expensive than on the heavily wooded, hummocky moraines. TOpography also exerts indirect controls upon intra- county development patterns, reflecting the problems associated with changing land uses. These controls are illustrated by the processes that have converted Suffolk's agricultural lands into urban lands over the last twenty years. Agriculture has been a major economic activity in the county since its early settlement in the seventeenth century. As recently as 1950 more than 120,000 acres were farmed (4,000 more than in 1940) which represented almost one-fourth of the county's total land area. But as the population grew from 276,000 in 1950 to 1.1 million in 1970, agricultural lands decreased to 73,000 acres. The suburban encroachment into agricultural lands created pressures upon the farmers who subsequently found themselves on the fringe of advancing urban lands. Hart (1976) has cited some of these pressures in a recent article where‘ he also identifies urban encroachment onto rural lands as a serious problem in Suffolk County. 28 An outwash plain in Brookhaven Township, the predominant landform in eastern Suffolk County. 29 City people who desire new houses, especially inexpensive new houses, can find them in quantity only in new, outlying developments. Political reality dictates that the burgeoning pOpulation of these new developments must be given city services...Taxes ‘must be raised in order to provide the necessary services, and the farmers who remain are eventually forced to sell their land because they cannot afford to pay urban- level taxes (Hart, l976:l). In this instance where agricultural land uses succumb to urban land uses, topography exerts an indirect influence upon suburbanization as it determines only the original use of the respective lands but not the subsequent uses which are determined more strongly by economics (Figure 11). Suffolk's plains and areas of moderate relief have accommodated the major portion of suburban develOpment since the 19503 while its morainal lands have generally resisted high-density suburban development for two reasons. First, the plains are easier to develop than the moraines in both economic and labor input. Second, in the past they have accommodated the greater amount of agricultural acreage while the moraines provided relatively little arable land. 'It has been shown that as such land uses become economically sensitive in the suburban-rural fringe zone, development is most likely to appear on the newly acquired, defunct agricultural lands. Of the county's four western towns, for example, the two northern towns of 30 FIGURE 11 (a) Remnant agricultural fields in western Suffolk County fringed by (a) residential and (b) in— dustrial lands. 31 Huntington and Smithtown, dominated by moderately dissected morainal topography recorded population densities of two thousand persons per square mile in 1970. The two southern towns of Babylon and Islip, located on the outwash plains recorded higher densities closer to four thousand persons per square mile. The overall effects of the topography upon suburban expansion in Suffolk County as a whole are only locally significant since most of the lands are characterized by plains and areas of moderate relief. Only to a limited extent have the direct and indirect controls of the geo- morphic environment shaped the patterns of intracounty suburban development. Recent POpulation Trends Papulation growth since the late 19403 has resulted in varying levels of develOpment throughout Suffolk County. At present, the western end of the county is heavily suburb- anized, while the eastern end remains relatively undeveloped and even semi-rural in character. For example, in 1972 the county grossed over $81 million from agricultural sales ‘making it New York State's leading county in agricultural production (Suffolk County Planning Department, 1975:296). Only 5 percent of the land area encompassed by the five western townships of Babylon, Huntington, Smithtown, Islip 32 and Brookhaven was devoted to agriculture in 1966, while 22 percent of the land area in the remaining five townships was farmed (NSRPB, 1968:18). From the remnant seventeenth century Dutch-English international boundary line which today divides Nassau from.Suffolk, the level of urban de- velopment shows a decrease from west to east across the county or conversely agriculture subsequently increases. Similar patterns of decrease from west to east exist for population density and percentage of built-up land for each of Suffolk County's ten townships. In 1970 each of the four westernmost townships recorded population densities in excess of two thousand persons per square mile. Densities in five of the six eastern townships, however, remained well below 320 persons per square mile. Brookhaven, the single eastern unit with a relatively high density of almost one thousand persons per square mile, is presently Long Island's fastest growing township, and represents a transitional zone between the urbanized western and semi-rural eastern ends. By comparison in 1966, the percentage of built-up or develOped land decreased from west to east with an average of 47 percent of the total land area in the western townships and 19 percent among the eastern ones (Figures 12 and 13). The spatial variation of built-up lands, agricultural lands and population densities across Suffolk County is summarized in Table 2 which suggests the existence of two 33 FIGURE 12 s°° 400073," SUFFOLK COUNTY: :1 - I970 POPULATION DENSITY E AND DISTANCE FROM n, 3000 - 3.9 NASSAU COUNTY K C 3 hi $¢$ ‘\°¢° m \ .\x Q; to“ tteteI Huntington recorded the second highest 3 2000 r- rheeien family inceneISISQOO) in I969 : tor Neeeee and Sutton Countiee. m P 2 8 l000 " a: u a P O l l l 0 IO M l L E 5 FROM NASSAU COUNTY FIGURE 13 '°°’ SUFFOLK COUNTY: ' I966 LAND USE AND — BUILT - UP 30% DISTANCE FROM ‘2’ AGRICULTURAL NASSAU COUNTY : . ..I C .- O .— '5 :2 MILES FROM NASSAU COUNTY 34 ama .a .m aammemmaa ”mammz Rona maaaaaopw auamcon soaumaneom .m magma one am: mama eaamnsoe mama. Ha mod .Nq. .amm moa deuce huaaoo xaommnm Auama u Aua u cmasv Ame u cameo one a H mm o Ha mamama nouaonm mma m 3 am aa aa aoaaamm mama mam om aa mm ma mm vaonunom mam aa aa am as maa :ouaamauaom mum me am am am we moonao>am cam n ma mm am How cm>mnxooam mmau a N am am an cBounuaam mmam a a am am mOa aaaaa aaoe a a ma mm mm :oaaaam amam a a an an ea coumcaucsm Apnea mo oaaz Hooch aoHSuanoan< THwUOH (manuaaam aou< oumnom mom mnemammv mo N mo N mama deuce moauamCon Ammaaz.oamsom :Hv eoaamasaom Caz oaaa.