Conference Agenda

Emission and backscattering signals of land surfaces at different frequencies have distinctive responses to soil and vegetation physical states. The use of multi-frequency combined active and passive microwave signals provides complementary information to better understand and interpret the observed signals in relation to surface states and the underlying physical processes. Such a capability also improves our ability to retrieve surface parameters and states such as soil moisture, freeze-thaw dynamics and vegetation biomass and vegetation water content for ecosystem monitoring. We present here a prototype Community Land Active Passive Microwave Radiative Transfer Modelling platform (CLAP) for simulating both backscatter (T_B) and emission (σ⁰) signals of land surfaces, in which the CLAP is backboned by an air-to-soil transition model (ATS) (accounting for surface dielectric roughness) integrated with the Advanced Integral Equation Model (AIEM) for modelling soil surface scattering, and the Tor Vergata model for modelling vegetation scattering and the interaction between vegetation and soil parts. The CLAP was used to simulate both ground-based and space-borne multi-frequency microwave measurements collected at the Maqu observatory on the eastern Tibetan plateau. The ground-based systems include a scatterometer system (1-10 GHz) and an L-band microwave radiometer. The space-borne measurements are obtained from the X-band and C-band Advanced Microwave Scanning Radiometer 2 (AMSR2) radiation observations. The impact of different vegetation properties (i.e., structure, water and temperature dynamics) and soil conditions (i.e., different moisture and temperature profiles) on the microwave signals were investigated as well as by CLAP simulation for understanding factors that account for diurnal variations of the observed signals. The CLAP is expected to improve our capability for understanding and applying current and future multi-frequency space-borne microwave systems (e.g. those from ROSE-L and CIMR) for land monitoring.

As the spatial resolution of space-borne spectrometers is rapidly improving and moving towards sub-kilometer scale, three dimensional (3D) cloud effects become more prominent in the retrieval of atmospheric trace gases. Currently in the Sentinel-5P (S5P) nitrogen dioxide (NO₂) product (3.6 km x 5.6 km² resolution) the Fast Retrieval Scheme for Clouds from the Oxygen A band (FRESCO) algorithm is used to retrieve a one dimensional (1D) horizontal homogeneous Lambertian cloud layer for cloud correction. However, in reality clouds are 3D objects, they are not spatially homogeneous in brightness and they can have effects on neighboring clear-sky pixels by casting shadows on lower clouds or on the ground surface or by scattering light into the pixels.

In the S5P NO₂ retrieval algorithm the retrieved slant column density is translated to a vertical column density (VCD) by correcting for the light path using pre-calculated air-mass factors (AMF) from a radiative transfer model, using surface and cloud parameters as input. When a cloud shadow is cast over a clear-sky pixel the downward light intensity is reduced, altering the average observed light path. This lowers the sensitivity of the measurement for the lower atmospheric layers and thus changes the AMF. As this effect is not accounted for in the current 1D AMF calculation, the AMF used in the retrieval is different from the true AMF. The effect would be similar to an overestimation of the surface albedo in the AMF calculation, in turn potentially leading to an overestimation of the vertically resolved AMF at surface levels and an underestimation of the NO₂ VCD, when sufficient NO₂ is present in the lower troposphere.

This study attempts to observe such cloud shadow effects in the AMF and VCD fields in the S5P NO₂ data with focus on hot-spot areas during winter when generally more clouds are present and the cast cloud shadow surface areas are relatively high due to higher solar zenith angles. SUOMI-NPP VIIRS data can be used to identify the pixels affected by cloud shadows. Further steps are to design a correction method for these cloud shadow effects in the NO₂ algorithm.

With the rapid economic growth, China's ports and shipping industry has achieved unprecedented development, while aggravating air pollution. Due to the large number of domestic inland river vessels, limited legislation for emission control and no monitoring infrastructure, information on inland river vessel emissions is very limited. Taking the Yangtze River in the region of Nanjing as research area, a ship emission inventory was compiled based on real-time information received from Automatic Identification System (AIS) signals, ship-related basic data provided by China Classification Society (CCS) database and relevant data from field research. A method is developed to use AIS signals to calculate ship emissions per vessel. The total estimated ship emissions have been calculated in the observation area from September 2018 to August 2019 for NO_x, SO₂, PM₁₀ and PM_2.5. The calculated inland ship NOx emissions have been compared to the total NOx emissions in the same region derived from TROPOMI observations using DECSO. The DECSO algorithm (Daily Emissions Constrained by Satellite Observations) is an inversion algorithm using TROPOMI data. By comparing the ship emission inventory with DECSO, the result shows a consistent spatial distribution with riverine cities having higher NO_x pollution than non-riverine cities. With this comparison we analyzed the relative impact of ship emissions on densely populated regions around the river. The same method of deriving ship emissions will be used for the port of Rotterdam, to derive the impact of ship emission in this region.

The Arctic is experiencing rapid climate change. The increasing temperature not only reduces the sea-ice extent but will also have doubled the number of lightning flashes by the end of the century. The increase of lightning will cause more wildfires. Both contributions give rise to nitrogen oxides (NO_x) emissions.

In this study, we track and estimate three-year (2019-2021) Arctic NO₂ emissions by combing the TROPOspheric Monitoring Instrument (TROPOMI) observations, Visible Infrared Imaging Radiometer Suite (VIIRS) data, and the Vaisala’s Global Lightning Dataset (GLD360).

The consecutive overlapping orbits of TROPOMI passing over the Arctic provide unique opportunities for tracking the lightning NO₂ (LNO₂) and calculating both LNO₂ lifetime and production efficiency. Previous studies focused on the LNO₂ emissions in the tropical and mid-latitude regions and estimated the global LNO_x within the range of 2 to 8 T N yr^-1. This study can add the missing LNO₂ productions in high latitudes and highlight the potential of TROPOMI as well as future satellite missions for monitoring Arctic NO_x emissions.

During the Research on the Simulation and Mechanism of the impacts of Black Carbon on Climate and Environment atmospheric measurement campaign carried out near Nanjing, China in June 2018, a lightweight, accurate nitrogen dioxide (NO₂) sensor was attached to a quadcopter to measure vertical profiles of NO₂. Between 1 and 14 June 2018, ∼50 vertical NO₂ profiles were measured inside the planetary boundary layer up to an altitude of 900-1300 meters during 13 subsequent measurement days. Six NO₂ soundings were conducted on a daily basis at approximately 8 AM (morning), 12 & 4 PM (afternoon), 8 PM (evening) and 12 & 4 AM (night). The NO₂ measurements were calibrated using a scaling factor derived from a side-by-side inter comparison with a commercial NO₂ analyzer operated by NUIST prior to the start of the campaign. These measurements clearly demonstrate the diurnal cycle of NO₂, including the emergence of elevated concentrations close to the surface during the night and early morning and the mixing of the boundary layer from sunrise onward resulting in flat NO₂ vertical profile shapes with lower concentrations. As a result, this type of measurement could play an important role in the validation of future geostationary satellites since the diurnal cycle of NO₂ will have an impact on the accuracy of the satellite retrievals.

An accurate weather forecast requires abundant and reliable input data in the process of ingestion into Numerical Weather Prediction Model (NWPM). The sources of data are heterogeneous: ground-based radars, Global Navigation Satellite System (GNSS) products, radiosonde, in-situ weather stations, and many more.

Unluckily, in several countries, meteorological data are scarce. One example is sub-Saharan Africa, where there is, on average, only one weather station every 26,000 square kilometers. This number is eight times lower than the World Meteorological Organization (WMO) recommendation. Moreover, According to the World Bank Organization, 54% of African weather stations cannot produce reliable results, while 71% of upper-air weather stations cannot capture accurate data.

One solution can be provided by exploiting space-borne Synthetic Aperture Radar (SAR) images. In the field of Interferometric SAR (InSAR), particularly for deformation monitoring, the atmospheric delay on the radar signal has always been considered a disturbance to be mitigated. However, this additional delay is a direct consequence of different conditions in pressure, temperature, and humidity of the medium traveled by the signal (the troposphere). Thus, it might be helpful information for meteorological models.

The advantages of SAR in meteorology are manifolds. First of all, large-scale measurements spanning thousands of square kilometers are beneficial to NWPM. Moreover, SAR can provide maps at very high resolution, in the order of half a kilometer, which is currently unmatched by any other instrument.

This poster will present a technique to estimate a set of Atmospheric Phase Screens (APS) from a stack of C-Band Sentinel-1 SAR data. The algorithm exploits the so-called Phase Linking algorithm, which can optimally estimate the interferometric phase over distributed and permanent targets. The joint exploitation of the two is the critical factor enabling the generation of dense and uniform maps spanning thousands of kilometers, even over highly decorrelating areas such as tropical forests.

An error in the knowledge of the satellite’s orbits during the acquisitions will induce an aberration in the estimated APS called Orbital Phase Screen (OPS). The OPS will corrupt the low spatial frequencies of the APS, therefore it must be removed. The proposed workflow includes a cross-calibration that relies on a network of GNSS stations installed on the imaged area. The cooperation of SAR and GNSS allows for a reliable estimation of the OPS and its subsequent removal from the atmospheric product. This procedure leads to an error-free APS without the risk of filtering the low spatial frequency components of the APS itself.

The procedure is first tested in South Africa by generating delay maps as large as 210,000 square kilometers. The area shows severe decorrelation and steep topography. Still, the proposed algorithm could produce a reliable estimate of the differential atmospheric delay. The orbital correction routine is also employed, and the derived APS are validated using an external NWPM (GACOS). Some spatial statistics are derived from the delay maps, and we show that they follow theoretical models in the literature.

Finally, we propose a second validation site: Sweden. In this area, the seasons and in particular snow formation and melting determine the correlation level between SAR images. Nevertheless, our approach is still able to produce satisfactory results. Several tens of GNSS stations are available in this area, allowing for a very accurate OPS estimation and subsequent validation of the APS maps.

Synthetic aperture radar (SAR) remote sensing has gained a prominent position concerning remote sensing of forest scenarios, largely due to its all-weather observation capabilities, large coverage, and accuracy [1]–[3]. Long wavelength fully polarimetric SAR systems are particularly interesting concerning the exploration of the vertical structure of the illuminated media, like forests, ice, and snow, by virtue of the combination of penetration capabilities and sensitivity to different scattering mechanisms [4]–[10]. SAR tomography (TomoSAR) technology can obtain three-dimensional vertical backscattering power distribution and reconstruct the vertical structure of the illuminated media[5], [7]. Different from the common SAR imaging mode, with the aid of a certain of SAR acquisitions, TomoSAR can resynthesize aperture perpendicular to the slant-range direction and providing a powerful technical tool for reconstructing the three-dimensional structure. That is also a key technology for the forthcoming Earth Explorer mission BIOMASS to map global forest height and above over ground biomass (AGB), as well as underlying topography[11]. In this context, the TomoSense experiment was organized by the European Space Agency (ESA) to provide the scientific community with unprecedented data to study the features of radar scattering from temperate forests, comprising tomographic and fully polarimetric SAR surveys at P-, L-, and C-band, acquired in mono- and bistatic mode by simultaneously flying two aircraft [10], [12]. The data were acquired at the Kermeter site in the National Park Eifel in North-Rine Westphalia in Germany. The TomoSense dataset also includes Terrestrial Laser Scanning (TLS) and Airborne Lidar Scanning (ALS) products to provide a reliable reference for validation.

In this paper, we analyze TomoSense data to present an experimental study related to the use of two different approaches to structural analyses of forested areas. The first one consists in using multi-baseline data to form a tomographic reconstruction of the forest vertical profile, using well known methods from SAR tomographic processing. [5]-[7]. The second one takes advantage of the phase histogram approach, which under some circumstances allows for an estimation of forest structure using single-baseline data [13]. The two approaches are here compared concerning their capability to correctly estimate forest structure and forest height.

[1] R. F. Hanssen, Radar interferometry: data interpretation and error analysis, vol. 2. Springer Science & Business Media, 2001.

[2] M. Liao, T. Balz, F. Rocca, and D. Li, “Paradigm changes in Surface-Motion estimation from SAR: Lessons from 16 years of Sino-European cooperation in the dragon program,” IEEE Geosci. Remote Sens. Mag., vol. 8, no. 1, pp. 8–21, 2020.

[3] F. Rocca et al., “Three-and Four-Dimensional Topographic Measurement and Validation,” Remote Sens., vol. 13, no. 15, p. 2861, 2021.

[4] S. R. Cloude, Polarisation: Applications in Remote Sensing. Oxford University Press, 2009.

[5] S. Tebaldini and F. Rocca, “Multibaseline polarimetric SAR tomography of a boreal forest at P-and L-bands,” IEEE Trans. Geosci. Remote Sens., vol. 50, no. 1, pp. 232–246, 2011.

[6] S. Tebaldini, T. Nagler, H. Rott, and A. Heilig, “Imaging the internal structure of an alpine glacier via L-band airborne SAR tomography,” IEEE Trans. Geosci. Remote Sens., vol. 54, no. 12, pp. 7197–7209, 2016.

[7] S. Tebaldini, “Algebraic Synthesis of Forest Scenarios From Multibaseline PolInSAR Data,” IEEE Trans. Geosci. Remote Sens., vol. 47, no. 12, pp. 4132–4142, 2009.

[8] N. Labriere et al., “In situ reference datasets from the TropiSAR and AfriSAR campaigns in support of upcoming spaceborne biomass missions,” IEEE J. Sel. Top. Appl. Earth Obs. Remote Sens., vol. 11, no. 10, pp. 3617–3627, 2018.

[9] T. Fatoyinbo et al., “The NASA AfriSAR campaign: Airborne SAR and lidar measurements of tropical forest structure and biomass in support of current and future space missions,” Remote Sens. Environ., vol. 264, p. 112533, 2021.

[10] S. Tebaldini et al., “The Tomosense Experiment: Mono-and Bistatic Sar Tomography of Forested Areas At P-, L-, and C-Band,” in 2021 IEEE International Geoscience and Remote Sensing Symposium IGARSS, 2021, pp. 7955–7958.

[11] S. Quegan et al., “The European Space Agency BIOMASS mission: Measuring forest above-ground biomass from space,” Remote Sens. Environ., vol. 227, pp. 44–60, 2019.

[12] M. M. d’Alessandro and S. Tebaldini, “SAR Correlation Tomography for Vegetation Analysis with ESA TomoSense Data,” in 2022 IEEE Radar Conference (RadarConf22), 2022, pp. 1–6.

[13] G. H. X. Shiroma and M. Lavalle, “Digital terrain, surface, and canopy height models from InSAR backscatter-height histograms,” IEEE Trans. Geosci. Remote Sens., vol. 58, no. 6, pp. 3754–3777, 2020.

Satellite observations of greenhouse gases over the Arctic and boreal regions are important for a better understanding of the changing natural carbon cycle and monitoring anthropogenic emissions. However, high latitudes pose significant challenges to reliable space-based observations of carbon dioxide (CO2). In addition to large solar zenith angles and frequent cloud coverage, snow-covered surfaces absorb strongly in the near-infrared wavelengths that are used for retrievals of CO2. Because of the resulting low radiances of the reflection measured by the satellite in nadir geometry, the retrievals over snow may be less reliable and are typically filtered or flagged for potentially poor quality.
We present the results of a feasibility study for examining how to improve satellite-based remote sensing of CO2 over snow-covered surfaces. Our primary goal is to support the development of the upcoming Copernicus Anthropogenic CO2 Monitoring Mission (CO2M). CO2M is planned to be operational in 2025 and it will provide quantitative information of anthropogenic CO2 emissions from cities and large production facilities to help meet the carbon emission reduction targets agreed in the Paris Agreement. Our findings are also applicable to other missions that retrieve CO2 from reflected sunlight, for example OCO-2 and TanSat.
As a part of the feasibility study, extensive radiative transfer (RT) simulator development is also undertaken. RaySca, a novel RT simulator aims to be a computationally fast model of polarized radiation within planetary atmospheres in visual, near-infrared and shortwave-infrared wavelength bands. Modeling the polarization of the radiation will enable further research of atmospheric remote sensing with planetary surfaces and atmospheric aerosols with complex reflection and scattering properties. Within the ESA Dragon co-operation, we plan to continue the research and development of atmospheric RT models in the context of further improving greenhouse gas retrievals from space, specifically in polluted conditions.

