The Microwave Land Surface Emissivity(MLSE)atlas and instantaneous simulation of all-sky/all-surface MLSE are important prerequisites for satellite data assimilation.A ten-day/month synthesized FengYun-3D MLSE atlas(N...The Microwave Land Surface Emissivity(MLSE)atlas and instantaneous simulation of all-sky/all-surface MLSE are important prerequisites for satellite data assimilation.A ten-day/month synthesized FengYun-3D MLSE atlas(New_FY3D)was constructed by the two global MLSE daily product datasets,clear-sky(FY-3D1)and clear/cloudy(FY-3D2),which were retrieved from the same FY-3D MicroWave Radiation Imager(MWRI)Level-1 brightness temperature(BT)data from 2021 to 2022,respectively.Then,a set of global MLSE label samples based on the New_FY3D,including 14 surface geophysical parameters,was obtained for an instantaneous global MLSE simulation at a 0.10°spatial resolution by adopting the extreme gradient boosting(XGBoost)machine learning method.Finally,the FengYun-3F(FY-3F)MWRI-II BT simulations using the Advanced Radiative Transfer Modeling System(ARMS)based on the above different MLSE products were evaluated.The results show that the New_FY3D atlas performs well,and the BT simulation at the top of atmosphere is better than that of FY-3D1,FY-3D2,and the international mainstream TELSEM2(Version 2.0 for a Tool to Estimate Land Surface Emissivities in the Microwaves)atlas.Surface roughness,vegetation coverage,land cover type,and snow cover are vital parameters for MLSE simulation.The XGBoost model can accurately simulate all-sky/all-surface MLSE instantaneously over the frequency range 10.65–89.0 GHz.The average simulation determination coefficients(R^(2))under clear-sky and cloud-sky conditions are 0.925 and 0.901,respectively,and the average root-mean-square errors(RMSEs)are 0.018 and 0.021,respectively.Large simulation errors occur in permanent wetland,ice and snow,and urban and built-up areas.With a standard deviation of 6.6 K,the BT simulation based on an XGBoost simulated MLSE is better than those based on New_FY3D and TELSEM2.展开更多
Accurate retrieval of atmospheric vertical profiles is critical for improving weather prediction and climate monitoring.However,the complexity of atmospheric processes in cloudy regions poses challenges compared to th...Accurate retrieval of atmospheric vertical profiles is critical for improving weather prediction and climate monitoring.However,the complexity of atmospheric processes in cloudy regions poses challenges compared to those of clear sky scenarios.This study presents a novel framework that integrates Bayesian optimization and machine learning approaches to retrieve atmospheric vertical profiles—including temperature,humidity,ozone concentration,cloud fraction,ice water content(IWC),and liquid water content(LWC)—from hyperspectral infrared observations.Specifically,a Bayesian method was used to refine ERA5 reanalysis data by minimizing brightness temperature(BT)discrepancies against FY-4B Geostationary Interferometric Infrared Sounder(GIIRS)observations,generating a high-quality profile database(~2.8 million profiles)across diverse weather systems.The optimized profiles improve radiative consistency,reducing BT biases from>40 K to<10 K in cloudy regions.To further overcome the limitations of the Bayesian method,we developed a Transformer-Resnet hybrid model(TERNet),which achieved superior performance with RMSE values of 1.61 K(temperature),5.77%(humidity),and 2.25×10^(–6)/6.09×10^(–6)kg kg^(–1)(IWC/LWC)across the entire vertical levels in all-sky conditions.The TERNet outperforms both ERA5 in cloud parameter retrieval and the GIIRS L2 product in thermodynamic profiling.Independent verification with radiosonde and Cloud-Aerosol Lidar and Infrared Pathfinder Satellite Observations(CALIPSO)datasets confirms the framework's reliability across various meteorological regimes.This work demonstrates the capability of combining physics-informed Bayesian methods with data-driven machine learning to fully exploit hyperspectral IR data.展开更多
