Background:
Water quality reflects both land use and geologic factors. However, specific water quality factors may be influenced by local or regional conditions which themselves may change on a seasonal basis (Johnson et al. 1997). We found that much of the variation in total nitrogen and NO3-N concentrations in Saginaw streams during the summer could be explained by land use, specifically the percentage of rowcrop agriculture in a watershed (Johnson et al. 1997). In addition, we found that land use in the riparian zone and land use throughout the entire watershed explained similar amounts of variation in TN, NO2+NO3-N, PO4-P, and alkalinity, while riparian zone land use alone explained more of the variance in total suspended solids and TP.
We are predicting critical aspects of stream water chemistry data using SWAT, (Srinivasan and Arnold 1994) a model designed to interface with a GIS database to model the influence of land cover and land uses on sediment, water, and chemical yields from ungauged basins. SWAT simulates continuous-time water quality responses so that catchment hydrology can be assessed in daily time steps over long periods. SWAT consists of several hydrologic GIS-based models and other database access tools integrated to form a basin-scale hydrologic model. Routing through the watershed is simplified by the creation of subwatersheds based upon homogeneous topographical, soil, and land cover characteristics. SWAT has been validated at several spatial scales ranging from 18 to 9000 km2 (Arnold et al. 1993, Srinivasan and Arnold 1994). We have established this GIS-SWAT interface for the Saginaw watersheds (Haro et al. 1996 abstract). Our early results suggest that the model works well at the scales of interest to this research.
Objectives:
To
obtain predictions of seasonal watershed outputs of water, nitrogen, phosphorus,
and sediment
Examine differences between watersheds in relation to landscape factors such as landuse and geology
Methods:
·
Use
SWAT, the Soil and Water Assessment Tool and ArcView to perform modeling
·
Use
STATSGO soil data, 30m Digital Elevation Model, and 30m resolution landuse
·
Incorporate
USGS county-level cropping practices into landuse data
·
Use
sample data for stream reach properties where available
·
Calibrate
flow with USGS gage data for numerous sites throughout watersheds
·
Calibrate
sediment and nutrients with grab samples taken at various times in each
watershed
·
Alter
properties to fit each scenario in turn
Early
Resu
1)
Screen
Capture of SWAT output, with SWAT-created basins, nodes, and rivers
2)
Landuse
of Whitewater River basin, with blow-up of two basins, illustrating diversity
4)
Results
of a SWAT run for Organic Nitrogen
