Wetland Mapping Research Articles

Advancements and opportunities to improve bottom-up estimates of global wetland methane emissions

Abstract Wetlands are the single largest natural source of atmospheric methane (CH4), contributing approximately 30% of total surface CH4 emissions, and they have been identified as the largest source of uncertainty in the global CH4 budget based on the most recent Global Carbon Project (GCP) CH4 report. High uncertainties in the bottom-up estimates of wetland CH4 emissions pose significant challenges for accurately understanding their spatiotemporal variations, and for the scientific community to monitor wetland CH4 emissions from space. In fact, there are large disagreements between bottom-up estimates versus top-down estimates inferred from inversion of atmospheric CH4 concentrations. To address these critical gaps, we review recent development, validation, and applications of bottom-up estimates of global wetland CH4 emissions, as well as how they are used in top-down inversions. These bottom-up estimates, using (1) empirical biogeochemical modeling (e.g., WetCHARTs: 125 - 208 TgCH4 yr-1); (2) process-based biogeochemical modeling (e.g., WETCHIMP: 190 ± 39 TgCH4 yr-1); and (3) data-driven machine learning approach (e.g., UpCH4: 146 ± 43 TgCH4 yr-1), were all subject to large uncertainties (~80 TgCH4 yr-1). Major tropical and arctic wetland complexes are regional hotspots of CH4 emissions. However, the scarcity of satellite data over the tropics and northern high latitudes offer limited information for top-down inversions to improve bottom-up estimates. Recent advances in surface measurements of CH4 fluxes (e.g., FLUXNET-CH4) across a wide range of ecosystems including bogs, fens, marshes, and forest swamps provide an unprecedented opportunity to improve existing bottom-up estimates of wetland CH4 estimates. We suggest that continuous long-term surface measurements at representative wetlands, high fidelity wetland mapping, combined with an appropriate modeling framework, will be needed to significantly improve global estimates of wetland CH4 emissions. There is also a pressing unmet need for fine-resolution and high-precision satellite CH4 observations directed at wetlands.

Deep neural network and transfer learning for annual wetland vegetation mapping using sentinel-2 time-series data in the heterogeneous lake floodplain environment

ABSTRACT The lake-floodplain wetlands are characterized by high biodiversity, difficult access, and significant environmental changes. Traditional remote sensing mapping methods struggle to generate consistent time-series data on wetland vegetation communities. Current research has endeavoured to address this issue through the application of deep learning methodologies. However, a significant limitation of these models is their reliance on a substantial volume of training samples, which contradicts the difficulty and high cost of obtaining samples from the lake-floodplain wetlands. Whether it is possible to construct a transferable deep learning model under small sample conditions and apply it to the mapping of lake-floodplain wetlands is an urgent issue that needs to be addressed. To solve this problem, this study first constructed a deep neural network (DNN) designed specifically for mapping complex lake-floodplain wetland vegetation under conditions of limited sample size. Subsequently, using 2021 as a reference year, a novel histogram threshold method was proposed to identify the unchanged samples for the target transfer years of 2019, 2020, 2022, and 2023. Finally, annual wetland vegetation mapping was performed in Poyang Lake using DNN and sample transfer learning (STL). The results showed that high-quality annual time-series data of wetland vegetation can be generated using the constructed DNN and STL, with all overall accuracies exceeding 80%. The histogram threshold method, which combines SAD and NDVI indicators of key phenological period, can effectively solve the problem of difficulty in determining the unchanged samples in transfer learning for heterogeneous lake wetlands. Furthermore, the performance of STL based on the constructed DNN model was significantly superior to those based on support vector machine and random forest algorithms for mapping annual wetland vegetation communities using limited training samples. This study demonstrates that the effective application of DNN and STL will be highly beneficial for long-term monitoring of vegetation in lake-floodplain wetlands, particularly where sample availability is limited.

Utilizing LISS-4 satellite imagery and support vector machine for mangrove and wetland mapping in part of coastal Maharashtra, India

Utilizing LISS-4 satellite imagery and support vector machine for mangrove and wetland mapping in part of coastal Maharashtra, India

Integrating Local and Global Features in a Convolutional Vision Transformer for Wetland Mapping

Integrating Local and Global Features in a Convolutional Vision Transformer for Wetland Mapping

The trajectory of wetland change in China between 1980 and 2020: hidden losses and restoration effects.

