CHAPTER 11 Modeling Applications GIS and H&H model integration allows the users to be more productive. Users can devote more time to solving the problems and less time on the mechanical tasks of inputting data and interpreting reams of model output. More than just text outputs, models become automated system-evaluation tools. GIS integration saves time and money. GIS integration is ideally suited to solve the computer modeling puzzle. 2097_C011.fm Page 193 Thursday, December 9, 2004 12:35 PM Copyright © 2005 by Taylor & Francis LEARNING OBJECTIVES The learning objectives of this chapter are to classify the methods of linking water, wastewater, and stormwater system computer models with GIS, and to understand the differences between various linkage methods. MAJOR TOPICS • GIS applications in H&H modeling • Modeling application methods • Interchange method • Interface method • Integration method • Examples and case studies of the preceding methods LIST OF CHAPTER ACRONYMS AML Arc Macro Language BASINS Better Assessment Science Integrating Point and Nonpoint Sources COE Corps of Engineers (U.S. Army) DEM Digital Elevation Model GUI Graphical User Interface H&H Hydrologic and Hydraulic HEC Hydrologic Engineering Center (U.S. Army Corps of Engineers) HSG Hydrologic Soil Group HSPF Hydrologic Simulation Program — FORTRAN NPS Nonpoint Source NRCS Natural Resources Conservation Service (U.S.) SCS Soil Conservation Service (U.S.) (Now NRCS) SWAT Soil and Water Assessment Tool SWMM Storm Water Management Model TMDL Total Maximum Daily Load TR-20 Technical Release 20 VBA Visual Basic for Applications WMS Watershed Modeling System This book focuses on the four main applications of GIS, which are mapping, monitor- ing, modeling, and maintenance and are referred to as the “4M applications.” In this chapter we will learn about the applications of the third M (modeling). TEMPORAL-SPATIAL MODELING IN WESTCHESTER COUNTY GIS software ArcInfo Modeling software Rational method and kinematic wave model 2097_C011.fm Page 194 Thursday, December 9, 2004 12:35 PM Copyright © 2005 by Taylor & Francis Lateral flow is an explicit component of the kinematic wave routing equation, yet some kinematic wave models such as HEC-HMS do not include lateral flow. The lateral flow component can be accounted for by using GIS. Until recently, it was considered impossible to conduct time-varying computations within a GIS. In this case study, lateral flows were derived from GIS output for each segment of the stream and at each time interval of the rain storm and were routed using the kinematic wave routing method. Rather than a raster GIS that uses a constant cell size, a vector GIS was used to define hydrologic response units that divide the stream channel into segments that vary in size according to the combined characteristics for land use, slope, and soil type. This approach permitted vector-based spatially distributed modeling of stream flow during storm events. GIS was used to map and visualize contributing areas around a stream channel. During each calculation of the discharge, a graphical image of the watershed and contributing areas was captured as a Graphics Interchange Format (GIF) image. A series of these images were displayed in sequence to produce a continuous animation (Gorokhovich et al., 2000). H&H MODELING No one said mathematical modeling would be easy, but the preparation of input data and interpretation of output results required by the ever-changing complexities of sophisticated hydrologic and hydraulic (H&H) models have been mind boggling. The good news is that model building and interpretation of results are now easier than ever before thanks to recent advances in computer hardware and software. The assembly of model input data by traditional, manual map measurements is just too time consuming and difficult to justify, now that personal computers are affordable and powerful. It is much easier to see a color-coded map of surcharged sewers to pinpoint the areas of flooding and surcharging than to digest reams of computer output, especially for nonmodelers. Nonmodelers are those people who are not expert modelers but who need to know the modeling results, such as clients of consulting firms, project managers, elected officials, and regulatory agencies. After all, as advocated in Chapter 1 (GIS Applications), GIS technology is an effective means of bridging the gap between information and its recipients. Rapidly developing computer technology has continued to improve modeling