SPATIO -TEMPORAL VARIABILITY OF POTENTIAL EVAPOTRANSPIRATION IN SRI LANKA AND ITS APPLICATIONS IN AGRICULTURAL PLANNING

摘要

Potential evapotranspiration (PET) is a climatic variable which represents the "water demand of the atmosphere". In Sri Lanka, PET values approximated to the nearest point observation of Pan evaporation are used for agricultural planning such as estimation of crop water requirement, irrigation scheduling etc. Hence planners are hindered with lack of reasonable approximations of PET values for their respective irrigation schemes, which lead to inaccurate planning, resulting in low water productivity.Therefore, a study was carried out to identify the spatial and temporal variability of PET over the Island in order to facilitate their usage in agricultural planning. Daily Pan evaporation values collected during the period 1970-2003 for 49 locations within the Island were used for the study. Pan evaporation values were converted into PET values using the pan coefficient. Locations of Pan evaporation measuring points were geo-referenced and surface models indicating the spatial distribution of daily PET values were developed for each month, covering the entire Island, using ArcView GIS Software. A subset of the total dataset which was not used for the analysis, was used to validate the interpolated model surfaces against the calculated PET. Further, arelationship with PET and altitude was established. The study highlights the spatio-temporal variability of PET and proposes representative values for the different ecological zones for each calendar month. It also reveals that the lowest mean PET values for all the ecological zones occur in December, while the highest evaporation losses occur in March. Up Country Wet and Intermediate zones exhibit lowest PET values almost throughout the year. The study shows that in regions above 1200 m msl, the effect of temperature on atmospheric evaporation demand is more prominent compared to that in areas below. Since correlation between interpolated PET and calculated point PET are more than 80%, such values can be safely used in agricultural planning when point PET values are not available.