Lead Authors:
Lisamarie Windham-Myers, U.S. Geological Survey
Wei-Jun Cai, University of Delaware
Contributing Authors:
Simone R. Alin, NOAA Pacific Marine Environmental Laboratory
Andreas Andersson, Scripps Institution of Oceanography
Joseph Crosswell, Commonwealth Scientific and Industrial Research Organization
Kenneth H. Dunton, University of Texas, Austin
Jose Martin Hernandez-Ayon, Autonomous University of Baja California
Maria Herrmann, The Pennsylvania State University
Audra L. Hinson, Texas A&M University
Charles S. Hopkinson, University of Georgia
Jennifer Howard, Conservation International
Xinping Hu, Texas A&M University, Corpus Christi
Sara H. Knox, U.S. Geological Survey
Kevin Kroeger, U.S. Geological Survey
David Lagomasino, University of Maryland
Patrick Megonigal, Smithsonian Environmental Research Center
Raymond G. Najjar, The Pennsylvania State University
May-Linn Paulsen, Scripps Institution of Oceanography
Dorothy Peteet, NASA Goddard Institute for Space Studies
Emily Pidgeon, Conservation International
Karina V. R. Schäfer, Rutgers University
Maria Tzortziou, City University of New York
Zhaohui Aleck Wang, Woods Hole Oceanographic Institution
Elizabeth B. Watson, Drexel University
Expert Reviewer:
Camille Stagg, U.S. Geological Survey
Science Lead:
Raymond G. Najjar, The Pennsylvania State University
Review Editor:
Marjorie Friederichs, Virginia Institute of Marine Science
Federal Liaisons:
Zhiliang Zhu, U.S. Geological Survey
Authors wish to thank their respective funding agencies, including the U.S. Geological Survey LandCarbon Program, NASA Carbon Monitoring System Program (NNH14AY671 for Windham-Myers), and the National Science Foundation Division of Ocean Sciences (OCE 1238212, 1637630, and 1237140 for Hopkinson).

Tidal Wetlands and Estuaries

Literature summaries and data compilations discussed in this section enable estimates to be made of carbon stocks and fluxes in North American tidal wetlands and estuaries. Accuracy in quantifying stocks and fluxes in tidal wetlands and estuaries is a function of the accuracy in estimated area (extent) and in estimated stocks and fluxes per unit area. For North America, estimates involve areas, sediment carbon stocks, and the following fluxes: the net change in the carbon stock of tidal wetland soils, tidal wetland exchange of CO2 with the atmosphere (i.e., NEE), tidal wetland exchange of CH4 with the atmosphere, tidal wetland carbon burial, lateral exchange of carbon between tidal wetlands and estuaries, and estuarine outgassing of CO2. Additionally, because the conterminous United States (CONUS) contains a more robust estuarine dataset of most stocks and fluxes, a separate analysis is presented for this region that includes estimates of estuarine NEP, burial, and export of organic carbon to shelf waters.

15.4.1 Tidal Wetland and Estuarine Extent

A synthesis of recent compilation efforts is used to estimate the areas of tidal wetlands and estuaries, and the accuracy of these estimates varies among countries of North America (see Table 15.1). In CONUS, a tidal wetland distribution is estimated using the full salinity spectrum of tidal wetland habitats mapped by the U.S. Fish and Wildlife Service National Wetlands Inventory (USFWS NWI; Hinson et al., 2017). However, in Mexico and Canada, only saline wetlands are available at a national scale, as mapped by the Commission for Environmental Cooperation (CEC; CEC 2016). Hence, tidal wetland areas in Mexico and Canada are likely underestimated. Estimates for the estuarine area of North America use a global segmentation of the coastal zone and associated watersheds known as MARCATS (MARgins and CATchments Segmentation; Laruelle et al., 2013). The MARCATS product is available globally at a resolution of 0.5 degrees and delineates a total of 45 coastal regions, or MARCATS segments, eight of which are in North America. Some CONUS-only applications use estuarine areas from the National Estuarine Eutrophication Assessment survey (Bricker et al., 2007), which is based on geospatial data from the National Oceanic and Atmospheric Administration (NOAA) Coastal Assessment Framework (NOAA 1985). The Coastal Assessment Framework includes a high-resolution delineation of the U.S. coastline in this area and delineates 115 individual estuarine subsystems. Seagrasses are considered separately because of their distinct sediment carbon stocks, even though they overlap in area with estuaries. Seagrass area across North America is estimated according to CEC (2016), using web-available map layers.

