摘要
为了定量预报农田未来7 d土壤墒情变化特征,提升农业气象服务工作质量与效率,该文旨在研发基于网络化的辽宁省农田土壤墒情监测与预报系统。该系统基于VC++及Fortran程序语言设计,以改进的CERES-MAIZE模型中的土壤水分子模型为基础,通过程序设计及各接口功能的实现,自动调用辽宁省自动土壤水分观测站的监测当日土壤含水量数据和中央气象台预报指导产品,实现数据的网络化获取和业务模型的实时运行,提升土壤墒情模拟的准确性、时效性和便利性。结果表明,预报准确率随预报日期增加而呈现降低的趋势,越临近实际监测日期土壤墒情预报情况与实际情况拟合越好;等级干旱预报准确率最低值为70.1%,最高值为81.9%,系统对于辽宁省农田土壤干旱级别的预报具有较高准确率。
A system was designed for monitoring current soil moisture and predicting soil moisture changes in the next 1-7 days in Liaoning province. In this system, data of soil volumetric water content from 54 automatic observation stations and weather forecast data from Central Meteorological Observatory in Liaoning were used. The soil moisture data of 8 soil layers was included, and they were different among different stations. The parameters of stations like longitude, latitude, altitude, etc were obtained from the observation stations. The soil parameters like field capacity, soil bulk density, wilting humidity were obtained by measuring soil samples when the stations were established. The next 7 days weather forecast data from the National Meteorological Center of CMA included daily maximum temperature, minimum temperature, precipitation, and cloudiness. A regression model of the sunshine and cloudiness was established based on the information of 54 stations during 1981 to 2010. All the data above were input into the system for monitoring and forecasting the soil moisture. The system was developed by VC＋＋ and Fortran based on soil water dynamic and CERES-MAIZE model. The model simulated the runoff process, infiltration process, evapotranspiration process and root water uptake process to obtain the soil moisture data of 8 layers. The system was used to predict the next 1-7 days soil moisture from June to August 2015 in order to verify the accuracy of the forecast. Results showed that when the absolute error between forecast and measured relative soil moisture was less than 3%, the highest accuracy was 52.9% for the next 1-d forecast. When the difference was less than 5%, the highest accuracy was 70.3%. When the absolute error was less than 10%, the accuracy was 78.5% for the 7 d forecast. The highest accuracy was 91.1% for the next 1-d forecast. The prediction accuracy decreased with increasing forecast date. In addition, drought level was predicted using the data of June-August, 2015. According to the threshold of relative soil moisture, 5 type of drought grade included severe drought, moderate drought, light drought, suitable, saturation with relative soil moisture less than 40%, 40% to 50%, 50% to 60%, 60% to 90%, and larger than and equal to 90%. The lowest value of drought level accuracy was 70.1%, and the highest value of drought level accuracy was 81.9%. Liaoning province suffered varying degrees of drought in the summer of 2015, which had impacted growth of maize and other field crops. The system played an important supporting role in agriculture meteorological service, as proved by its application in July 24. The study provides an effective system for improving the quality and efficiency in agriculture meteorological service.
出处
《农业工程学报》
EI
CAS
CSCD
北大核心
2016年第18期140-146,共7页
Transactions of the Chinese Society of Agricultural Engineering
基金
国家自然科学基金(41330531)
辽宁省气象科学研究所农业气象预报技术创新团队
2015年中央财政"三农"服务专项重点建设任务
关键词
土壤
干旱
模型
网络化
墒情
预报系统
soils
drought
models
web-based
moisture
forecast system
