| 引用本文: | 李军玲,李梦夏,褚荣浩,等.夏玉米花期高温干旱复合胁迫监测预警与
穗粒数响应机制研究[J].灌溉排水学报,2026,45(6):27-39. |
| LI Junling,LI Mengxia,CHU Ronghao,et al.夏玉米花期高温干旱复合胁迫监测预警与
穗粒数响应机制研究[J].灌溉排水学报,2026,45(6):27-39. |
|
| 摘要: |
| 【目的】精准评估夏玉米花期高温干旱复合胁迫风险并构建可靠的监测预警技术方法。【方法】以河南省为研究区，整合农业气象观测站实测数据与CLDAS格点数据等多源数据，构建耦合温度胁迫指数（TSI）和土壤湿度修正系数的高温干旱综合胁迫指数（HTI），基于2025年中国气象局智能网格预报产品，开展预警方法前瞻应用，并利用5地市28个县的夏玉米实地调查数据开展灾后实况后验分析，系统探讨了花期高温干旱复合胁迫监测指标的构建方法及预警技术的适用性。【结果】在CLDAS_2mT和CLDAS_20cmRSM数据订正中支持向量机模型表现最优，订正后观测值与格点值相关系数（R）达0.999 8~1.0，均方根误差（RMSE）较原始数据降低60%以上；HTI对高温干旱复合胁迫的综合表征能力优于TSI，其与穗粒数极显著负相关（r=-0.520 2，P＜0.001），对穗粒数的模拟RMSE降低6.04%，R绝对值提升5.5%。2017—2019、2022、2024年为复合胁迫较重年份，2025年复合胁迫较往年加重，高风险区集中于豫东、豫南及豫中部分地区，多数区域高温与干旱呈显著协同放大效应，且高风险区穗粒数显著低于低风险区。【结论】基于支持向量机模型订正的1 km分辨率CLDAS数据可满足区域尺度高温干旱复合胁迫评估需求；HTI能有效量化复合胁迫强度，穗粒数是复合胁迫影响产量的关键限制因子；2025年豫东、豫中及豫南高风险区需重点防控，基于TSI、土壤湿度修正系数和HTI构建的预警方法可为夏玉米抗逆稳产提供有效技术支撑。 |
| 关键词: 夏玉米；高温干旱复合胁迫；穗粒数；多源数据融合；花期；风险预警 |
| DOI：10.13522/j.cnki.ggps.2025361 |
| 分类号: |
| 基金项目: |
|
| Monitoring and early warning of heat-drought stress at the flowering stage of summer maize and its impact on kernel number per ear |
|
LI Junling, LI Mengxia, CHU Ronghao, XUE Changying,
MA Zhihong, ZHANG Xihe, LI Shuyan
|
|
1. CMA·Henan Agrometeorological Support and Applied Technique Key Laboratory, Zhengzhou 450003, China;
2. Henan Institute of Meteorological Sciences, Zhengzhou 450003, China
|
| Abstract: |
| 【Objective】Flowering stage is a critical stage for summer maize, during which concurrent heat and drought stress could severely undermine its yield formation. We developed a reliable method for monitoring and assessing the risk of combined heat and drought stress for summer maize during this growth stage.【Method】The study area is Henan Province. Multi-source data, including ground-based agrometeorological observations and CLDAS (China Land Data Assimilation System) grid data, were integrated to construct a comprehensive Heat-Drought Stress Index (HTI) by coupling the Temperature Stress Index (TSI) with soil moisture correction coefficients. Based on the 2025 intelligent grid forecast products of the China Meteorological Administration, an early warning method for heat–drought stress was developed. Field survey data from 28 counties in five regions across the province were used for post-disaster hindcast analysis, and the feasibility of the monitoring indicators for heat-drought stress at the flowering stage, as well as their applicability for early warning, was systematically evaluated.【Result】The Support Vector Machine (SVM) model is optimal for correcting the CLDAS_2mT and CLDAS_20cmRSM data, with the correlation coefficient (R) between observed and grid values reaching 0.999 8-1.0, and the Root Mean Square Error (RMSE) reduced by more than 60% compared with the raw data. The HTI outperformed the TSI in characterizing heat-drought stress, showing a highly significant negative correlation with kernel number per ear (r = -0.520 2, P＜0.001). Compared with the TSI, the HTI reduced the RMSE of the modelled kernel number per ear by 6.04% and increased the absolute value of R by 5.5%. The model identified 2017—2019, 2022 and 2024 as years with severe heat-drought stress, with the stress intensity in 2025 exceeding previous levels. Risky zones with high stress were mainly in the eastern, southern and central parts of the province. Most regions exhibited a significant synergistic effect of heat and drought, and the kernel numbers per ear in high-risk zones were significantly lower than those in low-risk zones.【Conclusion】The 1-km resolution CLDAS data corrected by the SVM model meet the requirements for regional-scale assessment of heat-drought stress. The HTI can effectively quantify stress intensity and demonstrates that kernel numbers per ear is the key yield-limiting factor. In 2025, high-risky areas were mainly in the eastern, centra and southern parts of the province. The early warning method developed using TSI, soil moisture correction coefficients, and HTI provides a robust support for enhancing stress resilience and ensuring a stable summer maize yield. |
| Key words: summer maize; combined heat and drought stress; kernel number per ear; multi-source data fusion; flowering stage; risk early warning |
