Search

A field campaign consisting of atmospheric boundary layer (ABL) observations and tracer experiments was conducted in September 2016 in a hilly region of the Gobi Desert in northwest China. High-frequency wind shears were observed between 70 and 200 m based on both 100-meter tower data and wind profiles. The ABL observations indicated that sub-layers of different atmospheric stability present within the ABL weakened the vertical transfer of momentum and tracer gas, resulting in low wind speed and high surface concentration. The CALMET model showed thermal heterogeneity to some extent in the temperature fields and mixing height distributions, and large deviations in the friction velocity between the model and observations were noted. The comparison of the CALPUFF model with the tracer measurements indicated a general tendency toward under-prediction of the concentration in the far distance because of wind shears and vertical thermal discontinuity. The results of the SCIPUFF model showed the same trend as those of CALPUFF, and comparisons demonstrated a good consistency of the two models.

CALPUFF is an atmospheric source-receptor model recommended by the US Environmental Protection Agency (EPA) for use on a case-by-case basis in complex terrain and wind condition. As the bulk of validation of CALPUFF has focused on long-range or short-range but long-term dispersion, we can not gauge the reliability of the model for predicting the short-term emission in near-field especially complex terrain, and sometimes this situation is important for emergency emission. To validate the CALPUFF’s application in such condition, we carried out a tracer experiment in a near-field complex terrain setting and used CALPUFF atmospheric dispersion model to simulate the tracer experiment in real condition. From the centroid trajectory comparison of predictions and measures, we can see that the model can correctly predict the centroid trajectory and shape of tracer cloud, and the results also indicate that sufficient observed weather data only can develop a good wind field for near-field. From the concentration comparison in each arc, we can see the model underestimate horizontal extent of tracer puff and can not reflect the irregular characters showed in measurements. The result of global analysis is FOEX of −25.91%, FA2 of 27.06%, FA5 of 61.41%. The simulations shows that the CALPUFF can simulate the position and direction of tracer cloud in near-field complex terrain but underestimate over measurements especially in peak concentrations.