Use of GIS, Geo-Based Programs, and Computer Models for Watershed and Site Analyses – Part 2
There is a wide variety of information that is sought by resources engineers and planners, erosion control professionals, and material specifiers, depending on their objectives and required level of detail. Therefore, the following discussions will be a broad brush related to typical watershed studies and will not be specific to the professions previously described. The mentioning of a computer program in this article is neither an endorsement of the program nor an indicator of its quality but is presented as a service to the reader.
Erosion, the detachment of soil particles, occurs by the action of water, wind, and glacial ice. Only erosion caused by water will be considered here. Water erosion occurs when raindrops, spring runoff, or floodwaters wear away and transport soil particles. Erosion is a complex natural process that has often been accelerated by human activities such as land clearance, agriculture, construction, surface mining, and urbanization. This article focuses primarily on the actual erosion and its impacts on downstream water quality.
Eroded sediment acts as both a physical and chemical pollutant. As a physical pollutant, sediment decreases turbidity in the receiving waters, which limits the penetration of sunlight into the water column, thereby limiting or prohibiting growth of algae and rooted aquatic plants. Sediment can also result in high levels of aggradation in rivers and may destroy riparian habitats and wetlands or alter fish spawning patterns. Sediment can also be a chemical pollutant because it can transport nutrients (i.e., phosphorus and nitrogen) and other contaminants such as heavy metals and pesticides. These contaminants can be released into the water column thus degrading water quality or be ingested by aquatic organisms and bioaccumulate in fish. If contaminated sediments deposit in rivers, water-quality degradation may persist even when other sources of pollution are controlled.
Since erosion has far-reaching environmental and economic impacts, significant effort has been directed toward developing analytical tools (computer models) that predict erosion and sediment-associated chemical runoff. The database of information required by these tools is relatively complex. With the advancement of GIS technology, the development and use of these computer models has become more efficient and economically feasible. This paper begins with a description of the methods and tools used (in conjunction with GIS) to estimate the amount of sediment eroded (sediment yield). The sediment yield can be used by sediment transport models to calculate the resulting suspended sediment concentrations as well as scour and deposition in rivers and reservoirs (see previous article). Next is a discussion of the models used to predict water quality in rivers and reservoirs. These models predict the transport of contaminants and nutrients associated with the water column, the suspended sediment, and the bed material. Since stormwater in urban areas is generally transported by a system of pipes and/or covered or uncovered drains, a discussion of the computer programs developed for stormwater sewer system follows. Groundwater models are then presented because groundwater tends to flow to rivers where it emerges as streamflow during periods of little or no rain. This groundwater carries contaminants and nutrients leached from the soil into the river thus impacting water quality. A discussion of watershed models that incorporate two or more of the components described above and in the previous article concludes this paper. These models are generally integrated with GIS and demonstrate the effectiveness of GIS in watershed analyses.Surface erosion and mass movements within a watershed produce sediment available for transport. The actual amount of eroded material passing a given point in a watershed (usually the outlet) within a given amount of time is termed sediment yield. Methods for estimating sediment yield were first developed for analysis of agricultural practices. Many different methods are now available to estimate sediment yield for basins with a variety of land uses. Some of the more popular methods are described below.
- Universal Soil Loss Equation (USLE). USLE is perhaps the most widely used method for estimating soil erosion. Although the equation was originally developed for small agricultural areas, its use has been extended to basins with other land uses. Average annual sheet and rill erosion is computed as the product of a rainfall erosion index (R), a soil erodibility factor (K), a slope length and steepness factor (LS), a vegetative cover factor (C), and an erosion control practice factor (P).
- Modified Universal Soil Loss Equation (MUSLE). This method modified the R in the USLE in order to predict erosion from a single storm event. The modified R represents the product of runoff volume and peak discharge for a single storm event.Revised Universal Soil Loss Equation (RUSLE). This method was developed to update and extend the USLE for non-agricultural applications and to incorporate additional data collected after development of the original USLE.