Table 15.1 reveals the relative areas of tidal wetlands, estuaries, and seagrasses of North America, in addition to how these ecosystems are distributed by subregion and country. Estuaries of North America cover about 10 times the area of tidal wetlands. About half the tidal wetlands of North America are salt marsh, a third are mangrove, and the remainder is split roughly between tidal fresh marsh and tidal fresh forest. The high-latitude region is characterized by a large estuarine area, about 60% of North America’s total estuarine area, but has only a few percent of the continent’s tidal wetland area and seagrass area. The Gulf of Mexico (GMx), on the other hand, is home to most of North America’s tidal wetlands and seagrasses, with 58% of each. The Atlantic Coast and GMx each have about 10% of the total estuarine area, and the Atlantic coast has about half the tidal wetland area and seagrass area of GMx. The Pacific Coast is similar to the high-latitude subregion with a relatively small area of tidal wetlands and seagrasses (although these areas may be undermapped), and it has an estuarine area about 50% greater than that of GMx. Tidal wetlands of North America reside mainly in CONUS (as salt marsh) and Mexico (as mangroves). Similarly, seagrasses are found mainly in coastal waters of CONUS and Mexico. Estuarine area is not available by country, except for CONUS, which is estimated to have 21% of North America’s total estuarine area.

15.4.2 Tidal Wetland and Estuarine Stocks

Estimates of tidal wetland and estuarine carbon stock in the upper 1 m of sediment or soil were made by using estimates of the carbon density (mass ­carbon per unit volume) from large synthetic datasets. Cross-site comparisons of soil carbon stocks in tidal wetlands illustrate very little range in carbon densities in North America both downcore and among tidal wetlands of varied salinity, vegetation structure, and soil types. Hence, for all tidal wetlands except GMx mangroves, a single estimate of carbon density, 27.0 ± 13 kg organic carbon per m3, was used based on a comprehensive review of the literature (Chmura 2013; Holmquist et al., 2018a; Morris et al., 2016; Nahlik and Fennessy 2016; Ouyang and Lee 2014). For mangroves in GMx, a value of 31.8 ± 1.3 kg organic carbon per m3 was used (Sanderman et al., 2018). A review of seagrass SOC densities (CEC 2017; Fourqurean et al., 2012; Kennedy et al., 2010; Thorhaug et al., 2017) revealed more variance within and between regions, with some notably high soil carbon densities in GMx. Best estimates (and ranges) of 2.0 ± 1.3 kg organic carbon per m3 were used for the Atlantic Coast and high-latitude subregions, 3.1 ± 2.4 kg organic carbon per m3 for GMx, and 1.4 ± 1.2 kg organic carbon per m3 for the Pacific Coast. For organic carbon density in estuarine sediments, a carbon density of 1.0 ± 1.2 kg organic carbon per m3 was used based on a mean value of organic carbon mass fraction (0.4% organic carbon in waters shallower than 50 m; Premuzic et al., 1982; Kennedy et al., 2010) and a dry bulk density average of 2.6 g per cm3 from Muller and Suess (1979). The assumed carbon densities and areas led to carbon stocks in the upper 1 m of 1,410, 354, and 122 Tg C for tidal wetlands, estuaries, and seagrasses, respectively, with a total carbon stock of 1,886 ± 1,046 Tg C.

Net Change in Tidal Wetland Soil Carbon Stock

An estimate of tidal wetland carbon stock loss could only be made using the loss rate for saltwater wetlands in CONUS, as loss rates in other parts of North America and for tidal fresh wetlands are not available. However, CONUS saltwater wetlands make up the overwhelming majority of North American tidal wetlands (see Table 15.1), so applying the CONUS saltwater wetland loss rate to all North American tidal wetlands is not unreasonable. The use of a loss rate of CONUS vegetated saltwater wetlands of 0.18% per year between 1996 and 2010 (Couvillion et al., 2017) and estimated mass of carbon in the upper meter of tidal wetland soils (i.e., 1,362 Tg C) resulted in an overall annualized loss rate of 2.4 Tg C per year. For CONUS only, which holds 1,019 Tg C, the loss rate is 1.8 Tg C per year. Expert judgement assigned 100% errors to these losses because they are deeply uncertain due to annualized episodic events (e.g., Couvillion et al., 2017), difficulty in mapping loss, and difficulty in assessing the rate and fate of carbon from disturbed tidal wetlands (Ward et al., 2017; Lane et al., 2016).

