Appendix C
Analysis of Density Assumptions and Feasibility of Committee Scenarios

In this appendix, the plausibility of the committee’s scenarios of increasing compact development by 2050 is analyzed through a comparison of the scenarios with trends in residential development. The purpose of this exercise is to understand how substantial the changes from the trends would have to be to achieve a doubling of density of either 25 percent (Scenario 1) or 75 percent (Scenario 2) of new residential development. Two different data sources are available with which to measure trends in development patterns, and there are different ways of measuring the density of residential development. The committee used both of the main data sources and two different ways of measuring density to test the assumptions behind its scenarios. The plausibility of the scenarios is tested by simply calculating how much change would be required from the trends to achieve the projected level of compact development by 2050 and then asking whether such change appears achievable. Conducting this analysis requires first a review of data sources and density measures. The second section uses the data sources and density measures to provide a summary of development trends. The third section presents the comparison between the trends and the scenarios, examines various ways in which the scenario target densities could be achieved, and provides a summary of the committee’s conclusions.

DATA SOURCES AND DENSITY MEASURES

As discussed in Chapters 1 and 2, residential density is typically calculated as the number of persons, households, or dwelling units (DUs) divided by some unit of land area [e.g., acre, square mile (mi2)].



The National Academies | 500 Fifth St. N.W. | Washington, D.C. 20001
Copyright © National Academy of Sciences. All rights reserved.
Terms of Use and Privacy Statement



Below are the first 10 and last 10 pages of uncorrected machine-read text (when available) of this chapter, followed by the top 30 algorithmically extracted key phrases from the chapter as a whole.
Intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text on the opening pages of each chapter. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

Do not use for reproduction, copying, pasting, or reading; exclusively for search engines.

OCR for page 211
Appendix C Analysis of Density Assumptions and Feasibility of Committee Scenarios In this appendix, the plausibility of the committee’s scenarios of increas- ing compact development by 2050 is analyzed through a comparison of the scenarios with trends in residential development. The purpose of this exercise is to understand how substantial the changes from the trends would have to be to achieve a doubling of density of either 25 percent (Scenario 1) or 75 percent (Scenario 2) of new residential development. Two different data sources are available with which to measure trends in development patterns, and there are different ways of measuring the density of residential development. The committee used both of the main data sources and two different ways of measuring density to test the assumptions behind its scenarios. The plausibility of the scenarios is tested by simply calculating how much change would be required from the trends to achieve the projected level of compact development by 2050 and then asking whether such change appears achievable. Conducting this analysis requires first a review of data sources and density measures. The second section uses the data sources and density measures to provide a summary of development trends. The third section presents the comparison between the trends and the scenarios, examines various ways in which the scenario target densities could be achieved, and provides a summary of the committee’s conclusions. data sources and density measures As discussed in Chapters 1 and 2, residential density is typically calculated as the number of persons, households, or dwelling units (DUs) divided by some unit of land area [e.g., acre, square mile (mi2)]. 211

OCR for page 211
212 Driving and the Built Environment Ideally, the denominator should be a measure of net residential acreage to closely match the land use, but such data are not readily available.1 Data Sources The principal data source for the numerator of the density calculation is the decennial U.S. Census of Population and Housing. As discussed in Chapter 2, two national sources of data can be used to track changes in land use—the denominator—over time, each with its pros and cons. The first is the National Resources Inventory (NRI), collected by the Natural Resources Conservation Service of the U.S. Department of Agriculture. The NRI surveys hundreds and sometimes thousands of sample points in each U.S. county and reports, among other data, the acres of land in each of a series of land-cover categories. The NRI combines two of these categories, urban and built-up areas and rural transportation land, to derive an estimate of developed land (see Box C-1 for detailed definitions). For the purposes of this study, the NRI’s land- cover data were aggregated to correspond to U.S. census designations for metropolitan areas as defined in 1999.2 For the density calculation, the NRI’s urban and built-up land category was used as the denominator— referred to in this appendix as urban acres—and intercensal estimates of housing units at the metropolitan level as the numerator. The advantage of using the NRI data is that they provide true land- cover-based estimates of aggregate density, excluding land—even that within city limits—upon which development has not occurred. The NRI data also include urban and built-up land beyond census-defined “urbanized areas” where a very large share of recent development has 1 Net residential acres include only residential land. Nonresidential land uses are excluded, as are local streets and parks (Downs 2004). Gross residential acres, by comparison, include all land within a geographic area, regardless of its uses. Unfortunately, there is no simple or well-established way to relate these two density measures because conditions vary widely from one location to another. 2 The most recent year for which detailed inventory data from the NRI are available is 1997.

