Volume 190, 15 February 2021, 116738
Author links open overlay panel, , , , , , , ,
The anthropogenic disturbance of carbon and nitrogen cycles globally has resulted in the rapid increase of atmospheric greenhouse gas (GHG) concentrations, especially with respect to carbon dioxide (CO2), methane (CH4), and nitrous oxide (N2O); and triggered a series of environmental issues such as global warming (Ciais and Sabine, 2013). The accurate quantification of GHG sources and sinks has been determined to be a critical task for the current identification of potential ways of mitigating and managing global warming (Griscom et al., 2017; Walsh et al., 2017). Most conventional investigations regarded global river ecosystems as considerable sinks of terrestrial carbon (C) and nitrogen (N) which may be transported into the adjacent oceans by the river network (Boyer et al., 2006; Regnier et al., 2013; Li et al., 2017). However, increasing evidence indicates that the global river ecosystem is a vital source of GHG, despite its smaller surface area compared to other ecosystems (terrestrial and marine ecosystems) (Cole et al., 2007; Bastviken et al., 2011; Quick et al., 2019). The magnitude and spatial variations of riverine GHG emissions have become one of the focal points of current studies, but it unfortunately, continue to be shrouded in uncertainty (Raymond et al., 2013; Hu et al., 2016; Stanley et al., 2016).
Over the past number of decades, many field investigations and dataset compilations have been aimed at quantifying the contribution of riverine GHG emissions to the global GHG budgets, and exploring the spatial variations of different GHG components (Bastviken et al., 2011; Raymond et al., 2013; Hu et al., 2016; Stanley et al., 2016). For instance, Raymond et al. (2013) employed new approaches toward the individual simulation of the surface area, gas transfer velocity and CO2 partial pressure (_p_CO2) for global rivers, and eventually estimated that global river ecosystems emit approximately 1.8 Pg C into the atmosphere every year. Another global synthesis suggested that most river ecosystems are supersaturated with CH4, and release approximately 26.8 Tg CH4 into the atmosphere every year (Stanley et al., 2016). For N2O, a recent study implemented the IPCC method whereby the N2O emission rate was calculated by multiplying the anthropogenic N loading by the corresponding emission factor, and estimated the total N2O emission from global rivers to be 46.5–55.5 Gg N2O/yr (Hu et al., 2016). Generally, these studies systematically evaluated the magnitude and spatial patterns of riverine GHG emissions, and certainly established the critical contribution of river systems to global GHG emissions. However, large uncertainties with respect to current global estimations have persisted, perhaps as a result of methodology bias, data scarcity, and inadequate understanding of the environmental factors driving riverine GHG emissions at a global scale.
One of the challenges of global riverine GHG estimation is the identification of key environmental factors that influence the riverine GHG emissions on a large scale. The biological, chemical, and physical processes of each type of GHG (CO2, CH4 and N2O) varied, and were thus driven by diverse environmental factors. A field survey exceeding eight years in Quebec suggested a strong correlation between terrestrial dissolved organic carbon (DOC) concentration and aquatic CO2 emission (Lapierre et al., 2013). In addition to the terrestrial carbon input, other environment factors such as annual precipitation, flow velocity, and wind speed may also be important factors affecting the riverine CO2 emissions (Butman and Raymond, 2011; Raymond et al., 2013; Li et al., 2019). Previous studies have indicated that riverine CH4 is primarily formed in anaerobic sediment by methanogens, and quickly oxidized by methanotrophs bacteria before being released into the atmosphere (Comer-Warner et al., 2018; Borges et al., 2019). Thus, sediment carbon storage and dissolved oxygen (DO) concentration play pivotal roles in the production and consumption of riverine CH4. In addition, the riverine CH4 emission may be affected by other factors, e.g., water/sediment temperature, trophic status and wind speed (Guérin and Abril, 2007; Stanley et al., 2016; Wang et al., 2018). Compared with CO2 and CH4, riverine N2O emission is more correlated with the anthropogenic N input, including agricultural N-fertilizer or industrial nitrogen-rich wastewater (Yu et al., 2013; Turner et al., 2015). Meanwhile, riverine N2O as a byproduct of two microbial processes, nitrification (NH4+ to NO3−) and denitrification (NO3− to N2O and finally N2), is affected by other environmental conditions such as DOC, pH, and DO (Rosamond et al., 2012; Mekonnen and Hoekstra, 2015). The key factors controlling riverine GHG emission could vary depending on the study regions or scales. For example, Hotchkiss et al. (2015) found that riverine CO2 emissions are dominantly influenced by terrestrially derived CO2 inputs in small rivers but are gradually more affected by aquatic internal metabolisms in the wake of increasing river sizes. Similar findings for riverine CH4 and N2O emissions have been reported in previous studies (Marzadri et al., 2017; Borges et al., 2019). The environmental factors that control the degassing of riverine GHGs continue to be complex and unclear worldwide.
