Conference Agenda

This research and collaboration project aims at the calibration and validation (Cal/Val) of the European Sentinel-3, Sentinel-6 and the Chinese HY-2 satellite altimeters using two permanent Cal/Val facilities: (1) the Permanent Facility for Altimetry Calibration established by ESA in Crete, Greece and (2) the National Altimetry Calibration Cooperation Plan of China. Other satellites, such as the Guanlan, CryoSat-2, CFOSAT, CRISTAL, etc., may also be supported by these Cal/Val infrastructures.

Satellites are being calibrated and monitored using uniform, standardized procedures and protocols while exploiting trusted and indisputable reference standards at both Cal/Val infrastructures in Europe and China. At present, the PFAC, Greece implements the action plan established by ESA for Fiducial Reference Measurements for Altimetry and reports its Cal/Val results along with their FRM uncertainty.

Through the ESA Dragon-5 project, the FRM procedures, protocols and best practices, will be updated, upgraded and followed at both Cal/Val facilities in Europe and China. Calibration of altimeters is accomplished by examining satellite observations in open seas against reference measurements. Comparisons are established through precise satellite positioning, water level observations, GPS buoys and reference models (geoid, mean dynamic topography, earth tides, troposphere and ionosphere) all defined by Cal/Val sites. The final uncertainty (FRM status) for altimeter bias will be attributed to several individual error sources, coming from observations in water level, atmosphere, absolute positioning, reference surface models, transfer of heights from Cal/Val sites to satellite observations, etc.

During this second Dragon5 year, the following tasks are being carried out:

• Two European Cal/Val sites in the west and south coast of Crete, Greece have been employed for the calibration of HY-2B satellite altimeter;
• Both sites already attain the FRM strategy in terms of their infrastructure, instrumentation, operations, and data processing and analysis;
• The Wanshan Cal/Val facility comprises four sites at Wailingding, Dangan, Zhiwan and Miaowan islands. Multi-disciplinary scientific and ancillary equipment (i.e., GNSS receivers, tide gauges, GNSS buoys, meterorological stations, solar photometers, etc.) support the calibration and validation services;
• Absolute calibration of the HY-2B sea-surface height bias has been determined by European and Chinese Cal/Val infrastructures following diverse and independent methodologies;
• The ESA FRM4ALT guidelines are put into action to estimate the FRM uncertainty for each Cal/Val site;
• Relative calibration of the HY-2B altimeter is carried out with respect to Sentinel-6A MF and Sentinel-3 altimetry mission via crossover analysis.
• Cal/Val results are cross-examined at the two Cal/Val facilities.

The main outcomes and conclusions of this Dragon5 joint work for the 2^nd year of collaboration, are:

• Both European and Chinese Cal/Val sites have become operational and implement independent calibration methodologies for European and Chinese satellite altimetry missions (i.e., Sentinel-6A MF, Sentinel-3A/B, HY-2B/C, etc.);
• Standardization of operations and data processing are followed by both Cal/Val infrastructures in Europe and China using different instrumentation/models and processing algorithms/strategies, while deviations because of geographic location (i.e., bathymetry, tides, etc.) are reported;
• Products of HY-2B satellite altimeter differ depending on the data repository used.;
• The performance of the HY-2B and HY-2C satellite altimeter is within its mission requirements.

Respect to the objectives and schedule of our project, the mid-term report will include on-going activities and results of Bufeng-1 data processing, calibration workflow, and validation of the calibrated results on hurricane winds, soil moisture, and sea level measurements. The presentation has three parts. Firstly, a short introduction will be given about Bufeng-1 and recent Chinese GNSS-R missions. Secondly, by utilizing the Bufeng-1 Normalized Bistatic Radar Cross Section (NBRCS), earth reflectivity, and range measurements, the preliminary results show that BuFeng-1 has a high agreement compared with other observations on severe sea surface winds, soil moisture, and sea level. In this presentation, the measurements of Bufeng-1 will be aligned with SFMR collected hurricanes, SMAP derived soil moisture, and DTU18 sea level models. Then, the validations of the accuracy and correlation coefficients will be analyzed to discuss the limitations and issues for the future research. For the last part, we will give the outlook about our future works of the objectives and the future plan of Chinese GNSS-R missiions.

This report provides an overview of the recent progress on the cross-calibration and validation of CSES/Swarm satellite magnetic field and plasma measurements. The main results are as follows:

(1) The first comprehensive comparison of ion density (Ni) in the topside ionosphere measured by the Langmuir probe (LP) and faceplate (FP) of the thermal ion imager on board Swarm satellites were performed. Results show a systematic difference between the LP and FP derived Ni values, and the systematic difference shows prominent dependences on solar flux, local time, and season. Although both Ni datasets show generally good linear regression with electron density (Ne) measurements from the incoherent scatter radar (ISR) located at Jicamarca, the Ni derived from LP shows an additional dependence on the solar flux, while such dependence cannot be seen in the FP-derived Ni. More light ions (e.g., H +), diffusing down from the plasmasphere to the Swarm altitude, seem to cause the overestimation of Ni from LP during low solar activity. A linear relation between the Swarm LP-derived Ni and ISR Ne is derived, and such a function is recommended to be implemented into further updates of the Swarm LP plasma density data.

(2) A detailed analysis for the correlation between electron density (Ne) and temperature (Te) at the topside ionosphere were carried out. In situ measurements from four satellites have been utilized, including the China Seismo-Electromagnetic Satellite (CSES), Swarm A and B, as well as the earlier Challenging Minisatellite Payload (CHAMP) satellite. Observations from the four satellites show generally consistent relationship between the Ne and Te at the topside ionosphere. When Ne is low, the Te is negative correlated with Ne, while the slop of negative relation becomes shallower or even reverses to a positive relation after Ne exceeds a certain threshold. Interestingly, two abnormal features of the Swarm Te measurements are observed: a) when Ne is lower than 1×10¹¹m⁻³, Te sometimes becomes very scatter at low and middle latitudes; b) when Ne is larger than 1×10¹¹m⁻³, Te is grouped into two branches at the equatorial and low latitudes. Further analysis reveals that the flags used in the Swarm Level-1 B plasma density product cannot well distinguish the two abnormal features of Te, implying further efforts are needed for the Swarm Te data calibration.

