A machine learning approach to quantify land model parameter uncertainty
Land surface models are essential tools for understanding and predicting terrestrial processes and climate-carbon feedbacks in the Earth system, but uncertainties in their future projections are poorly understood. Improvements in physical process realism and the representation of human influence arguably make models more comparable to reality, but also increase the degrees of freedom in model configuration leading to parametric uncertainty in projections. In this work we quantify the contribution of a subset of parameter choices to land model uncertainty using a machine learning approach. Using a perturbed parameter ensemble with the Community Land Model (CLM), we train a neural network to predict CLM output given a set of parameter values as input. We focus on parameters controlling biophysical features such as surface energy balance, hydrology, and carbon uptake. Validation and out-of-sample tests are used to assess the predictive skill of the network, and we utilize feature importance and partial dependence methods to better interpret the results. We then calibrate land surface model parameter values by comparing emulated model output with observations across multiple relevant metrics, including global carbon and water flux benchmarks. Parameter posterior distributions are produced using a Markov Chain Monte Carlo approach. By sampling from these distributions and running future climate simulations, we can then estimate the contribution of land model parameter uncertainty in future projections of climate change.