Understanding wetland change is critical to establishing and implementing international conservation and management conventions. With such knowledge, supporting sustainable development, making management decisions, improving policies, and conducting scientific research become possible. However, consistent information on changes in Chinese wetlands has been unavailable. We applied the hybrid object-based and hierarchical classification approach to ∼53,000 scenes of Landsat images acquired between 1980 and 2020 and created a national wetland mapping product (China_Wetlands) for six periods (e.g., 1980, 1990, 2000, 2010, 2015, and 2020). China_Wetlands revealed diverse changes in Chinese wetlands and their trajectories in response to climate change and human impacts over the past four decades. Specifically, there was a substantial shrinkage in wetland area before 2015, with a small rebound between 2015 and 2020. The net loss was ∼60.9×103km2, which represents 12% of the area in 1980. However, the loss of natural wetlands was hidden by human-made wetland gain with an offset of 15.6×103km2. Additionally, the expansion of surface water extent approximately 14.0×103km2 obscured the loss of vegetated wetlands. Wetland loss in hotspot areas (e.g., Sanjiang Plain and Yangtze River Delta) should not be neglected. The sustainable management and effective conservation of wetlands in China should target wetland areas, landscape structures, and small wetlands delivering important ecosystem services. Moreover, the conversion of wetland types and the invasion of alien species need to be monitored and regulated. China_Wetlands will be a critical wetland dataset for ecological research and the assessment of national and global environmental objectives (e.g., the United Nation's sustainable development goals).

Ecological analysis and multi-scenario simulation of Yellow River Delta wetland under clearing of Spartina alterniflora

In 2022, China released the Special Action Plan for the Control of Spartina alterniflora (2022-2025) with the goal of completely clearing S. alterniflora from China's coastal wetlands. It is crucial to conduct a thorough analysis of wetlands under artificial intervention and obtain reliable future wetland simulation results. The Yellow River Delta wetland is one of China's important wetland reserves, has faced severe impacts from S. alterniflora invasion over the past decade. This study focuses on the Yellow River Delta wetland, utilizing Landsat imagery and a combined object-oriented and pixel-based approach to achieve wetland mapping from 2013 to 2023. Through wetland analysis, we designed scenarios for natural development, ecological priority, farmland protection, and combined development. The patch generating land use simulation model was used to simulate wetland scenarios for 2030 and analyze factors influencing wetland use cover change and future conversion patterns. The results indicate that between 2013 and 2022, native vegetation was impacted by the invasion of S. alterniflora, with Suaeda glauca being the most affected. In 2023, the removal of S. alterniflora was effective, reducing its area by 42.37km2 compared to pre-removal conditions. In all three scenarios, the area of Phragmites australis increased and is expected to dominate the wetlands, while the live space of other native salt marsh vegetation has also been restored. However, the predictions indicate that S. alterniflora still poses a potential threat of expansion in the future, necessitating continuous monitoring of its dynamics. We recommend implementing zoned management of the wetland, balancing ecological restoration in coastal areas with agricultural development in the western regions.

Wetland classification based on depth-adaptive convolutional neural networks using leaf-off SAR imagery

The recent development of deep learning (DL) techniques has created opportunities for classifying wetlands from remote sensing data (mainly optical data). However, the methods for accurately and efficiently classifying large-scale wetlands using DL and radar data that can be more effective than optical data still needs evaluation. In this study, we developed an end-to-end depth-adaptive convolutional neural network (CNN) for mapping wetlands using leaf-off time-series Sentinel-1 Synthetic Aperture Radar (SAR) imagery along with ancillary data. We examined the inclusion of multi-land cover proximity information and a CNN-based self-supervised SAR denoising procedure for enhancing wetland classification accuracy. The depth-adaptive CNN based on U-Net architecture was designed to classify wetland classes (emergent wetland, scrub-shrub wetland, forested wetland, and open water) in Delaware, U.S. while achieving optimization between model complexity (network depths) and accuracy. Results show that our proposed DL method (OA = 0.93, MIoU = 0.60) not only produced a higher classification accuracy than the traditional RF method (OA = 0.89, MIoU = 0.18) but also had a significantly reduced computational cost compared to established state-of-the-art CNNs (e.g., DeepLabV3+ and DANet) without loss of accuracy. The inclusion of multi-land cover proximity information (especially distances to forest and water) and the CNN-based self-supervised SAR denoising procedure can both enhance wetland classification accuracy, especially for forested wetland using traditional RF methods. These results demonstrated the novelty and efficiency of our proposed DL method for classifying wetlands by combing denoised SAR imagery and ancillary information, which provides insights on integration of DL approach and radar data for supporting operational wetland mapping at large spatial scales.

