Neighborhood proportions

For both the independent variables, we reclassified the categorical values (land use, n = 4; SES, n = 3) into separate binary surfaces, where “1” indicated presence and “0” indicated absence. Then, we applied a neighborhood moving-window analysis, called focal statistics. Calculating the mean statistic within a 2500-m radius on the binary surfaces resulted in proportions. We assigned the proportions of land use and SES to the centroids of the hexagons that resulted from the EHSA for each SGA/LBWT. The 2500-m neighborhood estimated the proportions of each land use or SES class within the distance defined for the EHSA described above.