Urbanization driven convergence of surface water quantity and distribution

Background/Question/Methods

Hydrography, the spatial distribution of surface water, is a first-order property of landscapes.  In this study, we use hydrographic patterns at the continental United States scale to test 1) the influence of land cover change on surface water abundance and 2) the more general urban convergence hypothesis that predicts urbanization homogenizes landscape structure across cities compared to their native landscapes.  We compared the coverage, number, size, and type of water bodies in urban landscapes to agricultural and natural land in 100 metropolitan statistical areas (MSA) with a range of population sizes and occupying 20 eco-regions across the continental US.   Using the 2006 National Land Cover Data, we classified the land cover of each MSA as ‘natural’, agricultural, and urban open, low, medium, and high density.  Using water body coverage from the National Hydrography Database, we calculated the proportion of area covered by water, the density of water bodies, the median size, and the size frequency distribution (described by the exponent β) within each land cover type.  The convergence hypothesis predicts that variance will be lower in urban lands, and that urban lands will have more surface water than wildlands in dry regions, but less surface water in wetter regions. 

Results/Conclusions

Lakes and ponds were the dominant water body type within all land covers, but urban lands had more reservoirs and fewer intermittent water bodies than natural lands.  Areal coverage by surface water was 47%, 64%, and 74% lower in low, medium and high density urban land, respectively, than in natural and agricultural land, which had comparable surface water distributions.  Numerical density of water bodies showed similar patterns.  However, median size and β were higher in urban lands, indicating lower relative abundance of small water bodies.  The preferential loss of small water bodies suggests human modification thru elimination or building, as opposed to selection, is the cause of the observed differences.   Coefficients of variation for all parameters were also lower for urban land covers.  Differences between urban and wildland water body cover and density were strongly negatively correlated with wildland hydrography, but these differences were positive (i.e., urban>wildland) in areas with low wildland surface water cover.  These results support the urban convergence hypothesis and illustrate the scope and magnitude of urban hydrographic change.