(3) Based on the in-orbit magnetic field data of China Seismo-Electromagnetic Satellite (CSES) and Swarm Bravo satellite, some researches on the cross calibration and correction technology were carried out. The condition applied is that two satellites pass by in a relatively short period of time and through spatial location at a relatively close range, and set different spatial-temporal scale standards, combined with Kp index to screen for geomagnetic quiet periods. Then with the help of CHAOS model, indirect analysis was realized. Furthermore, the difference between the in-orbit data and model value was visualized, and the phenomenon and possible reason of data variation with time and geomagnetic latitude variation were analyzed. According to the analysis results from 2019 to 2020, the scalar magnetic field detection payloads of the two satellites have maintained long-term stability in-orbit. Both scalar magnetic field data are in good agreement with CHAOS model and relatively consistent and stable. The difference between the data and the model is mainly distributed in the geomagnetic high latitude region. The results of the study can evaluate the reliability of the satellite magnetic field data and the consistency of multiple satellites detection results. Applying them to the field of in-orbit data processing and analysis may improve data accuracy and reliability, and further optimize the data processing method, which may provide a methodological reference for doing similar evaluation and analyzation subsequently.

Sea ice is a highly sensitive indicator of past and present climate change. The demand for getting comprehensive, continuous, and reliable sea ice information from multi-source satellite data is growing as a result of climate change and its impact on environment, regional weather conditions, and on human activities such as operations in ice-covered ocean regions. This paper provides an overview of the Dragon 5 project dealing with synergistic monitoring of arctic sea ice by multi-source remote sensing data. For operational ice charting, new methods were developed to classify sea ice types from SAR imagery and CFOSAT SWIM data. Operationally used dual-polarization C-band wide-swath data were complemented by corresponding L-band images, and the benefit of an L- and C-band combination for ice type separation and ice feature detection for winter and summer conditions was assessed. The discrimination ability of sea ice types at small incidence angle provided by CFOSAT SWIM data was investigated and analyzed. The polarimetric backscatter behavior of sea ice in L-, S-, and C-band SAR images was compared to spatially and temporally coincident airborne SAR campaign data. A second topic included implementation and development of sea ice concentration (SIC) estimation and SIC noise reduction algorithms with the Chinese microwave radiometers such as e.g. the HY-2 Microwave Radiometer and the FY-3 Microwave Radiation Imager. We investigated the brightness temperature signatures of different surface types in various sea ice and weather conditions. The uncertainty and error statistics of the retrieved SIC are determined using validation data from in-situ measurements and high-resolution SAR satellite data. Thirdly, we proposed sea ice thickness retrieval algorithms from SAR and altimeter data (e.g. CryoSat-2, Sentinel-3 and HY-2). Especially, the consistency and intermission bias were compared and assessed by using different altimeters, upward looking sonar (ULS) instruments and Operation IceBridge (OIB) data. A method of merging multiple altimeter data to improve the temporal-resolution of the ice thickness product was proposed. Another effort was to develop robust and automated methods for iceberg detection in sea ice and on the open ocean. The study analyzed and evaluated the capability of the proposed methods using different radar frequencies, and in dependence of spatial resolution, incident angle, and the surface conditions around the icebergs.

High Mountain Asia (HMA), including the Tibetan Plateau and the adjacent mountains, is the region with the largest ice masses in the world. Known as the "Water Tower of Asia", meltwater from snow and glaciers feeds the major of Asia's rivers and is an important freshwater resource, affecting the regional water cycle and ecology. Here, we combine Earth System Observations (ESOs) and a novel eco-hydrological model to deliver a new understanding of the cryosphere and water cycle of key water towers of High Mountain Asia (HMA). We focus on blue (runoff) and green (evapotranspiration) water interactions in HMA, to integrate water supply changes due to a vanishing cryosphere with the effect of vegetation to dampen or amplify those changes, especially in periods of droughts. Our investigationstrategy has three steps, and we have progressed substantially on all three.

In the first step, we aimed to substantially advance our understanding of cryospheric, vegetation and land surface changes from remote sensing observations at benchmark sites. We have generated snow cover and glacier extent, glacier mass balance, glacier flow velocity, soil moisture, snow water equivalent, snow and glacier albedo, snow and glacier radiation balance. We studied snow cover variability and its impact on glaciers in the high mountain ranges of the Tarim Basin using multi-temporal remote sensing data. We carried out snow cover monitoring by using Landsat 8 / OLI clear sky condition data with a spatial resolution of 30m. Due to the complex terrain of the mountainous area, in view of the general overestimation problem of the normalized snow cover index method in the Himalayas, we used the support vector machine (SVM) classification method to select the snow cover training samples of different terrain, shadow and other conditions on a scene-by-scene basis to monitor snow accumulation from 2013 to 2020. Using Sentinel-2 as a reference, our retrievals gave a correlation coefficient above 0.95, and root mean square error about 0.1%. We produced a Global Daily-scale Soil Moisture Fusion Dataset (GDSMFD) for the period (2011-2018) at 25km spatial resolution by applying the Triple Collocation Analysis (TCA) and Linear Weight Fusion (LWF) methods. The data set was evaluated against in-situ measurements at 331 sites worldwide, including 57 sites in China, including all the permanent observatories on the Tibetan Plateau. We retrieved glacier albedo in the Western Nyainqentanglha Mountains (WNM) with MODIS data to characterize its spatiotemporal variability from 2001 to 2020. Glacier albedo experienced large inter-annual fluctuations, with an important decreasing trend of 0.043±2.2×10-4 per decade. A new parameterization of snow albedo was developed by combining WRF estimates of snow depth and age with MODIS retrievals of snow albedo, and led to significant relative reductions in RMSE and increases in correlation coefficients in WRF predictions of air temperature, albedo, sensible heat flux and snow depth.
Two advanced algorithms to retrieve land surface albedo and aerosol amount and properties were developed to improve: a) the characterization of the land surface background to separate the surface and atmospheric signals in aerosol retrievals and b) the separation of direct and diffuse irradiance in the retrieval of land surface albedo. The correlation of snow and ice albedo with aerosol loading was documented by comparing the evolution of Aerosol Optical Depth (AOD) at the ITP Nam Co observatory with glacier albedo in the Western Nyainqentanglha Mountains (WNM) during 2009 – 2018. We optimized the procedure to extract a Digital Elevation Model (DEM) from ZiYuan-3 (ZY-3) Three-Line-Array (TLA) stereo images and estimated the geodetic mass balance of glaciers in two areas of the Nyainqentanglha Mountains (NM) using ZY-3 DEMs and the C-band Shuttle Radar Topography Mission (SRTM) DEM in the periods 2000–2013, 2013–2017 and 2000–2017.
A time series of glacier surface velocity in the Parlung Zangbo Basin, southeast Tibetan Plateau, was generated applying the normalized image cross-correlation method to Sentinel 2 (S2) MSI and Landsat-8 (L8) OLI image data from 2013 to 2020.
Enhanced-resolution passive microwave satellite data (PMW) were used to investigate the Tibetan Plateau Lake Ice Phenology (PMW LIP). The Freeze Onset (FO), Complete Ice Cover (CIC), Melt Onset (MO), and Complete Ice Free (CIF) dates were derived for 109 lakes, including 22 lakes for the period 1978 to 2021 and 87 lakes for 2002 to 2021.

