Getting Answers Quickly Continued from page 27 base in ArcSDE as well as tables in a CAMA (computer-assisted mass appraisal) system that contains ownership, property classification, and valuation information for more than 248,000 parcels. Each record in the parcel geodatabase and CAMA table is identified by a unique parcel number, which was used to link these tables to obtain the ownership and property class information. Calculating the Flood Extent Runoff water from a hydrologic unit (watershed) is always drained to a particular stream. Thus, the stream floods the area within its watershed. The flooded areas in the county were divided into six watersheds for flood boundary calculation. Watershed boundaries were derived from a resampled (32-foot) sink-free DEM using the Hydrology tools in the ArcGIS Spatial Analyst extension. U.S. Geological Survey stream gauge sites were visually identified from the orthophotos. The ground elevation (m-values measured in feet) at these locations were obtained from the DEM. The flood boundary was calculated by determining the maximum flood elevation in feet (H) for a watershed, then identifying the areas in the DEM having values less than H. The maximum flood elevation is calculated as where SG is the maximum stream gauge reading in feet. Aerial photos of a subdivision in Cobb County before (top) and after (bottom) the flood H = M + SG Inundation areas were calculated separately for each watershed. The DEM grid (8-foot resolution) clipped by watershed boundary, was level sliced to identify the area below the maximum flood elevation. The flood cover raster was then converted to a polygon feature to obtain the flood boundary. For generating the preliminary flood boundary, slope of the streams was not taken into account. Modifications were made to the flood boundary using observations from the field. Flood boundaries from all the watersheds were then combined into a single polygon. The flood-affected parcels were identified by intersecting the flood boundary with the parcel feature layer. Parcels completely falling within the flood boundary were classified as fully inundated. The minimum, mean, and maximum floodwater heights for each parcel were calculated by estimating the respective elevations from the DEM (using the zonal statistics tool), according to the following formulas on page 29. Determining the Flood Boundary Deriving the flood boundary as fast as possible was the main goal of the EOC request. Necessary modifications were made to the flood boundary using the field observations. The corrected flood boundary was then used to identify inundated parcels. Stream gauges served as the primary source of water level information during the flood event. Readings from seven gauges located on various streams/rivers were used to determine the elevation of floodwater at these locations. The crux of any terrain-based analysis is a three-dimensional representation of the terrain. Hence the accuracy of the output is highly dependent on the terrain 28 ArcUser Spring 2010 model. A highly accurate digital elevation model (DEM) with a horizontal resolution of 8 feet, which was derived from the county’s 2-foot contours and stored as a float doubleformatted raster image, is available from the county’s enterprise GIS data warehouse. To verify the flood boundary and make corrections during field verification, the original 2-foot contour data was used. A line feature class of the streams was used to identify the stream centerlines and names. Orthophotos and oblique aerial photography were used in conjunction with photographs taken from aircraft to verify the flood boundary and the floodwater heights at known locations. The county maintains a parcel geodata- www.esri.com