Inland waters play a vital role in the global carbon cycle, acting as dynamic reactors that process vast amounts of carbon. The flux of CO2 outgassing from these waters is substantial, comparable to land-atmosphere and land-ocean exchanges, yet poorly understood in the context of global change impacts. Accurately quantifying CO2 evasion from inland waters is crucial for comprehending carbon cycle feedbacks with climate. Current global estimates are still being refined, with significant regional variations. China, despite covering only ~7% of the global land surface, possesses some of the largest rivers and half of the world’s reservoirs, making it a critical area for study. Its lakes are globally significant and have undergone substantial changes. Understanding CO2 evasion from China’s inland waters is therefore paramount for accurate global estimates and regional carbon budget assessments. China's rapid economic development over recent decades has caused significant environmental changes, altering terrestrial carbon dynamics. While terrestrial carbon sequestration and emissions have been extensively studied, lateral carbon export to inland waters and subsequent CO2 evasion remain relatively under-researched. This study aims to address this gap by quantifying and comparing CO2 emissions from Chinese streams, rivers, lakes, and reservoirs in the 1980s (before major anthropogenic impacts) and the 2010s (after extensive damming and land-use change). The analysis leverages an extensive spatio-temporal dataset to reconstruct past perturbations caused by rapid environmental and socioeconomic changes, offering insights into potential comparable changes in other regions globally.

Existing literature highlights the importance of inland waters in the global carbon cycle, emphasizing the need for improved estimates of CO2 evasion. Studies like Raymond et al. (2013) provide global estimates of CO2 emissions, but these are subject to uncertainties and regional variations. Other research focuses on specific regions, such as Africa (Borges et al., 2015) or the Amazon (Sawakuchi et al., 2017). However, there is a relative scarcity of studies on the impact of global change on CO2 emissions from inland waters over extended time periods. In the case of China, while there's substantial research on terrestrial carbon cycling (Fang et al., 2018; Lu et al., 2018; Tang et al., 2018), research on aquatic carbon dynamics and CO2 evasion is limited. Previous works suggest human activities have altered aquatic carbon dynamics in China (Yue et al., 2016), but the impact on CO2 evasion has been largely unexplored. Existing estimates of CO2 evasion from Chinese lakes and reservoirs (Li et al., 2018) are also comparatively recent and provide a context for this study's more comprehensive evaluation.

This study employed a multi-faceted approach combining in-situ measurements, remote sensing, and GIS analysis to estimate CO2 emissions from Chinese inland waters in the 1980s and 2010s. For the 1980s, when direct CO2 evasion measurements were unavailable, the researchers relied on estimated surface water pCO2 derived from hydrologic gauge-based water chemistry records (pH, water temperature, and alkalinity) from 1709 locations (1960-1985). Data quality was carefully assessed by comparing with other sources, correcting calculation errors caused by biased pH and organic alkalinity. For the 2010s, the study collected data from 1064 locations, using direct CO2 evasion data where available or estimating it based on dissolved CO2 concentration and gas transfer velocities. Hydrological data (discharge, velocity, width, slope), climate data (precipitation, temperature, wind speed), and land cover data were used for analysis and gas transfer velocity estimation. Inland water surface areas were determined using 507 satellite images (Landsat 5 and 7), national inventories, and GIS analysis. The areas were categorized into streams/rivers, lakes, and reservoirs. The stream network was extracted from centerlines and manually restored using maps and high-resolution satellite images. The 1980s network was reconstructed through backward updating. Strahler's ordering system was used for stream classification. For lakes and reservoirs, national inventories and satellite images were used, addressing limitations in older satellite imagery resolution for small water bodies through the use of inventory data. pCO2 was calculated using the CO2calc program. Gas transfer velocity (k) was estimated using different approaches for streams/rivers (Raymond et al. and Ulseth et al.) and lakes/reservoirs, considering size-dependency for the latter. Areal CO2 efflux (FCO2) was calculated using a formula incorporating k, Henry's constant, and the water-air gas concentration gradient. China was divided into six regions for spatial and seasonal variability analysis. The total CO2 efflux was calculated separately for streams/rivers (by Strahler order) and lakes/reservoirs (by size class), then aggregated. The contribution of terrestrial respiration to total CO2 efflux was estimated, distinguishing between soil respiration contributions that are already accounted for in terrestrial carbon sink estimates and other sources such as aquatic mineralization. Statistical analysis included linear regression, ANOVA, and Monte Carlo simulations for uncertainty analysis.

