The regional distribution, temporal evolvement and uprightness variation of soil humidity in plow layer during farming period in eastern Hexi Corridor is analyzed with statistical method. It is shown that there are three zones of soil humidity, and there is a remarkable seasonal change for the spatio-temporal distribution of soil humidity. The diminution of soil humidity is clearly related to the change of meteorological conditions such as air temperature, precipitation, sunlight and ground temperature. A prediction model of soil humidity in spring is established by using the method of optimization subclass neural network. This model may provide decision-making gist to combat drought calamities and improve the utilization of water resource. |