中国农业气象 ›› 2025, Vol. 46 ›› Issue (3): 305-314.doi: 10.3969/j.issn.1000-6362.2025.03.003

• 农业生态环境栏目 • 上一篇    下一篇

干旱区绿洲农田土壤水分自动观测数据修订

张鹏,李兴宇,丁文魁,齐月   

  1. 1.甘肃省武威市气象局/武威国家气候观象台,武威 733099;2.中国气象局兰州干旱气象研究所/甘肃省干旱气候变化与减灾重点实验室/中国气象局干旱气候变化与减灾重点实验室,兰州 730020
  • 收稿日期:2024-03-22 出版日期:2025-03-20 发布日期:2025-03-19
  • 作者简介:张鹏,E-mail: zhangpeng20207@163.com
  • 基金资助:
    国家自然基金项目(42175192);干旱气象科学研究基金项目(IAM202317;IAM202213);2022年兰州资源环境职业技术大学校级科技创新团队项目(X2022A−04);甘肃省气象局重点项目(Zd2024−D-3);中国气象局创新发展专项项目(CXFZ2024J056)

Revision of Automatic Soil Moisture Monitoring Data for Farmland of Oasis Regions in Arid Zones

ZHANG Peng, LI Xing-yu, DING Wen-kui, QI Yue   

  1. 1.Wuwei Meteorological Bureau of Gansu Province/Wuwei National Climatological Observatory, Wuwei 733099, China;2.Lanzhou Institute of Arid Meteorology, Chinese Meteorological Administration/Key Laboratory of Arid Climate Change and Reducing Disaster of Gansu Province/Key Laboratory of Arid Climate Change and Disaster Reducing of CMA, Lanzhou 730020
  • Received:2024-03-22 Online:2025-03-20 Published:2025-03-19

摘要:

基于20113−20227武威荒漠生态与农业气象试验站作物生长阶段0−50cm深度土壤重量含水率数据,利用等值线图分析干旱区农田自动观测和人工观测的土壤水分时空分布,通过均方根误差(RMSE)、绝对误差(E)和相对误差(RE)等统计指标对比两者差异及相关性,明确不同测定方法土壤水分动态变化特征,并由回归方程修订自动观测土壤水分数据。结果表明:自动观测和人工观测土壤水分的时空变化存在一定的差异性,但总体变化趋势大致相同,自动观测土壤重量含水率较人工观测偏低,其平均绝对误差为2.4个百分点各土层来看,多年观测的平均土壤重量含水率在0−10cm土层差值最大,10−40cm土层土壤重量含水率绝对误差小于5个百分点频率均能达到80%以上,40−50cm土层两种方式观测土壤重量含水率绝对误差小于5个百分点频率73.3%;多年观测的平均土壤重量含水率在2030cm土层差值。对自动观测的土壤水分数据进行修订后,各土层土壤重量含水率绝对误差小于5个百分点频率均达到80%以上,提高了干旱区农业气象自动化监测数据的准确性,可进一步为农田土壤水分干旱预警及监测提供一定技术支撑和理论依据

关键词: 土壤水分, 自动观测, 人工观测, 土壤重量含水率, 差异比较, 数据修订

Abstract:

Based on soil gravimetric water content data at depths of 050cm collected during crop growth stages at the Wuwei Desert Ecological and Agricultural Meteorological Experiment Station over the past 12 years (March 2011-July 2022), contour plots were employed to analyze the spatial and temporal distribution in soil moisture derived from both automatic and manual observations in farmland of arid region, the differences and correlations between these methods were quantified using statistical indices such as root mean square error(RMSE) and absolute error(E), employing statistical indices such as RMSE, absolute error(E), and relative error(RE), comparative analysis of the differences and correlations between the methodologies was conducted, and this study identified the dynamic changes in soil moisture using different measurement methods, and corrected the automatically observed soil moisture data using regression equations. The results showed these were some differences in the spatial and temporal variation of soil moisture between automatic and manual observations. However, the overall change trend in soil moisture was roughly the same. The soil gravimetric water content of automatic observation was lower than manual observation, and the mean absolute error (E) was 2.4 percent points(PP). From the perspective of each soil layer, the average soil gravimetric water content observed for many years was the largest in 010cm soil layer. The frequency of absolute difference E<5 percent points of soil gravimetric water content in 1040cm soil layer could reach more than 80%, and the frequency of absolute difference E<5 percent points of soil gravimetric water content observed in 4050cm soil layer was 73.3%. In addition, the error value in the soil gravimetric water content in 2030cm soil layer was the smallest under many years of observation. By correcting the soil moisture data of automatic observation, the frequency of absolute difference E<5 percent points of soil gravimetric water content in each soil layer reached more than 80%, which improved the accuracy of agricultural meteorological automatic monitoring data in arid areas, and further provided certain technical support and theoretical basis for farmland soil moisture drought warning and monitoring.

Key words: Soil moisture, Automatic monitoring, Manual observation, Soil gravimetric water content, Comparative analysis of difference, Data revision