The study revealed a substantial decrease in total CO2 emissions from Chinese inland waters between the 1980s and 2010s, declining from 138 ± 31 Tg C yr⁻¹ to 98 ± 19 Tg C yr⁻¹. This reduction is mainly attributable to decreased emissions from streams and rivers in most regions (31-56% decline), with the exception of the Tibetan Plateau, where there was an increase (18% for rivers, 81% for lakes). Headwater streams disproportionately contributed to riverine efflux in both periods, despite occupying only 34-38% of the total stream surface area. The decrease in riverine CO2 efflux was largely explained by reductions in FCO2, driven by factors such as increased forest cover (reducing labile organic matter and soil CO2), and decreasing cropland cover (decreasing soil erosion and terrestrial organic carbon inputs). Eutrophication and damming also played roles. Conversion of flowing rivers into reservoirs was responsible for a significant portion of the overall decrease in CO2 emissions. While the Tibetan Plateau showed increased emissions, this was linked to climate change-driven expansions of stream/river and lake surface areas, coupled with increased terrestrial carbon delivery. The study shows that the decrease in riverine FCO2 in eastern China, mostly driven by human-induced changes in land cover and land use, explained 58–97% of the decrease in riverine CO2 efflux. The FCO2 changes in lakes and reservoirs in these regions explained 82% of the change in CO2 efflux variability. Damming and water withdrawals were important in the arid NW China region. The decrease in river water surface area in this region explained 74% of the CO2 efflux change. In comparison with the 1980s, the total CO2 efflux decreased from 138 ± 31 Tg C yr⁻¹ to 98 ± 19 Tg C yr⁻¹ in the 2010s. Accounting for the soil respiration component reveals that other sources of carbon, not captured in terrestrial biomass accumulation measurements, contributed substantially to aquatic CO2 evasion.

The findings demonstrate that human activities and climate change have significantly altered CO2 emissions from Chinese inland waters. The unexpected decrease in overall emissions highlights the complex interplay between land-use changes, dam construction, and water management practices. The contrast between the decreasing emissions in most regions and the increasing emissions in the Tibetan Plateau underscores the regionally variable impacts of global change. The substantial contribution of headwater streams emphasizes the importance of considering smaller streams in global carbon cycle assessments. The study's quantitative assessment of the offsetting effect of CO2 emissions from inland waters on terrestrial carbon sinks has important implications for national carbon budgeting. The findings suggest that effective land management, including reforestation and agricultural practices, can substantially affect CO2 emissions from inland waters. The study’s findings emphasize the need for integrating aquatic carbon dynamics into terrestrial ecosystem carbon budgets and highlight the significant role of inland waters in the global carbon cycle.

This study provides a comprehensive assessment of temporal changes in CO2 emissions from Chinese inland waters, revealing a substantial overall decrease. The findings highlight the influence of both human-induced changes (land-use, damming) and climate change on CO2 evasion. Integrating aquatic carbon fluxes into national and global carbon budgets is crucial. Future research should focus on improving the accuracy of CO2 emission estimates, particularly for small water bodies and different carbon sources (including methane). Further investigation into the long-term effects of damming, eutrophication, and climate change on aquatic CO2 emissions is needed.

The study acknowledges several limitations. The reliance on estimated pCO2 for the 1980s introduces uncertainties. The data for the Tibetan Plateau in the 1980s is limited, and the assumptions made about small lakes and reservoirs in that period could affect the overall accuracy. The study focuses on CO2 emissions and doesn't fully account for other greenhouse gases like methane, which are also relevant to the overall carbon budget. The study highlights the difficulty of accurately attributing changes in emissions to specific drivers because of the simultaneous action of several factors.