College of Natural Resources
February 21, 2001
Mission
This study is deigned to evaluate the nitrogen concentration in the groundwater of the Tomorrow/Waupaca Rivers watershed that lies in the glaciated areas of central Wisconsin.
Importance
Groundwater contamination from agricultural activities has aroused wide attention. Forty-five percent of agricultural land in the watershed is non-irrigated production growing corn and hay, another 10% is irrigated production growing potatoes, peas, sweet corn, snap beans, soybeans, and cucumbers. Agricultural inputs on corn and potato fields, and animal wastes are believed to be the major contributors to water contamination in the watershed.
Project Summary
Artificial Neural Network (ANN) is a kind of statistical model. It uses a set of inputs, such as soil properties, land use, groundwater properties, and precipitation, and then outputs a set of predictions. It compares the predictions with actual data. If the predictions are not close, then the ANN model will modify the weights within the model and it will run again until it reaches a satisfactory answer.
Some new algorithms for the ANN model have already been completed, and a software prototype of the system has been developed. In addition, some data for the watershed, such as digital elevation model, soil properties, land use, hydrology, and precipitation, has been collected. The completion of this database will supply rich information on watershed research and management. Also, this spatial database will allow new advanced research tools like Geographic Information Systems (GIS) to be used more efficiently in research and watershed management.Furthermore, the spatial database will be put onto the Internet. This will allow local citizens to browse the spatial and attribute properties of the Tomorrow/Waupaca watershed simply by clicking on their Internet browsers. Thus, people will have access to data about the watershed and can learn more in a very convenient way.
People who have a detailed interest in groundwater contamination within the watershed, will have options through the Internet to input specific data. For example, people who have an adequate data set can input the data from their Internet browsers and run models on the Internet to predict the nitrogen concentration of groundwater. In later research, people will even be allowed to build models on the Internet.
Contact
Information
|
Dr.
Hangsheng Lin, hlin@uwsp.edu |
Sheng Wang, |