Spatiotemporal discrepancies between satellite- and inventory-derived estimates of global ammonia emissions

Abstract Bottom-up inventories and satellite-constrained estimates of ammonia emissions often diverge, but the conditions driving these discrepancies remain unclear. Here we compare monthly global estimates at 0.1° resolution from 2008 to 2016 using a normalised discrepancy index. Satellite-constrained emissions are higher on average, with a land-mean index of 0.077. The discrepancy strengthens during warm months and in agricultural areas, reaching 0.229 in May, and increases with near-surface air and dew-point temperatures. Differences are larger in subsistence, low-input and rainfed systems than in high-input or irrigated systems, whereas associations with soil properties, precipitation and evaporation are weak. These patterns indicate that current inventories do not fully resolve the effects of local weather, land use and management timing. This study provides insights into bridging the top-down and bottom-up estimates into closer agreement.