Carbon Stocks and Accumulation Rates in Salt Marshes of the Pacific Coast of Canada

Abstract. Tidal salt marshes are known to accumulate blue carbon at high rates relative to their surface area and have been put forth as a potential means for enhanced CO2 sequestration. However, estimates of salt marsh carbon accumulation rates are based on a limited number of marshes globally and the estimation of carbon accumulation rates require detailed dating to provide accurate estimates. We address one data gap along the Pacific Coast of Canada by estimating carbon stocks in 34 sediment cores and estimating carbon accumulation rates using 210Pb dating on four cores from seven salt marshes within the Clayoquot Sound UNESCO Biosphere Reserve and Pacific Rim National Park Reserve of Canada (49.2° N, 125.80° W). Carbon stocks averaged 80.6 ± 43.8 megagrams of carbon per hectare (Mg C ha−1) between the seven salt marshes, and carbon accumulation rates averaged 146 ± 102 grams carbon per square meter per year (g C m−2 yr−1). These rates are comparable to those found in salt marshes further south along the Pacific coast of North America (32.5–38.2° N) and at similar latitudes in Eastern Canada and Northern Europe (43.6–55.5° N). The seven Clayoquot Sound salt marshes currently accumulate carbon at a rate of 54.28 Mg C yr−1 over an area of 46.94 ha, 87 % of which occurs in the high marsh zone. On a per-hectare basis, Clayoquot Sound salt marsh soils accumulate carbon at least one order of magnitude more quickly than the average of global boreal forest soils, and approximately two times larger than rates for forests in British Columbia. However, because of their relatively small area, we suggest that their carbon accumulation rate capacity could best be considered as a climate mitigation co-benefit when conserving for other salt marsh ecosystem services.


List of Tables
Maximum corrected depth of excess 210  Chapter 1.

Introduction
Coastal, vegetated ecosystems, such as eelgrass meadows, mangroves, and tidal salt marshes, have recently been recognized for their ability to store large amounts of carbon, or "blue carbon," within their soils and sediments (IPCC 2014;Howard et al. 2014). While blue carbon ecosystems cover approximately 0.2 % of the ocean surface, previous studies have suggested that they could be responsible for up to 50 % of total ocean carbon burial (Duarte et al. 2005). The estimated, average per-area carbon sequestration rate is between 30 to 50 times greater than that of terrestrial forests (McLeod et al. 2011). Globally, blue carbon ecosystems have been estimated to sequester between 75.3 and 224.2 Tg C yr -1 (Duarte et al. 2013).
The high carbon storage and accumulation capacity per-area of coastal ecosystems have been investigated because of the potential for blue carbon to provide climate mitigation co-benefits, when managed for other ecosystem services provided by coastal wetlands, such as storm surge attenuation, coastal erosion control, habitat for commercially important species, and ecotourism (Howard et al. 2017). Climate change mitigation refers to efforts to reduce the negative impacts of anthropogenic climate change by either reducing carbon dioxide (CO2) and other greenhouse gas emissions or enhancing natural carbon sinks to increase the rate at which CO2 is removed from the atmosphere. Climate change mitigation for blue carbon resources the limitation of habitat destruction by human activity, because blue carbon ecosystems naturally store accumulated carbon in their soils for centuries or millennia (Duarte et al. 2005). This carbon can be released when the ecosystem is degraded (McLeod et al. 2011, Pendleton et al 2012. To better inform policies that identify priority areas for conservation, more precise measurement of carbon stocks and accumulation potential are needed (Howard et al. 2017).
Global estimates of salt marsh area, carbon stocks, and carbon accumulation rates (CAR) are subject to large uncertainties. Duarte et al. (2013) noted a 20-fold uncertainty in global estimates of salt marsh area, ranging from 22,000-to 400,000 km 2 .
This uncertainty is attributed to ambiguous classification schemes for wetlands. For example, some classification systems consider freshwater and saltwater marshes in the same category (Duarte et al. 2013). Similarly, the estimated, global soil carbon stock of all salt marshes ranges between 0.4 and 6.5 Pg C, a 16-fold range (Duarte et al. 2013).
Currently, the average global CAR estimate for salt marshes is 244.7 ± 26.1 g C m -2 yr -1 (Ouyang and Lee 2014), but recent reviews of salt marsh CAR estimates disproportionately represent certain areas of the world (Ouyang and Lee 2014;Chmura 2003 An additional factor that limits CAR quantification is the extensive use of 137 Cs radioisotope dating or a marker horizon method, which have the potential for producing overestimates of sediment accumulation rates when compared to radioisotope dating methods such as 210 Pb (e.g. Callaway et al. 2012;Johannessen and MacDonald 2016).
For example, of 143 studies reviewed by Ouyang and Lee (2014), only three did not use either a 137 Cs or marker horizon method, not including studies which did not specify.
Dating using these methods have been demonstrated to produce CAR estimates up to 26 % higher than 210 Pb in California salt marshes (Callaway et al. 2012 hectares of the west coast of Vancouver Island ( Figure 1). The region is part of the temperate rainforest biome with high annual rainfall (3270 mm y -1 ) and average annual temperature of 9.5 °C (Environment Canada, 1981-2010. The mean tidal range in Tofino is 2.14 m (Fisheries and Oceans 2016).
We collected 34 sediment cores from seven marshes during Summer, 2016, to determine their carbon storage and accumulation rates (Table 1; Figure 1). These

