A proposed roadmap for upscaling approaches in biogeochemistry: insights from conceptual carbon modeling to estimate soil respiration at regional scales

Large scale biogeochemical models are an essential tool to predict the export of carbon, nutrients and contaminants in terrestrial and aquatic ecosystems. The representation of biogeochemical fluxes (e.g., soil carbon fluxes) in earth system models is still a source of uncertainty due to the fact that scientific understanding is obtained at small scales (e.g., laboratory and field), whereas biogeochemical processes are controlled by complex interactions across a wide range of spatial scales (from microorganisms to regional scales). Reliable regional biogeochemical predictions thus require proper bottom-up approaches to scale up the theoretical rate parameters. Recently, an upscaling method has been developed in hydrological modeling utilizing 'transfer functions' that connect model parameters to inputs, and scale-adaptive averages of model parameters. In our study, we develop a roadmap to adopt similar approaches in regional biogeochemical modelling based on this successful hydrological upscaling practice. As a case study, we focus on estimating soil respiration using a simple first-order carbon model with two pools. Despite linear assumptions in the model, non-linear dependence of carbon decay rate constant on soil moisture dynamics (simulated using a conceptual hydrologic model) results in non-linear carbon fluxes such as soil respiration. Given theoretical values for the amount of respiration, we demonstrate the bias in the estimated amount of respiration flux at both sub-grid and regional scales. Our results indicate that improper upscaling approaches might result in significant soil respiration estimation error, which directly impacts decision making related to climate change impacts on regional carbon fluxes. Next step would be to utilize proper transfer functions and averaging rules based on the interactions between soil properties and process parameters to improve biogeochemical predictions across scales.