STUDIES ON EVAPORATION FROM OPEN WATER SURFACES IN TROPICAL CLIMATE

Abstract

Evaporation is quite often a deciding factor in the water resources planning and management in arid and semi-arid tropics. New dimensions to future planning and management have been added by the issue of global warming. Inspite of its significance, however, precise estimation of evaporation still remains one of the challenging tasks because evaporation involves complex interactions and interrelationships of various meteorological factors, which vary from region to region and, also temporally. A number of studies have been reported on evaporation from open water surfaces for temperate regions, but systematic studies for tropical regions are few. So, keeping in view the needs to study different aspects of evaporation for semi arid tropical climates, the present investigations were carried out using six years of daily data of various meteorological parameters for the study area, Udaipur in Rajasthan, India. Various aspects of evaporation were investigated. Estimates of actual vapour pressure and actual evaporation are needed for the investigations. As these are not directly measured, these have been estimated indirectly. Actual vapour pressure values have been obtained using the dew point temperature while the Penman combination model has been used as a standard method to obtain actual evaporation from open water surfaces. Three FAO-56 models of estimating vapour pressure from temperature and relative humidity and, four evaporation models, with less data requirements, namely Jensen-Haise Model, Stephens-Stewart Model, Priestly-Taylor Model and Brutsaert Model, have been evaluated for their performance and suitability. Results indicate that calibration of the models for local conditions significantly improves their performance. It has been observed that for the semi arid study region in tropical India, calibrated Jansen-Haise Model is found to be the best approach where data availability is limited. Evaluation of the three FAO-56 vapour pressure models indicates that all the three models predict the vapour iii pressure reasonably well with Model-1 performing better than the remaining two. Further, it has been observed that variable inputs of vapour pressure (obtained as outputs from different vapour pressure models) do not significantly affect the estimates of evaporation, indicating that the energy balance component plays more significant role than the mass transfer component in semi arid tropics. However, since the existing models of vapour pressure are empirical in nature and since there are no clear cut guidelines presently for selecting the best model, a rational approach has been developed for estimating air vapour pressure from temperature and relative humidity, based on clear theoretical considerations assuming various non-linear variations. Pan coefficients have been derived for the mesh covered as well as open Class-A Pans for the study area, for different time scales such as daily, monthly and seasonal, using the least square optimization method. Since 0.7 is commonly used as pan coefficient, errors caused by use of0.7 as pan coefficient for different time scales have been evaluated. The errors are observed to be very high, being more than 50% for monsoon and winter months. Evaluation of ANN technique for modelling evaporation and vapour pressure has been carried out. An ANN model has been developed for open water evaporation. Daily data of maximum temperature, minimum temperature, maximum relative humidity, minimum relative humidity, sunshine hours and wind velocity were used as inputs. It has been observed that the ANN model can predict evaporation accurately and better than the calibrated Jansen-Haise model. Further, it has been observed that evaporation can also be reasonably modeled using limited data of temperature variables only. However, the performance of such an ANN model is better when maximum temperature and minimum temperature are used as inputs rather than only mean temperature. An attempt has also been made to evaluate the potential of ANN for vapour pressure modelling using only mean temperature and only mean relative humidity as inputs. However, ANN modelling of vapour pressure using only mean temperature data has been found to be not feasible. IV To study the impact of water hyacinth on the evaporation rates, a field experiment was also conducted using two Class-A pans with water hyacinth being grown in one pan and only water in the other. Evapotranspiration rates of water hyacinth are observed to be very high and it has been found that on an average water hyacinth ET losses are more by 4.64 for day time, 2.7 for night time and 3.17 for 24-hr time scale, than the pan evaporation. Keeping mview the global concern for studying the impact ofclimate change on water resources, analysis of sensitivity of evaporation to rising temperature has been carried out using a simple climatic variability approach wherein a hypothetical increase of 1°C in daily mean temperature has been assumed. It has been observed that rising temperature would cause changes in the regimes ofvarious other meteorological parameters. Very high variations in vapour pressure deficit (VPD) are expected during monsoon. As aresult, daily evaporation rates for the study area are expected to rise significantly (upto 30%) during the monsoon and to some extent (upto 8%) during the winter. However, contrary to the common perception, they are expected to fall by about 5% during the summer. Further, it has been observed that the changes in evaporation regime would be influenced most by the changes in the regime of vapour pressure deficit followed by maximum temperature, net long-wave radiation, saturation vapour pressure and net radiation in that order.