Interestingly, some recent studies have reported that the magnitude of riverine GHG emissions could be correlated with stream order. Butman & Raymond (2014) determined the identical reverse dependency of riverine _p_CO2 and stream order, according to a field measurements database (USGS, The United States Geological Survey) of the 4,138 hydrological stations in the United States. Based on a long term (2010–2017) field survey, Marescaux et al. (2018) found significant variations in riverine GHG (CO2, CH4, and N2O) emissions among the different stream orders of Seine River. Similarly, the dependencies of GHG fluxes and stream order were observed in the streams of the US Corn Belt and Congo River network (Turner et al., 2015; Borges et al., 2019). However, the variations of riverine GHG emissions with increasing stream order at the global scale remain unknown. These studies may offer some theoretical insight, thus allowing for the simultaneous quantification of the global riverine GHG emissions based on the possible identification of negative or positive correlations between riverine GHG emissions and stream orders at the global scale.
In this study, we aimed to quantify the magnitude and spatial variations of global riverine GHG emissions, as well as discuss the underlying mechanisms. We developed a global database recording of GHG measurement from 642 rivers, and explored the spatial patterns of global riverine GHG emissions and their relationship with environmental factors (i.e., dissolved oxygen, DOC, etc.). We also aimed to provide a new estimation of GHG emissions from global rivers.
Section snippets
We compiled GHG emission observations (including concentration, partial pressure and flux) of global river ecosystems (i.e., streams, rivers) from studies published during 1974–2019, by searching keywords including ‘CO2’, ‘CH4’, ‘N2O’, and either ‘river’, or ‘stream’ using Google Scholar, Web of Science and CNKI database (China national knowledge infrastructure, http://www.cnki.net/). As a result, this global database consists of 5948 ecosystem GHG measurements of 595 rivers from 208 published
The global riverine GHG (CO2, CH4, and N2O) fluxes exhibit high variability (Fig. 1). The average global riverine CO2, CH4 and N2O fluxes are 441.1±36.3 mg CO2/m2 h, 5.6±0.4 mg CH4/m2 h, and 0.2±0.02 mg N2O/m2 h, respectively. Among the different types of biomes, the highest CO2 flux was found in tundra regions (1538.3±398.0 mg CO2/m2 h). The highest riverine CH4 and N2O flux occurred in tropics (10.3±1.3 mg CH4/m2 h) and temperate (0.2±0.02 mg N2O/m2 h) regions, separately. The average
This study reports headwater streams ecosystem as an important component of global GHG emissions. High concentration nutrients (DOC, DIC and NO3−) input and a low dissolved oxygen level jointly stimulate the emission of GHG from headwater streams. Additionally, the GHG emissions of global river ecosystems may have the potential to offset climatic benefits of terrestrial carbon (or ocean) sinks, a possibility that has not been sufficiently incorporated into the Earth System Models. Future
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.
This work was supported by the Second Tibetan Plateau Scientific Expedition and Research Program (2019QZKK060602), the National Key R&D Program of China (2017YFA0604803), and Natural Science Foundation in China (31988102). We would like to thank Editage (www.editage.cn) for English language editing. The compiled global database can be found in Appendix B, containing riverine GHG measurements and related environment information (e.g., DO, DOC, DIC, etc.). Any additional data supporting the
- H.M. Baulch et al.
- J.J. Beaulieu et al.
- J.J. Beaulieu et al.
- A. Borges et al.
- E.W. Boyer et al.
- D. Butman et al.
- Ciais, P. & Sabine, C. in Climate Change 2013: The Physical Science Basis. Contribution of Working Group I to the Fifth...
- J.J. Cole et al.
- S.A. Comer-Warner et al.
- J.T. Crawford et al.
- B.R. Deemer et al.
- L. Esters et al.
- B.W. Griscom et al.
- F. Guérin et al.
- S.E. Hinshaw et al.
- E. Hood et al.
- E.R. Hotchkiss et al.
- M. Hu et al.
- B. Jähne et al.