|:: ma: mesa 35 distinct internal zones. The townships in the Inner Zone (western zone) include Babylon, Huntington, Islip and Smithtown. Each has: 1) population densities in 1970 above the mean value for the entire county; 2) percentages of built-up lands in 1966 above the mean value for the entire county; and 3) percentages of agricultural lands in_ 1966 below the mean value for the entire county. The Outer Zone (eastern zone) includes the six remaining townships, and although it displays slightly less statistical uniformity in actual land use percentages, overall simi- larities do tend to unify the area. The members of this zone are characterized by: 1) population densities in 1970 below the mean value for the entire county; 2) percentages of built-up lands in 1966 below the mean value for the entire county, with the single exception of Shelter Island; and 3) percentages of agricultural lands in 1966 below the mean value among the 36 peripheral towns of Brookhaven,‘ Shelter Island and East Hampton, but below the mean value among the central towns of Riverhead, Southold and Southampton. The different levels of development existing in these two zones today result from intensive countywide suburban- ization that has taken place over the last quarter-century. The evolution of these present patterns is best explained by analyzing the recent population data for the county's ten townships shown in Table 3. From 1950-1960, all five western townships experienced population increases greater than 100 percent, while the eastern townships grew by less than 60 percent. During the next decade, all of the western townships exhibited decreases in their rates of increase, 'with only two (Smithtown and Brookhaven) maintaining rates in excess of 100 percent; the others decreased by an average 120 percent. In contrast, two eastern townships (Southold and Shelter Island) increased their populations by 11 percent and the remaining three averaged relatively modest declines (19 percent). Using LILCO's 1977 population data, current growth rates indicate significant shifts in the county's develoament patterns. Over the last seven years, the Inner Zone townships have experienced sharp declines recording rates of less than 13 percent. But in the Outer Zone, pop- ulation increases ranged from slightly less than 15 percent 37 Ammueaaumm coda n :35 am .a .aaaa .ooaaa seam maaaasoov eme.maa.~ aaa mmo.a~a.a am ama.mmm aaa ama.ma~ annoy amesoo xaoaasm amm.m mm aam.a ma aam.a ma aaa.a mamaaa aauaaem mmo.¢a mm oma.oa «a amm.m o3 mam.m :ouaamm uaam maa.aa ma som.ma mm mam.ma aa mam.aa maonaaom mam.ea mm «ma.mm mm mao.a~ am mao.aa aoaaaaausom 3mm.am ma ama.ma mm aam.ea ma maa.a mmasam>am ama.amm mm mmm.mam mma aoa.aoa Asa mam.ae ca>aaaooam aam.mma m amm.aaa mma mam.om oaa maa.a~ :souauaam mma.~am ma omm.ma~ am ama.maa maa mm¢.aa aaaaa nom.am~ a mam.mmm ma aom.~aa Nam mmm.ma eoaaamm ama.aam a «ma.mom am amm.m~a .mma moa.ae commaaucsm saaaa mmeaao oaaa ameaau omaa «memeo omaa e e e nmmaiomaa "owamno ufioouom one aoaumaseom Hooch masonsoh .m manna 38 to 38 percent. Rapidly suburbanizing Brookhaven, for example, gained over twelve thousand residents in 1976. In summary, in terms of population growth since the 19603 the previously rapid-growing Inner Zone has experienced dramatic decreases, while the Outer Zone has recorded only minor decreases and presently exceeds the western zone by 14 percent (Figure 14). Within the individual townships of the Outer Zone, an analysis of annual population growth rates for the 1960-1970 period and the 1970-1977 period reveals significant changes in the patterns of current intrazonal suburbanization.4 From 1960-1970, the highest annual rates were found in the northwestern parts of Brookhaven Township where eighteen of its forty-six minor civil divisions (MED) recorded percentage increases above the zone's mean value of 13 percent. Through- out the entire township, annual rates of increase ranged widely from 2 to 25 percent, with only one MCD (South Yaphank) experiencing a population decrease. Annual rates occurring in the remaining five townships also varied widely (Figure 15). During the first seven years of the 19703, the MCDs recording the highest annual growth rates throughout the zone were located in the eastern half of Brookhaven Township 4Annual population growth rates have been calculated and are used here to enable direct comparison of 1960 trends ‘with 1970 trends. The use of only the first seVen years for 1970 rates is adequately representative. FIGURE 1 SUFFOLK COUNTY no g. POPULATlo GROWTH ATES ‘50 BY ZONE \940—-\977 \30 : no (By zona averages) 4 IOO 2 so ‘2 70 so ,\\\ l -‘\ .. I” \\\.“'. [1 \\°. 30 II P\ OUTER ZONE. II II INNER ZONE \0 O \960 \370 I977 40 which represents an eastward shifting in the cluster of high rates found in the 19603. This shift is indicated specifi- cally by the township's seven eastern MCDs that experienced increases in their annual population growth rates in the 19703. MOre significant is the apparent spillover Of high increase rates into four MCDs in western Southampton (Riverside, Ramsenburg-Speonk, West Hampton and West Hampton Beach) and one MCD in central Riverhead Township (Roanoke). Relative to these population changes in the western townships, the eastern townships have experienced less significant growth rates (Figure 16). From.the percentage changes in the eastern zone's western and between the 19603 and 19703, a spatial trend of systematic suburban expansion is evident. During the 19603 the highest rates were clustered in northwestern Brookhaven Township, adjacent to the heavily developed Inner Zone. In the 19703, however, the formerly high- increase MCDs experienced decreases in their rates Of growth while neighboring ones to the east experienced increases. Development is therefore occurring in a syste- matic pattern as MCDs adjacent to rapid growth areas them- selves experience rapid growth in successive time periods after growth has increased, peaked, and then subsided in the rapidly growing MCDs. In combination, the patterns Of intertownship and intratownship growth indicate that suburbanization is 41 CO. :2: .235 malo~ QIO. mlv mIo o :2: 30.. hzwummm oz<.._m. cub-51m .