The Greenhouse Gases Monitoring Instrument (GMI) on board the Chinese Gaofen-5 (GF-5) satellite provides rich observation data for the global remote sensing of atmospheric CO₂. To meet the high-precision satellite retrieval needs of atmospheric CO₂, this paper designs a coupled bidirectional reflectance distribution function (BRDF) CO₂ retrieval (CBCR) method, which describes the surface reflectance characteristics by the BRDF, corrects for atmospheric scattering based on full physics retrieval theory, and ensures the stable retrieval of multiple parameters and atmospheric CO₂ by enriching prior constraints. Theoretical analysis shows that the influence of atmospheric scattering induced by the surface bidirectional reflectance characteristics is significantly related to the aerosol optical depth (AOD), solar zenith angle (SZA) and viewing zenith angle (VZA). The validation of GMI CO₂ retrievals shows that the CBCR method significantly reduces the influence of the surface bidirectional reflectance characteristics under high AOD and high SZA conditions, decreasing the atmospheric CO₂ retrieval error from 0.58±5.64 ppm to -1.33±3.13 ppm, and the correlation with the temporal variation of actual atmospheric CO₂ increases from 34.7% to 76.8%. Our CBCR method can correct the influence of atmospheric scattering induced by the surface bidirectional reflectance characteristics on atmospheric CO₂ retrievals, and this work demonstrates that describing the surface reflectance characteristics by the BRDF is a promising idea in the field of satellite CO₂ retrievals.

After the successful launch of Aeolus, which is the first spaceborne wind lidar developed by the European Space Agency (ESA), on 22 August 2018, we deployed several ground-based coherent Doppler wind lidars (CDLs) to verify the wind observations from Aeolus. By the simultaneous wind measurements with CDLs at 17 stations over China, the Rayleigh-clear and Mie-cloudy horizontal-line-of-sight (HLOS) wind velocities from Aeolus in the atmospheric boundary layer and the lower troposphere are compared with those from CDLs. To ensure the quality of the measurement data from CDLs and Aeolus, strict quality controls are applied in this study. Overall, 52 simultaneous Mie-cloudy comparison pairs and 387 Rayleigh-clear comparison pairs from this campaign are acquired. All of the Aeolus-produced Level 2B (L2B) Mie-cloudy HLOS wind and Rayleigh-clear HLOS wind and CDL-produced HLOS wind are compared individually. For the inter-comparison result of Mie-cloudy HLOS wind and CDL-produced HLOS wind, the correlation coefficient, the standard deviation, the scaled mean absolute deviation (MAD) and the bias are 0.83, 3.15 m s⁻¹, 2.64 m s⁻¹ and −0.25 m s⁻¹, respectively, while the y=ax slope, the y=ax+b slope and the y=ax+b intercept are 0.93, 0.92 and −0.33 m s⁻¹. For the Rayleigh-clear HLOS wind, the correlation coefficient, the standard deviation, the scaled MAD and the bias are 0.62, 7.07 m s⁻¹, 5.77 m s⁻¹ and −1.15 m s⁻¹, respectively, while the y=ax slope, the y=ax+b slope and the y=ax+b intercept are 1.00, 0.96 and −1.2 m s⁻¹. It is found that the standard deviation, the scaled MAD and the bias on ascending tracks are lower than those on descending tracks. Moreover, to evaluate the accuracy of Aeolus HLOS wind measurements under different product baselines, the Aeolus L2B Mie-cloudy HLOS wind data and L2B Rayleigh-clear HLOS wind data under Baselines 07 and 08, Baselines 09 and 10, and Baseline 11 are compared against the CDL-retrieved HLOS wind data separately. From the comparison results, marked misfits between the wind data from Aeolus Baselines 07 and 08 and wind data from CDLs in the atmospheric boundary layer and the lower troposphere are found. With the continuous calibration and validation and product processor updates, the performances of Aeolus wind measurements under Baselines 09 and 10 and Baseline 11 are improved significantly. Considering the influence of turbulence and convection in the atmospheric boundary layers and the lower troposphere, higher values for the vertical velocity are common in this region. Hence, as a special note, the vertical velocity could impact the HLOS wind velocity retrieval from Aeolus.

Aeolus has the capability to measurement wind profiles and aerosol optical properties profiles synchronously, which provide the possibility of the observation of aerosol transport and advection. Based on the observation of ALADIN (Atmospheric Laser Doppler Instrument), combined with the data of CALIOP (Cloud-Aerosol Lidar with Orthogonal Polarization), ECMWF (European Centre for Medium-Range Forecasts) and HYSPLIT (Hybrid Single-Particle Lagrangian Integrated Trajectory model), a long-term large-scale Saharan dust transport event which occurred between 14 and 27 June 2020 is tracked and the possibility of calculating the dust mass advection is explored. We evaluate the performance of ALADIN and CALIOP on the observations of dust optical properties and wind fields and explore the possibility of tracking the dust events and calculating the dust mass advection with the combination of satellite and model data. The dust plumes are identified with the AIRS/Aqua Dust Score Index and with the vertical feature mask product from CALIOP. The emission, dispersion, transport and deposition of the dust event are monitored using the data from AIRS/Aqua, CALIOP and HYSPLIT. With the quasi-synchronized observations by ALADIN and CALIOP, combined with the wind field and relative humidity, the dust advection values are calculated. From this study, it is found that the dust event generated on 14 and 15 June 2020 from the Sahara in North Africa dispersed and moved westward over the Atlantic Ocean, finally being deposited in the western Atlantic Ocean, the Americas and the Caribbean Sea. During the transport and deposition processes, the dust plumes are trapped in the northeasterly trade-wind zone between latitudes of 5^∘ and 30^∘ N and altitudes of 0 and 6 km. Aeolus provided the observations of the dynamics of this dust transport event in the Saharan Air Layer (SAL). From the measurement results on 19 June 2020, the dust plumes are captured quasi-simultaneously over the emission region (Western Sahara), the transport region (middle Atlantic) and the deposition region (western Atlantic) individually, which indicates that the dust plume area over the Atlantic on the morning of this day is quite enormous and that this dust transport event is massive and extensive. The quasi-synchronization observation results of 15, 16, 19, 24 and 27 June by ALADIN and CALIOP during the entire transport process show good agreement with the Dust Score Index data and the HYSPLIT trajectories, which indicates that the transport process of the same dust event is tracked by ALADIN and CALIOP, verifies that the dust transport spent around 2 weeks from the emission to the deposition and achieved the respective observations of this dust event's emission phase, development phase, transport phase, descent phase and deposition phase. Finally, the advection values for different dust parts and heights on 19 June and on the entire transport routine during transportation are computed. On 19 June, the mean dust advection values are about 1.91±1.21 mg m⁻² s⁻¹ over the emission region, 1.38±1.28 mg m⁻² s⁻¹ over the transport region and 0.75±0.68 mg m⁻² s⁻¹ over the deposition region. In the whole lifetime of the dust event, the mean dust advection values were about 1.51±1.03 mg m⁻² s⁻¹ on 15 June 2020, 2.19±1.72 mg m⁻² s⁻¹ on 16 June 2020, 1.38±1.28 mg m⁻² s⁻¹ on 19 June 2020, 1.60±1.08 mg m⁻² s⁻¹ on 24 June 2020 and 1.03±0.60 mg m⁻² s⁻¹ on 27 June 2020. During the dust development stage, the mean advection values gradually increased and reached their maximum on 16 June with the enhancement of the dust event. Then, the mean advection values decreased during the transport and the deposition of the dust over the Atlantic Ocean, the Americas and the Caribbean Sea.

Satellite altimetry provides the means for global monitoring of sea level, sea ice and
inland waters with accuracy of mm/yr. Absolute calibration of satellite altimeters by
external, permanent and independent ground facilities is necessary to safeguard
accuracy and reliability of those satellite observations for climate change monitoring.

The main objective of the Dragon V project (ID 59198) is to standardize procedures
for calibration/validation of European and Chinese satellite altimeters. Such
procedures are to follow the guidelines prescribed by the strategy for Fiducial
Reference Measurements for Altimetry, developed by the European Space Agency.

One of the fundamental quantities that needs to be calibrated in satellite altimetry is
the backscatter coefficient (sigma-naught); a parameter related to wind observations
at sea and which constitutes an indispensable parameter in climate change models.
At the moment, there is no European or Chinese Cal/Val facility exclusively dedicated
to sigma-naught calibration.

This work presents the first steps towards design, implementation and validation of
corner reflectors for absolute and direct sigma-naught calibration of satellite
altimeters. First, the pros and cons of corner reflectors are given and compared
against active transponders. Then, the geometrical shape of the corner reflector is
examined as it controls performance characteristics, such as radar cross section with
respect to radar elevation and azimuth, maximum gain achieved, side lobe
attenuation, its durability to outdoor conditions, etc. Finally, the need for designing a
corner reflector capable to support sigma-naught calibration for multi-mission and
multi-frequency satellite altimeters is presented.

Understanding the temporal dynamics of vegetation biophysical and structural characteristics such as the fraction of absorbed photosynthetically active radiation (FAPAR) and plant area index (PAI) is valuable for agricultural and forestry applications. The installation of automated in situ measurement networks can better characterise these parameters over a complete growing season in comparison to ad-hoc field campaigns.

For Component 2 of the Copernicus Ground Based Observations for Validation (GBOV) service, a wireless quantum sensor network have been installed at Mediterranean vineyard vegetation (Valencia Anchor Station, Spain). This network will supplement manual field data collections (DHP and LI-COR LAI-2200C measurements), which have been periodically collected throughout the growing season, allowing the performance of the automated systems to be assessed against established and accepted in situ measurement techniques.

The instrument deployment, data processing and filtering methods and performance against manual measurements are presented. An evaluation of the temporal dynamics of the biophysical variables from the automated systems and the implications of a temporally dense, but spatially limited, dataset for the validation of moderate spatial resolution satellite products are also discussed.

Sea ice drift (SID) is the key to understanding sea ice dynamics and critical for navigation safety. This study focused on a comprehensive analysis of SID retrieval based on spaceborne synthetic aperture radar (SAR) data in the Arctic. A state-of-the-art method combining feature tracking and pattern matching techniques was applied to sequential Sentinel-1 (S1) SAR data to derive SID from the central Arctic to the Fram Strait in different seasons. The SAR retrievals were compared with drifting buoy data for validation. For temporal intervals of S1 data of 16 to 24 hours, 13,586 collocations were collected in the winter and spring seasons, yielding a 0.00 cm/s bias for the drift velocity magnitude and 0.32 degrees for direction with the corresponding root mean square error (RMSE) of 0.50 cm/s and 4.96 degrees. The applied method outperforms the maximum cross-correlation method for rapidly drifting sea ice. Using temporal intervals of S1 data of less than 16 hours, we retrieved SID in the summer and autumn seasons. 644 collocations yield a comparison with a bias of 0.52 cm/s and 4.62 degrees for the drift magnitude and direction, respectively. The corresponding RMSE values are 1.85 cm/s and 20.73 degrees. The comparisons present better results than the operational SAR-based SID product and consistent seasonal trends in drift velocity with the coarse-resolution product. We also analyzed the variations in SAR retrievals and further estimated appropriate temporal intervals, making it feasible to conduct long-term SID retrievals based on spaceborne SAR data at high spatial resolution in the Arctic.

Abstract

The rapid increase of offshore wind installations in the south-China sea near the coast triggers a new demand for studying the effects of horizontal wind speed gradients and the wind power variation within the coastal zone. The advent of Synthetic Aperture Radar (SAR) data offers an opportunity to map wind speed gradients and wind farm wakes with high spatial resolution. We have retrieved wind maps at 10 m above mean sea level (m.s.l.) from Sentinel-1 SAR and Envisat Advanced SAR observations. Generally, the speed of the prevailing south-easterly winds and wind power declined about 8% and 22%, respectively. Although the southern offshore wind farms (OWFs) were not in operation before December 2019, the wind velocity deficit at the upstream side of northern OWFs were between 8-12 %. After the southern OWFs became online, the region between OWFs is subjected to wind wakes and coastal upwelling effects. The coastal upwelling phenomena speeds up the wind at the downstream sides of OWFs that reduce the wind wakes up to 8-10%. The wind wakes extended 20 km beyond the southern OWFs.

1. Introduction

A Synthetic Aperture Radar (SAR) is a side-looking satellite sensor, which can be used to visualize the fine spatial details of the wind flow close to the coast. SAR observations have been utilized for a wide range of applications ranging from wind resource assessments [2], identifying offshore wind farm wakes and coastal wind speed gradients [3]. Several studies have validated the wind speeds retrieved from SAR observations with respect to other datasets and typically found the root mean square error to be less than 2 m/s. Ahsbahs et al. show good agreement through comparison of 15 Sentinel-1A wind maps against light detection and ranging (LiDAR) measurements at the west coast of Denmark [4].

Wind wakes are defined as areas of reduced wind speed at the downstream side of the offshore wind farms (OWFs) because of energy extraction by the wind turbines. Wakes can extend several tens of kilometers, therefore, the wakes can interact with adjacent OWFs and have severe consequences for the power production, as the wind power is proportional to the cube of the wind speed.

The wind speed within the coastal zone is also affected by (i) the surface discontinuity at the coastline, (ii) the influence of onshore topography, and (iii) thermal gradients [5]. The magnitude of wind velocity deficits at the downstream side of OWFs is thus a combination of wind wakes and the effects of horizontal coastal wind speed gradients. Hasager et al. concluded that the winds in the coastal zones have larger spatial gradients than further offshore and many other wind phenomena occur in coastal zones [6]. Owda et al. have studied the effects of coastal gradients for many OWFs in northern European seas and found strong gradients inversely proportional with the distance to shore. They decomposed the wind gradient effects from wind wakes based on SAR observations before commissioning of the OWFs [3].

2.1. Data

2.1.1. Sentinel 1A/B & Envisat

Sentinel 1A/B is a constellation of two different satellites, Sentinel-1A (2014-present) and Sentinel-1B (2016-2021), sharing the same orbital plane at the mean altitude 693 km. Envisat (2002-2012) carried an Advanced Synthetic Aperture Radar (ASAR) instrument at the altitude 800 km. The satellite data were acquired using C-band SAR sensors operating at 5.405 GHz. The satellites are in a near-polar, sun-synchronous orbit with a 6 and 35 day repeat cycle for the Sentinel-1 constellation and Envisat, respectively. In this study, VV co-polarized images with extra wide (EW) for Sentinel-1 and wide swath mode for Envisat were used.

2.2. Study area

Our study area is in southeast China and at 120 to 120.8 longitude and 34 to 34.6 latitude. The area has four OWFs (Figure 1). Table I presents the characteristics of each OWF. Table II presents available SAR scenes based on commissioning date of first operated southern OWFs (Datang Binhai). We refer southern OWFs term to Datang Binhat and Spic Binhat South H3 and northern OWFs term to SPIC Binhai North H1 and H2.

2.3. Methodology

2.3.1. SAR wind retrieval

The SAR radar observables relate to the local near-surface wind speed using an empirical equation called a geophysical model function (GMF)[7]. .

2.3.2 Mean wind speed, deficit and power variation calculation

A grid of 60×50 km is overlaid over the entire study area and used to retrieve SAR wind measurements with the regular grid spacing 1.5 km between each rectangular bin inside the grid. Figure 1 illustrates the boundary of the used grid. Based on the 86 available SAR scenes, the wind rose for a point close to the coast shows that the prevailing wind direction is from the southeast. In this study, we have taken in our analysis only the scenes with wind directions between 90 and 180 degrees “southeast”. The mean wind speed (U) is computed for our area of interest based on the selected scenes of each period in Table II. Furthermore, the relative wind speed difference to the mean upstream wind (∆U) is calculated for the entire grid. The wind power density (P) is estimated using a simple power equation. Equations 1, 2 refer to ∆U (%) and P (Watt/m²), respectively.

4.0 Results

the spatial variation of the wind speed is strong near the curved coast. It also shows the effects of natural wind speed gradients on the wind power potential. The high spatial resolution of SAR can provide valuable information about the wind speed variation from far offshore to the coastal zone. The results have shown strong coastal gradients as the winds approach the coastline. Along a transect line at the southern region of the area investigated, it shows the speed is reduced about 8 % over the 32 km distance. In terms of wind power estimation, the power reduction along the same transect line is about 22%. The region between the southern and northern OWFs within our area of interest is subject to different wind velocity deficits before and after commissioning southern OWFs

The German Wadden Sea is an area of great economic and ecological importance. Apart from being the largest German National Park, tourism and fishery benefit from the specific characteristics of this intertidal region. One of those specific characteristics are strong morphodynamics, which is why the bathymetry of the region undergoes frequent changes, including an eastward movement of islands. Major shipping routes to harbors such as Hamburg and Bremerhaven cross the German Wadden Sea; hence, a reliable monitoring of sand banks and shallow waters does not only give clues about the state of the region, but is also crucial for marine security. Since the German Wadden Sea is frequently overflown by satellites, radar images acquired at different water levels can be used for its surveillance. Algorithms for an automatic extrapolation of waterlines from these radar data exist, but they strongly depend on image quality, and they require careful manual optimization.

