Urban growth reduces open space in and around cities, impacting biodiversity and ecosystem services. Using land-cover and population data, we examined land consumption and open space loss between 1990 and 2000 for all 274 metropolitan areas in the contiguous United States. Nationally, 1.4 million ha of open space was lost, and the amount lost in a given city was correlated with population growth (r(272) = 0.85, P<0.001). In 2000, cities varied in per capita land consumption by an order of magnitude, from 459 m2/person in New York to 5393 m2/person in Grand Forks, ND. The per capita land consum Using land-cover and population data, we examined land consumption and open space loss between 1990 and 2000 for all 274 metropolitan areas in the contiguous United States. Nationally, 1.4 million ha of open space was lost, and the amount lost in a given city was correlated with population growth (r(272) = 0.85, P<0.001). In 2000, cities varied in per capita land consumption by an order of magnitude, from 459 m 2 /person in New York to 5393 m 2 /person in Grand Forks, ND. The per capita land consumption (m 2 /person) of most cities decreased on average over the decade from 1,564 to 1,454 m 2 /person, but there was substantial regional variation and some cities even increased. Introduction Urban population in the United States has climbed dramatically in recent decades, from 84.5 million in 1950 to 226 million in 2000 [1] . At the same time, many of these new residents settled in suburbs farther from the city center at relatively low density [2] , resulting in a drastic expansion in urbanized area [3] , [4] . The expansion of urban area can have significant ecological impacts [5] – [7] . The amount of open space is reduced [8] , fragmenting natural habitat as well as reducing the recreational and other amenities people can enjoy from open space [9] – [11] . Open space is defined here as agricultural land and more natural land-cover such as forest and grassland, including both remnant patches within a city as well as larger patches at the city's fringe. Rapid urban growth has raised public concern about the ecological and social impacts of “sprawl,” which has been defined as new settlements relatively far from city centers, as opposed to “densification,” which raises the density of already developed neighborhoods [2] . One manifestation of this concern is the rapid growth of ballot initiatives to authorize bonds to fund the protection of open space. In this study we quantify the total amount of open space lost to development for US cities, as well as the developed area per capita, which we refer to as “per capita land consumption.” Moreover, most studies that have compared the density of development in different cities have used city-wide average, rather than tracking the full spectrum of development types within an urban area. Of particular interest is to what extent a small fraction of households consume large amounts of developed area while the rest of the population is concentrated in dense settlements that use much less land per capita. We quantify land inequality using the Gini index (G), which varies from 0 (most equal) to 1 (most unequal). Measurement of this “land inequality” quantifies the heterogeneity of complex urban landscapes. Note that per capita land consumption and land inequality are independent quantities, and it is possible for a city with high per capita land consumption to have either low or high levels of land inequality. In this paper we measure the open space lost from urban growth in all 274 metropolitan statistical areas (MSAs) in the contiguous United States from 1990 to 2000 ( Figure 1 ). We present our results in two steps, first discussing MSA-wide metrics that summarize some aspect of pattern into a single number for each MSA and then discussing fine-scale metrics that quantify variation in pattern within each MSA. Figure 1 Metropolitan areas examined in this study. For the largest metropolitan areas, the type of zoning scheme is shown (see text for details). Inset shows land-cover data for the Los Angeles area, with areas built up in 1990 (gray), areas of new development between 1990 and 2000 (red), and Results MSA-Wide Analysis Most MSAs in the United States increased in population from 1990 to 2000 ( Figure 2A ), with the greatest increase in southern California, the New York MSA, and the Atlanta MSA. Cities in the “Rust Belt,” running from upstate New York through western Pennsylvania and to Ohio, lost population. Land-cover in 1990 varied greatly across the country ( Figure 2B ), with the Midwest dominated by agriculture and the West dominated by grassland/shrublands. In contrast, the eastern U.S. and the Pacific Northwest are dominated by forests, while wetlands occur at greatest extent in Florida and the Gulf coast. The only significant contrast is between the “Reform” and “Traditional” zoning categories, with “Reform” cities slightly increasing in the proportion of houses in high-density neighborhoods while “Traditional” cities decreased. The change in proportion of houses in high density neighborhoods was not correlated with the median house value, the city size, or the degree of conservation funding. Land inequality Neighborhoods containing a relatively small proportion of urban residents at low density usurp most of the developed area in a city. In our study, we have classified an entire MSA as one of the four categories, using the category of the majority of the urban area in the Pendall et al. study. For example, Pendall et al. classified most of the New York MSA as “Traditional,” but classified the New Jersey suburbs as “Exclusion.” In our study, the entire New York MSA is classified as “Traditional.” This reduces the number of MSAs in our study with a zoning category to 49. MSA-Wide Analysis All data was clipped to MSA boundaries and projected to an Albers Equal-Area projection. For each MSA we tabulated open space loss and population change.