This repository contains modeling data (source code link and information, model input and output files) used in the analysis reported in the manuscript entitled “Power sector investment implications of climate impacts on renewable resources in Latin America and the Caribbean”, currently under review in Nature Communications.
- Source Code
This analysis employed the GCAM-LAC model (a research branch of the open-source Global Change Analysis Model ─ GCAM). The source code of GCAM-LAC is available at:
Silvia R. Santos da Silva, Mohamad Hejazi, Gokul Iyer, Thomas B. Wild, Matthew Binsted, Fernando Miralles-Wilhelm, … Chris R. Vernon. (2020, September 24). Silviameteoro/GCAM-LAC: GCAM-LAC_v5.1.3_Santos_da_Silva_et_al_2020 (Version GCAM-LAC_v5.1.3_Santos_da_Silva_et_al_2020). Zenodo. http://doi.org/10.5281/zenodo.4048788
- System Requirements for the source code installation
Any recent 64-bit operating system: Windows 10, Mac OSX 10.14, and Linux
Additional software and libraries include:
- Java 8+
- C++ compiler to compile source code with C++14 standard support such as GCC 5+ or Microsoft Visual Studio 2015+
- R 3.3+ for running the GCAM data system
- The boost C++ library 1.56+
- Xerces C++ library 3.1.1+
No non-standard hardware is required, however it is recommended that the user has 8 GB RAM.
- Installation Guide
The C++ source code is compiled according to these guidelines:
http://jgcri.github.io/gcam-doc/gcam-build.html
After compiling the source code, it is necessary to install the GCAM data system (written in R). Instructions to install the GCAM data system are available at:
https://github.com/JGCRI/gcamdata/wiki
Installation time should take about 15-20 minutes to compile the C++ code and another 30 minutes to build the GCAM data system.
- Instructions to run the model
Instructions to run the model can be found at (item 2 onwards) :
http://jgcri.github.io/gcam-doc/user-guide.html
Additional material can be found here: http://www.globalchange.umd.edu/data/gcam/GCAM_tutorial_2019.pdf
To run the specific scenarios of this study, it is necessary to use the configuration files shown in "Model Inputs/configuration_files" (see the README file within that folder). (**** Note that it is not necessary to download these configuration files. When the source code is downloaded, these files are downloaded with the code. ****)
Time required for each simulation varies as some scenarios are more challenging to solve than others. This varies from about 50 min to 2 hours on a "normal" desktop computer (assuming a minimum of 8 GB RAM).
- Outputs
Model outputs from each scenario is stored in an XML database. Each database requires about 2GB of space and stores a large number of parameters (most of them were not used in this work). Given that all databases produced for this study require more than 100 GB of space, this dataset is not hosted here. However, in the folder "Model Outputs" we provide all specific parameters (organized by scenario names) used in this analysis.