Modelling forest-atmosphere exchanges of carbon and water using an improved hydro-biogeochemical model in subtropical and temperate monsoon climates

Forest-atmosphere carbon exchanges are crucial yet challenging to quantify accurately due to scaling uncertainties in site observations. Process-based models that mechanistically represent coupled carbon, nitrogen, and water cycling processes are theoretically capable of reducing uncertainties in forest carbon flux quantification, thereby improving predictions of multiple ecosystem variables relevant to achieving the United Nations Sustainable Development Goals (SDGs) by 2030. Thus, we enhanced the CNMM-DNDC model by developing a forest-specific growth module incorporating key processes (photosynthesis, allocation, respiration, mortality, litter decomposition) based on Biome-BGC formulations. Compared with the original model, evaluation against 8-year (2003–2010) eddy covariance data from three Asian forests showed significant improvements in the updated model. At daily and annual scales, normalized root mean square error decreased by 46% and 54% for gross primary productivity (GPP), and 65% and 37% for ecosystem respiration (ER), respectively, though net ecosystem carbon dioxide exchange (NEE) improvements were less pronounced due to error offsetting. Sensitivity analysis identified specific leaf area, fraction of leaf nitrogen in Rubisco and annual leaf and fine root turnover fraction as most influential eco-physiological parameters, with solar radiation, humidity and air temperature as dominant meteorological drivers. The model’s ability to capture daily and inter-annual carbon flux variations demonstrates its potential for regional-to-global greenhouse gas assessments, while highlighting the need for component-specific validation to avoid error masking in net flux calculations.