Inland Waters

14.5.1 A Global Carbon Cycle Perspective

Understanding the fluxes of carbon through inland waters in the context of the global carbon cycle remains an active area of research today. Of particular interest are 1) terrestrial carbon fluxes to inland waters; 2) carbon transformations within inland waters, especially movement into storage reservoirs and the atmosphere; and 3) carbon fluxes to coastal waters and large inland lakes. Using Equation 14.1 assessment of components of the inland water carbon cycle can begin at the global, regional, and U.S. scales.

Globally, the component with the least uncertainty is the flux of carbon to coastal waters. Estimates of DOC flux to the coast, for instance, have remained around 0.2 ± 0.05 Pg C per year for the last 30 years, although these estimates often are based on the same underlying dataset (Dai et al., 2012; Meybeck 1982; Seitzinger et al., 2005). The DIC flux of 0.35 Pg C per year has been shown to result from strong linkages between lithology and climate, coupled with better global products for these drivers (Hartmann et al., 2014b). Global estimates of the POC flux to coastal waters have changed because of a large and evolving anthropogenic signal from POC trapping behind dams, with a total flux of 0.15 Pg C per year (Galy et al., 2015; Spitzy and Ittekkot 1991; Syvitski and Milliman 2007). The sum of DOC, DIC, and POC fluxes results in a C_export of 0.7 Pg C per year.

New global and ecosystem-specific estimates of CH₄ and CO₂ exchanges with the atmosphere have been facilitated by the growth of databases that capture measurements of these GHGs and by the ability to scale up estimates of inland water area and gas transfer velocity (Abril et al., 2014; Bastviken et al., 2011; Borges et al., 2015; Butman and Raymond 2011; Lauerwald et al., 2015; Raymond et al., 2013). New research suggests that Arctic and boreal lakes and ponds may release 16.5 Tg C per year (Wik et al., 2016), more than double previous estimates (Bastviken et al., 2011) for a similar range of latitudes. Evidence now shows that lake and river size, topography, land cover, and terrestrial productivity affect the total carbon dynamics in freshwaters (Butman et al., 2016; Holgerson and Raymond 2016; Hotchkiss et al., 2015; Stanley et al., 2016). However, these relationships are based on limited empirical data, and, although progress is being made, a mechanistic understanding that links landscapes to inland water carbon fluxes is still lacking (Hotchkiss et al., 2015). Furthermore, the fluxes of CH₄ and CO₂ per unit area of water surface are extremely high for very small streams and ponds (Holgerson and Raymond 2016), but these systems are not easily detected with remote sensing and have very few high temporal frequency studies (Feng et al., 2015; Koprivnjak et al., 2010).

Carbon dioxide flux from inland waters to the atmosphere (C_emissions) at the global scale is due to mostly large river systems and currently is estimated at 1.8 to 2.2 Pg C per year (Raymond et al., 2013). Recent data from the Amazon suggest that total global emissions could be as high as 2.9 Pg C per year (Sawakuchi et al., 2017). Carbon burial represents another large removal process for aquatic carbon. Global inland water burial estimates are fairly uncertain, ranging from 0.2 to 0.6 Pg C per year as C_burial (Battin et al., 2009b; Tranvik et al., 2009). Assuming that the carbon stock of inland waters is not changing with time and using compiled values only (Raymond et al., 2013) lead to the maximum possible terrestrial input being approximately 3.7 Pg C per year (Raymond et al., 2013), which represents the total carbon needed to balance the loss through coastal export, burial, and gas emissions. Internal primary production and respiration are known contributors to gas emissions, as well as burial. Therefore, verifying this 3.7 Pg C per year currently is not possible due to the diversity of terrestrial and inland water ecosystems, temporal variability of fluxes, and lack of studies of small end-member ecosystems.

