John,
here is some text and references in addition to my previous document. Both documents include some incomplete references, I will find the exact citation when you let me know which ones you are going to include. These notes are also at our website
regards,
Helena
2.3 Hydrologic modeling
- Lumped versus distributed models
Traditionally, watersheds have been represented as homogeneous units with terrain, soil, and cover conditions described by averaged values. GIS has provided the tools to compute these averaged values efficiently which allowing the user to split the large watersheds into smaller hydrologic units (usually subwatersheds or half-subwatersheds) and incorporate partial spatial variability into the watershed based simulations. These models (e.g., SWAT: Arnold et al. 1993) simulate a broad spectrum of processes (surface and subsurface water flow, sediment and pollutant transport, etc.) with continuous time simulation. The results represent averages for entire watersheds/subwatersheds, and often provide support for management at a regional level, which involves, for example, identification of watersheds with high risk composition of land use, or designation of watershed level conservation areas. The watershed based models have been linked to GIS for a number of years and several are currently available in on-line versions (SWAT - Srinivasan and Arnold 1996, BASIN - EPA, LTHIA2 - - Long Term Hydrologic Impact Assessment - Engel et al. 1999).
GIS tools for processing and analysis of spatial data have stimulated the development of a new generation of process-based models which simulate water flow as two dimensional function usually represented by a raster (TIN is being used too) (SIMWE: Mitas and Mitasova 1998, CASC2d: Julien, Ogden, Saghafian, 1995, MIT models: Brass, Willgoose, Moglen, SIBERIA, CHILD 1992-1998)). These models predict the water flow (water depth, discharge) at any point in landscape, not just the watershed outlet (as it is with the watershed based models). The averaged values of landscape characteristics have been replaced by their distributed representation (usually raster) and the models can then simulate not only the impact of specific landuse practices but also their spatial pattern and location within the watershed (Figure XX). Spatially distributed process-based models are thus providing new insights into the interactions between land use/land cover and water flow and water quality. This approach has also revealed substantial gaps in theory of sediment and pollutant transport processes in complex landscapes. New approach to field experiments integrated with spatial modeling is needed to improve our understanding of spatial interactions influencing water resources. (and for example bring the quantitative accuracy of sediment load predictions, which is currently at about 50-150%, to acceptable and useful levels.
4. Education
An example of Web based learning environment about the watersheds:
www.eot.org/TTF/Learn/lrn02-riverweb2.html
www.ncsa.uiuc.edi/Cyberia/RiverWeb
Some additional References:
2.1 New GIS data, their management and delivery: We should include references to work done at Purdue lead by Bernie Engel
Comprehensive Water Resources data sets (with focus on Indiana) are at Purdue
2.2 Water resources/water quality models linked with GIS at Purdue
Other useful references
Johnston, D.M. and A. Srivastava. 1999. Decision support systems for design and planning: the development of HydroPEDDS (Hydrologic Performance Evaluation & Design Decision Support) system for urban watershed planning, 6th International Conference on Computers in Urban Planning and Urban Management (CUPUMS'99), Venice, Italy.
Maidment, D. R., 1996, Environmental modeling within GIS, in GIS and Environmental Modeling: Progress and Research Issues, edited by M. F Goodchild, L. T. Steyaert, and B. O. Parks, GIS World, Inc., pp. 315-324.
References which can be omitted
Mitas ...1996 Interacting fields (is covered by 1997 and 1998 papers)