- Streams are disproportionately significant contributors to increases in greenhouse gas (GHG) effluxes in river networks. In the context of global urbanization, a growing number of streams are affected by urbanization, which has been suggested to stimulate the water-air GHG emissions from fluvial systems. This study investigated the seasonal and longitudinal profiles of GHG (N2O, CH4, and CO2) concentrations of Jiuxianghe Stream, a headwater stream undergoing urbanization, and estimated its GHG diffusive fluxes and global warming potentials (GWPs) using the boundary layer method. The results showed that N2O, CH4, and CO2 concentrations in Jiuxianghe Stream were 0.45–7.19 μg L−1, 0.31–586.85 μg L−1, and 0.16–11.60 mg L−1, respectively. N2O, CH4, and CO2 concentrations in the stream showed 4.55-, 23.70-, and 7.68-fold increases from headwaters to downstream, respectively, corresponding to the forest-urban transition within the watershed. Multiple linear regression indicated that NO3−–N, NH4+–N, and DOC:NO3−–N accurately predicted N2O and CO2 concentrations, indicating that N nutrients were the driving factors. The Jiuxianghe Stream was a source of atmospheric GHGs with a daily GWP of 7.31 g CO2-eq m−2 d−1 on average and was significantly positively correlated with the ratio of construction land and forest in the sub-watershed. This study highlights the critical role of urbanization in amplifying GHG emissions from streams, thereby augmenting our understanding of GHG emissions from river networks. With global urbanization on the rise, streams experiencing urbanization are expected to make an unprecedentedly significant contribution to riverine GHG budgets in the future.
- River and reservoir ecosystems have been considered as hot spots for GHG (greenhouse gas) emissions while their specific hydrological and biogeochemical processes affect GHG concentrations; however, few studies integrated river–reservoir systems to identify the dominant drivers of GHG concentrations and flux changes associated with these systems. In the present study, we examined the seasonal variations in GHG concentrations in the surface water of three river-reservoir systems in the Seine Basin. The levels and seasonal variations of GHG concentrations exhibited distinct patterns among reservoirs, upstream, and downstream rivers. The concentrations of CH4 (methane) in the reservoirs were notably higher than those observed in both upstream and downstream rivers and showed higher values in summer and autumn, which contrasted with CO2 (carbon dioxide) concentrations, while N2O (nitrous oxide) concentrations did not show an obvious seasonal pattern. A high mole ratio of CH4/CO2 was found in these reservoirs, with a value of 0.03 and was more than 30 and 10 times higher than that in the upstream and downstream rivers, respectively. The three river–reservoir systems were oversaturated with GHG during the study period, with the average diffusive fluxes (expressed as CO2eq: CO2 equivalent) of 810 ± 1098 mg CO2eq m−2 d−1, 9920 ± 2413 mg CO2eq m−2 d−1, and 7065 ± 2704 mg CO2eq m−2 d−1 in the reservoirs, upstream and downstream rivers, respectively. CO2 and CH4–CO2 were respectively the dominant contributors to GHG diffusive fluxes in river and reservoir sections, while N2O contributed negligibly to GHG diffusive fluxes in the three river–reservoir systems. Our results showed that GHG concentrations and gas transfer coefficient have varying importance in driving GHG diffusive fluxes among different sections of the river–reservoir systems. In addition, our results also show the combined effect of reservoirs and upstream rivers on the water quality variables and hydrological characteristics of downstream rivers, highlighting the future need for additional investigations of GHG processes in the river–reservoir systems.
- Intermittent rivers in semiarid and arid regions, constituting over half of the world's rivers, alternate the carbon cycle interactions among the biosphere, hydrosphere, and atmosphere. Inadequate quantification of flow duration and river water surface area, along with overlooked CO2 emissions from dry riverbeds, result in notable inaccuracies in global carbon cycle assessments. High-resolution remote sensing images combined with intensive field measurements and hydrological modelling were used to estimate and extract the flow duration, river water surface area and dry riverbed area of Huangfuchuan, an intermittent river watershed that acts as a major tributary of the Yellow River in semiarid Northwest China. CO2 emission rates and partial pressures in water and air across the watershed were in-situ measured. In 2018, the flow duration of Huangfuchuan increased from less than 5 days in the first-order tributary to 150 days in the sixth-order mainstream. River water surface area estimated by remote sensing extraction plus the hydrodynamic model simulation varied from 3.9 to 88.6 km2 under 5 %–95 % discharge frequencies. CO2 emissions from the water-air interface and dry riverbed in 2018 were estimated at 582.3 × 103 and 355.2 × 103 ton, respectively. The estimated total annual emission (937.5 × 103 ton) aligns closely with the range of emissions (67.3 × 103–1377.2 × 103 ton) calculated for the water-air interface alone, derived using DEM river length and hydraulic geometry method. This similarity can be attributed to the overestimation of flow duration and flow velocity, as well as the over- or under-estimation of river water surface area and slope. The new method proposed in this study has large potential to be applied in estimating CO2 emissions from data-scarce intermittent rivers located in mountainous regions and provides a standardized solution in the estimation of CO2 emission. Results of this research reveal the spatiotemporal distribution of CO2 emissions along an intermittent river system and highlight the substantial role of dry riverbed in carbon cycle.
