作为天气系统的主要组成部分,三维云仿真在军事、航空等领域都起着重要作用.目前主流的边界体积层次结构(Bounding Volume Hierarchy,BVH)在处理形状不均匀且体积较大的云时存在渲染效率低下的问题,为此提出一种基于优化BVH算法的云产...作为天气系统的主要组成部分,三维云仿真在军事、航空等领域都起着重要作用.目前主流的边界体积层次结构(Bounding Volume Hierarchy,BVH)在处理形状不均匀且体积较大的云时存在渲染效率低下的问题,为此提出一种基于优化BVH算法的云产品渲染方法.将WRF(Weather Research and Forecasting,天气研究与预报)模型网格点中的数据作为云基元,利用Z-order Hilbert曲线对其进行空间排序,结合云基元密度优化BVH算法,提高计算效率.提出ONS(Overlapping Node Sets,重叠节点结构)降低数据存取耗时.优化BVH算法能够减少不必要的光线和三角形面之间的相交测试次数,并解决边界体无效重叠问题.仿真实验显示,SAH(Surface Area Heuristic,表面积启发式)成本较同类最优算法可提升15.6%,EPO(Effective Partial Overlap,有效重叠部分)可提升10%,构建时间减少100%以上,在任意云场景中优化BVH算法的计算效率较同类算法都有显著提高,表明其能实现WRF云产品的快速渲染.展开更多
为提升低空风切变预报精度,本文综合运用欧洲中期天气预报中心第五代再分析资料[European Centre for Medium-Range Weather Forecasts(ECMWF)fifth-generation reanalysis data,ERA5]和美国国家环境预报中心(National Centers for Envi...为提升低空风切变预报精度,本文综合运用欧洲中期天气预报中心第五代再分析资料[European Centre for Medium-Range Weather Forecasts(ECMWF)fifth-generation reanalysis data,ERA5]和美国国家环境预报中心(National Centers for Environmental Prediction,NCEP)的FNL全球再分析资料(Final Operational Global Analysis)、先进星载热发射和反射辐射仪全球数字高程模型以及兰州中川机场的实况观测资料,采用中尺度数值天气预报模式(Weather Research and Forecasting Model,WRF)、WRF结合计算流体动力学(Computational Fluid Dynamics,CFD)方法、长短期神经网络(Long Short-Term Memory,LSTM)方法,对2021年4月15-16日兰州中川机场的两次风切变过程进行模拟分析。结果表明:(1)在小于1 km的网格中使用大涡模拟,WRF模式在单个站点风速模拟任务中表现更好,但在近地面水平风场风速模拟效果上,不如WRF模式结合计算流体力学模型方案;(2)对于飞机降落过程中遭遇的两次低空风切变的模拟,WRF-LES和WRF-CFD两种模式都可以模拟出第一次低空风切变,而第二次受传入模式的WRF风速数据值较小的影响,两种模式风速差都没有达到阈值,需要在后续工作中进一步验证;(3)低风速条件(6 m·s^(-1))下,基于LSTM的单变量风速预测模型平均绝对误差基本维持在0.59 m·s^(-1),能较好地把握不同地形与环流背景条件下风速变化的非线性关系,虽然受到WRF误差和观测要素不全的限制,多变量风速预测能在保证平均绝对百分比误差小于6.60%的情况下,以更高的计算效率和泛化能力实现风速预测。本文不仅验证了WRF-CFD和WRF-LES耦合方案在风场和低空风切变预报中的差异,还探讨了基于LSTM的风速预测的可行性和准确性,期望为提高风场模拟精度,缩短精细风场模拟时间提供新的视角和方法。展开更多
全球气候变化背景下,精确模拟区域碳通量及CO_(2)浓度分布有着十分重要的现实意义.本文基于WRF-GHG(Weather Research and Forecasting Model with Greenhouse Gases Module)模式,综合考虑人为碳排放、陆地生态系统碳交换、海洋碳交换...全球气候变化背景下,精确模拟区域碳通量及CO_(2)浓度分布有着十分重要的现实意义.本文基于WRF-GHG(Weather Research and Forecasting Model with Greenhouse Gases Module)模式,综合考虑人为碳排放、陆地生态系统碳交换、海洋碳交换和生物质燃烧碳排放的影响,对2022年中国及其周边地区陆地生态系统碳通量及大气CO_(2)浓度进行在线模拟,并利用OCO-2/OCO-3卫星观测资料评估模式性能.结果表明:(1)WRF-GHG模式整体模拟效果良好(R=0.7424,BIAS=1.3860×10^(-6)),但在低纬度地区的模拟效果略差于中纬度地区,表明该模式目前在亚热带和热带的适用性有限,需要进一步优化;(2)中国区域内,人为碳排放和陆地生态系统源碳交换呈现出显著的季节性特征,其中,人为源CO_(2)排放(全年11031 Tg)在各个排放源中占据主导地位,陆地生态系统(全年-900 Tg)可以吸收约8.2%的全年人为源排放,生物质燃烧源(全年65 Tg)排放则仅为人为源排放的0.6%;(3)模拟区域内,CO_(2)浓度高值区主要分布在我国胡焕庸线以东地区、日本和南亚地区等,在各排放源对CO_(2)浓度的贡献中,人为源排放的贡献量级(1×10^(-6)~100×10^(-6))最高,因而其主导了CO_(2)浓度的空间分布特征.展开更多
随着传统化石能源面临枯竭的问题日益加剧,使用太阳能进行光伏发电成为世界各国能源结构调整的重要方向,如何进一步提高光伏发电功率的预测精度成为亟待解决的问题。为提高光伏功率短期预测的准确性和可靠性,提出一种耦合太阳辐射预报...随着传统化石能源面临枯竭的问题日益加剧,使用太阳能进行光伏发电成为世界各国能源结构调整的重要方向,如何进一步提高光伏发电功率的预测精度成为亟待解决的问题。为提高光伏功率短期预测的准确性和可靠性,提出一种耦合太阳辐射预报模式系统(weather research and forecasting model for solar energy,WRF-Solar)及辐照度订正的光伏短期预测模型,先使用WRF-Solar进行动力降尺度天气数值预报,得到包含辐照度等在内的未来气象因子,再利用随机森林对预报辐照度进行订正,在此基础上运用长短期神经网络、反向传播神经网络和逐步聚类分析建立光伏功率短期预测模型,利用某40 MW光伏电站的实际运行数据进行模型对比分析。结果表明,使用随机森林模型订正后的辐照度更接近真实值,平均绝对误差率下降了56.06个百分点;与另外2种模型预测结果对比发现,长短期神经网络模型预测效果最好,平均绝对百分比误差降低了4.13个百分点,说明组合模型能够进一步提高功率预测的精度。展开更多