- Pacific Southwest Interagency Committee (PSIAC) Method. Unlike the USLE and its variations, the PSIAC method estimates total annual sediment yield, not just sheet and rill erosion. The method is intended primarily for planning purposes and results in a range of expected yield values. The procedure considers nine factors that depend on surface geology, soils, climate, runoff, topography, ground cover, land use, channel erosion, and upland erosion. The procedure was developed for watersheds in the western United States greater than 30 km2 (10 mi.2); however, it has been applied to smaller basins.
- Regional Methods. Regional sediment yield analyses have been prepared for several areas of the US. These methods are often a good preliminary estimate of expected yield and a check on some of the other methods’ results. The USDA Soil Conservation Service (now the Natural Resources Conservation Service) developed generalized sediment yield rate maps for the western US in 1974. Dendy and Bolton developed two regression equations relating sediment yield to drainage area and mean annual runoff based on data from about 800 reservoirs throughout the US. The Los Angeles District Method, developed by the US Army Corps of Engineers (ACE), predicts unit debris yield values for “n-year” flood events for the design and analysis of debris-catching structures in southern California watersheds.
GIS is a useful tool for sediment yield studies. Typically, maps of a basin’s soils, vegetation, land use, and other parameters (also called “coverages”) can be brought into a GIS and queries performed to extract the data needed for the above listed sediment yield methods. For example, the length slope and steepness factor in the USLE equation or the geology, soils, or ground cover variables in the PSIAC equation can all be estimated from GIS coverages. In some areas, these types of coverages may have already been prepared by local or federal government agencies. For some areas and/or watersheds, however, existing maps must be digitized and linked to a table or database in order to prepare the coverage. Usually some averaging of values for a certain parameter is performed over a given basin or sub-basin. The sediment yield equations can be programmed in the GIS application, the necessary parameters extracted from the appropriate coverages, and the results stored in a new coverage and/or database. GIS can also be used to graphically present the results from the yield analyses.
The water-quality models discussed in this section simulate the transport of water-quality factors in surface waters and do not simulate overland flow or runoff. These parameters are inputs to these models. Some of the models that have been applied in a wide variety of analyses include the Enhanced Stream Water Quality Model with Uncertainty Analysis (QUAL2EU) supported by the Environmental Protection Agency (EPA), the Water Quality Analysis Simulation Program (WASP5) that is approved by EPA, and HEC-5Q, developed by ACE’s Hydrologic Engineering Center.
QUAL2EU simulates the transport of dissolved water-quality constituents in branching streams and well-mixed lakes and can be operated as either a steady-state or dynamic model with respect to time. The constituents that can be modeled are dissolved oxygen, biochemical oxygen demand, temperature, algae as chlorophyll a, organic nitrogen, ammonia, nitrite, nitrate, organic phosphorus, dissolved phosphorus, coliforms, one nonconservative constituent, and three conservative constituents. The model assumes that the major transport mechanisms, advection and dispersion, are significant only in the main direction of flow and performs computations for the water column only (i.e., dissolved constituents). Neither sediment nor water quality in the riverbed is included. The model inputs include the stream network description; the division of streams into headwaters, reaches, and junctions; and 26 physical, chemical, and biological parameters for each reach. QUAL2EU includes uncertainty analysis to assist with model calibration and sensitivity analyses of the input parameters. QUAL2EU has been especially useful for analyzing the effects of nutrients on algal concentration and dissolved oxygen.
WASP5 can simulate contaminant fate in rivers and lakes by using a compartment modeling approach that can be applied in one, two, or three dimensions. The water-quality constituents that can be modeled include dissolved oxygen, biochemical oxygen demand, nutrients, bacterial contamination, organic chemicals, heavy metals, and sediment. The fate of the contaminants, dissolved and adsorbed onto sediments, in both the water column and the riverbed is simulated. WASP5 uses advection, dispersion, settling, resuspension and sedimentation, and evaporation and precipitation. The inputs include the definition of computational segments (compartments), hydraulic coefficients, inflow or circulation patterns, cross-sectional areas, mixing length, dispersion coefficients, loads, boundary concentrations, and initial concentrations. Users can easily substitute their own subroutines in the program. WASP5 has been used to model eutrophication and pollution of the Great Lakes, pollution of the James River Estuary, volatile organic pollution of the Delaware estuary, and heavy metal pollution of the Deep River in North Carolina.HEC-5Q is used primarily for analyzing water flows and water quality in reservoirs and associated downstream river reaches. It can perform detailed simulations of reservoir operations, such as regulating outflows through gates and turbines, and vertical temperature gradients in reservoirs. Water-quality analysis includes water temperature, three conservative and three non-conservative constituents, dissolved oxygen, and a phytoplankton option.
