Sea surface temperature(SST)is one of the important parameters of global ocean and climate research,which can be retrieved by satellite infrared and passive microwave remote sensing instruments.While satellite infrare...Sea surface temperature(SST)is one of the important parameters of global ocean and climate research,which can be retrieved by satellite infrared and passive microwave remote sensing instruments.While satellite infrared SST offers high spatial resolution,it is limited by cloud cover.On the other hand,passive microwave SST provides all-weather observation but suffers from poor spatial resolution and susceptibility to environmental factors such as rainfall,coastal effects,and high wind speeds.To achieve high-precision,comprehensive,and high-resolution SST data,it is essential to fuse infrared and microwave SST measurements.In this study,data from the Fengyun-3D(FY-3D)medium resolution spectral imager II(MERSI-II)SST and microwave imager(MWRI)SST were fused.Firstly,the accuracy of both MERSIII SST and MWRI SST was verified,and the latter was bilinearly interpolated to match the 5km resolution grid of MERSI SST.After pretreatment and quality control of MERSI SST and MWRI SST,a Piece-Wise Regression method was employed to correct biases in MWRI SST.Subsequently,SST data were selected based on spatial resolution and accuracy within a 3-day window of the analysis date.Finally,an optimal interpolation method was applied to fuse the FY-3D MERSI-II SST and MWRI SST.The results demonstrated a significant improvement in spatial coverage compared to MERSI-II SST and MWRI SST.Furthermore,the fusion SST retained true spatial distribution details and exhibited an accuracy of–0.12±0.74℃compared to OSTIA SST.This study has improved the accuracy of FY satellite fusion SST products in China.展开更多
本文基于卫星遥感的观测海表面温度(Sea Surface Temperature,SST)数据和自然资源部第一海洋研究所全球0.1°分辨率海浪-潮流-环流耦合数值预报模式(The surface wave-tide-circulation coupled ocean model developed by First Ins...本文基于卫星遥感的观测海表面温度(Sea Surface Temperature,SST)数据和自然资源部第一海洋研究所全球0.1°分辨率海浪-潮流-环流耦合数值预报模式(The surface wave-tide-circulation coupled ocean model developed by First Institute of Oceanography,MNR,China,FIO-COM)的预报结果,采用线性回归模型和长短期记忆神经网络(Long Short-Term Memory,LSTM)对SST预报结果进行误差校正。利用2016—2021年的数据开展了一系列对比试验,线性回归模型基于局部线性的假设实现对下一时刻误差的预报,LSTM利用2016—2020年共56个月的历史偏差数据训练模型,使用2021年的数据进行检验。结果表明,线性回归模型和LSTM神经网络都可以改善预报结果,LSTM神经网络相对于线性回归模型的效果更好,SST误差降低70%左右;与线性回归模型相比,经LSTM校正后的各点的偏差的概率密度分布集中在0附近。LSTM方法在统计意义上优于线性拟合且结果更稳定,可进一步推广到海洋数值预报多要素偏差校正。展开更多
Tropical cyclone(TC)activities in the North Indian Ocean(NIO)peak in May during the pre-monsoon period,but the TC frequency shows obvious inter-annual variations.By conducting statistical analysis and dynamic diagnosi...Tropical cyclone(TC)activities in the North Indian Ocean(NIO)peak in May during the pre-monsoon period,but the TC frequency shows obvious inter-annual variations.By conducting statistical analysis and dynamic diagnosis of long-term data from 1948 to 2016,the relationship between the inter-annual variations of Indian Ocean SST and NIO TC genesis frequency in May is analyzed in this paper.Furthermore,the potential mechanism concerning the effect of SST anomaly on TC frequency is also investigated.The findings are as follows:1)there is a broadly consistent negative correlation between NIO TC frequency in May and SST in the Indian Ocean from March to May,with the key influencing area located in the southwestern Indian Ocean(SWIO);2)the anomalies of SST in SWIO(SWIO-SST)are closely related to a teleconnection pattern surrounding the Indian Ocean,which can significantly modulate the high-level divergence,mid-level vertical motion and other related environmental factors and ultimately influence the formation of TCs over the NIO;3)the increasing trend of SWIO-SST may play an essential role in the downward trend of NIO TC frequency over the past 69 years.展开更多
