Crop Evapotranspiration Study On Evolution Rule And Forecast Model In The West Of Liaoning Province

Accurate estimates of crop water requirement is of great significance to thedevelopment of water-saving agriculture. It is an important method recommended byFAO to refer to crop evapotranspiration to estimate crop water requirement.Therefore, to estimates the crop water requirement, the prediction accuracy of thereference of crop evapotranspiration plays a vital role.In this paper, the following research is based on the meteorological data in thewest of Liaoning province from16meteorological stations from1960~2006:Climate tendency coefficient method is used to analyze interannual variabilityand seasonal variation of meteorological elements in western Liaoning. The resultsshowed that: there was a very obvious upward trend in the annual temperature inwestern Liaoning Province; the data of annual wind speed and annual averagesunshine hours showed a significant downward trend; relative humidity showed aweak upward trend; annual rainfall and the amount of evaporation both showed adownward trend. The region’s seasonal changes and interannual variations arebasically the same.The Penman-Montieth formula was used to estimate month reference cropevapotranspiration (ET0) from various weather stations in western Liaoning; climatetrend coefficients were used to analyze the inter-annual variation characteristics andseasonal variation characteristics of ET0; gray relational analysis method was used toanalyze the meteorological elements influencing changes of ET0. The results showedthat: in every region except Fuxin, the ET0showed a significant downward trend; interms of season changes, a significant downward trend was showed. In terms ofmeteorological factors, whether the inter-annual influence factors or seasoninfluence factors, the most important meteorological factors are wind speed andsunshine.Based on the ET0of Fuxin from1960to2005, by using the gray forecastingmethod, gray GM (1,1) model, gray metabolic GM (1,1) model (XGM (1,1) model),and metabolic GM (1,1) model (PXGM (1,1) model) based on the particle swarmoptimization are established. The application results showed that: as far as theprediction of ET0is concerned, the three prediction models demonstrated goodapplicability. The modeling process of gray model GM (1,1) is simple and easy tounderstand, but the model’s own shortcomings as well as the particularity of the data of the region led to the model’s low accuracy; XGM (1,1) model improved thedefects of the gray GM (1,1) model. It combined the metabolism with the originalgray GM (1,1) model,which rusulted in the fact that the combination is moreaccurate than the original gray GM (1,1) model. The accuracy has been greatlyimproved. PXGM (1,1) model is the combination of the PSO algorithm andmetabolism thought to optimize parameters of the model using the PSO algorithm.Practice showed that the accuracy of the model, compared with that of XGM (1,1)model, has been greatly improved. XGM (1,1) model and PXGM of (1,1) models arebuilt on the basis of the gray GM (1,1) model, and both are made out of improvedgray GM (1,1) model. The model is applied to the prediction of ET0in westernLiaoning Province, which has shown a good adaptability and provided a new methodto prediction of ET0.