Our second step of investigation focuses on generating glacier-specific altitudinal surface mass balance profiles that provide patterns of changes in glacier mass balance at the project study sites. Our approach is based on high-quality digital elevation change and glacier surface velocity datasets along we estimated ice thickness by applying the continuity equation. We derived multidecadal altitudinal mass balance profiles and quantified the equilibrium line altitude and accumulation area ratio for over 5000 glaciers across High Mountain Asia. We applied high resolution Pleiades, Deimos, and UAV datasets to retrieve precise glacier thinning and velocity datasets for the project selected catchments. . These results provide a crucial dataset to validate the eco-hydrological land surface model including its glacier components.

In the third, integrative step, we produce simulations of the land-surface interactions across the cryosphere, hydrosphere and biosphere of the selected study catchments. We used the land-surface model Tethys-Chloris, which describes both vegetation biophysics and cryospheric processes such as snow and ice melt, snow gravitational redistribution and snowpack processes. We have setup the model for five of the study catchments to date, forced with downscaled ERA5 data . We carefully validated the model simulations with the multiple datasets obtained in step one and two. The latent heat flux by snow sublimation and evapotranspiration can account for water losses as high as ice melt.

Part 1, Glacier velocity estimation based on Sentinel-2 observations at the Karakoram.

Twin satellites of Sentinel-2A/B provide 5-day repeat observation to the Earth and are capable of deriving glacier velocity with high-temporal resolution. Here we take the Karakoram as a study site and proposed a data processing procedure of deriving quasi-monthly glacier flow velocity fields by performing offset-tracking technique to 4 years of Sentinel-2 observations. We performed offset-tracking of each acquisition to its next three almost cloud-free acquisitions to increase the number of redundant observations. Flow speed and direction referenced method is taken to remove the wrong value of offset-tracking. Then an iterative SVD method solves the glacier velocity and removes the observation with a large residual. From Oct 2017 to Sep 2021, our results capture plenty of surged glaciers starting and/or ending their surging phases. Several glaciers show speed up one year ago before their surging phase. Rimo south glacier experienced a full surging phase during our study period and last for about two years, the maximum speed exceeded 9m/day. The normal type of glaciers also presented annually speed up and slow down, with acceleration starting in late April or earth May, and ending before September. We estimated RMSE of estimated velocity is about 0.11-0.29 m/day.

Part 2, Sea ice motion detection using Sentinel-1 imagery feature tracking.

Applying feature tracking techniques to Sentinel-1 imagery generates high resolution sea ice motion fields. However, the bad matching vectors still exist after the NNDR (Nearest Neighbor Distance Ratio) test and contaminate the derived motion fields, which need to be identified and filtered out. We propose two algorithms to eliminate such wrong matching vectors. The first employs the matching results derived by the maximum cross-correlation (MCC) method as the reference motion fields to evaluate such wrong matches. The second method employs the local spatial consistency presumption of sea ice motion fields. A Voronoi diagram is applied to slice the overlapping area of two SAR images into many fractions, and each fraction extends its size 50% outward to calculate the regional mean sea ice flow vector and standard deviation. Any vector within the fraction that exceeds 3 times the regional standard deviation will be recognized as an outlier and filtered out. Two methods are tested to two cases with strong rotation or irregular sea ice motion fields derived from Sentinel-1 imagery. The overall accuracy of our two methods is 93.9% and 98.7%, and they sacrifice 6.12% /1.22% of the correct vectors to filter out 100.0% / 94.12% of the wrong vectors for the MCC referenced filter and Voronoi fragmented filter, respectively.

Part 3, Incidence angle normalization of Sentinel-1 backscatter imagery for Greenland ice sheet.

This study proposes an incidence angle normalization method for dual-polarized Sentinel-1 image for Greenland Ice Sheet. A multiple linear regression model is trained using the ratio between backscatter coefficient differences and incidence angle differences of quasi-simultaneously observed ascending and descending image pairs. Regression factors include geographical position and elevation. The precision evaluation of the ascending and descending images suggests better normalization results than the widely-used cosine-square correction method for HH images and little improvement for the HV images. Another dataset of GrIS Sentinel-1 mosaics in four 6-day repeating periods in 2020 is also employed to evaluate the proposed method and yield similar results. For HH images, the proposed method performs better than the cosine-square method, reducing 0.34 dB RMSE on average. The overall accuracy of our proposed method is 0.77 dB and 0.75 dB for HH and HV images, respectively. The proposed incidence angle normalization method can benefit in applying wide-swath and high temporal resolution Sentinel-1 images for producing backscatter mosaic images for GrIS.

Forests play a critical role in the Earth's ecosystem and strongly impact the environment. Under the threat of global climate change, remote sensing techniques can provide information for a better understanding of the forest ecosystems, early detection of forest diseases, and both rapid and continuous monitoring of forest disasters. This project concerns the topic of ecosystems and spans the subtopics estimation of forest quality parameters and forest and grassland disaster monitoring. The aim is to study and explore the application of multi-source remote sensing technology in forest parameter extraction and forest disaster monitoring, using data fusion of satellite images, drone-based laser scanning and drone-based hyperspectral images. The research contents are tree species classification, forest parameter estimation, and forest insect damage detection.

1. Work performed

We applied for satellite images through ESA and Ministry of Science and Technology of the China. These included images acquired from RADARSAT-2, WorldView-3, Sentinel-1/2, and Gaofen series. These data cover several study areas in China and Sweden, including Gaofeng, Yantai, Fushun, Lu'an, Wangyedian, Genhe and Pu'er in China and Remningstorp in Sweden. Field investigations were carried out in Gaofeng, Fushun, Lu'an and Remningstorp. In these study ares, drone-based multispectral and hyperspectral images and laser scanning data were also acquired and studied.

(1) Satellite images

The satellite images acquired for the study areas in China are:

RADARSAT-2, one image covering Fushun and one image covering Qingyuan.

Sentinel-1, time-series images from 2019 to 2021, covering Gaofeng and Wangyedian.

