Research On The Influence Of Meteorological Conditions On The Growth And Harvesting Period Of Spring Tea In Zhejiang Provinc

The tea is an important cash crop in China.Spring tea is the main raw material for making famous and high-quality tea.Accurate and efficient picking date and yield prediction are very important for high-quality tea production.The paper studied the effects of different meteorological conditions on tea picking date,tea bud growth state and tea quality after tea sprouting and during tea picking based on field experiments and data over the years.At the same time,the meteorological data,soil data and field management data of many years in the tea planting area are collected,Combined with the measured data in the experimental area,the parameters of the Aqua Crop model are calibrated,the applicability of the model in the picking date and yield of different tea varieties in Zhejiang province is evaluated,and the prediction models of the picking date of‘White Leaf 1’,‘Longjing 43’and‘Longjingqunti’are established.Finally,based on the Aqua Crop model,using object-oriented programming technology,Development the software of tea picking date and yield prediction.The main conclusions are as follows:（1）The analysis of the effects of different meteorological factors on the growth state and yield of tea buds during tea picking shows that the growth of tea buds and the yield and quality of tea have been affected by the temperature since the germination of tea trees,and the temperature has the greatest impact on it a few days before picking.The yield of each variety of tea first increased and then decreased with the delay of picking time.The content of tea quality components of different varieties was affected differently by picking time.For‘Longjing 43’and‘Zhongcha 108’,the earlier they enter the picking date,the better their quality.For‘Jiukeng,’‘White Leaf 1’and‘Zhonghuang 1’,it will take some time for their quality to reach the best.Among the biochemical quality parameters of tea,amino acids and caffeine are the most affected by meteorological factors.In addition to the temperature,precipitation,wind speed,humidity and other factors also have a certain impact on the growth and quality of tea.Different varieties of tea trees are affected by different meteorological factors in different periods.Under the condition of ensuring the normal growth of tea trees,the lower the temperature a few days before picking,the higher the quality of tea.The analysis of tea picking date and quality and meteorological conditions during tea growth shows that the growth degree day of tea picking date is affected by meteorological factors and changes with the base temperature.The temperature in February,especially in late February,is the key factor affecting the growth degree day of’Longjing 43’picking date;The picking date of’Longjingqunti’is not only affected by the temperature in February,but also affected by the precipitation in January.（2）Based on the planting specifications and growth parameters of different tea varieties,the parameter sensitivity analysis,calibration and verification of Aqua Crop model were carried out with using different simulation cycles,and a set of tea trees simulation parameters of Songyang tea and Anji white tea were obtained.When simulating the tea yield of Songyang County with one year simulation cycles and Anji County with spring simulation cycles,the relative errors were 1.98%and 0.99%,CRM were-0.0129 and-0.0049,RMSE were0.0325t·hm^-2 and 0.0018 t·hm^-2,NRMSE values were 2.20%and 1.10%,d were 0.84 and 0.88,R² were 0.76 and 0.73,respectively.The overall simulation results were excellent.It proves the applicability of Aqua Crop model in the yield of different tea varieties in Zhejiang Province.（3）Based on Aqua Crop model,the growth degree day,stepwise regression and machine learning prediction model of three tea varieties‘White Leaf 1’,‘Longjing 43’and‘Longjingqunti’during picking date were constructed by using the methods of growth degree day,stepwise regression and machine learning.Using the growth degree day prediction model to predict the tea picking date of three tea trees varieties,the back substitution test MAE are1.1 d,2.1 d and 1.1 d respectively;The MAE of stepwise regression model is 0.7 d,0.7 d and0.9 d respectively,which is higher than the prediction result of growth degree day model.The MAE of‘Longjing 43’tea picking date is 1.3 d by machine learning method,but the stability of prediction result is poor due to less data used for program training.（4）Referring to the principle of Aqua Crop model and its open source code,the tea picking date and yield prediction software is developed to separate the calculation function of the model for the picking date,simplify the input data and operation process,and realize many functions such as data management,picking date prediction,model parameter calibration,yield prediction and cultivation strategy evaluation for different users,It plays a rapid and efficient auxiliary decision-making role in tea production.