Simulation Study On The Model Of The Yield Of Greenhouse Muskmelon

The applying of appropriate crop simulation model will be important to the producting of high quality and efficiency, making best of protected cultivation and environment controlling scientificly and rationally. So does the melon production under protected cultivation.The experiments of the diffirent sowing date, training types were carried out in the plastic tunnel of Guancun in Yangling and the experiment of the training type the number of fruit left, the nitrogen levels with diffirent film covering and the diffirent node which the fruit located were carried out in greenhouse and plastic tunnel seperately in the year of 2008 and 2009. Based on the relationship between growth and environmental factors in muskmelon, leaf area index simulation model was constructed according to the solar heat product (TEP) as the only standard; the dry matter partitioning of muskmelon model was construct accroding to the rule which partitioning index change with the TEP; the harvest time module was constructed accroding to the rule which harvesting time changed with TEP, and the yield simulation model of thick-skinned melon was formed.Leaf area index model of diffirent training type of thick-skinned melon was constructed basing on the effect of TEP on the formation of leaf area, and this model was tested using the data of diffirent training type and fuits number left. It showed that R2 and the RMSE between predicted and observed data based on the 1:1 line was 0. 8581,and 0.3882,respectively.A simulation model for dry matter partitioning of muskmelon under protected cultivation was developed using the data of experiment No.3 based on the relationships partitioning index and the(TEP), and the model was tested using data of nitrogen levels with diffirent film covering. The dry matter partitioning model, yield forming model and harvest time simulation model could predict the yield and the dry weight of diffirent part of thick-skinned melon. It showed that R2 of stem dry matter weight between predicted and observed data based on the 1:1 line was 0.8593,RMSE was 8.1㎏/667㎡; R2 of leaf dry matter weight between predicted and observed data based on the 1:1 line was 0.8639, RMSE was 11.72㎏/667㎡; R2 of fruit dry matter weight between predicted and observed data based on the 1:1 line was 0.7673, and RMSE was 201.61Kg/667m2, respectively. A simulation model showed that R2 of yield of muskmelon between predicted and observed data based on the 1:1 line was 0.9419,RMSE was 203.18㎏/667㎡;R2 of fruit diameter (NO-N3) between predicted and observed data based on the 1:1 line was0.9776,RMSE was1.67cm,respectively.Prediction error of havest time was in -3～2 days in shaanxi province.