Shulou(Shulou.com)11/24 Report--

Satellite remote sensing technology is a rapidly developing high and new technology, and the information network it has formed is constantly providing people with a large number of scientific data and dynamic information.

Original author:

Oleg Dubovik, Atmospheric Optics Laboratory, University of Lille, France

Gregory L. Schuster,NASA Langley Research Center, Hampton, Virginia, USA

Xu Feng, School of Meteorology, University of Oklahoma, Norman, Oklahoma, USA

Hu Yongxiang, School of Physics and Astronomy, University of Leicester, UK

Hartmut B ö sch, National Earth observation Centre, University of Leicester, Leicester, UK

Jochen Landgraf,SRON Dutch Space Research Institute, Utrecht, the Netherlands

Li Zhengqiang, Institute of Aerospace Information Innovation, Chinese Academy of Sciences, Beijing, China

Compilation: Tang Poetry

In the past 50 years, satellite remote sensing has become one of the most effective tools for measuring the earth at the local, regional and global spatial scales. These space-based observations are non-destructive and can quickly monitor the ambient atmosphere, its underlying surface and the ocean mixed layer.

In addition, satellite instruments can observe toxic or dangerous environments without putting people or equipment at risk. Large-scale continuous satellite observations complement detailed (but sparse) field observations and provide unparalleled measurements of volume and content for theoretical modeling and data assimilation.

01. Satellite remote sensing technology is developing rapidly

At present, there are a large number of very important applications that rely on satellite data. Atmospheric observation is used for weather forecast, environmental pollution monitoring, climate change and so on. Ocean surface remote sensing is used to monitor coastline dynamics, sea surface temperature and salinity, marine ecosystems and carbon biomass, sea level change, maritime traffic and fisheries, shallow water currents and bottom topographic maps. Satellite remote sensing has greatly promoted the exploration of mineral resources, monitoring of floods and droughts, soil moisture, vegetation, deforestation, forest fires, agricultural monitoring, urban planning and so on.

Finally, social science efforts to investigate global crises, such as the COVID-19 pandemic, benefit from satellite remote sensing data sets, which classify the human environment using a variety of targeted visualization and then link these observations to various socio-economic data sets.

In addition, satellite remote sensing provides an effective tool for collecting global information, such as:

Planetary topography

Atmospheric profile temperature, water vapor, carbon dioxide and other trace gases

Mineral and chemical composition of the surface and atmosphere

The properties of the cryosphere, such as snow, sea ice, glaciers and melting ponds

Particle and electromagnetic properties of the thermosphere, ionosphere and magnetosphere

Earth remote sensing has also contributed to the development of technological levels, which has contributed to the development of deep space remote sensing missions, such as Voyager and Cassini-Huygens space research missions.

In the early stages of the development of observation satellites, satellite sensors are often designed for specific targets.

For example, in the 1970s a range of instruments were introduced: Landsat and Advanced very High Resolution Radiometer (AVHRR) instruments are designed to monitor land surfaces and clouds, the Total Ozone Mapping Spectrometer (TOMS) instruments focus on observing total column ozone, and High Resolution Infrared radiation sounder (HIRS) instruments support weather forecasting and climate monitoring.

The deployment of these tasks provides unique data for each target topic, and these tasks have been recognized by the corresponding scientific community. These missions have been extended for many years to obtain climate significant data records. Based on accumulated experience, updated missions are continuously improved (in terms of technology and logistics).

The gratifying results of these missions encouraged the design and launch of increasingly advanced instruments with a wider range of observations over the two decades from 1990 to 2010.

For example, tropospheric pollution measurements (MOPITT), orbiting carbon Observatory (OCO) and greenhouse Gas observing Satellite (GOSAT) missions were deployed to launch several thermally enhanced infrared detectors, such as AIRS,TES,IASI,IMG and CRIS, for carbon dioxide (CO2) and methane (CH4) to monitor weather forecasts and atmospheric conditions of climate change.