Field Sampling
Within each marsh, sediment cores were extracted along linear transects perpendicular to the low tide shoreline following the methodology of Howard et al. (2014). Coring spots were approximately evenly spaced along the transect (between nine and 24 meters apart) from land to sea, and attempted to sample from both the low and high marsh zones  If a spot contained a mixture of these species, the majority percent cover of high or low marsh species was used to determine whether the spot was low or high marsh. Carex lyngbyei were often found throughout both strata and so were not considered unique to one zone. These designations are defined by the presence or absence of low marsh or high marsh vegetation--particularly the high marsh plants

Grindelia integrifolia and Potentilla anserina, which grow in a narrow elevation range in
Clayoquot Sound (Jefferson 1973). The high marsh species' ranges align approximately with the mean extreme high-water line of estuarine marshes in Clayoquot Sound, while low marsh encompasses elevations between the mean lower high water and the mean extreme high-water lines (Jefferson 1973as cited in Deur 2000Weinmann et al. 1984).
This method was groundtruthed using detrended correspondence analysis to verify that vegetation assemblages reflected distinct low and high marsh zones (Hill and Gauch 1980; see section 4.4.1).
Sediment cores were collected using a simple percussion coring technique in which a length of two-inch (57 mm) diameter, PVC vacuum tubing fitted with a plastic core catcher (AMS Inc.) was hammered into the ground until the depth of refusal. Depth of refusal (DoR) is considered a reasonable proxy for sampling to the maximum depth of organic accumulation (Fourqurean et al. 2014b). At one site (GBK) a steel sledge corer (AMS Inc.) was used to extract four cores, but mechanical problems required switching to the simpler method described above. All cores were stored upright between sampling until their return to the laboratory where they were photographed, logged, and stored under refrigeration at a Parks Canada laboratory in Vancouver, British Columbia.

Estimating Marsh Areas
ArcMap 10.3 tools were used with 50 x 50 cm resolution aerial orthophotos taken in July 2014 (Government of British Columbia) to obtain area estimates of high and low marsh zones. The difference between high marsh and low marsh was delineated by eye between darker-green, denser high marsh vegetation and lighter-green, salt-tolerant, and less-dense low marsh vegetation. This method was groundtruthed using the detrended correspondence analysis (e.g. Hill and Gauch 1980) of vegetation survey data and was found to accurately categorize 94 % of the cores into the correct marsh zone (see Discussion section 4.4.1).

