A method and results for increasing spatial and temporal resolution of NH­₃ emissions from chemical fertilizer usage (NH₃-CFU) are presented. Such emissions are used as inputs to chemical transport models (CTMs) to estimate particulate matter concentrations and reactive nitrogen deposition. Typically, emission inputs are obtained from the National Emissions Inventory (NEI). NH₃-CFU are estimated by combining annual fertilizer sales reported at county level with fertilizer-specific emission factors. However, inputs to CTMs are required at finer spatial resolutions and hourly temporal scale. The Sparse Matrix Operator Kernel Emissions (SMOKE) model is used to bridge resolution gaps between NEI and CTMs by use of spatial surrogates and temporal factors, which could be further improved.

Spatial surrogates within SMOKE are developed by estimating the ratio of cropland within 4x4 km² grid to net cropland within county area. This approach does not consider crop fertilizer requirements that result in spatial heterogeneity in NH₃ emissions at sub-county resolutions. The Improved Spatial Surrogate (ISS) method was developed to modify the existing spatial surrogate within SMOKE by incorporating annual cropland distribution and crop-specific nitrogen demands. Results for a test domain of 4x4 km² grids over Central Illinois indicate large variations in grid based differences in estimates between SMOKE and ISS. Such differences range between -10% - 120% with 58% of the grid cells exhibiting more than ± 10% difference. Applicability of the ISS method is currently being tested by upscaling to (1) 4x4 km² over Midwest USA and (2) 12x12 km² across continental USA.

Hourly temporal factors in SMOKE are currently estimated by equally disaggregating emissions within each crop season proportionately to the nitrogen applied. This excludes influences of local weather and soil conditions. In this research, the process based DeNitrification DeComposition (DNDC) model was employed to model daily variations in NH₃-CFU within the test domain. For the years 2002-2011, mean NH₃ emissions from DNDC were within ±15% of SMOKE and ISS estimates. Inter-annual temporal patterns were similar in distribution but varied in magnitudes by ±20%. However, individual emission peaks on days post fertilization were 2.5-8 times greater than those estimated by SMOKE.

By providing alternate approaches to bridge spatial and temporal resolution between NEI and CTMs, this study could assist in improving modeling predictions of atmospheric particulate matter and deposition of reactive nitrogen.