Other satellite imagers have also been deployed to observe the air, land and sea, and to support interdisciplinary research, such as single-view MODIS, MERIS and SGLI, dual-view ATSR and AATSR, multi-view MISR radiometer and POLDER polarimeter.

In addition to these more traditional passive observations, active measurements such as CloudSat radar and CALIPSO lidar are deployed to monitor the vertical structure of clouds and aerosols, which is important for a variety of atmospheric applications.

All these previous efforts have provided valuable insights to build an understanding of the true value limitations and potential of satellite remote sensing. In fact, the elastic development of space instrument technology and the vigorous development of informatics have created an unprecedented situation. The limitations of hardware, data acquisition and processing have been greatly weakened, and more advanced satellite sensor designs can be developed and deployed.

In addition, the scientific community has acquired a considerable amount of satellite data, accumulated a great deal of experience in managing and analyzing data, and has a truly realistic vision of the possible achievements of existing satellite data sets. and understand the necessary steps to improve the effectiveness of satellite data in the future.

On the other hand, the community is aware that the basic challenges of remote observation are endless. For example, separating the signal from noise to retrieve a specific set of geophysical variables and accurate instrument calibration is an ongoing challenge.

Technological advances have improved the information content of observations, but the data are never sufficient to uniquely describe all the geophysical parameters of interest; as science advances, the list of observable quantities required continues to grow.

Therefore, remote sensing is still a fundamentally ill-posed problem, which needs to be properly defined and constrained by theoretical models, prior knowledge and auxiliary observation. These are important considerations when designing new scientific goals.

02. Major challenges in the development of satellite remote sensing

The overall major challenge for the development of satellite remote sensing is to find innovative and affordable technologies and measurement concepts to solve new problems and to deal with the problems exposed by satellite remote sensing experiments over the past half century. Specifically, several complementary aspects are expected to be addressed:

1. One of the main advantages of satellite remote sensing to improve the spatial and temporal coverage and resolution of satellite observation is that it can quickly observe large areas of the earth. At the same time, the current coverage limits of available satellite data are also obvious.

For example, low Earth orbit (LEO) polar orbit imagers usually achieve global coverage in at least one day (but mostly two days or more), so many natural phenomena with high temporal and spatial variability have not been fully captured.

In this regard, high-orbit geostationary observations (GEO) address this limitation by providing frequent day and night observations of the same object.

However, there is a tradeoff between spatial coverage and satellite image resolution (usually the higher the coverage, the lower the spatial resolution). For many applications, it is desirable to achieve observations with wide spatio-temporal coverage and high spatial resolution, but it is also very challenging.

Therefore, the design of satellite observations may require the synergy of new innovations, auxiliary data and complementary observations to solve specific objects and related problems.

two。 Although the high capability of satellite observation has been clearly recorded, the data provided by our satellite instruments is limited to many applications. Therefore, it is desirable and planned to deploy new sensors with enhanced capabilities.

For example, it has been clearly recognized that multi-angle polarimeters (MAP) provide the most appropriate data for characterizing the detailed columnar characteristics of atmospheric aerosols and clouds, and attention to polarization data in aerosol and cloud characterization is expected to increase significantly over the next decade.

European and U.S. space agencies plan to launch several advanced polarization missions in the coming years, including 3MI (Multi-View Multi-Channel Multi-polarization Imaging Mission) on MetOp-SG, Aerosol Multiangle Imager (MAIA) instruments, Spex (Planetary Exploration Spectropolarizer) and Super-Angle Rainbow Polarimeter (HARP), as part of the NASA PACE mission, Multispectral Imaging Polarimeter (MSIP) / Aerosol-UA MAP instruments as part of the Copernicus CO2M mission, etc.

In addition, the National Space Administration of China (CNSA) has invested heavily in polarization sensors. CNSA has recently launched several polarization remote sensing instruments, including MAI / TG-2, CAPI / TanSat, DPC / GF-5 and SMAC / GFDM, and plans to launch POSP, PCF and DPC-Lidar in the next few years.

The concepts of these instruments, their technical design and algorithm development have been deeply discussed and tested using airborne prototypes, and the concepts of these instruments are discussed in detail by Duibovik et al.

