Analysis Of Farmland Microclimate Characteristics And Soil Moisture Dynamic Simulation In Low Hilly Red Soil Region

Studying on soil temperature and humidity variation, soil water balance characteristics and soil moisture prediction models of typical red soil experimental field has scientific significance to deepen the understanding of characteristics of soil moisture transfer and relieving seasonal drought in red soil region. This paper choses peanut fields and watermelon fields of typical red soil hillslope in Yingtan Jiangxi as the research objects. The measured elements included microclimate, soil temperature, humidity and relevant soil parameters.After studying the microclimate features, soil temperature and moisture dynamic characteristics in 2014-2015, the application of soil water balance model and BP neural network model was discussed in soil moisture dynamics simulation. The results will provide a reference for guidancing local residents to improve water use efficiency and to prevent drought effectivly in red soil region. The primary conclusions are as follows:(1)Under the condition of different soil layers, slopes and weather conditions, the variation characteristics of soil temperature are different. In terms of soil temperature variation characteristics in each layer, surface soil temperature is greatly influenced by air temperature, and the variation periods change more dramatically than bottom soil layers, it shows that surface soil temperature changes more sharply. Soil temperature of upslope experimental field is generally higher than footslope experimental field at the same period. The 24h soil temperature variation curves of 10cm,20cm depth appear sinusoidal trend, and 10cm depth curve has the maximum amplitude. Under typical weather conditions, soil temperature differences are fewer in bottom layer and more in middle-upper layers, and the differences of soil temperature between day and night are fewer in rainy days and more in cloudy and sunny days. Among all meteorological factors that affects soil temperature, air temperature, relative humidity and wind speed has greater grey correlation degree with soil temperature.(2)Soil moisture is greatly influenced by rainfall and its seasonal changes are obviously, and soil moisture appears up and down oscillated fluctuation trend during March to early July in different soil depth. The soil moisture of shallow layer in experimental field is generally lower but its fluctuation and variability is greater, and more susceptible to the impact of surface water status. However, deep soil moisture performes a greater difference due to the influence of soil physical properties, different planting patterns, locking water capacity of crop roots, experimental location and so on. Soil moisture content is greatly influenced by rainy days, cumulative rainfall, air temperature and evapotranspiration of one period. The same experimental cell at different time has consistent trendency of soil moisture content. Among all meteorological factors that affects soil moisture, relative humidity, precipitation and daily maximum temperature has larger grey correlation degree with soil moisture.(3)The simulation results of lmeter deep soil water storage by using soil water balance model show that the consequence of slope experimental field is better than hillside, peanut experimental field(A zone, D zone) use logarithmic form soil water stress formula to simulate soil moisture slightly better, while watermelon experimental field(B zone, C zone) adopt the power function form might be better. Soil water storage analog values are different degrees higher than the measured value during mid-May to mid-June in four experimental fields, the causes are owing to the lower estimates of evapotranspiration during simulation process.(4)Soil water content of 10cm and 100cm depth was simulated by using BP neural network model. The results show that modeling effects in footslope is better than upslope at the depth of 10cm, and the prediction accuracy of watermelon field is better when under the same slope. At 100cm depth, soil moisture simulation values of upslope are better, under the same gradient, peanut experimental fields own higher forecasting precision. For different crops, analog values of soil water storage exist a certain differences, the soil water storage of peanut experimental field is more than watermelon experimental field at the depths of lm during the same period.