Marsh plants beat salinity at predicting methane, and blue-carbon ledgers were only half audited

A study led by Boston University doctoral candidate Emily Wilson and published in the Proceedings of the National Academy of Sciences finds that the plant species growing in a tidal marsh predict its methane emissions better than salinity, the variable researchers have leaned on for decades ^[1][2]. That matters because the climate value assigned to protecting or restoring a marsh is a net figure: carbon dioxide removed and stored in soils, minus methane emitted ^[3].

The evidence base is a compilation rather than a new field campaign. Wilson assembled more than 2,000 methane measurements from 87 published studies worldwide ^[5]. Running machine learning models on that pool, Wilson, professor Robinson Fulweiler and fellow doctoral candidate Sawyer Balint found that plant species alone explained 62% of the variability in methane emissions ^[6]. Adding latitude, season and salinity raised that to 71% ^[7], an increment of nine percentage points for three additional variables ^[8]. Wilson describes plant species as "a robust proxy for methane fluxes, outperforming all previously described proxies" ^[10].

The salinity convention is not arbitrary; it was cheap and it correlated. But researchers now know methane can be produced even under salty conditions, which complicates the long-standing salinity-methane relationship ^[12]. The practical consequence is on the policy side. According to Wilson, "the default assumption for a lot of these policies and frameworks is that if it's above a certain salinity then you just don't think about methane emissions" ^[4]. That is a threshold rule standing in for a measurement, and it sits on the emissions side of a ledger whose storage side gets counted.

The asymmetry is the story. Soil carbon accumulation is what marsh accounting is built to capture; methane flux is the term that is expensive to measure directly and therefore gets a shortcut. Plant species are easier to identify than fluxes are to measure, and they are often already mapped ^[11], in the field or from aerial imagery ^[16]. Fulweiler frames the appeal in cost terms: accurate carbon budgets matter for determining net carbon balance and for carbon finance markets, and "a plant proxy is a low-cost and efficient way to estimate methane" ^[13].

Two caveats travel with the result. First, even the best-performing combination leaves 29% of variability unexplained ^[9], so this is a screening tool, not a substitute for flux measurement where the stakes are high. Second, the claim to novelty is about scale, not mechanism. Fulweiler's lab spent more than a decade testing environmental conditions thought to drive greenhouse gas emissions and found that most were not related, while plants stood out because they integrate long-term conditions ^[14]. What was missing, Wilson says, was consensus that species could estimate fluxes beyond one or a few marshes ^[15].

Watch whether standards bodies and national inventories replace salinity cutoffs with species layers, and how quickly, because the switch would reprice projects rather than merely refine them. Watch also the nine-point contribution of latitude, season and salinity ^[8]: if season carries real weight, then when a survey is conducted starts to matter as much as what it maps.