The cloud liquid water content(LWC)of the Tibetan Plateau(TP)is crucial for cloud water conversion.There are very few accurate observations of the LWC on the TP.This makes our estimation of the LWC and precipitation i...The cloud liquid water content(LWC)of the Tibetan Plateau(TP)is crucial for cloud water conversion.There are very few accurate observations of the LWC on the TP.This makes our estimation of the LWC and precipitation inaccurate on the TP.This paper introduces an indirect estimation scheme for the LWC profile obtained using a monochromatic radiative transfer model(MonoRTM)and microwave radiometers(MWRs)on the TP.The LWC estimation method was improved using an optimization of the difference between the simulated and observed brightness temperature(TB)at specific microwave channels that are sensitive to liquid water.The accuracy of the LWC estimation method depends heavily on the value of the cloud-base environment humidity criterion(CBEHC).Our experiment confirmed that the default CBEHC value of 95%is unsuitable for the TP.For the rainfall scenarios,the optimization method suggested the use of CBEHC values of 81%,76%,and 83%for Mangya,Nagqu,and Qamdo stations,respectively.The new CBEHC values produced a 30 K improvement in the TB simulation when compared to that of 95%CBEHC under rainfall conditions.This demonstrates the robustness of the LWC estimation scheme and its significant improvement in LWC estimation on the TP.For no-rainfall scenarios,the original Karstens model remained suitable for Nagqu station.An adjustment of the CBEHC to 94%for Mangya station resulted in a 1 K improvement of its TB simulation.Qamdo station had a 2.5 K improvement when the CBEHC was adjusted to 98%.The relationship between the simulated TB simulation error and the maximum relative humidity of the radiosonde profiles weakened after CBEHC optimization.Thus,the innovative method proposed in this article provides a practical estimation method for LWC in the TP region.This LWC estimation method has a higher potential for rainfall days than no-rainfall days.Under no-rainfall conditions,the accuracy of the proposed LWC estimation method is sensitive to TB errors included in its measurement and simulation.An accurate estimation of LWC for no-rainfall conditions relies more on the equipment and radiation model.展开更多
Accurate calibration of China's new generation ground-based polarimetric radar(GR) network is crucial yet challenging. Although application of the Dual-frequency Precipitation Radar(DPR) of the Global Precipitatio...Accurate calibration of China's new generation ground-based polarimetric radar(GR) network is crucial yet challenging. Although application of the Dual-frequency Precipitation Radar(DPR) of the Global Precipitation Measurement Core Observatory for GR assessment is well-established, current methodologies are inherently limited. Focusing on three GRs in the Guangdong-Hong Kong-Macao Greater Bay Area(GBA)—strategically selected for their high overlapping coverage(>65%) and distinct from single GR or less dense coverage studies—this work introduces key refinements by integrating innovative enhancements into the volume-matching method(VMM), reflecting a systematic approach to mitigating potential error sources. Specifically, we integrate: 1) a novel frequency correction method that adapts to DPR-observed precipitation phase and type, replacing assumption-based polynomial fitting;and 2) a precise beam time-difference matching approach(accuracy < 1 s) to minimize temporal mismatch errors, which improves upon coarser time averaging methods. Furthermore, we developed statistically robust, optimized threshold criteria based on systematic sensitivity analyses using 11 quality control factors, including precipitation type, bright band effects, and attenuation correction limitations. These criteria establish an enhanced protocol designed to minimize errors arising from instrumental, frequency, and scanning differences. Application of this enhanced methodology to the GBA GRs(2021–2023) yielded a significantly improved matching accuracy(correlation coefficient, CC: 0.90–0.95;standard