OCR for page 211
213 Analysis of Density Assumptions and Feasibility of Committee Scenarios Box C-1 definition of developed land Developed land: A combination of land cover/use categories, including urban and built-up areas and rural transportation land. Urban and built-up areas consist of residential, industrial, commer- cial, and institutional land; construction sites; public adminis- trative sites; railroad yards; cemeteries; airports; golf courses; sanitary landfills; sewage treatment plants; water control struc- tures and spillways; other land used for such purposes; small parks (less than 10 acres) within urban and built-up areas; and highways, railroads, and other transportation facilities if they are surrounded by urban areas. Included are tracts of less than 10 acres that do not meet the above definition but are completely surrounded by urban and built-up land. Two size categories are recognized in the NRI: (a) small built-up areas of 0.25 acre to 10 acres and (b) large urban and built-up areas of at least 10 acres. In 1997, both size categories accounted for 78 percent of total developed land. Rural transportation land consists of all highways, roads, railroads, and associated rights-of-way outside urban and built-up areas, including private roads to farmsteads or ranch headquarters, logging roads, and other private roads, except field lanes. In 1997, this category accounted for the remaining 22 percent of total developed land. Sources: NRCS 2002; NRCS 2003.

OCR for page 211
214 Driving and the Built Environment occurred, but at densities below the census’s threshold for identifying urbanized blocks (Downs 2004). The main disadvantage of the NRI is that it does not provide accurate estimates of land cover below the county level. It therefore cannot be used either to describe the pattern of development within regions (e.g., the continuity of development) or to provide a fine- grained breakdown of development density within a metropolitan area (e.g., the share of housing built above 8 DUs per acre). The alternative data source is the decennial U.S. Census of Population and Housing, which reports (among many other variables) the number of housing units and the amount of land area for the entire United States. The advantage of these data is that they have greater spatial accuracy than the NRI can provide for small areas, with reports available down to small blocks based on the 100 percent enumeration of housing units. They can therefore be used to classify small areas—including census tracts, as is done here—into a range of density categories for core-based statistical areas (CBSAs).3,4 By doing so, it is possible to estimate the amount of housing in each metropolitan and micropolitan area that is built at what are presumably the lowest density levels, as 3 Data on housing, measured at the census tract level, are drawn from the National Neighborhood Change Database, a special tabulation of data from the 1970 through 2000 censuses that allocates population and housing to the boundaries of the 2000 census. The 2000 census Geocorr application from the University of Missouri was used to collect data on the land area of all U.S. census tracts. Then, for each census tract, the number of housing units was divided by the total land area to yield an average (gross) density figure. Each tract in 1990 and 2000 was allocated to a density category (e.g., <1 acre, 1 to 1.9 acres), summing the number of housing units and land area in tracts in each density range. 4 A CBSA is defined as a geographic entity consisting of a county or counties associated with at least one core (urbanized area or urban cluster) of at least 10,000 in population, plus adjacent counties having a high degree of social and economic integration with the core as measured through commuting ties with the counties containing the core (OMB 2000). Metropolitan and micropolitan statistical areas are the two categories of CBSAs. A metropolitan statistical area is a CBSA associated with at least one urbanized area that has a population of at least 50,000. A micropolitan statistical area is a CBSA associated with at least one urban cluster that has a population of at least 10,000 but less than 50,000. Both metropolitan and micropolitan statistical areas make up the central county or counties containing the core, plus adjacent outlying counties having a high degree of social and economic integration with the central county as measured through commuting.

OCR for page 211
215 Analysis of Density Assumptions and Feasibility of Committee Scenarios well as at higher density levels.5 The disadvantage of census tract–level data, however, is that such data are a gross measure of available land within census tracts (i.e., do not net out land uses for purposes other than residential development), thus in all likelihood underestimating the density of development in the most extensive tracts. For this reason the census data are referred to as gross acres in this appendix. Because of these pros and cons, both data sources are used here to examine density levels in light of recent development trends and how much change from current trends the committee’s scenarios imply. Density Measures One other distinction should be made. Two methods of defining densities are possible. The conventional method (A) simply divides total DUs (or persons) by total land area to get the density of the average acre of land. The alternative method (B) is to weight developments of different densities by the number of DUs they contain to get the density of the average DU (see Box C-2 for an example). The differences and similarities between the two sources of land data and the two methods of calculat- ing density are illustrated in Tables C-1 through C-3. Table C-1 provides a summary of changes in density patterns from 1987 to 1997 and 1990 to 2000 on the basis of density figures from the NRI and the census, respectively. Tables C-2 and C-3 provide more detail on each. Method A for calculating density can be used with both the NRI and the census measures of land area, but Method B is possible only with the more spatially detailed census data. Densities are higher when calculated with land use data from the NRI than with those from the census, presumably because the urban acres reported by the NRI include less undeveloped land than do the gross acres reported by the census. For example, the average density 5 It should be noted, however, that census tracts vary in size, becoming larger at metropolitan area boundaries.