Soil Carbon Density and Carbon Stocks
For each marsh, average carbon stocks (Mg C ha -1 ) were estimated in each sediment core by first measuring the soil carbon density (SCD, g C cm -3 , Eq. (1)) on onecm thick sample intervals over the length of each core. SCD is the mass of carbon found in a cubic centimetre of soil at a given depth and is the product of the organic carbon content % C and the dry bulk density (DBD): where DBD represents the weight of one cc volume of soil that was dried for no less than 72 hours at 60°C.
Organic carbon content (%C) was estimated either using loss-on-ignition (LOI, Eq. (2)) or using CN Elemental and coulometric analysis (Froehlich 1980). An LOI test was performed on every 1 cm subsample by homogenizing samples with a mortar and pestle, combusting them at 550°C for four hours, weighing, and combusting again at 1000°C for two hours (Heiri et al. 2001). The percentage mass loss-on-ignition (%LOI) was estimated as: where DWi is initial dry weight and DWf is the dry weight after burning. The %C was also estimated by measuring total carbon (%TC) and inorganic carbon ( Figure A1): Next, the carbon stock of a core was estimated from the sum of all 1-cm intervals in each core (Eq 4): Where i = the depth of the top of a 1-cm subsection in cm, n = the depth of the top of the deepest subsection of the core (cm), and SCDi = the SCD of each subsection i in grams C cm -3 .
Carbon stocks were calculated both in megagrams per hectare (Mg C ha -1 ) --the typical unit used in carbon stock assessment (Fourqurean et al. 2014a) --and in total Mg C for high and low marsh to compare the estimates for each marsh zone.
First, to calculate the average carbon stock for all marshes in megagrams C per hectare, all core C stock estimates were averaged across each marsh and scaled up: Where x = the number of cores in a marsh.
A Kruskal-Wallis test of significance for sample groups of unequal variances was used to test for significant differences between C stockMarsh (Mg C ha -1 ) between the seven marshes studied. Lastly, the Clayoquot Sound average C stock, C stockCS, was computed by averaging the C stockmarsh estimates from all seven marshes.
Characteristics for low and high marshes were estimated and compared, including total C stock, DBD, %C, SCD, and DoR (Welch's t test). Lastly, the total C stock for low marsh C stockLowCS was estimated by averaging each site's low marsh core C stock estimates and multiplying by the total estimated low marsh area in Clayoquot Sound. The same was done to estimate the total high marsh C stock, C stockHighCS.

Carbon Accumulation Rate
Carbon accumulation rates (CARs) were estimated in five cores from the CBE, CRF, GBK (2), and TMF sites, by multiplying the sediment accumulation rates (SAR) by the SCD (Eq. (6)): SARs were calculated from age models determined using 210 Pb dating.
Subsamples from each of the five cores were dated using Polonium-210 alpha counting by Core Scientific International (Winnipeg, Canada) and MyCore Scientific (Dunrobin, Canada). Using a constant rate of supply model, age-depth models were constructed, and SARs estimated (Oldfield and Appleby 1984;Rowan et al. 1994; see Appendix B).
Some core compaction (maximum 40 %) occurred during the coring process, which would affect our estimated accumulation rates. We corrected for this compaction by applying a correction factor for each core (Eq. (7)): and used it to find the uncompacted depth (Eq. 8)) of any given subsample (Fourqurean et al. 2014a): The uncompacted depths were used only to calculate SAR (cm yr -1 ), which was then used to calculate CAR (see equation 6).
The regional average CAR in from Clayoquot Sound, CARCS, was calculated as the average of all five cores with 210 Pb dating. The total CAR for a marsh with a dated core was calculated by multiplying the high marsh core CAR times the high marsh area.
Low marsh CAR for each site used the one low marsh dated core multiplied by the site's low marsh area. Regional average CAR for the high and low marsh zones specifically were estimated using the average of the four, 210 Pb dated high marsh cores to represent the high marsh and the one low marsh core to represent the low marsh zone.

Soil Properties
Depths of refusal ranged from five cm in the low marsh of SWC to a maximum of  Soil carbon densities averaged 0.037 ± 0.17 g C cm -3 for all sites, and site-wide average SCDs ranged from 0.020 to 0.055 g C cm -3 (Table 1)

Carbon Storage and Marsh Area
The seven marshes ranged in size from 0.51 to 27.42 ha, with a total area of 46.93 ha (  Figure 5).
Using our estimates of marsh area, we calculate that C stockCS is 4709 ± 136 Mg C, 70 % of which is stored in the high marsh.

Carbon Accumulation Rates
Carbon accumulation rates averaged 146 ± 102 g C m -2 yr -1 at the four sites from which 210 Pb dating was completed. The low marsh core at GBK had the lowest CAR of 37 g C m -2 yr -1 . CAR in the four high marsh cores ranged from 75 g C m -2 yr -1 at TMF to 264 g C m -2 yr -1 at CBE ( Figure 5). The SAR ranged from 0.142 cm yr -1 at the GBK low marsh to 1.322 cm yr -1 at GBK high marsh ( Approximately 87 % of the total, annual CAR is in the high marsh, while this area represents only 58% of the total marsh area. Compaction factor for GBK 1-4 is average of the other 3 from GBK because hole depth was not measured due to infilling after the corer was withdrawn. No significant difference in CAR was found when using the minimum (0 %) and maximum (40 %) compaction from other cores (p > 0.05).