- In order to foresee the impact of permafrost thaw on CO2 emissions by high-latitude rivers, in-situ measurements across a permafrost and climate/vegetation gradient, coupled with assessment of possible physico-chemical and landscape controlling factors are necessary. Here we chose 34 catchments of variable stream order (1 to 9) and watershed size (1 to >105 km2) located across a permafrost and biome gradient in the Western Siberian Lowland (WSL), from the permafrost-free southern taiga to the continuous permafrost zone of tundra. Across the south-north transect, maximal CO2 emissions (2.2 ± 1.1 g C-CO2 m−2 d−1) occurred from rivers of the discontinuous/sporadic permafrost zone, i.e., geographical permafrost thawing boundary. In this transitional zone, fluvial C emission to downstream export ratio was as high as 8.0, which greatly (x 10) exceeded the ratio in the permafrost free and continuous permafrost zones. Such a high evasion at the permafrost thawing front can stem from an optimal combination of multiple environmental factors: maximal active layer thickness, sizable C stock in soils, and mobilization of labile organic nutrients from dispersed peat ice that enhanced DOC and POC processing in the water column, likely due to priming effect. Via a substituting space for time approach, we foresee an increase in CO2 and CH4 fluvial evasion in the continuous and discontinuous permafrost zone, which is notably linked to the greening of tundra increases in biomass of the riparian vegetation, river water warming and thermokarst lake formation on the watershed.
- River CO2 emissions, which contribute 53 % of the basin's overall carbon emissions, are essential parts of the global and regional carbon cycles. Previous CO2 flux calculates are mostly based on single samples collected during ice-free periods; however, little is known about the effects of freeze–thaw cycles on the river CO2 flux (FCO2) of inland rivers in alpine regions. Based on one year-round monthly continuous field sampling, we quantified the FCO2 and determined their driving factors in typical rivers during different freeze–thaw periods in the Qinghai Lake Basin (QLB) using the thin boundary layer model (TBL) and the path analysis method. The findings indicated that (1) the average FCO2 in the typical rivers was 184.98 ± 329.12 mmol/m2/d, acting as a carbon source during different freeze-thaw periods, and showed a decreasing trend with completely thawed periods (CTP, 303.15 ± 376.56 mmol/m2/d) > unstable freezing periods (UFP, 189.44 ± 344.08 mmol/m2/d) > unstable thawing periods (UTP, 62.35 ± 266.71 mmol/m2/d); (2) pH, surface water temperature (Tw) and total alkalinity (TA) were the dominant controlling factors during different freeze–thaw periods. Interestingly, they significantly affected FCO2 more before completely frozen than after frozen, with Tw and TA changing from having promoting effects to having limiting effects; (3) in addition, dissolved carbon components indirectly affected FCO2, primarily through the indirect effects of pH and Tw in the UTP; wind speed (U) directly promoted FCO2 in the CTP; and Ca2+ and dissolved inorganic carbon (DIC) were susceptible to indirect effects, which promoted/limited the release of FCO2 in the UFP, respectively. Our results reveal the changes of FCO2 and the factors influencing it in inland rivers within alpine regions during different freeze–thaw periods, thereby offering valuable support for carbon emission-related studies in alpine regions.
- Headwater streams transport nutrients, sediment, and mineral-rich groundwater downstream. In High Mountain Asia (HMA), headwater streams also funnel glacier and snow melt to sustain continuous water supply for the downstream region. These channels remain poorly mapped because of their inaccessibility and because they are smaller than the resolution of Landsat (30 m) and Sentinel-2 (10 m). In this study, we assessed the ability of 3 m resolution PlanetScope imagery to detect the proglacial headwaters downstream of all high-altitude glaciers larger than 5 km2 in HMA. We created 3000 manually labeled image tiles to train and evaluate computer vision (CV) against techniques common in the hydrologic remote sensing literature, specifically normalized difference water index (NDWI) thresholding and random forests (RF). Results indicate that CV best detects the headwater streams with >0.60 F1-scores, nearly 0.20 points higher than RF and 0.45 points higher than thresholding. We also assessed how errors in CV propagate to derived hydrologic information, exemplified by the biogeochemically critical measurement of stream surface area. We found that CV classifications produced surface areas with 0.98 R2, 0.01 km2 MAE, and 0.02 km2 RMSE against manually labeled surface areas. We also observed the best CV performance during the spring season with 30% more skillful classification performance than in summer and fall. Our results prove the ability of PlanetScope imagery to detect and map headwater streams accurately and at scale, and that classification errors stemming from the imagery or the CV methods do not greatly impair our ability to quantify stream surface area meaningful for biogeochemical exchange and hydrology studies.
© 2020 Elsevier Ltd. All rights reserved.