Watershed water quality modeling using integrated fuzzy modeling approach with HSPF model and radar rainfall data
Abstract
Watershed water quality modeling is very essential for watershed management activities. These modeling activities and resulting simulations were used as the main resource for several decision making processes. In this research study, uncertainties associated with the watershed water quality modeling were handled with fuzzy modeling approach in three different phases. Geospatial data analysis is an integral part of watershed modeling. Several geospatial data (such as Digital elevation model, land use etc) were used with Geographical Information Systems (GIS) for model development. As a case study, Coffee Creek watershed (a part of eight digit watershed HUC 04040001) located in North West Indiana (Chesterton region) was considered for this research. Coffee Creek watershed is listed in 303(d) document published by US EPA under impaired watersheds category and recommended the state agency to develop a TMDL report. Fecal Total Maximum Daily Loads (TMDL's) was developed for Coffee Creek watershed using a traditional HPSF (Hydrologic Simulation Program Fortran) modeling with integrated fuzzy based approach. HSPF computer simulation models are popularly used in watershed modeling. Through mathematical equations expressing physical relationship between several variables associated with watershed, HSPF model simulate watershed rainfall runoff processes. Fuzzification was implemented for non-point sources to decide the water quality load allocations in this study. This fuzzy based approach proposed in current study is promising and shows the potential of facilitating TMDL decision making. Uncertainties were found at different levels of watershed quality management. Meteorological information used in HSPF model is usually a point rainfall observed using a weather station. In the second phase, in this research, a new approach is proposed by using spatial rainfall information from NEXRAD (Next Generation RADAR) to improve the model performance in decision making instead of point rainfall data. In public domain no software are available to convert NEXRAD information into spatial rainfall information of the required watershed. A reflectivity-rainfall conversion software was developed for the study. Uncertainness was noticed at reflectivity-rainfall conversion level too. To handle this problem, in phase III, fuzzy logic based approach was used in deciding the coefficient values used for radar signal to rainfall rate conversion for Chicago radar observations. This research study shows the advantages of using fuzzy model approach at different levels in watershed quality management. Proposed approach is very simple and can be easily generalized to use for different watersheds.
Degree
M.S.E.
Advisors
Viswanathan, Purdue University.
Subject Area
Geographic information science|Environmental engineering|Artificial intelligence
Off-Campus Purdue Users:
To access this dissertation, please log in to our
proxy server.