Evaporation and Modeling Water Availability in the Savannah River Basin.

摘要

Four methods for estimating lake evaporation are presented and compared. Three are mass transfer methods that incorporated high resolution satellite imagery obtained from the MODIS sensor on the Terra and Aqua satellites. The fourth is the traditional pan method, which used monthly derived pan coefficients and pan evaporation measurements from a Class A evaporation pan located in Clemson, SC.;The four evaporation methods were used to estimate lake evaporation from the five major lakes in the Savannah River Basin for a period of more than one decade. Analysis of the evaporation rates clearly illustrated the uncertainty in daily, monthly, and yearly lake evaporation estimates derived from using different evaporation parameterizations. Results showed significant differences in the seasonal evaporation patterns between the mass transfer methods, which produced peak evaporation in the early fall months, and the pan method, which showed peak evaporation in the middle of the summer. In fact, there was virtually no correlation between the daily pan and daily mass transfer evaporation rates. Monthly and yearly evaporation estimates started to become more correlated, though the extent of the correlation varied from lake-to-lake, with the correlation increasing with decreasing lake depth.;Uncertainty in lake evaporation estimates was present. The effect of this uncertainty and its role in water-availability predictions within the Savannah River Basin were evaluated. Basin hydrologic modeling under historical and future water use scenarios were simulated for 70 and 57 years, respectively. The results showed significant uncertainty in the predicted available water during low-flow conditions, which corresponds to basin drought periods. Under normal-flow conditions, uncertainty in lake evaporation estimates did not show uncertainty in the water-availability predictions, due to an abundant supply of water during these conditions. For all lakes and evaporation methods presented, uncertainty in water availability increased with increasing water consumption.;Basin scale return periods were determined for an extreme hydrologic event, defined as each lake falling within 50 % of the annual available storage volume. Under the historical water use scenario, the observed uncertainty in the predicted return periods was approximately 7 years, while the future water use scenario experienced an uncertainty of 22 years. This represents a 214 % increase in the uncertainty in predicted water availability, due to population and industry growth, along with uncertainty in evaporation estimates. This type of uncertainty limits the predictive capabilities of the current basin model and will ultimately constrain the development of resilient drought and water-management plans within the Savannah River Basin.