\.Nm. I. 0N9 .mw 2410 ZO_._.mHZDOQ vfiommam 3” $50”; 44 the seventeen year period, and reflects a significant over- all shift eastward in the county's population. From the preceding Observations concerning the spatial and temporal dimensions Of recent suburban development in Suffolk County, it is hypothesized that at any given time, a township's location with respect tO all Others determines both its rate Of growth and population density. TO test this hypothesis, an analysis Of these three variables (location, growth and density) is carried out for all ten townships. Assuming that the boundary line with congested Nassau County marks Suffolk's zone Of maximum.suburban de- velopment, a township's straight-line distance from it rep- resents relative location tO the suburban core. The 'measured distance from Nassau County is then used to deter- mine the strength Of its relationship to the respective 1977 population density and the percentage population change from 1970-1977 for each township. Computing the rank-order Correlation coefficients, a strong inverse relationship Of -.90 is Obtained between distance and density, and a strong positive relationship Of +.73 between distance and percent- age population change. In summary,'suburbanization has diffused eastward across Suffolk County from Nassau County into the western portion Of the Outer Zone. Presently, Suffolk County's eastern townships demonstrate the greatest potentials for future development because Of their relatively low population 45 densities, high population growth rates, and low percentages Of built-up lands. These townships also accommodate most Of the county's agriculture which has been shown tO be a tenuous economic activity. It is highly sensitive tO the pressures Of expanding urbanization. The eastern half Of the county is topographically uniform, predominated by outwash plains as the Harbor Hill and Ronkonkoma moraines are highly interrupted and less evident here than in the west. But the so called "tidal.wave" Of urbanization has yet to crest in these transitional townships Of the Outer Zone. By identifying some Of the significant variables active in the county's past development trends, an esti- mation Of future development situations is possible. The following chapter examines both past and future trends by applying three simple Monte Carlo simulation models tO suburbanizing Suffolk County. CHAPTER IV THE SUFFOLK COUNTY SIMULATION MODEL OF SUBURBAN DEVELOPMENT Simulation MOdels in Geographic Research Since its inception and introduction as a social science research tOOl by Hagerstrand in 1952, the Mbnte Carlo simu- lation model of diffusion has become a common analytical technique for the study Of various geographic problems (Bryant, 1975:88). The continued use of such probabilistic modeling has been encouraged by King (1969) who states: The development Of this probability analysis in relation to patterns Of locations and interactions as they vary over space and time is a challenge for future geographic research. It should not go unanswered (King, 1969: 230). Before investigating the application Of the MOnte Carlo simulation.model to the study Of suburban expansion, a brief review Of several significant applications in geography is provided. Taaffe, Garner, and Yeates (1963) proposed a progressive series Of simulation models in their study Of Chicago's peripheral commuter patterns. The models were based upon 46 47 empirical Observations and applied tO the city's western suburbs; the purpose was to simulate the distribution of the area's commuters. From the actual patterns Taaffe et a1. determined that although one primary factor (distance) *was nonrandomly related "to place Of employment decisions," the majority Of influences upon these decisions were ran- domly distributed. Since the actual distribution contained "a strong random.element, static versions of Hagerstrand's Mbnte Carlo simulation Of diffusion were used which enabled the identification Of purely random from purely nonrandom processes. Four models were presented with the probability field Of MOdel I arranged according to the influences Of population size and distance upon commuting, the primary factors. After this first simulation, progressive modifi- cations were integrated for each successive run to incorp- orate additional probabilities if the existence Of a non- random.process was indicated from a preceding run. Under the design controls set up, patterns developed from.aon- random processes were easily distinguished from those de- veloped by random processes. Such determinations not only benefit the understanding Of local (Chicago) patterns Of intraurban commutation, but also aid in the develOpment Of relatedgeneral theories concerning urban systems. Mbrrill (1962; 1965a; 1965b) employed similar probabil- istic methods to Taaffe et al. in several articles in the 19603. In his noted study of Negro migration and ghetto 48 expansion in Seattle, he prOposes a very simple MOnte Carlo simulation model (Merrill, 1965a). He begins by identifying the expansion Of the ghetto "as a spatial diffusion process in which Negro migrants gradually penetrate the surrounding white area." A number Of relevant elements needed tO con- struct a probabilistic model Of "ghettoization" were selected by empirical Observations, including: ...natural increase Of the Negro population; Negro immigration into the ghetto; the nature Of the re- sistance tO Negro out-migration and its relation to distance; land values and housing characteristics; and the population size limits Of ggggination blocks (MOrrill, 1965a: A probability field was constructed incorporating these empirical factors into the simulation framework. The con- ventional generation Of random.numbers was then carried out to produce several simulations Of ghetto expansion from 1940-1964. In discussing his results Mbrrill indicates that the simulated patterns Obtained need not match the actual patterns Of ghettoization precisely, since minor variations are acceptable in this type Of modeling. He notes: Simulation is a valuable technique in science and technology, in which a model is constructed to depict arti- ficially certain major features Of some real process. 49 And MOrrill continues: We dO want the model tO generate a pattern... that corresponds in its characteristics to the real pattern, and we can satisfy our- selves Of the correspondence by visual and statistical tests. The purpose and hOpe are to discover and illustrate the nature Of the process (MOrrill, 1965az349). In addition to their application to problems Of human spatial behavior such as innovation diffusion, migration and commuting, Monte Carlo simulation models have been used extensively in the study Of urban growth and land development. Garrison (1962) is credited with introducing such models into the field Of urban spatial research. Since then Chapin, Donnelly and Weiss (1965) have utilized simulation techniques tO study residential development patterns and later Yuill (1970) devised a comprehensive spatial simulation model Of urban growth. TO investigate the expansion Of Seattle's urban fringe, MOrrill (1965b) again used a MOnte Carlo simulation model very similar to the one used in his study Of the city's Negro ghetto. This model is less detailed in design and more general in con- text, and is based upon MOrrill's hypothesis: ...that in spatial detail, expansion Of the urban fringe may be described as a spatial diffusion proceSs in which de- velopment Of new properties is essentially random.in direction (land being equal) (MOrrill, 1965b:186). 