Sentinel-2, time-series cloud-free images from 2019 to 2021, covering Gaofeng, Lu'an, Wangyedian, Genhe, and Pu'er.

Gaofen-1/2/6, 295 images from 2021 to 2022, covering Gaofen, Genhe and Pu'er.

The satellite images acquired in Remningstorp, Sweden are:

WorldView-3, one SWIR image (June 2021).

Sentinel-2, time-series cloud-free images from 2018 to 2021.

RADARSAT-2, one image each in 2020 and 2021.

Pleiades, one image (29 Apr 2021).

(2) Field investigations data

Field investigation of forest parameters was conducted in Gaofeng, China. The inventory recorded diameter at breast height, tree height, under branch height, and the coordinates of the plots.

Spectral information was collected from healthy and pine nematode-infested forests at different stages in the Fushun and Lu'an study areas.

The forest information of the sample plots in Remningstorp was updated. A controlled experiment was conducted for bark beetle infestation, and the infestation symptoms were recorded.

(3) Technical progress

Tree species classification. We proposed three deep learning models using drone-based hyperspectral data: an improved prototype network (IPrNet), a CBAM-P-Net model of the prototype network combined with an attention mechanism, and a Proto-MaxUp+CBAM-P-Net model of the CBAM-P-Net combined with a data enhancement strategy. We developed ACE R-CNN, an attention mechanism, edge detection and region-based instance segmentation algorithm, to accurately identify individual-tree species using UAV LiDAR and RGB images. The performance of these models was demonstrated in the Gaofeng study area. A tree species classification method based on multi-temporal Sentinel-2 data was developed and compared with the classification using mono-temporal data. The performance was verified at Remningstorp.

Forest parameters extraction. We proposed a mean-shift individual-tree crown segmentation algorithm based on canopy attributes using UAV oblique photography data, and developed an individual-tree biomass estimation model fusing multidimensional features, which has good performance in the Gaofeng study area. We proposed a method to automatically extract high-resolution tree height products by combining ZY-3 stereo images and DEM. We developed a forest aboveground biomass estimation model using Sentinel-2 data and tree height data, which obtained accurate forest aboveground biomass maps in the Wangyedian study area. The potential of using PolSAR data acquired at C- and X-band was investigated to estimate forest aboveground biomass at a test site in southern Sweden. The polarization decomposition method was used to RADARSAT-2 and TerraSAR-X data for estimating forest aboveground biomass.We also investigated the potential of time-series TanDEM-X for monitoring forest growth by using the phase height data. We proposed methods to quantify the effects of thinning and clear-cuts on the phase height and apply the methods on detecting silvicultural treatment.

Detection of forest biotic disturbance. About D. tabulaeformis, we proposed a spectral-spatial classification framework combining drone-based hyperspectral images and RGB images to identify damaged tree crowns. For Bursaphelenchus xylophilus, we analyzed the spectral characteristics of two tree species (Pinus tabulaeformis and Pinus koraiensis) in the study areas of Weihai and Fushun during different infection stages. Sensitive bands were selected and a detection model was constructed to identify the infection stages of Bursaphelenchus xylophilus. For European spruce bark beetles (Ips typographus [L.]) infestation, methods of early detecting infestations were proposed using drone-based multispectral images. We investigated how early the infestation can be detected after an attack. We also compared the sensitivity of Sentinel-2 and WorldView-3 SWIR images in detecting early-stage infestations.

(4) Collaborative Research

Co-supervising 1 PhD student.

One joint research paper (accepted for publication in Ecological Indicators), one manuscript, and one published conference paper (IGARSS 2022).

2. Future Plans

(1) Data acquisition

We would like to apply for TanDEM-X and WorldView-3 images covering the study areas of Gaofeng, Genhe, and Pu'er. We plan to obtain drone-based multispectral and hyperspectral imagery covering Remningstorp. We will continue the controlled experiment and field observation of bark beetle infestations.

(2) The research contents

For tree species classification, we will explore deep learning models for individual-tree and stand-scale tree species classification using WorldView-3 and Sentinel-2 imagery.

For tree forest parameters, we will explore crown extraction methods combining satellite imagery and LiDAR, and monitor regional biomass dynamics using Sentinel-1 SAR data. We will develop a method of detecting forest biomass change using RADARSAT-2 imagery.

For forest insect damage detection, we will explore the early identification method of Bursaphelenchus xylophilus using the acquired hyperspectral data. We will study early identification methods of Ips typographus [L.] based on multispectral and hyperspectral images from UAVs.

(3) Cooperation plan:

Co-supervising 1~2 PhD students.

Co-publishing 2~3 research papers.

Co-organizing an international summer school on forest parameters and deforestation mapping using remote sensing data.

Monitoring the status and dynamics of forest is a major issue in the frame of current climate change analysis, as carbon stock variations for the biosphere represent the major source of uncertainties within the global carbon cycle. Synthetic Aperture Radar (SAR) is an active remote sensing device, able to image the reflectivity of wide environments from space, in a systematic way, independently of weather or light conditions. The penetration of electromagnetic waves into vegetated media makes SAR a unique tool for forest 3-D remote sensing applications. Series of SAR acquisitions are used in this study according to two different processing modes.

The first one uses the coherent information contained in a set of SAR images, acquired with a Polarimetric diversity (PolSAR), Interferometric, i.e. spatial, diversity (InSAR), or Tomographic, i.e. multi-baseline InSAR, diversity (TomoSAR), 3-D imaging purposes. Some very significant results have been found regarding the characterization (forest height, underlying grund topography, and Above Ground Biomass) of tropical forests measured at P band, and participate to the preparation of the upcoming ESA BIOMASS mission. Estimated quantities were further analyzed by comparison with other sources of information, such as the GEDI spaceborne lidar acquisitions , or by evaluating the geophysical properties of the retrieved topographic indicators. At higher frequency bands, the difference between the correlation time of radar echoes measured over forests and the revisit time of a spaceborne SAR platform does not allow to apply classical repeat pass imaging techniques. This working group experienced different approaches able to cope with this serous limitation, and based on the model-based analysis of single-pass InSAR pairs on the one side, and on the reconstruction of classical repeat-pass information from a time series of InSAR pairs. Both approaches led to very promising results at L and X bands.

The second processing mode used in this project is related to the incoherent analysis of time series of SAR images, in order to detect changes an relate them to specific properties of forest. In particular the Tropisco service, providing operational deforestation maps derived from Sentinel 1

data, has been launched recently Details may be found at the following links

https://www.spaceclimateobservatory.org/tropisco-amazonia
https://www.spaceclimateobservatory.org/tropisco-south-east-asia

This joint project combines the use of field, airborne and drone remote sensing, and spectroradiometer data to validate and calibrate innovative new satellite sensors from CNAS, ESA and NASA for forest inventory, assessment and health monitoring applications in China and the UK.

