论文部分内容阅读
以改进本地作物播种期土壤含水量和耕层地温的预报服务及提升农业生产安排决策能力为目的。利用内蒙古河套灌区1980~2011年年春季和作物播种期间气温、降水、风速、地温、土壤相对湿度等资料,采用相关分析、回归分析、Mann-Kendall趋势检验等方法 ,在土壤温湿因子诊断分析基础上,建立不同层次逐日土壤温度和逐旬相对湿度预报模型。结果表明:1980年以来河套灌区大部地区春播期平均土壤相对湿度在0~10 cm、10~20 cm和20~30 cm土层均呈下降趋势,影响的主要因子依次为前一旬土壤相对湿度、当旬的平均气温和降水量;河套灌区气温及0 cm、5 cm和10 cm地温变化趋势相同,均呈现上升的趋势,影响的主要因子为平均气温和平均风速;建立了各层土壤相对湿度预测模型84个和各层地温预测模型36个,均通过信度检验(P≤0.05);土壤相对湿度模型回代和预报检验准确率分别大于85%和80%,有的甚至超过90%;地温模型回代检验平均误差为1.9~2.3℃,2011、2012年预报检验平均误差为2.1~2.5℃。模型输出结果更能反映当地作物适宜播种期间土壤温湿匹配效果,预报精度达到了一定的水平,可用于干旱地区土壤相对湿度和地温的预报。
Aiming at improving the prediction service of soil moisture and topsoil temperature at the planting time of local crops and improving the decision-making ability of agricultural production arrangements. Based on the data of temperature, precipitation, wind speed, ground temperature and soil relative humidity during the spring of 1980 ~ 2011 and sowing date in Hetao Irrigation Area of Inner Mongolia, correlation analysis, regression analysis, Mann-Kendall trend test and other methods were used in the analysis of soil temperature and moisture Based on the establishment of different levels of daily soil temperature and ten-day relative humidity forecast model. The results showed that the average soil relative humidity of spring sowing period in most parts of Hetao Irrigation District decreased from 0 to 10 cm, from 10 cm to 20 cm and from 20 cm to 30 cm since 1980, with the main factors affecting the soil relative humidity Humidity, average temperature at ten days and precipitation. The temperature of 0 cm, 5 cm and 10 cm in the Hetao Irrigation District showed the same trends of temperature, showing a rising trend, the main factors affecting the average temperature and average wind speed; 84 relative humidity prediction models and 36 ground temperature prediction models all passed the reliability test (P≤0.05). The soil relative humidity model was more than 85% and 80% respectively, and even more than 90% %; The average error of the ground temperature model back to the test was 1.9 ~ 2.3 ℃, the average error of forecast in 2011 and 2012 was 2.1 ~ 2.5 ℃. The results of model output can better reflect the effect of soil temperature and moisture matching during the sowing of the local crop, and the prediction accuracy reaches a certain level, which can be used to predict the soil relative humidity and ground temperature in arid areas.