50 In establishing the probability Of development for specific sites at the urban fringe, he notes: ...marginal land, distance from major arterials, schools, and shopping centers, as well as proximity to existing major developments, were found to be important...(MOrrill, 1965b:187). These factors were then built into his model and a simulation was rendered which estimated the spatial patterns Of develop- ment for several time periods. Concluding with an evaluation of the "fit" between the simulated and the actual patterns, the author reasserts that the method itself "does not ask for exact replication," but attempts to generate growth patterns which are spatially similar to reality. Malm, Olsson, and.Warneryd.(l966) introduced a more intricate model Of urban growth based upon a camplex series Of mathematical derivations concerning the function Of distance in spatial theory. This model also utilizes tra- ditional Monte Carlo methods to simulate urban sprawl through time. The probability matrix was formulated primarily from the inversion Of an actual cost grid for the construction Of new residential units in a part Of GOthenburg, Sweden. Follow- ing the design Of Morrill's (1965a) migration model, a re- sistance factor was introduced to incorporate the effects Of ground conditions at potential development sites into the simulation. Urbanization was then estimated by standard procedure, that is, the matching Of generated random numbers 51 to the appropriate ranges found within the cells Of the matrix. The resistance values, which range from.ane tO three, were used to indicate the number Of "hits" required to develop a cell. For example, cells delimiting sites where bedrock predominated the surface were assigned a resistance value Of two, since construction would be more difficult here than at sites characterized by sand and gravel which received a value Of one. And sites having clay at the surface were assigned the highest resistance value, three. The resultant simulated patterns Of the city's growth from 1920-1940 are in "fair agreement" with the respective empirical patterns. Hence the model's ability to "sketch" the general features Of real-world processes is demonstrated satisfactorily. In conclusion the authors evaluate its overall acceptability for study- ing the geographic problem Of urban sprawl. From the above review Monte Carlo simulation is considered an effective research method useful in the analysis Of various spatial and temporal problems, parti- cularly, development and land use change inside urban areas. The technique itself can be used both to identify the causative factors underlying spatial trends and to simulate, in a predictive manner, future situations result- ing from those trends. These capabilities suggest its applicability to further investigate suburbanization processes and patterns in Suffolk County. 52 Suburbanization: A MOdel and Simulations Since the end Of WOrld War Two, Suffolk County has experienced rapid and extensive suburban growth. Figure 18, compiled from.USGStopographica1 maps, shows the extent of the county's built-up lands for 1949 and 1969. From the map gross development patterns are identifiable, as are several factors which have apparently influenced the evolution Of the county's suburban landscape. First, distance from New York City has evidently been a primary factor in this evolution. The largest portion Of the county's total development took place within the Inner Zone while the Outer Zone experienced very little growth during the same twenty year period. As discussed in the preceding chapter, however, current intracounty development patterns indicate a reversal Of earlier trends. A second apparently important factor is adjacency to existing development. The expansion or sprawl Of built-up places into surrounding lands accounted for a large part Of subsequent development with only several instances Of isolated spontaneous development occurring throughout the county. In effect localized sprawl has filled-in much Of the interstitial land left vacant by "leap-frog" develop- ment, a common situation in urban fringe areas (Hart, l976:l). In addition to these primary factors suburbanization 53 has also been affected by the interplay of various socio- economic, political, and environmental parameters. Land use zoning, for example, exerts direct and immediate con- trols upon forthcoming development. Land use patterns are partially created or shaped by legislative action. In other words, the zoning codes enacted by municipalities are contributory to the resultant develop- 'ment. One action usually complements the Other (NSRPB, 1966:27). A political mechanism shaping local suburban growth is illustratedby Suffolk County's Farmland Acquisition Program” By enabling the County to purchase the development rights to agricultural lands from.their private owners, the program is an attempt to prevent the sporadic development Of those lands in eastern Suffolk County. Its success would preserve presently idle agricultural lands in their natural state, and allow the implementation Of a controlled development process to ensure more efficient land use for both the present and future. In essence, both zoning and land develop- ment programs reflect legislative schemes that are inherently short-term in design, channelling growth in response tO prevailing public desires and concerns. Such policies, however, are oftentimes devised in direct response to environ- mental situations, rendering land use designations that are permanent. For instance, local topographic conditions can limit or restrict certain landscapes from.specific uses as 54 illustrated by the undeveloped sections Of western Suffolk County which, for the most part, identify the hilliest areas (Figure 18). Suburbanization in Suffolk County is influenced pre- dominantly by: 1) distance from.New York City; 2) adjacency to existing develOpment; 3) land use policies; and 4) topography. The first two factors (distance and adjacency) represent universally applicable parameters while the last two rep- resent purely local parameters, applicable only within the study area itself. Since one Objective Of this study is toward the formulation Of a general theory Of suburban growth that is not overly hindered by local constraints, the latter factors Of land use policies and tOpography are omitted from further consideration. The following compara- tive analyses Of actual and simulated patterns Of recent develOpment examine only the