deviation,STD: 1.2–1.6 dB). A unique contribution of this work is the quantitative estimation of historical calibration errors and operational stability, which was achieved by linking VMM results with operational GR calibration and maintenance records. This analysis revealed decreasing STD trends and identified specific calibration-related events, such as an underestimation of approximately 2.43 dB for the Shenzhen radar following calibration in 2023. Consequently, the refined methodology provides a robust framework for ongoing GR network monitoring and offers a validated pathway for authenticating China's Fengyun-3G(FY-3G) satellite precipitation measurement radar(PMR) data.展开更多
针对卫星遥感因其被动观测卫星的云识别和CTH(Cloud Top Height)产品的精度通常有所欠缺,而主动观测卫星虽然提供了高精度的CTH和云识别信息,但观测范围有限的问题,提出一种GA-LightGBM(Genetic Algorithm-Light Gradient Boosting Mach...针对卫星遥感因其被动观测卫星的云识别和CTH(Cloud Top Height)产品的精度通常有所欠缺,而主动观测卫星虽然提供了高精度的CTH和云识别信息,但观测范围有限的问题,提出一种GA-LightGBM(Genetic Algorithm-Light Gradient Boosting Machine)模型。该模型利用哨兵五号(S5P:Sentinel-5P)、第五代再分析数据(ERA5:Fifth generation ECMWF atmospheric reanalysis of the global climate)、CALIPSO(Cloud-Aerosol Lidar and Infrared Path nder Satellite Observation)的数据,分别进行云识别和CTH预测。使用2018年6月-2020年12月的数据训练模型,并应用2021年全年的数据测试模型性能。实验结果表明,在测试集中,云识别模型的准确率为86%,能很好地识别出云和晴空;云顶高度反演模型的平均绝对误差(MAE:Mean Absolute Error)为1.26 km,均方根误差(RMSE:Root Mean Square Error)为1.87 km,决定系数R2为0.7971,反演结果与真实值存在较好的一致性,证明了方法的有效性。展开更多
激光雷达LiDAR(Light Detection and Ranging)能够精准地还原被测物体的3D结构,是遥感领域最具革新性的技术之一。近几十年来,LiDAR技术取得了快速的发展,并极大地推动了生态与地学领域的相关研究。本文系统回顾并展望了LiDAR硬件和算...激光雷达LiDAR(Light Detection and Ranging)能够精准地还原被测物体的3D结构,是遥感领域最具革新性的技术之一。近几十年来,LiDAR技术取得了快速的发展,并极大地推动了生态与地学领域的相关研究。本文系统回顾并展望了LiDAR硬件和算法的最新发展及其在生态与地学领域的应用。首先,LiDAR的硬件呈现出多样化、高精度的发展态势,特别是近些年无人驾驶技术的成熟极大丰富了近地面LiDAR平台的类型;其次,深度学习、同步定位与地图构建SLAM(Simultaneous Localization And Mapping)、大模型等人工智能技术的发展极大推动了LiDAR算法的进步,使得点云配准、点云分割与分类、点云与多源数据融合等算法不断推陈出新;最后,本文详述了LiDAR在内陆地形测绘、海洋测绘、地质灾害监测、森林结构测量、树木枝干结构网络、3D辐射传输及场景重建、森林微气候模拟、智慧农业、生物多样性、城市与建筑,以及行星测量11个生态与地学分支领域的应用。未来,随着硬件、算法、及LiDAR大数据的进一步发展,LiDAR将继续推动生态与地学的研究,并有望在更多领域发挥重要作用。展开更多
基金supported by the National Natural Science Foundation of China(Grant No.U2242211)the Hunan Provincial Natural Science Foundation Major Project(Grant No.2021JC0009).
文摘The Microwave Land Surface Emissivity(MLSE)atlas and instantaneous simulation of all-sky/all-surface MLSE are important prerequisites for satellite data assimilation.A ten-day/month synthesized FengYun-3D MLSE atlas(New_FY3D)was constructed by the two global MLSE daily product datasets,clear-sky(FY-3D1)and clear/cloudy(FY-3D2),which were retrieved from the same FY-3D MicroWave Radiation Imager(MWRI)Level-1 brightness temperature(BT)data from 2021 to 2022,respectively.Then,a set of global MLSE label samples based on the New_FY3D,including 14 surface geophysical parameters,was obtained for an instantaneous global MLSE simulation at a 0.10°spatial resolution by adopting the extreme gradient boosting(XGBoost)machine learning method.Finally,the FengYun-3F(FY-3F)MWRI-II BT simulations using the Advanced Radiative Transfer Modeling System(ARMS)based on the above different MLSE products were evaluated.The results show that the New_FY3D atlas performs well,and the BT simulation at the top of atmosphere is better than that of FY-3D1,FY-3D2,and the international mainstream TELSEM2(Version 2.0 for a Tool to Estimate Land Surface Emissivities in the Microwaves)atlas.Surface roughness,vegetation coverage,land cover type,and snow cover are vital parameters for MLSE simulation.The XGBoost model can accurately simulate all-sky/all-surface MLSE instantaneously over the frequency range 10.65–89.0 GHz.The average simulation determination coefficients(R^(2))under clear-sky and cloud-sky conditions are 0.925 and 0.901,respectively,and the average root-mean-square errors(RMSEs)are 0.018 and 0.021,respectively.Large simulation errors occur in permanent wetland,ice and snow,and urban and built-up areas.With a standard deviation of 6.6 K,the BT simulation based on an XGBoost simulated MLSE is better than those based on New_FY3D and TELSEM2.