OCR for page 211
216 Driving and the Built Environment Box C-2 two methods of defining densities: an example Two methods of defining densities are used in this study—one weights by population, the other by acres. A simple example illus- trates the difference. Assume a hypothetical metropolitan area of 10 acres in which 5 people live. But 1 person lives on the periphery on a 9-acre lot, while the other 4 people live in the metropolitan center on 1⁄4-acre lots. The density to which the average acre is developed is 5 people/ 10 acres = 0.5 person per acre. However, the average person in the metropolitan area does not experience a density of 0.5 person per acre but something much higher. That average is (4⁄5 person × 4 persons per acre) + (1⁄5 person × 1 person/9 acres) = 3.20 + 0.02 = 3.22 persons per acre. The two measures differ because the lot sizes vary within the metropolitan area and because the former measure weights the density by acres, while the latter weights the density by persons. If everyone in the hypothetical metropolitan area lived on a 2-acre lot, for example, the average density would be 0.5 person per acre under either approach. Both approaches are legitimate. The population-weighted density measure provides a better sense of the density to which the average metropolitan resident is exposed rather than the density to which the average acre in the metropolitan area is developed. The former is more likely to reflect the local circumstances in which the typical person lives. In the hypothetical metropolitan area, for example, it would be misleading to say that the typical resident behaved as if he or she lived in a neighborhood of 0.5 person per acre when four-fifths of the population lives at 4 persons per acre.

OCR for page 211
217 Analysis of Density Assumptions and Feasibility of Committee Scenarios TABLE C-1 Summary of Recent Trends in the Density of Development Patterns Based on the NRI and the U.S. Census Based on NRI data, 1987–1997 (urban acres) Density of average acre (Method A) Density of average acre: 1987 1.86 DUs/acre Density of average acre: 1997 1.66 DUs/acre Density of average new acre developed: 1987–1997 0.99 DUs/acre New development as a share of 1997 average 60 percent as dense as the average Based on census data, 1990–2000 (gross acres) Density of average acre (Method A) Density of average acre: 1990 1.13 DUs/acre Density of average acre: 2000 1.11 DUs/acre Density of average new acre developed: 1990–2000 0.96 DUs/acre New development as a share of 2000 average 86 percent as dense as the average Density of average DU (Method B) Density of average DU: 1990 3.02 DUs/acre Density of average DU: 2000 2.89 DUs/acre Density of average new DU developed: 1900–2000 1.93 DUs/acre New development as a share of 2000 average 67 percent as dense as the average Note: See Tables C-2 and C-3 for details. TABLE C-2 Change in Dwelling Units by Urban Acre, 1987–1997, Based on NRI Urban Acresa Year Number of Dwelling Units (DUs) DUs/Urban Acre 1987 75,307,500 40,543,000 1.86 1997 87,412,760 52,776,100 1.66 Density of average acre, new and existing development, 1997: 1.66 Density of average new acre developed, 1987–1997: 0.99b Density of average new acre developed as a share of new and existing development in 2000: 60 percentc a Note that urban acres correspond to counties within metropolitan statistical areas and consolidated metropolitan statistical areas as defined by the U.S. census prior to 2000. Thus, they are consistently defined for the 1987–1997 time frame used here. 0.99 = (87,412,760 − 75,307,500)/(52,776,100 − 40,543,000). b (0.99/1.66) × 100 = 60 percent. c

