Search

Drained wetlands are thought to be carbon (C) source hotspots, and rewetting is advocated to restore C storage in drained wetlands for climate change mitigation. However, current assessments of wetland C balance mainly focus on vertical fluxes between the land and atmosphere, frequently neglecting lateral carbon fluxes and land‐use effects. Here, we conduct a global synthesis of 893 annual net ecosystem C balance (NECB) measures that include net ecosystem exchange of CO2, along with C input via manure fertilization, and C removal through biomass harvest or hydrological exports of dissolved organic and inorganic carbon, across wetlands of different status and land uses. We find that elevating water table substantially reduces net ecosystem C losses, with the annual NECB decreasing from 2579 (95% interval: 1976 to 3214) kg C ha−1 year−1 in drained wetlands to −422 (−658 to −176) kg C ha−1 year−1 in natural wetlands, and to −934 (−1532 to −399) kg C ha−1 year−1 in rewetted wetlands globally. Climate, land‐use history, and time since water table changes introduce variabilities, with drainage for (sub)tropical agriculture or forestry uses showing high annual C losses, while the net C losses from drained wetlands can continue to affect soil C pools for several decades. Rewetting all types of drained wetlands is needed, particularly for those formerly agriculture‐used (sub)tropical wetlands where net ecosystem C losses can be largely reduced. Our findings suggest that elevating water table is an important initiative to reduce C losses in degraded wetlands, which could contribute to policy decisions for managing wetlands to enhance their C sequestration. [ABSTRACT FROM AUTHOR]

As climate change intensifies, understanding the dynamics of lake evaporation is imperative, especially in semi‐arid regions where water resources are already scarce. This study examines the regulatory role of lake depth on evaporation rates, focusing on a terminal lake in a semi‐arid region: Dali Lake in China. Using the Complementary Relationship Lake Evaporation model, we simulated the heat and temperature lag time of Dali Lake, an 8 m deep lake, due to its heat storage capacity. This approach was validated through moderate‐resolution imaging spectroradiometer (MODIS)‐based surface temperatures of Dali Lake and adjacent Ganggenor Lake. Dali Lake, by storing heat during the warmer months, maintains lower surface temperatures compared with the shallower Ganggenor Lake. Under the same climatic conditions, Dali Lake has an annual evaporation of 980 mm, which is 45 mm less than that of Ganggenor Lake, which has an annual evaporation of 1024 mm. To further study the impact of lake depth, we simulated the heat storage and evaporation of Dali Lake during the Holocene, when the lake reached up to 34 m average depth, representative of the maximum depth reached by Dali Lake. During the Holocene, under constant climate conditions, the annual evaporation would be 44 mm/year less than the average evaporation from 1984 to 2016. Average annual evaporation decreased with increasing depth, showing a significant reduction during warmer months, while the release of heat during the ice‐cover period did not result in additional evaporation. Our results highlight the important relationship between lake depth and evaporation under climate change, emphasizing the necessity for depth‐specific water management strategies in semi‐arid regions. [ABSTRACT FROM AUTHOR]

Previous studies have shown that regional deep learning (DL) models can improve runoff prediction by leveraging large hydrological datasets. However, training a DL regional model using all data without screening may degrade local performance. This study focuses on constructing enhanced local models through the utilization of cross‐basin data. To this end, we propose an approach that employs a novel training strategy to optimize DL model training for specific basins. The approach measures the impact of any one basin's gradient on the loss of the basin of interest, providing insights into the relationships between different basins. The approach was validated using 531 basins from the CAMELS dataset. Results suggest that local performance degradation is a common occurrence in regional models, and imbalanced data are likely to result in a specific pattern dominating the entire regional model. In comparison to a regional model simply trained with all basins, the median Nash‐Sutcliffe efficiency (NSE) for our models is 0.031 higher. In particular, the increase in NSE can exceed 0.2 for some dry basins. Our findings indicate that this novel DL strategy can significantly improve model performance in specific basins using large hydrological datasets, while mitigating local performance loss. Plain Language Summary: In the realm of deep learning, incorporating more data into the training process typically results in a more potent model. Conventionally, large datasets have been employed to train regional models with the intention of predicting rainfall‐runoff processes across all basins. However, such regional models often encounter performance degradation when applied to local basins. This degradation can be attributed to the extraction of overly general features, leading to a loss of specificity. In this paper, our objective is to harness information from a large dataset to establish a more robust local model. We have introduced a method that autonomously learns the similarities in rainfall‐runoff behavior among basins. Subsequently, we utilize this learned similarity to selectively choose data that proves advantageous for training the local model. Our results demonstrate that this strategy can significantly enhance runoff prediction, particularly in arid basins. Key Points: Training a DL regional model using all data without screening may degrade local performanceProper selection of training data is crucial for enhancing DL model training for individual basins, especially arid basinsData from different basins might act as mutual noise during the training process [ABSTRACT FROM AUTHOR]

The Zoige peatlands are the largest peatland area in China, and the largest high‐altitude peatland in the world. As with many peatlands worldwide, degradation from land management and climate change mean that the intact Zoige peatland area has decreased, potentially reducing the carbon (C) sink function and ecosystem services. This review summarizes current knowledge of the impacts of land‐use and climate change on the Zoige peatland C cycle in a global perspective and identifies future research and management directions. The existing literature suggests that artificial drainage carried out to lower water tables and improve grazing has a significant impact on the peatland C cycle. Drained and degraded areas may act as a net C source, through increased CO2 emissions, although the overall C balance of the Zoige peatlands is likely still a net C sink. Future climate change may also impact upon the peatland C cycle. Warming of 2°C may significantly reduce the strength of the C sink of intact peatland areas, which may shift the overall Zoige peatland C cycle balance to a net C source. The effect of warming on degraded Zoige peatlands is a major uncertainty, although the global literature suggests warming effects may be greater in degraded peatlands. Restoration of degraded peatlands (by blocking drains) may help reverse some of the impacts of degradation and gradually recover C sink function. However, there are fewer studies in Zoige peatlands than elsewhere. We conclude with several specific suggestions for future research on the peatland C cycle. This article is categorized under:Paleoclimates and Current Trends > Modern Climate ChangeAssessing Impacts of Climate Change > Observed Impacts of Climate ChangeClimate, Ecology, and Conservation > Observed Ecological Changes [ABSTRACT FROM AUTHOR]