Faculty Interaction

Indrajeet Chaubey

Department of Agricultural and Biological Engineering
Purdue University

Brief Project Description

Factors such as population growth, climate change, urbanization and the growing demand for biofuels raise concerns that the availability of fresh water may be an issue bigger than the supply of oil in decades to come. Of a population of roughly 6.7 billion globally, more than a billion already lack access to clean water, according to World Bank estimates. Local water shortages in the U.S. have made headlines from South Florida to Northern California.

Indrajeet Chaubey and colleagues, using ITaP research computing resources, are developing new ways to better understand the myriad factors—and the interplay among them—that can affect water quality and availability. The goal is to improve how we manage precious water resources.

Chaubey, a Purdue agricultural and biological engineering and earth and atmospheric sciences professor, is an ecohydrologist who's examining watersheds in Arkansas, where he worked before coming to Purdue, and around Indiana. His research is part of a U.S. Department of Agriculture Cooperative State Research, Education, and Extension Service (CSREES) program looking at different types of watersheds in locations nationwide.

Chaubey and colleagues make extensive use of the USDA's Soil Water Assessment Tool (SWAT). They use modeling to extrapolate from the data they collect and to ask “What if?” questions about future water quality and availability given various changes in and interactions among conditions. Those conditions can include weather and climate, naturally, but also things like sediment and pesticide transport and changes in land cover from, for example, the installation of filter strips, grazing or crop rotation, the planting of previously marginal lands to grow biofuels, urban growth, and more.

 

The researchers run their models tens of thousands of times while adjusting parameters to capture the complexity and time scales involved. The work could take years to run even on a very powerful desktop computer, but the high-throughput computing resources available through ITaP's Rosen Center, such as Purdue's Condor pool, can reduce that to weeks, days or hours. The researchers can run the models more frequently and examine many more scenarios as a result, gaining a much more comprehensive picture.

Condor is a distributed computing system developed at the University of Wisconsin to tap idle processors for large research jobs. Purdue and partners operate a multi-campus Condor pool, called DiaGrid, with tens of thousands of processors in supercomputing clusters, campus computer labs and office machines offering, in aggregate, hundreds of teraflops of performance.

The Condor pool, which Purdue also makes available on the National Science Foundation's TeraGrid research network, allows Chaubey's group to run lots of scenarios at the same time by parceling them out to thousands of processors, while handling job scheduling and collecting the results. ITaP research scientists have developed software tools for scheduling and distributing jobs that enhance Condor and increase its effectiveness. The Rosen Center also assisted in getting SWAT running on Condor, porting SWAT from Windows to Condor's primarily Linux environment and tuning the software for the group's input files, for mining the data efficiently and for producing output files in the desired format.

The researchers have done some of the most comprehensive watershed modeling ever in analyzing the impact of a variety of Best Management Practices (BMPs) on water quality and availability in a watershed. Ultimately, the goal is to be able to perform such analyses on a regional scale, where the results can have real policy implications.

 

 

More information:

https://engineering.purdue.edu/~ichaubey/index.html

Last updated 06/03/09