Comparisons between marshes and strata
C stockHighCS is significantly higher than C stockLowCS (p < 0.05). This is largely attributable to differences in the DoR between high and low marshes ( Figure 5). While the average DoR of high marsh cores is significantly higher than the average DoR of low marsh cores (p < 0.05), no significant differences were found between average DBD, average % C, or a core's average SCD in high versus low marsh cores (p > 0.05) ( Figure 5).
The Kruskal-Wallis test found significant differences between each of the seven C stockMarsh estimates (p < 0.05; K = 12.67). This result shows that each of the marsh average carbon stock estimates vary enough from one another that a single site average cannot be assumed to represent the average carbon stocks of all marshes in the region.  Chapter 4. Discussion

Carbon Stocks-Comparisons
The C stockCS averaged 80.6 ± 43.8 Mg C ha -1 , which is roughly half the global estimate for the top meter of salt marsh soils of 162 Mg C ha -1 (Duarte et al. 2013). These

Carbon Accumulation Rates-World Comparisons
While the Clayoquot Sound regional average CAR of 146 g C m -2 yr -1 appears lower than the global average of 245 g C m -2 yr -1 , this difference is not statistically significant (Welch's t-test, p > 0.05). Clayoquot Sound's average CAR is also comparable to CAR estimates from both its latitude band and the other sites within its biogeographical region. Even though Clayoquot Sound's CAR appears lower than the estimate of 315 g C m -2 yr -1 calculated for the latitude range 48.4-58.4° N (Ouyang and Lee 2014), this difference is not statistically significant, most likely due to the high variability within this latitude range (SEM ± 62.9 g C m -2 yr -1 ). The median value for the 48.4-58.4° N range (153.5 g C m -2 yr -1 ) is comparable to Clayoquot Sound's average CAR. The latitudinal average appears to be inflated by two high CAR estimates of 793 and 1133 g C m -2 yr -1 (Andrews et al. 2008). At the same time, the average Clayoquot Sound CAR is also not significantly different (Welch's t test, p < 0.05) from the NE Pacific average of 174 g C m -2 yr -1 (SEM ± 45.1 g C m -2 yr -1 ), which were estimated from eight data points in California, USA (Ouyang and Lee 2014). These results both underscore that while there is site-to-site variability in CAR, on the scale of 10-degree latitude bands or biogeographical region, Clayoquot Sound's average CAR is close to expected values for its region and its latitude.

210 Pb and 137 Cs Dating
Clayoquot Sound's average CAR is slightly-but not significantly-lower than the other regions of North America and its latitude band, and some of this difference may be attributable to the method used to measure sediment accumulation rate. Previous researchers have argued that using 137 Cs dating to establish age models can result in elevated SARs, and therefore also CARs that are biased high (Johannessen and MacDonald 2016). This overestimation can be a point of concern when making global estimates of salt marsh CAR because the dating method may artificially elevate estimated carbon sequestration potential. All of the accumulation rates from the NE

Figure 6 Comparison of Clayoquot Sound CAR with other salt marsh studies compiled by Ouyang and Lee (2014) grouped by regions as defined by that study.
All eight available data points from the NE Pacific region south of 38.2 °N are shown; North Europe data points are the minimum (Skallingen, Denmark) and maximum of that dataset (Scheldt, Netherlands). Data from single sites are unfilled shapes, while filled-in shapes represent averages. *= high marsh; **= low marsh; No asterisks= not specified. [