To date, there are no commercial GIS extensions linked directly to these water-quality models. However, this is an area of active research. Jennifer Benaman, Neal Armstrong, and David Maidment developed an ArcView script that performs the following tasks used to develop a WASP5 model: (1) write the input file information, (2) format the input information into the proper WASP5 file, (3) execute the WASP5 subprogram of dissolved oxygen and biological oxygen demand modeling and model calibration, (4) process the WASP5 output, and (5) assist the user in output visualization. GIS can be an extremely powerful tool in preparing the model inputs, comparing the results to measurements, and displaying the model results. Both QUAL2EU and WASP5 have been incorporated into larger watershed models that are linked to GIS data. These models will be discussed at the end of this article.
Some pollutant transport models have been developed entirely within GIS. Most of the GIS programs are equipped with a macro language that allows the user to write models within the GIS application. In addition, external procedures written in such programming languages as C/C++ or FORTRAN can be executed by a macro, thus making the modeling process efficient. Water-quality models developed in GIS have been used to assess nutrient loads in New Jersey rivers.
GIS is gaining a foothold in the area of storm sewer hydraulic analysis. Storm sewer computer simulation packages model the flow of rainfall runoff that is collected by, and flows through, a municipal storm sewer system. The principal goal of the analysis is to determine the capacity of the sewer system to handle peak flow periods, and whether inlets or manholes might become flooded, or catchments might overflow, due to insufficient capacity. Such models require a map representation of the sewer network involved. Some stormwater distribution programs have been incorporated into larger watershed models. These models are discussed at the end of this article.
For several years, various modeling systems such as Haestad StormCAD (www.haestad.com/), and HYDRA (www.scisoftware.com/) have had some level of integration with (non-GIS) computer-aided-design (CAD) systems. Usually this allows the user to edit the layout and attributes of the storm sewer system in AutoCAD, for example. Then, the hydraulic analysis program can read this geometric information to perform the hydraulic analysis, and even output information back into AutoCAD.
In recent years, with the increasing popularity of GIS, many programs added features to allow their software to import and export to GIS formats, such as Arc/INFO coverages, ArcView shape files, and MapInfo files. But, as was done with CAD earlier, now the trend is to provide a higher level of integration with GIS packages to allow the user to edit both graphical and attribute information (such as friction factors) in the GIS program, return to the hydraulic analysis program to run the analysis/simulation, and then view some of the results (such as resulting water levels) using a GIS program.
One advantage of using GIS applications for hydraulic analysis is that they are increasingly being used as part of asset management databases. Municipalities are using GIS to track locations and characteristics of various installations. For existing storm sewer systems, the GIS layout of the system may already exist; it may be an easy matter to then visually modify the layout for use by the model. In addition to the geometric layout, information such as pipe types, sizes, and installation dates will often be in the GIS datasets. Friction factors, which may not be explicitly present in the GIS dataset, can be derived from this other information. At the very least, street and topographical maps are easy to obtain in GIS format and the storm sewer system can be overlaid on these. The ability of GIS programs to provide color-coding of various attributes (such as pipe sizes) enables quick visual verification of the system.
One of the models that is integrated extensively with GIS is MOUSE GIS, which works in conjunction with the program MOUSE, both published by the Danish Hydraulic Institute (DHI) (www.dhi.dk/). MOUSE is designed to simulate surface runoff, flow, water quality, and sediment transport in urban catchments and sewer systems. MOUSE GIS runs inside the popular ArcView program by Environmental Research Systems Institute (ESRI, www.esri.com/) for the entry of pipes, nodes, pumps, weirs, and catchments. The MOUSE GIS interface also provides a utility that simplifies networks to facilitate computation by MOUSE. An export file is then created for use by MOUSE. The analysis is run in MOUSE (which can be launched by clicking an icon from within ArcView), and in addition to the information available from MOUSE, some results can be returned to ArcView. From within ArcView, the user can view the maximum pipe filling (which is visually represented by color coding of pipes according to their maximum fill level), maximum water levels in manholes (again color coded), and time series water-level in manholes.
