Conversion of coastal wetlands, riparian wetlands, and peatlands increases greenhouse gas emissions: A global meta-analysis

Land-use/land-cover change (LULCC) often results in degradation of natural wetlands and affects the dynamics of greenhouse gases (GHGs). However, the magnitude of changes in GHG emissions from wetlands undergoing various LULCC types remains unclear. We conducted a global meta-analysis with a database of 209 sites to examine the effects of LULCC types of constructed wetlands (CWs), croplands (CLs), aquaculture ponds (APs), drained wetlands (DWs), and pastures (PASs) on the variability in CO₂, CH₄, and N₂O emissions from the natural coastal wetlands, riparian wetlands, and peatlands. Our results showed that the natural wetlands were net sinks of atmospheric CO₂ and net sources of CH₄ and N₂O, exhibiting the capacity to mitigate greenhouse effects due to negative comprehensive global warming potentials (GWPs; −0.9 to −8.7 t CO₂-eq ha⁻¹ year⁻¹). Relative to the natural wetlands, all LULCC types (except CWs from coastal wetlands) decreased the net CO₂ uptake by 69.7%−456.6%, due to a higher increase in ecosystem respiration relative to slight changes in gross primary production. The CWs and APs significantly increased the CH₄ emissions compared to those of the coastal wetlands. All LULCC types associated with the riparian wetlands significantly decreased the CH₄ emissions. When the peatlands were converted to the PASs, the CH₄ emissions significantly increased. The CLs, as well as DWs from peatlands, significantly increased the N₂O emissions in the natural wetlands. As a result, all LULCC types (except PASs from riparian wetlands) led to remarkably higher GWPs by 65.4%−2,948.8%, compared to those of the natural wetlands. The variability in GHG fluxes with LULCC was mainly sensitive to changes in soil water content, water table, salinity, soil nitrogen content, soil pH, and bulk density. This study highlights the significant role of LULCC in increasing comprehensive GHG emissions from global natural wetlands, and our results are useful for improving future models and manipulative experiments.