1.Vegetation parameter distribution using airborne LiDAR data - UK

Work with airborne LiDAR has been applied for the estimation of mensuration variables, such as Yield Class, Top Height, Basal Area, Volume and mean DDB plus other variables like fractional cover and LAI, over 120K ha of woodlands across the North of England. Derived point cloud metrics were combined with current growth models in sub-areas of 30x30 m² inside each forest stand. As different parts of the country were surveyed in different years, the growth models helped us to reduce this temporal decorrelation and to bring all estimates to 2022. The comparison of results against field observations produced R² above 0.95 for all estimates. Afterwards, Sentinel-1 has been analysed to detect the perimeter of areas affected by wind damage after the Arwen storm in November 2021. Then, volume lost has been calculated as the intersection of polygons representing wind damage and pixels containing volume estimates.

2. Assessing the effects of drought and stress on productivity using satellite data, field experiments and biomass modelling - UK

Monitoring of drought stress processes in Sitka spruce plantations in Kielder Forest using time-series analysis of different vegetation indices derived from Sentinel-2 and Landsat. In parallel, polytunnel experiments with control and drought groups for this species tracked hyperspectral changes in reflectance to determine early onset drought indicators and identify differing drought tolerances between breeding population clonal types. This process helps us to isolate relevant pigment changes under mild-to-moderate drought stress and to identify drought sensitive indices. Additional work with NDVI, derived from MODIS and Sentinel-2 satellites, was used to drive a biophysical vegetation model to study carbon sequestration over commercial Sitka Spruce plantations. Future work will look to integrate more drought sensitive indices into existing models to improve model accuracy under drought conditions.

Other attempts to detect signs of plant stress in the time series of Vegetation Indices (VIs) computed from Sentinel-2, involved the use of Principal Component Analysis applied to the pixel time series of the VIs to learn features of the data common to most observations (e.g. be that phenological or instrumental) and the learned principal components are used as basic functions in a simple linear regression model to filter out these signals. Remaining features in the observations that cannot be well explained by these components may be indicative of stress.

3.Forest Cover Mapping using Sentinel-2 and GF-6 Data, Pu’er Study area, China

The research area focuses on the surroundings of Pu'er City in Yunnan Province, using Sentinel-2 images analysed through the Google Earth Engine (GEE) platform to extract the spectral features, texture features, and terrain features, combined with field survey data, airborne remote sensing data, and terrain data. A classification data set containing the optimal features was obtained by feature screening. The object-oriented and pixel-based classification methods were used to respectively carry out random forest classification. The results show that the classification accuracy of the object-oriented classification method is higher than that of the pixel-based classification method, with an overall classification accuracy of 88.21% and the Kappa coefficient of 0.865. This was followed by a comparison and verification of classification results. The classification results are compared at pixel level with other published land cover products (including the Dynamic World, ESA, ESRI) to analyze their area consistency and spatial consistency. Accuracy evaluation of all products was carried out combining Pu'er airborne hyperspectral data and LULC ground truth data. Finally, the product inconsistency factors were analysed to improve the quality of classified products. Furthermore, the GF-6 data will be used for Pu'er forest classification, and the classification results will be compared with the Sentinel-2 classification results to obtain the optimal forest covered map.

4. Multi-Scale Biomass Mapping Using Airborne LiDAR Data - China

The LiDAR biomass index (LBI) was extended to airborne LiDAR data for multi-scale forest biomass estimation. Through suitability compensating the laser point clouds of each tree and using a small number of trees measured in the field for model calibration, robust and highly accurate results were obtained. The method was verified by the field measurement data of 20 analytical trees, 133 sample plots and 39 subcompartments with larch plantations in three forest farms. Good performances were demonstrated (R2=0.98 and RMSD=11.85 kg at tree scale, R2=0.77 and RMSD=28.74 t/ha at plot scale and R2=0.86 and RMSD= 144.15 t at stand scale). Through comparing with the existing methods of predicting DBH based on tree height and crown width to calculate biomass, the proposed method shows higher accuracy and obvious versatility among different forest farms. In comparison with the existing LiDAR metrics method, it can obtain more detailed results, but only a small amount of measurement data was needed to calibrate the model.

5. Forest gap identification based on UAV LiDAR

The remote sensing of UAV LiDAR can quickly obtain the three-dimensional spatial information of the forest. The study site is located in the Puer Sun River Reserve in Yunnan, China. The canopy height model (CHM) was derived from the point cloud data of UAV LiDAR. The fixed threshold method was used to identify forest gaps in CHM. The reference data from visual interpretation of images was used for accuracy assessment of forest gap identification. The overall accuracy of the fixed threshold method was 92%, and the spatial distribution of the gap was aggregation. The forest gaps in the study site area were mainly small and medium gaps, showing that there were fewer disturbance events. The spatial distribution of forest gaps and its spatial characteristics in small area subtropical natural forests can be mapped by UAV LiDAR data. Forest gap information from UAV LiDAR can be used for the accuracy assessment and validation for the forest gap derived from GF-7 satellite imagery for large area.

Past year still saw that COVID restrictions hampered FTIR and MAX-DOAS operations, yet we reached some very important milestones. Foremost, after passing the TCCON application and evaluation process, Xianghe has become an official operational TCCON site in September 2021. Data from June 2018 onwards are publicly available from the TCCON data archive (https://tccondata.org/). Also the teams readily implemented the transition from the GG2014 to the GGG2020 TCCON retrieval processing suite, data of which became publicly available in May 2022.

Xianghe FTIR (both TCCON and NDACC-type), and UV-VIS ground-based remote sensing measurements have been and continue to be used for a wide array of research topics, be it satellite validation (S5P, OCO-2/3, TANSAT, etc.), finding novel retrieval strategies (for instance for O₃) or network retrieval strategy harmonisation studies (NDACC HCHO and NO₂). Its location next to Beijing provides an excellent testing ground for retrieval algorithms (satellite and ground-based) as one wants to test ones product under as many conditions as possible (from remote pristine to heavily urbanized, across all continents). This is important for uncovering imperfections in the algorithms, which can then be evaluated and remedied. For instance, when comparing Xianghe TCCON CO₂ dry air mole fractions with OCO-2 and OCO-3 measurements in North-China, we found that the mean bias is close to 0, but that the OCO-3 snapshot area mode (SAM) is about 1.0 ppm overestimated.

