python_excel_differences.txt

The primary differences between the Python and Excel versions of the SWEET methane emissions model are:

1) The Python model allows for composting and anaerobic digestion of paper and cardboard.
2) The Python model landfills waste that is rejected by combustion/incineration facilities.
3) The Python model uses 2019 IPCC defaults for compost and combustion/incineration when input data is not available.

Decay engine (checked against SWEET v4.0.6 by scripts/excel_engine_check.py; tests/test_excel_engine_parity.py replays the results):

4) Given the same inputs, the annual engine generates exactly the same methane as Excel.
5) Its emissions match Excel exactly too. Until 2026-10 Python oxidised the gas a flare fails to destroy, which Excel lets out through the flare as is; that made Python about 0.8% low at a landfill with 60% capture.
6) The monthly engine, which the TRACE pipeline runs, is within 0.5% of Excel over a site's life and within 4% in any one year. The largest gaps are in the first year after closing and in leap years.
7) Excel is wrong in one cell: 'I. CH4'!C64 (site 1, food, 2015) holds the 2016 formula, so Excel's 2015 food methane is really 2016's.