importance of distance and adjacency upon suburbanization. _ TO begin the simulation a grid is placed arbitrarily over the entire county, with each internal cell delimiting a development site Of approximately 1 square mile (2.6 square kilameters). Excluded from this grid are the south- shore barrier islands, Gardiners, Shelter, and Fishers 55 Islands, and the extreme eastern tip of East Hampton Town- ship; these areas have little if any developable land. By generalizing the built-up areas existent in both 1949 and 1969 tO conform tO this grid several important factors concerning suburban development become apparent (Figure 19). Since 1950, development has consumed almost all of the available land within the county's Inner Zone. In fact, without certain limitations Of the physical environment and land use restrictions, this zone would be totally developed because Of its proximity tO New York City and urbanized Nassau County. Extensive suburbanization, as mentioned above, has taken place only in the western end of the Outer Zone, specifically in western Brookhaven Township. The major portion Of this zone remains undeveloped, suggesting that it will be forced to accommodate any future county growth. The ubiquity Of built-up lands throughout the Inner Zone, correlated with the localized and limited deVelOpment found within the Outer Zone, indi- cates that suburbanization has been a highly nonrandom process. When compared to these actual patterns, simu- lated patterns of deVelOpment generated exclusively fram random methods will identify the extent Of nonrandommess in theempirical process. First, to simulate random development, each cell with- in the previously constructed grid that contains over one- half square mile Of land is numbered. Cells identifying 56 n. o. n o a 332:2... 3m. 3 33...: 2.8 g arm. 3 52.226.» .200 a 0E0 20:44:25 3 om~3._.ZDOU xaommsm amaze: 1|4||mo JHZDOQ xaoumsm .2352“. 57 built-up land in 1949 are omitted, since the attempt here is to simulate development from.l949-1969. A determination is then made Of the number Of cells actually developed throughout the county over the twenty year period, and random numbers are drawn until the appropriate number of cells (200) is developed. The method itself is basically that used in simple random sampling procedures. Comparing the random development situation to the empirical situation illustrates the high degree of dependence upon nonrandom factors in the actual process (Figure 20). It is there- fore necessary tO incorporate the previously identified, nonrandomly distributed factors Of distance and adjacency into a second simulation model, and thus gauge their spatial significance. This second model is constructed by the same Monte Carlo techniques used in the first, but two manipulations Of the numbered matrix are entered. To account for a de- velopment site's distance from.New York City, the study area is divided from west to east into three sections Of equal length, approximately 25 miles (40 kilometers). The western section (Sector I) being nearest to New York City (and adjacent to Nassau County), has the highest probability for development. Consequently, each cell within this sector is assigned a range Of four consecutive random numbers to increase its chance Of development as random numbers are drawn. Cells located in the middle section (Sector II) 58 are assigned a slightly narrower range Of three consecutive numbers, while cells located in the remaining section (Sector III) receive only single number designations. In effect, development becomes less likely with increasing distance from New York City, as suggested by the actual patterns illustrated in Figure 19. Accounting for adjacency tO existing development is accomplished by assigning simple resistance values to each cell throughout the entire county according to relative location. An undeveloped cell ad- jacent either vertically, horizontally, or diagonally, to one that was developed in 1949, required only one hit by a drawn random number to develop. Nonadjacent cells re- quired two hits before they were developed. The resulting sbmulation Of growth is shown in Figure 21, illustrating a probable development pattern constructed from.random processes, and controlled by the influences Of distance and adjacency. Although it is far from a replication Of the empirical situation found in 1969, the simulation does suggest a significant degree Of overall correspondence between the model and reality. Only in a few scattered areas was develOpment simulated where it did not actually occur, primarily in the eastern parts Of Sector II. However, this discrepancy is explained by the arbitrary division Of ‘the study area into three sections. Smaller divisions would produce more spatially similar patterns. In Sector III development was neither over nor under estimated in terms Of 59 total area consumed. As a result Of the omission Of tOp- ographic and land use restrictions, simulated development does appear in many Of the presently undeveloped parts Of Sector I. Since the amounts Of estimated develOpment are generally equivalent tO the actual amounts within each sector, the model erred only in locational aspects Of growth, but succeeded in identifying the significant spatial characteristics. Therefore, this model seems applicable for a third simulation that can be used tO investigate future growth trends in Suffolk County's Outer Zone, an area Of high development potential. This quasi-predictive simulation also examines only the effects Of distance and adjacency upon suburbanization, and assumes that the total amount Of land developed in ~ the Outer Zone through 1990 will be approximately equal to the amount developed throughout the entire county from 1949-1969. The assumption is based upon the relatively large tracts Of vacant, developable land distributed throughout the zone and the well developed transportation network that allows easy access tO all major centers. From this Operational assumption, it is also possible to estimate growthpatterns for two ten year periods: 1970-1980 and 1980-1990. Following previous designs, the grid covering this zone is divided into three sections whose extents are de- termined by the intersections Of the Brookhaven, Riverhead, 60 and Southold Township boundary lines with the northern shoreline. Ranges Of four and three consecutive numbers are assigned tO the cells located in the western and middle (Sectors Ioz and 1102) respectively. Since exten- sive development is assumed for the entire zone, the cells in the eastern section (Sector Illoz) are assigned a range Of two consecutive numbers, reducing the overall variation Of development probabilities between all three sections. Resistance