摘要
根据烟台某一气象站提供的2003年与2005年气象资料及土壤温度数据,对此地区土壤温度的变化特征及其预报模型进行了研究。结果表明:各个土层土壤温度变化规律基本一致,1月份土壤温度最低,到7月份达到最高值,同时也表现出土壤温度年变化振幅随深度增加而减小的规律。基于气象因子的多元回归模型对深层土壤温度预测精度不够理想,而BP人工神经网络模型(ANN-BP)相比多元回归模型则能显著提高预测精度,是一种良好的土壤温度预测模式。
According to the weather data and data of soil temperature which were supplied by a weather station of Yantai,change characteristics of soil temperature and its forecasting model were studied of this area.The results showed that the variation laws of soil temperature at the different soil layers are consistent,and soil temperature fell in the minimum in January,the maximum of soil temperature appear in July.It also showed that the amplitude of soil temperature decreased with the increase of soil depth.The precision of multiple linear regression models based meteorological factors to estimate soil temperature was low.The model of ANN-BP being a good method to be applied to estimate soil temperature,can greatly improve estimation accuracy.
出处
《农业系统科学与综合研究》
CSCD
2010年第4期487-492,共6页
System Sciemces and Comprehensive Studies In Agriculture
基金
山东教育科研计划资助项目(J07YF160)
关键词
气象因子
土壤温度
变化特征
预测模型
meteorological factors
soil temperature
change characteristics
forecasting model