In recent years, Europe, USA, and China have started to implement the concept of creating an SI-traceable Satellite – SITSat. A key objective is to enable high-accuracy on-orbit radiometric calibration of other satellite sensors traceable to the SITSat. Cross-calibration is considered as an effective approach to transfer the radiometric reference from the high-accuracy satellite (e.g., SITSat) sensor to other satellite sensors which require periodical calibration. The Simultaneous Nadir Overpass (SNO) method based on strict matching of observations is the ideal and most accurate way to cross-calibrate and compare satellite sensors. However, high-resolution spaceborne sensors tend to have relatively small swaths and thus fewer opportunities to have ideal matching conditions with a reference sensors. However, relaxing the matching conditions; times, view/illumination angles etc to increase the number of matches inevitably leads to increased uncertainty.

The key step in cross-calibration is to establish the proper reference i.e. Top-of-Atmosphere (TOA) radiance/reflectance, corresponding to the observation value of the satellite being monitored. For pseudo invariant calibration sites (PICS), the relatively stable surface and atmospheric conditions help ensure the accuracy of cross-calibration results. Similarly, TOA reflectance models established for PICS sites can also ensure high-repeatability of such sites as a radiometric reference in cross-calibration. For RadCalNet sites, autonomous observation instruments are deployed to obtain surface and atmospheric parameters at the time of overpass of a satellite. RadCalNet can provide satellite operators with SI-traceable TOA spectrally-resolved reflectance derived over a network of sites, with associated uncertainties. Therefore, RadCalNet sites are able to serve as calibration references for high-medium resolution satellite, with uncertainties limited by the radiometric values assigned to them from characterization against another sensor, at present ground based observations. However, with the concept of SITSats the opportunity arises to provide the reference calibration to the RadCalNet site from a spaceborne sensor and use the surface instrumentation to provide a monitor/correction function instead of the absolute reference. Here we describe a new benchmark transfer calibration method for high-resolution spaceborne sensors, which uses RadCalNet sites measurements as the intermediate radiometric reference value.

In this study, the TOA reflectance models of RadCalNet sites were constructed using satellite observation data with high radiometric calibration accuracy. Then the model was used to correct the RadCalNet standard TOA reflectance products. The corrected RadCalNet TOA reflectance was then used as an intermediate radiometric reference, which can be traced back to the reference satellite sensor. The corrected RadCalNet TOA reflectance was then used to calibrate the monitored satellite sensors. Through the uncertainty analysis of this method, the uncertainty of cross-calibration between the reference satellite and the satellite to be calibrated caused by the relaxation of time matching constraint can be reduced.

Taking Baotou site in China and RVP site in US, the proposed method is validated, and the application demonstration is carried out by using the Chinese and European satellites (i.e., Sentinel-2A/2B, GF-1, GF-6 and SV-1). The preliminary uncertainty analysis results show that this method can achieve obtain high-precision calibration coefficients, and the calibration uncertainty is 3.5%-4%.

Plant canopy characteristics are essential indicators of plant growth status, of which the two most commonly reported are the leaf area index (LAI) and leaf chlorophyll content (LCC). They determine the interception and absorption rates of solar radiation by vegetation, thereby implying the plant productivity and yield. They are also the two main driving variables in several ecological and crop models that provide practical guidance for precision agriculture and aid monitor the state and functioning of terrestrial environments. Such biophysical products could indirectly be estimated from the optical Earth Observation dataset through surface reflectance. Nevertheless, the non-destructive estimation from different satellites have varying levels of spatial and spectral resolution. Therefore, the validation of such biophysical products parameters and improving remote sensed characterization of vegetation biophysical properties are of great importance to ensure they meet the requirements for specific applications. The UAV4VAL project objective is to evaluate the capability of UAVs as a source of reference data for validating decametric surface reflectance and vegetation biophysical products like LAI, with a specific focus on the European Sentinel-2 and Chinese Gaofen-6 missions.

In the first year, we focused on the LAI product retrieval from both Sentinel2 and GF6 data and validated with in-situ LAI collection and UAV image. We collected 17 ground LAI measurements at Taizi Mountain, China (30.916°, 112.866°) on 31th October 2020, and gathered the GF6 and Sentinel2 imagery over the same period. LAI-2200C plant canopy analyser and digital hemispherical photography was used for obtaining in-situ LAI. In addition, the drone images were collected by a P4 Multispectral camera with 5 bands. To evaluate whether UVA imagery can well bridge the scale gap between ground measurements and satellite imagery, we first generate LAI validation maps from UAV by constructing a regression model between in-situ LAI and UAV vegetation index(VI). Second, the Sentinel-2 retrieved LAI maps was implemented using SNAP and a similar hybrid LAI retrieval process was applied on GF6, the hybrid LAI retrieval combine the advantages of physical-methods with the learning accuracy and flexibility of non-parametric regression algorithms. Finally, we compared the traditional in-situ LAI validation and the validation using UAV LAI maps on Sentinel-2 imagery.

The results of this study showed Atmospherically resistant vegetation index (ARVI) is the best VI for UAV-based LAI retrieval with R2 of 0.66 and RMSE of 1.00. The validation of Sentinel-2 LAI products using the UAV LAI map outperformed traditional in-situ validation. The UVA validation on LAI has RMSE of 1.566 and MAE of 1.238 while the in-situ validation has higher RMSE and MAE of 2.17 and 1.61, respectively. The ground validation of GF6-derived LAI also got poor result with low R2, i.e., ~0.1 to ~0.15 and high RMSE, i.e., ~1.4 to ~1.6 between GF6-derived LAI and in-situ LAI. However, the GF6 image of the study area was highly influenced by cloud and the LAI retrieval from GF6 still needs further exploration. In addition, a vegetation indices (VIs) sensitivity test between NDVI, MTCI and MCARI based on PROSAIL simulated dataset. MTCI showed the highest correlation with Canopy Chlorophyll Content (CCC), with 0.8577 for UAV sensors and 0.8493 for GF6 sensor. MTCI was recommended for Chlorophyll retrieval in the future. The next step will focus on the in-situ biophysical parameters and UVA image collection in the UK site and compare the LAI-retrieval results between GF6 and Sentinel2.

