Sensitivity of modeled forest carbon exchange in Germany to climate forcing differences between E-OBS and ERA5-Land

Process-based models are widely used to estimate forest carbon exchange, however, their outcomes depend strongly, among other factors, on the choice of climate forcing data. This study assesses the sensitivity of modeled carbon exchange in German forests to differences between the E-OBS and ERA5-Land climate datasets using the ecosystem LandscapeDNDC model for the period 2011–2023. Simulations were conducted on a spatially explicit 10 × 10 km grid across Germany for the four dominant tree species groups (beech, oak, spruce, and pine). Within each grid cell, 15 representative sampling points per species were selected and used to extract the vegetation and soil properties. Forest structure was initialized using biomass and canopy height derived from Planet products, soil properties from SoilGrids, and climate forcing was provided alternatively by E-OBS and ERA5-Land. Preliminary results indicate that on average, ERA5-driven simulations produce higher gross primary production (GPP) due to higher precipitation amounts, but also higher total ecosystem respiration (TER) associated mainly with increased minimum temperatures, reflecting warmer nighttime conditions and enhanced ecosystem respiration. As the relative increase in TER exceeds that of GPP, net ecosystem exchange (NEE) is reduced under ERA5-Land forcing compared to E-OBS. Despite this general reduction, NEE exhibits considerable spatial variability across Germany, including both positive and negative deviations. Overall, both climate datasets reproduce the large-scale spatial patterns of GPP, TER, NEE, with consistent regional gradients across Germany. These findings demonstrate that climate forcing choice can significantly influence modeled forest carbon balances at national scale, with important implications for greenhouse gas inventories and forest carbon accounting.