资源环境科技发展态势分析平台

Statistical models of high-resolution climate models are useful for many purposes, including sensitivity and uncertainty analyses, but building them can be computationally prohibitive. We generated a unique multiresolution perturbed parameter ensemble of a global climate model. We use a novel application of a machine learning technique known as random forests to train a statistical model on the ensemble to make high-resolution model predictions of two important quantities: global mean top-of-atmosphere energy flux and precipitation. The random forests leverage cheaper low-resolution simulations, greatly reducing the number of high-resolution simulations required to train the statistical model. We demonstrate that high-resolution predictions of these quantities can be obtained by training on an ensemble that includes only a small number of high-resolution simulations. We also find that global annually averaged precipitation is more sensitive to resolution changes than to any of the model parameters considered.

Plain Language Summary We rely on climate models to make future projections of important climate phenomena. Many physical processes occur on scales smaller than the grid size of the model and are therefore encoded via parameters that approximate the subgrid-scale processes. High-resolution models are more accurate and also more computationally expensive. We use machine learning to provide a mapping between the input parameters and two important globally and annually averaged climate model outputs: energy balance and precipitation. The machine learning algorithm is able to learn the mapping for the high-resolution model by only using a small number of high-resolution runs, but supplemented with much cheaper lower resolution runs, thereby greatly reducing the computational expense. The mapping allows one to make high-resolution predictions for different values of input parameters and also can be used to estimate the parameter values given observed data. These steps are important in improving the model's ability to simulate the climate system.