Spatiotemporal prediction of alpine wetlands under multi-climate scenarios in the west of Sichuan, China.

The alpine wetlands in western Sichuan are distributed along the eastern section of the Qinghai-Tibet Plateau (QTP), where the ecological environment is fragile and highly sensitive to global climate change. These wetlands are already experiencing severe ecological and environmental issues, such as drought, retrogressive succession, and desertification. However, due to the limitations of computational models, previous studies have been unable to adequately understand the spatiotemporal change trends of these alpine wetlands. We employed a large sample and composite supervised classification algorithms to classify alpine wetlands and generate wetland maps, based on the Google Earth Engine cloud computing platform. The thematic maps were then grid-sampled for predictive modeling of future wetland changes. Four species distribution models (SDMs), BIOCLIM, DOMAIN, MAXENT, and GARP were innovatively introduced. Using the WorldClim dataset as environmental variables, we predicted the future distribution of wetlands in western Sichuan under multiple climate scenarios. The Kappa coefficients for Landsat 8 and Sentinel 2 were 0.89 and 0.91, respectively. Among the four SDMs, MAXENT achieved a higher accuracy (α = 91.6%) for the actual wetland compared to the thematic overlay analysis. The area under the curve (AUC) of the MAXENT model simulations for wetland spatial distribution were all greater than 0.80. This suggests that incorporating the SDM model into land change simulations has high generalizability and significant advantages on a large scale. Furthermore, simulation results reveal that between 2021 and 2100 years, with increasing emission concentrations, highly suitable areas for wetland development exhibit significant spatial differentiation. In particular, wetland areas in high-altitude regions are expected to increase, while low-altitude regions will markedly shrink. The changes in the future spatial distribution of wetlands show a high level of consistency with historical climate changes, with warming being the main driving force behind the spatiotemporal changes in alpine wetlands in western Sichuan, especially evident in the central high-altitude and northern low-altitude areas.

A novel spatio-temporal vision transformer model for improving wetland mapping using multi-seasonal sentinel data

Wetlands mapping using remote sensing data is a challenging task due to the spectral similarity of wetlands, the fragmented nature of these landscapes, and seasonal variations in wetlands. To address these limitations, this study proposes a novel spatio-temporal vision transformer (ST-ViT) model for an accurate wetland classification using seasonal data. The ST-ViT model was trained using multi-seasonal Sentinel-1 (S1) and Sentinel-2 (S2) data acquired during the spring, summer, and fall of 2020 in a study area located in Newfoundland and Labrador, Canada. The performance of the ST-ViT model was evaluated against the validation dataset, achieving an overall accuracy (OA) of 0.950 and F1-score (F1) of 0.934, outperforming other deep learning models such as random forest (RF), hybrid spectral network (HybridSN), etc. The model demonstrated strong classification capabilities among most wetland classes, with some challenges in distinguishing between spectrally similar classes like bogs and fens. Moreover, the integration of spatio-temporal features enabled the reduction of feature mixing between wetland classes, particularly during different seasons. The ST-ViT model provides an accurate wetland distribution map in different seasons, supporting critical decision-making processes related to wetland conservation and environmental monitoring.