methods for water, wastewater, and stormwater systems. GIS applications provide an accurate and manageable way of estimating model input parameters such as node demands, sewage flows, and runoff curve numbers. GIS-based modeling, as a side benefit, also provides an updated database that can be used for nonmodeling activities such as planning and facilities management. GIS data Land use (from low-altitude infrared color photography), USGS DEM (1:24,000 scale), NRCS soils (1:12,000 and 1:24,000 scales), streams, and watersheds Study area 0.14 mi 2 (0.36 km 2 ) Malcolm Brook watershed, Westchester County, New York Organization New York City Department of Environmental Protection 2097_C011.fm Page 195 Thursday, December 9, 2004 12:35 PM Copyright © 2005 by Taylor & Francis There are two fundamental requirements in most H&H modeling projects: a suitable model and input data for the model. It is often difficult to select a modeling approach because of trade-offs between models and data. For instance, a detailed model requires a large amount of input data that is often too difficult to obtain or is too expensive. On the other hand, a simple model that requires little data may not provide a detailed insight into the problem at hand. Modelers must, therefore, use an optimal combination of model complexity (or simplicity) and available data. The recent growth in computational hydraulics has made it increasingly difficult for practitioners to choose the most effective computational tool from among a variety of very simple to very complex H&H models. Fortunately, thanks to the advances in GIS applications, creation of input data sets is easier than ever before. This chapter serves as a guide to help professionals select the most appropriate GIS applications for their modeling needs. It presents an overview of the GIS and computer modeling integration approaches and software. The chapter also shows how to estimate the physical input parameters of H&H models using GIS. The chapter largely uses watershed hydrologic modeling examples to explain modeling integration concepts. However, the integration methods presented here are equally applicable to modeling of water and wastewater systems. Water system modeling applications and examples are presented in Chapter 12 (Water Models). Sewer system modeling applications and examples are presented separately in Chapter 13 (Sewer Models). APPLICATION METHODS There are two types of hydrologic models: lumped and distributed. Lumped- parameter models lump the input parameters of a study area over polygons and use vector GIS applications. Distributed models distribute the input parameters of a study area over grid cells and use raster GIS applications. Application of GIS technology to H&H modeling requires careful planning and extensive data manip- ulation work. In general, the following three major steps are required: 1. Development of spatial database 2. Extraction of model layers 3. Linkage to computer models H&H models, databases, and GIS applications are critical in efficiently and effec- tively completing large modeling studies. The models and GIS can be linked to other databases for data sharing purposes. For example, data can be imported from other sources like AutoCAD, edited and modified within the application, and exported or linked to other databases. Database files can be saved in .dbf format and linked or imported into Microsoft Access for further data manipulation. For example, ArcView GIS provides Open Database Connectivity (ODBC) features that can be used to link ArcView tables with other tables and queries in Microsoft Access or other database programs, without actually going through any import and export exercises. Such procedures eliminate the data redundancies and user errors typically associated 2097_C011.fm Page 196 Thursday, December 9, 2004 12:35 PM Copyright © 2005 by Taylor & Francis with such cumbersome tasks. This has proven to be beneficial and time-saving during alternative analysis of systems where a large number of scenarios are modeled and reviewed using the same data connectivity and templates for maps (Hamid and Nelson, 2001). A useful taxonomy to define the different ways a GIS can be linked to computer models was developed by Shamsi (1998, 1999). The three methods of GIS linkage defined by Shamsi are: 1. Interchange method 2. Interface method 3. Integration method Figure 11.1 shows the differences between the three methods. Each method is discussed in the following text with the help of examples and case studies. As described in Chapter 1 (GIS Applications), advanced applications such as the capa- bility to automatically create a model input file always require writing a computer program using a scripting language like Visual Basic for Applications (VBA). INTERCHANGE METHOD Preprocessing is defined as the transfer of data from the GIS to the model. Postprocessing is defined as the transfer of data from the model to the GIS. The interchange method employs a batch-process approach to interchange (transfer) data Figure 11.1 Three methods of GIS applications in H&H modeling. 