• McCullough, Ian M.
  University of Maine Graduate School
  

Water clarity is an ideal metric of regional water quality because clarity can be accurately and efficiently estimated remotely on a landscape scale. Remote sensing of water quality is useful in regions containing numerous lakes that are prohibitively expensive to monitor regularly using traditional field methods. Field-assessed lakes generally are easily accessible and may represent a spatially irregular, non-random sample. Remote sensing provides a more complete spatial perspective of regional water quality than existing, interest-based sampling;however, field sampling accomplished under existing monitoring programs can be used to calibrate accurate remote water clarity estimation models. We developed a remote monitoring procedure for clarity of Maine lakes using Landsat Thematic Mapper (TM) and Moderate-Resolution Imaging Spectroradiometer (MODIS) satellite imagery. Similar Landsat-based procedures have  been implemented for Minnesota and Wisconsin lakes, however, we modified existing methods by incorporating physical lake variables and landscape characteristics that affect water clarity on a landscape scale. No published studies exist using MODIS data for remote lake monitoring owing to the coarse spatial resolution (500 m) (Landsat:3O m), however, daily image capture is an important advantage over Landsat (16 days). We estimated secchi disk depth during 1990-2010 using Landsat imagery (1,511 lakes) and during 2001-2010 using MODIS imagery (83 lakes) using multivariate linear regression (Landsat: R3=0.69-0.89; 9 models; MODIS:R2=0.72-0.94; 14 models). Landsat is useful for long-term monitoring of lakes > 8 ha and MODIS is applicable to annual and within- year monitoring of large lakes (> 400 ha).