Meassurement Report: Methane and NOx emissions from natural gas cooking stoves: The case of Chile and Colombia

We measured emission rates of NOx (NO+NO₂), CO₂, CO and CH₄ (mg/min) for each stove under three operating modes:

Continuous leaks (“Steady-State Off”): burners off, kitchen sealed with forced air circulation for ~10 min to quantify CH₄ leaking from closed valves.

On/Off events: short pulses of CH₄ measured immediately after turning a burner on or off.

Combustion (“Steady-State On”): one burner lit with a pot of water on it to mimic normal cooking, kitchen vacated and sealed for 6–10 min while concentrations of CO₂, CO, NOₓ and CH₄ rose.

To calculate emission rates, we treated the kitchen as a well-mixed chamber of known volume V0 with a known air exchange rate lambda. Assuming negligible chemical reactions, changes in gas concentration over time follow a simple mass balance:

V0 * dCi/dt = Ei - lambda * V0 * (Ci - Ci_b)

where Ci is the gas concentration inside the kitchen, Ci_b is the background concentration, and Ei is the emission rate. Sensor readings (mole fractions) were converted to concentrations using the measured temperature and pressure. We closely followed the procedure of Kashtan et al. (2023); further details are given in the Supplementary Material.

This method was applied to any period where the stove operated in steady state, allowing us to determine an average emission rate Ei_bar (mol/min) for each operating mode and for individual burners. Absolute emission rates (mol/min or mg/min) can be combined with typical use times to estimate total emissions. They can also be converted into emission factors (kg/TJ) by normalizing by the amount of natural gas burned, estimated from the combined carbon emitted as CO₂, CO and CH₄ and the known fuel composition and lower heating value (LHV) in each city (Bogotá and Santiago).

The enclosed volume V0 was obtained from geometric measurements of the kitchen, subtracting the volume occupied by furniture and appliances; this approach agrees within ~3% with tracer-gas methods. The air exchange rate lambda (1/min) was determined from the decay of CO₂ concentration after combustion experiments: once the burners were turned off and the kitchen sealed, we recorded CO₂ decay and fitted an exponential curve using the mass balance equation above with Ei = 0.