基于SARIMA模型的五道沟地区0~320 cm土层
季尺度地温预测研究

蒋鑫平; 王启猛; 刘 猛; 王发信; 吕海深; 陈 雨; 李 杰; 王振龙

引用本文:	蒋鑫平,王启猛,刘猛,等.基于SARIMA模型的五道沟地区0~320 cm土层季尺度地温预测研究[J].灌溉排水学报,2024,43(2):54-60.
	JIANG Xinping,WANG Qimeng,LIU Meng,et al.基于SARIMA模型的五道沟地区0~320 cm土层季尺度地温预测研究[J].灌溉排水学报,2024,43(2):54-60.

【打印本页】【下载PDF全文】【查看/发表评论】【EndNote】【RefMan】【BibTex】

←前一篇|后一篇→

过刊浏览高级检索

本文已被：浏览 99次下载 99次	码上扫一扫！
分享到：微信更多字体:加大+\|默认\|缩小-
基于SARIMA模型的五道沟地区0~320 cm土层季尺度地温预测研究
蒋鑫平，王启猛，刘猛，王发信，吕海深，陈雨，李杰，王振龙
1.河海大学，南京 210098；2.水利部淮河水利委员会，安徽蚌埠 233000； 3.安徽省（水利部淮委）水利科学研究院，合肥 230088

摘要:

【目的】探讨五道沟地区地温季尺度变化趋势和突变特征，建立SARIMA地温预测模型。【方法】基于五道沟水文实验站1964—2022年长系列实测地温资料，采用线性回归、Sen's斜率估计、MK检验等方法，开展0~320 cm土层地温季尺度变化趋势和突变特征研究，建立不同土层深度（0~320 cm）地温SARIMA预测模型。【结果】①春季、冬季0~160 cm土层地温呈显著上升趋势；夏季除0、10 cm土层外其他土层地温均有显著下降趋势；秋季0、20 cm土层地温具有显著上升趋势；320 cm土层地温在冬季具有显著下降趋势。②春季0、10、20、40、160 cm土层地温分别在2006、2013、2012、2015、2018年发生突变，突变后增温趋势显著；320 cm土层地温在1984年前后开始显著降低。③地温数据的预测值与实测值拟合优度均?0.95，不同土层地温预测模型均有较好的预测能力，且随土层深度增加预测精度提高，MAE随土层深度增加由1.666下降至0.390，RMSE随土层深度增加由2.139下降至0.525。【结论】SARIMA模型精度较高，可用于淮北平原地区地温模拟预测。

关键词: 地温；变化特征；时间序列；SARIMA模型

DOI：10.13522/j.cnki.ggps.2023223

分类号:

基金项目:

Spatiotemporal temperature variation in soil in Wudaogou area and its modelling using the SARIMA model

JIANG Xinping, WANG Qimeng, LIU Meng, WANG Faxin, LYU Haishen, CHEN Yu, LI Jie, WANG Zhenlong

1. Hohai University, Nanjing 210098, China; 2. Huaihe River Water Conservancy Commission, Ministry of Water Resources, Bengbu 233000, China; 3. Anhui Province (Huaihe Commission, Ministry of Water Resources) Institute of Water Resources Science, Hefei 230088, China

Abstract:

【Objective】Soil temperature is not only important for hydrological processes but also plays an imperative role in crop growth and soil biochemical reactions. Understanding its spatiotemporal variation is crucial to improving soil and hydrological management. The purpose of this paper is to investigate the applicability of the SARIMA model to model spatiotemporal change in temperature across the entire soil profile.【Method】The study is based on temperatures measured from 1964 to 2022 across a 0-320 cm profile located at the Wudaogou Hydrological Experimental Station, in Anhui province, China. Linear regression, Sen's slope estimation, MK test and other methods are used to analyze the seasonal change in temperature in different soil layers, and to establish the SARIMA model.【Result】① In spring and winter, the temperature in 0-160 cm soil layer had been in increase from 1964 to 2022 at significant levels. Except in the 0-10 cm soil, summer temperature in other soil layers had been in decrease from 1964 to 2022 at significant levels. In the fall, the temperature had been increasing in the 0 and 20 cm soil layer, but decreasing in other soil layers. ② The temperature in depths of 0, 10, 20, 40, and 160 cm had endured sudden drops in spring in 2006, 2013, 2012, 2015 and 2018, followed by significant increases. Since 1984, temperature in the 320 cm soil layer had begun to decrease significantly. ③ The correlation between measured and predicted temperature was >0.95. With the increase in soil depth, the correlation increases, MAE decreases from 1.666 to 0.390, and the RMSE decreases from 2.139 to 0.525.【Conclusion】The SARMA model is accurate to model spatiotemporal change in soil temperature across the entire 0-320 cm soil profile in Huaibei Plain area.

Key words: soil temperature; characteristics of change; time series; SARIMA model