values are entered by the same method as before to incorporate the effects Of adjacency to existing development into the model. Random numbers are then drawn and the resulting qualitative estimate Of the probable distribution Of developed lands across Suffolk County's Outer Zone is given in Figure 22. As the model is based solely upon previous growth trends, it remains purely speculative in nature. Thus, the patterns rendered in Figure 22 are estimations at best, since in reality future suburbanization will be dependent upon a number Of factors not considered here, such as regional economics, political actions, and.demographic trends. The model does, however, suggest the continued importance Of two Of the most locally significant influences upon suburbanization, namely distance from New York City and adjacency to exist- ing development. By sketching these future patterns accord- ing to prescribed probabilistic methods, the model is a valuable tOOl that can be effectively used for the planning 61 FIGURE 20. w "I‘VE-3': 4mm: Pi-Iii *XZI. HIPQM'” . Waeo‘o'.(aizm¢m_ _ rt .1 k , a s .a .2 atom r ., . legg‘guw- W" W.”- v r SUFFOLK COUNTY: RANDOM SIMULATION OF DEVELOPMENT, |949 - I969 . ~4‘J' .‘I' ':,.‘7 / amt-f ~ , ., _ "I ll. -1! 15.7mm ng'K Immun- nuneneu I a a In I! e Indicate: simulated development. FIGURE 2|. I949— I969 SUFFOLK COUNTY: CONTROLLED SIMULATON OF DEVELOPMENT, 55"” I SECTOR 1: monom- e Indicates simulated development. I970- I990 FIGURE 22. OUTER ZONE: SIMULATION OF DEVELOPMENT, '5'»? gfififlfih momma-h "is game Tea 3333?“. I Vow-or I980 O ISSO-ISSO elndlcatee simulated development l970- 62 Of Suffolk County's impending suburban landscape. Model Evaluations The two simulation models presented above have estimated development trends in Suffolk County from 1950-1990. Although the performance Of the first model (Figure 21) indicates a high degree Of spatial similarity tO the empirical situation, several additional considerations are suggested. By far the greatest error was found in the locational abilities Of the first model; accurate amounts of development were simulated for each Of the three con- structed sectors, but not in the appropriate places within them. This error, however, is easily accounted for since the model examined only two Of the numerous factors which influence suburban growth. The incorporation Of topographic, political, and socioeconomic parameters into the model would most likely improve overall efficiency. Manipulations of the actual sizes Of both the individual probability cells and the resistance sections could also prove beneficial in future investigations. These modifications would then provide the basis for a predictive simulation model to be formulated with the ability to forecast future development patterns in a highly reliable manner. CHAPTER V SUMMARY AND RECOMMENDATIONS To summarize, this study has identified urbanization as the central theme Of the population dynamics evidenced throughout the world today. Urbanization in the United States, however, is a multidimensional process that goes beyond the simple shifting Of the nation's urban-rural population components. During the last quarter-century, for instance, suburban regions have been growing far more rapidly than their urban counterparts. Since the Bureau Of the Census estimates that in 1975 about 39 percent Of the population resided in the suburbs, a definite need has arisen for more geographically oriented research on this newly emergent "outer city". In response to this need, the population patterns Of the New York Metropolitan Region (NYMR)--the nucleus Of Gottmann's Megalopolis -- were first examined for the three decades between 1940-1970, revealing that population change within the member counties depended upon three major factors: distance from.the New York City CBD; population density; and existing level Of development. For example, population growth rates through time were 63 64 systematically lower in the more urbanized counties and relatively higher in the less built-up, suburban counties. These regional patterns clearly reflect the contemporary trend towards the deconcentration and decentralization Of people and activities from.city tO suburb, or more speci- fically "the urbanization Of the suburbs" (Muller l976:l). The gross regional patterns analyzed within this first Objective, suggested local variations in suburbanization processes that warranted a more detailed investigation Of one Of the NYMR's low-density, outer suburban counties. Nassau and Suffolk Counties are located on Long Island; as Of 1972 they became the nation's first all suburban Standard Metropolitan Statistical Area with a combined population Of 2.5 million people. The island itself is politically divided into four counties: the western two (Kings and Queens) are boroughs Of New York City, while to the east, Nassau and Suffolk are independent counties Of New York State. The island is Often viewed as a linear projection Of land extending eastward from urbanized New York City to semi-rural Suffolk County. This situation Of different conditions Of development existing at Opposite ends of the island has been an impor- tant theme investigated throughout this study. Population densities, for example, decrease from.fifty thousand persons per square mile at the western end to less than five hundred persons per square mile at the eastern end. 65 The total amount of built-up lands also follows a similar decreasing pattern from west to east. Since recent suburban development has consumed almost all Of Nassau County's developable lands but only those in the western half Of Suffolk County, the Nassau-Suffolk SMSA provides a prime area for the study Of the process Of suburbanization. With suburban growth expected to continue in this region through-I out the remainder Of this century, and Suffolk County accommodating more than 90 percent Of that growth, this county was selected as the study area to provide an "in situ" laboratory Of both past and present development trends. The relatively large tracts of undeveloped land located in the county's Outer Zone also suggested a qualitative investigation into probable future patterns Of development for this area Of high growth potential. Simple renditions Of Monte Carlo simulation models were presented which helped to identify two Of the most locally significant factors affecting suburbanization, namely: distance from New York City and adjacency to existing development. These probabilistic models were constructed and demonstrated to be effective in the esti- ‘mation Of existing develOpment patterns throughout Suffolk County, as well as future patterns in the undeveloped townships Of the Outer Zone. As