论文部分内容阅读
基于多层误差反传网络结构模型和一维时间序列拓展的方法，发展了一种新的旱季农田土壤湿度预报模式。该模式不仅预报准确性高，而且不受中期降水预报准确性的影响。一般只要具备386以上的计算机条件即可进行工作，十分便于业务预报推广。可为我国干旱地区或季节的农业区土壤水资源的合理利用和防旱抗旱，减少光、热农业气候资源的浪费提供有效的新方法。同时也为充分利用宝贵的土壤湿度观测资料开劈了途径。
Based on multi-layer error backpropagation network structure model and one-dimensional time series expansion method, a new dry season soil moisture forecasting model was developed. This model not only has high prediction accuracy but also is not affected by the accuracy of mid-term precipitation forecasting. As long as generally have more than 386 computer conditions to work, is very easy to promote business forecasts. It can provide an effective new method for the rational utilization of soil water resources, drought-resistance and drought-resistance and the waste of light and hot agricultural climate resources in agricultural areas in arid areas or seasons. It also opens the way for full use of valuable soil moisture observations.