RESULTS AND DISCUSSION
For this assessment we were not able to find acceptable imagery for the entire State of Minnesota for any given
year, but were able to acquire clear Landsat imagery state-wide for the ~1990 and ~2000 time periods when imagery
from two or three years were used. In remote areas where lake data is limited we improved our results by using two
or more images from the same path and date. By using paths of connected images we were able to find enough SDT
data points to calibrate the imagery while also increasing the range of available data. In this and previous studies, we
found R2 = ~ 0.8 and standard error = ~ 0.3 for images within the index period (July 15 – September 15). Figure 3
shows predicted versus observed TSI(SDT) for Landsat Path 29 Rows 26-28, August 28, 2001 images with N = 124
and R2 = 0.89, and Figure 4 presents predicted versus observed TSI(SDT) for Landsat Path 27 Rows 26-30,
September 12, 2000 images with N = 229 and R2 = 0.82. The differences in R2 appear to be due to differences in
water clarity in the different regions of the State. Path 29 does not extend into the extremely eutrophic lakes of
southern Minnesota and does not have a representation of data from lakes higher than around 65 to 70 TSI. Path 27
extends throughout the entire State and has a larger representation of lakes in southern Minnesota with TSI values
higher than 65 to 70. The data from lakes with higher TSI values tend to have more scatter, probably because of
sampling inconsistencies and more temporal and in-lake variation of eutrophic lakes. Because TSI is a logarithmic
transformation of SDT, small errors in measurements of less than 0.5 feet have a much larger effect in more
eutrophic lakes than in clearer lakes.
To date we have completed draft copies of the ~ 1990 and ~2000 Minnesota lake water clarity atlases. These
atlases have water clarity data for over 10,000 lakes and represent the lake water clarity of the State of Minnesota for
the ~1990 and ~2000 time periods. The ~2000 Minnesota water clarity map is presented as Figure 5. This map
indicate that lake water clarity is relatively poor in southern and southwestern Minnesota and becomes progressively
clearer toward northeastern Minnesota. We are using these water quality atlases to assess spatial and temporal
patterns in lake water clarity based on surrounding land use and cover by using a GIS to link the lake clarity data
with land use features.