文摘Sea surface temperature(SST)is one of the important parameters of global ocean and climate research,which can be retrieved by satellite infrared and passive microwave remote sensing instruments.While satellite infrared SST offers high spatial resolution,it is limited by cloud cover.On the other hand,passive microwave SST provides all-weather observation but suffers from poor spatial resolution and susceptibility to environmental factors such as rainfall,coastal effects,and high wind speeds.To achieve high-precision,comprehensive,and high-resolution SST data,it is essential to fuse infrared and microwave SST measurements.In this study,data from the Fengyun-3D(FY-3D)medium resolution spectral imager II(MERSI-II)SST and microwave imager(MWRI)SST were fused.Firstly,the accuracy of both MERSIII SST and MWRI SST was verified,and the latter was bilinearly interpolated to match the 5km resolution grid of MERSI SST.After pretreatment and quality control of MERSI SST and MWRI SST,a Piece-Wise Regression method was employed to correct biases in MWRI SST.Subsequently,SST data were selected based on spatial resolution and accuracy within a 3-day window of the analysis date.Finally,an optimal interpolation method was applied to fuse the FY-3D MERSI-II SST and MWRI SST.The results demonstrated a significant improvement in spatial coverage compared to MERSI-II SST and MWRI SST.Furthermore,the fusion SST retained true spatial distribution details and exhibited an accuracy of–0.12±0.74℃compared to OSTIA SST.This study has improved the accuracy of FY satellite fusion SST products in China.
基金funded by the National Natural Science Foundation of China[grant number 42105063]the Youth Training Project of the Key Laboratory for Meteorological Disaster Monitoring and Early Warning and Risk Management of Characteristic Agriculture in Arid Regions[project number CAMT-202302]a funded project of Hengyang Normal University[project number 2022QD11].
文摘本文基于卫星遥感的观测海表面温度(Sea Surface Temperature,SST)数据和自然资源部第一海洋研究所全球0.1°分辨率海浪-潮流-环流耦合数值预报模式(The surface wave-tide-circulation coupled ocean model developed by First Institute of Oceanography,MNR,China,FIO-COM)的预报结果,采用线性回归模型和长短期记忆神经网络(Long Short-Term Memory,LSTM)对SST预报结果进行误差校正。利用2016—2021年的数据开展了一系列对比试验,线性回归模型基于局部线性的假设实现对下一时刻误差的预报,LSTM利用2016—2020年共56个月的历史偏差数据训练模型,使用2021年的数据进行检验。结果表明,线性回归模型和LSTM神经网络都可以改善预报结果,LSTM神经网络相对于线性回归模型的效果更好,SST误差降低70%左右;与线性回归模型相比,经LSTM校正后的各点的偏差的概率密度分布集中在0附近。LSTM方法在统计意义上优于线性拟合且结果更稳定,可进一步推广到海洋数值预报多要素偏差校正。
基金National Natural Science Foundation of China(41965005,41790471,42075013)Key R&D Plan of Yunnan Province Science and Technology Department(202203AC100006)National Natural Science Foundation of Yunnan Province(202201AS070069)。
文摘Tropical cyclone(TC)activities in the North Indian Ocean(NIO)peak in May during the pre-monsoon period,but the TC frequency shows obvious inter-annual variations.By conducting statistical analysis and dynamic diagnosis of long-term data from 1948 to 2016,the relationship between the inter-annual variations of Indian Ocean SST and NIO TC genesis frequency in May is analyzed in this paper.Furthermore,the potential mechanism concerning the effect of SST anomaly on TC frequency is also investigated.The findings are as follows:1)there is a broadly consistent negative correlation between NIO TC frequency in May and SST in the Indian Ocean from March to May,with the key influencing area located in the southwestern Indian Ocean(SWIO);2)the anomalies of SST in SWIO(SWIO-SST)are closely related to a teleconnection pattern surrounding the Indian Ocean,which can significantly modulate the high-level divergence,mid-level vertical motion and other related environmental factors and ultimately influence the formation of TCs over the NIO;3)the increasing trend of SWIO-SST may play an essential role in the downward trend of NIO TC frequency over the past 69 years.