Continued glacier mass loss in High Mountain Asia impacts freshwater supply in and beyond the mountains. Previous studies have shown large spatial variations in glacier mass balance in this region, but the reasons for this variability are not well understood. We developed a new index based on satellite-derived surface characteristics to discriminate winter- and summer-accumulation type glaciers across High Mountain Asia. Combined with the existing mass balance data, it is found that the accumulation type is closely related with accumulation type. Glacier regions that gain mass predominantly from summer snow have thinned on average nearly four times faster than those gaining most mass in winter (-0.43 ± 0.12 m water equivalent (w.e.) a^-1 vs -0.10 ± 0.06 m w.e. a^-1 from 2000 to 2018). The results highlight the importance of the seasonality of snowfall for the glacier mass budget emphasizing that accurate precipitation fields are paramount to quantify future glacier changes reliably in this region.

The monitoring and forecasting of global water cycle under climate changes indeed require enhancement of satellite remote sensing products in both of spatial resolution and accuracy. To strengthen the ability of microwave remote sensing in global water cycle studies and seek for new opportunities of satellite missions, we put forward research contents as follows in the second year of project implementation:

(1) Continuous L-Band Soil Moisture (SM) Datasets from SMOS and SMAP Observations

The Soil Moisture Ocean Salinity (SMOS) and Soil Moisture Active Passive (SMAP) are two existing satellites capable of providing L-band observations at a global scale. Although both satellites have independently performed calibrations, there are some differences in brightness temperature (TB) between the two. Intercalibrations were conducted to develop a consistent SMOS-SMAP TB, then the multi-channel collaborative algorithm (MCCA), which utilizes information from collaborative channels expressed as an analytical form of brightness temperature at the core channel to rule out the parameters to be retrieved, is adopted to develop a consistent L-band soil moisture dataset. Inter-comparison with other SM products (MT-DCA version 5, and DCA, SCA-H, and SCA-V from SMAP Level-3 products version 7) shows an analogous spatial pattern. The MCCA derived SM had the lowest ubRMSD (about 0.058 m3/m3) followed by DCA (0.061 m3/m3), and an overall Pearson’s correlation coefficient of 0.702 (DCA performed best with R=0.746) when evaluated against in situ observations from 19 dense soil moisture networks. The MCCA generates vegetation optical depth (VOD) at both vertical and horizontal polarization, which were found to have a good linearity with the live biomass and canopy height, though partial saturation exists in the relationship with live biomass of tropical forests but not canopy height. The polarization difference of VOD mainly located at densely vegetated and arid areas. It is important to note that this continuous L-band SM and polarized VOD dataset is expected to improve our understanding of the water-transport process in the soil-vegetation continuum.

(Submitted to Remote Sensing of Environment)

(2) Continuous X-Band Soil Moisture (SM) Datasets from FY-3 Series Observations

Long term SM data with stable and consistent quality are critical for global environment and climate change monitoring. SM products from L-band observations have proven to be optimal global estimations. Although X-band has a lower sensitivity to soil moisture than that of L-band, Chinese FengYun-3 series satellites (FY-3A/B/C/D) have provided sustainable and daily multiple SM products from X-band since 2008. This research developed a new global SSM product (NNsm-FY) from FY-3B MWRI from 2010 to 2019, transferred high accuracy of SMAP L-band to FY-3B X-band. The NNsm-FY shows good agreement with in-situ observations and SMAP product and has a higher accuracy than that of official FY-3B product. At selected dense in-situ networks, it is found that NNsm-FY has a relatively good performance with median CC of 0.66 and median ubRMSE of 0.046 m3/m3, With this new dataset, Chinese FY-3 satellites may play a larger role and provide opportunities of sustainable and longer-term soil moisture data record for hydrological study.

(Submitted to Scientific Data)

The main objective of the project is to develop prototypes of real-time remote sensing (RS) land data assimilation systems (LDAS) for monitoring the water cycle in the silk road endorheic river basins and EUROCORDEX-domain. This will provide a synergic and innovative way to integrate RS data from NRSCC and ESA into terrestrial system models for better quantifying the water cycle at the watershed/regional scale. The objective will be achieved through the following sub-objectives: i) Retrieval of key water cycle variables from multi-source RS data (WP1); ii) Development of real time RS LDAS to integrate RS data into terrestrial system models (WP2); iii) Calibration/validation of terrestrial system models using RS retrievals of key water cycle variables (WP3); iv) Parameter estimations for terrestrial system models based on the LDAS (WP3); v) Closing and quantifying the water cycle at the watershed/regional scale based on the LDAS (WP4).

Two LDAS will be developed in the project, one for the silk road endorheic river basins (LDAS_Silk) and one for EUROCORDEX-domain (LDAS_EU). LDAS_Silk will be based on the recently developed watershed system model and a common software for nonlinear and non-Gaussian land data assimilation (ComDA). LDAS_EU will be based on the recently developed Terrestrial System Modeling Platform (TSMP) and Parallel Data Assimilation Framework (PDAF). Multi-source RS data, from visible to thermal infrared and microwave, will be used to retrieve key ecohydrological variables, such as evapotranspiration (ET), snow coverage area (SCA), snow water equivalent (SWE), snow depth (SD), soil moisture (SM), lake and glacier extents, irrigation, and vegetation density and structure. These data will be used as forcing data, calibration and validation data, and for assimilation into the two LDAS.

In this presentation, the mid-term progress on the project will be reported.

Because of the sheer size of grasslands, grassland degradation makes a far more important contribution than cropland degradation to the global environmental change phenomenon of desertification, which is land degradation in dry areas. About 42% of China's national land area consists of grasslands, and if all the grasslands in the world were brought together they would cover 50% more than the total area of Europe. Yet since there are no reliable estimates of the scale and degree of grassland degradation in the world, the global extent of desertification is equally uncertain, and this also makes it difficult to monitor compliance with the Land Degradation Neutrality (LDN) target of the Sustainable Development Goals (SDGs).

