HighlightsAERMOD and SCIPUFF were employed to back-calculate farm-level PM10 emission rates based on inverse modeling.Both AERMOD and SCIPUFF did not capture the diurnal and seasonal variations of farm-level PM10 emission rates.AERMOD modeling results were affected by wind speed, with higher wind speed leading to higher emission rates.Higher numbers of receptors and PM10 measurements with greater time resolution may be recommended in the future.Abstract. Air pollutant emissions from animal feeding operations (AFOs) have become a serious concern for public health and ambient air quality. Particulate matter with aerodynamic equivalent diameter less than or equal to 10 µm (PM10) is one of the major air pollutants emitted from AFOs. To assess the impacts of PM10 emissions from AFOs, knowledge about farm-level PM10 emission rates is needed but is challenging to obtain through field measurements. The inverse dispersion modeling approach provides an alternative way to estimate farm-level PM10 emission rates. In this study, two dispersion models, AERMOD and SCIPUFF, were employed to back-calculate farm-level PM10 emission rates based on hourly PM10 concentration measurements at four downwind locations in the vicinity of a commercial egg production farm in the southeast U.S. Onsite meteorological data were simultaneously recorded using a 10 m weather tower to facilitate the dispersion modeling. The modeling results were compared with PM10 emission measurements from two layer houses on the farm. Single-area source, double-area source, and double-volume source were used in AERMOD, while only single-point source was used in SCIPUFF. The inverse modeling results indicated that both SCIPUFF and AERMOD did not capture the diurnal and seasonal variations of the farm-level PM10 emission rates. In addition, the AERMOD modeling results were affected by wind speed, and higher emission rates may be predicted at higher wind speeds. The single-point source for SCIPUFF, the plume rise simplification for AERMOD, and insufficient concentration measurement resolution in response to temporal changes in wind direction may have added uncertainties to the modeling results. The results of this study suggest that more receptors covering more representative downwind locations should be considered in future modeling for farm-level emissions assessment. Moreover, ambient data collection with greater time resolution (e.g., less than one hour) is recommended to capture diurnal and seasonal patterns more rigorously. Only in this way can researchers achieve a better understanding of the effectiveness of inverse dispersion modeling for estimation of pollutant emission rates. Keywords: AERMOD, Animal feeding operations, Egg production, Farm-level emission rate, Inverse dispersion modeling, PM10, SCIPUFF.