Relationship between skill at forecasting the Great Plains Low Level Jet and skill at forecasting mesoscale convective systems in 4km WRF runs

- Indicates paper has been withdrawn from meeting
- Indicates an Award Winner
Monday, 3 February 2014: 5:00 PM
Room C202 (The Georgia World Congress Center )
Brian Joseph Squitieri, Iowa State University, Ames, IA; and W. A. Gallus Jr.

The Low Level Jet (LLJ) is often poorly simulated by forecast models. Given the inaccuracy of both the timing and the placement of the terminus of the LLJ and its maximum winds (both in a horizontal extent and with height), it is understandable that errors often exist in Mesoscale Convective System (MCS) forecasts. It is thus necessary to simulate the LLJ with high horizontal and vertical spatial resolution in order to investigate details of the LLJ depth, height and magnitude and to see how variability in these features affects MCS development. Twenty individual cases involving LLJ-influenced nocturnal MCS's were simulated with the Weather Research and Forecasting Model (WRF). Each case was simulated using the same 400X400 km domain centered in the eastern Great Plains with 4 km grid spacing and 50 defined eta levels (with 28 eta levels established below 850 mb) . Given that Planetary Boundary Layer (PBL) characteristics influence LLJ development, changes in the PBL were investigated by using the Yonsei University (YSU), Mellor-Yamada-Janjic (MYJ) and Mellor-Yamada-Nakanishi-Niino (MYNN) PBL schemes in the WRF simulations, with all schemes employed under both the Thompson and the WRF Single-Moment (WSM6) microphysics schemes. To compare high temporal and vertical spatial resolution LLJ simulations with observations, 60 m resolution wind data from the Lamont, OK Vertical Profiler (Southern Great Plains Central Facility) was used to calculate the Mean Absolute Error (MAE) in wind direction, magnitude and vertical placement of the LLJ at a given point. Rapid Update Cycle 0-hour analyses (Rapid Refresh Model 0-hr analyses for dates in 2012/2013) were used in the absence of a dense wind profiler network across the plains for high temporal and horizontal spatial resolution analysis, with the MAE computed for the WRF winds with respect to the RUC-Analysis winds. To determine the forecast skill in the placement, timing and magnitude of the MCSs in the WRF simulations, an objective verification was conducted for WRF simulated rainfall using STAGE IV observational rainfall data to compute Equitable Threat Scores (ETSs). Previous research suggests that LLJ depths and magnitudes of peak winds are often under-forecasted by the models, mainly due to poor simulations of the PBL. (Coniglio et al, Weather and Forecasting, 2013) found that multiple PBL schemes (including the MYJ) in the WRF portray a boundary layer with a shallow and moist bias. However, implementing the MYNN scheme greatly reduced the shallow/moist bias in the PBL. A significant research task is to test the MYNN scheme against the MYJ and YSU to see if the MYNN greatly reduces the MAE for the depth, peak wind magnitude, and height of the peak wind magnitude, thus improving the forecast skill for the LLJ. As such, the main goal of this research is to determine if an increase in forecast skill with the LLJ directly correlates with an increase in forecast skill with the MCSs under different PBL schemes.