One‐at‐a‐Time Parameter Perturbation Ensemble of the Community Land Model, Version 5.1

• Published in: Journal of advances in modeling earth systems Vol. 17; no. 8
• Main Authors: Kennedy, D., Dagon, K., Lawrence, D. M., Fisher, R. A., Sanderson, B. M., Collier, N., Hoffman, F. M., Koven, C. D., Kluzek, E., Levis, S., Lu, X., Oleson, K. W., Zarakas, C. M., Cheng, Y., Foster, A. C., Fowler, M. D., Hawkins, L. R., Kavoo, T., Kumar, S., Newman, A. J., Lawrence, P. J., Li, F., Lombardozzi, D. L., Luo, Y., Shuman, J. K., Swann, A. L. S., Swenson, S. C., Tang, G., Wieder, W. R., Wood, A. W.
• Format: Journal Article
• Language: English
• Published: Washington John Wiley & Sons, Inc 01.08.2025
• Subjects: Automation; Biogeochemistry; Calibration; Carbon; Climate change; CLM5.1; community land model; Hydrology; Land use planning; Nitrogen; Parameter estimation; parameter perturbation experiment; Parameters; PPE; Simulation; Software; Uncertainty
• ISSN: 1942-2466; 1942-2466
• DOI: 10.1029/2024MS004715

Comprehensive land models are subject to significant parametric uncertainty, which can be hard to quantify due to the large number of parameters and high model computational costs. We constructed a large parameter perturbation ensemble (PPE) for the Community Land Model version 5.1 with biogeochemistry configuration (CLM5.1‐BGC). We performed more than 2,000 simulations perturbing 211 parameters across six forcing scenarios. This provides an expansive data set, which can be used to identify the most influential parameters on a wide range of output variables globally, by biome, or by plant functional type. We found that parameter effects can exceed scenario effects and that a small number of parameters explains a large fraction of variance across our ensemble. The most important parameters can differ regionally and also based on the forcing scenario. The software infrastructure developed for this experiment has greatly reduced the human and computer time needed for CLM PPEs, which can facilitate routine investigation of parameter sensitivity and uncertainty, as well as automated calibration. Plain Language Summary The Community Land Model includes a large set of numerical settings that help describe attributes of the various components of the land system. Each setting has a default value, but we know that other values may also be reasonable within a certain range. We ran a large set of simulations, increasing and decreasing each setting independently to better understand its influence on model outputs, such as plant productivity and the water cycle. We repeated these experiments across a range of scenarios, including present‐day conditions and introducing (or removing) various aspects of climate change. We found that changing certain model settings could influence our results as much as the influence of climate change, itself. We also found that the most influential settings varied by geographic region. Understanding the influence of all of these settings can help us improve our model and also help us gauge our confidence in model predictions. Key Points We constructed a parameter perturbation ensemble of the Community Land Model, v5.1, perturbing 211 parameters across six forcing scenarios Parameter effects can exceed scenario effects and parameter effect rankings differ by biome and based on the forcing scenario The software infrastructure developed in our experiment has greatly reduced the human and computer time needed for constructing future PPEs