suburbanization has been shown to be diffusing eastward across Suffolk County, several critical probleme 66 come into view which most certainly demand additional research and investigation. First, although the county's Farmland Acquisition Program is an ambitious attempt to stem the tide Of advancing suburban sprawl, it does little to encourage the economic vitality Of those farmers who presently farm.some Of the most valuable land in New York State and the New York City area. The loss Of such lands, predominantly fruit and truck crop oriented, would no doubt have negative affects upon local and regional seasonal crop availabilities. Second, Long Island is basically an environmentally sensitive area that has recently been exhibiting signs Of ecological devastation. In some parts of Nassau and Suffolk County, for example, intensive post WOrld war Two urbanization and industrialization have severely contaminated local groundwater reserves, thus endangering the island's sole fresh water resource (Galant, 1977:28). Continued growth, characterized by more concrete, more asphalt, and less Open space, would definitely intensify these problems. Finally, the most pressing problem now threatening suburbia in general reflects the simple dilemma Of the diminishing urban tax base in today's complex metro- politan region. It has been indicated that although suburban counties have prospered tremendously over the past three decades by developing their local job bases, 67 ...those bases are built upon city- earned incomes and their growth really is the growth Of the city 3 economy beyond its political boundaries (Stern, l976:E7). Therefore, as suburban Suffolk County continues to grow, it is inadvertently drawing vital resources away from New York City and weakening the economic structure Of the entire urban region. MOre geographic research into the environmental and socioeconomic problems which result from contemporary suburbanization is needed. It is hoped that this study has suggested some possible directions and methodologies toward that frontier. APPENDIX 68 APPENDIX LILCO's Population Survey The following is taken from.the Long Island Lighting Company's 1977 Population Survey and explains their methods Of estimation. Essentially, LILCO's annual estimate Of present population is determined from census figures and active residential electric meters. For each community a factor, people per meter, established at the time Of the most recent census, is ‘multiplied by the recorded number Of meters added or subtracted each year tO yield the annual population change in each community. The estimates are also weighted by type Of dwelling unit to compensate for the increasing trend to apartments, condo- miniums and senior citizen housing, which have fewer occupants per dwelling unit. In addition, establishments that have group accommodations are contacted to insure a correct count Of this segment Of the 69 population. For example, contacted were the state hospitals in Suffolk County where the present population of 10,295 represents a decrease Of 12,413 persons since the 1970 census. An illustration Of the accuracy Of this method is Obtained by comparing the findings of the U. S. Bureau Of the Census special census on Brookhaven Township to the LILCO estimate for 1975. The census reported 320,677 in April, 1975; LILCO estimated a January 1 population Of 321,150. BIBLIOGRAPHY . 7O BIBLIOGRAPHY Books Bergman, Edward and POhl, Thomas. The Geography Of The .New York Matropolitan Region. Dubuque, Iowa: Kendall/Hunt Publishing CO., 1975. Blalock, Hubert M. Jr. Social Statistics. New York: McGraw-Hill, Inc. 1972. Browning, Clyde E , ed. Population and Urbanized Area Growth in Megalopolis, 1950-1970. University Of North Carolina, Studies in Geography No.7. Chapel Hill: 1974. Carey, George W; New York-New Jersey: A Vignette of the Metropolitan Region. Annals Of the Association Of American Geographers Comparative MetrOpOlitan AnalygégéProject. Cambridge: Ballinger Publishing CO., . Clawson, Marion. Suburban Land Conversion in The United States: (An Economic and Governmental Process). Baltimore: John HOpkins Press, 1971. Detwyler, Thomas R. and Marcus, Melvin, eds. Urbanization and Environment: The Physical Geography Of the City. Belmont, CA: Duxbury Press, 1972. Dobriner, William.M. The Suburban Community. New York: G. P. Putnam's Sons, 1958. . . Donnelly; Chapin; and Waiss. A Probabilistic MOdel for Residential Growth. Urban Studies Research MOnO- graph, Institute of Research in Social Science, University Of North Carolina.- Chapel Hill: 1964. Gans, Herbert J. The Levittowners. New York: Pantheon Books, 1959. 71 Gottmann, Jean. Megalopolis, The Urbanized Northeastern Seaboard Of the U. S. New York: Twentieth Century Fund, 1961. Gottmann, Jean and Harper, R., eds. Metropolis on the Move: Geographers Look at Urban Sprawl. New York: JohnWiley and Sons, Inc., 1967. Guetzkow, Harold Steere, ed. Simulation in Social Science; Reagings. Englewood Cliffs, N.J.: Prentice-Hall, 96 . Hammersley, J. M. and Handscomb, C. C. MOnte Carlo Methods. London: Methuen and CO., Ltd., 1964. Hoover, Edgar M. and Vernon, Raymond. Anatomy Of a'Metro- polis: The Changing Distribution Of People and Jobs Within the New York Metropolitan Region. Cambridge: Harvard University Press, 1959. Johnson, James H., ed. Suburban Growth-Geographical Processes at the Edge Of the Western City. New York: John Wiley and Sons, Ltd., 1974. Kamer, Pearl'M. Manufacturing in Suburbia: the Nassau- Suffolk Employment Experience. 'Mineola, N.Y.: Nassau County Planning Commission, 1970. King, Leslie J. Statistical Analysis in Geography. Englewood Cliffs, N.J.: Prentice-Hall, Inc., 1969. Milgram, Grace. The City Expands: A Study Of the Conversion Of Land from Rural tO Urban Use. Institute for Environmental Studies, University Of Pennsylvania. Philadelphia: 1967. Masotti, Louis H. Suburbs, Suburbia and Suburbanization: A Bibliography. MOnticellO, Il: Council Of Planning Librarians, 1974. Meyer, David R. Urban Change in Central Connecticut: From Farm tO Factory to Urban Pastoralism. Annals Of the Association Of American Geographers Comparative MetrOpOlitan Analysis Project. Cambridge: Ballinger Publishing CO., 1976. Northam, Ray. Urban Geography. New York: John Wiley and Sons, Inc., 1975. 