| 张静文1, 张竞成1*, 张雪雪1, 黄玉娟1, 郭安红2, 吴开华1.耦合气象影响因素和Logistic方程的水稻纹枯病发病等级动态预测模型研究[J].植物保护,2022,48(3):172-180. |
| 耦合气象影响因素和Logistic方程的水稻纹枯病发病等级动态预测模型研究 |
| A dynamic forecasting model for the severity of rice sheath blight by coupling meteorological factors with Logistic equation |
| 投稿时间：2021-04-19 修订日期：2021-05-24 |
| DOI：10.16688/j.zwbh.2021238 |
| 中文关键词: 水稻纹枯病 Logistic方程 温度 湿度 预测模型 |
| 英文关键词:rice sheath blight Logistic equation temperature humidity forecasting model |
| 基金项目:浙江省科技计划（LGN20D010003）；国家自然科学基金（42071420）；国家重点研发计划国际合作项目（2019YFE0125300） |
|
| 摘要点击次数: 260 |
| 全文下载次数: 217 |
| 中文摘要: |
| 本文基于植保、气象等数据研究水稻纹枯病发病等级-时间动态的预测方法和模型。利用2010年－2016年湖南省12个县的植保调查数据和气象观测值,以水稻纹枯病流行机理为基础将Logistic方程与构建的温度影响模块和湿度影响模块耦合,建立Logistic-RICEBLA病害预测模型。通过对模型参数进行调优、训练和验证,实现对水稻纹枯病发病等级的动态预测。结果表明,Logistic-RICEBLA模型能够较好地响应温度、湿度等气象条件的变化,模型预测结果与实际的水稻纹枯病发病等级-时间变化曲线具有较高的一致性。经验证,模型预测结果在单时相上精度达到R2=0.68,RMSE=1,容错准确率P_bias=88%,表明预测值与实际发病等级的误差基本控制在±1级范围内。在多时相整体趋势的验证方面,模型预测的病害流行曲线下面积(AUDPC)与病害实际发展的AUDPC保持高度一致性,决定系数(R2)达到0.86,表明模型能给出纹枯病在水稻不同生育期发病等级变化的整体趋势。本研究构建的Logistic-RICEBLA模型能由简单的气象数据和植保数据驱动,对水稻纹枯病发病等级进行动态预测,有助于在植保管理中及时掌握区域中病害发生发展的趋势,为水稻病害统防统治等防控工作提供参考。 |
| 英文摘要: |
| Based on disease survey and meteorological data of 12 counties in Hunan province over 2010－2016, this study developed a temporal dynamic disease forecasting model for rice sheath blight. According to the epidemic mechanism of rice sheath blight disease, a temperature impact module and a humidity impact module were constructed, which were coupled with the Logistic equation to form the Logistic-RICEBLA disease forecasting model. The optimization, training and verification of model parameters were conducted to achieve the dynamic forecast of the disease in the study area. The results suggested that the Logistic-RICEBLA model could respond to variations of temperature and humidity. The forecasting results were highly consistent with the actual disease development curve of rice sheath blight. The model accuracies at single phase were as followed: R2=0.68, RMSE=1 and P_bias=88% (tolerated accuracy), indicating that the forecasting error fell within ±1 severity grade. With regard to the model performance at multiple phases, the area under the disease prevalence curve (AUDPC) was used to assess the capability of the model in forecasting the general development trend of the disease. The predicted AUDPC was highly consistent with the actual AUDPC, and the coefficient of determination (R2) reached 0.86, suggesting that the model is able to predict the overall trend of disease development over multiple phases. The developed Logistic-RICEBLA model could be driven by simple meteorological data and disease survey data. The disease forecasting information will facilitate the understanding of the disease development trend in the region, which is important in guiding the prevention and control practices of rice diseases. |
| 查看全文 查看/发表评论 下载PDF阅读器 |
| 关闭 |
|
|
|