Neural networks that use images as input and also produce images as output are called U-Nets, because of their network architecture. A U-Net consists of encoding and decoding steps, shrinking the image size in the former, and reconstructing it in the latter. To maintain informational value, the number of channels rises with decreasing image size. Setting up a neural network to detect waterlines showed to be a promising approach. While no accurate predictions could be achieved, the use of additionally generated data could compensate for the small dataset of radar images.

The U-Net in this project consists of three encoding and two decoding steps and uses 128x128 images as input and output. Each step consists of a large (15x15) and a small (3x3) Convolutional Layer with Dropout, Batch Normalization, and a nonlinear Rectified Linear Unit (ReLU) function. During each encoding step, the number of channels is increased. After each of the first two encoding steps, a Maxpool Layer halves the image dimension, resulting in a virtual image size of 32x32 with 64 channels. Before each decoding step, an Upward Convolutional Layer reverts the downsizing of the Maxpool Layer, and the informational value from the comparable encoding step is concatenated to the decoding channels. With increasing virtual image size, the channels are reduced and finally transformed into an output mask consisting of only one channel.

Although further training with radar images improved the result both qualitatively and quantitatively, major problems detain an improvement with the given model and datasets: the size of the dataset of radar images is not sufficient to predict waterlines in a given Wadden Sea area at acceptable accuracy, and the improvement through augmentation is limited. Furthermore, the dependence of radar contrast on weather conditions may hinder the use of one single big image.

Coastal regions represent areas where a large portion of the world’s population lives. Hence, many people rely to some extent on coastal and marine ecosystems and resources for food, building materials, building sites, and agricultural and recreational areas, while utilizing coastal areas as a dumping ground for sewage, garbage, and toxic wastes. The human-induced phenomena, added to extreme natural events, lead to an ever-increasing pressure on such regions. As a result, harsh coastal environments can be formed where wetlands, mudflats, mangroves, marshes etc. are present altogether. Monitoring the impact of such phenomena on land-sea dynamic processes and supporting an effective coastal area management is a major scientific challenge, with an increasing importance due to growing urbanization, industrialization, and transportation. In this context, space-borne synthetic aperture radars (SARs) sensors gain great importance since they allow obtaining high-resolution imagery collected during almost all-weather conditions and captured during day and night. Moreover, the use of SAR multi-polarimetric imaging modes allows obtaining improved monitoring accuracy with respect to the optical and single-polarization cases.

In this study, a multi-frequency and multi-polarimetric approach is proposed to study the properties of harsh coastal environments. The multi-polarimetric approach allows studying the different scattering mechanism of the region of interest, while using different sensors at different frequencies allows discriminating among different aspects of the same scenario. The study area is the Solway Firth coastal region, located along the western coastal boundary between Scotland and England, that represents a very harsh coastal environment composed of marshes, mudflats, agricultural crops, hill farming and shallow water rich in sediments. Moreover, it is severely affected by erosion processes induced by storm surges during the rainy season. C- and X-band fine-resolution quad-polarimetric SAR satellite measurements, collected over the Solway Firth area by Radarsat-2 and Cosmo-SkyMed Second Generation (CSG) missions, respectively, are used. The purpose of the study is to analyze the considered scattering scenarios in terms of a two-fold analysis: an intensity-based multi-frequency backscattering analysis and a polarimetric analysis. The latter is performed using two properties, the polarization signature and the co-polarised phase difference.

The properties of three different scenarios, i.e., grasslands, mudflats and sea water are analyzed using the two different sensors and the two polarimetric properties. Preliminary results show that the proposed analysis allows improving the understanding of the harsh coastal environment scattering processes. Hence, this kind of techniques may support the development of advanced and robust scattering-based algorithms for coastal management purposes, in order to monitor and mitigate human- and nature-induced processes.

The Synthetic Aperture Radar (SAR), owing to its day-night and almost all-weather imaging capabilities together with its fine spatial resolution, is a valuable tool to observe the oceans and monitoring oil pollutions. Mineral oil films appear in the SAR image plane as spots darker than the sea surface background because of their suppression of capillary waves. However, mono-molecular biogenic surfactants, which are produced for instance by plankton or fishes, give rise to radar signatures similar to that of mineral oil films. Hence, to fully understand the link between the actual oil slick and the dark patch observed in the SAR image plane, it is necessary to analyze the underlying scattering process from theoretical aspects and distinguish mineral oil spills from such false alarms in SAR imagery in a robust way.

Mono-molecular oil films call for a resonance-type damping of short gravity and capillary waves that is well-described by the so-called “Marangoni damping”, which is used - together with a reduced input wind modeled through a reduced friction velocity to account for the effects on long wave part – to reduce the sea surface spectrum. In this study, sea surface scattering with and without surfactants is predicted using the two-scale boundary perturbation model (BPM) and the advanced integral equation model (AIEM) augmented with two different damping models, i.e., the Marangoni one and the model of local balance (MLB). Numerical predictions are showcased for both oil and biogenic slicks. The two scattering models result in significantly different predictions according to the slick type and the considered damping model.

Numerical predictions are contrasted with actual SAR measurements collected at X-band by the German TerraSAR-X sensor over oil and biogenic slicks of known origin. Experimental results show that: 1) When dealing with slick-free sea surface, the two-scale BPM and AIEM result in predicted NRCS values at both polarizations that exhibit non-negligible differences up to an incidence angle of about 40°. Those differences are negligible (less than 1 dB) at larger incidence angles. The NRCS predicted by BPM results in the best agreement with the measured one at low incidence angles, while AIEM results in a PR that best fits actual measurements; 2) the two-scale BPM augmented with the Marangoni damping model is more suitable for predicting the NRCS and the damping ratio of biogenic slicks; 3) the AIEM combined with the damping MLB results in a better agreement with SAR measurements collected over oil slicks.

This study is supported by the ESA-NRSCC Dragon-5 cooperation project “Monitoring harsh coastal environments and ocean surveillance using radar remote sensing sensors” (ID 57979).

Terra Nova Bay (TNB) is situated in the western Ross Sea, between the Drygalski ice tongue (75◦24’S, 163◦30’E) and Cape Washington (74◦39’S, 165◦25’E), Antarctica. The area is confined to a narrow strip of coastal waters to the south of Mario Zucchelli Station (MZS) (Italy), extending approximately 9.4 km in length and generally within 1.5 – 7 km of the shore, comprising an area of 29.4 km².

The site typically remains ice-free in summer, which is rare for coastal areas in the Ross Sea region, making it an ideal and accessible site for research into the near-shore of the region. The coastline of TNB is characterized predominantly by rocky cliffs, with large boulders forming occasional beaches. During the winter season, this coastline is the area where land ice meets the sea ice. Sometimes it is very difficult to determine a precise line that can be called “coastline,” due to the dynamic nature of the sea and the sea ice. In addition, these areas are often strongly affected by extreme weather and sea conditions and by the continuous fusion processes, therefore observing the coast from space in a continuous and effective way is of fundamental importance to support the planning and management of this coastal zone. In fact, polar coastal zone monitoring is an important task in sustainable development and environmental protection and the analysis of the time-variability of the coastline is a task of primarily importance.

This study focused on the analysis of the time variability of the morphology of the coastline of Terra Nova Bay (TNB), Antarctica, using a time series of C-band Sentinel-1 DP (HH-HV) Interferometric Wide (IW) swath mode single-look complex (SLC) SAR scenes collected from 2016 to 2022. The methodology to extract the coastline is based on a global threshold constant false alarm rate (CFAR) approach that is tested on both ice-free and ice-covered sea conditions. Once the optimal setting if found, the methodology is applied to a larger time series of S1 scenes to monitor the time-variability of selected features belonging to the extracted coastline.

The surface wave investigation and monitor- ing (SWIM) aboard the China-France Oceanography Satellite (CFOSAT), a pioneer conically scanning wave spectrometer, was successfully launched on October 29, 2018. Its innovative configuration composed of one nadir and five rotating near-nadir beams is designed to simultaneously observe the directional wave spectrum at a global scale. In this study, we systematically implement the spectral analysis of the radar backscattering with the periodogram technique to obtain the fluctuation spectrum for each azimuth direction. The 2-D fluctuation spectrum of the three spectral beams (θ = 6◦ , 8◦ , and 10◦ ) combines all the azimuth directions within one entire rotation of 360◦. The case study demonstrates that the wave features (peak wavelength and direction) are roughly consistent between the estimated fluctuation spectrum and the collocated WaveWatch III wave slope spectrum. A marked up-to-downwave asymmetry of the fluctuation spectrum with larger spectral level in the upwave direction for all the three spectral beams is observed. A ratio is defined between the fluctuation spectrum within the [0◦ , 180◦ ] sector relative to the [180◦ , 360◦ ] sector. Statistics display that this ratio is greater than 1 when it denotes the up-to-downwave ratio and smaller than 1 for the down-to-upwave ratio. This observed spectrum asymmetry is linked to the asymmetric modulation from upwind to downwind. In addition, we employ such finding to help remove the 180◦ wave direction ambiguity from a practical point of view. Preliminary results of the direction ambiguity removal display a bias of 41.3◦, 40.6◦, and 36.7◦ for the beams. The 10◦ beam shows slightly better performance compared to the other two beams in terms of bias and standard deviation. This shall lay a strong basis for the operational implementation of such algorithm to resolve the direction ambiguity.

The Surface Waves Investigation and Monitoring (SWIM) instrument onboard the China France Oceanography Satellite (CFOSAT) can retrieve directional wave spectra with a wavelength range of 70~500 m. This study aims to validate the partitioned integrated wave parameters (PIWPs) from SWIM, including partitioned significant wave height (PSWH), peak wave period (PPWP), and peak wave direction (PPWD), against those from National Data Buoy Center (NDBC) buoys. With quasi-simultaneous spectra from two NDBC buoys 13 km away from each other near Hawaii, the methods of comparing PIWPs from two sets of spectra were discussed first. After cross-assigning partitions according to the spectral distance, it is found that wrong cross-assignments lead to many outliers strongly impacting the estimate of error metrics. Three methods, namely comparing only the best-matched partition, changing the threshold of spectral distance during cross-assignment, and maximum likelihood estimation of root-mean-square error (RMSE) of PIWPs, were used to reduce the impact of potential wrong cross-assignments. Using these methods, the SWIM PIWPs were validated against NDBC buoys. The results show that SWIM performs well at finding the spectral peaks of different partitions with the RMSE of PPWPs and PPWDs of 0.9 s and 20°, respectively, which can be a useful complement for other wave observations. However, the accuracy of PSWH from SWIM is not that good at this stage, probably because the high noise level in the spectra impacts the result of the partitioning algorithm. Further improvement is needed to obtain better PSWH information.

Characterizing the uncertainties in buoy ocean wave records is critical not only for understanding the limitations of in situ wave measurements, but also for interpreting the implied accuracies of the remotely sensed products in which these buoy data are used as validation references. This letter preliminarily assesses the error of long-period swell heights (Hss) representing specific directional wave partition energy observed from deep-water buoys moored in the northeast Pacific. We propose a buoy Hss error estimation method by combining dual and triple collocation using data derived from buoys, two kinds of space-borne radars and numerical simulations. Compared to traditional methods, the proposed approach can reveal “absolute” errors (with respect to the underlying truth) from buoy Hss, accepting and then confirming that swell heights from buoy, satellite and model are all uncertain. This study simultaneously employs ocean swell products derived from synthetic/real aperture radars (Sentinel-1A/B and CFOSAT/SWIM) and WaveWatch III ocean wave model hindcasts to diagnose the accuracy of the Hss values observed by buoys of National Data Buoy Center (NDBC) and Coastal Data Information Program (CDIP) during the period from July 2019 to October 2021. We quantify that the NDBC’s 3-m heave-pitch-roll buoy (CDIP’s Waverider buoy) recorded Hss have root-mean-square error of 0.17 m (0.12 m), or have about 10.65% (7.06%) uncertainty relative to the mean Hss value (approximately 1.6 m). Our findings imply that the reference value uncertainties should be taken into account when understanding direct satellite Hss validation against buoy in situ.

Satellite altimetry has been developed for several decades and obtained abundant information on Marine environment change. At present, the SAR (Delay-Doppler) altimeter has become one of the main loads of satellite altimeters, such as the Sentinel-3 series and Cryosat-2 satellites. The measurement accuracy of SAR altimeter in the open sea is relatively high. However, due to the influence of land, island, and other factors, there are still some problems in the nearshore area, which limits the application of satellite altimeters in this area. Based on the waveform theory of SAR altimeter and the systematic analysis of different types of waveform retracking algorithms, this paper proposes a waveform retracking data processing strategy based on neural network waveform classification and discusses the waveform retracking algorithms suitable for different sea surface types. Based on the Sentinel-3 altimetry data, the accuracy of the algorithm and the change of altimetry sea level is analyzed by using the sea surface height data of the tide station and buoy.

The HY-2B satellite was launched in October 2018 as China second marine dynamic environmental satellite. It is equipped with a traditional dual-frequency altimeter, which can accurately observe marine dynamic environmental elements including sea surface height, wind field and significant wave height. In coastal areas, the precision of range corrections such as sea state bias, ionosphere delaying correction and tropospheric delaying correction provided by SGDR data have declined due to the influence of coastal "pollution" on the altimetry system. Aiming at this problem, this paper carries out a study about the improvement of the coastal altimetry range corrections for HY-2B altimeter. Based on the 20hz sea surface observation, the high-frequency sea state bias model is constructed, the deviation of the wet troposphere correction is modified by a composite method, and the error of the ionosphere correction and the dry troposphere correction are reduced by high-frequently filtering. Finally, the effectiveness of the new range correction is validated by comparing and analyzing the SSH before and after improving range corrections.

In July of 2020, the orbit of CryoSat-2 was modified to allow for repeated overlaps with ICESat-2. Following a year of coincident orbits with parallel observations by radar from CryoSat-2, and lidar from ICESat-2 allows for direct comparison between these systems. Using 136 orbit segments from the northern hemisphere, constrained to the Pacific and Atlantic oceans as well as the Bering Sea, we compare the significant wave height (SWH) observations. By utilizing the coincident orbits, we can compare observations between altimeters of the same sea state within a constrained time lag (less than four hours), allowing for comparison within longer stretches of the orbits. This is crucial to assess the level of agreement between observations, owing to the constantly changing ocean surface. With the comparison between the systems, as well as discussing the inherent benefit of each system, we can assess the possibilities of alternate methods for ocean surveying. From the available data, SWH up to 10 m has been used for the analysis, enabling this comparison to be done at various sea states.
We have used three methods with the ICESat-2 data in the comparison, with the first being the standard ocean data output (ATL12) as produced by the ICESat-2 team. This is compared with a method where modeling of the individual surface waves is used as an assessment of the SWH. It has been shown before to be possible to use the geolocated photons from ICESat-2 to assess these waves, which is then beneficial to compare with the radar altimeter of CryoSat-2. Functioning as a baseline for the wave approach, we are using the standard deviation of the ocean surface, the same method as in ATL12, however with the same filtering as for the wave-based model.
From this, we have described the differences between the altimeters and show a high correlation, with correlations between the models and CryoSat-2 SWH of 0.97 for ATL12, 0.95 for the observed waves model, and 0.97 for the standard deviation model. There has been found a mean deviation relative to the observed SWH for each model, deviating more at SWH larger than 2.5 m, but generally between -10 cm and 16 cm for SWH smaller than 2.5 m for all models. Compared with CryoSat-2 there was found an increasing deviation along with increasing SWH, along with a larger variance. In general, the SWH observed from ICESat-2 is found to agree with observations from CryoSat-2, within limitations due to cloud coverage. Observing the individual surface waves from ICESat-2 is therefore seen to provide additional observed properties of the sea state for global observations.

