Environmental Modelling & Software 26 (2011) 817e821 Contents lists available at ScienceDirect Environmental Modelling & Software journal homepage: www.elsevier.com/locate/envsoft Short communication A GIS-based decision support system for integrated flood management under uncertainty with two dimensional numerical simulations Honghai Qi a, *, M.S. Altinakar b,1 a b National Center for Computational Hydroscience and Engineering (NCCHE), University of Mississippi, University, MS 38677, USA NCCHE, University of Mississippi, University, MS 38677, USA a r t i c l e i n f o a b s t r a c t Article history: Received 19 January 2010 Received in revised form 19 November 2010 Accepted 23 November 2010 Available online 8 January 2011 A new decision support system has been developed for integrated flood management within the framework of ArcGIS based on realistic two dimensional flood simulations. This system has the ability to interact with and use classified Remote Sensing (RS) image layers and other GIS feature layers like zoning layer, survey database and census block boundaries for flood damage calculations and loss of life estimations. It also provides a user friendly interface which allows construction of user defined criteria, such as stage-damage curves, running computations and visualization of the results. Monte Carlo Simulation method is used to take into account uncertainties in various variables and parameters, and event tree analysis is used to estimate the population dynamics. The analysis of a dam break flood management strategy for Sinclair Dam in Georgia, USA is chosen as a case study to demonstrate the capabilities of the decision support system. The test results compared with HEC-FDA software indicate that this new system provides a very versatile and reliable environment for estimating various flood damage, and may greatly enhance decision making process for future design of the flood proofing facilities. Ó 2010 Elsevier Ltd. All rights reserved. Keywords: Integrated flood management Decision support system 2D simulation ArcGIS Remote sensing Census block 1. Introduction Natural or man-made disasters produce damages with varying degrees of intensity in a given area. Flooding, a frequently occurring hazard, costs human hardship and economic loss whose intensity varies in space. The flood management can be considered as a spatial problem (Simonovic, 2002; Luino et al., 2009). Flood Protection can be achieved through various structural measures such as dykes, diversion channels, reservoirs, and non-structural measures such as flood warning, mass evacuation and etc. Spatial representation of the consequences of these flood mitigation alternatives provides a better insight into their relative effectiveness and facilities decision making (Jonkman et al., 2008; Ernst et al., 2008). Usually, comparison of different flood protection measures and evaluation of their impacts are based on multiple criteria (Bouwer et al., 2009; Pingel and Watkins, 2010). * Corresponding author. Present address: NMP Engineering Consultants, Inc., Executive Plaza III, Suite 300, 11350 McCormick Rd, Cockeysville, MD 21031, USA. Tel.: þ1 410 771 99808; fax: þ1 410 771 9809. E-mail addresses: hqi@nmpengineering.com (H. Qi), altinakar@ncche.olemiss. edu (M.S. Altinakar). 1 Tel.: þ1 662 915 3783; fax: þ1 662 915 7796. 1364-8152/$ e see front matter Ó 2010 Elsevier Ltd. All rights reserved. doi:10.1016/j.envsoft.2010.11.006 These criteria usually include loss of life and urban flood damage. Recent advancements in reliable two dimensional (2D) flood modeling, geographical information system and remote sensing technology open a promising new path for flood hazard management since it takes into account of spatial variability of flood hazard. The main objective of the research described in this paper is to develop a decision support system in widely used GIS environment. Based on the computational results of a 2D flood analysis model, crucial criteria that are usually implemented for evaluating flood hazard, such as loss of life and flood damage (in urban and rural area) can also be calculated within the framework of this system, by interacting with other GIS feature layers like census block layer and remote sensing image. The case study of the dam break flood case of Sinclair Dam in Milledgeville of Georgia, USA is used to test the capabilities of the system. The versatile environment for construction different criteria and use of other GIS feature layer demonstrate that the GIS-based decision support system can provide the user better decision making aid. The results of the case study validated by HEC-RAS and HEC-FDA software (US Army Corps of Engineers, 2010) clearly show that this new approach based on 2D simulation results allows the stake holders to have a better appreciation of the consequences of the flood. 