72 Putnam, Stephen H. An Empirical MOdel Of Regional Growth: with an Application to the NE Magalopolis. Regional Science Research Institute, MOnograph Series, NO. 6. Philadelphia: 1975. Quante, WOlfgang. The Exodus of Corporate Headquarters fr;2.New York City. New York: Praeger Publishers, 9 . Regional Plan Association Of New York. Spread City. New York: 1962. Reilly, William R., ed. The Use Of Land: A Citizen's Policy Guide tO Urban Growth. A Task Force Report Sponsored by the Rockefeller Brothers Fund. New York: Thomas Y. Crowell CO., 1973. Rugg, Dean S. Spatial Foundations Of Urbanism. Dubuque, IA: William C. Brown Publishers, 1972. Taaffe, Edward; Garner, Barry; and Yeates, Maurice. The Peripheral Journey to Work: A Geographical Con- sideration. Northwestern University MetrOpOlitan Transportation Series. Evanston, Il: Northwestern University Press, 1963. Ullmann, John E., ed. The Suburban Economic Network, Economic Activity, Resource Use, and the Great Sprawl. Praeger Special Studies in Economic, Social an; Political Issues. New York: Praeger Publishers, l9 7. . Vernon, Raymond. Metropolis 1985: An Introduction Of the Findings Of the New York Metropolitan Regional Study. Cambridge: Harvard University Press, 1960. Weiss, Shirley F., et a1. Residential DevelOper Decisions: a focused view Of the urban growth process. University Of North Carolina, Center for Urban and Regional Studies, Institute for Research in Social Science. Chapel Hill: 1966. Yeates, Maurice H. and Garner, Barry J. The North American City. New York: Harper and Row Publishers, 1971. Zschock, Dieter R., ed. Economic Aspects Of Suburban Growth: Studies Of the Nassau-Suffolk Planning Region. State University Of New York Economic Research Bureau. Stony Brook: 1969. 73 Articles Blumenfeld, Hans. "The Tidal wave Of MetrOpOlitan Expansion." Journal of the AmeriCan Institute Of Planners 20 (1954): 3-14. Brodsky, H. "Land Development and the Expanding City." Annals Of the Association Of American Geographers 63 (1973): 159-1661 Bryant, C. R. "The Conversion Of Arable Land to Orchards: A Simulation Experiment Using Simple Probability Surfaces." Gaografiska Annaler 57B (1975): 88-99. Chapin, Donnelly and weiss. "A MOdel for Simulating Residential Development." Journal Of the American Institute Of Planners 31 (1965): 120-125. Clawson, Marion. "Urban Sprawl and Speculation in Suburban Land." Land Economics 38 (1962): 99-111. Crouch, W. W. and Dinerman, B. "Decentralization in Magalopolis." In Urban AmeriCa-in the Twentieth Centu . PP. 194-200. Edited’by M.'D. Speizman. ew or : Thomas Y. Crowell CO., 1968. Galant, Richard. "NO Real Answers on Cancer A ents in Wells." Newsday, 3 February 1977, p. 2 . Garrison, William L. "Toward Simulation MOdels Of Urban Growth and Development." In Proceedings Of the ICU Symposium.in Urban Geography. Edited—byTKEut Naggerg. Lund: Lund Studies in Geography 243, Hagerstrand, Torsten. "The Propagation Of Innovation Waves." Lund Studies in Geography 4B (1952): 1-21. . "A MOnte Carlo Approach to Diffusion." 'Annils Of the Association Of American GEOgraphers 6 (1965) : 43-67. Hart, John Fraser. "Central Tendency in Areal Distributions.’ Beonomic Geography 30 (1954) : 48-59. . "Urban Encroachment on Rural Areas." Geographical Review 66 (1976) : 1-17. 74 Holmes, John H. "External Commuting as a Prelude to Suburbanization.‘ Annals of the Association, Of American Geograp ers : - Hudson, John. "Density and Patterns in Suburban Fringes." Annals Of the Association of American Geographers 63 (I973) : 28-39. Kessler, Robert E. "A Three-Year Plan Of Action." Newsday, 3 February 1977, p. 28. Malm., Roger; Olsson, Gunner; and Werneryd, OlOf. "Approaches to Simulations Of Urban Growth. " Gaografiska Annaler 483 (1966) :9 22. ‘Mitchell, James K. "Adjustment to New Physical Environ- ments Beyond the Fringe." Gaographical Review 66 (1976) :18-31. Morrill, Richard. "Simulation Of Central Place Patterns Over Time." Lund Studies in Geography 24B (1962) :109-120. ' (a). "The Negro Ghetto: Problems and Alterna- tives." GaographiCal'ReView 55 (1965) :339-361. (b). "Expansion Of the Urban Fringe: A Simulation Experiment. " Pa ers and Proceedin s Of the Regional Science Association 15 (I965) :155-I99. MOrris, Tom. "Planning Report Details LI Growth." Newsday, 18 March 1977, P. 17. Newling, Bruce E. "Urban Growth and Spatial Structure: Mathematical MOdels and Empirical Evidence. Geographical Review 56 (1966) .213 225. "The Spatial Variation Of Population Densities‘ GeOgraphical Review 59 (1969) :242- 251. Pryor, Robin J. "Delineating Outer Suburbs and the Urban ~Fringe." Geggrafiska Annaler 59 (1969) :242- 251. Schnore, Leo F. "Municip a1 Annexations and the Growth Of MetrOpOlitan Suburbs; 1950-1960. " American Journal of Sociology 67 (1962) :406-417. 75 Sterne, Michael. "Economically, Suburbs and City Can't Escape Each Other." New YOrk'Times, 13 June 1976, p. E7. RepOrts and‘Papers Fuller, Myron L. "The Geology of Long Island." U. S. Geological Survey Professional Paper No. 82, 1914. Kamer, Pearl M. "The Long Island Economy: Anatomy of Change." Technical supplement. Hauppauge, N.Y.: Nassau-Suffolk Regional Planning Board, 1971. Long Island Lighting Company. "Population Survey; Current Population Estimates for Nassau and Suffolk Counties.‘ 'Mineola, N.Y.: 1977. ! Muller, Peter O. "The Outer City: Geographical Con- sequences of the Urbanization of the Suburbs." Annals of the Association of American Geographers Resource Paper No. 75-2. Washington D.C.: 1976. Nassau-Suffolk Regional Planning Board. "Existing Land Use." Hauppauge, N.Y.: 1968. "U. S. Census '70." Hauppauge, NY:1971. ‘ . "Population Estimates and Projections:l975- $995." Interim.Report Series: 1. Hauppauge, NY: 976. Roseman, Curtis C. "Changing Migration Patterns Within the United States." Annals of the Association of’ American Geographers Resource Papers for College Geography No. 77-2. Washington DC: 1977. Suffolk County. "Report to the Suffolk County Legislature from.the Select Committee on the Acquisition of Farmlands," November 1974. Suffolk County Planning Commission. "The Suffolk County Catalog, 1975." Hauppauge, NY: 1975. Yuill, Robert S. "A General Model for Urban Growth: A Spatial Simulation." University of Michi an Geo- graphy Publication No. 2. Ann Arbor: l 70. 76 Theses Carroll, John E. "Land Use Change and Natural Resource Problems on Eastern Long Island." M. A. Thesis, western Michigan University, 1968. Graff, Michael Arthur. "Urban Encroachment on Agricultural Land in East Asia: The Japan Mbdel." M.A. Thesis, Michigan State University, 1971. Neithercut, Mark E. "The Response of Rural Settlement to a Local Hazard System: A Model and Simulation." M. A. Thesis, Michigan State University, 1977. Unterberg, Lester. "An Analysis of Some Aspects of the Changing Population of Nassau and Suffolk Counties, L.I. for the Period 1930-1957." 'M.A. Thesis, Clark University, 1961. Maps Army Map Service. New York, N.Y.: New Jersey; Connecticut Topographic Quadrangle. Series l:250,000. Washington, DC: 1946. United States Geological Survey. New York, N.Y.: New Jersey; Connecticut Topographic Quadrangle. 'Washington, DC: 1969.