Sea ice plays an important role in global climate change, shipping, navigation and the extraction of natural resources, and influences the detection of other ocean phenomena; for example, sea wave retrieval requires the removal of sea ice ‘pollution’. The Surface Wave Investigation and Monitoring instrument (SWIM) on the China-France Oceanography Satellite (CFOSAT) is a new type of sensor with a small incidence angle detection mode that is different from traditional remote sensors. Sea ice monitoring at small incidence angles has rarely been studied. Therefore, this research focuses on sea ice monitoring in the Arctic based on SWIM data from October 2019 to April 2021. Sea ice type is the key parameter of Arctic Sea ice monitoring. Six waveform features are extracted from the SWIM echo data at small incidence angles, then sea ice classification based on multi-feature combinations is carried out using the chosen KNN classifier. Thus, the optimal classifier-feature assembly at each incidence angle is developed, and the highest overall accuracy is up to 81% at 2°. Sea ice recognition is useful for extracting sea ice density, sea ice extent, sea ice edge and other parameters, and is also applied for the detection of other ocean phenomena. Therefore, based on the above work, a method to distinguish between sea ice and sea water is established. Eleven features are first extracted, and applied for the analysis of the waveform characteristics using the cumulative probability distribution and mutual information measurement. The optimal classifier is the KNN method with Euclidean distance and k equal to 11. Feature combinations are also used to separate sea ice and sea water based on the KNN method to select the optimal combination. Thus, the optimal classifier-feature assembly at each small incidence angle is established, and the highest overall accuracy reaches 97.1%. Moreover, the application of the optimal classifier–feature assemblies is studied. The overall accuracies of sea ice recognition using the optimal classier–feature assemblies in three stages of sea ice development are higher than 90 percent, and the highest reaches 99.9%. Sea ice extents and edges can also be extracted using this method. The consistency of sea ice extents with NSIDC is higher than 94% (the highest is 98.2%), and the accuracies of daily sea ice edge products are higher than NSIDC. Our work not only confirms the ability of sea ice classification and recognition based on the new SWIM data with high accuracies, but also studies the application of SWIM data in sea ice services. SWIM data can be used as a new data source for operational sea ice monitoring.

Icebergs impose a risk on ship navigation and offshore structures. Images from satellite Synthetic Aperture Radar (SAR) which can be acquired independent of light conditions and cloud coverage, are widely used for monitoring icebergs. Automated detection of icebergs in SAR images is normally carried out by utilizing constant false alarm rate (CFAR) filters, which compare the intensity of individual pixels or cluster of pixels against the intensities of the adjacent pixels (i.e. the background) and adaptively set a threshold based on several assumptions regarding the statistical distribution of the background clutter.

Five different CFAR filters were tested for iceberg detection in open water. The algorithms were applied to both a C- and an L-band SAR image acquired over the Labrador Sea in July 2019. The SAR images were dual-polarized (HH and HV) as used at the operational ice and iceberg charting centers and were resampled to the same pixel spacing. The performance of the CFAR filters was assessed by comparing the automated detections to 230 icebergs manually identified in a coinciding optical Sentinel-2 image. The F-score was used for quantifying the success rate as a trade-off between false and missed detections.

Comparing the frequency bands, the L-band data reveal a slightly worse overall performance than C-band. At C-band, the highest F-score was obtained for a CFAR detector based on the gamma distribution, whereas for L-band data, a CFAR detection based on the log-normal distribution resulted in the highest F-score. This indicates that for a given sea state, the sea clutter distributions differ between C- and L-band. The magnitude of the F-score shows high variability, dependent on the PFA and type of algorithm. This is demonstrated by our poster, considering also the runtime of the single algorithms. Potential problems of iceberg detections in SAR images acquired at C- and L-band will be mentioned.

Snow albedo is a significant factor in the land surface energy balance and the water cycle. It is usually parameterized as functions of snow-related variables in land surface models (LSMs). However, the default snow albedo scheme in the widely used Noah LSM shows evident drawbacks in land-atmosphere interactions simulations during snow processes on the complex topographic Tibetan Plateau (TP). We firstly demonstrate that the improved Noah snow albedo scheme performs well in relation to near-surface meteorological elements estimates after including MODIS albedo products and explicit considering snow depth (SD) as an additional factor. Then, we comprehensively evaluate the performance of the improved snow albedo scheme implemented in the coupled WRF/Noah in simulating additional eight snow events on the TP. The modeling results are compared with WRF run with the default Noah scheme and in-situ observations. The improved snow albedo scheme significantly outperforms the default Noah scheme in relation to air temperature, albedo and sensible heat flux (SH) estimates, by alleviating cold bias estimates, albedo overestimates and SH underestimates, respectively. This in turn contributes to more accurate reproductions of snow event evolution. The averaged RMSE relative reductions (and relative increase in correlation coefficients) for air temperature, albedo, SH and SD reach 27% (5%), 32% (69%), 13% (17%) and 21% (108%) respectively. These results demonstrate the strong potential of our improved snow albedo parameterization scheme for snow event simulations on the TP. Our study provides a theoretical reference for researchers committed to further improving the snow albedo parameterization scheme.

Glacier albedo determines the net shortwave radiation and therefore affect glacier energy and mass balance. Glaciers in the Western Nyainqentanglha Mountains (WNM) are one of the most important fresh water resources for the people around Lhasa, however, the relationship between albedo and mass balance is still unclear due to absence of high spatial resolution and accurate glacier products. In this study, we firstly optimized the procedure to retrieve glacier mass balance for Chinese high resolution stereo images (ZY-3/TLA) uniquely provided by the Dragon Program, and then analyzed its change in the WNM during 2000-2017. Secondly, we improved albedo retrieval method for Sentinel 2/MSI, Landsat 8/OLI and MODIS data, and then analyzed long term variability of albedo in the WNM. Finally, explored their relationships according to these two results.

The results showed that: 1) ZY-3 TLA data can generate 5 m spatial resolution DEM and is very promising to extract high accurate mass balance estimates. 2) The glaciers in the WNM experienced accelerated mass loss in 2000-2017, and the thinning rates in the ablation regions were apparently larger than in the accumulation regions. 3) The improved glacier albedo retrievals were in good agreement with field observations and gave a better accuracy in terms of spatial and temporal coverage. 4) Glacier albedo experienced large inter-annual fluctuations and a significant decreasing trend in 2001-2021. 5) Good correlation between albedo and mass balance indicates that decreasing albedo is a key driver of mass loss in this region.

Global surface soil moisture (SSM) products with higher accuracy are needed urgently for agricultural water resource management, environment, and climate analysis applications such as the global climate change monitoring, drought monitoring and vegetation growth monitoring. Temporal and spatial sampling by the space-borne instruments designed to retrieve SSM is limited by the orbit of the satellite and by the operation of the sensor system. This constraints the generation of global, daily SSM data products. To obtain a global SSM product with higher coverage and accuracy this study integrated five SSM products with good performance in global coverage and accuracy, i.e., the SSM retrievals from the data acquired by the Soil Moisture and Ocean Salinity (SMOS), Advanced Scatterometer (ASCAT), FengYun 3-B (FY3-B), ESA-CCI and Soil Moisture Active and Passive mission (SMAP). These five SSM data products were retrieved using different algorithms, but they were combined to produce a (2011~2018) time-series of daily global SSM by applying the TCA and Linear Weight Fusion (LWF). First, we merged the global SMOS, FY3-B, and ASCAT SSM products from 2011 till 2018 using the TCA-based LWF algorithm. Then, the first merged SSM product, ESA-CCI, and SMAP SSM products were 2^nd merged using the same fusion method but for the period 2015~2018. The Global Daily-scale Soil Moisture Fusion Dataset (GDSMFD) with 25km spatial resolution (2011~2018) was produced. Finally, we used five metrics to evaluate and compare the SMOS, FY3-B, ASCAT, ESA-CCI, SMAP, the GDSMFD SSM products against in-situ soil moisture measurements at the sites of ten observation networks, which belong to the International Soil Moisture Network (ISMN). Results indicated that the GDSMFD was consistent with in-situ soil moisture measurements, the minimum of root mean square error values of GDSMFD was only 0.036 cm3/cm3. Moreover, the GDSMFD had a good global coverage with mean Global Coverage Fraction (GCF) of 0.672 and the maximum GCF of 0.837. GDSMFD performed well in accuracy and global coverage fraction, making it valuable in applications to the global climate change monitoring, drought monitoring and hydrological monitoring. GDSMFD product was released at the National Tibetan Plateau Data Center (DOI: 10.11888/Terre.tpdc.271935)

Glacier and snow are sensitive indicators of regional climate variability. In the early 21st century, glaciers in the West Kunlun and Pamir regions showed stable or even slightly positive mass budgets, and this is anomalous in a worldwide context of glacier recession. The analysis was focused on the high mountain ranges surrounding the Tarim Basin, where the spatial distribution of snow cover is quite variable. The study was based on multi-temporal remote sensing data to monitor glacier and snow cover area in the Tarim Basin high mountain area. MODIS data was used to calculate the Normalized Difference Snow Index (NDSI) and a threshold was applied to extract the Tarim Basin glacier area and seasonal snow cover. Between 2000 and 2020, the total area of the Tarim Basin Glacier declined at a rate of 0.94% per year. Because of differences in atmospheric circulation and environmental conditions, changes in the glacier area of the Tarim Basin show differences across five sub-regions. The rate of glacier area loss was fastest (2.98% per year) in the East Tian Shan, while Pamir Mountains and East Kunlun had the slowest rates, i.e. 0.50% per year and 0.81% per year, respectively.

Little is known about how rain, snow- and ice melt vary sub-seasonally and along the altitudinal gradient in high-elevation watersheds. Such basins play a key role in sustaining water supply to mountain communities and downstream ecosystems in High Mountain Asia.

We simulate mountain hydrology using a land surface model that constrains energy and mass fluxes using physical representations of both cryospheric and biospheric processes at 100 m spatial resolution. We study the upper Langtang catchment (~4000-7000 m a.s.l.) in the Nepalese Himalayas, and simulate a detailed water balance for one hydrological year (2018/2019), revealing the relative importance of precipitation, snow, ice, soil moisture and vegetation for different elevations and seasons. We use the model to study how snow and glacier processes affect the hydrological cycle and how vegetation can mediate water yield from the high mountains of a glacierized Himalayan catchment downstream. This bridges the modelling gap between snow- and glacier dynamics, which generate the runoff, and vegetation processes, which interfere with runoff production and water uses at lower elevations.

Our land surface modelling approach provides detailed insights into the importance of each of the energy and mass balance components for the catchment water budget and reveals high altitudinal and subseasonal variability in the hydrologic partition of the high-elevation Langtang valley. Our simulations indicated that the depletion of the cryospheric water budget is dominated by snow melt, but at high elevations primarily dictated by snow and ice sublimation. At the catchment scale we found that water loss through evapotranspiration, dominated by snow sublimation at high elevations and evapotranspiration from vegetation at the lowest altitudinal zone, exceeds the water production from ice melt by > 50%. This shows that vegetation is relevant in determining the amount of runoff transferred further downstream, even for high elevation, extensively glacierized Himalayan catchments.

Glaciers are key components of the Asian water towers and provide water to large downstream communities for domestic, agricultural and industrial uses. In the Nepal Himalaya, the Indian Summer Monsoon dominates climate, and results in a complex meteorology and simultaneous accumulation and ablation that complicate the quantification of snow processes. Assessing solid precipitation input, especially in the upper accumulation area (> 6000 m a.s.l.), remains key to understanding recent mass losses. Catchment-scale glacio-hydrological modelling in the Himalaya has to date mostly relied on temperature-index or intermediate-complexity enhanced temperature-index methods, but recent studies have shown that such approaches can lead to inaccurate amounts of melt, especially at high elevations where refreezing, sublimation and avalanches influence the snow depth variability. The Trakarding–Trambau Glacier system experienced significant mass loss over the last decades, and recent field measurements of meteorology and glacier change present the opportunity to examine these problems with physically-based and spatially-resolved atmospheric and glacio-hydrological modelling.

We combine a novel non-hydrostatic atmospheric model (NHM; atmospheric core of the cryosphere-oriented regional climate model NHM-SMAP) and an advanced land surface model at cloud-permitting hyper-resolution (~ 100 m) to explore the role of snow processes in the water balance of this glacierized catchment. We force the land-surface model of the catchment with dynamically downscaled, hourly outputs from NHM for the 2018-2019 hydrological year.

We evaluate the NHM output using available in-situ meteorological observations and evaluate the land surface model skills and process representation with in-situ mass balance observations, remotely sensed surface elevation change and snow cover. Coupling of the two types of models is unprecedented in the Himalaya, and holds promise to reveal processes that cannot be explicitly assessed by simpler models or forcing data. We investigate the contribution of sublimation and precipitation partition to the glacier mass balance and catchment runoff, and analyze the difference in mass balance and its drivers between the debris-covered and debris free-glaciers. To place this very novel type of simulations into the context of current research, we compare our NHM-forced simulations with simulations forced by station data and ERA5-Land reanalysis. Finally, we evaluate the effect of spatial resolution (50 m, 100 m, 200 m) on model performance and process representation.

Our results highlight the potential of sophisticated models based on the calculations of energy and mass fluxes to unravel the complex processes that shape the response of Himalayan catchments, and provide an assessment of their skills as a function of spatial resolution.

Supraglacial debris thickness is an important control on the surface melt rates of debris-covered glaciers, which are common features of the cryosphere in High-Mountain Asia. Here we present a new dataset of altitudinally-distributed supraglacial debris thickness for the region, generated using remote sensing and numerical modelling techniques. Our modelled debris thickness data are consistent with in-situ data by less than 0.1 m 79% of the time at 14 glaciers, and show similar altitudinal patterns and central values. We show that debris thickness increases as surface velocity decreases, and that debris is thicker on glaciers in a more advanced stage of their debris-cover evolution.

Recent advances in remote sensing approaches have enable large-scale and glacier-specific assessments of glacier volume changes in response to climatic changes. However, these geodetic-differencing thinning patterns integrate the distinct influences of mass balance processes (ie ablation and accumulation) and ice dynamics (advection and flux divergence), rendering them useful for glacier model calibration or validation only at the glacier or larger scales.

In this study, we leverage high-quality digital elevation change and glacier surface velocity datasets along with multi-model ice thickness estimates to isolate the annualized local glacier mass balance for glaciers. Our implementation of the continuity equation builds on recent advances and is applicable to derive altitudinal or distributed mass balance profiles at local and regional scales. We demonstrate the applicability and utility of this approach for several case studies, yielding understanding for glacier health and vital target datasets for improved glacier model calibration and validation.

First, we show that a basic implementation of the continuity equation, applicable with existing regional-scale glacier-change datasets, can yield important insights into glacier health. Specifically, we derive multidecadal effective altitudinal mass balance profiles and leverage these to quantify the equilibrium line altitude (ELA) and accumulation area ratio (AAR) for over 5000 glaciers across High Mountain Asia, and additionally determine the ablation balance ratio, a metric contextualizing the glacier mass balance in terms of the rate of mass turnover.

Second, we apply the continuity equation to derive mass balance gradients for a global subset of glaciers. For this, we leverage new higher-quality large-scale glacier change datasets relevant for the recent five-year period (2016-2020). Over this timeline, firn density profiles and mass balance gradients generally remain stable, making these results suitable for large-scale glacier model calibration.

In a third demonstration, we apply the continuity equation to precise glacier thinning and velocity datasets derived from high resolution photogrammetric products (e.g. derived from Pleiades, Deimos, and UAV datasets) for selected catchments in High Mountain Asia with high-quality ground observations. These short-timescale applications (1-2 years) require more sophisticated flow corrections, but provide annualized specific net mass balance information at high spatial resolution (2m) and are able to represent areas inaccessible by traditional measurements (e.g. avalanche cones). As such, these results provide crucial targets for advanced glacier models of high process complexity.

Taken together these three examples showcase the utility of the continuity equation to bridge the gap between traditional glaciological measurements and new remote sensing datasets. The approach can estimate fully-distributed net annual mass balance at high spatial resolution and over broad domains, independent of traditional melt and mass balance models. As such, our results provide vital new, extensive calibration targets that reduce equifinality risk posed by models calibrated to geodetic measurements alone. These target datasets (e.g. ELAs, AARs, mass balance gradients, local mass balance) are suited for glacier models of varying complexity and process representation.

