Conference Agenda

Sea ice plays an important role in the Earth's climate system, accurately identifying and monitoring melt ponds provides important information for understanding the sea ice evolution process. This study aims to identify the melt ponds in the Canadian Arctic Archipelago and estimate theri depths. To achieve this, Landsat-8 TOA data and ICESat-2 data were used. A multi-layer neural network and a multi-layer perceptron were adopted to successfully achieve accurate classification and depth estimation of melt ponds based on Sentinel-2. Meanwhile, the spatiotemporal variations of melt pond coverage and depth in the Canadian Arctic Archipelago in the last nine years were analyzed.

The near-instantaneous radial velocity of a target can be obtained using the Doppler effect of SAR, which is widely used in ocean current retrieval. However, in sea ice drift velocity measurements, only a Doppler centroid estimation algorithm in frequency domain has been studied, so whether there is a better algorithm is worth exploring. In this study, based on Sentinel-1A IW data, three Doppler centroid estimation algorithms applied to ocean current retrieval are selected. Combined with the characteristics of the TOPS mode, made two applicability adjustments to each algorithm, and finally applied the three algorithms to sea ice radial surface velocity measurements. The first adjustment is to explore and determine the optimal parameters. The second adjustment is to use parallel computing to improve the efficiency, which is improved by an average of 43.55%. In addition, the deviation of Doppler centroid estimation bias correction is verified using rainforest data, and the deviation is controlled at approximately 3 Hz. Based on the three algorithms, five sets of experiments are carried out in this study. By analyzing and comparing the results of each algorithm, it is found that the results of the three algorithms are relatively consistent, among which the correlation Doppler estimation algorithm has the advantages of high efficiency and high precision and is the most suitable method for sea ice drift measurement among the three methods. However, for SAR images with abnormal speckles caused by human activities, the sign Doppler estimation algorithm can effectively remove abnormal speckles and ensure the smoothness of the image with better adaptability.

The China-France Oceanography Satellite SCATterometer (CSCAT) can observe radar backscatter values on the same sea surface at multiple incidence angles, and is often used to estimate the ocean near-surface wind. However, CSCAT utilizes a novel scanning mechanism and the wind vector cell has a spatial resolution is 25km or 12.5 km, which limit the study of high-resolution land and sea ice monitoring. To address this issue, this paper constructs a geometric model of the main lobe-to-ground projection relationship and generates the enhanced-resolution radar images. CSCAT data are applied to three main image reconstruction algorithms (SIR, AART, and MART), and experiments are performed in the Iceland and Hudson Bay, and verified by Sentinel-2 optical remote sensing data. The experiments show the geometric model for CSCAT improves the spatial resolution from traditional 25km to 5 km, and the SIR-reconstructed images are characterized by higher accuracy and better suppression of noise than are those obtained with the AART and MART methods. Therefore, this study extends the application of domestic remote sensors and provides data support for high-resolution applications, such as land and sea ice monitoring.

Images from satellite Synthetic Aperture Radar (SAR) systems are widely used for iceberg monitoring. Icebergs can be detected in SAR images if the difference (the “contrast”) between the backscattered radar intensities from an iceberg and from the surface around it is statistically significant. In our presentation, we focus on sea ice surfaces. For test sites from the Northern Hemisphere (Belgica Bank, Northeast Greenland) and the Southern Hemisphere (Prydz Bay), we manually identified icebergs and determined the backscattering coefficients averaged over the iceberg area and over an area of the sea ice around it. For Belgica Bank, we used ALOS-2 PALSAR-2 (L-band) ScanSAR, and Sentinel-1 (C-band) extra wide swath imagery. For Prydz Bay, we used ALOS PALSAR quad-polarimetric and Radarsat-2 dual-polarimetric SAR imagery. We found that the intensity contrast depends on the radar frequency, the incidence angle, and the sea ice surface characteristics. In our poster, we will present examples and summarize the results for each of our test sites. The findings are valuable for developing strategies and algorithms for automated iceberg detections as required by operational sea ice and iceberg monitoring services, considering the use and combination of recent and upcoming SAR satellite missions.

