结合门控循环单元算法与探空仪数据的人工增雨效果模拟分析OA
Simulation Analysis of Artificial Precipitation Enhancement Effects By Combining Gated Recurrent Unit Algorithm with Radiosonde Data
人工增雨效果受云层微观物理过程复杂性和多尺度气象要素耦合作用影响,其动态响应机制难以量化表征,导致模拟增雨量与实际效果存在显著偏差.为此,提出结合门控循环单元算法与探空仪数据的人工增雨效果模拟分析方法.通过无人机下投探空仪获取高分辨率的气象数据,在集合经验模态分解(EEMD)方法作用下,获取不同时间尺度的本征模函数(IMF)分量和残余分量,表征不同时间尺度的气象特征.构建双向门控循环单元(BiGRU)模型,利用其强大的时序建模能力,输入不同时间尺度的气象特征,对人工增雨过程中的复杂气象条件进行动态模拟和预测.实验结果表明,该方法能够有效捕捉催化作业后的降雨量变化,模拟的增雨量曲线与实际人工增雨量曲线更加接近.同时在不同催化剂、投放剂量和高度下,人工增雨效果模拟分析的相关系数均在0.85以上.该方法模拟结果能够为人工增雨作业提供科学的决策依据,并为干旱区生态系统恢复和水资源可持续管理提供技术支撑.
The effectiveness of artificial precipitation enhancement is influenced by the complexity of microphysical processes of cloud and the coupled effects of multi-scale meteorological factors,making it difficult to quantitatively characterize its dynamic response mechanisms.This leads to significant discrepancies between simulated rain enhancement amounts and actual outcomes.To address this,we proposeda simulation analysis method for artificial rain enhancement that combines the gate recurrent unit algorithmwith radiosonde data.By deploying radiosondes via unmanned aerial vehicles to collect high-resolution meteorological data,the ensemble empirical mode decomposition(EEMD)method was applied to extract intrinsic mode function(IMF)components and residual components at different temporal scales,thereby characterizing meteorological features at various temporal scales.A bidirectional gate recurrent unit(BiGRU)model was constructed,leveraging its robust temporal modeling capabilities to input meteorological features at different time scales,enabling dynamic simulation and prediction of complex meteorological conditions during artificial precipitation enhancement processes.Experimental results showed that this method can effectively capture changes in precipitation after catalytic operations,with simulated precipitation curves more closely matching actual artificial precipitation curves.Additionally,under different catalyst types,dosage levels,and release altitudes,the correlation coefficients for simulated artificial precipitation enhancement effects all exceeded0.85.The simulation results of this method can provide scientific decision-making support for artificial precipitation enhancement and technical support for ecosystem restoration and sustainable water resource management in arid regions.
尤宏宇;鞠洋;边立峰;陈曦;左宇明;常智莹
大兴安岭地区气象局,黑龙江加格达奇 165000黑龙江省人民政府降雨办公室,黑龙江哈尔滨 150030大兴安岭地区气象局,黑龙江加格达奇 165000大兴安岭地区气象局,黑龙江加格达奇 165000大兴安岭地区气象局,黑龙江加格达奇 165000塔河县气象局,黑龙江塔河 165200
资源环境
门控循环单元探空仪数据人工增雨模拟分析多尺度分解降雨量
Gate recurrent unitradiosonde dataartificial precipitation enhancementsimulation analysismulti-scale decompositionprecipitation
《四川环境》 2026 (4)
161-167,7
黑龙江省气象局科技创新发展项目"基于MapReduce+Spark实现区域站数据状态可视化的研究"(HQFZ2022011).
评论