Soil moisture plays an important role in controlling water and energy exchange between the land and the atmosphere. Characterizing large scale soil moisture is important for many applications, e.g, agricultural and water resources management, drought and flood forecasting. Assimilation (DA) of remotely sensed soil moisture observations into land surface models (LSMs) can improve soil moisture estimation, however, in many studies assimilation of remotely sensed soil moisture improves evapotranspiration prediction hardly. Most LSMs have an over-simplified representation of groundwater dynamics, and the propagation of remote sensing information into neighboring and deeper soil zones depends on the lateral flow and subsurface physical processes represented in the model. In this study, we assimilated soil moisture information into the stand-alone Community Land model (CLM) and the land surface-subsurface model CLM-ParFlow, components of the Terrestrial Systems Modeling Platform (TSMP). The CLM-ParFlow uses Richard’s equation to simulate variably saturated three-dimensional flow in the subsurface and uses a two dimensional kinematic wave approximation for overland flow and river routing. The experiment is conducted for a temperate region (150 km × 150 km) in Western Germany, with a horizontal grid resolution of 500 m, for the period from Mar 2018 to Nov 2018. The SMAP soil moisture data are assimilated with the Ensemble Kalman Filter (EnKF) on a daily basis. We compared the simulated soil moisture content with in situ soil moisture measurements derived from Cosmic Ray Neutron Sensors (CRNS), and simulated ET with observations by Eddy Covariance (EC) stations. It is found that soil moisture characterization improved by DA, but metrics are not better for CLM-ParFlow than CLM stand-alone. Nevertheless, spatial soil moisture patterns by CLM-ParFlow look more realistic than the ones simulated by CLM. DA was able to further improve the characterization of soil moisture contents and improves ET estimation under drought conditions. In addition, with coupled land surface-subsurface models the impact of soil moisture assimilation on simulated groundwater levels and river discharge, which are not represented well by classical land surface models, could also be evaluated.

This work focuses on relatively accurate evaluating the penetration depths of both C-band and X-band radar for glacier areas of the south-eastern Tibet Plateau and further estimate multi-temporal glacier mass balance for the largest lake-terminating glacier in the TP based on the geodetic method. Our results demonstrate that there are either an underestimation of 60% or an overestimation of 202% for the previous C-band penetration corrections. We also found that the rate of mass loss of Yanong Glacier has doubled since 2000, and the interannual mass change has shown a highly volatile and accelerating trend.

Organic farming is an agricultural system that is a priority in EU Mitova (2014), it shows clear environmental advantages in terms of environmental toxicity and the use of biological resources (Nemecek et al. 2006). Cereals occupy a particularly important place in organic farming. They are the main arable crops from which baby and dietary foods are produced and are very in demand on both our and international markets (Atanasova et al. 2014).

Perhaps the most common alternative cereal is the einkorn, which has already found its place in organic farms and among consumers (Konvalina 2011). The einkorn is an alternative for farmers who can incorporate another crop into their crop rotation, which guarantees them stable yield in conditions of sharp climate change. Due to its advantages, the mezza is not only an extremely valuable plant, as a healthy product of high biological value, but its cultivation does not require the use of plant protection products and mineral fertilizers (Eisele & Korke, 1997).

Ground data and drone Unmanned Aerial Vehicle (UAV) footage were collected during the agricultural year 2020-2021 on a biologically certified field with an einkorn located in central southern Bulgaria in the land of Byala Reka village, Parvomai Municipality, Plovdiv region. The boundaries of the field are determined with the help of the farmer who grows the crop in Google Earth Pro.

The field was divided into three separate parts depending on the condition and development of the harvest at the end of March 2021, when the crop was en masse entered the fraternal phase (BBCH 29). The field status was established using the EOS Crop Monitoring platform, in which a KMZ field boundary file was uploaded and the Vegetation Index (VI) NDVI was generated, on the basis of which the field was divided into three separate parts with high NDVI values, those with average NDVI values and low NDVI values. In the selected three different parts of the field, 3 GPS points were generated in the EOS Crop Monitoring platform. On the field were established three squares with sides 10 m × 10 m in the corners of which were permanent twelve permanent sites markers marked with 12 GPS points.

Ground data was collected in the spindle phases (BBCH 45) and milk maturity (BBCH 75), and biological yield was also taken into account when reaching technological maturity (BBCH 99). On the day of filming with UAV, all the lint plants and weed plants were counted in each of the 12 permanent test sites. All the limp plants and weeds were collected, measuring the fresh and dry biomass.

In May in phase spindle (BBCH 45) and June Milk Maturity (BBCH 75) UAV Wingtra (was used with multispectral camera Micasense and Sony RGB camera, The UAV capture data was processed with the Pix4D software. Vegetation indices EVI, MSAVI, NDVI, Chlorophyll Index Green and Chlorophyll Index RedEdge and other 29 VI were generated using the same software.

The results for the indicators characterizing the terrestrial mass of organic sowing were obtained and analyzed in three statistically proven differences in the values of the vegetation index NDVI generated with data from the Sentinel-2 satellite through phase fraternation (BBCH 25) respectively – 0.86; 0.74 and 0.63. From the ground data collected during the spindle phase, it was found that the average weight of the fresh green mass of einkorn gives way to the fresh weight of weeds. In the spindle phase, the most suitable VI for characterization of the sowing of einkorn grown in the conditions of organic farming are VARI and BIM. While VI such as GLI, HI, GRVI and GLAI could be used to assess the degree of entanglance in the crop.

During the milk maturity phase (BBCH 75) of seed development, the sowing of the weeds, most of which are in the initial phases of development. In the milk maturity phase, the vegetation indices SI, BIM, BIS and VVI have no proven differences in the three differences in sowing. They may describe sowing by indicator % dry matter in plants and number of plants and weeds in m2.

In conclusion, it can be said that in terms of indicators characterizing the state of sowing, it is appropriate to perform a filmed with UAV during the spindle phase, and not during the milk maturity phase, since during this phase much of the chlorophyll in plants has already been withdrawn and you cannot well characterize the state of the sowing.

In terms of yield and productivity elements, it is found that data obtained from the BDs during the milk maturity phase are more appropriate to characterize the elements of productivity and yield. Of all the BSI studied, only BSI was found to have a strong positive correlation with yield, and VARI was in a medium negative correlation with yield. During your milk maturity phase, which are in a strong correlation with yield are CVI, SCI and Chlorophyl index.

The crop growth condition in the spring of 2022 in Ukraine was attracted the attentions from agricultural community in the world. Thanks to the global coverage of the second generation FENGYUN polar orbiting satellite, the normalized difference of vegetation index NDVI retrieved from FY3C VIRR and FY3D MERSI were used to monitor the crop growth condition in Ukraine. The NDVI difference model between the current value and the historical mean in the past 5 year and the time series of NDVI at present and the historical mean were used to closely monitor the changes of the crop vegetation from March to June when was the winter crop growth season in 2022. Based on the NDVI difference model, the spatial condition of crop growth was mapped every dekad since the early March and the crop growth condition was categorized into five classes, such as worse, poor, normal, favorable, and good. The NDVI value averaged for the entire country of Ukraine and five states in the north and east neighboring with Russia were used for following the crop growth cycle. The results showed that the crop growth in Ukraine from March to early July 2022 was in a condition of "poor in the early season and better in the late season" that was better than the multi-year average. The generated NDVI time series curve presented a slight lag in April, a gradual increase from the end of April to the end of June, and a slight decrease in NDVI values in the beginning of July but still higher than the past 5 years average. It proofs that the growth of winter crops in Ukraine was not seriously affected.

Monitoring agricultural process of Jiansanjiang Farm based on multi-source remote sensing data

Abstract

Food security is an important foundation of national security, Jiansanjiang farm has a total cultivated land area of 776000 hectares, the average annual grain output accounts for about 1/11 of Heilongjiang Province, 1/100 of the country, An important grain production base in China, in order to fully and timely understand the progress of spring ploughing in Jiansanjiang, to ensure food security, to Jiansanjiang Branch under the jurisdiction of 15 farms as the research object, based on sentinel-2 satellite April 19, April 24, Images from April 29 and images from the Landsat8 satellite on April 28. Monitored the progress of spring preparations for paddy fields on fifteen farms. Based on the random forest algorithm and expert prior knowledge, the images of each period are divided into three categories: undisturbed, irrigated and flooded. According to the classification results, the growth rate of irrigated plots between April 19 and April 24 was faster, and the process of flooded was slower; As of April 29, the proportion of the flooded in Jiansanjiang Farm that has been the promotion of large-scale mechanization operations has increased rapidly, accounting for about 90% of the total paddy field area, and the spring preparation of paddy fields has basically ended

Key words:Food security;Random forest;Supervised classification;Jiansanjiang farm

Microwave retrieval of soil moisture is an underdetermined problem, as microwave emission from the landscape is affected by a variety of surface parameters. Increasing observation information is an effective means to make retrievals more robust. In this study, a multi-temporal and multi-angular (MTMA) approach is developed using SMOS (Soil Moisture and Ocean Salinity) satellite L-band data for retrieving vegetation optical depth (VOD_p, p indicates the polarization (H: horizontal, or V: vertical)), effective scattering albedo (ω_p^eff), soil surface roughness (Z_p^S), and soil moisture (SM_p). The main conclusions are as follows: this paper for the first time at the global scale produced a polarization-dependent SMOS VOD_p and ω_p^eff products, and their global spatial patterns follow global vegetation distributions; the retrieved surface roughness (Z_p^S) range from 0.04 to 0.22 cm, and its spatial distribution is partially different from the existing roughness products/auxiliary data from SMOS and SMAP (Soil Moisture Active Passive). The retrieved MTMA SM_p shows generally high correlations with in-situ measurements with overall correlation coefficients of more than 0.75, and the overall ubRMSE of MTMA-SM_H (0.050 m³/m³) and MTMA-SM_V (0.054 m³/m³) are also lower than that of SMOS-IC Version 2 (V2) (referred to as SMOS-IC) (0.058 m³/m³) and SMOS-L3 (SMOS Level 3) (0.066 m³/m³) products. Therefore, it is concluded that by incorporating multi-temporal SMOS data, the proposed method can be used to systematically retrieve soil moisture, VOD and additional surface parameters (effective scattering albedo and surface roughness were retrieved in addition in this study) with comparable or better performance of soil moisture than that of SMOS-IC and SMOS-L3.

The distributed hydrological models are important tools potentially used for policy planning and decision-making in terms of water-soil balance at the catchment level in different environmental conditions. However, the model calibration and validation present a crucial challenging task in poorly gauged basins, e.g. many river basins in Africa. Our study contributed to providing an operational framework to calibrate hydrological models by using distributed geospatial remote sensing data. The Soil and Water Assessment Tool (SWAT) model was calibrated in monthly steps using only twelve months of satellite-based actual evapotranspiration (ETa) geospatially distributed in the 37 sub-basins of the Lake Chad Basin in Africa. The identification of influential model parameters was done based on global sensitivity analysis by applying the Sequential Uncertainty Fitting Algorithm–version 2 (SUFI-2), incorporated in the SWAT-Calibration and Uncertainty Program (SWAT-CUP). This technique is designed to deal with spatially variable parameters and estimates either multiplicative or additive corrections applicable to the entire model domain, which limits the number of unknowns while preserving spatial variability. Fifteen influential parameters were selected for calibration based on the sensitivity analysis. The optimized parameters set could achieve the best model performance judging by the high Nash–Sutcliffe Efficiency (NSE), Kling–Gupta Efficiency (KGE), and determination coefficient (R²). Four sets of ET were tested for SWAT model calibration, i.e. ETMonitor, GLEAM, SSEBop and WaPOR. Overall, the calibration performance was very good, especially when matching the SWAT ET calculated with Hargreaves-equation based potential ET (ETp), to the ETMonitor ET and GLEAM ET, with performance metrics R²> 0.9, NSE>0.8 and KGE>0.75. The ETMonitor ET product was finally adopted for the SWAT model calibration in this study for further application, since it showed the best calibration results. The calibrated SWAT model were further validated by comparing its outputs with the total water storage change (TWSC) derived from GRACE and surface soil moisture from ESA – CCI product. The validation during 2010-2015 using total water storage derived from GRACE gave an acceptable performance, i.e. R2=0.56 and NSE=0.55. The evaluation against the ESA – CCI soil moisture showed NSE=0.85.

Net ecosystem productivity (NEP) is an important indicator to describe ecosystem function and the global carbon cycle. In this paper, the Carnegie Ames Stanford approach (CASA) model was optimized, and the NEP value of the European terrestrial ecosystem was estimated by coupling the soil respiration model. The results showed that the R² between the estimated value of NEP and the observed value increased from 0.252 to 0.403, and the RMSE decreased from 84.557 gC∙m^-2∙month^-1 to 64.466 gC∙m^-2∙month^-1 after optimizing the maximum light use efficiency () of the CASA model parameters using the vegetation classification data. After further optimizing the optimal temperature, R² increased to 0.428, and the RMSE decreased to 63.720 gC∙m^-2∙month^-1. These results have shown that it is an effective method to improve the NEP estimation accuracy by optimizing and the optimal temperature to improve the CASA model. On this basis, the spatial and temporal changes in NEP in various regions in Europe were analyzed using the optimization results. The results show that NEP in Europe is in the spatial distribution pattern of Western Europe > Southern Europe > Central Europe > Eastern Europe > Northern Europe. The monthly changes in NEP in all regions show a unimodal curve with summer as the peak, and the annual overall value is positive (i.e., it shows a carbon sink). The research results can deepen the understanding of the carbon source/sink distribution in Europe and provide a reference for carbon cycle research and carbon balance policy formulation in the region.

Land cover data is an important information in natural resource survey, land management, environmental monitoring, etc. It is of great significance to evaluate its accuracy and reveal its category confusion characteristics for many scientific fields. In this study, nine urban agglomerations were selected as the study area, and samples were collected by visual interpretation of Google Earth's high-resolution images. Then, the spatial distribution characteristics and classification accuracy of four land cover data products (GlobeLand30, FROM-GLC, GLC-FCS30 and CCI-LC) were analyzed quantitatively and qualitatively. The results show that all products have achieved good results in the classification of urban agglomeration features, among which the overall accuracy of FROM-GLC is the highest, reaching 82.34%, and CCI-LC is relatively low, with the overall accuracy of 78.09%. Further analysis shows that the classification accuracy of various data products for different land types is different. The classification accuracy of farmland, forest and other large and concentrated land types is higher, while the accuracy of shrubs, wetlands and other small and scattered land types is relatively low. The research results can help users choose data products according to their needs, and provide reference for data producers to improve product accuracy.

Net primary productivity of vegetation refers to the ability of green plants to fix and convert inorganic carbon into organic carbon by using sunlight for photosynthesis. It is not only an important variable to characterize plant activity and has practical significance in crop yield estimation, forest stock volume survey, grassland yield and ecosystem material circulation, but also the main factor to determine the carbon source and sink of the ecosystem and regulate the ecological process, and is mainly affected by environmental factors such as climate and land use.

Africa is the region with the least greenhouse gas emissions but the most affected. The report of United Nations Development Programme (UNDP) pointed out that CO₂ and other greenhouse gases emitted by developed and developing countries will have a serious impact on Africa, especially the sub-Saharan region in the future. Africa has the lowest greenhouse gas emissions of all continents (except Antarctica), but the worst effects of climate change on the stability of African ecosystems come first. Given that NPP is one of the key indicators to characterize the health of ecosystems, it is crucial to analyze the temporal and spatial variation trends of African vegetation NPP, which is of practical significance for ecological protection in Africa. Therefore, using the long time-series data of global NPP from 1981 to 2018, this paper will solve the following problems: (1) using trend analysis and coefficient of variation to analyze the change trend of African NPP; (2) using anomaly index analysis and Mann-Kendall test to study African NPP; (3) using wavelet analysis to explore the periodic variation and temporal patterns of African NPP.

The results suggest that: (1) The annual NPP has a significant change trend with a total of 48.56% of the pixels NPP changing. Among them, the NPP reduced in 32.44% of the pixels and significantly decreased in 29.37% of the land area, mainly concentrated in the Sahara Desert to the north of 15°N. The NPP increased in only a small part of the region, approximately 16.13%, and the NPP increased significantly in 12.08% of the areas, mainly in the north and south sides of the tropical rainforest area. (2) 46.27% of the pixels have low degree of NPP fluctuation, which are mainly concentrated in the Sahara Desert to the north of 15°N, northeastern East Africa and western South Africa. 39.49% of the regions with high degree of NPP fluctuation are mainly located in the north and south sides of the equatorial tropical rainforest. Among them, the Central African region with the equator as the center and extending about 5° from north to south has the highest NPP fluctuation. (3) The period from 1981 to 2018 can be divided into four stages. In 1981, NPP in Africa was generally higher than the average level, indicating that Africa's carbon sink capacity was strong at this stage. From 1982 to 1995, the NPP during different seasons in Africa was basically lower than the average level, indicating that Africa's carbon sequestration capacity was low during this period. In particular, NPP declined significantly during 1987-1992. From 1996 to 2018 (except for 2015 and 2016), the NPP of Africa in different seasons were basically higher than the average level, showing an overall upward trend, indicating that Africa’s carbon sink capacity was increasing. (4) Seasonal NPP increased over time, and there were mutation points in both annual and four-season NPPs in Africa, all occurring around 1995. (5) On the annual scale, NPP has a short period of 4-8 years, 15-21 years and 23-35 years, and a long period of 42-62 years, and exists on the time scale of 7 years, 20 years, 29 years and 55 years. Significant oscillations, of which the 55-year cycle has the strongest signal, is the first main cycle, and the second and third cycles are 29 and 20 years, respectively.