14.5.2 Comparison Between Global and U.S. Carbon Fluxes

The fluxes of carbon from the United States (CONUS and Alaska) represent those with the highest confidence reported here and will be evaluated against those at the global scale. A comparison of global versus U.S. estimates of aquatic carbon fluxes shows similar patterns in the relative magnitude of carbon flux pathways. Applying the conservative global estimate for carbon burial of 0.2 Pg C per year (Tranvik et al., 2009), carbon emissions across the air-water interface are 60% of the total flux at the global scale and 63% at the U.S. scale (see Equation 14.1 and Figure 14.2). In contrast to estimates in SOCCR1, these results suggest that half of all aquatic carbon fluxes are releases of gases to the atmosphere. At the global and U.S. scales, lateral fluxes from land to coasts represent 24% and 26% of the total, respectively. It is important to note that globally, POC entrapment through burial, if assumed to be 0.2 Pg C per year, is nearly 6% of the total flux of carbon from inland waters. This amount increases to 16% if the burial term is considered to be 0.6 Pg C per year (Battin et al., 2009b). The range of estimates for the proportion of carbon entering sediments (i.e., 6% to 16%) globally bounds the more refined modeling for CONUS that suggests burial is 10% of the total.

Global and U.S. CO₂ emissions equal 17 and 13.6 g C per m² per year, respectively, indicating that CO₂ emissions from U.S. inland waters are 20% less than the global average per unit land area. Carbon burial per unit area varies from 1.5 to 4.5 g C per m² per year, very similar to the 1.9 g C per m² per year estimate obtained for CONUS and Alaska. Overall, per unit area, the total carbon flux at the global scale is 25% greater (at 24.8 g C per m² per year) than the 20.6 g C per m² per year estimated for the United States. The discrepancies between the U.S. and global areal fluxes increase if recently estimated values (Sawakuchi et al., 2017) are used for the comparisons (see Table 14.1). These discrepancies may be due to differences in methodologies but also may reflect spatial variability in inland water ecosystem type. For example, the importance of tropical systems for carbon fluxes may drive the distribution of inland water fluxes at the global scale, even though tropical areas represent only a very small fraction of the ecosystems within CONUS.

14.5.3 Regional Differences of U.S. Carbon Fluxes

Carbon fluxes from inland waters differ across regions in CONUS, and the relative contributions of each flux component vary across space (Butman et al., 2016). In particular, lateral fluxes from the eastern portion of the Mississippi River basin are larger than gaseous emissions, while carbon burial dominates lake fluxes in the river’s lower basin. Carbon dioxide emissions are dominant in systems that have steep topography and more acidic waters. Emissions of CO₂ are highest in the western regions of the Pacific Northwest, where both rainfall and topography drive large carbon inputs from primary production and topography enhances gas transfer (Butman et al., 2016). Inorganic carbon fluxes in the form of bicarbonate are large within watersheds with large areas of agriculture in the upper Midwest, an effect attributed to agricultural liming (Oh and Raymond 2006). Regional variability in inland water carbon fluxes is driven by the available inputs of carbon from variable land cover, as well as precipitation that facilitates the physical movement of that carbon from groundwater, soils, and wetlands.

14.5.4 North American Carbon Fluxes in Context

Total carbon fluxes from inland waters of North America were estimated using the results of the Regional Carbon Cycle Assessment and Processes (RECCAP) effort (see Table 14.1, p. 576) for emissions and lateral fluxes based on the scaling of empirical data (Hartmann et al., 2009; Mayorga et al., 2010; Raymond et al., 2013). The average burial rate of carbon based on land cover from CONUS and Alaska was used herein for calculations (Clow et al., 2015). The total carbon flux from inland waters is estimated to be 507 Tg C per year. About 48% of this carbon, or 247 Tg per year, consists of emissions across the air-water interface from both lentic and lotic systems. The lateral flux of carbon to the coast is 105 Tg C per year, or 21% of the total. This estimate compares well with recent results derived from a spatially explicit coupled hydrological-biogeochemical model that suggest 96 (standard deviation 8.9) Tg C per year move laterally to coastal systems in North America (Tian et al., 2015). Finally, the burial of carbon within inland waters is estimated to be nearly 30% of the total flux, at 155 Tg C per year. These estimates are based on modeled export of carbon to coastal systems and broadly scaled estimates for CO₂ emissions derived from sparse datasets at high latitudes (Hartmann et al., 2014a; Raymond et al., 2013) and are considered uncertain.