OCR for page 211
TABLE C-3 Change in Dwelling Units by Gross Density of Census Tract, 1990–2000, U.S. Census Data for Metropolitan and Micropolitan Areas Density of Average New Acre Developed (Method A) Average Percent Total Acresa Change Range— Average Number Persons/ of Total 1990–2000 DU/gross DU/gross Persons/ square Change: acre acre acre mile 1990 2000 1990–2000 1990 2000 Percentage <1 0.6 1.50 960 62,779,845 72,862,155 10,082,310 75.5 75.7 76.5 1–1.9 1.4 3.50 2,240 10,050,874 11,914,815 1,863,941 12.1 12.4 14.1 2–2.9 2.4 6.00 3,840 4,713,174 5,494,329 781,155 5.7 5.7 5.9 3–3.9 3.5 8.75 5,600 2,460,770 2,739,475 278,705 3.0 2.8 2.1 4–7.9 5.5 13.75 8,800 2,296,743 2,443,374 146,631 2.8 2.5 1.1 12.0 30.00 19,200 798,832 832,758 33,926 1.0 0.9 0.3 8+ Total 83,100,239 96,286,906 13,186,668 100.0 100.0 100.0 Density of average acre, new and existing development, 2000 DU/gross acre: 1.11b Persons/square mile: 1,776c Density of average new acre developed: 1990–2000 DU/gross acre: 0.96b Persons/square mile: 1,536c Density of average new acre developed as a share of new and existing development in 2000: 86 percent

OCR for page 211
Density of Average New Dwelling Unit Developed (Method B) Total Number of Dwelling Units Change Range— Average Persons/ Percent of Total 1990–2000 DU/gross DU/gross square Change: acre acre miled 1990 2000 1990–2000 1990 2000 Percentage <1 0.6 960 37,667,907 43,717,293 6,049,386 40.1 41.0 47.5 1–1.9 1.4 2,240 14,071,224 15,680,741 2,609,517 15.0 15.6 20.5 2–2.9 2.4 3,840 11,311,617 13,186,390 1,874,773 12.0 12.4 14.7 3–3.9 3.5 5,600 8,612,696 9,588,164 975,468 9.2 9.0 7.7 4–7.9 5.5 8,800 12,632,087 13,438,555 806,468 13.5 12.6 6.3 12.0 19,200 9,585,988 9,993,097 407,109 10.2 9.4 3.2 8+ Total 93,881,519 106,604,240 12,722,721 100.0 100.0 100.0 Density of average DU, new and existing development, 2000 Density of average new DU developed: 1990–2000 DU/gross acre: 2.89e DU/gross acre: 1.93e Persons/square mile: 4,624f Persons/square mile: 3,083f Density of average new DU developed as a share of new and existing development in 2000: 67 percentg a The number of acres is estimated by dividing the number of DUs in each density category (see above) by the aggregate density of that category. For example, in 2000, the average number of dwelling units <1 DU/acre was 43,717,293, and the aggregate density for that category was 0.6. Thus, the DUs in that category required a total of 72,862,155 gross acres. b DU/average acre is calculated by dividing the total number of DUs in 2000 (see above) by the total number of acres in 2000. c Persons/square mile is calculated by multiplying the average DU/gross acre by 2.5 persons/household by 640. d Assumes 2.5 persons/household. There are 640 acres in 1 square mile. e The average DU/gross acre is calculated by multiplying the percentage of DUs by the average DU/gross acre for each category and dividing the total by 100. f The average number of persons/square mile is calculated by multiplying the average DU/gross acre by 2.5 persons/household by 640. g (1.93/2.89) × 100 = 67 percent.

OCR for page 211
220 Driving and the Built Environment (calculated by using Method A) was 1.66 DUs per urban acre in 1997 (when NRI data are used) but only 1.11 DUs per gross acre in 2000 (when census data are used). Densities are lower when calculated with Method A than with Method B because dense developments account for a smaller proportion of developed acres than of DUs. On the basis of the 2000 census data, for example, the average acre in a metropolitan or micropolitan area was developed to a density of 1.1 DUs per acre (Method A), while the average DU was located in a tract with a density of 2.89 DUs per acre (Method B). Densities are lower using Method A because in 2000 only 24 percent of the acres were in census tracts that were developed to densities of at least 1 DU per acre, while 59 percent of the DUs were in tracts with at least 1 DU per acre. development trends The trends in density are similar regardless of which data source or method of calculating density one uses. Over the decade of the 1990s, average densities declined because the average density of new develop- ment was lower than the average density of the existing development.6 On the basis of the NRI data, the density of the average acre (Method A) fell from 1.86 DUs per acre in 1987 to 1.66 DUs per acre in 1997 because the new development between 1987 and 1997 was built to a density of only 0.99 DUs per acre. On the basis of the census data, the density of the average acre fell less sharply, from 1.13 DUs per acre in 1990 to 1.11 in 2000 because the density of new development was 0.96 DUs per acre, only slightly less than the density of the existing development in 1990. The more detailed census data reveal some of the patterns behind these averages. The density of the average acre (Method A) declined between 1990 and 2000 because more than three-quarters (76.5 percent) of acres developed in this period were in tracts with 6 The NRI data measure changes from 1987 to 1997, the most recent year for which such detailed data are available.

