The Drivers of Spatial and Temporal Variability of Potential Evaporation Across CONUS: Laying Poor Parameterizations to Rest With A First-order, Second-moment Variability Analysis
Abstract
Widely used approaches for estimating potential evaporation (Eo) in hydrologic models employ temperature-driven schemes that omit or indirectly parameterize the influence of other dominant drivers of Eo variability. Errors in Eo estimation are important because they affect the quantification of evapotranspiration, a major water balance variable to which Eo is an upper limit, and consequently hydrologic analyses for objectives such as streamflow prediction, drought monitoring, and climate change impact assessment. Estimating Eo on the basis of physically appropriate forcings is essential to reduce extraneous modeling uncertainty and recognize key sources of variability, and thereby to improve dependent hydrologic analyses. To understand the sensitivity of Eo variability to its drivers, we present a mean-value, second-moment uncertainty analysis, as applied to a 30-year, CONUS-distributed dataset of daily synthetic pan evaporation. For drivers, we use six North American Land Data Assimilation System (NLDAS) variables: temperature, specific humidity, station pressure, wind speed, and downwelling shortwave and longwave radiation fluxes. Accounting for both the sensitivity of Eo to its drivers and their observed variabilities, the variability of Eo is decomposed across CONUS at various time scales into contributions from each driver. We find that, contrary to the assumption of much hydrologic practice, temperature is not always the most significant driver of temporal variability in Eo, particularly at sub-annual time scales. Instead, depending on the region and the season, one of four drivers dominates. In many regions, parameterizations based solely on temperature should be avoided at any time scale. This result has clear implications for modeling Eo in operational hydrology and within analyses of climate change and variability. This presentation describes the analysis concept and summarizes the results across CONUS.ank of each driver's contribution to variability of Eo at the inter-annual time-scale. Results vary significantly according to time-scale.
- Publication:
-
AGU Fall Meeting Abstracts
- Pub Date:
- December 2012
- Bibcode:
- 2012AGUFM.H44A..06H
- Keywords:
-
- 1616 GLOBAL CHANGE / Climate variability;
- 1818 HYDROLOGY / Evapotranspiration;
- 1873 HYDROLOGY / Uncertainty assessment;
- 9350 GEOGRAPHIC LOCATION / North America