Studying soil moisture at a national level through statistical analysis of NASA NLDAS data

Abstract

The purpose of this research is to enable better understanding of current environmental conditions through the relations of environmental variables to the historical record. Our approach is to organize and visualize land surface model (LSM) outputs and statistics in a web application, using the latest technologies in geographic information systems (GISs), web services, and cloud computing. The North American Land Data Assimilation System (NLDAS-2) (http://ldas.gsfc.nasa.gov/nldas/; Documentation: ftp://hydro1.sci.gsfc.nasa.gov/data/s4pa/NLDAS/README.NLDAS2.pdf) drives four LSM (e.g., Noah) (http://ldas.gsfc.nasa.gov/nldas/NLDAS2model.php) that simulate a suite of states and fluxes for central North America. The NLDAS-2 model output is accessible via multiple methods, designed to handle the outputs as time-step arrays. To facilitate data access as time series, selected NLDAS-Noah variables have been replicated by NASA as point-location files. These time series files or ‘data rods’ are accessible through web services. In this research, 35-year historical daily cumulative distribution functions (CDFs) are constructed using the data rods for the top-meter soil moisture variable. The statistical data are stored in and served from the cloud. The latest values in the Noah model are compared with the CDFs and displayed in a web application. Two case studies illustrate the utility of this approach: the 2011 Texas drought, and the 31 October 2013 flash flood in Austin, Texas.