The Bohai Sea and its surrounding areas are rich in oil and natural gas, and play an important role in the industry, agriculture and economy. However, the Bohai Sea suffers severely from the sea ice in the winter. The Geostationary Ocean Color Imager (GOCI) is the first geostationary orbit ocean color satellite, providing high spatial and temporal resolution for extraction of sea ice parameters in the Bohai Sea. Based on GOCI data, a systematic and standardized method is developed for extracting sea ice parameters. This method can perform normalized preprocessing on the GOCI raw data, including atmospheric correction, relative radiation correction, and sea ice or cloud masking. Subsequently, it extracts relevant sea ice parameters, including sea ice concentration, sea ice thickness, and sea ice drift velocity. The unique advantage of GOCI is its geostationary orbit and short imaging interval (1 hour), which enables tracking the daily drift of the sea ice in the Bohai Sea. Using this method, sea ice parameters are retrieved in the Bohai Sea in winter from 2017 to 2021, and the retrieval accuracy meets the sea ice forecast demand. Finally, we extract a long time series dataset of sea ice parameters from 2011 to 2020 (December to March of the following year), and conduct a statistical analysis of the long-term sea ice changes in the Bohai Sea, which is consistent with the information formally released by the State Oceanic Administration. The sea ice extent and thickness in the Bohai Sea reached their maximum in 2012 and their minimum in 2019, respectively. The sea ice growth during each winter follows the same pattern: the sea ice forms in late December, reaches its maximum extent in January, begins to shrink in early February, and disappears completely by early March. The sea ice drift velocity is largely influenced by the wind and currents, without significant rules of inter-annual or annual changes. The extraction of these parameters will provide initial field data of the sea ice for sea ice forecasting in the Bohai Sea. Furthermore, it will provide valuable data support for sea ice monitoring and ocean environmental research, helping to better understand the trends in oceanic changes and ultimately contribute to the preservation of the health and stability of marine ecosystems.

Sea ice in the Yellow Sea and Bohai Sea affects maritime transportation and economic activities every winter. Hence, monitoring sea ice concentration and thickness, the key parameters, is vital. HY-1C and HY-1D are the ocean color satellite series that provide optical data in the morning and afternoon, respectively. In the afternoon, rising temperatures may cause slight melting on the sea ice surface, which may hamper optical detection. Therefore, HY-1C is more suitable for Bohai sea ice monitoring than HY-1D. Its onboard Chinese Ocean Color and Temperature Scanner (COCTS) has ten spectral bands for retrieving sea ice concentration and thickness. This study proposes a systematic and standardized method for extracting sea ice parameters based on HY-1C data. The raw COCTS data undergoes normative pre-processing, which includes geometric correction, atmospheric correction, radiometric calibration, and sea ice masking. Then, sea ice concentration and thickness are retrieved. For sea ice thickness, the linear correlation between MODIS shortwave broadband reflectance and HY-1C band reflectance is analyzed. Then, a linear regression equation is established between MODIS shortwave broadband reflectance and HY-1C band reflectance to obtain shortwave broadband reflectance from HY-1C data. Subsequently, based on the theoretical model of sea ice thickness and shortwave broadband reflectance, the Bohai Sea ice thickness is calculated. Sea ice concentration is extracted using the shortwave reflectances of sea ice and sea water. Three methods are used to calculate the shortwave reflectance of sea water: standard, mean, and direct assignment. Two methods are used to calculate the shortwave reflectance of the sea ice: the standard method and mean method. Six sea ice concentration results from these method combinations are obtained and compared. The comparison shows that using the direct assignment method for sea water shortwave reflectance and the standard method for sea ice shortwave reflectance yields the most accurate results relative to the original image. Hence, this approach is adopted for sea ice concentration extraction. Using these methods, we have monitored sea ice in the Yellow Sea and Bohai Sea from 2021 to 2023. This project provides a systematic and standardized method for inverting sea ice thickness and concentration based on HY-1C data. It provides initial fields of sea ice parameters for sea ice forecasting in the Yellow Sea and Bohai Sea, which is vital for shipping, transportation, and resource development.