本文选取GRAPES_MESO(Global/Regional Assimilation PrEdiction System-Mesoscale version)模式和WRF(Weather Research and Forecasting Model)模式在国产鲲鹏(KUNPENG)平台上开展数值模式计算特征分析,并与英特尔(X86)平台进行对比,...本文选取GRAPES_MESO(Global/Regional Assimilation PrEdiction System-Mesoscale version)模式和WRF(Weather Research and Forecasting Model)模式在国产鲲鹏(KUNPENG)平台上开展数值模式计算特征分析,并与英特尔(X86)平台进行对比,探讨数值模式在鲲鹏平台上资源使用、计算瓶颈、热点函数等方面的改进空间。结果表明:经过适配后,两个模式在国产KUNPENG平台上能得到与英特尔X86平台一致的计算结果,呈现出较好的并行扩展性;两个模式对CPU的使用率均较高,计算瓶颈主要集中在后端CPU瓶颈,对节点的整体内存使用率适当,后续优化主要集中在代码效率、算法、访存等方面。在KUNPENG平台上,可以考虑通过优化集合通信的Collective Sync、Allreduce和Wait算法,来改善GRAPES_MESO模式的MPI的通信效率;可通过优化GCR算法、以uct、ucg为代表的集合通信热点、以expf、powf等为代表的数学函数、malloc内存操作等热点函数对GRAPES_MESO模式进行优化。展开更多
[Objective]Precipitation events caused by Super Typhoon Doksuri in Fujian Province were simulated and evaluated based on the WRF model to provide a reference for typhoon precipitation simulation and forecasting in sou...[Objective]Precipitation events caused by Super Typhoon Doksuri in Fujian Province were simulated and evaluated based on the WRF model to provide a reference for typhoon precipitation simulation and forecasting in southeast coastal areas of China.[Methods]The next-generation mesoscale numerical weather prediction model WRF V4.3(The Weather Research and Forecasting Model)was used to simulate the precipitation caused by Typhoon Doksuri in Fujian Province in 2023.Observations from 86 meteorological stations with hourly rainfall records were used to evaluate the model’s performance.Six evaluation indices were used,including the correlation coefficient(R),root mean square error(RMSE),mean absolute error(MAE),equitable threat score(ETS),probability of detection(POD),and false alarm ratio(FAR).[Results](1)The temporal and spatial evolution of precipitation during Typhoon Doksuri was effectively captured by the WRF model.Precipitation intensity increased gradually from July 27 to 29,2023,with the heaviest rainfall concentrated in the northern and eastern coastal areas of Fujian Province.(2)Significant differences in model performance were observed in terms of R,RMSE,and MAE.The largest errors occurred in Putian City,while smaller errors were found in southwestern Fujian Province.The evaluation result of all six indices showed that the WRF model performed best in simulating daily precipitation compared to hourly,three-hourly,six-hourly,and twelve-hourly precipitation.(3)The R95p index indicated that the WRF model successfully captured the overall spatial distribution of extreme precipitation.However,extreme precipitation intensity was overestimated in certain coastal areas.