2097_C011.fm Page 197 Thursday, December 9, 2004 12:35 PM Copyright © 2005 by Taylor & Francis between a GIS and a computer model. In this method, there is no direct link between the GIS and the model. Both are run separately and independently. The GIS database is preprocessed to extract model input parameters, which are manually copied into a model input file. Similarly, model output data are manually copied in a GIS as a new spatial layer for presentation-mapping purposes. Script programming is not necessary for this method, but it may be done to automate some manual operations such as derivation of runoff curve numbers (described in the following text). This is the easiest method of using GIS in computer models and is the most commonly utilized method at the present time. In this method GIS is essentially used to generate model input files and display model output data. Any GIS software can be used in the interchange method. A GIS with both vector and raster capabilities provides more interchange options. Representative examples of the interchange method are described in the following subsections. Subbasin Parameter Estimation Most H&H models need input data for subbasin parameters such as area, overland flow width, and slope. If subbasins are represented as polygons, GIS can automatically calculate the area as a layer attribute. Overland flow width can be measured interac- tively in any direction (e.g., along a stream or a sewer) by using the measurement tools available in most GIS packages. For example, ArcView 3.x provides a Measure Tool that can be used for on-screen measurement of distance. It also provides a Drawing Tool that can be used for on-screen measurement of polygon area. Subbasin slope can be estimated from digital elevation models (DEMs) as described in Chapter 4 (DEM Applications). GIS estimation of subbasin runoff curve number, a critical input parameter in many rainfall-runoff models, is perhaps the best example of the interchange method. 2097_C011.fm Page 198 Thursday, December 9, 2004 12:35 PM Copyright © 2005 by Taylor & Francis Runoff Curve Number Estimation Runoff curve numbers are a set of standard empirical curves that are used to estimate stormwater runoff. Accurate estimation of curve number values is critical to accurate runoff modeling because the quantity of runoff is very sensitive to runoff curve number values. GIS estimation of subbasin runoff curve number, a critical input parameter in many rainfall-runoff models, is perhaps the best example of the interchange method. The estimation approach is based on the land use, hydrologic soil group (HSG), and runoff curve number relationships developed by U.S. Natural Resources Conservation Service (NRCS), formerly known as the Soil Conservation Service (SCS). These relationships are available in the form of runoff curve number tables (U.S. Department of Agriculture, 1986). These tables provide runoff curve numbers for a large number of land uses and four hydrologic soil groups: A, B, C, and D. They also list average percent imperviousness values for various land-use classes. Table 11.1 shows selected data from the NRCS tables for percent impervi- ousness and runoff curve number for typical land-use classes. If existing land-use classifications are not consistent with SCS taxonomy, they may be replaced with an equivalent SCS land-use class shown in Table 11.2. Table 11.1 SCS Runoff Curve Numbers Land-Use Class Average Percent Imperviousness Runoff Curve Number for Hydrologic Soil Group ABCD Fully Developed Urban Areas (Vegetation Established) Open space (lawns, parks, golf courses, etc.) — 49 69 79 84 Paved parking lots, roofs, driveways, etc. — 98 98 98 98 Paved streets and roads — 98 98 98 98 Dirt roads — 72 82 87 89 Commercial and business 85 89 92 94 95 Industrial 72 81 88 91 93 Residential: lots 1/8 acres or less (town houses) 65 77 85 90 92 Residential: lots 1/4 acres 38 61 75 83 87 Residential: lots 