Temperature controls many processes of NH3 volatilization. For example,urea hydrolysis is an enzymatically catalyzed reaction described by the Arrhenius equation. Diet crude protein (CP) controls NH3 emission by affecting N excretion. Our objectives were to use the Arrhenius equation to model NH3 emissions from beef cattle (Bos taurus) feedyards and test predictions against observed emissions. Per capita NH3 emission rate (PCER),air temperature (T),and CP were measured for 2 yr at two Texas Panhandle feedyards. Data were fitted to analogs of the Arrhenius equation: PCER = f(T) and PCER = f(T,CP). The models were applied at a third feedyard to predict NH3 emissions and compare predicted to measured emissions. Predicted mean NH3 emissions were within - 9 and 2% of observed emissions for the f(T) and f(T,CP) models,respectively. Annual emission factors calculated from models underestimated annual NH3 emission by 11% [f(T) model] or overestimated emission by 8% [f(T,CP) model]. When T from a regional weather station and three classes of CP drove the models,the f(T) model overpredicted annual NH3 emission of the low CP class by 14% and underpredicted emissions of the optimum and high CP classes by 1 and 39%, respectively. The f(T,CP) model underpredicted NH3 emissions by 15,4,and 23% for low,optimum,and high CP classes,respectively. Ammonia emission was successfully modeled using T only,but including CP improved predictions. The empirical f(T) and f(T,CP) models can successfully model NH3 emissions in the Texas Panhandle. Researchers are encouraged to test the models in other regions where high-quality NH3 emissions data are available.