OCR for page 211
221 Analysis of Density Assumptions and Feasibility of Committee Scenarios below 1 DU per gross acre (see Table C-3). The density of the average DU (Method B) declined slightly less sharply because only 47.5 percent of the new DUs developed during this period were built in census tracts with below 1 DU per gross acre and another 20 percent in tracts with between 1 and 2 DUs per gross acre. Less than one-third (31.9 percent) of new DUs were built in tracts having 2 or more DUs per gross acre (see Table C-3). In short, under either method, the majority of new development occurred in the lowest density categories. comparison of trends and scenarios The committee’s two scenarios assume that densities are doubled for 25 percent (Scenario 1) and 75 percent (Scenario 2) of all new residential housing built and replaced from 2000 to 2030 and to 2050. How difficult would it be to achieve these densities? The committee addressed this question by using the density data described previously to project current trends toward low-density development forward to 2050 and then comparing the higher densities assumed in the two scenarios with the baseline. Although the same calculations were conducted for 2030 as well, only the 2050 projections are presented here because the durability of the housing stock means that decades are required for the projected changes in development patterns to take hold. The committee recognizes that the location of more compact development within a metropolitan area is likely to affect the reduction in vehicle miles traveled (VMT). In the committee’s simple projections reported in Chapter 5, however, the VMT estimates are sensitive only to the average density of urban areas, not the spatial distribution of that density within urban areas. Thus while the committee uses the doubling of density in 25 or 75 percent of new development as a shorthand for describing its two scenarios, there are many ways in which an equivalent change in average urban density could be achieved. The committee assumed that much of the new development would occur at the urban fringe where developable land is available. But

OCR for page 211
222 Driving and the Built Environment some amount of infill is also likely, particularly in Scenario 2, which assumes that a large share (75 percent) of new development is at higher densities. Calculations Using NRI Data On the basis of the data from the NRI, which pick up lower densities at the urban fringe and leapfrogged development, if all new and replacement housing were to be built at the current average density of new development (the base case scenario)—0.99 DUs per urban acre— the average density of the housing stock would fall from 1.66 DUs per urban acre in 2000 to 1.29 to 1.39 DUs per urban acre by 2050 (see Table C-4). If 25 percent were developed at double the density (1.98 DUs per urban acre) (Scenario 1), average densities would decline from 1.66 to only between 1.43 and 1.49 DUs per urban acre, a 7 to 11 percent increase from the 2050 baseline. Average densities still decline in this scenario because even with the doubling of density for 25 percent of new development, the average density of new development from 2000 to 2050 (1.24 DUs per urban acre)7 is still below the average density of existing development in 2000 (1.66 DUs per urban acre). In the upper-bound scenario (Scenario 2), which assumes 75 percent of new and replacement housing is built at these higher densities, the average density in 2050 would actually increase from the 2000 average of 1.66 DUs per urban acre to between 1.69 and 1.7 DUs per urban acre, a 22 to 32 percent increase from the 2050 baseline (see Table C-4). Average densities would be higher in 2050 than they were in 2000 because the doubling of density for 75 percent of new development would increase the average density of new development (1.73 DUs per urban acre)8 above the average density of existing development in 2000 (1.66 DUs per urban acre). (1.98 × 0.25) + (0.99 × 0.75) = 1.24. 7 (1.98 × 0.75) + (0.99 × 0.25) = 1.73. 8

OCR for page 211
TABLE C-4 Changes in Density to Meet Targets of Committee Scenarios by 2050 NRI Data Census Data (Method B Only)a Density Percent Density Percent Level Change Level Change Number of Units (DU/urban from Number of Units (DU/gross from (millions) acre) Base Case (millions) acre) Base Case Base case Base case 86.64–90.35 1.66 Original housing 86.64–90.35 2.9 Original housing stock in 2050 stock in 2050 62.44–105.37 0.99 New and replacement 62.44–105.37 1.9 New and replacement units units 152.79–192.01b 1.29–1.39 — Total in 2050 152.79–192.01b 2.35–2.49 — Total in 2050 Scenario 1 (25 percent) Scenario 1 (25 percent) 86.64–90.35 1.66 Original housing 86.64–90.35 2.9 Original housing stock in 2050 stock in 2050 15.61–26.34 1.98c New, more compact 15.61–26.34 3.8c New, more compact 46.83–79.02 0.99 New, not more 46.83–79.02 1.9 New, not more compact compact 152.79–192.01 1.43–1.49 7.2–10.8 Total in 2050 152.79–192.01 2.61–2.69 8–11.1 Total in 2050 (continued on next page)

