PARENT SESSION
Oral Session #65: Aquatic Ecology: Stream and Lake.
Presiding: C. Osenberg
Wednesday, August 7. 1:00 PM to 4:45 PM. Coconino Meeting Room, TCC.
Assessing regional lake water clarity using Landsat: the role of inter-lake variability.
Nelson, Stacy*,1, Soranno, Patricia1, Cheruvelil, Kendra1, Batzli, Sam2, Skole, David2, 1 Michigan State University, East Lansing, MI2 Department of Geography and Basic Science and Remote Sensing Institute, East Lansing, MI
ABSTRACT- Remote sensing has been effectively used to measure water clarity in several single-lake studies. However, applying these approaches to large numbers of lakes within a large geographic region has been challenging. Our objectives were to: (1) develop a model to predict lake water clarity from Landsat data using 93 calibration lakes across the state of Michigan, and (2) examine how the distribution of water clarity across the 93 lakes, as measured by Secchi disk transparency (SDT), influences the model. Our regression model of field-collected SDT data against Landsat-7 ETM+ data resulted in an r2 of 0.43 (p <0.001), which was substantially lower than many previous single-lake studies. To examine the role of lake SDT distribution, we simulated a calibration dataset with a different SDT distribution. The percent of lakes that had a SDT value less than 1.5 m was increased from 8.6% (original dataset) to 47% (simulated dataset). The regression model for the simulated dataset resulted in an r2 of 0.82 (p <0.001). Our results show that Landsat can be used to measure water clarity across a large number of lakes with a wide range of SDT values. However, the regression models are sensitive to the distribution of SDT values used in the calibrated dataset, and must be taken into account when developing regional models to predict lake water clarity using remote sensing.
KEY WORDS: Lakes, Water Quality, Remote Sensing, Secchi disk depth