15.4.3 Tidal Wetland and Estuarine Fluxes

Tidal Wetland Net Ecosystem Exchange

Presented in Table 15A.1 are annual estimates of NEE in North America based on continuous measurements, focusing primarily on eddy covariance approaches and high-frequency datasets from static chamber deployments to reduce uncertainty. A total of 16 sites were compiled, including restored wetlands, all of which are in CONUS and mostly along the Atlantic Coast. This limited dataset indicates that NEE varies greatly within and among sites, ranging from the highest annual uptakes in a mangrove ecosystem (–1,200 g C per m2 per year) to the greatest annual losses in a mudflat (1,000 g C per m2 per year) and in a sequence of tidal marshes in Alabama (400 to 900 g C per m2 per year; Wilson et al., 2015). Excluding the restored sites and mudflats from the Hudson-Raritan estuary in New Jersey, as well as the static chamber data from Alabama, the mean NEE at the continuously monitored sites (n = 11 of 16) was negative, indicating uptake of atmospheric CO2 by tidal wetlands. Comparing annual values from the 11 sites (comprising 22 annual datasets) yields coast-specific estimates of NEE: –133 ± 148 g C per m2 per year on the Pacific (one site, 3 years), –231 ± 79 g C per m2 per year on the Atlantic (seven sites, 1 to 3 years), and –724 ± 367 g C per m2 per year in GMx (three sites, 1 to 5 years). Integrating these estimates by area of tidal wetlands on each of North America’s three coasts, the NEE estimate is –27 ± 13 Tg C per year. For CONUS only, NEE is –19 ± 10 Tg C per year.

Tidal Wetland Carbon Burial

Rates of carbon burial in wetland soils and sediments are associated with specific temporal scales depending on calculation methods. Typically, carbon burial is calculated as the product of soil carbon density (i.e., the mass of carbon stored in soil per unit volume) multiplied by accretion rate (i.e., the vertical rate of soil accrual and thus change in volume), which is measured by a variety of dating techniques that span multiple time frames (e.g., marker horizons; radioactive isotopes including those of cesium (137Cs), lead (210Pb), and carbon (14C); pollution chronologies; and pollen stratigraphy). Carbon burial is thus a rate of carbon accumulation in tidal wetland soils over a specific time period (typical units are g C per m2 per year). This measure integrates all carbon pools present, both “old” and “new,” and both autochthonous and allochthonous sources.

Table 15.2 lists carbon burial estimates for salt marshes summarized by Ouyang and Lee (2014), excluding short-term accretion cores (e.g., marker horizons). Identified were 125 cores in North America, about half of which are along the Atlantic Coast and the rest roughly spread evenly among the three other subregions. Mean carbon burial estimates vary considerably among the four subregions, with the lowest rates along the Atlantic Coast, intermediate rates along the Pacific Coast, and the highest rates in the high-latitude subregion and GMx. The ­spatially integrated burial rate was computed for each subregion by multiplying its mean burial rate by its tidal wetland area, thus using an assumption that the salt marsh burial rate applies to tidal freshwater and mangrove systems. The spatially integrated burial rate (±2 standard errors) across North America is 9.1 ± 4.8 Tg C per year, with more than 75% in GMx, owing to its large tidal wetland area (see Table 15.1) and high carbon burial rate (see Table 15.2). For CONUS alone, assuming equivalent distributions of rates among coasts and vegetation types, carbon burial is estimated to be 5.5 ± 3.6 Tg C.

Table 15.2. Carbon Accumulation Rate (CAR) and Associated Data for Tidal Estuarine (Salt and Brackish) Marsha

Region n Mean CAR ± 2σb (g C per m2 per year) Regional Tidal Wetland Burialc ± 2σ (Tg C per year)
High Latitudes 25 301 ± 155 0.5 ± 0.2
Atlantic Coast 59 126 ± 87 1.4 ± 1.0
Pacific Coast 18 173 ± 92 0.6 ± 0.3
Gulf of Mexico 23 293 ± 210 6.6 ± 4.7
North America 125 236 ± 124 9.1 ± 4.8

Notes
a From Ouyang and Lee (2014).
b 2σ = 2 standard errors.
c Regional burial calculated for all tidal wetland types regardless of salinity or vegetation type.
d Key: n, number of sites; g C, grams of carbon; Tg C, teragrams of carbon.