Xuzhou is one of the main producing areas of garlic in the country. Accurate and rapid acquisition of garlic spatial distribution information plays a very important role in estimating garlic production and daily prices. This paper takes Xuzhou as the research area, and based on the Google Earth Engine (GEE) cloud platform and Sentinel-2 data, the training samples are determined through visual interpretation and field inspection, and the NDVI index time series curve of typical crops in the study area is calculated. Construction and spectral index feature construction. After comparing the three classification algorithms of random forest classification, classification regression tree and support vector machine, the classification performance of different algorithms is evaluated, and the accuracy is verified. Among them, the random forest algorithm has obvious advantages over other algorithms. In the research of land object classification, the random forest algorithm has obvious advantages. Compared with the other two algorithms, the overall accuracy is 37.4‰ and 87.2‰ higher, and the kappa accuracy is 53.3‰ and 122.3‰ higher than the other two algorithms, respectively.

Key words: Remote Sensing Extraction; Google Earth Engine; Random Forest; Feature Extraction

Agricultural systems are the main consumers of freshwater resources at global scale, consuming 60 % to 90 % of the total available water. While the growing demand for agricultural products and the resulting intensification of their production will increase the dependency on available freshwater resources, this sector will become even more vulnerable because of the intensifying impacts of climate change. Detailed knowledge about soil moisture can help to mitigate these effects. Nevertheless, high resolution (space & time) surface soil moisture data for regional and local monitoring (down to precision farming level) are still challenging to obtain. By using current as well as future Synthetic Aperture Radar (SAR) satellite missions (e.g. Sentinel-1, ALOS-2, NISAR, ROSE-L), this knowledge gap can be filled. SAR observations are suitable for regional and local soil moisture estimations, but with a global extent. While the increasing resolution and total number of SAR recordings will contribute to an improvement of the estimation in general, the computational costs as well as the local memory capacity on the other hand become a limiting factor in processing the vast load of data. Here, on-demand cloud-based processing services are one way to overcome this challenge. This is especially interesting as most of the severely affected regions have limited access to computational resources.

Based on the alpha approximation approach of Balenzano et al. 2011, we developed an automated workflow for estimating soil moisture using temporal and spatial high-resolution Sentinel-1 timeseries. The workflow is established within the cloud processing platform Google Earth Engine (GEE), providing a fast and applicable way for on-demand computation of soil moisture for individual time periods and areas of interest around the globe. The algorithm was tested and validated over the Rur catchment (Germany); with an area of 2,354 km², it comprises a great diversity in agricultural cropping structure as well as topologies. A total of 711 individual Sentinel-1A and Sentinel-1B dual-polarized (VV + VH) scenes in Interferometric Wide-Swath Mode (IW) and Ground Range Detected High Resolution (GRDH) format are used for the analysis from January 2018 to December 2020. Using all available orbits (both ascending and descending), a temporal resolution of one to two days could be achieved with a spatial resolution of 200 m. The workflow includes multiple steps: despeckling, incidence angle normalization, vegetational detrending and low-pass filtering. The results were validated against eight Cosmic-Ray Neutron Stations (CRNS). In total, the method achieves an unbiased RMSE (uRMSE) of 5.84 % with an R² of 0.46. Looking at individual months, the highest correlation can be achieved in the months April and October with R² values range between 0.65 to 0.7, while the lowest correlation is observed in July and January, with R² values ranging between 0.15 o 0.2. Looking at individual landuse, the method achieves the best results for pastures, with an uRMSE of 0.42 and an R² value of 0.63.

Knowledge of Soil properties in agricultural fields allows more efficient use of resources, but this kind of information is rarely available. The qualitative information included in existing soil maps is often insufficient for site-specific management strategies, for these purposes, the quantitative estimation of soil properties reached through time-consuming and expensive on-site investigations is necessary.

Remote sensing data can be applied to acquire, in a cost-effective way, quantitative information about soil. The leaves and the soil are the main elements influencing the spectral reflectance of the image pixel. The position and intensity of the reflectance peaks are associated with absorption, high spectral resolution is essential to resolve the spectral features of interest with high accuracy. Existing instruments (eg. Sentinel 2, etc.) have not been designed to provide high spectral and spatial resolution capabilities in the spectral range of 400-2500nm to fulfil these needs. The assessment of soil variables from multispectral remote imagers is hindered by inadequate spectral resolution, multispectral satellite data are mainly used for qualitative assessments. The hyperspectral sensors, having the capability to observe the full spectrum between 400-2500nm with a better spectral resolution are more desirable for soil spectroscopy purposes.

This project represents the initial phase of retrieving topsoil properties with multiplatform and multi- hyper-spectral EO data, in particular hyperspectral data, using machine learning and multivariate regression.

The study areas are located in “Quzhou County”, a county of Hebei Province, China, administered by Handan Prefecture. Data collection on the ground has been carried out in synchronous with satellite observations. From about 50 fields, for a total of 95 , between 2019 and 2020, measurements of topsoil properties like Soil Organic Matter (SOM), pH, Effective Phosphorus, Available Potassium, and Total Nitrogen, were retrieved.

Satellite data from European Space Agency (ESA) Sentinel-2 and the Italian Space Agency (ASI) Hyperspectral Precursor and Application Mission (PRISMA) are being utilized.

PRISMA was launched on 22 March 2019, the instrument is a prism spectrometer, the design is based on a push broom sensor type observation concept providing hyperspectral imagery (around 250 bands) at a spatial resolution of 30 m. The spectral resolution is about 12 nm in a spectral range of 400-2500 nm (VNIR and SWIR regions). Panchromatic imagery is provided along with the Hyperspectral cube, at a spatial resolution of 5 m, and is co-registered with the latter to allow testing of image fusion techniques.

Sentinel-2 is an Earth observation mission part of the European program Copernicus. Two satellite (S2A and S2B) compose a constellation allowing a 5 days revisit frequency. Each satellite is carrying a single multi-spectral instrument (MSI), with 13 spectral channels, in the range 400-2500nm, that acquires optical imagery at a spatial resolution from 10 to 60 m.

The site's phenology was retrieved by studying the temporal series of vegetation indices such as NDVI and NBR2 obtained by the Sentinel-2 L2A images from 2019 to 2021. The crop phenology was applied to identify the days of the year when the bare soil is visible [1].

PRISMA images have been co-registered with Sentinel-2 data through an Automated and Robust Open-Source Image Co-Registration Software (AROSICS) with the aim to improve their georeferencing.

To improve the retrieval of soil properties the images have been preprocessed to produce fused data with high spatial and spectral resolutions [1]. Both the pan-sharpening, with the panchromatic PRISMA images, and the multispectral/hyperspectral fusion, with the Sentinel 2 reference image used for co-registration, have been attempted to increase the spatial resolution of the hyperspectral cube to 5 and 10 m respectively [2].

Bibliography

[1]

N. Yokoya, T. Yairi and I. Akira, “Coupled Nonnegative Matrix Factorization Unmixing for Hyperspectral and Multispectral Data Fusion,” IEEE TRANSACTIONS ON GEOSCIENCE AND REMOTE SENSING, February 2021.

[2]

A. ARIENZO, G. VIVONE and A. GARZELLI, "Full Resolution Quality Assessment of Pansharpening: Theoretical and Hands-on Approaches," IEEE Geoscience and Remote Sensing Magazine, May 2022.

[3]

N. Mzid, F. Castaldi, M. Tolomio, S. Pascucci, R. Casa and S. Pignatti, "Evaluation of Agricultural Bare Soil Properties Retrieval from Landsat 8, Sentinel-2 and PRISMA Satellite Data," Remote Sensing MDPI, 2022.

Recently, Remote Sensing (RS) information has been involved in numerous hydrological modelling applications as a tool for quick collection of geophysical data. This data should be handled with care, being aware of its characteristics, such as spatial and temporal resolutions and target area composition.
In this work, evapotranspiration estimates over heterogeneous crops are subjected to a scale analysis. These have been obtained from a distributed energy-water balance model (FEST-EWB) employing, among others, high-resolution remotely-sensed Land Surface Temperature (LST) and vegetation data. FEST-EWB is calibrated, pixel-wise, on measured LST, by minimizing, for each pixel within the domain of interest, the model error against the satellite-retrieved LST. The case study is a Sicilian vineyard, with some observations days in the summer of 2008. During these days, meteorological and energetical fluxes data were gathered through an eddy-covariance station, whereas airborne instruments collected LST and vegetation data are obtained at the high resolution of 1.7 metres.
After a preliminary calibration on LST data and validation on energy fluxes, the scale analysis is performed: model results are compared after both input and output aggregation. In reference to some common satellite products, four coarse resolutions were selected for the analysis: 10.2 m (in reference to Sentinel’s 10 m), 30.6 m (Landsat, 30 m), 244.8 m (MODIS visible, 250 m) and 734.4 m (MODIS, 1000 m). Firstly, modelled surface temperature and evapotranspiration are upscaled to each scale by progressive averaging. Then, model inputs are aggregated to the same spatial resolutions and the model is calibrated anew, obtaining independent results directly at the target scale. The results of the two procedures are found to be similar, positively reporting the model flexibility in providing accurate products for heterogeneous areas even at lower spatial resolutions.

Urbanization is continuing at an unprecedented rate in many cities across the globe. Timely and reliable information on the sprawl of settlements are important to support sustainable planning. Earth observation has been playing a crucial role to map land cover changes associated with urbanization (Ban & Yousif, 2016). Many studies have been conducted to demonstrate the potential of Synthetic Aperture Radar (SAR) and multispectral data for urban change detection (e.g. Ban & Yousif, 2012; Bovolo & Bruzzone, 2015; Bruzzone & Prieto, 2000; Gamba et al., 2006).
In recent years, several deep learning methods using fully Convolutional Neural Networks (CNNs) have been used to detect changes in multi-temporal satellite imagery. In particular, the vast amount of high resolution (10–20 m) imagery collected by the Sentinel-2 MultiSpectral Instrument (MSI) mission has been used extensively for urban change detection (e.g., Daudt et al., 2018; Papadomanolaki et al., 2021). More recently, we demonstrated that fusion of Sentinel-1 SAR and Sentinel-2 MSI data can improve urban change detection results (Hafner et al., 2021). While our dual stream network using late fusion achieved improvements over input-level fusion, recent work demonstrated that multi-modal deep neural networks suffer from their greedy nature, meaning that they tend to rely on just one modality while the other modality remains largely unused (Wu et al., 2022). In turn, greedy learning negatively affects the model’s generalization ability, which Wu et al. (2022) overcame by proposing an algorithm to balance the conditional learning speeds between modalities during training.
For this study, we investigate the greedy nature of deep neural networks for multi-modal learning from Sentinel-1 SAR and Sentinel-2 MSI data. To that end, we add a Multi-Modal Transfer Module (MMTM) to different levels of the U-Net encoder and decoder. Using squeeze and excitation operations, the MMTM enables intermediate modality fusion for effective multi-modal learning (Joze et al., 2020). The modified dual stream U-Net is then trained on an urban change detection task to investigate the model’s greedy nature using the concept of conditional utilization rate introduced in Wu et al. (2022). The conditional utilization rate for a given modality in a multi-modal setup is the gain on the accuracy when a model has access to the modality in addition to another modality. Finally, we investigate methods to overcome the greediness of multi-modal deep neural networks, including the proposed algorithm in Wu et al. (2022). All experiments are conducted on the SpaceNet 7 dataset, consisting of time series of monthly Planet mosaics and corresponding building footprint annotations for 60 sites covering unique geographies around the world (Van Etten et al., 2021). Sentinel-1 SAR and Sentinel-2 MSI images were downloaded from Google Earth Engine to replace the Planet mosaics (Gorelick et al., 2017). This research is progressing well, and the urban change detection results will be finalized, validated and presented at the Dragon 5 mid-term symposium.

Ground object detection and retrieval from a large-scale remote sensing image are very important to manage smart cities and support sustainable cities. We present a novel problem of visual grounding in remote sensing images. Visual grounding aims to locate the particular objects (in the form of the bounding box) in an image by a natural language expression. The task already exists in the computer vision community. However, existing methods mainly focus on natural images rather than remote sensing images. Compared with natural images, remote sensing images contain large-scale scenes and the geographical spatial information of ground objects (e.g., longitude, latitude). The existing method cannot deal with these challenges. To address the drawback, we design a new method, namely GeoVG. In particular, the proposed method consists of a language encoder, image encoder, and fusion module. The language encoder is used to learn numerical geospatial relations and represent a complex expression as a geospatial relation graph. The image encoder is applied to learn large-scale remote sensing scenes with adaptive region attention. The fusion module is used to fuse the text and image features for visual grounding. We evaluate the proposed method by comparing it to the state-of-the-art methods. Experiments show that our method outperforms the previous methods by a large margin.

In the face of Synthetic Aperture Radar (SAR) image object detection with different distributions of training and test data, traditional supervised learning methods cannot achieve good detection performance. Domain adaptation (DA) method has been shown to have the ability to solve this problem, but existing DA object detection algorithms all use adversarial DA theory for the detection task, which is ineffective in solving object regression localization in the detection task. In this article, to better solve the above problem, an automatic SAR image ship detection method based on feature decomposition crossing different satellites is proposed. The feature extraction layer of backbone network is divided into low level and high level, where domain-invariant feature extractors are designed for the local features extracted from the low level and the global features extracted from the high level, respectively.We argue that the local and global features extracted from source domain and target domain contain domain-specific features (DSF) for adversarial DA and domain-invariant features (DIF) that contribute to object regression localization. Then, we decompose the local features and global features into DSF and DIF via vector decomposition method. For DSF counterpart, we introduce adversarial DA attention for feature alignment. DIF from the local features are fused into the backbone network for high-level global feature extraction. Finally, by using region proposal network and adversarial domain classifier, we can get the accurate bounding box and object class of SAR image objects. Extensive experiments prove that the proposed method outperforms state-of-the-art methods in terms of detection performance.

In the fall of 2021 took place one of the biggest eruptions in the volcanic Canary Islands, on the La Palma island. This paper proposes the use of multispectral Sentinel 2 time series data to monitor pre and post event activity and assess vegetation’s damages. The analysis will be carried out between July 2021 and February 2022. The area of interest is the Cumbre Vieja, an active volcanic ridge on the island. It covers dozens of craters and cones, overlaying the southern half of la Palma island. To study the impact of the volcanic eruption on the vegetation, we will use two vegetation indices, the normalized difference vegetation index (NDVI) and the nonlinear version of this index (kNDVI). The NDVI index is extremely widespread because through it highlitghts changes in vegetation due to both natural disturbances such as wild fires, plant changes and human activities, such as deforestation. Therefore, it can show us how the vegetation was before the eruption and how it looks both during and after the eruption. We expect to find a correlation between the distance from the eruption and changes in the normalized difference vegetation index (NDVI), more precisely we expect to see that NDVI increases with the increase in distance from the eruption. We aim to demonstrate that the nonlinear version of this index (kNDVI) consistently improves accuracy in monitoring key parameters and also that it enables more accurate measures. In the end our goal is to assess these indices against NHI (Normalized Hotspot Indices), a system that performs the automated monitoring of volcanic thermal anomalies.

As an important oceanic physical process, mesoscale eddies play a key role in the processes of ocean mass transport and energy exchange, such as biogeochemical cycles, marine ecosystems, and marine heat balance.

