Academic journal article Cartography and Geographic Information Science

Effects of Threshold Drainage Area and Study Region Size on HAZUS Flood Analysis

Academic journal article Cartography and Geographic Information Science

Effects of Threshold Drainage Area and Study Region Size on HAZUS Flood Analysis

Article excerpt

Introduction

HAZUS (abbreviation for HAZards United States) is a GIS-based natural hazard loss estimation software package developed and freely distributed by the United States Federal Emergency Management Agency (FEMA). The current version of HAZUS, HAZUS-MH (HAZUS-MultiHazard), can model three types of hazards: flooding, hurricanes, and earthquakes. Based on inventory data of buildings, vehicles, transportation systems, utility systems, critical facilities, hazardous materials, agricultural commodities, and demographics (1), as well as specified hazard parameters, HAZUS locates hazard areas and estimates the potential physical, social, and economic costs based on engineering knowledge of the damage and hazard loss estimates (FEMA 2005). Hazard loss estimation provides a basis for hazard preparation, mitigation strategy development, and response and recovery planning, and it is essential to decision making at all levels of government. Equipped with information about the potential loss, hazard managers are able to better prioritize mitigation measures to reduce further disaster losses, and local communities can allocate limited resources more efficiently to responding to, and recovering from, disaster events.

This study intends to study how the HAZUS parameter of threshold drainage area affects flood modeling results. A threshold drainage area (thereafter drainage threshold) is a parameter that a user specifies to delineate a synthetic stream network from a Digital Elevation Model (DEM). It is the minimum accumulation area draining into a stream network. Decreasing drainage threshold will generate a more detailed stream network with additional tributaries. Theoretically, the minimum drainage threshold is only limited by the resolution of available DEM data (FEMA 2009a). Past studies of hydrological modeling often regard a drainage threshold as optimal if the accordingly developed drainage network matches the real-world stream network (Gall et al. 2007; Moussa and Bocquillon 1996; Montgomery and Foufoula-Georgiou 1993). The realistic drainage threshold may change with local relief, soil, geology, and climate conditions (Colombo et al. 2007; Giannoni et al. 2005; Roth et al. 1996). Nevertheless, stream networks developed in HAZUS are for flood modeling purposes and not necessarily the same as real-world stream networks. Apparently, appropriate drainage thresholds for HAZUS flood modeling largely depend on user's need (FEMA 2009b). A small drainage threshold delineates small stream reaches and thus a maximum coverage of potential floodplains, though the offshoot floodplains may be spurious (Muthukumar 2005). As most hydrological studies aim to use real drainage thresholds to delineate stream networks, literature on the effects of specified threshold parameter on hydrological modeling is limited. Moussa and Bocquillon (1996) and Stepinski and Collier (2004) reported that the total length of stream networks decreases with increasing drainage threshold. Li et al. (2008) found that, as drainage threshold increases, the outflows of basins became slower, the discharge of flood peak decreases, and the basin mean concentration time becomes longer. Liu et al. (2003) indicated that using a large threshold value will result in a longer flow time and a prolonged flow response at the end of the flow path.

Another objective of this study is to investigate the effects of study region size on flood modeling. A HAZUS study region is the area defined by a user in the beginning of analysis to allow HAZUS to prepare inventory and DEM data (for delineating stream networks). Note that users are allowed to select stream reaches of interest as a hazard scenario to conduct floodplain delineation and loss estimation. We expect that different sizes of DEM data will produce different stream networks and, consequently, different flood modeling results. Limited studies have mentioned the effects of study region size (channel network extent) on hydrological modeling. …

Search by... Author
Show... All Results Primary Sources Peer-reviewed

Oops!

An unknown error has occurred. Please click the button below to reload the page. If the problem persists, please try again in a little while.