The Terrestrial Water Cycle: Modeling and Data Assimilation across Catchment Scales

Abstract

Hydrologic science is currently undergoing a revolution in which the field is being transformed by the multitude of available new data streams. Historically, hydrologic models that have been developed to answer basic questions about rainfall–runoff relationships, surface water and groundwater storage and fluxes, land– atmosphere interactions, and so forth, have been optimized for previously data-limited conditions. However, with the advent of remote sensing technologies and increased computational resources, the environment for water cycle researchers has fundamentally changed to one where there is now a flood of spatially distributed and time-dependent data. This transition from a “data poor” to a “data rich” environment has allowed for an increase in the scope and scale of research questions and practical problems that can be addressed, but also requires that we fundamentally change our approaches to solving these problems. Largely due to a lack of historical data, the mean states and fluxes in the terrestrial water cycle remain poorly characterized. Development of diagnostic and predictive frameworks for characterizing the mean states and their natural variability is a crucial first step in understanding how they may be altered under anthropogenic climate change. The buildup of greenhouse gases during the past century is postulated to be responsible for global warming through increased radiative forcing of the earth system. Changes in the energy balance of earth are likely to affect other climatic factors besides temperature, including the components of the terrestrial water cycle. For example, atmospheric moisture amounts (and thus precipitable water) in the Northern Hemisphere are generally observed to have increased since 1973 (Ross and Elliott 2001). An increase of a few percent in atmospheric moisture is expected to lead to stronger rainfall rates (Trenberth et al. 2003). In terms of evaporation, even though a very robust finding in all climate models with global warming is for an increase in potential evapotranspiration (IPCC 2001), recent studies (e.g., Peterson et al. 1995; Brutsaert and Parlange 1998; Ohmura and Wild 2002; Roderick and Farquhar 2002) indicate there is still a debate about whether land evaporation will increase or decrease under a warming climate. In any case, a reduction in or an augmentation of the global terrestrial evaporation, combined with the increasing trend in global precipitation, will have to be balanced by changes in runoff and soil moisture, and this will effectively spin down or accelerate the hydrological cycle (Ramanathan et al. 2001). Milly et al. (2002) investigated the changes in risk of large floods using both streamflow observations and numerical simulations of anthropogenic climate change. They found that the frequency of large floods increased substantially during the twentieth century, and that this trend would continue into the twenty-first century. However, the frequency of floods with shorter return periods did not increase significantly. Therefore, evidence would seem to indicate that a potential consequence of global warming is the acceleration of the hydrological cycle. The ability to detect * Organizers of the International Workshop on Terrestrial Water Cycle.