Conference Agenda

Spatio-temporal patterns of Suspended Particulate Matter (SPM) and Chlorophyll-a (Chla) concentrations, have been assessed from the Ocean and Land Color Instrument (OLCI) over the whole European coastal waters from 2016 to 2023. The semi-analytical algorithm of Han et al. 2016 has been used for SPM estimation, while Chla has been computed based on an optical classification approach proposed by Tran et al. 2023, that combines several Chla algorithms. The generated products have been validated using an extensive dataset of in-situ measurements. Chla and SPM climatologies have been generated at the scale of Europe, and the temporal patterns (seasonal variability, long term trend, and irregular component) have been described using the Census-X-11 time series decomposition method.

By utilizing Level 2A products (particle optical properties and numerical weather prediction data) and Level 2C products (numerical weather prediction wind vector assimilated with observed wind component) provided by the Atmospheric Laser Doppler Instrument (ALADIN) onboard the Aeolus mission, and Level 2 vertical feature mask (VFM) products provided by Cloud-Aerosol Lidar with Orthogonal Polarization (CALIOP) onboard Cloud-Aerosol Lidar and Infrared Pathfinder Satellite Observation (CALIPSO) mission, three remote ocean areas are selected and the optical properties at 355 nm of marine aerosol are derived. The combined analysis of marine aerosol optical properties at 355 nm and instantaneous co-located wind speeds above the remote ocean areas are conducted. Eventually their relationships are explored and discussed at two sperate vertical atmospheric layers (0-1 km and 1-2 km, correspond to the heights within and above marine atmospheric boundary layer (MABL)), revealing the marine aerosol related atmospheric background states. Pure marine aerosol optical properties at 355 nm are obtained after quality control, cloud screening and backscatter coefficient correction from the ALADIN observations. The spatial distributions of marine aerosol optical properties and wind speed above the study areas are presented and analysed, respectively, at two vertical layers. The statistical results of the marine aerosol optical properties along with the wind speed grids at two vertical layers together with the corresponding regression curves fitted by power law functions are acquired and analysed, for each remote ocean area. The optical properties present increasing trends with wind speed in all cases, implying that the atmosphere of the two vertical layers will both receive the marine aerosol input produced and transported by the wind and the turbulence. The marine aerosol enhancement caused by the wind speed at the lower layer is more intensive than at the higher layer. As derived data from ALADIN, the averaged marine aerosol optical depth and the averaged marine aerosol lidar ratio at 355 nm are acquired and discussed along the wind speed range. The marine aerosol optical properties distributions, wind speed bins, and the marine aerosol variation tendencies along wind speed above the individual study areas are not totally similar, implying that the development and evolution of the marine aerosol above the ocean might not only be dominated by the drive of the wind, but also be impacted by other meteorological and environmental factors, e.g., atmospheric stability, sea and air temperature, or relative humidity. Combined analysis on the aerosol optical properties and wind with additional atmospheric parameters above the ocean might be capable to provide more detailed information of marine aerosol production, entrainment, transport and removal.

The main objective of the Dragon V project (ID 59198) is to standardize procedures for calibrating European and Chinese satellite altimeters. Calibration and Validation (Cal/Val) actions should follow the guidelines prescribed by the Fiducial Reference Measurements for Altimetry strategy, developed by the European Space Agency for standardizing procedures and results. One of the fundamental quantities that needs to be calibrated in satellite altimetry is the backscatter coefficient (sigma-naught). This is a satellite measurement related to wind observations at sea and constitutes an important and indispensable parameter for climate change models. At the moment, there is no European or Chinese Cal/Val facility dedicated to sigma-naught calibration.

This work presents the progress made in the design, analysis and validation of corner reflectors for the absolute and direct calibration of the backscatter coefficient in satellite altimeters. Requirements and specifications (i.e., material, dimensions, etc.) for manufacturing such corner reflectors have been defined. These are tailored for calibrating Ku and Ka-band satellite altimeters. Finally, the ground location where these corner reflectors are to be installed has been selected because of its low clutter level and capability of calibrating multiple satellites.