Lakes comprise approximately 1.8% of the Earth's surface, and up to 40%-50% in certain areas of the Arctic and subarctic regions. Frozen lakes represent approximately 59.55% of the total lake area in the Northern Hemisphere, as determined by the January 0°C isotherm. Lake ice serves as a key Environmental Climate Variable (ECV) in the Global Climate Observing System (GCOS), where ice extent, phenology, thickness, and type are crucial indicators for assessing climate change and ecological research. However, global warming is causing a decline in ice coverage, with delayed freeze-up dates and earlier ice breakup dates. The reduction of lake ice has significant implications for lake ecosystems, including biodiversity, biogeochemical processes, and greenhouse gas emissions.

Satellite remote sensing has become a widely employed tool for lake ice monitoring owing to its broad spatial coverage, frequent observation cycles, and high precision. The optical remote sensing data acquired by the Moderate Resolution Imaging Spectroradiometer (MODIS) features high spatial resolution and enables bi-daily monitoring. In contrast, passive microwave data remains unaffected by weather and cloud cover, and both data types present unique advantages for large-scale lake ice monitoring. Accordingly, we leveraged both MODIS optical remote sensing and passive microwave data to investigate lake ice in the Northern Hemisphere.

We utilized MODIS NDSI data to monitor lake ice for over 23,000 lakes across Eurasia.We applied a series of cloud removal methods to process the data and effectively reduce the cloud cover. Using the cloud-free MODIS data, we extracted the long time series of lake ice coverage data from 2002 to 2022. The dataset was verified to have high accuracy and can effectively monitor the changing trends of lake ice.To classify lakes with different lake ice cover trends, we have developed a convolutional neural network-based method for time series classification of lake ice cover. This method effectively categorizes lakes into four different types.

Based on passive microwave data, we employed the nearest neighbor algorithm to reduce the impact of mixed pixels and extracted brightness temperature data for 753 lakes in the Northern Hemisphere from 1978 to 2020. We then derived corresponding lake ice phenological parameters and verified the results to have a high degree of accuracy.Based on the analysis of the dataset, the following results were obtained: lakes freeze earlier, melt later, and have longer ice periods as latitude increases. Lakes located between 28°N and 40°N have longer ice cover durations compared to those located north of 40°N, mainly due to the prevalence of lakes at low latitudes on the Qinghai-Tibet Plateau. Above 45°N, at the same latitude, the average ice cover duration of North American lakes is longer than that of Eurasian lakes.

Meanwhile, we analyzed the changes in lake ice phenology of lakes in Northern Europe, Qinghai-Tibet Plateau, and Mongolian Plateau using passive microwave data, and investigated their associations with climate.The results showed that the lake ice changes in the three regions were significantly correlated with the corresponding regional temperature changes. Among them, in Northern Europe, the temperature change had a more sensitive impact on the lake ice phenology. There were still other factors that influenced the lake ice changes. However, in the northern region of the Tibetan Plateau, the ice period of many lakes has increased since 2000. There are many factors contributing to this phenomenon, such as the decrease of Kara Sea ice ，the winter North Atlantic Oscillation (NAO) and early spring Antarctic Oscillation (AAO) anomalies.

Glacier surface albedo is one of the most important parameters to determine the net shortwave radiation and therefore affect glacier energy and mass balance. Glaciers in the High Mountain Asia (HMA) are the main water resource for local ecology and local people (~one billion), and have accelerated mass loss in the past 20-years. Due to a good sensitivity to climate, better understanding of the spatiotemporal variability of glacier albedo can help us to understand the mass balance and the response of glacier to climate in this region. With our retrieval method developed for Sentinel, Landsat and MODIS data, we firstly generated half-monthly glacier albedo by MODIS surface reflectance data, and then analyzed its change from 2001 to 2020. The results show that the glacier albedo experienced a decline over the entire region, but with a distinct spatial and seasonal differences. In the westerly-dominated regions, glacier albedo shows a slight decrease even increase in the Hindu Kush and West Himalaya, while in the monsoon-dominated and transition regions, it shows large decrease with the rapidest change in the Inner Tibetan Plateau. Autumn albedo shows the quickest decrease, while the lowest is observed in spring. Good correlation between albedo and mass balance indicates that decreasing albedo is indeed a key driver of mass loss in this region.

