Web-based stochastic modeling and visualization
Location
Eccles Conference Center
Event Website
http://water.usu.edu
Start Date
4-1-2014 6:45 PM
End Date
4-1-2014 6:50 PM
Description
Stochastic modeling is a valuable tool to account for and investigate the effects of variability associated with hydrologic modeling. However, when working with high resolution, distributed, hydrologic models several challenges exist which limit the feasibility of performing sophisticated stochastic analysis. Among these challenges is the processing power required to run the hundreds or thousands of simulations necessary. Distributed, hydrologic models often require significant time to run a single simulation. Running thousands of simulations using traditional computing methods can be prohibitive. Another challenge associated with stochastic modeling is the data processing required to explore the results. Spatially distributed data from thousands of simulations must be compared and statistically analyzed using geoprocessing tools. Finally, synthesizing and visualizing the results is also a challenge when the volume of data is large. These challenges limit the amount of stochastic analysis that can be done using traditional modeling resources. We have addressed these challenges by developing a web-based hydrologic modeling framework where existing models can be modified, run, and visualized. Here we present a stochastic risk analysis for flooding using the Gridded Surface Subsurface Hydrologic Analysis (GSSHA) model. The computational barrier to stochastic modeling is overcome by leveraging cloud-computing resources through Windows Azure Cloud Services. The geoprocessing required to synthesize the results are automated using 52°North web processing services (WPS). Results are visualized in a customized mapping environment using the Google Earth plugin. Through this framework the barriers to stochastic modeling are lowered enabling it to become a viable tool in hydrologic modeling.
Web-based stochastic modeling and visualization
Eccles Conference Center
Stochastic modeling is a valuable tool to account for and investigate the effects of variability associated with hydrologic modeling. However, when working with high resolution, distributed, hydrologic models several challenges exist which limit the feasibility of performing sophisticated stochastic analysis. Among these challenges is the processing power required to run the hundreds or thousands of simulations necessary. Distributed, hydrologic models often require significant time to run a single simulation. Running thousands of simulations using traditional computing methods can be prohibitive. Another challenge associated with stochastic modeling is the data processing required to explore the results. Spatially distributed data from thousands of simulations must be compared and statistically analyzed using geoprocessing tools. Finally, synthesizing and visualizing the results is also a challenge when the volume of data is large. These challenges limit the amount of stochastic analysis that can be done using traditional modeling resources. We have addressed these challenges by developing a web-based hydrologic modeling framework where existing models can be modified, run, and visualized. Here we present a stochastic risk analysis for flooding using the Gridded Surface Subsurface Hydrologic Analysis (GSSHA) model. The computational barrier to stochastic modeling is overcome by leveraging cloud-computing resources through Windows Azure Cloud Services. The geoprocessing required to synthesize the results are automated using 52°North web processing services (WPS). Results are visualized in a customized mapping environment using the Google Earth plugin. Through this framework the barriers to stochastic modeling are lowered enabling it to become a viable tool in hydrologic modeling.
https://digitalcommons.usu.edu/runoff/2014/2014Posters/5