Abstract
Abstract. Land 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 increased parametric uncertainty in projections. In this work we design and implement a machine learning approach to globally calibrate a subset of the parameters of the Community Land Model, version 5 (CLM5) to observations of carbon and water fluxes. We focus on parameters controlling biophysical features such as surface energy balance, hydrology, and carbon uptake. We first use parameter sensitivity simulations and a combination of objective metrics including ranked global mean sensitivity to multiple output variables and non-overlapping spatial pattern responses between parameters to narrow the parameter space and determine a subset of important CLM5 biophysical parameters for further analysis. Using a perturbed parameter ensemble, we then train a series of artificial feed-forward neural networks to emulate CLM5 output given parameter values as input. We use annual mean globally aggregated spatial variability in carbon and water fluxes as our emulation and calibration targets. Validation and out-of-sample tests are used to assess the predictive skill of the networks, and we utilize permutation feature importance and partial dependence methods to better interpret the results. The trained networks are then used to estimate global optimal parameter values with greater computational efficiency than achieved by hand tuning efforts and increased spatial scale relative to previous studies optimizing at a single site. By developing this methodology, our framework can help quantify the contribution of parameter uncertainty to overall uncertainty in land model projections.
Highlights
Land models were originally developed to provide lower boundary conditions for atmospheric general circulation models but have evolved to simulate important processes such as carbon cycling, ecosystem dynamics, terrestrial hydrology, and agriculture
We plot the results of the permutation feature importance tests for PC1 of gross primary production (GPP) and latent heat flux (LHF) as bar charts in Fig. 7, with larger bars reflecting greater prediction error and implying important information is stored in that parameter
We find the permutation results are different for the first modes of GPP and LHF, where the skill of the GPP emulator is dominated by one parameter in particular, kmax, and none of the other parameters is important
Summary
Land models were originally developed to provide lower boundary conditions for atmospheric general circulation models but have evolved to simulate important processes such as carbon cycling, ecosystem dynamics, terrestrial hydrology, and agriculture. Including these societally relevant processes helps provide insight into potential impacts on humans and ecosystems and introduces additional sources of uncertainty in model predictions. The model spread in predictions of land– atmosphere carbon fluxes was found to be strongly tied to process representation, such as model treatment of carbon dioxide fertilization and the nitrogen cycle This situation has remained unchanged in the latest iteration of the Climate Model Intercomparison Project (CMIP6) (Arora et al, 2020). The trained networks are applied to globally estimate optimal parameter values with respect to observations, and these optimal values are tested with CLM5 to investigate changes in model predictive skill
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
