Study On The Mothod Of Monitoring And Forecasting Soil Moisture Content (Drought)

Because shortage of water resource is becoming a global problom, governments of many countries are paying much more attention to using water resource economically and efficiently. It is of great theoretical and practical significance to realize accurate measuring, calculating and forecasting of soil moisture content (drought) by adopting advanced sensor technique and space-time analysis method.This paper summarizes the current study status of detecting and forecasting soil moisture content (drought) at home and abroad. It is indicated that the study content of soil moisture content (drought) consists of detecting, calculating and forecasting of drought. On the basis of analyzing experiment data, the main problom for detecting, calculating and forecasting of soil moisture content (drought) is studied in this paper.The main contents of study and conclusions are summarized as follows:1. The experiments on space distribution characteristics and transformation tendencies of soil moisture were studied during the past time. Based on the statistical analysis of soil moisture data and parameter tests, the conclusion is drawn that space distribution of soil moisture conforms to logarithmic normal distribution or approximately logarithmic normal distribution. Space transformate tendency of distribution of soil moisture in region is related to the geographical position.2. Based on the data of correlative analysis and tests of soil moisture in different depths (within 20cm under the earth), it is proved that there is a linear correlation between layers.3. The burying position of soil water sensors in drought monitoring is studied. The analysis data and tests indicate that:Moisture rates in soil body at the dots of 20, 30 and 40cm under the earth are linearly correlative; it is also linearly correlative at the dot of 50 and 60cm under the earth; moisture rate in soil body beneath 60cm under the erath is stable during spring season; it is linearly correlative in the range of 1m under the earth during flood season.The suitable moisture sensor's laying place is at the dots of 10 30 60 and 100cm under the earth.4. Through analysing the experiment data by numeric taxology theory, it is proved that space cluster analysis method of information on soil moisture content is suitable in the position choosing of soil moisture watching stations.5. The number of soil moisture watching stations in a region range is discussed theoretically in this paper. Davies-Bouldin index, an index that is often adopted in small amount classification, is proved to be suitable for the determination of the number of watching stations.6. Geography statistic analysis theory is used for calculating and forecasting of soil moisture content in a region range firstly. Space distribution, transformation tendency, anisotropic direction & magnitude and auto correlation distance are displayed by GIS software and it has been proved that space variance and anisotropy exist in soil moisture content. Simple Crick method is fit for soil moistureforecasting in region and the forecasting error in half-variation triangle function model is less.7. Drought measurement experiment in the field indicated that soil drought has no obvious space tendency, little anisotropy and little space auto correlation distance. IDW is suitable for drought distribution prediction in the field. Voronoi map can provide visualized tool for the drought data distribution and check in the field. Distance between dots should be uniform in the field measurement. Sampling number should be estimated by certain believe level, estimate accuracy and variance coefficient.8. A short-term (a week, a day) model for forecasting soil moisture content is established based on BP ANN. The result is proved by practical calculating data and examinations that the model has many advantages, such as high accuracy, little error and easily to be popularized.