Multi-temporal image analysis of wetland dynamics using machine learning algorithms

Wetlands play a crucial role in enhancing groundwater quality, mitigating natural hazards, controlling erosion, and providing essential habitats for unique flora and wildlife. Despite their significance, wetlands are facing decline in various global locations, underscoring the need for effective mapping, monitoring, and predictive modeling approaches. Recent advances in machine learning, time series earth observation data, and cloud computing have opened up new possibilities to address the challenges of large-scale wetlands mapping and dynamics forecasting. This research conducts a comprehensive analysis of wetland dynamics in the Thatta region, encompassing Haleji & Kinjhar Lake in Pakistan, and evaluates the efficacy of different classification systems. Leveraging Google Earth Engine, Landsat imagery, and various spectral indices, we assess four classification techniques to derive accurate wetland mapping results. Our findings demonstrate that Random Forest emerged as the most efficient and accurate method, achieving 87% accuracy across all time periods. Change detection analysis reveals a significant and alarming decline in Haleji & Kinjhar Lake wetlands over 1990–2020, primarily driven by agricultural expansion, urbanization, groundwater extraction, and climate change impacts like rising temperatures and reduced precipitation. If left unaddressed, this continued wetland loss could have severe implications for aquatic and terrestrial species, water and soil quality, wildlife populations, and local livelihoods. The study predicts future wetland dynamics under different scenarios - enhancing drainage for farmland conversion (10–20% increase), increasing urbanization (10–20% expansion), escalating groundwater extraction (7.2m annual decline), and climate change (up to 5 °C warming and 54% precipitation deficit by 2050). These scenarios forecast sustained long-term wetland deterioration driven by anthropogenic pressures and climate change. To guide conservation strategies, the research integrates satellite data analytics, machine learning algorithms, and spatial modeling to generate actionable insights into multifaceted wetland vulnerabilities. Findings provide a robust baseline to inform policies ensuring sustainable management and preservation of these vital ecosystems amidst escalating human and climate threats. Over 1990–2020, the Thatta region witnessed a 352.8 sq.km loss of wetlands, necessitating urgent restoration efforts to safeguard their invaluable ecosystem services.

Agronomic and economic effects of wetlands on crop yields using precision agriculture data

CONTEXTWetland drainage has become an increasingly important conservation issue in the Prairie Pothole region of North America. Financial incentives for annual crop production have driven wetland drainage for decades, and the removal of wetlands has detrimentally impacted key wetland ecosystem services such as wildlife habitat and carbon sequestration. Past studies which model the farmer's decision to drain wetlands often operate on the assumption that drained wetlands will produce similar yields to upland regions of the field. OBJECTIVEOur objective is to estimate the effects of wetlands and their buffer zones on crop yields, farm financial performance, and incentives for wetland drainage in the Prairie Pothole Region. METHODSWe combine precision yield data and detailed wetland mapping data from 36 fields in the Black and Dark Brown soil zones of Saskatchewan, Canada to estimate the agronomic impacts of wetlands and their buffer areas on crop yields. Then, we incorporate these yield effects into a farm accounting model with three wetland drainage scenarios to estimate the annual per cultivated acre net benefits of wetland drainage in the study area, and compare these results to those estimated without wetland yield effects. RESULTSWe find that yields in wetland basins are relatively lower than the field's average yield, with substantial variability with respect to crop type, soil zone, and annual precipitation. Wetland drainage can mitigate these yield effects, but yields in drained wetland basins still fail to meet the field average yield. These yield effects can extend more than 50 m beyond the wetland boundary. We find that these effects substantially impact the net benefits of wetland drainage. The returns from wetland drainage increase when yield effects are considered. On average, full wetland drainage within the study area increases net benefits by $17 to $33 per cultivated acre relative to full wetland restoration. SIGNIFICANCEThe results demonstrate the importance of considering wetland and buffer zone yield effects in wetland drainage decisions, improve our understanding of wetland costs, and potentially inform policy development and the design of incentives for wetland conservation in agricultural landscapes.

Wetlands Mapping and Monitoring with Long-Term Time Series Satellite Data Based on Google Earth Engine, Random Forest, and Feature Optimization: A Case Study in Gansu Province, China

Wetlands Mapping and Monitoring with Long-Term Time Series Satellite Data Based on Google Earth Engine, Random Forest, and Feature Optimization: A Case Study in Gansu Province, China

Mobilizing Evidence-based Knowledge for Sustainable Wetlands Co-management and Co-governance amidst increasing Anthropogenic and Environmental Stressors: Key Lessons from Mityana District, Uganda

