Lake and Reservoir Evaporation Estimation: Sensitivity Analysis and Ranking Existing Methods

Water when harvested is commonly stored in dams, but approximately up to half of it may be lost due to evaporation leading to a huge waste of our resources. Estimating evaporation from lakes and reservoirs is not a simple task as there are a number of factors that can affect the evaporation rate, notably the climate and physiography of the water body and its surroundings. Several methods are currently used to predict evaporation from meteorological data in open water reservoirs. Based on the accuracy and simplicity of the application, each of these methods has advantages and disadvantages. Although evaporation pan method is well known to have significant uncertainties both in magnitude and timing, it is extensively used in Iran because of its simplicity. Evaporation pan provides a measurement of the combined effect of temperature, humidity, wind speed and solar radiation on the evaporation. However, they may not be adequate for the reservoir operations/development and water accounting strategies for managing drinking water in arid and semi-arid conditions which require accurate evaporation estimates. However, there has not been a consensus on which methods were better to employ due to the lack of important long-term measured data such as temperature profile, radiation and heat fluxes in most lakes and reservoirs in Iran. Consequently, we initiated this research to find the best cost−effective evaporation method with possibly fewer data requirements in our study area, i.e. the Doosti dam reservoir which is located in a semi-arid region of Iran. Our study site was the Doosti dam reservoir located between Iran and Turkmenistan borders, which was constructed by the Ministry of Water and Land Reclamation of the Republic of Turkmenistan and the Khorasan Razavi Regional Water Board of the Islamic Republic of Iran. Meteorological data including maximum and minimum air temperature and evaporation from class A pan were acquired from the Doosti Dam weather station. Relative humidity, wind speed, atmospheric pressure and precipitation were acquired from the Pol−Khatoon weather station. Dew point temperature and sunshine data were collected from the Sarakhs weather station. Lake area was estimated from hypsometric curve in relation to lake level data. Temperature measurements were often performed in 16−day periods or biweekly from September 2011 to September 2012. Temperature profile of the lake (required for lake evaporation estimation) was measured at different points of the reservoir using a portable multi−meter. The eighteen existing methods were compared and ranked based on Bowen ratio energy balance method (BREB). The estimated annual evaporation values by all of the applied methods in this study, ranged from 21 to 113mcm (million cubic meters). BREB annual evaporation obtained value was equal to 69.86mcm and evaporation rate averaged 5.47mm d-1 during the study period. According to the results, there is a relatively large difference between the obtained evaporation values from the adopted methods. The sensitivity analysis of evaporation methods for some input parameters indicated that the Hamon method (Eq. 16) was the most sensitive to the input parameters followed by the Brutsaert−Stricker and BREB, and radiation−temperature methods (Makkink, Jensen−Haise and Stephen−Stewart) had the least sensitivity to input data. Besides, the air temperature, solar radiation (sunshine data), water surface temperature and wind speed data had the most effect on lake evaporation estimations, respectively. Finally, all evaporation estimation methods in this study have been ranked based on RMSD values. On a daily basis, the Jensen−Haise and the Makkink (solar radiation, temperature group), Penman (Combination group) and Hamon (temperature, day length group) methods had a relatively reasonable performance. As the results on a monthly scale, the Jensen−Haise and Makkink produced the most accurate evaporation estimations even by the limited measurements of the input data. This study was carried out with the objective of estimating evaporation from the Doosti dam reservoir, and comparison and evaluation of conventional method to find the most accurate method(s) for limited data conditions. These examinations recognized the Jensen−Haise, Makkink, Hamon (Eq. 17), Penman and deBruin methods as the most consistent methods with the monthly rate of BREB evaporation estimates. The results showed that radiation−temperature methods (Jensen−Haise and Makkink) have appropriate accuracy especially on a monthly basis. Also deBruin, Penman (combination group), Hamon and Papadakis (temperature group) methods produced relatively accurate results. The results revealed that it is necessary to calibrate and adjust some evaporation estimation methods for the Doosti dam reservoir. According to the required input data, sensitivity and accuracy of these methods, it can be concluded that Jensen−Haise and Makkink were the most appropriate methods for estimating the lake evaporation in this region especially when measured data were not available.