PPT Slide
We now have a formalized framework for a water-quality
monitoring effort to support efficient statistical assessment of
relative contributions of pollutant runoff from different
land-use/land cover categories:
Let Yi represent the pollutant concentration recorded at sampling
station i, and let Xi1 ... XiM represent the percents of total land
area draining to sampling station i that are in each land-use
category 1...M. Then the regression
Y = B1X1 + ... + BMXM + e
will yield empirical estimates of average loading coefficients
B1...BM for each land-use category. The sampling framework
minimizes multicollinearity between the X’s.