1/2 acres 25 54 70 80 85 Residential: lots 1 acres 20 51 68 79 84 Residential: lots 2 acres 12 46 65 77 82 Developing Urban Areas Newly graded areas 77 86 91 94 Idle lands 77 86 91 94 Nonurban Areas Row crops (straight row, poor condition) — 72 81 88 91 Row crops (straight row, good condition) — 67 78 85 89 Meadow (continuous grass, no grazing) — 30 58 71 78 Brush (fair condition) — 35 56 70 77 Woods (fair condition) — 36 60 73 79 Farmsteads (buildings, driveways, and lots) — 59 74 82 86 2097_C011.fm Page 199 Thursday, December 9, 2004 12:35 PM Copyright © 2005 by Taylor & Francis A vector layer for subbasin runoff curve numbers is created by overlaying the layers for subbasins, soils, and land use to delineate the runoff curve number polygons. Each resulting polygon should have at least three attributes: subbasin ID, land use, and HSG. The NRCS landuse-HSG-curve number matrix (Table 11.1 or Table 11.2) can now be used to assign runoff curve numbers to each polygon according to its land use and HSG. Polygon runoff curve number values are then area-weighted to compute the mean runoff curve number for each subbasin. These subbasin-runoff curve numbers can now be entered into the model input file. The runoff curve number estimation technique is shown in Figure 11.2 to Figure 11.5. Figure 11.2 shows the layers for subbasins and land use. Figure 11.3 shows the layers for subbasins and HSGs. Figure 11.4 shows the computed runoff curve number polygons. Figure 11.5 shows the average runoff curve numbers for the subbasins. Some H&H models also need an input for the subbasin percent imperviousness, which can be estimated in GIS using the NRCS runoff curve number tables. A layer for the subbasin percent imperviousness can be created by overlaying the layers for subbasins and land use to delineate the percent imperviousness polygons. The land- use percent imperviousness matrix can then be used to assign percent impervious values to the polygons. Finally, polygon percent imperviousness values are area- weighted to compute the mean percent imperviousness value for each subbasin. Water Quality Modeling Data Estimation Some urban-runoff quality models such as the TRANSPORT Block of SWMM need curb length as a model input parameter, which can be estimated from a GIS layer of road or street centerlines. Subbasin curb length can be estimated by performing Table 11.2 Equivalent SCS Land-Use Classes Land-Use Class SCS Equivalent Land-Use Class Average Percent Imperviousness Runoff Curve Number for Hydrologic Soil Group ABCD Parks, cemeteries, ball fields, etc. Open space 0 39 61 74 80 High-density residential Average lot 506–1012 m 2 (0.125–0.25 acres) 51 69 80 87 90 Medium-density residential Average lot 1349–2024 m 2 (1/3–1/2 acres) 28 56 71 81 86 Low-density residential Average lot 4047–8094 m 2 (1–2 acres) 16 49 66 78 83 Schools Average lot 1349–2024 m 2 (1/3–1/2 acres) 28 56 71 81 86 Agricultural Row crops average 0 64 75 83 87 Disturbed Newly graded areas 35 77 86 91 94 2097_C011.fm Page 200 Thursday, December 9, 2004 12:35 PM Copyright © 2005 by Taylor & Francis an overlay analysis of street and subbasin layers. For example, in ArcView 3.x, curb length can be estimated by performing a theme-on-theme selection using roads as the target theme and subbasins as the selector theme. Alternatively, for large areas such tasks can be handled more efficiently in ArcView’s Network Analyst extension, which has been designed for transportation network applications. Network Analyst is an optional extension that must be purchased separately. Figure 11.2 Layers for subbasins and land use for runoff curve number estimation. Figure 11.3 Layers for subbasins and HSGs for runoff curve number estimation. 2097_C011.fm Page 201 Thursday, December 9, 2004 12:35 PM Copyright © 2005 by Taylor & Francis Demographic Data Estimation Demographic data can be used for the following modeling tasks: 1. Estimation of quantity and quality sanitary sewage flow 2. Estimation of present and potential future development 3. Creation of auxiliary layers for visualization and presentation purposes Figure 11.4 Computed runoff curve number polygons. Figure 11.5 Subbasin runoff curve number map. 2097_C011.fm Page 202 Thursday, December 9, 2004 12:35 PM Copyright © 2005 by Taylor & Francis [...]... layers have more than these five land-use classes However, this is not a problem in GIS because the existing classes can be easily consolidated into the five land-use classes mentioned above Figure 11. 