Alfred Kalyanapu

Assessment of GPU computational enhancement to a 2D flood model

This paper presents a study of the computational enhancement of a Graphics Processing Unit (GPU) enabled 2D flood model. The objectives are to demonstrate the significant speedup of a new GPU-enabled full dynamic wave flood model and to present the effect of model spatial resolution on its speedup. A 2D dynamic flood model based on the shallow water equations is parallelized using the GPU approach developed in NVIDIAs Compute Unified Development Architecture (CUDA). The model is validated using observations of the Taum Sauk pump storage hydroelectric power plant dam break flood event. For the Taum Sauk flood simulation, the GPU model speedup compared to an identical CPU model implementation is 80x-88x for computational domains ranging from 65.5 k to 1.05 M cells. Thirty minutes of event time were simulated by the GPU model in 2 min, 15 times faster than real time. An important finding of the analysis of model domain size is the GPU model is not constrained by model domain extent as is the CPU model. Finally, the GPU implementation is shown to be scalable compared with the CPU version, an important characteristic for large domain flood modeling studies.

CHANNEL MORPHOLOGY TOOL (CMT): A GIS-BASED AUTOMATED EXTRACTION MODEL FOR CHANNEL GEOMETRY

This paper describes an automated Channel Morphology Tool (CMT) developed in ArcGIS 9.1 environment. The CMT creates cross-sections along a stream centerline and uses a digital elevation model (DEM) to create station points with elevations along each of the cross-sections. The generated cross-sections may then be exported into a hydraulic model. Along with the rapid cross-section generation the CMT also eliminates any cross-section overlaps that might occur due to the sinuosity of the channels using the Cross-section Overlap Correction Algorithm (COCoA). The CMT was tested by extracting cross-sections from a 5-m DEM for a 50-km channel length in Houston, Texas. The extracted cross-sections were compared directly with surveyed cross-sections in terms of the cross-section area. Results indicated that the CMT-generated cross-sections satisfactorily matched the surveyed data.

GIS-BASED PREDICTION OF HURRICANE FLOOD INUNDATION

A simulation environment is being developed for the prediction and analysis of the inundation consequences for infrastructure systems from extreme flood events. This decision support architecture includes a GIS-based environment for model input development, simulation integration tools for meteorological, hydrologic, and infrastructure system models and damage assessment tools for infrastructure systems. The GIS-based environment processes digital elevation models (30-m from the USGS), land use/cover (30-m NLCD), stream networks from the National Hydrography Dataset (NHD) and soils data from the NRCS (STATSGO) to create stream network, subbasins, and cross-section shapefiles for drainage basins selected for analysis. Rainfall predictions are made by a numerical weather model and ingested in gridded format into the simulation environment. Runoff hydrographs are estimated using Green-Ampt infiltration excess runoff prediction and a 1D diffusive wave overland flow routing approach. The hydrographs are fed into the stream network and integrated in a dynamic wave routing module using the EPAs Storm Water Management Model (SWMM) to predict flood depth. The flood depths are then transformed into inundation maps and exported for damage assessment. Hydrologic/hydraulic results are presented for Tropical Storm Allison.

Integration of SWMM into a Dam Break, Hurricane, and Extreme Flood Modeling and Damage Assessment Framework

SWMM5 has been seamlessly integrated with a Geographic Information System (GIS) for simulation of inundation and analysis of consequences resulting from extrem…

GIS-based 1-d diffusive wave overland flow model

This paper presents a GIS-based 1-d distributed overland flow model and summarizes an application to simulate a flood event. The model estimates infiltration using the Green-Ampt approach and routes excess rainfall using the 1-d diffusive wave approximation. The model was designed to use readily available topographic, soils, and land use/land cover data and rainfall predictions from a meteorological model. An assessment of model performance was performed for a small catchment and a large watershed, both in urban environments. Simulated runoff hydrographs were compared to observations for a selected set of validation events. Results confirmed the model provides reasonable predictions in a short period of time.