Wetlands (covering about 1.5–1.6 billion hectares globally), are critical biodiversity and livelihood hotspots. Wetlands further replenish the global economy with $47.4 trillion/year worth of ecosystem services. By jealously guarding wetlands, progress toward sustainable development goals, and livelihood welfare are possible. Unfortunately, despite the promulgation of wetland governance mechanisms, 35 percent of the global natural wetlands have been lost since the 1970s. This could be worse in undocumented or explored wetland zones situated in remote tropical regions. In this study, we bring to the fore insights from 286 documents sourced from Scopus and engagements from 105 citizens in Mityana, to (i) map wetlands (including the current vulnerabilities and threats), and (ii) co-develop a wetlands management action pathway that could create sustainable co-management possibilities and sustainable livelihood futures. Findings revealed that although research on wetlands has increased for the last 31 years, since 2021, it has plummeted. In Uganda, wetland research and scholarship is predominantly situated around the Lake Victoria region. Most research focuses on natural or biological sciences. Emerging policy themes and trending research topics are shifting from key wetland management paradigms. From a total of 105 sampled wetlands scattered across fourteen (14) sub-counties in the Mityana district, critical wetland issues were unraveled. Mityana is crossed by two wetland systems (Lake Wamala and River Mayanja dominated by permanent papyrus and seasonal swamps respectively. Wetlands offer unique livelihood, cultural assets/capitals, and ecological benefits (including cultural/aesthetics meaning). An unfathomable rate of degradation is evident. Anthropogenic factors are the predominant threat drivers, especially eucalyptus planting. The loss of culturally valuable wetlands has increased socio-cultural-ecological grief, such as around Lake Wamala. Micro-level management actions are increasing, albeit mainly around accessible permanent wetlands. Most riparian wetland sedentary populations expressed willingness and interest in the co-management and governance of community wetlands. More robust actions and pathways are needed to create avenues for community co-management. The co-developed the sustainable wetlands management action pathway (SWeMAP) provides seven (7) coherent steps, including critical social science insights that could aid sustainable wetlands governance and management across geographies. As wetlands in Uganda have been gazetted as critical to sustainable development, the urgent co-development and financing of micro-level wetland action plans, including situational inventories could help create avenues for sustainable wetlands management.

Wetland mapping in the Liaohe River Estuary using multi-source remote sensing image feature selection

ABSTRACT As a crucial node in the East Asia–Australasia migratory bird network, the Liaohe River Estuary Wetland (LREW) contributes significantly to coastal ecological balance and biodiversity. Accurate and timely mapping of the LREW is essential for monitoring ecological changes and understanding shifts in the wetland’s ecosystem structure and functions. In this study, an innovative approach that integrates radar (Sentinel-1), optical (Sentinel-2), and DEM data on Google Earth Engine (GEE) is proposed to achieve a more optimized and accurate wetland mapping. Recursive Feature Elimination (RFE) is employed to improve feature collection, facilitating the construction of an optimal feature set for Random Forest (RF) classification to extract detailed LREW information. Mapping conducted from 2017 to 2023 using the proposed approach demonstrates its effectiveness and stability in mapping the LREW, achieving overall accuracies ranging from 89.84% to 93.73% and Kappa from 0.85 to 0.91. Compared to single-source methods, the integration of multi-source data substantially enhances mapping accuracy, while RF-RFE effectively eliminates redundant features, further refining classification precision. Texture features (Sum Average, SAVG) and the Water Index 2019 (WI2019) were found to significantly influence wetland mapping. The importance of different feature types is ranked as follows: spectral features > vegetation/water body index > PCA features > SAR texture features > red-edge index > other spectral index > spectral texture features > backscatter coefficient > H-Alpha. During the study period, the wetland area increased from 849.60 km2 to 979.49 km2. This expansion was caused primarily by wetland protection policies that led to the dismantling of coastal aquaculture ponds, resulting in a reduction of 90.28 km2 in waterbodies over 7 years, with many areas transitioning to tidal flats and Suaeda salsa habitats. In conclusion, the proposed method is effective in mapping the LRE wetlands and provides reliable data support for wetland ecological protection and restoration.