High precision magnetometer (HPM) has worked successfully more than 5 years to provide continuous magnetic field measurement since the launch of CSES. After rechecking these years data, it is necessary to make an improvement for fluxgate magnetometer (FGM) orthogonal calibration (to estimate offsets, scale values and non-othogonalities) and alignment (to estimate three Euler angles). The following efforts are made to achieve this goal: For orthogonal calibration, we further considered the FGM sensor temperature correction on offsets and scale values to remove the seasonal effect. Based on these results, Euler angles are estimated along with global geomagnetic field modeling and then the latitudinal effect for east component is improved. After considering above improvement, we can prolong the updating period of all calibration parameters from daily to 10 days, without the separation of dayside and nightside data. These algorithms will be helpful to improve HPM routine data processing efficiency and data quality to support more scientific studies.

We propose a GNSS-R soil moisture (SM) retrieval algorithm based on land surface clustering using the twin satellites A/B of BuFeng-1 (BF-1). Similar to other semi-empirical algorithms, this algorithm incorporates vegetation and roughness parameters. However, it introduces empirical clustering as an alternative to quantitative calculations. Vegetation and roughness, recognized as significant factors influencing GNSS scatter signals, are utilized to categorize the land surface into distinct classes. The opportunity observations of spaceborne GNSS-R presents a challenge in obtaining a sufficient number of valid observations within a grid cell at the theoretical spatial resolution of approximately 3.5 km to 20 km over land. This limitation hampers the establishment of robust empirical relationships. Consequently, our algorithm avoids pixel-by-pixel fitting and instead establishes empirical relationships between SM and GNSS-R observations within each class. A global comparison between the algorithm's results and the 36-km soil moisture product from the Soil Moisture Active Passive (SMAP) mission reveals a correlation coefficient (R) of 0.82 and an unbiased root mean square error (ubRMSE) of 0.070 cm³·cm⁻³.

Quantitative estimation of the Leaf Area Index (LAI) from remote sensing imagery is crucial for monitoring vegetation growth and assessing the ecological environment. Gaofen-6, a high-resolution remote sensing satellite launched by China, offers a valuable tool for vegetation monitoring due to its high spatial resolution and spectral coverage. Accurate LAI estimation from Gaofen-6 imagery can provide essential information for crop management, land use planning, and climate modelling. A number of studies have explored the LAI estimation from Gaofen-6 using machine learning algorithms and have achieved nice results. However, these studies are non-torableable to different locations and hardly applicable to complex vegetation structures. The physically-based Look-up Table (LUT) approach is relay on radiative transfer models (RTMs), it takes into account the fundamental principles of light interaction with vegetation, which can result in more accurate and reliable LAI as well as other vegetation biophysical parameters estimations. So far, to our knowledge, there's no research that has attempted to use the LUT method to invert LAI on Gaofen-6 imagery. Therefore, in this study, we explore the LUT for LAI retrieval on Gaofen-6 and validate the LAI with both in-situ measurements and Drone-based LAI estimations.

In this study, one wide-field view (WFV) scene of Gaofen-6 over Taizishan Forest Park (30.91-30.92°N, 112.87-112.88°E), China is used. The LUT method was implemented in R and applied to the subset of the Gaofen-6. The Gaofen-6-based LAI inversion result achieved a comparable result with Sentinel-2 LAI inversion. The latter has RMSE of 1.02 and 0.59 when evaluated with UAV-based LAI reference map and in-situ measurements, while Gaofen-6 achieved RMSE of 1.49 and 0.89. The estimated LAI value ranges between 0 and 5 in the study area, which is consistent with our prior knowledge and ground measurements.

Overall, our study is one of the few that have implemented a LUT-based inversion approach on Gaofen-6 data. However, due to the lack of ground information, there is a certain gap between the Gaofen-6 LAI map and the UAV-based LAI estimation. In the future, we will continue to supplement the measurement of ground information and seek a method to invert a higher-precision Gaofen-6 LAI map over other study areas like Whythum forest in the UK.