In recent years, these automated ocean eddy identification and tracking algorithms can be divided into two categories: Eulerian- and Lagrangian-based approaches. Based on instantaneous sea surface height (SSH) or velocity field, the major circular structures of mesoscale eddies can be detected, which are called as Eulerian eddies. Previous SSH-based and sea level anomaly-based (SLA-based) methods have shown the best performance due to their ability to avoid extra noise and excess eddy detections. Although SSH-/SLA-based methods worked well at the basin scale, the calculation efficiency decreases distinctly at the global scale, mainly due to the high-order computation complexity on contour iterations. With the improvement of data resolution, existing hardware fails to meet the requirement of long-time-scale global eddy recognition due to the increase of the number of SSH/SLA contours. Thus, based on satellite altimeter data, a highly effective orthogonal parallel algorithm for identifying and tracking global eddies is proposed. Surprisingly, this algorithm is ~100 times faster than the previous SSH-based method on global eddy detection without reducing the accuracy of mesoscale eddy recognition. According to this orthogonal parallel algorithm, the global mesoscale eddy dataset for the past 28 years (1993-2020) was generated, which provides a data foundation for the subsequent study of mesoscale eddies. In terms of the mesoscale eddy dataset, an automatic recognition method of global dipole eddy pairs that consist of two counter-rotating eddies moving together for a period of time within a limited space distance is proposed by using the K–D tree for cutting space. Simultaneously, the transmission modes and characteristics of dipoles are analyzed, including the characteristics of long life, high propagation speed, and entanglement trajectory. In addition, an algorithm named EddyGraph for tracking mesoscale eddy splitting and merging events is proposed and the corresponding data set of eddy trajectories in the Northwest Pacific is available, which would fill the gaps in data sets to support studies on eddy splitting and merging in the Northwest Pacific.

Different from other methods based on instantaneous flow fields, Lagrangian eddies are the cumulative results of the state of the fluid within a given time scale, which can maintain material coherence over the specified time intervals. Firstly, by using the elliptic Lagrangian Coherent Structures, the boundary of a black-hole eddy was extracted based on the data of the geostrophic flow velocity field in the Western Pacific Ocean. Combined with multi-source satellite remote sensing data and in-situ data, the consistency of material transport in the horizontal direction and the coherence of material in the vertical direction of the vortex are analyzed and verified. The results show that the black hole vortex boundary can describe material transport more objectively than the Euler vortex boundary on a longer time scale. Then, the Lagrangian eddies in the western Pacific Ocean are identified and analyzed. By introducing the Niño coefficient, the lag response of the Lagrangian eddy to El Niño is found. Through normalized chlorophyll data, it is observed that Lagrangian eddies can cause chlorophyll aggregation and hole effects. These findings demonstrate the important role of Lagrangian eddies in material transport. Nevertheless, although Lagrangian eddies work well at estimating material transport, the high calculation cost during the integration process has become a bottleneck, especially when the data resolution is improved or the study area is enlarged. Therefore, SLA-based orthogonal parallel detection of global rotationally coherent Lagrangian eddies is built, whose runtime is much faster than that of a previous nonparallel method. Finally, a data set of long-time-scale global Lagrangian eddies is established.

Furthermore, an integrated marine visualization system, named i4Ocean, has been presented. The system is designed and implemented to investigate and study physical marine processes by visualizing and analyzing spatiotemporal marine data. Notably, these actions are realized by providing various GPU-based interaction and visualization techniques for displaying multidimensional data. The system achieves three goals: high visibility, good performance and interactive capabilities. The techniques of z-coordinate calibration and sphere rendering, which restore the most authentic marine environment, provide excellent feedback for oceanographers. The efficient ray sampling technique including a preintegrated transfer function and adaptive sampling methods, increases the rendering efficiency of ocean data. By further introducing a transfer function, users can extract the region of interest in the system and analyze diverse marine phenomena. A data-centric approach was adopted to guide the design of the transfer function by analyzing the scalar field and its properties. The best parameters of the transfer function were obtained to maximize the visibility of important features, which helps to analyze mesoscale eddies of typical ocean phenomena.

Detecting forest insect damage before the visible discoloration (green attacks) using remote sensing data is challenging, but important for damage control. In recent years, the European spruce bark beetle (Ips typographus, L.) has damaged large amounts of forest in Europe. However, it is still debatable how early the infestations can be detected with remote sensing data. Some studies showed a spectral difference between healthy and green-attacked spruce trees at the plot level, while others showed that spectral differences existed before attacks. Therefore, a hypothesis is proposed that no spectral difference can be identified between green-attacked forests compared to healthy forests if the differences do not exist before the attacks. In this study, we tested this hypothesis using Sentinel-2 and WorldView-3 SWIR images on 24 healthy plots and 24 plots with mild, moderate, and severe attacks. In the results, the severely attacked plots did not show significant spectral differences in the Sentinel-2 images until August, and the sensitivity was found in the blue, red, red-edge, and SWIR band. Only the red band showed a significant difference between the healthy and moderately attacked plots in August, and only the blue, red, and SWIR band showed significant differences in September, October, and November. No significant differences were observed in the WorldView-3 images at the plot or individual tree level. We accepted the hypothesis that green attacks do not show spectral differences with the healthy forests when the differences do not exist before the attacks. We concluded that the SWIR bands were sensitive to attacks in the Sentinel-2 images with 10 m resolution, but not in the WorldView-3 images with 3.7 m resolution. Further studies are needed to explore the methodology of using WorldView-3 SWIR images for the early detection of forest infestation.

Accurate and fine multiple tree species supervised classification based on few-shot learning has attracted close attention from researchers, because the sample collection is often hindered in forests. Prototypical networks (P-Nets), as a simple but efficient few-shot learning method, have significant advantages in forest tree species classification. Nevertheless, the overfitting phenomenon caused by the lack of training samples is still prevalent in few-shot classifiers, which brings challenges to training accurate classification models. In this study, we proposed a novel Proto-MaxUp (PM) framework to minimize the issue of overfitting from the perspective of data augmentation and a feature extraction backbone for tree species classification. Taking Gaofeng Forest Farm (GFF) in Nanning City, Guangxi Province, as the study area, nine tree species, cutting site, and road were classified. First, by analyzing the effects of a series of popular data augmentation methods and their combinations in different parts of the P-Net, several effective data augmentation pools were established. Then, the pools aforementioned were combined with PM to obtain the best classification performance. To verify the robustness and validity of the proposed strategy, we applied PM to the other four popular public hyperspectral datasets and achieved excellent results. Finally, this efficient data augmentation method was used in different feature extraction backbones. The results show that the classification accuracy was greatly improved with the optimal backbone (overall accuracy (OA) and Kappa, are 98.08% and 0.9789, respectively), and the difference between training accuracy and test accuracy is less than 2%. It is concluded that the accurate and fine classification for multiple tree species can be realized by the PM data augmentation strategy and backbone proposed in this article.

Pine wilt disease (Bursaphelenchus xylophilus, L.) is one of the major forest diseases in China. The pine wilt nematode infects trees in the Pinus genus, and the infected trees usually die within three months. Since it was discovered in 1982 in China, it has been spreading rapidly and now spread in 19 provinces in China, covering an area of 1.8 million hectares, causing significant damage to the forest and ecological environment. The nematode cannot travel outside the wood independently but is spread by the main insect vector pine sawyer longhorn-beetles (Monochamus spp.,L.) during feeding and oviposition. Therefore, removing the trees infected by the pine wilt nematode from the forest as early as possible is essential to control the spread.

This study aims at developing methods to detect infections using hyperspectral drone images. We inventoried 391 pines in middle east China and recorded them as healthy or early-, middle-, late-stage infected trees. The hyperspectral drone images were obtained with 0.11 m resolution and wavelength from 400 to 1000 nm, covering from red band to near-infrared (NIR). We used the successive projections algorithm (SPA) to select the sensitive bands and the support vector machine (SVM) algorithm to classify trees into different health statuses.

The results showed that when the infection developed into the middle and late stages, the tree crowns showed different signatures from the healthy ones, while during the early stage, the spectral signatures were similar to healthy ones, which decreased the detection accuracy. The classification resulted in high accuracy during the middle and late-stage infection, while separating healthy and early-stage infection was challenging. The spectral signature showed a decreasing ratio between the red and red-edge bands during the infection. We will develop the method using the derivative of the spectral signature to achieve accurate early detection in the future.

Crop classification is one of the most important applications in polarimetric SAR images. Time-series polarimetric SAR images have the characteristics of reflecting the changes of various scattering features of crops in different growth periods. However, since time-series polarimetric SAR needs to combine multiple single polarimetric SAR images, the redundancy between features is multiplied. In this paper, aiming at the problem of feature redundancy, the method of similarity measurement is used to select features from two dimensions of polarimetry and time respectively to reduce feature redundancy. Since the sample size of SAR feature images applied in supervised classification is small, which makes it not suitable for multiple downsampling in CNN, so a suitable classifier based on Transformer is designed. Preliminary experiments on the full polarimetric data verified the effectiveness of the proposed method.

Forest height is an important information for evaluating and analyzing forest resources. Accurate estimation of forest height has great significance for forestry management and ecological research. Multi-band InSAR utilizes the penetration differences of different SAR wavelength, and can obtain the elevations of the underlying topography and the forest canopy surface at the same time. Therefore, it has the ability to extract the forest height. However, in practice, although the long-wavelength InSAR has better penetrating ability than short-wavelength InSAR, its signals are still affected by vegetation scatterers resulting the phase center deviates from the underlying surface. And the extracted elevation of long-wavelength InSAR cannot be directly used as the underlying topography for forest height extraction. On the other hand, the penetration ability of short-wavelength InSAR to vegetation layer cannot be ignored. There is a certain deviation between the phase center and the canopy surface, and the extracted elevation of short-wavelength InSAR cannot be used as a digital surface model (DSM) representing the forest canopy elevation. Therefore, from the perspective of obtaining accurate underlying topography and DSM, the paper proposed a forest height extraction method based on multi-band InSAR data. In this method, an subapeture decomposition approach was used to obtain the underlying topography based on long-wavelength InSAR data. Moreover, based on the gap penetration characteristics of InSAR signal, a DSM penetration bias compensation method for short-wavelength InSAR data was developed based on multi-layer model. Eventually, the forest height was obtained by the difference between the extracted DSM and the underlying topography. To verify the method, the experiment was carried out based on the multi-band airborne InSAR data.

Measuring forest height on a large scale is of importance to forest resource management and biomass estimation. For deformation monitoring applications of radar interferometry, sufficient time series short-baseline polarimetric SAR interferometry (PolInSAR) data have been achieved, and since PolInSAR is limited by interferometric spatial-temporal baseline, using this type of data to estimate forest height will cause a large error. In this paper, a total of five ALOS-2 PALSAR-2 data were obtained in Saihanba forest farm, and the semi-empirical model based on the simplified Random Motion over Ground (RMOG) model and the machine learning algorithm combined with multiple features were used to evaluate the potential for forest height estimation using time series short-baseline PolInSAR data. Experimental results show that: (1) Compared with the semi-empirical model, machine learning algorithms can take full advantage of the multi-feature information of the data and achieve better estimated performance. (2) After combining polarimetric and interferometric characteristics, the phenomenon of overestimation in low regions and signal saturation in high regions can be effectively improved. An R² of 0.44 and RMSE of 3.08m was achieved for inversion result of forest height in pixel size of 90m×90m.

Synthetic Aperture Radar Tomography (TomoSAR) provides an unprecedented opportunity to characterize volumetric environments such as forested areas using 3-D electromagnetic reflectivity maps. Classical 2-D SAR imaging capabilities can be extended to 3-D using acquisitions performed from slightly shifted trajectories, and a coherent synthesis along an additional aperture in elevation. As shown by experiments based on the use or airborne SAR sensors, TomoSAR and its multi-polarization version, PolTomoSAR, is able to characterize various kinds of forests (tropical, temperate, boreal) and may be used to estimates forest height, above ground biomass, underlying ground topography, canopy structure…

However, the application of TomoSAR using spaceborne devices is hindered by the time lag separating successive SAR acquisitions, whose value, on the order of a few days, depend on orbital considerations and on laws of physics. For radars operating at higher carrier frequencies, i.e. at L, S, C, X, Ku bands and above, the correlation time over vegetated environments rarely exceeds hours or minutes, limiting the 3-D analysis through repeat-pass TomoSAR to temporally stable targets, such as those encountered in urban scenarios. A possible solution to this limitation consists in using single-pass interferometers, consisting of two or more SAR sensors measuring, at the same time, the observed scenes from different positions, i.e. in a bistatic configuration. Simultaneous SAR acquisitions permit to solve highly limiting problems related to temporal decorrelation, whereas slight modifications of the relative trajectory between the sensors allows to describe an aperture in elevation and to successfully apply SAR tomographic focusing. Another advantage of this operation mode is that the acquisition of a tomographic stack may be spanned over a large period of time, provided that the structure of the observed medium does not change drastically, i.e. generally months.

This paper illustrates the principles of incoherent bistatic tomography, shows the different processing steps of this technique, which significantly differ from the ones employed to perform repeat-pass TomoSAR. Relevant solutions to operational challenges, linked to the imperfect knowledge of the scene geometry, irregular baseline sampling, and even missing data are presented and validated using airborne data sets using geophysical parameter estimation procedures from the state of the art.

Theoretical aspects will be complemented by an analysis of real data from the ESA campaign TomoSense, where bistatic data were collected at L- and C- Band over the forest site of the Eifel Park, North-West Germany, by flying two airplanes in close formation. Preliminary analyses at L-band were already carried out and gave promising results. A new model-based approach for 3-D reconstruction was developed and implemented; the outcome was then compared to the tomographic profiles produced by standard repeat pass tomography. The structure of the forest was properly recovered in both cases. Quantitative analyses about the visibility of the ground and the forest height were also carried out; the discrepancy of the forest height with respect to the LiDAR map resulted in about 2.3m (1 sigma).

Ground subsidence caused by the exploitation of mineral resources is not only an important factor to be considered in the development and utilization of land space, but also an obvious indication for the area of underground illegal mining. The mining distribution of mineral resources is wide and scattered, so it is very necessary to quickly and accurately identify and extract the spatial distribution of mining subsidence in large areas.In this paper, the multitemporal difference interferometric phase diagram of subsidence mining area is obtained by using synthetic aperture radar differential interferometry (DInSAR) technology, FCN-8s, PSPNet Deeplabv3 and U-Net models are used to train the network. The results show that the U-Net model has high detection accuracy and takes short time. 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 ECA-UNet results show that compared with the traditional model, the intersection union ratio (IOU) corresponding to mining subsidence is increased by 2.54%.

The deformation of reservoir bank slopes and water conservancy and hydropower engineering facilities is related to the operation safety of water conservancy and hydropower projects. Based on the multi-temporal Sentinel-1 images, the time series InSAR technique was used to carry out a demonstration application in the Xiaolangdi Multipurpose Dam Project. Four small areas with large deformation rate monitoring points were found on the slope of the reservoir bank in the key area of the north bank of Xiaolangdi. The deformation rate of the monitoring points mainly concentrated in -10mm~ -25mm/yr. In the central area of the crest of the Xiaolangdi Dam, a large deformation accumulation area with an annual deformation amount of -60mm/year was found in the satellite line of sight. Before January 2021, the deformation amount increased rapidly. From January to March 2021, the deformation is slowing down. The InSAR technique can quickly obtain the deformation information of water conservancy and hydropower engineering facilities and reservoir bank slopes, which can be used as a common monitoring method for monitoring the safe operation of water conservancy projects.

The collapse of buildings is the main cause for casualties after earthquakes. Real-time and accurate positioning of the building areas are crucial to make an effective implementation of emergency rescue after an earthquake. This work carries out an investigation into a building recognition method using a single post-earthquake SAR image with the complex background and the scatter distribution pattern. In the SAR image, the sparse and scattered building identification, the large parts of highlighted mountains and the large-scale structures significantly affect the identification of building areas. In order to solve these problems in assessing building damage, we propose to use sparse villages in low spatial resolution SAR images, scattered buildings in high spatial resolution SAR image, oriented buildings and a set of construction features, and develop an algorithm based on the recognition of linear ridgelines and the planar highlighted mountain elements. We apply the algorithm to identify point structures and planar structures, which are used to build the identification of buildings after an earthquake. In this report, we will present a novel recognition method based on spatial association modelling and self-group spatial magnetic expression for performing building damage survey and assessment, this method makes use of the background of building objects combined with semantic features to great extent. It comprises of the following components: target restoration recognition using geometric feature matching, the local space matching recognition algorithm of standardized shape primitives, the ridgeline detection algorithm combining terrain features and linear elements, the recognition method based on layered local texture feature sequence, the oriented building recognition based on linear feature tracking and the effective feature optimization. We also present preliminary assessment results on the building damage of the 2021 Haiti earthquake.

