Parameterizing Heat and Moisture Fluxes using a New Three-Dimensional Turbulence Velocity Scale

Better representations of surface-layer (SL) exchanges of heat and moisture are crucial for land-surface and boundary layer sub-models, as these two processes significantly influence the prediction accuracy of meteorological and air quality models. In this field-scale study, we introduce a similarity function-independent three-dimensional turbulence velocity scale (e_*), which is a function of friction and convection velocities and the Monin-Obukhov lengthscale, into SL parameterizations for sensible heat (H) and latent heat (LE) fluxes. We use two different methods estimating H and LE fluxes: (1) Monin-Obukhov Similarity Theory (MOST), and (2) a bulk-Richardson (Ri_b) approach suggested recently. We evaluate the impact of introducing e_* on the parameterized fluxes using one year (i.e., 1 January through 31 December 2018) of measurements sampled 10 m above ground level (AGL) from two micrometeorological towers in the southwestern U.S: a 10-m tower in southeastern Arizona and a 200-m tower in northwestern Texas.

Results indicate that incorporating e_* into the MOST and Ri_b parameterizations significantly improve the accuracy of the parameterized fluxes. For the site in Arizona, the root-mean square error (RMSE) for H and LE is 35 W m^-2 and 98 W m^-2, respectively, when using the traditional MOST parameterizations. However, when we use the MOST parameterizations that incorporate e_* scaling, the RMSE is reduced to 23 W m^-2 and 73 W m^-2 for H and LE, respectively. Furthermore, the slope of the line of best fit (m_b) between the parameterized and observed values approaches unity when is incorporated into MOST parameterizations with m_b values of 0.88 and 1.16 for H and LE, respectively, compared to 1.17 and 1.55, respectively, for the traditional MOST parameterizations. Similar improvements are found when using the observations from the 200-m tower site in Texas. Our findings, coupled with recently-reported findings in less arid regimes, help to motivate the potential use of in parameterizations of SL fluxes, which form the groundwork for the land surface and boundary layer sub-models used in weather forecasting, air pollution, and climate models.