818 H. Qi, M.S. Altinakar / Environmental Modelling & Software 26 (2011) 817e821 2. Developing multiple flood management decision making criteria 2.1. Loss of life estimation with census block boundary According to Graham (1999), loss of life resulting from flooding is highly influenced by three factors: 1) The number of people occupying the floodplain, which is also called Population at Risk (PAR); 2) The amount of warning that is provided to the people exposed to dangerous flooding and 3) The severity level of the flooding. For this research, spatial variability of the above factors is taken into account for each location of the area of interest, which is a completely two dimensional approach. Fig. 1 shows the flowchart to estimate the loss-of-life due to a flood within the decision support system. In order to determine the PAR value, the census block data, which is usually a vector polygon layer (for example, in TIGER format) are used in GIS environment (Poser et al., 2009). Census blocks are small in area showing population variation. After importing this layer into ArcGIS, the population density is first calculated by using the total population of each census block and its area. This feature polygon layer is then converted to a raster layer which has the same cell size as the flood computation results. The cell value, which represents the PAR living and working inside each cell, is reclassified according to the product of population density and the cell area. This operation would obtain a raster layer showing the PAR distribution (Qi et al., 2005). The typical definition of warning time of a flood is the length of time from when the first public warning is issued until the flood wave reaches the first person in the PAR (Aboelata et al., 2002). The flood severity definition is usually associated with the flood depth. Low, medium and high severity can be categorized according to Graham (1999). For instance, the equation for retrieving the life loss of a sub category layer which satisfies a medium flood severity with no warning time in the domain is: ½LossLifeŠ ¼

½DŠ < Hhigh &ð½DŠ > ¼ Hlow Þ &ðð½ATŠ Wissue Þ < Wnw Þ  ½CsŠ  Rf (1) where [LossLife] [D], [AT] and [Cs] represent raster layer of loss of life, flood depth, arrival time and census block information respectively, Hhigh/Hlow are the limits for high/low severity flood depth, Wissue is the initial time of a public warning, Wnw is the time limit for no warning, and Rf is the corresponding fatality rate. The total life loss distribution layer is then obtained by summing up the entire sub-category layers. Parameters of Hhigh, Hlow, Wissue, Wnw and Rf can be found from Graham (1999). GIS and Remote Sensing Census Block boundary layer User Awareness level of population 2.2. Urban flood damage calculation with remote sensing image The second criterion used in the evaluation of the alternatives is the dollar value of damage to the flooded structures within the region of interest. Field surveys or interviews and expert panel opinions are the two primary sources of the data used to develop depth-damage relationships for structural and its content groupings and alternative residence types. Flood depth here refers to the depth of the flooding above or below the first floor of the structure. The percentage damage to the structure refers to the percent of the total depreciated replacement cost of the structure that is damaged (US Army Corp of Engineers, 1997). Remote Sensing (RS) is the recently developed science and technology of obtaining information about an object, area or phenomenon through the analysis of data acquired by a device that is not in contact under investigation (Lillesand, 1999). The satellite image can provide important information by showing various urban land cover features, like vegetation, residential area or water body. Since different land feature types have their inherent spectral reflectance and emittance properties, the RS image is usually classified so that all the pixels in this image fall into certain land over classes or themes. Each class of the land features manifests a unique Digital Number (DN) value. With overlaying of this classified RS image with the flood inundation image, the flood damage calculation can be achieved using arithmetic and relational raster map algebra (Qi et al., 2006). 3. Case study of a dam break flood analysis A dam break flood analysis of Sinclair Dam near Milledgeville, Georgia of the United States has been chosen to test the capability of the designed decision support system. Completed in 1953 by the Georgia Power Company, the Sinclair Dam is located on Oconee River upstream of historical Milledgeville, GA, which was the state capital of Georgia from 1803 to 1868. The dam consists of concrete non-overflow and spillway sections, flanked by reinforced earthen embankments. The span of the dam is 2988 ft (910.74 m). The dam height is 104 ft (31.70 m). Sinclair Dam is classified as a high-hazard dam. The river at the dam drains a total basin area of 13,805 km2. The volumes at the normal storage level and the maximum storage level are 333,000.0 acre-foot (410,749,457 m3) and 490,000.0 acrefoot (604,406,109 m3), respectively. The spillway has a width of 870.0 ft (265.2 m). The maximum discharge is 479,000.0 cfs (13,563 m3/s). The floodplain downstream of the dam has a surface area of about 669 km2. The population (2000 Census) of Milledgeville is about 44,700 (Fig. 2). Census Raster layer (PAR) Number of fatalities Flood warning time CCHE2D FLOOD Flood arrival times Flood depths Time available for escape Fatality Rate Flood severity Empirical Relationship based on past events (USBR) Fig. 1. Flowchart showing the procedure to estimate probable loss-of-life due to a flood. H. Qi, M.S. Altinakar / Environmental Modelling & Software 26 (2011) 817e821 819 Fig. 2. Study area: Milledgeville, Georgia in Southeastern United States (Scale 1:3000). 