Hidden Deep Soil Carbon Emerges as a Hidden Cost of Coastal Wetland Restoration

Beneath the cordgrass marshes of Hangzhou Bay in eastern China lies a vast, dark reservoir of carbon that scientists have long assumed was safely locked away. A new field experiment suggests that assumption may be dangerously wrong. When restoration crews stripped out invasive plants and covered the ground with black plastic sheeting, the effects rippled far beyond the surface: salt levels plunged through the entire top metre of soil, microbial life surged in deep horizons that were thought to be biologically dormant, and substantial amounts of carbon vanished from the profile. The study, published in the journal Biogeosciences by Jingwen Gao of the Chinese Academy of Forestry and colleagues, is among the first to track how common restoration techniques reshape the hidden chemistry of deep coastal soils.

Coastal wetlands are celebrated as blue carbon powerhouses, capable of storing up to roughly 200 megagrams of carbon per hectare in their top metre of soil. Crucially, about 60 to 70 percent of that carbon sits below 30 centimetres, where waterlogged, oxygen-poor conditions slow decomposition and iron minerals bind organic molecules into stable complexes. Most restoration assessments, however, monitor only vegetation recovery and the top 30 centimetres of soil, implicitly treating everything beneath as inert. The Hangzhou Bay team set out to test whether that assumption holds when management interventions dramatically alter salinity and moisture at the surface.

The researchers established three field conditions in a wetland overrun since the 1980s by the invasive cordgrass Spartina alterniflora. Untouched invaded stands served as the control. In a second set of plots, crews removed all aboveground vegetation and covered the soil with 0.1-millimetre black polyethylene film to suppress regrowth and curb evaporative salt accumulation, a technique known as plastic mulching. In a third set, vegetation was removed and the soil mechanically tilled to 160 centimetres with a rotary cultivator to boost aeration and physical disruption, a technique called deep tillage. Five replicate plots of 20 by 20 metres were established for each treatment, separated by at least 50 metres to minimise spatial interference.

Eighteen months later, the team collected soil cores from each plot and sliced them into five depth intervals spanning 0 to 100 centimetres. The samples were analysed for pH, electrical conductivity, moisture, nitrogen pools, total and organic carbon, microbial biomass, and a suite of iron fractions extracted through sequential chemical procedures. The researchers also quantified iron-bound organic carbon, the fraction of soil carbon locked to iron oxides, and sequenced bacterial 16S rRNA genes from surface layers to characterise community shifts. Statistical models, including partial least squares path modeling, were then used to explore how the measured variables related to one another across depth.

The most striking finding was the sheer reach of plastic mulching. Salinity fell by 43 to 53 percent throughout the entire profile, and electrical conductivity dropped by 45 to 50 percent in deep soils compared with untouched stands. Deep tillage, by contrast, reduced surface salinity by only about 20 percent and had limited effect below. With the osmotic stress of salt lifted, microbial biomass carbon in deep soils under mulching climbed from roughly 25 percent above control levels at 30 to 50 centimetres to more than 100 percent at 50 to 100 centimetres. Meanwhile, microbial biomass in the surface 30 centimetres declined, revealing a wholesale redistribution of microbial life from the surface toward deeper horizons.

That microbial awakening coincided with substantial carbon losses. Total carbon under mulching fell by 19 to 35 percent across the profile relative to the control, and soil organic carbon declined by 34 to 65 percent, with the strongest losses at depth. Total nitrogen dropped even more sharply, by 51 to 74 percent. The carbon declines tracked a parallel weakening of mineral protection: poorly crystalline iron oxides, the highly reactive phases that scavenge organic matter, fell by 30 to 50 percent, and iron-bound organic carbon fell by 35 to 50 percent. Integrated across the full metre of soil, the fixed-depth estimate pointed to an apparent soil carbon stock decline of 65 plus or minus 12 megagrams of carbon per hectare over just 18 months.