7 shows an ArcView 3.x screenshot for subbasin and land-use layers for a SWMM-based sewer system model The land-use layer was prepared from classification of 30-m-resolution Landsat Thematic Mapper (TM) satellite... available to the public HEC-1 and HEC-2 are COE’s legacy DOS programs for hydrologic and hydraulic modeling, respectively Recently HEC-1 and HEC-2 have been replaced with Windows programs called the Hydrologic Modeling System (HECHMS) and River Analysis System (HEC-RAS), respectively HEC Geo-HMS and HEC Geo-RAS have been developed as geospatial hydrology toolkits for HEC-HMS and HEC-RAS users, respectively... models and switching between the models was hardly noticeable, and output from one model was used as input for another model The TR-55 model was embedded within the TR-20 model The output from TR-55 became input for TR-20 The TR-20 model provided discharge values for input to the HEC-2 model, and output from HEC-2 could be directly inserted into TR-20 as rating-table information All modeling applications. .. public-domain modeling system that uses the integration method A few integrated commercial systems are available starting at $5000 There are many in-house and proprietary integrated systems The number of public-domain and commercial integration packages is expected to grow steadily in the near future CHAPTER QUESTIONS 1 What are the GIS applications in computer modeling of water, wastewater, and stormwater. .. structures, and other control points The hydrologic results from HEC-GeoHMS are then imported by HEC-HMS, where simulation is performed HEC-GeoHMS works with Windows 95/98/NT/2000 operating systems HEC-GeoRAS HEC-GeoRAS (formerly named AV/RAS) for ArcView is an ArcView 3.x GIS extension specifically designed to process geospatial data for use with HEC-RAS The extension allows users to create an HEC-RAS import... method is the ArcInfo and HEC-1 and HEC-2 interface developed by Woolpert, Inc (Phipps, 1995) In HEC-1, the interface model uses vector layers for streams and open channels and the ArcInfo networking function to show the locations of and the relationships between different types of flows: sheet flow, shallow channel flow, and concentrated channel flow The ArcInfo and HEC-1 Figure 11. 9 GISHydro TxDOT extension... census block and block group layers for estimating subbasin population Copyright © 2005 by Taylor & Francis 2097_C 011. fm Page 204 Thursday, December 9, 2004 12:35 PM Land-Use Data Estimation Dry- and wet-weather flows from sewersheds depend on land use Most rainfallrunoff models, therefore, need land-use information In some models such as SWMM’s TRANSPORT Block, land-use type is input directly In others,... data for HEC-HMS and HEC-RAS models Free downloads of these programs are available from the HEC software Web site HEC-GeoHMS HEC-GeoHMS is an ArcView 3.x GIS extension specifically designed to process geospatial data for use with HEC-HMS It allows users to visualize spatial information, delineate watersheds and streams, extract physical watershed and stream characteristics, perform spatial analyses, and. .. Initially, the layer contained 19 land-use classes Subsequently, the initial land-use classes were reclassified into the five land-use types required for TRANSPORT input In order to refine the land-use classification, the layer was combined with the census-block layer Figure 11. 7 Estimating land use using interchange method Copyright © 2005 by Taylor & Francis 2097_C 011. fm Page 205 Thursday, December 9,... specific problems are required For example, MIKE SHE can be used for modeling runoff quality and quantity, including interaction with surface water systems, MIKE 11 can be used for detailed modeling of river hydrodynamics and water quality, and MIKE 21 can be used for modeling bays and estuaries Geo-STORM Integration An early example of ArcInfo 7.x and modeling integration is the Geo-STORM package developed . modeling methods for water, wastewater, and stormwater systems. GIS applications provide an accurate and manageable way of estimating model input parameters such as node demands, sewage flows, and runoff. Estimation Dry- and wet-weather flows from sewersheds depend on land use. Most rainfall- runoff models, therefore, need land-use information. In some models such as SWMM’s TRANSPORT Block, land-use type. hydrographs generated in the HEC-1 hydrologic model are then used to feed the HEC-2 hydraulic model. The ArcInfo and HEC-2 interface allows graphical display of the 2-, 1 0-, 2 5-, 5 0-, 10 0-, and 500-year frequency