This repository contains code for Balasus et al. (2024).
I use micromamba as a package manager. You can install it here. Then, from the command line, run:
micromamba create -f environment.yml
The data needed are (1) blended TROPOMI+GOSAT and (2) HRRR data (winds, temperature, surface pressure, precipitation). First, go to config.json
and specify the directories where each of these datasets can be stored. Then, run the corresponding Python script for each in scripts/
. I use SLURM to allocate resources for each job:
sbatch -J blended -p sapphire -t 0-24:00 --mem=1000000 -c 112 \
--wrap "source ~/.bashrc; micromamba activate ldf_env; \
python -B -m scripts.download-blended"
months=()
for i in {0..60}; do months+=( $(date -d "20190101+${i} month" +%Y-%m-%d) ); done
for i in {0..59}; do
sbatch -J hrrr -p sapphire -t 0-24:00 --mem=500000 -c 48 \
--wrap "source ~/.bashrc; micromamba activate ldf_env; \
python -B -m scripts.download-hrrr ${months[i]} ${months[i+1]}"
done
The file config.json
lists the four landfills along with their coordinates and GHGRP IDs. The Python script assemble.py
will oversample TROPOMI data for 2019-2023 for each of the landfills, apply a cross sectional flux analysis to infer emissions from the resulting plumes, and scrape relevant data from the EPA FLIGHT tool.
sbatch -J assemble -p sapphire -t 0-06:00 --mem=64000 -c 4 \
--wrap "source ~/.bashrc; micromamba activate ldf_env; \
python -B -m scripts.assemble"
The notebooks epa-ghgrp-figs.ipynb
and individual-landfill-figs.ipynb
then produce the figures for the paper.