Tidal Wetland CH4 Fluxes

While CH4 fluxes tend to be negligible from tidal wetlands with high soil salinities, emissions can increase considerably when sulfate availability is lower (as indexed by salinity; Poffenbarger et al., 2011). Based on the higher net radiative impact of CH4, climatic benefits of CO2 uptake and the sequestration illustrated by most of the sites in Table 15A.1 may be offset partially by CH4 release in lower-salinity tidal wetlands (Whiting and Chanton 2001).

Here are reported annual CH4 fluxes from tidal wetlands across North America (see Table 15A.2), with values from studies published in 2011 or earlier taken from Poffenbarger et al. (2011). For studies published after 2011, the same methodology was used as Poffenbarger et al. (2011) in analyzing CH4 flux data and reporting average annual CH4 emissions. If CH4 emissions were measured over all seasons of the year with the annual rate unreported, calculations were made by extracting emission rates from tables and figures and then interpolating between time points. Finally, although this was only the case in a few studies, for short-term studies lasting a few days to months over the growing season, average daily CH4 emissions were calculated and then converted to annual fluxes using the rate conversion factors determined by Bridgham et al. (2006). The compilation resulted in CH4 flux measurements at 51 sites in North America.

The compilation, illustrated in Figure 15.3, continues to support the role of salinity as a predictor of CH4 emissions observed by Poffenbarger et al. (2011). However, there is considerable variability among methods and sites in annual CH4 emissions in fresh and brackish (i.e., oligohaline and mesohaline) wetlands, indicating the need for further studies to help improve understanding of the drivers and sensitivities of CH4 fluxes in these common salinity ranges. Tidal wetlands in the salinity range of 0 to 5 practical salinity units (PSU; i.e., fresh-oligohaline) show an average (±2 standard errors) CH4 emission of 55 ± 48 g CH4 per m2 per year, whereas tidal wetlands in the salinity range of 5 to 38 PSU (i.e., mesohaline to fully saline) emit CH4 at an average rate of 11 ± 13 g CH4 per m2 per year. The spatially integrated tidal wetland CH4 emission rate, computed by multiplying the fluxes for fresh-oligohaline and mesohaline-saline systems by their respective areas (5,491 and 33,118 km2; see Table 15.1), results in 0.29 ± 0.27 and 0.35 ± 0.43 Tg CH4 per year, respectively, totaling 0.65 ± 0.48 Tg CH4 per year (0.49 ± 0.36 Tg C per year) across the entire salinity gradient. Hence, in North America, fresh-oligohaline and mesohaline-saline systems contribute about equally to the total flux, with the former having high per-unit-area flux rates and low area and the latter having low per-unit-area flux rates and high area.

   

Figure 15.3: Tidal Marsh Methane (CH4) Emissions Versus Salinity

Figure 15.3: Approaches to measuring atmospheric CH4 flux are coded by method as SC (static chamber) and EC (eddy covariance flux tower). CH4 flux is in grams (g); salinity is in practical salinity units (PSU). The dashed line denotes the demarcation of fresh and oligohaline marshes (0 to 5 PSU) versus mesohaline to saline marshes (5 to 35 PSU).

SHRINK

Lateral Fluxes of Carbon from Wetlands to Estuaries

A significant part of tidal wetland and estuarine carbon budgets is the lateral flux from tidal wetlands to estuaries, which is due mainly to tidal flushing. Twelve estimates of TOC (in both dissolved and particulate forms) exchange (per unit area of wetland) in tidal wetlands of the eastern United States were summarized by Herrmann et al. (2015), and the mean value and 2 standard errors derived in that study (185 ± 71 g C per m2 per year) were used herein. Similarly, four estimates of DIC exchange in eastern U.S. tidal wetlands were summarized in Najjar et al. (2018), with a mean (±2 standard errors) of 236 ± 120 g C per m2 per year. With only a small number of DIC flux measurements, the error was doubled. Hence, tidal wetland export of total carbon is estimated to be 421 ± 250 g C per m2 per year. Applying this to all North American tidal wetlands (see Table 15.1) yields a total export of 16 ± 10 Tg C per year; applied to CONUS wetlands only, the estimate of lateral export is 11 ± 7 Tg C per year.