(4)Despite accurately identifying the coastal regions of Fujian as being most affected,the WRF model failed to accurately simulate the spatial distribution and intensity of precipitation.The simulated precipitation centers showed discrepancies when compared with the observed centers.[Conclusion]Although the WRF model underestimated hourly precipitation,it successfully captured the temporal evolution and spatial distribution of rainfall caused by Typhoon Doksuri in Fujian Province.It reproduced the heavy rainfall centers in central Fujian Province,with daily precipitation peaks reaching up to 350 mm.This highlighted the severity of extreme rainfall caused by Typhoon Doksuri.展开更多
文摘作为天气系统的主要组成部分,三维云仿真在军事、航空等领域都起着重要作用.目前主流的边界体积层次结构(Bounding Volume Hierarchy,BVH)在处理形状不均匀且体积较大的云时存在渲染效率低下的问题,为此提出一种基于优化BVH算法的云产品渲染方法.将WRF(Weather Research and Forecasting,天气研究与预报)模型网格点中的数据作为云基元,利用Z-order Hilbert曲线对其进行空间排序,结合云基元密度优化BVH算法,提高计算效率.提出ONS(Overlapping Node Sets,重叠节点结构)降低数据存取耗时.优化BVH算法能够减少不必要的光线和三角形面之间的相交测试次数,并解决边界体无效重叠问题.仿真实验显示,SAH(Surface Area Heuristic,表面积启发式)成本较同类最优算法可提升15.6%,EPO(Effective Partial Overlap,有效重叠部分)可提升10%,构建时间减少100%以上,在任意云场景中优化BVH算法的计算效率较同类算法都有显著提高,表明其能实现WRF云产品的快速渲染.
文摘为提升低空风切变预报精度,本文综合运用欧洲中期天气预报中心第五代再分析资料[European Centre for Medium-Range Weather Forecasts(ECMWF)fifth-generation reanalysis data,ERA5]和美国国家环境预报中心(National Centers for Environmental Prediction,NCEP)的FNL全球再分析资料(Final Operational Global Analysis)、先进星载热发射和反射辐射仪全球数字高程模型以及兰州中川机场的实况观测资料,采用中尺度数值天气预报模式(Weather Research and Forecasting Model,WRF)、WRF结合计算流体动力学(Computational Fluid Dynamics,CFD)方法、长短期神经网络(Long Short-Term Memory,LSTM)方法,对2021年4月15-16日兰州中川机场的两次风切变过程进行模拟分析。结果表明:(1)在小于1 km的网格中使用大涡模拟,WRF模式在单个站点风速模拟任务中表现更好,但在近地面水平风场风速模拟效果上,不如WRF模式结合计算流体力学模型方案;(2)对于飞机降落过程中遭遇的两次低空风切变的模拟,WRF-LES和WRF-CFD两种模式都可以模拟出第一次低空风切变,而第二次受传入模式的WRF风速数据值较小的影响,两种模式风速差都没有达到阈值,需要在后续工作中进一步验证;(3)低风速条件(6 m·s^(-1))下,基于LSTM的单变量风速预测模型平均绝对误差基本维持在0.59 m·s^(-1),能较好地把握不同地形与环流背景条件下风速变化的非线性关系,虽然受到WRF误差和观测要素不全的限制,多变量风速预测能在保证平均绝对百分比误差小于6.60%的情况下,以更高的计算效率和泛化能力实现风速预测。本文不仅验证了WRF-CFD和WRF-LES耦合方案在风场和低空风切变预报中的差异,还探讨了基于LSTM的风速预测的可行性和准确性,期望为提高风场模拟精度,缩短精细风场模拟时间提供新的视角和方法。
文摘全球气候变化背景下,精确模拟区域碳通量及CO_(2)浓度分布有着十分重要的现实意义.本文基于WRF-GHG(Weather Research and Forecasting Model with Greenhouse Gases Module)模式,综合考虑人为碳排放、陆地生态系统碳交换、海洋碳交换和生物质燃烧碳排放的影响,对2022年中国及其周边地区陆地生态系统碳通量及大气CO_(2)浓度进行在线模拟,并利用OCO-2/OCO-3卫星观测资料评估模式性能.结果表明:(1)WRF-GHG模式整体模拟效果良好(R=0.7424,BIAS=1.3860×10^(-6)),但在低纬度地区的模拟效果略差于中纬度地区,表明该模式目前在亚热带和热带的适用性有限,需要进一步优化;(2)中国区域内,人为碳排放和陆地生态系统源碳交换呈现出显著的季节性特征,其中,人为源CO_(2)排放(全年11031 Tg)在各个排放源中占据主导地位,陆地生态系统(全年-900 Tg)可以吸收约8.2%的全年人为源排放,生物质燃烧源(全年65 Tg)排放则仅为人为源排放的0.6%;(3)模拟区域内,CO_(2)浓度高值区主要分布在我国胡焕庸线以东地区、日本和南亚地区等,在各排放源对CO_(2)浓度的贡献中,人为源排放的贡献量级(1×10^(-6)~100×10^(-6))最高,因而其主导了CO_(2)浓度的空间分布特征.