Similarly the number of satellite lidars and radars is expected to increase as active remote sensing instruments provide detailed information on vertical changes in the atmosphere. In fact, most major space agencies are implementing space-based lidar programs.

For example, NASA launched the Geoscience Laser Altimeter system (GLAS) on the ICESat satellite in 2003, the cloud aerosol lidar (CALIOP) with orthogonal polarization on the CALIPSO satellite in 2006, the cloud aerosol transport system (CATS) on the International Space Station in 2015, the Advanced terrain Laser Altimeter system (ATLAS) on ICESat-2 in 2018, and as part of the Life Planet Program (LPP). ESA launched the Aladdin wind lidar on the Aeolus satellite in 2018, and CNSA will launch the DPC-Lidar airborne CM-1 satellite in 2021, and the European / Japan joint EarthCARE satellite (expected to be launched in 2023) will include high-performance lidar and radar technologies that have never been flown in space before.

The success and prospect of these tasks make lidar an important part of the future observation system.

At the same time, in the complex environment, no single sensor can provide comprehensive information about the target object, so it is necessary to explore the synergy of complementary observation.

Even state-of-the-art multi-angle polarimeters cannot guarantee reliable 3D characterization of aerosols because of their limited sensitivity to vertical changes in aerosols and clouds.

Lidar and MAP instruments combine observations to provide a 3D representation of the atmosphere as shown in the figure, and the value of coordinating passive and active observations has been clearly recognized and taken into account in planning satellite missions. For example, the A-Train satellite constellation provides polarization, radiation, lidar and other supplementary data. Similarly, ongoing NASA aerosol and cloud, convection and precipitation (ACCP) studies consider coordinated observation through the deployment of passive (polarimeter, spectrometer, microwave radiometer) and active (lidar and radar) sensors.

In addition, the next generation of remote sensing inversion intends to explore collaborative inversion which depends on the results observed by different instruments.

For example, the main challenge in retrieving atmospheric aerosol properties is to distinguish light scattered by aerosol particles from clouds, atmospheric gases and underlying surfaces. Satellite sensors designed specifically for aerosol monitoring, such as MODIS (radiometer) or POLDER (polarimeter), may not have the best ability to remove clouds, gases, and surface pollutants.

Similarly, land reflectance observations usually need to eliminate scattering from atmospheric aerosols and gases; similarly, atmospheric gas monitoring may also be affected by aerosol and cloud pollution.

Therefore, it is always advisable to observe a variety of instruments with different sensitivities for clouds, aerosols and gases. The analysis of aerosol measurements contaminated by clouds and trace gases on inhomogeneous surfaces may benefit from the collaborative measurement of multiple instruments. Infrared images, lidar and radar observations can be used to limit the cloud part. Spectral data are highly sensitive to gas concentration, and high-resolution images help to reduce the uncertainty related to surface heterogeneity.

For example, simultaneous multi-angle polarization observations through MAP / CO 2 M and DPC2 deployed within the CO framework anticipate that the EU / Copernicus and GF-5 China missions will provide information for atmospheric correction and greenhouse gas monitoring.

In fact, carbon monoxide and other gases obtained by instruments such as OCO and GOSAT are only available under very few atmospheric aerosols.

In this regard, the addition of MAP observation 2 to Copernicus CO predicts that Mission M will improve GHG characterization in moderate and possibly highly aerosol situations. As part of the NASA PACE mission, SPEX and HARP rotator instruments are expected to supplement hyperspectral radiation data from OCI, thereby providing more accurate aerosol information and helping to retrieve ocean surface and subsurface properties.

Another example of a satellite that combines polarization measurements with high-resolution spectrum is the high-resolution and multimode integrated imaging satellite successfully launched by CNSA on July 3, 2020.

Finally, the cooperative use of satellite data with ground and suborbital target measurements is also an important consideration.