Estuarine CO2 Outgassing

The SOCCR2 assessment used the global synthesis of Chen et al. (2013), which combined field estimates of outgassing per unit area with the MARCATS areas. Most MARCATS segments were found to be sources of CO2 to the atmosphere, with the integrated flux over North America at +10 Tg C per year (see Table 15.3). Chen et al. (2013) did not provide error estimates, so expert judgment was used to provide a range. The MARCATS segments in North America contain only 25 individual flux estimates, 15 of which are along the Atlantic coast, and some segments have no measurements at all (in which case data from similar systems were used). There is a possibility of a 100% error in the North American flux, so the estimate was placed at 10 ± 10 Tg C per year. Reduced uncertainty may be possible for distinct regions, but this level of error indicates confidence bounds at a continental scale.

Table 15.3. Estuarine CO2 Outgassing for North Americaa,e

MARCATSb Segment No. CO2 Outgassingc (g C per m2 per year) Number of Systems CO2 Outgassing (Tg C per year)
1 129 3 4.4
2 11 3 0.1
3 174 0 1.1
9 96 2 3.1
10 118 15 4.0
11 –9 1 –0.3
12 –5 1 –0.2
13 –13 0 –2.1
Total North America 25 10.0
Approximate CONUSd (2, 9, and 10) 20 7.2

Notes
a Based on the Global Synthesis of Chen et al. (2013).
b MARCATS, MARgins and CATchments Segmentation.
c For regions 3 and 13, where no data were available within the segments, the methods of Chen et al. (2013) were used.
d CONUS, conterminous United States.
e Key: CO2, carbon dioxide; g C, grams of carbon; Tg C, teragrams of carbon.

A separate estimate was made of CONUS estuarine outgassing based on the SOCCR2 synthesis of CO2 flux estimates (see Table 15A.3) and the areas from the Coastal Assessment Framework (NOAA 1985). Because only one study was identified for the Pacific Coast, analysis was limited to the Atlantic and GMx coasts, which contain about 90% of the CONUS estuarine area (see Table 15.1). For the Atlantic coast, mean fluxes were first estimated in each of three subregions (GOM, MAB, and SAB) before multiplying by their respective areas. This was done because the outgassing per unit area increases toward the south. This analysis results in an outgassing of 10 ± 6 Tg C per year (best estimate ±2 standard errors), which is larger (but not significantly so) than the Chen et al. (2013) analysis for the three segments covering CONUS (i.e., 7 Tg C per year). The SOCCR2 synthesis is an improvement over Chen et al. (2013) by being based on a larger flux dataset and more accurate CONUS estuarine areas.

Estuarine CH4 Emissions

Only a very limited number of studies are known to be available and scalable for estimating net CH4 emissions in North American estuaries. In their global review, Borges and Abril (2011) report only three within North America (de Angeles and Scranton 1993; Bartlett et al., 1985; Sansone et al., 1998), ranging from 0.16 to 5.6 mg CH4 per m2 per day. Two recent studies with continuous sampling illustrate temporal and spatial variability. Relatively high emissions were observed in the Chesapeake Bay during summer (28.8 mg CH4 per m2 per day; Gelesh et al., 2016). In the Columbia River estuary (Pfeiffer-Herbert et al., 2016), summer emissions were estimated at 1.6 mg CH4 per m2 per day; 42% of the CH4 losses were to the atmosphere, 32% were to the ocean, and 25% were to CH4 oxidation. When scaled to a year, the estuarine CH4 fluxes from the above studies range from 0.04 to 8 g C per m2 per year, which is well below typical CO2 outgassing rates (e.g., the U.S. Atlantic Coast mean estuarine CO2 outgassing rate is 104 ± 53 g C per m2 per year, see Table 15A.3). Thus, estuarine CH4 outgassing is likely a small fraction of estuarine carbon emissions. To be comparable with North American tidal wetland CH4 emissions (~0.5 Tg CH4 per year), the mean estuarine CH4 emissions rate would need to be a conceivable rate of ~0.1 g CH4 m2 per year. Unfortunately, the lack of estuarine CH4 emissions data for North America—and any well-constrained relationship with salinity or other physical parameter—precludes the possibility of making a constrained estimate of estuarine CH4 emissions for North America.