OCR for page 211
TABLE C-4 (continued) Changes in Density to Meet Targets of Committee Scenarios by 2050 NRI Data Census Data (Method B Only)a Density Percent Density Percent Level Change Level Change Number of Units (DU/urban from Number of Units (DU/gross from (millions) acre) Base Case (millions) acre) Base Case Scenario 2 (75 percent) Scenario 2 (75 percent) Original housing 86.64–90.35 1.66 Original housing 86.64–90.35 2.9 stock in 2050 stock in 2050 New, more compact 46.83–79.02 1.98c New, more compact 46.83–79.02 3.8c New, not more 15.61–26.34 0.99 New, not more 15.61–26.34 1.9 compact compact Total in 2050 152.79–192.01 1.69–1.70 21.6–31.8 Total in 2050 152.79–192.01 3.07–3.13 23.3–33.2 Note: Scenario 1 assumes 25 percent of all new and replacement housing is built more compactly. Scenario 2 assumes 75 percent of all new and replacement housing is built more compactly. a Only Method B was selected to illustrate the scenario results using census data because the population-weighted density measure better reflects how the average resident experiences density (see Box C-2). b Assumes the same range of housing units projected in Chapter 4, Table 4-3. c Assumes density of new development is doubled (0.99 × 2 = 1.98 for NRI data; 1.9 × 2 = 3.8 for census data).

OCR for page 211
225 Analysis of Density Assumptions and Feasibility of Committee Scenarios Calculations Using Census Data According to the census data, if all new and replacement housing were to be built at the current average density of new development (the base case scenario)—1.9 DUs per gross acre—the average density of the housing stock would fall from 2.9 DUs per gross acre in 2000 to 2.35 to 2.49 DUs per gross acre by 2050 (see Table C-4). If 25 percent were developed at double the density (3.8 DUs per gross acre) (Scenario 1), the average density of new development (2.38 DUs per gross acre)9 would still be slightly below the density of the development that existed in 2000. As a result, average densities would fall slightly from 2.9 DUs per gross acre in 2000 to between 2.61 and 2.69 DUs per gross acre in 2050, an 8 to 11 percent increase from the 2050 baseline. In the upper-bound scenario (Scenario 2), which assumes 75 percent of new and replacement housing is built at these higher densities, new development is denser (3.33 DUs per acre)10 than existing development, so average density increases from 2.9 DUs per gross acre in 2000 to between 3.07 and 3.13 DUs per gross acre, a 23 to 33 percent increase from the 2050 baseline. As with the NRI data, the census data show that doubling density in 25 percent of new development is equivalent to assuming that the average new development is slightly less dense than the average existing development. In contrast, doubling density in 75 percent of new development is comparable with assuming that the average new development is moderately more dense than the average existing development. Seen in this way, both scenarios are substantial departures from existing patterns in that new development has historically been sig- nificantly less dense than existing development. In some metropolitan areas, however, even a growing area such as Phoenix, densities for new development are similar to or slightly higher than densities for existing development, so the 25 percent doubling scenario is possible. But as (3.8 × 0.25) + (1.9 × 0.75) = 2.38. 9 (3.8 × 0.75) + (1.9 × 0.25) = 3.33. 10

OCR for page 211
226 Driving and the Built Environment discussed in Chapter 5, Phoenix has unusual characteristics (e.g., bounded by desert, limits on development), so that the 25 percent doubling scenario for all metropolitan areas would require substantial policy intervention. A 75 percent doubling scenario would require even greater policy intervention and public acceptance because it assumes that new developments are denser than existing developments, and, if compared with a continuing trend of new, low-density development to 2050, significantly more dense than the 2050 baseline. The more detailed census data allow a test of the plausibility of the two scenarios by showing how much the different density categories would have to change to reach the target densities implicit in the scenarios. Meeting the target of Scenario 1—doubling density in 25 percent of new development—could be achieved (or nearly so) in a variety of ways. For example, one way would be to eliminate half of all new development in tracts in the lowest density category (less than 1 DU per gross acre), substantially reducing development in many urban fringe areas where lots are very large (see Strategy B in Table C-5). Alternatively, one could follow an infill strategy and double the densities in tracts with an average of 4 or more DUs per acre (see Strategy C in Table C-5). Reaching the target of Scenario 2—doubling density in 75 percent of new development—would require much more drastic measures, for example, eliminating all new DUs in tracts where development currently averages less than 1 DU per acre (see Strategy A in Table C-5) or doubling the density of tracts with an average of 3 or more DUs per acre, which would be a much more aggressive infill scenario (see Strategy D in Table C-5). Note that the comparisons made here are based on measures of the density of the average DU (Method B). Using an alternative way of measuring density (Method A)—by the density of the average developed acre rather than the density of the average DU—changes somewhat the ease (or difficulty) of the various ways of achieving the target densities. For example, infill scenarios are less effective. Even the aggressive infill scenario—doubling the den- sity of tracts with 3 or more DUs per acre—does not achieve either the