Since a significant scale difference exists between the field of view of ground station sensors and satellite sensors, the validation of satellite-retrieved land surface parameters is usually performed over homogeneous surfaces. However, due to typically inhomogeneous natural surfaces and the urgent need to evaluate satellite-retrieved land surface parameters over a broad range of representative land cover types, it is crucial to be able to evaluate those parameters over inhomogeneous surfaces. In an attempt to address this issue, a temporal variation method for evaluating the spatial representativeness of ground stations was proposed for kilometer-scale LST validation (Ma et al., 2021). In this method, a station’s spatial representativeness indicator (SRI) is defined as the LST difference between the ground radiometer’s FOV and the corresponding satellite pixel. In order to estimate the SRI, which is effectively the temperature difference due to spatial scale, a temporal variation model of SRI is established, which combines the temporal variation of LST and its main influence factors. Meanwhile, according to its definition, SRI can be used as a bridge to convert in-situ LST to the corresponding pixel scale. Therefore, the SRI allows to validate satellite LST against in-situ LST at the same spatial scale.

The method was applied in the validation of MODIS and AATSR LST. Based on Landsat TM/ETM+, the LST within the ground radiometer’s FOV and the corresponding MODIS and AATSR pixel were simulated at 16 Chinese stations. Then the annual variation of LST at the two spatial scales was modeled using the annual cycle model (ATC), from which SRI’s variation tendency ∆ATC was obtained. Using the random forest method, a temporal variation model was constructed for the fluctuation term (∆USC) around ∆ATC, which was based on surface condition parameters and instantaneous meteorological parameters. Results show that when the spatial representativeness of the ground station is ignored, the systematic bias is between -4.05 K and 5.08 K, and the standard deviation of the bias is between 1.11 K and 6.95 K, for MODIS daytime LST. After considering the stations’ spatial representativeness, the systematic bias is between -4.35 K and 1.17 K, and the standard deviation of the bias is between 0.61 K and 6.01 K. Here, the systematic deviation and the corresponding standard deviation are 1.43~5.34 K and 0.35~3.39 K, respectively, due the ground station’s spatial representativeness. For the AATSR daytime LST, when the spatial representativeness of the ground station is ignored, the systematic bias is between -3.57 K and 7.28 K, and the standard deviation of the bias is between 1.26 K and 6.35 K. After considering the stations’ spatial representativeness, the systematic bias is between -2.63 K and 4.36 K, and the standard deviation of the bias is between 0.28 K and 5.07 K. Here, the systematic deviation and the corresponding standard deviation are -1.95~5.6 K and 0.07~3.72 K.

It can be concluded that large systematic deviations and random errors can result from a lack of spatial representativeness of a ground station, which considerably reduces the meaningfulness of the validation results obtained on the satellite pixel scale. Therefore, it is recommended to always analyze and account for the spatial representativeness of ground stations at the satellite pixel scale, e.g. by using the proposed or another established method for validating LST.

Ma, J., Zhou, J., Liu, S., Frank-Michael Göttsche, Zhang, X., Wang, S., Li, M., 2021. Continuous evaluation of the spatial representativeness of land surface temperature validation sites. Remote Sensing of Environment 265, 112669. https://doi.org/10.1016/j.rse.2021.112669

Land surface temperature (LST) is a key variable in a wide variety of studies directly linked to land–atmosphere energy transfer and flux balances, as well as in a broad range of applications such evaporation monitoring, estimates of fire size, detection of volcanic activity, permafrost detection or monitoring of vegetation health. Furthermore, LST is considered by World Meteorological Organization (WMO) and the Global Climate Observing System (GCOS) as one of the essential climate variables (ECVs) for climate change monitoring. However, satellite LST acquisitions are often limited due to cloudy skies. Several methods have been proposed in the literature to estimate the under-cloud LST from thermal and passive microwave data: these are known as all-weather LST products. Thus, all-weather LST products are required for an accurate analysis on climate studies at global and local scale and climate change monitoring.

In this study we investigate the accuracy of an all-weather LST products produced within the Dragon 5 project ’All-weather land surface temperature at high spatial resolution: validation and applications’. The investigated LST product merges clear-sky MSG/SEVIRI LST at a spatial resolution of 5 km with the surface temperature of a Soil-Vegetation-Atmosphere (SVAT) model (Martins et al., 2019). This product is validated over KIT’s permanent validation site on the gravel plains at Gobabeb (Namibia) for years 2010 to 2012. This period includes the largest rainfall at Gobabeb in recorded history, which makes the product retrievals challenging due to the extreme atmospheric conditions but also due to the changes on biophysical surface properties, which are linked to surface emissivity and LST. Thus, the results will provide a comprehensive analysis of the all-weather LST product performance over a broader range of atmospheric and surface conditions.