Low Marsh CAR
Our low marsh core exhibits anomalously low CAR, however a single core does not posses sufficient statistical power to draw conclusions about differences between average low and high marsh CAR in Clayoquot Sound. While previous studies have found that low marsh CARs are consistently higher than CARs from high marsh areas (Adams et al. 2012;Callaway et al. 1996;Connor et al. 2001;Elsey-Quirk et al. 2011), our results show that CARs were significantly lower in the low marsh at GBK when compared with the high marsh at GBK, and with the high marsh cores from the other sites. A power analysis showed that at least nine total cores measured for CAR would be required to confidently compare the means of low marsh and high marsh cores. This was beyond the resources of our study, but future studies should consider this to investigate whether the low marsh CAR in Clayoquot Sound is consistently lower.
Evidence from past studies suggests that organic sediment accumulation drives marsh accretion, and that this biomass accumulation would be greater in the high marsh than the low marsh. Marsh soil accretion is the result of both organic deposition and inorganic sediment supply, and the relative contribution of each can vary over time (Drexler 2011).
A study from the US Pacific Northwest found a strong relationship between marsh standing biomass and soil carbon (Thom 1992). Additionally, a study of Louisiana salt marshes found that sediment accumulation varied with organic sediment input but not with inorganic input (Nyman et al. 2006). These both suggest that low marshes may experience higher inorganic sediment input, but the CAR would be lower because accretion would be driven by low-carbon, inorganic sediment.
Falling relative sea level (RSL) in Clayoquot Sound may influence marsh accretion dynamics in a way that has yet to be studied and would require additional work to quantify. Low marshes accumulate inorganic sediment primarily from tidal inundation, as particles fall out of suspension in the water or become trapped by the roots of low marsh vegetation (e.g. Connor et al. 2001). Salt marshes thus accumulate vertically in response to rising sea levels (Morris et al. 2002). The tide gauge at Tofino has measured a steadily falling relative sea level since observations began in 1905 (NOAA 2013a), which is most likely a consequence of tectonic uplift in the region (Mazzotti et al. 2008). Therefore, the mechanism of vertical accretion may be different from that observed in marshes experiencing rising sea level.
Lastly, the low marsh core's low CAR could simply be the product of small-scale variability of SAR and CAR due to variables we could not control. Previous studies of marsh accretion dynamics have demonstrated variability in SAR on scales as small as one meter due to such influences as recent ecological disturbance (Webb et al. 2013), water table height and soil drainage (Craft 2007), and variable mineral sediment deposition from freshwater drainage (Callaway et al. 2012).

Marsh Areas and Vegetation Survey
Some inaccuracy was expected when ground-truthing the area estimation method using vegetation survey, but this was minor. This approach to differentiating high and low marsh matched with vegetation data for 32 of 34 (94 %) of cores. Both CRF 1-2 and CRF 2-2 were classified as low marsh by vegetation survey but fell within the high marsh using the visual orthophotography method. These cores lie 16 m (CRF 1-2) and 12 m (CRF 2-2) away from the boundary with low marsh as measured using orthophotos, which is less than their distances from the nearest high marsh cores (17 m and 23 m, respectively). All other cores fell within the correct marsh zone.
A detrended correspondence analysis (Hill and Gauch 1980)  The distinction between a salt marsh and a bordering freshwater area has complicated efforts to classify marshes by salinity (Duarte et al. 2013), but this result shows that clustering of vegetation type corresponds reasonably well with each site's designation as high or low marsh (Figure 7). Additional work could clarify this high marsh-backshore boundary with greater precision (See recommendations section).

Identifying Measurement uncertainties
This study produced qualitative estimates of carbon stocks and accumulation rates using surface area and inferred depth as a proxy for the total volume of soil carbon stocks. These estimates are reasonably reliable but should not be considered a comprehensive account of carbon stocks, but rather a summary estimate of inferred carbon stock and accumulation rates based on the 34 samples recovered. This is true of both our overall estimates and our comparisons between high and low marsh. For example, a sufficiently powerful number of cores for statistical comparisons between high and low marsh SCD is 135, which was beyond the means of this study.
SCD estimates were the main source of uncertainty for carbon calculations.
While SCD was not estimated with a high degree of power, the significant difference in high and low marsh core depths was estimated with power approaching 1. This was most likely due to the relatively small differences in both DBD and %C between low and high marsh. While a much larger number of cores would be required to confirm if any significant difference exists between low and high marsh SCD, the difference in low marsh and high marsh core depths can be interpreted with a high degree of confidence.
C stockcore and accumulation rate values reported here include uncertainty propagation because each core's carbon stock was computed from uncertain SCD values and CAR were computed from SCD and uncertain, core-average SAR values. In turn, each C stockmarsh average also includes uncertainty propagation, as do the regional averages estimated for both stocks and accumulation rates.
Several other factors of the calculation of carbon stock and accumulation rate estimates contributed to the uncertainties in calculated values: avoiding channels and ditches while sampling, the accuracy of the method used for marsh area estimates (see section 4.4.2), and the relationship between %TC and %LOI (see Appendix A).
Avoiding ditches and channels likely biased our carbon stock estimates slightly high by roughly 5 %, however this is difficult to quantify. Vegetation survey data shows an overall average of 5 ± 9 % coverage classified as "bare" without vegetation, however this varied substantially between cores (ranging from 0 % to 30 % bare). Using unvegetated surface area as a rough estimate for the surface area of channels is not ideal, as in some places, particularly the low marsh, bare spots without vegetation were still covered with organic accumulation.
Lastly, the calculated relationship between %LOI and %TC relied upon a strong relationship between %LOI and %TC as measured by elemental analyzer (r 2 =0.97), which is a minor source of potential uncertainty. Measuring soil C by %LOI alone tends to overestimate because some compounds other than carbon, as well as structural water in clay minerals are volatilized at high temperatures (Schumacher 2002). This method produced a small amount of uncertainty, but the strength of the linear relationship between the measured %LOI and calculated %C quantities shows that this was minimal.