基金supported by the National Natural Science Foundation of China under Grant U2442219Fengyun Satellite Application Pioneer Program(2023)Special Initiative on Numerical Weather Prediction(NWP)Applications,the Civil Aerospace Technology Pre-Research Project(D040405)the Joint Funds of the Zhejiang Provincial Natural Science Foundation of China under Grant No.LZJMZ23D050003。
文摘Accurate retrieval of atmospheric vertical profiles is critical for improving weather prediction and climate monitoring.However,the complexity of atmospheric processes in cloudy regions poses challenges compared to those of clear sky scenarios.This study presents a novel framework that integrates Bayesian optimization and machine learning approaches to retrieve atmospheric vertical profiles—including temperature,humidity,ozone concentration,cloud fraction,ice water content(IWC),and liquid water content(LWC)—from hyperspectral infrared observations.Specifically,a Bayesian method was used to refine ERA5 reanalysis data by minimizing brightness temperature(BT)discrepancies against FY-4B Geostationary Interferometric Infrared Sounder(GIIRS)observations,generating a high-quality profile database(~2.8 million profiles)across diverse weather systems.The optimized profiles improve radiative consistency,reducing BT biases from>40 K to<10 K in cloudy regions.To further overcome the limitations of the Bayesian method,we developed a Transformer-Resnet hybrid model(TERNet),which achieved superior performance with RMSE values of 1.61 K(temperature),5.77%(humidity),and 2.25×10^(–6)/6.09×10^(–6)kg kg^(–1)(IWC/LWC)across the entire vertical levels in all-sky conditions.The TERNet outperforms both ERA5 in cloud parameter retrieval and the GIIRS L2 product in thermodynamic profiling.Independent verification with radiosonde and Cloud-Aerosol Lidar and Infrared Pathfinder Satellite Observations(CALIPSO)datasets confirms the framework's reliability across various meteorological regimes.This work demonstrates the capability of combining physics-informed Bayesian methods with data-driven machine learning to fully exploit hyperspectral IR data.
基金supported by the National Natural Science Foundation of China(Grant Nos.41975009 and U2442213).
文摘The cloud liquid water content(LWC)of the Tibetan Plateau(TP)is crucial for cloud water conversion.There are very few accurate observations of the LWC on the TP.This makes our estimation of the LWC and precipitation inaccurate on the TP.This paper introduces an indirect estimation scheme for the LWC profile obtained using a monochromatic radiative transfer model(MonoRTM)and microwave radiometers(MWRs)on the TP.The LWC estimation method was improved using an optimization of the difference between the simulated and observed brightness temperature(TB)at specific microwave channels that are sensitive to liquid water.The accuracy of the LWC estimation method depends heavily on the value of the cloud-base environment humidity criterion(CBEHC).Our experiment confirmed that the default CBEHC value of 95%is unsuitable for the TP.For the rainfall scenarios,the optimization method suggested the use of CBEHC values of 81%,76%,and 83%for Mangya,Nagqu,and Qamdo stations,respectively.The new CBEHC values produced a 30 K improvement in the TB simulation when compared to that of 95%CBEHC under rainfall conditions.This demonstrates the robustness of the LWC estimation scheme and its significant improvement in LWC estimation on the TP.For no-rainfall scenarios,the original Karstens model remained suitable for Nagqu station.An adjustment of the CBEHC to 94%for Mangya station resulted in a 1 K improvement of its TB simulation.Qamdo station had a 2.5 K improvement when the CBEHC was adjusted to 98%.The relationship between the simulated TB simulation error and the maximum relative humidity of the radiosonde profiles weakened after CBEHC optimization.Thus,the innovative method proposed in this article provides a practical estimation method for LWC in the TP region.This LWC estimation method has a higher potential for rainfall days than no-rainfall days.Under no-rainfall conditions,the accuracy of the proposed LWC estimation method is sensitive to TB errors included in its measurement and simulation.An accurate estimation of LWC for no-rainfall conditions relies more on the equipment and radiation model.