Mountains act as water towers, supplying crucial freshwater to downstream areas and affecting large populations particularly in High Mountain Asia. Yet, the propagation of water from HMA headwaters to downstream areas is not fully understood, as interactions in the mountain water cycle between the hydrosphere and biosphere remain elusive. Understanding how green water processes affect the availability of blue water from glaciers, snow and precipitation in High Mountain Asia is a pressing but challenging research question. Cyrosphere and biosphere dynamics are manifest in distinct manners across the extreme elevation range of catchments in High Mountain Asia due to the intra-annual variability of climate and associated ecosystems. It is therefore imperative to couple our understanding of blue and green water fluxes, which traditionally have been studied in an isolated way, and to examine these fluxes sub-seasonally and with elevation.

Land surface models are numerical models that account for these blue-green fluxes in a complete manner by solving the coupled fluxes of water, energy, and carbon between the land surface and atmosphere. However, most land surface models focus at the global or regional scale with horizontal grid dimensions of 0.25–1° (equivalent to about 25–100 km at mid-latitudes), and are thus not able to resolve land-surface energy and water fluxes at sufficient spatial detail to capture local topographic and microclimatic effects or lateral flows of water, as found in complex mountainous topographies such as glacierized watersheds in HMA. Due to the lack of observations or computational constraints, land surface models usually focus on specific processes and neglect the links between the cryosphere, hydrosphere and biosphere, or represent them in a simplistic way. This is problematic, for example because plant transpiration forms a major part of the green water flux even in high mountain areas with scarce vegetation, but the large variability in water-use strategies between different plants hinders a quantification based on simple parameterizations. Similarly, mountain glaciers, let alone debris-covered glaciers which are common in HMA, have been neglected in land surface models so far.

High resolution meteorological forcing information is usually less robust for mountain regions as station data are available at a few research sites only. Downscaling methods have seen improvements but either suffer from the lack of station data needed for statistical downscaling or from computational resources needed for dynamical downscaling.

Given the dramatic lack of high-elevation in-situ data in HMA, and the general difficulty of capturing land-atmosphere interactions in complex topographies well with field measurements, remotely sensed data of high spatiotemporal resolution offer a great opportunity to develop, calibrate and test land surface models, while reducing uncertainties in model initialization, simulation, and validation.

The increasing resolution and accuracy of remote sensing data, and a new generation of models representing the cryosphere, hydrosphere and biosphere within one modeling system and with the highest degree possible of physical representation, bring the possibility of a paradigm shift in the simulation of blue-green water interactions in high mountain catchments. As an example for an integrated approach to reveal blue-green water fluxes in a high mountain region we show how we apply a state-of-the-art land surface model (Tethys & Chloris) to the glacierized Langtang catchment in the Nepalese Himalayas and explain the use of high resolution earth observation data (e.g. glacier thinning and surface motion; glacier albedo; snow cover) to constrain the meteorological uncertainty and validate our model results.

High Mountain Asia (HMA) hosts the largest mass of ice outside the Polar Regions and provides water to large downstream communities. Glacier change has been highly diverse across the region over the last decades, with glaciers in the Pamirs experiencing near-neutral mass balance while fast rates of mass loss are observed in the Southeastern Tibetan Plateau (SETP). In a previous modeling study in the SETP, we found that precipitation phase changes associated with climate warming were a major accelerator of glacier losses, but this mechanism of mass loss acceleration has yet to be explored across the rest of HMA. Additionally, snow sublimation and gravitational redistribution are two processes known to influence glacier mass supply, but their relevance has not been systematically investigated at the catchment scale at distinct locations across HMA.