Blue Carbon vs. Boreal Forest for Climate Mitigation in Canada
The carbon storage potential of blue carbon ecosystems such as the salt marshes of Clayoquot Sound have been touted as a reason to conserve marshes against future degradation as part of a climate change mitigation strategy . Blue carbon ecosystems have been argued to accumulate and store carbon at rates several times higher than terrestrial forests per unit area (McLeod et al.

2011). Conservation of blue carbon ecosystems such as Canada's salt marshes
presents an opportunity not only to protect a significant carbon stock but also to ensure that these marshes continue to accumulate significant amounts of carbon. Given the tremendous importance of saltmarshes as fish habitat, nursery areas, and coastal buffer zone, carbon storage is an excellent co-benefit to other management activities.
Total soil C stocks per unit area from Clayoquot Sound salt marshes are lower than those of boreal forests. Clayoquot Sound salt marsh soil C stocks average 80.6 Mg C ha -1 , which is similar to the approximately 80 Mg C ha -1 estimated for Canada's boreal forest (Kurz et al. 2013). However, this estimate does not include the other C stock pools in forests, such as aboveground biomass; the soil pool is estimated to account for only 40 % of total forest carbon. This estimate for soil C stock also does not include organic material on the ground such as leaf litter, which, when included, brings the total soil C stock estimate up to 123 Mg C ha -1 .
However, salt marsh carbon accumulation rates are substantially higher per hectare. Boreal forest soils are estimated to accumulate 4.6 ± 2.1 g C m -2 yr -1 globally (Zehetner 2010as cited in McLeod et al. 2011, while Clayoquot Sound marshes accumulate 146 ± 102 g C m -2 yr -1 . Canada's boreal forest is estimated at 270 million hectares in area (Kurz et al. 2013), and total marsh area in Canada is approximately 44,000 ha (Bridgham et al. 2006 (Costanza et al. 2008). In addition to this, Breaux et al. (1995) estimate that the capitalized value of water purification from salt marshes can reach up to $15,000 USD ha -1 when compared with artificial water treatment. Habitat for recreational fishing is estimated to be worth $2,420 ha -1 (Bell 1997  Marshes with salinity between 5-18 emit enough methane in CO2 equivalent units to offset part of the CO2 equivalent value of carbon sequestration (Poffenbarger et al. 2011), while marshes with salinity >18 can be assumed to represent net carbon sinks with no significant methane emissions (Poffenbarger et al. 2011).Freshwater wetlands emit methane due to metabolic activity from methanogenic bacteria in the soil, but these bacteria are outcompeted by sulphate-reducing bacteria in saline soils, negating their methane emissions (Bartlett et al. 1987).

Ecosystem Services and Carbon Valuation
We endeavored to select sites with low salinities but could not verify the salinity of every site with precision and chose instead to focus solely on soil carbon stock and sequestration. The soil salinity of marshes in Clayoquot Sound are likely similar to known surface water salinity measured offshore near several sites. These measurements-taken in late May 2016 after a three-week period of low rainfall--range from 5.9, in Kennedy Cove approximately 100 m northwest of KCS (Postlethwaite et al. 2016, submitted) to 24 in Grice Bay near GBK.).

