Research ArticleENVIRONMENTAL STUDIES

Two centuries of settlement and urban development in the United States

See allHide authors and affiliations

Science Advances  03 Jun 2020:
Vol. 6, no. 23, eaba2937
DOI: 10.1126/sciadv.aba2937
  • Fig. 1 Maps of the FBUY at different levels of granularity.

    The maps depict national-, regional-, and local-scale processes of human development: (A) county-level FBUY within contemporary (Census 2010) boundaries used as constant units of analysis over time (counties where no built-up year is available are shown in gray); more detailed distributions of FBUY for the states of Colorado, Kansas, and Ohio within (B) county boundaries, (C) 2500-m grid cells, and (D) 250-m grid cells, respectively. (E) A detailed depiction of the 250-m resolution data for the cities of Denver, CO; Wichita Falls, KS; and Columbus, OH.

  • Fig. 2 HISDAC-US settlement layers derived from the ZTRAX data at fine granularity for different points in time.

    The layers are shown at 250-m spatial resolution for different points in time, 1810–2010, for Rockingham County, NH and surroundings, including: (A) FBUY layer in which raster cells are assigned the earliest built year recorded, and a time series of the number of built-up property records (BUILD) located within a raster cell in (B) 1810, (C) 1860, (D) 1910, (E) 1960, and (F) 2010. County boundaries of the 2010 census are shown in black.

  • Fig. 3 Settlement trends and multitemporal distributions for San Francisco, Atlanta, and Boston.

    (A) Time series of densification and expansion, calculated over 15-year time increments computed within metropolitan statistical area boundaries of 2010. (B) New built-up grid cells (indicated by black grid cells) during the given time periods. (C) Change in BUI (i.e., the sum of building indoor area per grid cell) during the given time periods, with warmer colors indicating greater change.

  • Fig. 4 Settlement trends describing different types of growth in rural and urban strata.

    Boxplots of semi-decadal distributions of (A) densification and (B) absolute expansion in rural and urban counties. (C) Graphic display of relative locations of newly and previously built-up grid cells to calculate midrange expansion, internal, and peripheral growth within the Greater Washington, DC area including Arlington, Bethesda, and Georgetown. (D) Trends of building indoor area (BIA) in urban and rural strata (counties), each broken down into whether the increase happened in newly built-up cells (midrange expansion), in previously built-up cells at the edge of larger BUAs (peripheral growth), or in previously built-up cells in inner parts of BUAs (internal growth). (E) Proportion of internal, peripheral, and internal-peripheral combined growth (i.e., growth in previously built-up cells) in relation to overall change for the two strata (rural and urban).

  • Fig. 5 Correlation measures between county-level population counts and settlement variables.

    The different plots show correlations over the time period from 1810 to 2010 between population and (A) the number of built-up property records (BUILD), (B) BUI, and (C) BUA. avg. corr., average correlation. In (D), correlation measures between county-level population change and absolute expansion (Abs. expansion) are shown for the same time period. Population counts are enumerated within historical county boundaries, while settlement measures (all land use classes) are calculated within contemporary county boundaries. Correlation measures are shown for various levels of temporal county boundary stability (e.g., the blue lines represent only counties whose area did not change more than 10% over time) to demonstrate the importance of compatible spatial units in spatiotemporal analysis. Average correlation coefficients are shown over all years in parentheses.

  • Table 1 Panel analysis results using different points in time, 1860–2010.

    SEs are given in parentheses. Statistical significance is provided (*P < 0.05, **P < 0.01, and ***P < 0.001). The different models tested are OLS, least squares dummy variables (LSDVs), and generalized least squares (GLS). LSDV and GLS are “within” estimators. BUILD is based on all land-use types; results for residential land-use types only are very similar.

    OLSOLSLSDVGLSGLS
    (1)(2)(3)(4)(5)
    Total populationTotal populationTotal
    population
    Total populationTotal population
    Number of built-up
    property records
    (BUILD)
    2.680***2.668***2.190***2.190***2.246***
    SE(0.007)(0.012)(0.058)(0.013)(0.014)
    Constant10503.8***13767.2***22513.3***22513.3***16123.0***
    SE(550.069)(815.994)(1068.970)(701.136)(1274.106)
    N109965900590059005900
    R20.9260.8980.9460.8640.873
    Adjusted R20.9260.8980.9290.8190.831
    County sampleAllConsistentConsistentConsistentConsistent
    Fixed effectsCountyCountyCounty, decade
    Clustered SENoNoYesYesYes

Supplementary Materials

  • Supplementary Materials

    Two centuries of settlement and urban development in the United States

    Stefan Leyk, Johannes H. Uhl, Dylan S. Connor, Anna E. Braswell, Nathan Mietkiewicz, Jennifer K. Balch, Myron Gutmann

    Download Supplement

    The PDF file includes:

    • Figs. S1 to S6
    • Tables S1 and S2
    • Legends for movies S1 and S2
    • References

    Other Supplementary Material for this manuscript includes the following:

    Files in this Data Supplement:

Stay Connected to Science Advances

Navigate This Article