Meeting Abstract

76.3  Tuesday, Jan. 6  Urbanization measures are not interchangeable: effects of spatial scale and habitat context TAVERNIA, B.G.*; REED, J.M.; Tufts University; Tufts University brian.tavernia@tufts.edu

A wide variety of metrics is used to quantify features of urbanization in ecological studies. Successful integration of results across studies using different metrics, however, requires a strong relationship among the metrics. Additionally, it is important that the relationships remain consistent across spatial scales of biological interest and across habitat matrix types. We examined the strength and nature of relationships between eight urban metrics at 1105 sites, including: population, agriculture cover, forest cover, wetland cover, dense residential cover, impervious surface cover, road length and green space cover. Values were measured at five spatial scales (100 m, 250 m, 500 m, 1 km, 2 km) and at one spatial scale (1km) for 100 urban sites within each of three habitat context (salt marsh, forest, freshwater marsh) within Massachusetts, USA. We found generally weak correlations between urbanization measures, with only 26/140 mean correlation coefficients exceeding 0.70, and 70/140 <0.30. Spatial scale did not significantly affect the strength of correlations, but habitat context did, with lower average values in salt marsh habitat. Similarly, principal components analysis showed that spatial scale did not affect the nature (relative relationships among variables) of the relationships between urbanization measures, but habitat context did. Our results show that in our study area no single metric adequately characterizes urban settings, making multiple measures and multivariate statistically approaches an invaluable tool in assessing the influence of urbanization on ecological phenomena.