Mapping and understanding degradation of alpine wetlands in the northern Maloti-Drakensberg, southern Africa

The alpine terrestrials of the Maloti-Drakensberg in southern Africa play crucial roles in ecosystem functions and livelihoods, yet they face escalating degradation from various factors including overgrazing and climate change. This study employs advanced Digital Soil Mapping (DSM) techniques coupled with remote sensing to map and assess wetland coverage and degradation in the northern Maloti-Drakensberg. The model achieved high accuracies of 96% and 92% for training and validation data, respectively, with Kappa statistics of 0.91 and 0.83, marking a pioneering automated attempt at wetland mapping in this region. Terrain attributes such as terrain wetness index (TWI) and valley depth (VD) exhibit significant positive correlations with wetland coverage and erosion gully density, Channel Network Depth and slope were negative correlated. Gully density analysis revealed terrain attributes as dominant factors driving degradation, highlighting the need to consider catchment-specific susceptibility to erosion. This challenge traditional assumptions which mainly attribute wetland degradation to external forces such as livestock overgrazing, ice rate activity and climate change. The sensitivity map produced could serve as a basis for Integrated Catchment Management (ICM) projects, facilitating tailored conservation strategies. Future research should expand on this work to include other highland areas, explore additional covariates, and categorize wetlands based on hydroperiod and sensitivity to degradation. This comprehensive study underscores the potential of DSM and remote sensing in accurately assessing and managing wetland ecosystems, crucial for sustainable resource management in alpine regions.

Fine-Resolution Wetland Mapping in the Yellow River Basin Using Sentinel-1/2 Data via Zoning-Based Random Forest with Remote Sensing Feature Preferences

Accurate wetland classification in the Yellow River Basin (YRB) is crucial for China’s ecological security, sustainable development, and wetland resource management. This calls for more sustained research on regional variations and studies on remote sensing features, especially with temporal considerations. To address this and the optimization of feature extraction as well as ranking, Sentinel-1 and Sentinel-2 images were used. Additionally, to achieve more precise wetland classification, the YRB was subdivided into four regions for random forest classification. The results show that different remote sensing indices effectively distinguish various wetland types and that key percentiles play a significant role in distinguishing wetland types. The 10 m refined wetland classification map for 2020, with an overall accuracy of 85%, demonstrates that this framework can meet the needs of conventional large-scale wetland analysis and management. The total area of wetlands in the YRB in 2020 was 33,554.67 km2, mainly distributed in the upper reaches of the YRB (71%), with seasonal marshes being predominant. The total water area reached 8538.64 km2, primarily distributed in the upper reaches of the YRB (57.40%). This high-resolution wetland map offers crucial insights and tools for monitoring, protecting wetland resources, and shaping policies, advancing regional sustainable development goals.

Wet-ConViT: A Hybrid Convolutional–Transformer Model for Efficient Wetland Classification Using Satellite Data

Accurate and efficient classification of wetlands, as one of the most valuable ecological resources, using satellite remote sensing data is essential for effective environmental monitoring and sustainable land management. Deep learning models have recently shown significant promise for identifying wetland land cover; however, they are mostly constrained in practical issues regarding efficiency while gaining high accuracy with limited training ground truth samples. To address these limitations, in this study, a novel deep learning model, namely Wet-ConViT, is designed for the precise mapping of wetlands using multi-source satellite data, combining the strengths of multispectral Sentinel-2 and SAR Sentinel-1 datasets. Both capturing local information of convolution and the long-range feature extraction capabilities of transformers are considered within the proposed architecture. Specifically, the key to Wet-ConViT’s foundation is the multi-head convolutional attention (MHCA) module that integrates convolutional operations into a transformer attention mechanism. By leveraging convolutions, MHCA optimizes the efficiency of the original transformer self-attention mechanism. This resulted in high-precision land cover classification accuracy with a minimal computational complexity compared with other state-of-the-art models, including two convolutional neural networks (CNNs), two transformers, and two hybrid CNN–transformer models. In particular, Wet-ConViT demonstrated superior performance for classifying land cover with approximately 95% overall accuracy metrics, excelling the next best model, hybrid CoAtNet, by about 2%. The results highlighted the proposed architecture’s high precision and efficiency in terms of parameters, memory usage, and processing time. Wet-ConViT could be useful for practical wetland mapping tasks, where precision and computational efficiency are paramount.