4. Results and validations The main objective of the floodplain analysis is to evaluate the consequences of the flood resulting from complete Sinclair Dam failure happened on a sunny weekday. The dam break case was assumed to occur at when the reservoir has normal storage level of water. Loss-of-life and urban damage analysis were carried out using the decision support system. The loss-of-life potential is quite high due to the assumption of sudden and complete failure of the entire dam. This is a highly improbable scenario, and the value should be regarded as an upper bound. A gradual failure would probably result in much smaller loss-of-life potential due to not only the reduced peak discharge, thus lower inundation depths and velocities, but also due to longer escape time for the PAR. In fact loss-of-life potential is extremely sensitive to the time when the alarm was given. Based on the user defined probability distributions, the Monte Carlo simulation method was used to take into account the uncertainties. The number of Monte Carlo runs was set to 5000 for each analysis. The Table 1 Input data for loss-of-life estimation with probability distribution functions. Category Name Primary value Probability distribution Related parameters Flood severity Warning time High Low Initial Normal Normal Uniform Variance ¼ 0.35 Variance ¼ 0.5 Range (15, 35) Census Adequate No PAR >20 ft (6.1 m) <15 ft (4.6 m) 20 min (after flood) >70 min <25 min Census raster Uniform Uniform Normal Fatality rate FR Default value N/A Range (60, 80) Range (15, 35) Variance ¼ 6% PAR N/A input data used for the spatial Monte Carlo Analysis of loss of life and flood damage for both urban and rural areas are presented in Tables 1 and 2. “No of houses” means numbers of house units within one cell of DEM used for the analysis. Fig. 3 Left shows a map of the expected values of the estimated loss-of-life potential after 5000 Monte Carlo Simulation runs. Although not shown here, one can easily create maps for other statistical properties such as standard deviation, skewness, kurtosis, etc. Right shows a map of the expected values of the estimated urban damage potential in US dollars based on the three categories of building stock: High Intensity Residential Area (HIRA), Low Intensity Residential Area (LIRA), and Commercial Industrial Transportation Area (CITA). This analysis was carried out using the urban flood damage toolbox of the GIS-based decision support system. The results of loss of life and urban damage computed by CCHE2D-Flood and Decision Support System (DSS) are summarized in Table 3. As compared with the results obtained from same scenario analysis by the widely used and accepted HEC-RAS dam break simulation and HEC-FDA, it is found that the CCHE2D-Flood and DSS system provide very close and accurate estimates (error <5%). The major reason for the difference results from the difference from 1D flood simulation by HEC-RAS and 2D flood simulation by CCHE2D-Flood. Downstream of Sinclair Reservoir sits within the wide FEMA 100-year floodplain of Oconee River, where there is not a well defined valley. 1D model assumes that the flow fills the entire cross section instantly. When flood flow enters a flat area with poorly defined channel, 1D flow hypothesis breaks down, and it may lead to significant errors in estimating arrival time, flow depth, and flow duration. At those locations, 2D model can better simulate the hydrodynamics of flood situation in flat area. The advantages of using the DSS system also lie in that various maps showing spatially 820 H. Qi, M.S. Altinakar / Environmental Modelling & Software 26 (2011) 817e821 Table 2 Input data for urban flood damage analysis with probability distribution functions. Category Name Primary value Prob. distribution Related parameters High intensity residential area (HIRA) No of houses Structure value Content value No of houses Structure value Content value No of houses Structure value Content value Other value HIRA 10 Units $98,000 $75,000 6 Units $120,000 $90,000 8 Units $150,000 $100,000 $30,000 f ¼ 0.68h4 þ 7.78h3 34.52h2 þ 78.7h þ 4.95 f ¼ 2.61h4 þ 25.37h3 78.42h2 þ 95.89h þ 4.31 f ¼ 0.81h4 þ 9.25h3 38.13h2 þ 71.20h þ 4.77 Triangular Normal Normal Triangular Normal Normal Triangular Normal Normal Normal Triangular min ¼ 8, max ¼ 14 Variance ¼ 0.45 Variance ¼ 0.35 min ¼ 4, max ¼ 10 Variance ¼ 0.45 Variance ¼ 0.35 min ¼ 6, max ¼ 12 Variance ¼ 0.25 Variance ¼ 0.35 Variance ¼ 0.4 min ¼ 5%f, max ¼ þ5%f min ¼ 5%f, max ¼ þ5%f Low intensity residential area (LIRA) Commercial industrial transportation area (CITA) Depth e % Damage relationship (h: flood depth in feet, f: damage value in $) LIRA