For example, the landscape and surface properties of dense urban centers tend to be highly heterogeneous. Therefore, atmospheric processes and dynamics are affected by local activities of height variation, and the observation of ambient air quality needs high time frequency and high spatial resolution. Small and cheaper so-called nano-or cube satellite constellations with different equatorial crossing times can improve coverage by increasing the number of orbital instruments.

In addition, surface measurements and conventional suborbital measurements can also enhance satellite retrieval in densely populated areas.

For example, Liu et al proposed a PM monitoring technique that uses ground measurements and geostatistical regression of PM concentrations in juxtaposed satellite observations. When measurements are unavailable or contaminated, chemical transport models are used to fill time and / or space gaps, thus contributing to regression. The MAIA project is using this method to classify aerosol components such as sulfate, nitrate, ammonium, black carbon, organic carbon and dust. Then spatial PM information and health records are further combined to better understand the relationship between aerosol pollutants and poor public health problems.

Solving the field-of-view differences that combine the observations of different instruments is another challenge. In this regard, the experience gained from using data from multiple instruments with different fields of view deployed in the A-Train constellation provides valuable insights for future tasks.

3. The quality of remote sensing retrieval algorithm, which is the most advanced data processing method of the next generation, is another key aspect that affects the quality of the final product. In fact, once the instrument is deployed, the quality of the observed data can not be fundamentally improved, and the retrieval algorithm is still improving.

The final remote sensing products may be significantly different, not only because of the intake of data from different instruments, but also due to improvements in the concept of retrieval. In this regard, the new generation of remote sensing retrieval algorithms have made great progress in the past decade.

For example, new algorithms often rely on fast and accurate atmospheric modeling (rather than using pre-calculated lookup tables or LUT) and are able to retrieve a large number of parameters. In addition, simultaneous retrieval of aerosol characteristics as well as surface and / or cloud characteristics has been implemented.

Finally, within the EU / Copernicus framework, the joint retrieval of carbon dioxide and aerosol characteristics is a promising method to reduce the effect of aerosol pollution on derived carbon dioxide products.

The basic logic of the evolution of retrieval algorithm itself shows that there is a high potential to use the synergism of different observations to improve the accuracy of retrieval.

In addition, the idea of developing independent algorithms for multi-functional instruments that can be applied to different observations or their combinations is becoming more and more popular. Generalized search of aerosol and surface properties (GRASP) is an example of this algorithm. The algorithm can be used for all kinds of satellite and ground passive and active measurements. It has also been successfully applied to the simultaneous cooperative inversion of lidar profile and cylindrical radiation measurements.

There are still some algorithmic challenges in accurate cloud satellite remote sensing. An accurate and efficient radiative transfer model is a prerequisite. Although the independent column approximation method is widely used to retrieve cloud droplet size and optical depth, the three-dimensional radiation transfer (RT) effect caused by cloud horizontal inhomogeneity (cloud top roughness) may be the reason for the retrieval deviation.

The 3D nature of clouds becomes more concerned when studying the interaction between clouds and aerosols (for example, at the edge of the cloud), starting with coupling their retrieval to a joint framework. Under this background, there is an urgent need to develop a fast and accurate 3D RT model for geometrical and optical complex media, combined with the spectral characteristics of gas absorption and the correct use of cloud particle scattering model. There is also a need to develop reliable three-dimensional radiation models to illustrate the horizontal heterogeneity of the surface in order to fully interpret all satellite images.

Another related outstanding problem is the construction of 3D cloud field to simulate 3D radiation field, which can be solved by a combination of active and passive sensors.

Many modeling and observational studies have shown that cirrus clouds play an important role in promoting weather and climate processes. Although optically thin, cirrus clouds have a global presence and regulate Earth's radiation and play an important role in the climate system. Cirrus particles have a highly irregular shape and their single scattering properties (such as single scattering albedo and scattering phase function) are significantly different from those of spherical particles.

If the algorithm cannot recognize these irregular shapes, they may lead to significant deviations in aerosol and cloud retrieval. Therefore, it is a very promising direction to identify representative cirrus particle models and incorporate them into aerosol retrieval.