Here we apply a mechanistic land-surface model at high spatial and temporal resolution (100m, hourly) at three glacierized catchments covering distinctive climates in HMA (Kyzylsu in the Northern Pamirs, Trakarding-Trambau in the Nepalese Himalayas, and Parlung No.4 in the SETP). We force the model with ERA5-Land reanalysis which was downscaled and bias-corrected with locally available meteorological observation. We constrain and evaluate our model with independent in-situ observations (ablation stakes, snow depth measurement) and remote-sensing observations (snow cover, surface elevation changes, glacier surface mass balance and albedo). Our goal is to quantify the importance of solid precipitation, snow sublimation, and gravitational snow redistribution on the glacier mass balance and in the catchment water balance. Our first modeling results highlight the challenges but also the added value of applying such sophisticated models in these remote areas characterized by extreme topography and scarce or altitudinally-biased local observations. We show how the choice of the precipitation phase scheme influences the seasonality of the simulated snowfall amounts and the overall glacier mass balance. We discuss the limitations associated with the use of reanalysis datasets and ways forward to better account for the spatial variability of key meteorological variables.

This work paves the way towards a better understanding on how snow accumulation processes will be affected by climate change and what the implications will be for glacier future evolution and high-elevation catchment hydrology.

Mountain glaciers are sensitive to climate variability and can be of great importance for downstream residents due to their hydrological significance. Synthetic Aperture Radar images are often used to monitor glaciers based on the backscatter coefficient, but the influence of satellite orbit and polarization when collecting images for wide regions were not well considered. We study the changes of wet snow in summer and firn in winter in West Kunlun Shan and the Tibet Interior Mountains by using Sentinel-1 C-band data acquired in the summer 2019 and winter 2019/20. We found that there is a clear threshold for the backscattering coefficient in the glacier area after using the maximum likelihood classification, and using this threshold allows monitoring of both wet snow and firn. Images from ascending and descending may differ greatly in summer for wet snow detection. This effect can be related to the orbit and therefore the different acquisition time and different air temperature in the morning and afternoon. Using the proposed method, we show that West Kunlun Shan has lower wet- snow-area ratio, but higher firn-area ratio than the Tibet Interior Mountains. In general, orbital produce greater identification differences than polarization.

In recent decades, the most significant mass losses in the Antarctic Ice Sheet (AIS) have been found to be driven by ocean-forced basal melting reducing the buttressing ability of ice shelves. The Ross Ice Shelf (RIS) is the largest cold water ice shelf on the AIS and buttresses both the West and East Antarctic Ice Sheet. Understanding the current dynamics of the RIS in a warming world is important as the ice shelf has a large control over the mass balance of the AIS. The RIS has been suggested to be in steady state but recently seasonal changes in sea ice cover have been found to elevate basal melt rates at the calving front of the RIS (Stewart et al., 2019). Understanding of the influence that short-term environmental variability, such as seasonal basal melt rates, have on the RIS dynamics and mass loss is not yet fully understood. In this project, further understanding is achieved through observing the RIS flow rates over seasonal and annual timescales using GNSS and satellite methods at different locations on the ice shelf. Quantifying the variability of the RIS flow rates provides critical information on the ice dynamics and how these could change in a warming world. Sensitivity experiments are also carried out using the Ice-sheet and Sea-level System Model (ISSM) to understand which short-term environmental forcings may be driving the observed velocity variations, and how these may impact the mass loss of the RIS.

China is the main contributor to global vegetation greening, and large-scale greening has been proven to be related to afforestation. However, with the rise in global temperatures, climate change has become an undeniable factor affecting regional vegetation changes. It is necessary to quantitatively evaluate the relative contributions of climate change and afforestation to China’s vegetation greening, and evaluating the vegetation recovery in forestry projects is conducive to future policy formulation and response to climate change. This study is based on meteorological observation data and satellite remote sensing data. Firstly, the greening of vegetation in China and eight forestry projects from 1982 to 2020 was monitored and evaluated. Then, the relative contributions of climate and afforestation initiatives to vegetation greening were quantitatively evaluated, and the future vegetation greenness change was predicted on this basis. The main research results are as follows:

During the study period, vegetation in China has significantly increased. Pixels with increasing trends accounted for 57% of the region, pixels with stable or unchanged trends accounted for 27% of the region, and pixels with decreasing trends accounted for 16% of the region. The pixels with a significant increase trend are mainly distributed in the Loess Plateau, Northeast Plain, and South China region, while the pixels with a significant decrease trend are mainly distributed in the Qinghai-Tibet Plateau and Northeast region. Due to differences in land use types, climate conditions, and topographic conditions in different regions, there are differences in the ecological implementation effects of the eight forestry project areas, and vegetation degradation is still relatively obvious in some forestry engineering areas.