Haestad Methods, publishers of popular hydraulic and hydrologic modeling software, already provides some integration of StormCAD with GIS: Storm sewer networks for StormCAD can be built and maintained from within ArcView or Arc/INFO. Haestad plans to increase the level of integration. The company announced in April 2000 its Geographic Engineering Systems (GES) product line, which will integrate GIS technology and its various modeling systems. Haestad has chosen ESRI’s Arc/INFO 8, a high-end GIS package, as its framework.
Groundwater systems are recharged by downward infiltration of rainfall and runoff through the overlying soil. The rate at which groundwater systems are recharged is a function of how much inflow, overland flow, and evapotranspiration occurs. Groundwater tends to flow into rivers where it emerges as streamflow, called baseflow, during periods of little or no rain. This groundwater carries contaminants and nutrients leached from the soil into the river thus impacting water quality.
Because of population, industrial, and agricultural growth, groundwater systems are becoming increasingly threatened and are often contaminated. The need to clean up this natural resource and the technological advances in computing have led to the development of groundwater modeling software to analyze and solve contamination problems.
Many software packages that simulate common features in groundwater flow and associated contaminant movement have been developed. Some of the most widely used models include: MODFLOW, developed by the US Geological Survey (USGS); GMS, developed by Brigham Young University; and Visual MODFLOW, developed by Waterloo Hydrogeologic Inc. MODFLOW is a three-dimensional finite difference groundwater flow model that allows users to simulate subsurface conditions such as steady-state or transient flow, confined or unconfined aquifers, recharge areas, evapotranspiration, hydraulic conductivities, and transmissivities. GMS and Visual MODFLOW are graphical user interfaces that have modified the core MODFLOW simulator. Documentation of more groundwater modeling software that deals with groundwater and contaminant movement can be found at http://scisoftware.com/.
The use of GIS and groundwater modeling is relatively new. Several inputs are needed to model groundwater. Factors such as streams, land use, rivers, vegetation, drains, wellhead locations, and topography are needed. In most cases this data resides in a GIS and will not have to be recreated. With the continuing development of GIS and the increasing availability of data, groundwater modeling software manufacturers have seen the advantages of integrating GIS data into their models. Several of the groundwater modeling programs now allow GIS data to be imported, exported, or both. The most common supported GIS data structure used in groundwater modeling is in Arc/INFO and ArcView format developed by ESRI. Raster and vector and DEM data can be obtained through several methods. A search engine is provided through the ESRI Web site or data can be found through the USGS Web site (www.usgs.gov/).As a result of combining GIS data into a groundwater model, the model development is more efficient and more accurate. GIS applications can be used to sort and organize the data by performing queries on statistical groundwater data, to display proximity studies, and to create overlays.
Watershed Models: An Integrated Analysis Approach
A watershed was described by John Wesley Powell in 1869 as “that area of land, a bounded hydrologic system, within which all living things are inextricably linked by their common water course and where, as humans settled, simple logic demanded that they become part of the community.” Since many of the processes associated with rainfall/runoff, erosion, and water quality within a watershed are interrelated, EPA and other agencies are emphasizing a watershed-based assessment and integrated analysis of the processes affecting erosion and water quality. This approach addresses both point and nonpoint-source pollution.
Point-source pollution results from the release of a contaminant from a specific location such as an industrial or sewage treatment plant. Nonpoint-source pollution is caused by rainfall moving over and through the ground. This flow picks up and transports pollutants to lakes, rivers, wetlands, coastal waters, and even groundwater (EPA). These pollutants include excess fertilizers, herbicides, insecticides, oil, grease, sediment, salt, bacteria, and toxic chemicals. States report that nonpoint-source pollution is the leading remaining cause of water-quality problems (see www.epa.gov/OWOW/NPS/qa.html).
