Apple Tree Identification And Evapotranspiration Process Simulation On The Loess Plateau

The implementation of the project has made great progress in the ecological construction of the Loess Plateau.The Chinese government has made great efforts to expand the planting area of economic forests of crops such as jujube and apple on the Loess Plateau and achieved remarkable benefits.However,due to the drought and water shortage in the Loess Plateau,the rapid development of apple industry has changed the previous water balance.Under dry conditions,the water consumption of orchard planting is usually greater than the water income,resulting in serious water shortage.Therefore,it is very important and necessary to evaluate the water suitability of Apple plantation on the Loess Plateau in order to ensure the sustainable development of Apple plantation and the stable growth of farmers’ income.Based on this,in this study,firstly,the sentinel 2A remote sensing image is used to accurately identify the apple trees in the suitable growing area of apple,and the spatial distribution map of apple trees is obtained.Then etmod model is introduced to simulate the actual evapotranspiration of apple trees in typical years.Finally,the water requirement and water use efficiency of apple trees were calculated to analyze the water adaptability.It can provide basis and guarantee for the sustainable development of Apple economic forest.The main research results are as follows:(1)Using sentinel 2A remote sensing image data,Download multi period and high-resolution remote sensing image,use multi-phase analysis method and support vector machine classification method to realize the fine recognition and extraction of fruit trees in the apple growing area,and verify the accuracy of classification results.The results show that the classification effect of SVM is better than that of MLE,and the classification is complete.The classification accuracy is high,and the phenomenon of misclassification and misclassification is relatively less.Based on this,we choose support vector machine classification to extract the information of remote sensing image.According to the spatial distribution map of fruit trees after classification,50% of the verification points of fruit trees were used to evaluate the spatial position accuracy.The evaluation results of confusion matrix are as follows: the overall accuracy is 82.25%,the user accuracy is 79.74%,and the mapping accuracy is 81.33%.The accuracy of recognition effect needs to be improved.(2)Based on the principle of two source evapotranspiration model,the ETMOD model for regional evapotranspiration simulation was developed.In the development of the model,the principle of crop evaporation is fully considered,and the runoff production infiltration module of the underlying surface is added to improve the accuracy of the model in calculating the actual evapotranspiration of fruit trees.The ETMOD model is used to estimate the actual evapotranspiration in a typical hydrological year,and the annual distribution of Lai and actual evapotranspiration is analyzed quantitatively.The simulation results show that E,T and ET are concentrated in April to September,accounting for 75%,90% and 85% of the annual total,respectively.E,T,ET and LAI are all increased first and then decreased,reaching the maximum value in July,and the change trend of LAI is the most similar to that of vegetation transpiration T.The evapotranspiration and actual evapotranspiration of vegetation in space are increasing from southwest to northeast.The evapotranspiration of soil,vegetation and total evapotranspiration in normal year are higher than that in dry year.(3)In order to evaluate the water demand of apple trees in typical level years,the water demand of key growth period of fruit trees is calculated based on the fourth chapter,and the water demand of fruit trees in typical level years is calculated by the ratio of water consumption and water demand.The results show that the actual water consumption of fruit trees cannot meet the water demand of apple trees in both dry and normal years,and cannot meet the growth of apple trees under the optimal conditions.From May to August in the low water year,the water requirement from flowering to young fruit development is better between 0.5 and 0.7.From April to June in the year of normal water,the water requirement from germination to new shoot growth is better than 0.5-0.6.In July,the water requirement of fruit trees is better,and in August,the water requirement of fruit expansion is better.Generally speaking,the water demand in normal year is better than that in dry year.(4)Through the simulation calculation of GPP and WUE,it can be seen that the trend of GPP and WUE is basically the same,both of them have a trend of rising first and then declining,reaching the maximum in August.In the two typical levels,the GPP was concentrated in May to September,accounting for more than 83% of the annual GPP,and less than 17 in the other eight months.Compared with the annual distribution of GPP,the annual distribution of WUE is mainly from April to October,among which only 5 months have WUE greater than 0.6.In the other five months,the water use efficiency was lower than 0.5.In August,the water use efficiency reached the maximum value,all of which were higher than 0.9.Generally speaking,the water adaptability of May to August can basically meet the growth of fruit trees,but the water adaptability of other months is poor,and the water use efficiency and water demand need to be improved.