In addition, the use of satellite data is highly consistent with the progress made in the Global Climate and Chemical Migration Model (CTM). For example, reliable aerosol retrieval can be absorbed into the Chemical Transport Model (CTM) to provide accurate aerosol loads when observations are not available.

At the same time, spectral and polarization information are highly sensitive to constrained aerosol types, and satellite data may provide additional restrictions for improving global emissions of atmospheric components in chemical transport models.

Therefore, the coordination of satellite data processing and existing model information is another promising research field to promote satellite remote sensing.

Finally, machine learning methods are increasingly used to extract patterns and insights from remote sensing and geospatial data. This branch of artificial intelligence is well suited for analyzing and interpreting Earth observation data and is attractive because it proposes ways to "learn" from the data, identify patterns and make decisions with minimal human intervention.

In particular, the emerging technologies of deep learning and deep neural networks have recently been used in remote sensing research, especially for processing and analyzing large amounts of data. These techniques reveal the potential to automatically extract spatio-temporal relationships and gain further knowledge that helps to improve the prediction and modeling of physical phenomena observed on multiple time scales.

These methods are very attractive for satellite data analysis, especially the combination of the versatility of data-driven machine learning and physical process models.

4. Achieving consistent satellite observations and continuous recording of long-term data long-term and high-quality recording of basic climate variables is essential for monitoring and studying Earth's climate change. A necessary condition for achieving this goal is the continuity of observations, which can be ensured only if high-quality data collection continues without being disrupted. Otherwise, the gaps in multi-instrument data records cannot be explained correctly, and the value of satellite records almost disappears.

Therefore, the absolute calibration of each instrument and the mutual calibration of multiple related sensors are still crucial to the success of almost all targets in satellite remote sensing. Calibrating many instruments is challenging, especially for small satellite constellations.

The National Academy of Sciences, the Academy of Engineering and the Medical School of the United States stressed the need to maintain long-term observations; the CLARREO (Climate absolute radiation and Refractive Index Observatory) mission is the first attempt to define a satellite mission committed to this goal. Similar to direct observation calibration, next-generation satellite products are critical to the traceability and consistency of today's instrument kits.

Conclusion therefore, more than half a century after the launch of the first satellite, remote sensing of the Earth from space has developed into a highly complex tool that powers basic science and supports day-to-day activities that are vital to humankind.

A large number of satellite instruments have been developed and launched, which provide a large amount of data for various needs. The number of satellite instruments and the quality and scope of information collected by satellites are constantly improving.

At the same time, the experience accumulated in the satellite remote sensing field reveals the challenges that need to be addressed in the future development. Although each observation of the atmosphere, land or ocean surface (as well as other remote sensing areas) may have specific and completely different problems, there are some common conceptual challenges in many satellite remote sensing disciplines.

Specifically, the improvement in data value and the efficiency of satellite remote sensing methods may be related to success in the following areas:

Improve the coverage and resolution of observation space and time records

Increase the information content of observations by deploying satellite instruments with enhanced capabilities and exploring the synergy of complementary observations, for example, the synergism of passive images with active vertical atmospheric profiles and hyperspectral spectroscopy, combine observations of different sensitivities obtained in different spectral ranges or on different time or spatial scales And combine satellite observation with suborbital observation and chemical observation to transmit the results of the model.

Develop the next generation of state-of-the-art data processing methods that rely on rigorous forward modeling and numerical inversion methods, taking into account a wide range of state parameter sets (for example, joint retrieval of parameters representing different atmospheric components (such as gases, aerosols, clouds and underlying surfaces) and the use of new solution techniques such as deep learning and neural networks

Achieve consistent satellite observations and continuity of long-term data sets by ensuring adequate compatibility and consistency of past, present and future data sets required for the accumulation of climate records.

This article comes from the official account of Wechat: new Research (ID:chuxinyanjiu), author: Tang Shi

Welcome to subscribe "Shulou Technology Information " to get latest news, interesting things and hot topics in the IT industry, and controls the hottest and latest Internet news, technology news and IT industry trends.