کد مقاله | کد نشریه | سال انتشار | مقاله انگلیسی | نسخه تمام متن |
---|---|---|---|---|
1049912 | 945647 | 2009 | 6 صفحه PDF | دانلود رایگان |

A wide variety of metrics is used to quantify features of urbanization in ecological studies. Selecting statistically independent measures of urbanization depends upon the nature (multivariate collinearity) and strength (correlation coefficient) of correlations between urban metrics. We evaluated the influence of landscape extent and habitat context, factors that commonly differ between studies, on correlations between urban metrics. We examined the nature and strength of relationships between urban metrics at 1105 sites within Massachusetts, USA, including: population, agriculture cover, forest cover, wetland cover, dense residential cover, impervious surface cover, road length and greenspace cover. At each site, values were measured at five extents: 100 m, 250 m, 500 m, 1 km, and 2 km radii buffers. We also investigated the influence of habitat context on correlations by measuring values with a 1 km radius buffer for 100 sites within each of three habitat contexts (salt marsh, forest, and freshwater marsh). Principal component analysis showed that spatial scale did not affect the nature of relationships, but habitat context did. The average strength of bivariate correlations significantly increased at larger extents, and was significantly lower in salt marsh habitat context. Our results indicate that no single set of urbanization metrics is universally applicable and underline the importance of using a suite of statistical techniques to characterize independent aspects of urban environments.
Journal: Landscape and Urban Planning - Volume 92, Issue 1, 15 August 2009, Pages 47–52