Devising a reliable method for mapping grassland degradation is therefore imperative, and is the aim of this project. In the first part of this project we have critically evaluated all previous research into mapping desertification with satellite data in order to devise such a method. We report eight main findings. First, grassland degradation, like desertification, is a very complex phenomenon, comprising multiple forms of soil degradation and vegetation degradation. Second, grassland degradation is poorly monitored at national scale in most countries. Third, even though monitoring dryland degradation is essential to estimate compliance with the LDN target, little progress in using satellite data and other types of "big data" to compensate for the lack of national data has been reported in the literature since the SDGs were finalized. Fourth, matching the spatial resolution of a sensor to the areal variability of each form of vegetation degradation and soil degradation is necessary to minimize spatial systematic errors, while matching the temporal resolution of a sensor to the turnover time of each component is necessary to minimize temporal systematic errors. Fifth, consequently, a multiple sensor approach is needed to monitor grassland degradation, since while medium (20-80 m) resolution data can be used for initial land use mapping, key grassland biophysical parameters(e.g. biomass, productivity, soil carbon content) estimation and the mapping of some forms of land degradation (such as sandy area expansion), very high (≤ 1 m) resolution data is often needed for mapping most forms of vegetation degradation (e.g. tree density, shrub encroachment) and soil degradation (e.g. gully expansion) in dry areas. Sixth, a multilayer multisensor approach, with land use classification and key grassland parameters estimation preceding degradation measurement, will optimize image processing time and the overall accuracy of grassland degradation assessment. Seventh, reliable methods are still lacking for using satellite data to measure wind erosion and soil compaction. Eighth, there is great scope to experiment with radar and LIDAR sensors to improve the accuracy and coverage of grassland degradation monitoring.

The continuous monitoring of natural multi-hazards, namely large landslides, is of fundamental importance to minimize and prevent actual and future risks. In this Dragon-5 project, Earth Observation (OE) data are used and integrated into various methods for the identification and monitoring of landslide phenomena at different spatio-temporal scales. The upper-Jinsha River region, composed of steep terrain and broken structures, is known to cause frequent landslide disasters. Distributed scatterer SAR interferometry (DS-InSAR) was used to monitor and analyze the Woda landslide area. In the study, 106 Sentinel-1A ascending and 102 Sentinel-1A descending images, covering a time period of almost 5 years (from 2014/11/05 to 2019/09/04 and from 2014/10/31 to 2019/09/11, respectively). The two-dimensional deformation of the landslide revealed that the maximum surface deformation rate in the normal direction is about -80 mm/yr, and in the east-west direction is about 118 mm/yr. According to the rescaled range (R/S) analysis, the Hurst index the deformation trend will continue for some time.

The Gilgit region, Pakistan, was also analyzed using a time series of Sentinel-1 images, covering a time period of about a year (2019 – 2020). Several deformation areas were mapped in the region. A 3D prototype platform, enabling the visualization of the deformation results of the typical active landslides, has been developed, allowing the identification of 3 landslides with potential risk.

The use of Artificial Intelligence (AI) techniques in EO data, to develop/optimize data processing and analysis methods, is also one of the major goals of this project. In this regard, the multitemporal difference interferometric phase diagram of subsidence mining area is obtained by using synthetic aperture radar differential interferometry (DInSAR) techno machine learning (ML) FCN-8s, PSPNet Deeplabv3 and U-Net models were applied to DinSAR data to extract mining subsidence. The results show that the U-Net model presents high detection accuracy and takes short time to run. In order to improve the semantic segmentation and extraction accuracy of mining subsidence, the efficient channel attention (ECA) module is introduced into the traditional U-Net model for training. The detection of earthquake fringes with Vision Transformer (ViT) technic was also tested in the scope of this project. In our methodology we use a dataset composed of 5110 interferograms from LicSAR dataset, that were used to apply DL techniques to recognize patterns. The F1-Score and AUC clearly show that the ViT outperforms VGG19, respectively, with 0.88 against 0.69, and 0.97 against 0.86. However, the major disadvantage of this method consists in the detection of false negatives in overlapped patches. Sometimes the model identifies as “no deformation” patches with deformation fringes but that change in the overlapped patches for the same interferogram. For a better confirmation the method is evaluated applied to the whole interferogram. It can be concluded that the results are higher in both models, proving that some false negatives are adjusted in overlapped patches. ViT remains the model with the best performance. The precision is now equal for both (1.0) and with a significant rise of the Recall and F1-Scores. As main conclusions, we can say that the new and different model proposed in this work to detect fringes in SAR interferograms achieves better results, comparing to Convolutional Neural Networks. This way, we are even closer of obtaining results that can create practical products that may help providing a faster response in earthquakes scenarios.

Monitoring structures of great heritage and historical value, more quickly and effectively, is also one of the major goals of our project. However, only the use of Artificial Intelligence techniques will allow to deal with the huge amount of data that will be generated. The Vilariça Valley, located in the north of Portugal is crossed by an active fault and 3 classified buidings (churches of Torre de Moncorvo, Freixo de Espada à Cinta and Foz Côa, in Portugal) are being used as test sites. More than 1000 Sentinel-1, ERS and Envisat images are already processed providing thousands of points in the villages where these buildings are located. However, only a few points were identified directly in the buildings and in their proximity. To increase the density of points in the building and in its proximity, a time series of 25 PAZ images was commissioned, with the support of ESA. In the near future, these results will be integrated with data from field sensors and LiDAR models, using AI techniques.

Based on the big archived Sentinel-1 satellite radar data during the last 7 years, we developed parallel processing techniques for the detection of subtle ground motion on a phase-stable point basis, so that InSAR deformation could be used for accurately determining tectonic motion or strains induced by anthropogenic activities. In particular, we strengthen our analysis results by joint analysis with 3D GPS data within the same time span if available. In this way, we are able to separate ground motion with long-wavelength errors of InSAR results, which is the major source of errors when tectonic motion is considered. Joint analysis of GPS and InSAR data also provides us a way to separate InSAR surface motion in the Line-Of-Sight direction into vertical and horizontal deformation in particular orientations. But it is worthless to have 3D measurements with InSAR only as it cannot effectively discern the signals in a nearly north-south direction.

We apply our method in various tectonic regions for seismic deformation detections, including the population densified North China Plain, the fast shale-gas development Sichuan basin, and the tectonic active Eastern Tibetan Plateau, etc. In this study, we reduce various InSAR errors, in particular the unwrapping errors as before, while the most important long-wavelength errors are overcome by GPS-assistant joint analysis in this version of processing, so that tectonic motion can be well detected and compared with GPS measurements.

The project aims to develop and apply innovative data analytic methods underpinned with machine (deep) learning technology to analyze and detect seismic anomalies from electromagnetic data observed by the SWARM and CSES satellites along with CSELF network. In this report we will present the progress of anomaly detection algorithms underpinned with Deep Neural Networks (DNN), which focuses on predicting Swarm data as long as possible from the Swarm historic data. We report our investigation into several architectures of Recurrent Neural Networks (RNN), in particular we investigate the development of a Long-Short Term Memory (LSTM) based methods. We present our design and implementation of the LSTM algorithm and predicted results of applying the algorithm to learn from the Swarm historic data. We will also describe error metrics for measuring the accuracy of Swarm data reconstruction and the methods of detecting anomalies in the Swarm data, which are in relation to three earthquakes.