EPA Stationary Combustion · GHG Emissions Quantification
Decides Scope 1 CO2, CH4, and N2O emissions from owned or leased boilers, furnaces, and other combustion sources by fuel type and activity data.
Overview
For any boiler, furnace, kiln, flare, or similar combustion source an organisation owns or leases, this package turns a fuel type and an activity quantity into a full emissions result: carbon dioxide, methane, and nitrous oxide in kilograms, fossil and biogenic CO2 reported separately, and a total CO2e figure in both kilograms and metric tons, each traceable to the specific Table 1 factors and equations that produced it.
It is built on the EPA GHG Emission Factors Hub's 2025 Table 1 for stationary combustion, covering solid, gaseous, and liquid fossil fuels, biomass fuels, Kraft pulping liquors, and mixed fossil-and-biomass waste streams such as municipal solid waste. Sources outside the owned-or-leased boundary are classified out of scope rather than calculated.
In practice it handles cases like: a leased boiler burning natural gas measured in standard cubic feet, a furnace running distillate fuel oil on a gallon basis, an incinerator burning municipal solid waste that needs a biogenic-fraction split between fossil and biomass CO2, and a facility drawing down fuel stock inventory where the quantity purchased differs from the quantity actually combusted.
Where the data doesn't support a calculation — an activity unit that isn't published for the fuel, a lower-heating-value figure for a fuel family with no published conversion, a mixed waste stream with no biogenic-fraction supplied, or a non-default heat or carbon content lacking supporting documentation — it stops and asks for the specific missing input rather than guessing or substituting a default.
Every run returns a structured verdict — classification tags, the method used, and the computed emissions figures — together with reason codes for any missing evidence, so results drop cleanly into downstream reporting or review workflows.
What it decides
Try it — live sandbox
Build a case and run it against the publisher's live production release — the exact rules a subscriber calls. Run the same case twice and the answer is identical: no LLM in the loop.
Run a case to see the deterministic decision — tags, routes, reasons, evidence, and facts. Same case in, same answer out, every time.
Integrate — two ways
Call it directly over HTTP, or connect your coding agent over MCP. Same decision, same metering, either way.
One authenticated POST — the same deterministic, metered evaluation the sandbox runs, no LLM in the loop.
# Request shape only. Replace sert_mk_YOUR_KEY and send every field from the Inputs contract above.
curl -X POST https://api.sertainly.ai/api/marketplace/v1/packages/stationary_combustion_emissions_quantification/evaluate \
-H "Authorization: Bearer sert_mk_YOUR_KEY" \
-H "content-type: application/json" \
-d '{ "case": { "reporting_period_start": "example", "reporting_period_end": "example", "source_type": "boiler" } }'