Climate change,natural disasters,pollution,and fast urbanization have made environmental security a more serious international issue.Timely,accurate,and multi-dimensional information is essential in the effective moni...Climate change,natural disasters,pollution,and fast urbanization have made environmental security a more serious international issue.Timely,accurate,and multi-dimensional information is essential in the effective monitoring and management of such complex challenges in the environment.The Earth Observation(EO)systems,including optical sensors,radar sensors,Light Detection and Ranging(LiDAR)sensors,thermal sensors,Unmanned Aerial Vehicle(UAV)sensors,and in-situ sensors,offer a good coverage of space and time,as well as provide useful information on land,water,and atmospheric processes.But the shortcomings or weaknesses of individual sensors,such as their vulnerability to weather conditions,spectral or spatial resolution,and gaps in time,can tend to limit their ability to provide a complete picture of the environment.One of the solutions has been multi-sensor fusion,which combines heterogeneous data and makes it more accurate,robust,and interpretable.This systematic review analyzes the latest methods of multi-sensor fusion,which are machine learning,deep learning,probabilistic models,and hybrid approaches,in terms of methodological principles,preprocessing needs,and computational frameworks.Applications in environmental security are highlighted,which include monitoring natural disasters,monitoring of climate and ecosystem,pollution monitoring,monitoring of land use change,and early warning systems.The review also covers evaluation measures,validation plans,and uncertainty measures,where a strict measure of evaluation is vital to making actionable decisions.Lastly,emerging issues,e.g.,data heterogeneity,computational needs,sensor interoperability,and prospects in the future,e.g.,AI-based adaptive fusion,UAVs and Internet of Things(IoT)integration,and scalable cloud-based systems,are discussed.The synthesis has highlighted the transformational capability of multi-sensor EO in terms of improving the environment in the context of environmental security and sustainable management.展开更多
Hydrological extremes,such as floods,droughts,and compound events,are extremely dangerous to human societies,ecosystems,and infrastructures,whose frequency and severity are affected by climate change more and more.Eff...Hydrological extremes,such as floods,droughts,and compound events,are extremely dangerous to human societies,ecosystems,and infrastructures,whose frequency and severity are affected by climate change more and more.Effective disaster preparedness,water resource management,and climate adaptation have to do with accurate prediction and extensive risk assessment.This review sums up recent progress in predictive modeling and risk assessment systems in the framework of hydrological extremes in the changing climatic conditions.Statistical and empirical techniques,including extreme value theory and nonstationary frequency analysis,give probabilistic information using historic records,whereas process-based models give an understanding of physical hydrological processes at different climate and land-use conditions.New information-based and hybrid methods that use machine learning and high-resolution data take advantage of the complexity and nonlinearities and enhance the predictive power.Hazard,exposure,vulnerability,and adaptive capacity risk assessment models allow predictive output to be translated into actionable decision support,with socio-economic aspects and analysis of the scenario.Case studies of various regions across the globe show the use of these techniques to address floods,droughts,and compound events,with success and current problems.The review also addresses current trends such as compound hazard,multi-hazard integration,AI-enabled modelling,and cross-sectoral decision support,and outlines research priorities of improving predictive capability and resilience.This review will inform researchers,policymakers,and practitioners by offering a synthesis of all the effects of the hydrological extremes in climate change to formulate sound strategies for alleviating these effects.展开更多
文摘Climate change,natural disasters,pollution,and fast urbanization have made environmental security a more serious international issue.Timely,accurate,and multi-dimensional information is essential in the effective monitoring and management of such complex challenges in the environment.The Earth Observation(EO)systems,including optical sensors,radar sensors,Light Detection and Ranging(LiDAR)sensors,thermal sensors,Unmanned Aerial Vehicle(UAV)sensors,and in-situ sensors,offer a good coverage of space and time,as well as provide useful information on land,water,and atmospheric processes.But the shortcomings or weaknesses of individual sensors,such as their vulnerability to weather conditions,spectral or spatial resolution,and gaps in time,can tend to limit their ability to provide a complete picture of the environment.One of the solutions has been multi-sensor fusion,which combines heterogeneous data and makes it more accurate,robust,and interpretable.This systematic review analyzes the latest methods of multi-sensor fusion,which are machine learning,deep learning,probabilistic models,and hybrid approaches,in terms of methodological principles,preprocessing needs,and computational frameworks.Applications in environmental security are highlighted,which include monitoring natural disasters,monitoring of climate and ecosystem,pollution monitoring,monitoring of land use change,and early warning systems.The review also covers evaluation measures,validation plans,and uncertainty measures,where a strict measure of evaluation is vital to making actionable decisions.Lastly,emerging issues,e.g.,data heterogeneity,computational needs,sensor interoperability,and prospects in the future,e.g.,AI-based adaptive fusion,UAVs and Internet of Things(IoT)integration,and scalable cloud-based systems,are discussed.The synthesis has highlighted the transformational capability of multi-sensor EO in terms of improving the environment in the context of environmental security and sustainable management.
文摘Hydrological extremes,such as floods,droughts,and compound events,are extremely dangerous to human societies,ecosystems,and infrastructures,whose frequency and severity are affected by climate change more and more.Effective disaster preparedness,water resource management,and climate adaptation have to do with accurate prediction and extensive risk assessment.This review sums up recent progress in predictive modeling and risk assessment systems in the framework of hydrological extremes in the changing climatic conditions.Statistical and empirical techniques,including extreme value theory and nonstationary frequency analysis,give probabilistic information using historic records,whereas process-based models give an understanding of physical hydrological processes at different climate and land-use conditions.New information-based and hybrid methods that use machine learning and high-resolution data take advantage of the complexity and nonlinearities and enhance the predictive power.Hazard,exposure,vulnerability,and adaptive capacity risk assessment models allow predictive output to be translated into actionable decision support,with socio-economic aspects and analysis of the scenario.Case studies of various regions across the globe show the use of these techniques to address floods,droughts,and compound events,with success and current problems.The review also addresses current trends such as compound hazard,multi-hazard integration,AI-enabled modelling,and cross-sectoral decision support,and outlines research priorities of improving predictive capability and resilience.This review will inform researchers,policymakers,and practitioners by offering a synthesis of all the effects of the hydrological extremes in climate change to formulate sound strategies for alleviating these effects.