文摘随着传统化石能源面临枯竭的问题日益加剧,使用太阳能进行光伏发电成为世界各国能源结构调整的重要方向,如何进一步提高光伏发电功率的预测精度成为亟待解决的问题。为提高光伏功率短期预测的准确性和可靠性,提出一种耦合太阳辐射预报模式系统(weather research and forecasting model for solar energy,WRF-Solar)及辐照度订正的光伏短期预测模型,先使用WRF-Solar进行动力降尺度天气数值预报,得到包含辐照度等在内的未来气象因子,再利用随机森林对预报辐照度进行订正,在此基础上运用长短期神经网络、反向传播神经网络和逐步聚类分析建立光伏功率短期预测模型,利用某40 MW光伏电站的实际运行数据进行模型对比分析。结果表明,使用随机森林模型订正后的辐照度更接近真实值,平均绝对误差率下降了56.06个百分点;与另外2种模型预测结果对比发现,长短期神经网络模型预测效果最好,平均绝对百分比误差降低了4.13个百分点,说明组合模型能够进一步提高功率预测的精度。
文摘本文选取GRAPES_MESO(Global/Regional Assimilation PrEdiction System-Mesoscale version)模式和WRF(Weather Research and Forecasting Model)模式在国产鲲鹏(KUNPENG)平台上开展数值模式计算特征分析,并与英特尔(X86)平台进行对比,探讨数值模式在鲲鹏平台上资源使用、计算瓶颈、热点函数等方面的改进空间。结果表明:经过适配后,两个模式在国产KUNPENG平台上能得到与英特尔X86平台一致的计算结果,呈现出较好的并行扩展性;两个模式对CPU的使用率均较高,计算瓶颈主要集中在后端CPU瓶颈,对节点的整体内存使用率适当,后续优化主要集中在代码效率、算法、访存等方面。在KUNPENG平台上,可以考虑通过优化集合通信的Collective Sync、Allreduce和Wait算法,来改善GRAPES_MESO模式的MPI的通信效率;可通过优化GCR算法、以uct、ucg为代表的集合通信热点、以expf、powf等为代表的数学函数、malloc内存操作等热点函数对GRAPES_MESO模式进行优化。
文摘[Objective]Precipitation events caused by Super Typhoon Doksuri in Fujian Province were simulated and evaluated based on the WRF model to provide a reference for typhoon precipitation simulation and forecasting in southeast coastal areas of China.[Methods]The next-generation mesoscale numerical weather prediction model WRF V4.3(The Weather Research and Forecasting Model)was used to simulate the precipitation caused by Typhoon Doksuri in Fujian Province in 2023.Observations from 86 meteorological stations with hourly rainfall records were used to evaluate the model’s performance.Six evaluation indices were used,including the correlation coefficient(R),root mean square error(RMSE),mean absolute error(MAE),equitable threat score(ETS),probability of detection(POD),and false alarm ratio(FAR).[Results](1)The temporal and spatial evolution of precipitation during Typhoon Doksuri was effectively captured by the WRF model.Precipitation intensity increased gradually from July 27 to 29,2023,with the heaviest rainfall concentrated in the northern and eastern coastal areas of Fujian Province.(2)Significant differences in model performance were observed in terms of R,RMSE,and MAE.The largest errors occurred in Putian City,while smaller errors were found in southwestern Fujian Province.The evaluation result of all six indices showed that the WRF model performed best in simulating daily precipitation compared to hourly,three-hourly,six-hourly,and twelve-hourly precipitation.(3)The R95p index indicated that the WRF model successfully captured the overall spatial distribution of extreme precipitation.However,extreme precipitation intensity was overestimated in certain coastal areas.(4)Despite accurately identifying the coastal regions of Fujian as being most affected,the WRF model failed to accurately simulate the spatial distribution and intensity of precipitation.The simulated precipitation centers showed discrepancies when compared with the observed centers.[Conclusion]Although the WRF model underestimated hourly precipitation,it successfully captured the temporal evolution and spatial distribution of rainfall caused by Typhoon Doksuri in Fujian Province.It reproduced the heavy rainfall centers in central Fujian Province,with daily precipitation peaks reaching up to 350 mm.This highlighted the severity of extreme rainfall caused by Typhoon Doksuri.
