Bridging the Temporal Mismatch between Remotely Sensed Land Use Changes and Field-based Water Quality/Quantity Observations
SUNY College of Environmental Science & Forestry
Web Address: http://www.aboutgis.com
This material is based upon work supported by the United States
Environmental Protection Agency under Award Number EPA 05 X-83232501-0.
The overarching objective is to assess water quality impacts due to expanding impervious surface cover. We target impervious surfaces because of their significant impact on hydrology, water quality and micro-climate. Our study site is the southwestern portion of Onondaga County.
Our approach targets the development of:
1. Enhanced detection and classification algorithms of impervious surfaces using remotely-sensed imagery; and
2. Watershed impact models designed to take advantage of up-to-date maps of imperviousness.
We developed the necessary image processing algorithms to automatically detect impervious surfaces from satellite images. Our methodology is designed to integrate multiple algorithms as opposed to existing methods that are limited into a single algorithm. In addition to a more accurate representation, the method produces spatially explicit accuracy metrics that link every impervious pixel to a specific accuracy metric. This allows subsequent hydrological modelers to use remote sensing products in a more efficient and educated manner without necessarily requiring image analysis expertise.
Suitability maps showing the likely areas of new impervious surface development were generated from the impervious surface area (ISA) maps.Using the suitability map as input, we generated projections of new ISA to the year of 2020. The projected land use change was used as input to the hydrology and nutrient loading model ReNuMa to predict the consequent changes in streamflow and nitrate export. We tested if improved ISA maps translated into better hydrological modeling by running the simulation with and without the ISA map (NLCD ISA maps or NLCD only). Similar results were obtained, indicating that ReNuMa, which is not spatially explicit and considers only the total IS in the catchment, has limitations in dealing with variability in spatial configuration of impervious cells. We are developing a truly spatially explicit model that considers the connectivity of impervious cells, which may be more appropriate to match the high detail of land use model outputs. We also took a closer look at nutrients within the Onondaga Creek watershed. Synoptic surveys of water chemistry revealed complex patterns of nutrient availability and uptake, with spatial and temporal complexity.
Back To Project List