Graduate Seminar

            Aircraft, surface-flux tower, and boundary layer radar wind profiler data were gathered during fair-weather boundary-layer heterogeneity missions to look at the role of surface heterogeneity in determining heat and moisture fluxes, and convective boundary layer (CBL) structure and evolution.  The observational results are compared with results from the Noah land surface model (LSM) and the Advanced Research WRF (ARW) numerical weather prediction model to assess the behavior of both models and to gain some insight into PBL and flux behavior.  Numerical simulations are initiated using soil moisture and temperature profiles derived from the High Resolution Land Surface Data Assimilation System (HRLDAS). Throughout the comparisons, the major challenge has been to understand the strengths and limitations of both the data and the modeling system, where the term “system” is used to include the model input data as well as the model physics.

             After discussing the model and observations, I focus on how we have evaluated the aircraft-derived fluxes.  While statistical tests have often been used as a measure of flux certainty, we have found that plots of sensible heat H vs latent heat flux LE for averaged data at different locations in the measurement domain are particularly useful for days with nearly uniform net radiation.  Under these conditions, H and LE for surface data fall on a straight line with a negative slope (DLE/DH), as do the simulated fluxes.  We have found that aircraft fluxes usually need to be horizontally averaged for this to occur.  Moreover, departures from the slope line seem to be due to statistical scatter or focusing of flux by atmospheric processes.  This enables us to isolate the fluxes the Noah LSM needs to replicate.

             Comparison of observed fluxes to those from the Noah LSM reveal biases similar to those we have seen in other “control” simulations:  too-high H but LE about right; horizontal variability that was attributed to differences in land cover were not captured because the input land cover was uniform.  The H biases translate into a too-deep PBL in the simulations.  However, the appearance and duration of low (cumulus) cloudiness in the model correspond to the observations for three of the four days.  This suggests that drying due to downward mixing of dry air above the PBL raises the condensation level enough to compensate for the bias in PBL depth.  WRF appears to replicate the observed CBL structure at the ~3 km and ~50 km scales.