基金National Key Research and Development Program of China (2023YFB3905801)。
文摘Accurate calibration of China's new generation ground-based polarimetric radar(GR) network is crucial yet challenging. Although application of the Dual-frequency Precipitation Radar(DPR) of the Global Precipitation Measurement Core Observatory for GR assessment is well-established, current methodologies are inherently limited. Focusing on three GRs in the Guangdong-Hong Kong-Macao Greater Bay Area(GBA)—strategically selected for their high overlapping coverage(>65%) and distinct from single GR or less dense coverage studies—this work introduces key refinements by integrating innovative enhancements into the volume-matching method(VMM), reflecting a systematic approach to mitigating potential error sources. Specifically, we integrate: 1) a novel frequency correction method that adapts to DPR-observed precipitation phase and type, replacing assumption-based polynomial fitting;and 2) a precise beam time-difference matching approach(accuracy < 1 s) to minimize temporal mismatch errors, which improves upon coarser time averaging methods. Furthermore, we developed statistically robust, optimized threshold criteria based on systematic sensitivity analyses using 11 quality control factors, including precipitation type, bright band effects, and attenuation correction limitations. These criteria establish an enhanced protocol designed to minimize errors arising from instrumental, frequency, and scanning differences. Application of this enhanced methodology to the GBA GRs(2021–2023) yielded a significantly improved matching accuracy(correlation coefficient, CC: 0.90–0.95;standard deviation,STD: 1.2–1.6 dB). A unique contribution of this work is the quantitative estimation of historical calibration errors and operational stability, which was achieved by linking VMM results with operational GR calibration and maintenance records. This analysis revealed decreasing STD trends and identified specific calibration-related events, such as an underestimation of approximately 2.43 dB for the Shenzhen radar following calibration in 2023. Consequently, the refined methodology provides a robust framework for ongoing GR network monitoring and offers a validated pathway for authenticating China's Fengyun-3G(FY-3G) satellite precipitation measurement radar(PMR) data.
文摘针对卫星遥感因其被动观测卫星的云识别和CTH(Cloud Top Height)产品的精度通常有所欠缺,而主动观测卫星虽然提供了高精度的CTH和云识别信息,但观测范围有限的问题,提出一种GA-LightGBM(Genetic Algorithm-Light Gradient Boosting Machine)模型。该模型利用哨兵五号(S5P:Sentinel-5P)、第五代再分析数据(ERA5:Fifth generation ECMWF atmospheric reanalysis of the global climate)、CALIPSO(Cloud-Aerosol Lidar and Infrared Path nder Satellite Observation)的数据,分别进行云识别和CTH预测。使用2018年6月-2020年12月的数据训练模型,并应用2021年全年的数据测试模型性能。实验结果表明,在测试集中,云识别模型的准确率为86%,能很好地识别出云和晴空;云顶高度反演模型的平均绝对误差(MAE:Mean Absolute Error)为1.26 km,均方根误差(RMSE:Root Mean Square Error)为1.87 km,决定系数R2为0.7971,反演结果与真实值存在较好的一致性,证明了方法的有效性。
文摘激光雷达LiDAR(Light Detection and Ranging)能够精准地还原被测物体的3D结构,是遥感领域最具革新性的技术之一。近几十年来,LiDAR技术取得了快速的发展,并极大地推动了生态与地学领域的相关研究。本文系统回顾并展望了LiDAR硬件和算法的最新发展及其在生态与地学领域的应用。首先,LiDAR的硬件呈现出多样化、高精度的发展态势,特别是近些年无人驾驶技术的成熟极大丰富了近地面LiDAR平台的类型;其次,深度学习、同步定位与地图构建SLAM(Simultaneous Localization And Mapping)、大模型等人工智能技术的发展极大推动了LiDAR算法的进步,使得点云配准、点云分割与分类、点云与多源数据融合等算法不断推陈出新;最后,本文详述了LiDAR在内陆地形测绘、海洋测绘、地质灾害监测、森林结构测量、树木枝干结构网络、3D辐射传输及场景重建、森林微气候模拟、智慧农业、生物多样性、城市与建筑,以及行星测量11个生态与地学分支领域的应用。未来,随着硬件、算法、及LiDAR大数据的进一步发展,LiDAR将继续推动生态与地学的研究,并有望在更多领域发挥重要作用。