The mechanistic story the authors propose centres on what soil scientists call the iron gate. In anaerobic wetland soils, iron oxides form organo-mineral complexes that resist microbial attack, effectively shielding carbon for decades or centuries. Rapid desalinization can alter the solubility and mobility of reactive iron phases through changes in ionic strength and pH, potentially dissolving those protective associations. At the same time, salt-stressed deep soil microbes, once freed from osmotic constraint, may become more active in carbon turnover. The path models were consistent with this picture: mulching was strongly associated with altered soil physicochemical conditions, which in turn were negatively associated with reactive iron pools and with soil organic and iron-bound carbon. In surface soils, bacterial community attributes were linked to carbon retention, while in deep soils, increased microbial biomass covaried with reactive iron decline and lower carbon.

The bacterial sequencing revealed dramatic community reorganisation in the surface layers. Untouched stands harboured elevated abundances of sulfur- and iron-cycling taxa such as Sulfurifustis and Acidibacter, tilled soils were enriched in Nocardioides and Nitrospira, and mulched soils were characterised by Thiobacillus and other taxa. Mantel tests showed that salinity and iron-carbon variables were the strongest correlates of community composition, explaining a large share of the variation. The authors are careful, however, to note the limits of their evidence: sequencing covered only the top 20 centimetres, and no direct measurements of carbon dioxide or methane fluxes, pore-water dissolved organic carbon, iron redox speciation, or lateral hydrological export were made. The lost carbon could have been mineralised to greenhouse gases, leached away as dissolved iron-organic complexes, or simply redistributed within the profile, and the study cannot distinguish among these fates.

Those caveats matter because the numbers are sobering when set against restoration gains. Aboveground biomass in restored coastal wetlands typically accumulates at only about 1 to 3 megagrams of carbon per hectare per year after native vegetation returns. A potential subsoil loss of 65 megagrams over 18 months could therefore dwarf the early carbon benefits of replanting, at least in the short term. Three trajectories remain possible for mulched sites: stabilisation as labile carbon pools are exhausted, deceleration as recovering vegetation rebuilds carbon inputs, or continued decline if weakened iron protection persists. Only multi-year monitoring with direct flux measurements can determine which path these sites actually follow, and whether they function as net atmospheric carbon sources.

The study also carries an important design lesson. Because both managed treatments involved vegetation removal before mulching or tillage, the contrasts reflect responses to integrated field management states rather than isolated single-factor effects, and the single-site results from Hangzhou Bay may not generalise to wetlands with different salinity, tides, or sediment textures. Still, the central message is hard to ignore: the magnitude and spatial extent of hydro-geochemical change, not disturbance intensity alone, appears to govern the fate of deep soil carbon after invasive plant removal. The authors argue that blue carbon accounting frameworks should incorporate full-profile soil monitoring to at least one metre, iron-associated carbon metrics as indicators of mineral protection, and multi-year measurements of gaseous and dissolved carbon fluxes. As coastal restoration accelerates worldwide in the fight against climate change, the invisible carbon beneath the marsh surface may prove to be the most important number nobody was counting.

Subject of Research: Effects of plastic mulching and deep tillage restoration interventions on deep-soil salinity, microbial biomass, and iron-bound carbon in a Spartina alterniflora-invaded coastal wetland

Article Title: Profile-wide desalinization is associated with increased deep-soil microbial biomass and reduced iron-bound carbon in coastal wetland restoration

Article References: Profile-wide desalinization is associated with increased deep-soil microbial biomass and reduced iron-bound carbon in coastal wetland restoration. (n.d.). https://doi.org/10.5194/bg-23-6671-2026

Image Credits: AI Generated

DOI: 10.5194/bg-23-6671-2026

Keywords: coastal wetlands, blue carbon, soil organic carbon, iron-bound carbon, Spartina alterniflora, soil salinity, microbial biomass, wetland restoration, deep tillage, plastic mulching, soil biogeochemistry, Hangzhou Bay