Recommendations
Further research into Clayoquot Sound carbon stocks and CAR can address the shortcomings of this study and inform future carbon dividend or offset policies. We recommend the following topics as research priorities for future work to quantify carbon in Clayoquot Sound: 1. Groundtruthing area estimates and marsh strata designations with a statistically powerful number of absolute elevation measurements 2. Collecting a greater number of dated cores from the low marsh to investigate differences between low marsh and high marsh CAR 3. Measuring methane emissions throughout the marsh using gas collection chambers, which could both quantify the GHG balance of salt marshes and assess the reliability of salinity as a proxy for methane emissions in the region.
Measuring absolute elevation using surveying equipment such as a tripodmounted level with a stadia rod would permit more precise groundtruthing of marsh area estimates and the designation of high and low marsh strata. Using measurements of absolute elevation would permit the definitions of high and low marsh strata to be directly related to their tidal inundation exposure. Results of these data, in turn, would provide a much more precise way to estimate the surface areas of high and low marsh, and to delineate the boundary between the high marsh and the backshore.
As stated in section 4.4.1, a greater number of low marsh cores would be required to establish a statistical difference in CAR between low and high marsh with confidence, so a future research plan wishing to investigate this should incorporate a larger number of dated cores for CAR analysis. The minimum number required would be 9 cores, based on the variability of other factors involved in computing CAR, but low marsh cores should be dated alongside a high marsh core from the same marsh or transect. This would help to control for site-specific factors affecting CAR. This work could also be compared against similar studies from areas of rising relative sea level to investigate the way marshes might respond to the falling sea level in Clayoquot Sound.
Lastly, quantifying methane emission from salt marshes is an important final step for determining the viability of carbon dividends or offset credits trading. This would require periodic, direct measurement of methane emissions from a number of sites in the region because methane emissions can vary over both short and long time periods (Bridgham et al. 2013).
Relating these methane emissions to salinity measurements would also help to verify the strength of that relationship in marsh soils, especially because soil salinity is also likely to fluctuate substantially through time based on the seasonality of precipitation in Clayoquot Sound. 82 % of annual rainfall in Tofino falls between the months of October and April (Environment Canada, 1981-2010, and soil salinity of marshes can vary by a factor of 10 through time (Bartlett et al. 1987). Quantifying salinity alongside methane would be vital for ensuring that carbon dividends or credits are not misallocated, and could also inform the use of salinity as a proxy in future blue carbon studies elsewhere in BC. This would help to avoid overvaluing of blue carbon's net soil carbon accumulation and, as a result, a net increase in anthropogenic carbon emissions resulting from implementing policy.

Chapter 5. Conclusions
Our work provides estimates of soil carbon stocks and accumulation rates from salt marshes on the Pacific coast of Canada, addressing the data gap within North We found lower carbon stock in the low marsh and an anomalously low CAR in the low marsh when compared with the high marsh, however we cannot determine if this CAR result is representative of the region's low marsh in general because it comes from only a single core. Soil properties such as SCD, DBD, and %C are not statistically different between marsh elevation zones. The anomalously low CAR may be due to chance because of small-scale variability in SAR, greater soil formation in the high marsh, or other region-specific factors such as a falling RSL may also influence CAR due to differences in vertical accretion dynamics within areas of emergent coastline.
While providing an important climate regulation ecosystem service, blue carbon alone does not provide a sufficient monetary incentive for marsh conservation, and should be regarded as a co-benefit of marsh conservation that seeks to preserve other, more valuable ecosystem services. Carbon storage in Clayoquot Sound marsh soils is valued at approximately $11,000 ha -1 when estimated with the current British Columbia carbon tax, however this is only 43 % of the estimated per-hectare value of storm and erosion management, habitat for commercially important species, and ecotourism provided by salt marshes. Despite it's relatively low value, blue carbon ecosystems should be regarded as carbon accumulation 'hot spots,' and the value of their carbon accumulation should be factored into management decision-making.
Lastly, further investigation of groundtruthing methods for area estimates and high and low marsh designations would allow more precise calculations of carbon storage and accumulation in marshes. This knowledge, alongside greater understanding of any variability between high and low marsh CAR and understanding of methane emissions, would help to inform the role of blue carbon in both local ecosystem services management and the larger global carbon cycle.