Pre-earthquake anomalous variations in Earth’s ionosphere and lithosphere were examined in ~800 km radius for two major earthquake episodes in China – M6.0 in Arzak, occurred on 19^th January 2020., and M6.3 occurred in Xizang on 22^nd July 2020. The study has built on a previously conducted Empirical Orthogonal Function and Principal Component Analysis (EOF and PCA), utilizing ESA’s satellite SWARM A, B, and C geomagnetic data. Eight observed significant PC time series were selected for modelling using a LSTM neural network architecture on a three-month and 1-year time scales, each of them is split into training and testing subsets. Strong departure from normal behaviour was noted on 9^th January 2020 in Arzak region, and on 14^th July 2020 in Xizang, corresponding to results previously obtained through EOF and PCA. Several additional anomalous events were observed in a period of two weeks and one month prior to the earthquake events, which further investigations are under way.

China Seismo-Electromagnetic Satellite (CSES-1) is Chinese first satellite that is dedicated to monitoring ionospheric disturbance caused by earthquakes. It transits in a solar synchronous orbit with an altitude of 507 km and revisits the same place every 5 days. In order to more effectively realize the coverage monitoring of Chinese domestic seismic belts, the satellite orbits has been specially optimized for the Chinese seismic belts. The payload of the satellite consists of eight kinds of scientific detection instruments.

Using the revisited orbit design of CSES-1, we analyzed the electron density (Ne) data of 10 orbits from 30 days before the earthquake to 15 days after the earthquake. Before analysis, we take the data that exceeds the mean value by 6 times, greater than and less than or equal to 0 as erroneous data, and replaced them with null values. After replaced the erroneous data, the average value of Ne of the 6 orbits before the earthquake is treated as the background field, which is quite consistent with the 27-day solar cycle. After the background field is obtained, this paper compares the Ne on the day of the earthquake and within 15 days after the earthquake with the background field to try to extract the disturbance signal in the ionosphere which may cause by the earthquake.

The above method was successfully applied to the Yangbi Ms6.4 earthquake in Yunnan on May 21, 2021 and the Qinghai Maduo Ms7.4 earthquake on May 22, 2021. The results show that, about 20 days before the earthquakes, the disturbance signal began to appear in the earthquake epicenter and around the seismogenic area. While the earthquakes approaches, the anomalies appear more and more frequently, and then disappear quickly after the earthquakes.

We apply this method to M ≥ 6 earthquakes in the world and M ≥ 5 earthquakes in the mainland of China and adjacent areas. The statistical results of these earthquakes show that:

1. The number of anomalies increases with the approach of earthquakes, both the global and Chinese regions. And the number of anomalies decreases rapidly after the earthquake.
2. The number of anomalies increased significantly from 10 days before the earthquake to 5 days after the earthquake.
3. With the increase of earthquake magnitude, the number of anomalies increased, and the appearing time of anomalies is also advanced, and the duration of the anomalies after the earthquake is also extension.

The observed InSAR deformation in the Sichuan basin is probably caused by hydraulic fracturing for shale gas production. Some speculations are made based on such deformation patterns. Firstly, the surface deformation could be caused by long-term fluid injection or pumping which lasted several months in poroelasticity medium. Secondly, such deformation may be due to m­ultiple induced seismicities caused by pore pressure diffusion or fluid migration to vulnerable faults. Thirdly, long-term shale gas development could change the underground fluid mass. Loss or gain of fluids would change upper crustal gravity and produce the elastic response of the crust. We test these hypotheses based on numerical analysis of surface deformation patterns. Currently, the poroelasticity effects may exist in many geophysical exploitation activities, including underground water extraction, shale gas development, enhanced geothermal systems, etc. There are two main methods for poroelasticity forward modeling. One is the analytic solution or semi-analytical solution. The other one is the numerical simulation. The former cannot model spatially complicated medium, while the numerical method could approximate the poroelasticity problem of the real stratum as much as possible. Following Rongjiang Wang’s poroelasticity semi-analytical solution, we enhance the accuracy of the numerical method and verify the consistency of the parameters in both solutions. We then make numerical simulations to model the observed InSAR deformation in the Sichuan basin.

The Sentinel-1 mission provides dual-polarization (VV and VH) images for free, but only the VV channel is widely used for deformation measurement due to its larger amplitude value. Through optimizing a cost function related to one pixel selection criteria, polarimetric persistent scatterer interferometry (PolPSI) offers us a tool to combine both VV and VH channels as one optimum channel. Like other single polarization channels, the optimum channel can be used to implement practical applications including deformation monitoring. In order to analyze how the VH channel helps improve the measurement results, two experiments over Barcelona and Alcoy in Spain were carried out. In these experiments, we use amplitude dispersion (DA) as the pixel selection criteria and employ coherent pixels technique (CPT) as PSI processing method.

For the physical interpretation of the optimization process by using the VH channel, a test site comprising structures with diverse geometrical features and orientations was selected. In many cases, the amplitude of VH channel is smaller than that of VV channel, but the DA value is improved thanks to the more stable amplitude provided by the VH channel, which allows PolPSI to select many additional pixels with good phase quality. For two Sentinel-1 datasets acquired from 2017 to 2021, the PS density in the optimum channel increased by around 130% compared with VV channel (under the condition: DA<0.25). Finally, the additional PSs with stable phase increase the coverage of the measurement area and the pixel linking network in CPT. Taking the experiment in Alcoy as an example (a city with landslides and consolidation settlements over small areas), results given by the optimum channel are more accurate than the ones provided by VV when compared with the available in-situ displacement data. All these results support using PolPSI to combine VV and VH channel for a better displacement measurement.

Active landslides pose a significant hazard in mining areas given their considerable potential to induce slope failures, which typically affect open pits and waste and tailing disposal facilities. In order to minimise the impact caused by slope failures in mining areas, much effort has been devoted in recent decades to the development of new approaches to obtain and update active landslides inventory maps with a particular focus on those approaches based on remote sensing. This work illustrates the potential of exploiting satellite InSAR and airborne LiDAR data, combined with data inferred through safety factor maps, to obtain and update inventory maps of active landslides in mining areas. The proposed approach is illustrated by analysing the region of Sierra de Cartagena-La Union (Murcia), a mountainous mining area in the southeast Spain. Firstly, we processed Sentinel-1 InSAR imagery acquired in both ascending and descending geometry during the period between October 2016 and November 2021. The obtained ascending and descending InSAR datasets were then post-processed to semi-automatically generate two active deformation areas (ADAs) maps. Subsequently, both two InSAR datasets were used to decompose the 2D LOS displacement into vertical and east-west components. Complementarily, open-access and non-customized LiDAR data were used to analyse surface changes. Furthermore, safety factor (SF) was calculated over the study area adopting an infinite slope stability model. Finally, the obtained InSAR-derived maps, the LiDAR-derived maps, the original inventory map and the SF map were jointly analysed to create a new active landslide inventory map. In a further step, the influence of rainfall on the activity of the mapped landslides was studied by analysing of the InSAR time series. The results highlight the effectiveness of different remote sensing techniques (i.e., InSAR and LiDAR) jointly with classical methods for slope stability evaluation to update inventory maps of active landslides in mining areas.

Multi-dimensional, long-term time series displacement monitoring is crucial for generating early warnings for active landslides and for mitigating geohazards. The synthetic aperture radar (SAR) interferometry method has been widely applied to achieve small-gradient landslide displacement monitoring; however, measuring the landslide displacement with a steep gradient has posed certain challenges. In comparison, the SAR offset tracking method is a powerful tool for mapping large-gradient landslide displacement in both the slant-range and azimuth directions. Nevertheless, there are some limitations in the existing SAR offset tracking approaches: (i) the measurement accuracy is greatly reduced by the extreme topography relief in high mountain areas, (ii) a fixed matching window from expert experience is usually adopted in the calculation of cross-correlation, (iii) estimating the long-term displacements between the SAR images from cross-platforms and with longer spatiotemporal baselines is a challenging task, and (iv) it is difficult to calculate the three-dimensional (3D) landslide displacements using a single SAR dataset. Additionally, only a few studies have focused on how to realize early warning of landslide deformation using SAR measurements. To address these issues, this paper presents an improved cross-platform SAR offset tracking method, which can not only estimate high-precision landslide displacements in two and three dimensions but also calculate long-term time series displacements over a decade using cross-platform SAR offset tracking measurements. Initially, we optimize the traditional SAR offset tracking workflow to estimate high-precision ground displacements, in which three improvements are highlighted: (i) an “ortho-rectification” operation is applied to restrain the effect of topography relief, (ii) an “adaptive matching window” is adopted in the cross-correlation computation, and (iii) a new strategy is proposed to combine all the possible offset pairs and optimally design the displacement inversion network based on the “optimization theory” of geodetic inversion. Next, robust mathematical equations are built to estimate the two-dimensional (2D) and 3D long-term time series landslide displacements using cross-platform SAR observations. The M-estimator is introduced into the 2D displacement inversion equation to restrain the outliers, and the total least squares criterion is adopted to estimate the 3D displacements considering the random errors in both the design matrix and observations. We take the Laojingbian landslide, Wudongde Reservoir Area (China), as an example to demonstrate the proposed method using ALOS/PALSAR-1 and ALOS/PALSAR-2 images. The results reveal that the proposed method significantly outperforms traditional methods. We also retrieve the movement direction of each pixel of the Laojingbian landslide using the proposed method, thus allowing us to understand the fine-scale landslide kinematics. Finally, we capture and analyze the acceleration characteristics of the landslide, perform an early warning of hazard, and forecast the future displacement evolution based on the 3D displacement time series coupled with the physical models of the rocks.

Since the 21st century, the urbanization of human living environment has been accelerated, and a large number of various bridge facilities have emerged. With the increase of operation time and daily load, some bridges have started to experience different degrees of settlement, deformation, cracks and uplift, which seriously affect the safety of daily use of bridges. Therefore, the use of a reliable technology for periodic bridge deformation monitoring is of great importance to prevent public casualties and property damage caused by bridge collapse.

Compared with traditional contact monitoring means (GPS, level, etc.), which have the shortcomings of long monitoring period and are easily affected by the environment, InSAR technology is a non-contact monitoring means, and the monitoring of bridges, tall buildings and other infrastructures by InSAR technology has the characteristics of all-weather detection, high accuracy, low cost, and does not affect bridge operation. High-resolution SAR data can be applied to bridge fine deformation monitoring work with the advantages of higher monitoring point density and sensitivity.

This research plan is based on a five-span large-span continuous box girder bridge with variable cross-section in Shenyang, Liaoning Province - Xinlipu Bridge over Hun River. The data sources used are 30 images from March 2015 to April 2017 provided by TerraSAR-X satellite and 29 images from August 2015 to June 2017 provided by COSMO-SkyMed satellite, and the data set is processed by SBAS-InSAR technique to obtain the deformation information in the LOS direction of the bridge. The least-squares linear fitting method is applied to extract the temperature influence factor by combining the structural characteristics and material properties of the bridge, and to construct a bridge thermal dilation model to separate the thermal dilation and trend deformation of the bridge. The bridge deformation is the result of the combined effect of periodic thermal dilation and linear trend-type deformation, so separating the thermal dilation from the trend deformation can help us better study the deformation characteristic mechanism of the bridge. Then the multi-source LOS directional thermal dilation is combined with the bridge structure and sensor geometry parameters, based on the natural neighborhood interpolation method, to obtain the bridge along the bridge directional thermal dilation field. Based on the time and space interpolation methods and the principle of singular value decomposition, the LOS trend deformation obtained from the multi-source SAR data is geometrically aligned, interpolated and fused to solve the bridge deformation along the bridge and vertical deformation. Finally, for the visualization of the deformation of the complex structure of the bridge, the deformation points of different structural parts of the bridge are separated and extracted, so that the deformation patterns of different structural parts of the bridge can be better analyzed.

The results show that the thermal dilation of continuous box girder bridges is very obvious, and the method of constructing a bridge thermal dilation model based on the least squares method to extract the temperature influence factors of the bridge monitoring points can separate the periodic thermal dilation from the long-term trend deformation of the bridge. Based on various temporal and spatial interpolation methods, the multi-source SAR data fusion method applying the principle of singular value decomposition can obtain accurate bridge vertical and longitudinal deformation information, and the results show that the main span of Xinlibao Bridge has obvious vertical deformation in the middle of the main span, and the secondary span also has vertical and longitudinal displacement. InSAR technology can be used as a conventional deformation monitoring tool to extract and analyze the time series deformation of various bridges, which provides a reliable technology and data support for bridge health inspection.

In this paper, facing the demand of volcanic activity analysis in Tianchi, Changbaishan, the existing time-series InSAR deformation monitoring method and volcanic point source model are improved, and a set of volcano monitoring scheme suitable for Changbaishan is proposed. Firstly, to address the problem of high vegetation coverage and deformation monitoring being greatly affected by vegetation decoherence, a time-series InSAR deformation monitoring method based on normalized difference vegetation index (NDVI) constraint is proposed. Based on 33 Envisat ASAR images between 2004 and 2010 and 19 ALOS PALSAR images between 2018 and 2020, the accurate surface deformation parameters of the Changbaishan Tianchi crater and the surrounding area were extracted using the small baseline subset technique (SBAS-InSAR). Due to the lack of level data between 2018 and 2020 for comparison, the surface deformation parameters between 2018 and 2020 were also extracted using the persistent scatterer technique (PS-InSAR). The two sets of results were cross-validated and analyzed together with the seismic activity data of the same period. Secondly, we systematically analyzed the three-dimensional geometric relationship between the volcanic surface deformation field and the radar line of sight direction, established a generalized projection conversion equation from the horizontal and vertical deformation of the volcano to the LOS direction, improved the original point source model based on the horizontal and vertical deformation respectively to a point source model based on the LOS direction deformation, and inverted the magma chamber parameters for each time period of Changbaishan Tianchi volcano. Finally, based on the inversion results of the improved point source model, the surface deformation field of Tianchi volcano was orthorectified. The orthorectified results were compared and analyzed with seismic monitoring and fluid geochemical monitoring data to accurately assess the changes of magma chamber of Tianchi Volcano, and to explore the process of volcanic activity in Tianchi, which changed from strong to weak around the end of the disturbance period and gradually became active in the last two years. The results of this paper show that the Tianchi volcanic magma chamber first experienced a brief expansion between 2004 and 2010, with the 3.7 earthquake on September 8, 2004 as the turning point, and then began to enter a fluctuating gradual contraction after the earthquake until it stabilized in 2008. The volcanic magma chamber of Tianchi showed a fluctuating gradual expansion state between 2018 and 2020, and the whole change process was cyclical, with extreme values of deformation once the summer season. Similarly, the temporal deformation of PS-InSAR also has a cyclical trend, which is consistent with the results of SBAS-InSAR.

搜索

复制

In recent years, as the mining scale of open-pit mines continues to expand, a large number of high and steep slopes are formed, leading to increasingly serious slope instability disasters. Landslide disasters are complex, concealed and hazardous, and it is difficult to grasp the accurate deformation development law. Therefore, it is very important to carry out large-scale, long time series and high-precision dynamic monitoring of the slopes of open pit mines to ensure the safe production of mines.

Traditional deformation monitoring technology has the disadvantages of low efficiency, small monitoring range, high labor cost, and inability to obtain a wide range of monitoring data. Therefore, based on the demand for efficient, accurate and near real-time landslide disasters monitoring technology, the interferometic synthetic aperture radar (InSAR) is widely used in the field of landslide monitoring for its advantages of short revisit period, high measurement accuracy, low weather influence and large monitoring range. For large complex landslides with fast sliding, the interferometry technique based on phase information is plagued by the phase unwarpping and is only applicable to slowly deforming landslides with small deformation gradients. However, the Pixel Offset-Tracking (POT) technique based on SAR amplitude information is not affected by the phase unwarpping and space-time decoherence problems, and can overcome the limitation that InSAR can only acquire one-dimensional deformation and measure two-dimensional deformation in azimuth and line of sight (LOS) simultaneously. Under the high-resolution data condition, the deformation solution accuracy of POT technology can reach the decimeter level.

In this paper, a total of 34 scenes of Cosmo-SkyMed SAR data from June 4, 2014 to December 18, 2016 were acquired to monitor the landslide of Fushun West Open Pit Mine using the time-series SAR offset tracking technique, analyze the development pattern of the open pit slope deformation and study its deformation evolution mechanism. In addition, this paper analyzes the degree of influence of three influencing factors, namely, search window, oversampling factor and step size, on reliable pixel point extraction by setting up several groups of comparison experiments, and then selects the most suitable parameters for the analysis of temporal offset slippage deformation by considering the running time and experimental effect. Finally, the obtained monitoring results were verified with GPS observation data, and the comparison results existed high consistency, which further verified the high feasibility and applicability of the pixel offset tracking method in the application of large complex multi-gradient landslide monitoring, and the research results have important reference significance for the slope stability monitoring of Fushun West Open Pit Mine.