Abstract
This paper presents EnsemFire v1.0, a global ensemble fire emission dataset that provides daily emissions at 0.1° × 0.1° spatial resolution for key air pollutants like fine particulate matter (PM₂.₅), black carbon (BC), organic carbon (OC), carbon monoxide (CO), ammonia (NH₃), nitrogen oxides (NO(x)), and sulfur dioxide (SO₂), greenhouse gases including carbon dioxide (CO₂) and methane (CH₄), and fire radiative power (FRP). EnsemFire integrates seven widely used biomass burning emission inventories, including five global datasets (GFAS, FINN, FEER, QFED, GBBEPx) and two regional products (EPA and CFFEPS). Our analysis reveals noticeable inconsistencies among these datasets, reflecting the large uncertainty in biomass burning emission estimates. By applying an ensemble approach, EnsemFire reduces this uncertainty and provides a more robust emission estimate. When used as input to the Unified Forecast System (UFS) model, EnsemFire significantly reduces simulation bias and improves the model performance to predict aerosol optical depth (AOD) compared to the control run that uses the default emission input. This dataset offers a valuable resource for atmospheric modeling, air quality forecasting, and climate research.