Surface Soil Moisture Estimation from Time Series of RADARSAT Constellation Mission Compact Polarimetric Data for the Identification of Water-Saturated Areas

Soil moisture is one of the main factors affecting microwave radar backscatter from the ground. While there are other factors that affect backscatter levels (for instance, surface roughness, vegetation, and incident angle), relative variations in soil moisture can be estimated using space-based, medium resolution, multi-temporal synthetic aperture radar (SAR). Understanding the distribution and identification of water-saturated areas using SAR soil moisture can be important for wetland mapping. The SAR soil moisture retrieval algorithm provides a relative assessment and requires calibration over wet and dry periods. In this work, relative soil moisture indicators are derived from a time series of the RADARSAT Constellation Mission (RCM) SAR compact polarimetric (CP) data over reclaimed areas of an oil sands mine in Alberta, Canada. An evaluation of the soil moisture product is performed using in situ measurements showing agreement from June to September. The surface scattering component of m-chi CP decomposition and the RL SAR products demonstrated a good agreement with the field data (low RMSE values and a perfect alignment with field-identified wetlands).

Marsh decrease was much faster than the water increase among the Yellow River Source wetlands during 1986–2022

Wetlands are valuable and sensitive ecosystems that make them imperative to tracking the dynamics in their extent for sustainable management under global warming. Here we focused on the Yellow River Source (YRS) wetlands, which is renowned for hosting one of the world's largest plateau peat bog, unfortunately, it had experienced sharp degradation, threatening the safety of water supply for approximately 110 million people of the lower Yellow River basin. However, the lack of long-term, dense time-series data makes it challenging to assess its evolution trends and driving factors. Therefore, we developed a decision tree sample migration method based on Euclidean distance and Land Surface Water Index, and successfully generated annual wetland mapping of YRS from 1986 to 2022 by utilizing the Landsat 5/7/8 datasets and Random Forest method. The average sample migration rate was 89.21 %, with an average overall accuracy of 95.49 %. We observed that the marsh area decreased by 2031 km2, marking a decline of 12.98 %, while the water area increased by 710 km2 (31.24 %) compared to 1986. Spatially, 10.96 % of marsh composition presents significant (P < 0.05) decline trend, which are mainly converted to grass (86 %), followed by impervious (10 %). There were 6.69 % of water composition showing significant (P < 0.05) increase trend, which are mainly sourced from impervious (82 %) and marsh (12 %). Grazing activities were more important driving forces than climate change for marsh degradation, while the water expansion was associated with recent rising temperature in YRS. The sample migration method is proved to be feasible, robust, and effective for long-term wetland mapping. We suggest that wetland decision-makers need to focus on marsh degradation and reduce grazing intensity, so that fostering the sustainable and healthy wetlands in the Qinghai-Tibetan Plateau.

CVTNet: A Fusion of Convolutional Neural Networks and Vision Transformer for Wetland Mapping Using Sentinel-1 and Sentinel-2 Satellite Data

Wetland mapping is a critical component of environmental monitoring, requiring advanced techniques to accurately represent the complex land cover patterns and subtle class differences innate in these ecosystems. This study aims to address these challenges by proposing CVTNet, a novel deep learning (DL) model that integrates convolutional neural networks (CNNs) and vision transformer (ViT) architectures. CVTNet uses channel attention (CA) and spatial attention (SA) mechanisms to enhance feature extraction from Sentinel-1 (S1) and Sentinel-2 (S2) satellite data. The primary goal of this model is to achieve a balanced trade-off between Precision and Recall, which is essential for accurate wetland mapping. The class-specific analysis demonstrated CVTNet’s proficiency across diverse classes, including pasture, shrubland, urban, bog, fen, and water. Comparative analysis showed that CVTNet outperforms contemporary algorithms such as Random Forest (RF), ViT, multi-layer perceptron mixer (MLP-mixer), and hybrid spectral net (HybridSN) classifiers. Additionally, the attention mechanism (AM) analysis and sensitivity analysis highlighted the crucial role of CA, SA, and ViT in focusing the model’s attention on critical regions, thereby improving the mapping of wetland regions. Despite challenges at class boundaries, particularly between bog and fen, and misclassifications of swamp pixels, CVTNet presents a solution for wetland mapping.