OCR for page 211
227 Analysis of Density Assumptions and Feasibility of Committee Scenarios TABLE C-5 Alternative Strategies for Achieving Higher Densities for New Development Implied in Committee Scenario Targets Baseline DUs/Acre Average Change in DUs, Percent Change in Acres, Percent (range) DU/Acre 1990–2000 of Total 1990–2000 of Total <1 0.6 6,049,386 47.5 10,082,310 76.5 1–1.9 1.4 2,609,517 20.5 1,863,941 14.1 2–2.9 2.4 1,874,773 14.7 781,155 5.9 3–3.9 3.5 975,468 7.7 278,705 2.1 4–7.9 5.5 806,468 6.3 146,631 1.1 8+ 12.0 407,109 3.2 33,926 0.3 Total 12,722,721 100.0 13,186,668 100.0 Strategy A: Eliminate All New DUs in Lowest-Density Tracts, Redistribute Proportionately DUs/Acre (range) Change in DUs Percent of Total Change in Acres Percent of Total <1 0 0 0 0 1–1.9 4,975,047 39 3,553,605 60 2–2.9 3,574,257 28 1,489,274 25 3–3.9 1,859,731 15 531,352 9 4–7.9 1,537,532 12 279,551 5 8+ 776,154 6 64,679 1 Total 12,722,721 100 5,918,461 100 Persons/Acrea Persons/Square Mileb DUs/Acre Density of average acre (Method A) Baseline 0.96 2.40 1,536 Strategy A 2.15 5.37 3,440 Percent change 129 129 129 25 percent higher density target 1.2 3.0 1,920 75 percent higher density target 1.68 4.2 2,688 Density of average DU (Method B) Baseline 1.93 4.82 3,085 Strategy A 3.13 7.82 5,008 Percent change 62 62 62 25 percent higher density target 2.41 6.02 3,856 75 percent higher density target 3.38 8.43 5,399 (continued on next page)

OCR for page 211
228 Driving and the Built Environment TABLE C-5 (continued) Alternative Strategies for Achieving Higher Densities for New Development Implied in Committee Scenario Targets Strategy B: Reduce by One-Half New DUs in Lowest-Density Tracts, Redistribute Proportionately DUs/Acre (range) Change in DUs Percent of Total Change in Acres Percent of Total <1 3,024,693 24 5,042,155 53 1–1.9 3,792,282 30 2,708,773 28 2–2.9 2,724,515 21 1,135,215 12 3–3.9 1,417,599 11 405,028 4 4–7.9 1,172,000 9 213,091 2 8+ 591,631 5 49,303 1 Total 12,722,721 100 9,552,564 100 Persons/Acrea Persons/Square Mileb DUs/Acre Density of average acre (Method A) Baseline 0.96 2.40 1,536 Strategy B 1.33 3.33 2,128 Percent change 38 38 38 25 percent higher density target 1.2 3.0 1,920 75 percent higher density target 1.68 4.2 2,688 Density of average DU (Method B) Baseline 1.93 4.82 3,085 Strategy B 2.39 5.97 3,721 Percent change 24 24 24 25 percent higher density target 2.41 6.02 3,856 75 percent higher density target 3.38 8.43 5,399 Strategy C: Double New DUs in Highest-Density Tracts (>4 DU/Acre), Redistribute Proportionately DUs/Acre (range) Change in DUs Percent of Total Change in Acres Percent of Total <1 5,411,511 43 9,019,185 75 1–1.9 2,334,358 18 1,667,398 14 2–2.9 1,667,088 13 698,787 6 3–3.9 872,610 7 249,317 2 4–7.9 1,612,936 13 293,261 2 8+ 814,218 6 67,852 1 Total 12,722,721 100 11,995,800 100