CITA Triangular Triangular min ¼ 5%f, max ¼ þ5%f Fig. 3. Map of the expected values of the estimated loss-of-life potential and urban flood damage obtained following a Monte Carlo simulation with 5000 runs (Scale 1:3000). Table 3 Summary of estimated loss-of-life potential and flood damage results compared with HEC-RAS and HEC-FDA analysis. Category Items CCHE2D-flood and DSS results HEC-RAS and HEC-FDA results Loss of life Number of fatalities 147 person 3.45  10 $ N/A (no such function) 3.50  108 $ 2.62  108 $ 2.71  108 $ 5.81  108 $ 5.72  108 $ 1.18  109 $ 1.19  109 $ Urban flood High intensity residential damage area (HIRA) Low intensity residential area (LIRA) Commercial, industrial transportation area (CITA) Total 8 distributed loss of life and flood damage information can be provide rather than damage report and plan performance in forms and charts by HEC-FDA. 5. Conclusions The present study shows risk and uncertainty analysis based on 2D numerical simulation results, GIS and remote sensing technologies can significantly improve the accuracy of flood hazard assessment. This approach efficiently assists in evaluation and ranking of flood control management strategies, and future design of flood proofing works. The resulting raster/vector maps of the case study showing spatial distribution of loss of life and flood damage can greatly enhance decision making process for future planning of emergency management operations. Currently, the studies are underway to incorporate the discharge-exceedance probability functions and flood frequency probability functions into the spatial risk and uncertainty analysis. It is also important to underline the fact that the currently available relationships used for estimating property damage are generally expressed as a function of flood depth only. The accuracy of predictions can probably be improved by taking into account other detailed information provided by 2D flood modeling, such as flood velocity and duration. Further research is needed to develop such improved damage relationships. References Aboelata, M. et al., GIS Model for Estimating Dam Failure Life Loss. Invited paper. In: The Tenth Engineering Foundation Conference on Risk-Based Decision making in Water Resources: Protection of the Homeland’s Water Resources Systems, Santa Barbara, California. November 2002. Bouwer, L.M., Bubeck, P., Wagtendonk, A.J., Aerts, J.C.J.H., 2009. Inundation scenarios for flood damage evaluation in Polder areas. Natural Hazards and Earth System Sciences 9, 1995e2007. H. Qi, M.S. Altinakar / Environmental Modelling & Software 26 (2011) 817e821 Ernst, J., Dewals, BJ., Giron, E., W Hecq, W., Pirotton, M., 2008, Integrating Hydraulic and Economic Analysis for Selecting Flood Protection Measures in the Context of Climate Change. In: 4th International Symposium on Flood Defence: Managing Flood Risk, Reliability and Vulnerability, Toronto, Ontario, Canada, May 6e8, 2008. Graham, W.J., 1999. A Procedure for Estimating Loss of Life Caused by Dam Failure. Report No. DSO-99e06. Dam Safety Office, US Bureau of Reclamation, Denver, CO. Jonkman, S.N., Bo ckarjovab, M., Kokc, M., Bernardinid, P., 15 May 2008. Integrated hydrodynamic and economic Modelling of flood damage in the Netherlands. Ecological Economics 66 (1), 77e90. Lillesand, Thomas M., 1999. Remote Sensing and Image Interpretation, Fourth ed. John Wiley & Sons, Inc. Press, ISBN 0-471-25515-7. Luino, F., Cirio, C.G., Biddoccu, M., Agangi, A., Giulietto, W., Godone, F., Nigrelli, G., September 2009. Application of a model to the evaluation of flood damage. GeoInformatica 13 (3), 339e353. Pingel, Nathan, Watkins Jr., David, 2010. Multiple flood source expected Annual damage computations. Journal of Water Resources Planning and Management 136 (3), 319e326. Poser, Kathrin, Kreibich, Heidi, Dransch, Doris, 2009. Assessing Volunteered Geographic Information for Rapid Flood Damage Estimation. In: 12th AGILE 821 International Conference on Geographic Information Science 2009. Leibniz Universität Hannover, Germany. Qi, Honghai, Altinakar, M.S., Ying, X., Wang, S.S.Y, 2005. Risk and Uncertainty Analysis in Flood Hazard Management Using GIS and Remote Sensing Technology. In: Proc.. of American Water Resources Association Annual Conference, Seattle, Nov 2005. Qi, Honghai, Altinakar M.S., Youngho Jeon, 2006, A Decision Support Tool for Flood Management under Uncertainty Using GIS and Remote Sensing Technology. In: Proceedings of 7th International Conference on Hydroscience and Engineering, Drexel University, Sept 2006, Philadelphia, United States. Simonovic, S.P., 2002, A Spatial Fuzzy Compromise Programming for Management of Natural Disasters, ICLR Research Paper Series e No. 24. US Army Corp of Engineers, 1997, Depth-Damage Relationships for Structures, Contents and Vehicles and Content-To-Structure Value Ratios (CSVRs) in Support of the Jefferson and Orleans Flood Control Feasibility Studies, Final Report, Baton Rouge, Louisiana, United States; US Army Corps of Engineers., 2010. HEC-FDA: Flood Damage Reduction Analysis User’s Manual, Version 1.2.5. Hydrologic Engineer Center, Davis, California, United States.