Effects of a Variety of WRF Urbanization Schemes on the Simulation of a Bifurcating Thunderstorm over Beijing

A Beijing urban-area bifurcating summer evening precipitating thunderstorm event is investigated. The case was observed during the Study of Urban Impacts on Rainfall and Fog/Haze (SURF) Project IOP on 22 July 2015. Radar precipitation reflectivity and AWS surface observations showed bifurcation of the storm as it approached and passed over the city. The AWS observations also showed that an hour before the storms reached Beijing, surface winds over the city had bifurcated around the city and thus the bifurcated storm-halves moved around the city in that flow.

Our previous observational and simulation studies over Beijing, NYC, and Atlanta have shown that urban induced building-barrier bifurcation exists under weak UHI conditions, but that strong UHIs produce storm-initiation over cities. Four twin WEF simulation sets (each with and without Beijing) were thus carried out: (a) BEP only; (b) BEP + unmodified BEM (called uBEM); (c) BEP + once modified BEM (with the cooling tower parameterization of Gutierrez, et al. 2015, J. of Solar Energy Engineering, called cBEM); and (d) twice modified dcBEM (like cBEM, but also with the new speed-dependent drag coefficient of Gutierrez, et al. 2015, Bound.-Layer Meteor.)

Results showed that the BEP only case with correctly reproduced the weak UHI, storm bifurcation, strong urban-induced deceleration (i.e., from 4 m/s over missing-Beijing to 2 m/s over the city). In the uBEM urban case, UHI intensity was overestimated (1.75 vs 0.5 K) due to the thermal forcing in the unmodified BEM parameterization. The new UHI value was thus greater than the observed climatological threshold that allows for rainfall bifurcation over Beijing. The precipitation bifurcation disappeared and was replaced by an erroneously produced urban-induced precipitation maxima over Beijing. This occurred as the cooling tower parameterization in cBEM correctly converted a fraction of the total anthropogenic sensible heat to latent heat. The excessive UHI frm uBEM was then corrected in the cBEM simulation, and the bifurcated precipitation pattern was again reproduced. Simulations with the new drag coefficient usually correct another error introduced in some urbanized WRF simulations, when their excess UHIs over-accelerated urban wind speeds.

These simulations thus show that (a) bifurcating thunderstorms are the most complex, and thus the most difficult, urban impacts to correctly simulate, as the model must first correctly reproduce the temperature and wind fields, (b) the new (most advanced) BEP + BEM formulation captures most urban impacts on wind velocity and precipitation, and (c) use of models with an incomplete urban parameterization can produce erroneous results.