Navigating Research Challenges to Estimate Blue Carbon Benefits From Saltmarsh Restoration

Abstract

Williamson et al. (2024) queried elements of our article in Global Change Biology (Mason et al. 2024), where we used data from 431 articles to quantify global and regional carbon benefits from saltmarsh restoration.

The first query concerns the risk of double counting some carbon through including CO₂ flux within net flux calculations. Some carbon could be double counted with our approach (Figure 1a) when the major source of organic carbon to the sediment is autochthonous. Two other approaches proposed by Williamson et al. (2024) (Figure 1b,c) could be used, but, like ours, over- or underestimate net carbon accumulation, whether the measure is of autochthonous or both autochthonous and allochthonous C accumulation. One measure (Figure 1a) may be consistent with IPCC methodology (Kennedy et al. 2014), although can be inappropriate for offsetting or carbon credit schemes where allochthonous inputs should be excluded (Needelman et al. 2018). Net ecosystem exchange (NEE) in C flux incorporates only autochthonous carbon and can be measured by eddy covariance (EC) (Figure 1c), revealing saltmarshes to be a CO₂ source or sink (±N₂O and CH₄) over the timescale of instrument deployment (years). However, NEE-only estimates may not equate to total carbon accumulation over longer timescales (decades) (Figure 1b) (Lovett, Cole, and Pace 2006), given longer term sediment processes such as remineralisation. Thus, NEE on its own (Figure 1c) can over/underestimate carbon storage. We agree that distinguishing between autochthonous and allochthonous carbon and accounting for carbon exchanged through lateral transport (see section 4.2) are critical next steps in the field of saltmarsh carbon. However, scarcity in published data restricted our ability to account for these processes.

Williamson et al. (2024) suggest basing NEE flux calculations on EC data and excluding chamber-based observations (Figure 1) as their time durations are restricted. We agree that EC is more temporally complete. Yet, EC is also spatially scant and regionally biased, precluding any global analysis. We used GHG flux observations from a range of methodologies including static (opaque or transparent) chambers and EC done on a short-term or seasonal basis. Shahan et al. (2022) showed combining EC and chamber methods improved estimates of net carbon fluxes. Mayen et al. (2024) found that the absence of observations during inundated periods did not influence annual flux rates.

We utilised a large dataset to calculate global carbon stock, identify environmental drivers of spatial variation, highlight current data gaps and discuss implications for policy. We welcome discussion concerning the net flux estimate we produced, but underline that this is just one component of a much larger analysis. In our global synthesis, we worked with an imperfect data set subject to geographical and temporal limitations. There is no perfect data set; consequent interpretations will each have different strengths and limitations. As such, we acknowledge the limitations of our approach and work with what we have until we have well-parameterised models to predict C gains or losses across different contexts.

Victoria G. Mason: conceptualization, formal analysis, investigation, methodology, visualization, writing – original draft, writing – review and editing. Annette Burden: conceptualization, funding acquisition, methodology, visualization, writing – original draft, writing – review and editing. Graham Epstein: conceptualization, writing – review and editing. Lucy L. Jupe: conceptualization, visualization, writing – review and editing. Kevin A. Wood: conceptualization, writing – review and editing. Martin W. Skov: conceptualization, methodology, supervision, visualization, writing – original draft, writing – review and editing.

The authors declare no conflicts of interest.

This article is a Response to the Letter by Williamson et al, https://doi.org/10.1111/gcb.17525, which was related to the paper of Mason et al, https://doi.org/10.1111/gcb.16943.