Climate change is the main factor affecting the recovery of vegetation in China. The contribution rates of climate change and human activities to vegetation recovery are 72.34% and 27.66%, respectively. In arid and semi-arid areas such as the Mongolian Plateau, Qinghai-Tibet Plateau, and Loess Plateau, precipitation is crucial for vegetation growth. Temperature has a significant promoting effect on vegetation growth in the southeast region because the region has abundant precipitation resources and higher temperatures are conducive to regional vegetation growth.

Only 14% of the regions with continuous NDVI growth are expected to continue to grow in the future, and the remaining regions show obvious anti-continuity (59% from increase to decrease, 22% from decrease to increase). The risk of vegetation degradation in the future is high. The impact of climate factors on vegetation is gradually weakening, while the impact of human activities on vegetation changes will become more complex. Although ecological engineering has played a positive role in the restoration of vegetation ecosystems, vegetation degradation in the Three North Shelterbelt Program (TNSF), Coastal Shelterbelt Program (CSP), and Shelterbelt Program for Liaohe River (SPLR) are still relatively obvious. This is related to the fragile regional ecological environment and the destruction of vegetation by agriculture, animal husbandry, and urbanization. Therefore, it is necessary to further strengthen the construction of ecological engineering to better maintain the effectiveness of these projects.

Lakes in the basin of the Yangtze River, play a fundamental role in regional bio-geochemical cycles and provide major services to the communities, provisioning services (drinking water, fishing) and biodiversity keeping. However, the extreme temporal and spatial variability of these massive but extremely shallow ecosystems prevents a reliable quantification of their dynamics with respect to changes in climate and land use. The final aim is to model, map and explain the distribution of biodiversity and their associated habitats, explaining spatio-temporal changes in biodiversity caused by biotic and abiotic factors. Within this dragon 5 project ID 58815, sensitive areas having a rich biodiversity including the East Dongting lake, Hunan Porvince (Xiaoxi, Daxi, Caisang,) and the disconnected lakes of the Anhui Province (Wuchang, Shengjin and Baiding Lakes) are considered.

For the epoch 2019-2023, by exploiting our house tool ExtractEO which is a software implementing automated end-to-end chains, water surfaces were detected over Sentinel-2 data using a multilayer perceptron algorithm and integrating the Global Surface Water database for sampling. Sentinel-2 water extents were generated from the Sentinel-2 time series and then densified by exploiting RADARSAT-2 and ICEYE SAR imagery. Validation of the processing chain was done by comparing water surface derived from S2 with the one obtained from a Pléiades NEO imagery with a resolution of 30 cm. Water levels were also monitored by exploiting Sentinel-3 altimetric data and validated by comparison with ICESat-2 known for its high precision.

Results obtained over the sensitive Anhui and Hunan province lakes, will be presented and discussed. Based on these preliminary results, guidelines for further investigation particularly for SWOT data exploitation will be presented.

During November 2022, Poyang Lake suffered from a severe drought disaster, and the water level at Xingzi Station receded to 6.48 meter, which set a new record low water level. In order to explore the impact of this extreme drought event on the hydrological patterns of Poyang Lake, we constructed a dataset of the water area in different periods by utilizing Sentinel-1 Synthetic Aperture Radar (SAR) images, with the advantages of high spatial–temporal resolution and all-day and all-weather working capacity. The relationship model between lake area and water level was constructed based on the data from hydrological stations in Poyang Lake. We found that the water level and water area showed strong correlation in recent years, especially at Xingzi station (R²=0.88). Therefore, we can make an early warning of the overall drought condition of Poyang Lake through the real-time water level of Xingzi Station, especially the change of food and environment of migratory birds' habitats. For purpose of assessing the drought disaster in Poyang Lake more accurately, we carried out the research on the precise classification of land cover. Afterwards, the algorithm was applied to estimate the yield of oilseed rape in Poyang Lake. Our research results can provide decision support for the relevant management departments for disaster early warning and assessment of Poyang Lake.