Many computer models have been developed to perform watershed assessments integrating rainfall/runoff (hydrologic models) with erosion models (sediment yield and sediment transport in streams) and identifying point and nonpoint pollution sources and assessing their impact on water quality within the watershed. These models include Better Assessment Science Integrating Point and Nonpoint Sources (BASINS), developed by EPA; Soil and Water Assessment Tool (SWAT), supported by USDA Agricultural Research Service; EPA’s Stormwater Management Model (SWMM); Hydrologic Simulation Program Fortran (HSPF), from EPA; Storage, Treatment, Overflow, Runoff Model (STORM), developed by the ACE Hydrologic Engineering Center; and the Agricultural Nonpoint Source Model (AGNPS), developed as a joint project between the USDA Agricultural Research Service and the Natural Resources Conservation Service. The BASINS, SWAT, and SWMM models are discussed in this article. Information on the other models can be obtained from their respective vendors.
BASINS integrates GIS, national watershed data, and environmental and modeling tools into one package. The main components of BASINS are (1) national databases (EPA’s Ecoregions, National Water Quality Assessment Study Unit Boundaries, State Soil and Geographic Database, EPA’s Water Quality Monitoring Stations and Data Summaries, Water Quality Stations and Observation Data, National Sediment Inventory Stations and Database, Listing of Fish and Wildlife Advisories, Gage Sites, Weather Station Sites, etc.); (2) assessment tools; (3) utilities including local data import, land use and digital elevation model reclassification, watershed delineation, and management of water-quality observation data; (4) watershed and water-quality models including HSPF and QUAL2EU; and (5) postprocessing output tools for interpreting model results. HSPF simulates nonpoint-source runoff and pollutant loadings for a watershed and performs flow and water-quality routing for river reaches. The water-quality constituents that can be modeled are dissolved oxygen, organic matter, temperature, pesticides, nutrients, salts, bacteria, sediment, and pH. Future versions will include the SWAT model, which is described below. BASINS’ databases and assessment tolls are directly integrated with an ArcView GIS environment. The simulation models run in a Windows environment, using data input files generated in ArcView.
SWAT is a watershed-scale hydrologic and water-quality model developed to predict the effects of alternative land use management decisions on water, sediment, and chemical yields. SWAT simulates hydrology, pesticide and nutrient cycling, bacteria transport, erosion, and sediment transport. Sediment yield is computed using MUSLE. The river processes in SWAT include channel flood and sediment routing and nutrient and pesticide routing and transformation modified from the QUAL2EU model. GIS interfaces have been developed for SWAT using both the Graphical Resources Analysis Support System (GRASS) and ArcView. These GIS interfaces prepare the model input, execute the model, and display the results.
SWMM is a comprehensive computer model for analysis of quantity and quality problems associated with urban runoff. Both single-event and continuous simulation can be performed on watersheds having storm sewers or combined sewers and natural drainage, for prediction of flows, water surface elevations, and pollutant concentrations. Many vendors have developed Windows interfaces to aid in the development and presentation of results of SWMM models. Some GIS interfaces have been developed to prepare the model input from GIS databases (such as PCSWMM GIS by Computational Hydraulics International).
Watershed analyses that include rainfall/runoff, hydraulic, erosion, and water-quality assessments are data intensive, and traditional approaches have involved many separate steps and analyses. Traditionally, each individual step was performed using a variety of tools and computer systems. The isolated implementation resulted in a lack of integration, limited coordination, and a time-intensive, expensive process. Thanks to the integrated tools discussed above, watershed analyses have become more efficient and economically feasible. By performing these types of watershed studies, erosion control professionals can evaluate the impacts of erosion control methods, perform sensitivity analysis on design parameters, and evaluate watershed treatments.
GIS provides an integrating framework that can store and process the massive amount of data required for watershed studies and display the inputs and results, or any combination thereof, in a spatial, user-friendly format. The implementation of GIS into watershed and erosion studies has opened a whole new realm of opportunities for water resources engineers, planners, and erosion control professionals.