The META 21 Integrated Assessment Model in GAMS and LHS Sampling

The META 21 Integrated Assessment Model (Dietz et al., 2021) represents a fairly comprehensive climate change simulation model incorporating a number of features: (a) a recent simple climate model; (b) down-scaling of temperature change to the country level; (c) integration of a number of bio-physical tipping points linked to rising temperatures; and (d) economic damages linked to rising sea levels and temperature using recent country-level estimates from the literature. The original implementation of META 21 was in Excel and linked to the @Risk Excel add-in for performing Monte Carlo-type analysis. This paper reflects a translation of META 21 into GAMS. One of the key purposes is to allow for ready incorporation of many of the features of META 21 into other models—notably CGE-based IAMs. To replace the features provided by the @Risk Excel add-in, this paper also introduces a software package that produces random deviates—and, similar to @Risk, uses the Latin Hypercube Sampling (LHS) approach, which is a stratified sampling technique intended to cover the entire sampling space efficiently and, in addition, orders the sample to reflect a desired correlation matrix for the sampled random variables.