OCR for page 211
229 Analysis of Density Assumptions and Feasibility of Committee Scenarios Persons/Acrea Persons/Square Mileb DUs/Acre Density of average acre (Method A) Baseline 0.96 2.40 1,536 Strategy C 1.06 2.65 1,696 Percent change 10 10 10 25 percent higher density target 1.2 3.0 1,920 75 percent higher density target 1.68 4.2 2,688 Density of average DU (Method B) Baseline 1.93 4.82 3,085 Strategy C 2.28 5.70 3,648 Percent change 18 18 18 25 percent higher density target 2.41 6.02 3,856 75 percent higher density target 3.38 8.43 5,399 Strategy D: Double DUs in Tracts Averaging >3 DUs/Acre, Redistribute Proportionately DUs/Acre (range) Change in DUs Percent of Total Change in Acres Percent of Total <1 4,792,239 38 7,987,065 73 1–1.9 2,067,223 16 1,476,588 13 2–2.9 1,485,169 12 618,820 6 3–3.9 1,950,936 15 557,410 5 4–7.9 1,612,936 13 293,261 3 8+ 814,218 6 67,852 1 Total 12,722,721 100 11, 000,996 100 Persons/Acrea Persons/Square Mileb DUs/Acre Density of average acre (Method A) Baseline 0.96 2.40 1,536 Strategy D 1.16 2.90 1,856 Percent change 21 21 21 25 percent higher density target 1.2 3.0 1,920 75 percent higher density target 1.68 4.2 2,688 Density of average DU (Method B) Baseline 1.93 4.82 3,085 Strategy D 2.51 6.27 4,016 Percent change 30 30 30 25 percent higher density target 2.41 6.02 3,856 75 percent higher density target 3.38 8.43 5,399 a Persons/acre equals number of DUs/acre times 2.5 persons/household. b Persons/square mile equals persons/acre multiplied by 640 acres.

OCR for page 211
230 Driving and the Built Environment 25 percent or the 75 percent target because doing so prevents much less land from being used for development than curtailing development in the lowest-density tracts where the average acres per DU are largest. Pursuing the latter strategy—eliminating all new development in tracts with densities below 1 DU per acre—results in achieving not only the 25 percent but also the 75 percent target. The reason is the large amount of acreage in the below-1-acre category (see the baseline in Table C-5). Because such large lots represented approximately 75 percent of the residential land developed between 1990 and 2000, this strategy would require an almost complete reversal of recent development patterns. Given the large amount of land available for development and lax land use policies in more rural counties, achieving such a goal would require extraordinary changes in land use policy and market trends. The analyses provided here suggest that both the 25 and the 75 per- cent targets represent a significant departure from recent trends, which have involved lower densities than the average for new devel- opment over the decade of the 1990s and for decades before that. Doubling the density of 25 percent of new development by 2050 will not raise densities above the current average and will raise them only about 7 to 11 percent above the 2050 baseline. In addition, precedents for such changes in density can be found even in growing areas such as Phoenix. Nevertheless, Phoenix is not typical of many growing metropolitan areas, and meeting the 25 percent target will require a trend change. Doubling the density of 75 percent of new development by 2050 will require densities above those of existing developments, significantly above (approximately 20 to 33 percent) the 2050 baseline. The committee disagreed about the feasibility of doubling the density of 75 percent of new development, even by 2050. Those members who thought it was possible questioned whether densities will keep declining. The combination of macroeconomic trends—likely higher energy prices and carbon taxes—in combination with growing public support for strategic infill, investments in transit, and higher densities

OCR for page 211
231 Analysis of Density Assumptions and Feasibility of Committee Scenarios along rail corridors, could result in considerably higher densities by 2050. Other members believed that the substantial infill or curbing of large-lot development at the urban fringe, or both, implied in achieving the targets in Scenario 2 would require such a significant departure from current housing trends, land use policies of jurisdictions on the urban fringe, and public preferences that they would be unrealistic absent a strong state or regional role in growth management. references Abbreviations NRCS Natural Resources Conservation Service OMB Office of Management and Budget Downs, A. 2004. Still Stuck in Traffic: Coping with Peak-Hour Traffic Congestion. Brookings Institution, Washington, D.C. NRCS. 2002. National Resources Inventory, 2002 and 2003 Annual NRI, Glossary of Key Terms. U.S. Department of Agriculture, Washington, D.C. www.nrcs.usda.gov/ technical/NRI/2002/glossary.html. Accessed Aug. 28, 2008. NRCS. 2003. National Resources Inventory, 2001 Annual NRI, Urbanization and Development of Rural Land. U.S. Department of Agriculture, Washington, D.C., July. www.nrcs.usda.gov/technical/NRI/2001/urban.pdf. OMB. 2000. Standards for Defining Metropolitan and Micropolitan Statistical Areas. Federal Register, Vol. 65, No. 249, pp. 82228–82238.