Date of this Version
Published in Remote Sensing of Drought: Innovative Monitoring Approaches, edited by Brian D. Wardlow, Martha C. Anderson, & James P. Verdin (CRC Press/Taylor & Francis, 2012).
Drought is a pervasive natural climate hazard that has widespread impacts on human activity and the environment. In the United States, droughts are billion-dollar disasters, comparable to hurricanes and tropical storms and with greater economic impacts than extratropical storms, wildfires, blizzards, and ice storms combined (NCDC, 2009). Reduction of the impacts and increased preparedness for drought requires the use and improvement of monitoring and prediction tools. These tools are reliant on the availability of spatially extensive and accurate data for representing the occurrence and characteristics (such as duration and severity) of drought and their related forcing mechanisms. It is increasingly recognized that the utility of drought data is highly dependent on the application (e.g., agricultural monitoring versus water resource management) and time (e.g., short- versus long-term dryness) and space (e.g., local versus national) scales involved. A comprehensive set of drought indices that considers all components of the hydrological–ecological–human system is necessary. Because of the dearth of near-real-time in situ hydrologic data collected over large regions, modeled data are often useful surrogates, especially when combined with observations from remote sensing and in situ sources.
This chapter provides an overview of drought-related activities associated with the North American Land Data Assimilation System (NLDAS), which purports to provide an incremental step toward improved drought monitoring and forecasting. The NLDAS was originally conceived to improve short-term weather forecasting by providing better land surface initial conditions for operational weather forecast models. This reflects increased recognition of the role of land surface water and energy states, such as surface temperature, soil moisture, and snowpack, to atmospheric processes via feedbacks through the coupling of the water and energy cycles. Phase I of the NLDAS (NLDAS-1; Mitchell et al., 2004) made tremendous progress toward developing an operational system that gave high-resolution land hydrologic products in near real time. The system consists of multiple land surface models (LSMs) that are driven by an observation-based meteorological data set both in real time and retrospectively. This work resulted in a series of scientific papers that evaluated the retrospective data (meteorology and model output) in terms of their ability to reflect observations of the water and energy cycles and the uncertainties in the simulations as measured by the spread among individual models (Pan et al., 2003; Robock et al., 2003; Sheffield et al., 2003; Lohmann et al., 2004; Mitchell et al., 2004; Schaake et al., 2004). These evaluations led to the implementation of significant improvements to the LSMs in the form of new model physics and adjustments to parameter values and to the methods and input meteorological data (Xia et al., 2012). The system has since expanded in scope to include model intercomparison studies, real-time monitoring, and hydrologic prediction and has inspired other activities such as high-resolution land surface modeling and global land data assimilation systems (e.g., the Global Land Data Assimilation System [GLDAS], Rodell et al., 2004; the Land Information System [LIS], Kumar et al., 2006).