15.4.4 Total Organic Carbon Budget for Estuaries of the Conterminous United States

The empirical model of Herrmann et al. (2015) was applied to quantify the TOC budget for CONUS estuaries (see Table 15.4). This model takes carbon and nitrogen inputs from a data-constrained watershed model and uses empirical relationships to compute burial and NEP. TOC export to shelf waters is computed by the difference. TOC input from rivers and tidal wetlands to CONUS estuaries is estimated to be 19.5 Tg C per year, with an average of 79% coming from rivers and the rest from tidal wetlands (not shown). Most of the input (74%) is exported from the estuary to the shelf, while 21% is remineralized to CO2 and 5% is buried in estuarine sediments. Like most estuaries worldwide (Borges and Abril 2011), CONUS estuaries are, in the aggregate, net heterotrophic. However, there are regional differences in NEP, with GMx estuaries remineralizing twice as much of the TOC input as Atlantic estuaries and Pacific estuaries metabolically neutral.

Table 15.4. Estuarine Areas and Organic Carbon Regional Budgets for the Conterminous United Statesa,c

Estuary Area (km2) Riverine + Tidal Wetland Input (Tg C per year) Net Ecosystem Production (Tg C per year) Burial (Tg C per year) Export to Shelf (Tg C per year)
Gulf of Mexico 30,586 12.6 ± 3.5 –2.2 ± 0.6 –0.3 ± 0.1 –10.1 ± 3.5
Pacific Coast 6,690 1.4 ± 0.2 0.0 ± 0.2 –0.2 ± 0.1 –1.2 ± 0.2
Atlantic Coast 37,764 5.5 ± 1.3 –1.8 ± 1.0 –0.5 ± 0.3 –3.2 ± 1.3
CONUSb 75,040 19.5 ± 3.8 –4.0 ± 1.2 –1.0 ± 0.3 –14.5 ± 3.7

Notes
a Positive values = input of organic carbon to estuaries; negative values = removal of organic carbon from estuaries. Source: model of Herrmann et al. (2015).
b CONUS, conterminous United States; best estimate and ±2 standard errors.
c Key: Tg C, teragrams of carbon.

15.4.5 Summary Budgets for Tidal Wetlands and Estuaries

The individual flux estimates above were combined into overall carbon budgets for tidal wetlands and estuaries of CONUS and the rest of North America. CONUS (see Figure 15.4a) has better constraints on the fluxes. Central estimates of CONUS tidal wetland carbon losses and gains are very close to balancing even though they were estimated independently; burial, lateral export, and loss of soil carbon stock are all found to be significant terms of carbon removal that balance carbon uptake from the atmosphere. For the estuarine CONUS balance, riverine carbon delivery at the head of tide was taken from Ch. 14: Inland Waters (41.5 ± 2.0 Tg C per year). Including the tidal wetland delivery (11 ± 7 Tg C per year), CONUS estuaries thus were found to receive a total of 53 ± 7 Tg C per year from upland sources. With about 15% (best estimate) of this input outgassed and only a few percent buried, the resulting net total carbon flux from estuaries to shelf waters is 40 ± 9 Tg C.

   

Figure 15.4: Summary Carbon Budgets for Tidal Wetlands and Estuaries

Figure 15.4: Budgets are given in teragrams of carbon (Tg C) for (a) the conterminous United States (CONUS) and (b) North America, with errors of ± 2 standard errors.

SHRINK

The North American carbon budget for tidal wetlands and estuaries (see Figure 15.4b) is similar to the CONUS budget except that most of the fluxes are larger. The net uptake of atmospheric CO2 by the combined system of tidal wetlands and estuaries is 17 ± 16 Tg C per year. The riverine flux of 105 Tg C per year from Ch. 14: Inland Waters was used and assigned an error of 25%. Lacking direct estimates of carbon burial in North American estuaries, the CONUS estimate was used (see Table 15.4, p. 614) and scaled to all North American estuaries; the error is doubled to reflect this extrapolation. The carbon flux from North American estuaries to the shelf waters, estimated as a residual, is 106 ± 30 Tg C per year.


See Full Chapter & References