ML-IAM: Supporting Data for Machine Learning Emulator of Integrated Assessment Models

Superseded — merged into https://doi.org/10.5281/zenodo.17390677. (ML-IAM repository). Please cite that record instead. This dataset contains supporting data files required for reproducing ML-IAM, a machine learning emulator for Integrated Assessment Models. The dataset includes: 1. Base year mappings for 189 IAM model versions (unique_models_all_with_base_year.csv) - Maps each IAM model to its base year for temporal alignment - Includes literature references for verification 2. Input/output variable classifications for 539 variables (variable_classification_1019.csv) - Categorizes AR6 database variables as model inputs or outputs - Includes units, definitions, and classification rationale These files are essential for preprocessing the IPCC AR6 Scenarios Database and training the ML-IAM models. They complement the source code available at https://doi.org/10.5281/zenodo.17390678. Related publication: "ML-IAM: Emulating Integrated Assessment Models With Machine Learning" (submitted to Geoscientific Model Development) The IPCC AR6 Scenarios Database used for model training is available at https://doi.org/10.5281/zenodo.7197970