Citrus Drought Early Warning Model Studies, And Data Sampling Algorithm

At present, in Chongqing,citrus yield rows in the first of all the kinds of fruit. To safeguard the interest of citrus farmers, the key point is to improve the yield and quality of citrus. Citrus growth and development is based on the water, so citrus orchard water management has a direct impact on the yield and quality of citrus. Because of global become warming and Chongqing is located in the subtropical, it is more prone to drought. Therefore, soil moisture and leaf water are used to predict the future water change of citrus. When data deviates from a range of indicators, it is necessary to take timely measures. This paper focuses on the construction of citrus drought early warning model, proposes citrus drought early warning decision support system for the decision making of irrigation of citrus orchards to reduce the negative impact of drought on citrus yield and quality.The study found that real-time status of citrus growing environment of soil moisture and leaf water directly impact on the yield and quality of citrus. Thus, the paper uses quantitative and qualitative analysis and other methods to construct the citrus drought warning model of soil moisture and leaf water. For the collected 28 factors that affected the soil moisture and leaf water, using extracting common factors and regression analysis, prediction equation based on public factors is constructed. It is used to calculate the prediction value of soil moisture and leaf water content. And then verified and modified by the really measured value and MAPE criteria of assessing and forecasting accuracy. Upon examination of the value of soil moisture content, the MAPE value y1 is about 12.1, so precision type belongs to accuracy. The MAPE value of citrus leaf water content is about 1.9, so precision type belongs to excellent. Namely two warning models are high accuracy, and can be used to predict citrus garden soil moisture and leaf water content situation of Chongqing Beibei, and the expert knowledge and the citrus drought warning criterion could integrated into the citrus drought warning system to realize citrus drought early warning information automatic generation and release.Due to the changeful climate and environment, the weighted coefficients value of water content prediction equation needed to be adjust constantly in order to accurately predict the future water situation. So, the predict data of Soil moisture and leaf water content which calculated by prediction equation should be saved in order to composite analysis with new data Collected latter and adjust the weighted coefficients value of water content prediction equation constantly. By analyzing the characteristics of predictive value of moisture content, it is found that changes in real-time data relatively gentle, so, saving some time representative data point for future amendments to the weight factor but all data is advisable and it can reduce the workload of the actual value of water collected. In this paper, data sampling algorithm is put forward:first data filtering by algorithm with dead zone limits, then representative data extraction by swing door SDT algorithm based on comparing the slope to reduce the views of comparing the slope and save processing time. The simulation results show that data sampling technology has achieved good results, and the error is small, and it can be used to adjust the weighted coefficients value to ensure much more accurate and reliable of water content prediction. This technology plays an very important role to improve citrus production.