TELL weather forcing sample data

This dataset contains three years of sample data generated by the sequence of processing scripts that convert the meteorology from IM3's climate simulations using the Weather Research and Forecasting (WRF) model into input files ready for use in the Total ELectricity Load (TELL) model. The first step in the processing chain spatially averages the gridded meteorology output from WRF into county mean values. The output of that processing step is a series of .csv files (one for every hour processed) with the county-mean value of six meteorological variables: T2, Q2, U10, V10, SWDOWN, and GLW. The second step then takes these county-level hourly values and population-weights them into an annual time-series for each of the balancing authorities (BAs) used in the TELL model.

The sample dataset includes three years of output for both processing steps: 2019, 2059, and 2099. The 2019 data is based on observed historical meteorology simulated by WRF. In contrast, the 2059 and 2099 sample data comes from IM3's future WRF runs under the RCP 8.5 climate scenario with SSP5 population forcing. Additionally, we include the full historical meteorology for a single BA (California Independent System Operator; CISO) in case you want to explore retraining the TELL machine-learning models using a longer historical record.

Finally, we include a sample output dataset from the United States version of the Global Change Analysis Model (GCAM-USA) that can be used in the TELL quickstarter notebook.

The code used to generate this sample dataset and more information about the wrf_to_tell processing pipeline can be found at: https://github.com/IMMM-SFA/im3components. More information about the TELL model can be found at: https://immm-sfa.github.io/tell/user_guide.html.