This project has had the following findings: regarding convective initiation (CI), a climatology of pristine nocturnal CI found a dual-peak pattern for initiation time, with one peak at 04 and the other at 07 UTC (Stelten and Gallus 2017). It also found that convection-allowing models run during PECAN were more deficient with location than with timing for CI. Additional analysis of CI events during PECAN found that cases near other MCSs and with frontal overrunning were more predictable than events occurring farther from existing convection, whether or not the CI was within a LLJ (Weckwerth et al. 2019).
Three studies were performed using modifications of the MYNN PBL scheme that included changes in eight closure parameters based on observations, an increase in mixing length for a stable environment with strong bulk wind shear, and an explicit accounting of turbulent potential energy that is non-zero under stable conditions. Marked sensitivity in 80 m winds was found for three of the eight closure parameters (Jahn et al. 2017a, b) with a reduction in error present when a 25% reduction was made to two of the parameters, but a large increase in error when the third parameter was reduced by the same amount. When the closure parameter changes were made in conjunction with the change in mixing length, an average 17% reduction in wind forecast error occurred in most cases. A 13% reduction occurred in all cases for the change accounting for turbulent potential energy. When the MYNN, modified MYNN, and YSU schemes were tested for LLJ cases during PECAN, the modified MYNN better simulated the LLJ winds than the original MYNN scheme, matching well the YSU results (Jahn and Gallus 2018). However, YSU had a warm and dry bias that was not present in the MYNN runs. The modified MYNN scheme worked best overall for simulating these events. Jahn and Gallus (2018) also found that RAP analyses agreed well with PECAN soundings overall for LLJ events. Ongoing work is examining the relationship between errors in late afternoon boundary layer characteristics associated with different PBL schemes in WRF and errors in depiction of the LLJ.
Convective system evolution and its predictability were examined in several other studies. Thielen and Gallus (2019) found that refinement of horizontal grid spacing from 3 to 1 km increased the number of linear systems simulated, although the number was still less than those observed, but a morphological skill score did not significantly improve, suggesting that lines were predicted at the wrong times or in the wrong cases. The increase in linear modes was due to a filling in of gaps in high reflectivity values at 1 km that were present at 3 km, likely due to stronger ascent all along the cold pool in 1 km runs. The work also suggested that the underprediction of bow echo and trailing stratiform squall line events primarily is due to deficiencies during the initiation and dissipation stages of MCSs, with less of a problem during the mature stage. Substantial variability was found as microphysics schemes were changed. Carlberg et al. (2018) found that a small WRF ensemble had some skill at predicting broad morphology categories but was less skillful for specific mode types. Squitieri and Gallus (2020) examined differences in cold pool structure in more detail for runs with 3, 1, and 0.333 km horizontal grid spacing, finding stronger and deeper cold pools with finer grid spacing, with some convergence between the 0.333 and 1 km results. Ongoing work is examining if differences in the simulated balance between shear and cold pool-generated horizontal vorticity in some PECAN cases may explain why the upscale evolution is simulated better in some cases in WRF than in others.
Finally, regarding predictability of convection, Vertz et al. (2020) found that for MCS events associated with LLJs in strongly forced environments, a statistically significant correlation existed between moisture and theta-E errors in the inflow regions of MCSs in WRF simulations and the displacement errors for initiating MCSs, such that larger dry biases resulted in greater downstream displacements. The relationship was much weaker for events with weaker synoptic forcing. Also, Gallus et al. (2019) used two Community Leveraged Unified Ensemble sub-ensembles, one with mixed physics and one without, to find that greater spread of solutions in short term QPF and reflectivity was present when mixed physics were used compared to only mixed initial and lateral boundary conditions, but skill did not differ substantially in the two ensembles.
More recently, additional funding has been provided to continue the work but by adding in FV3 simulations and idealized CM1 simulations to better understand what is happening when the model runs display some common failures, such as depicting lines without stratiform rain when trailing stratiform rain is observed, or showing broken lines of cells when squall lines are observed. These runs will also explore in more detail how upscale growth is occurring in nocturnal systems.
REFEREED PUBLICATIONS FROM THIS WORK:
Carlberg, B., W. A. Gallus, Jr., and K. Franz, 2018: A preliminary examination of WRF ensemble prediction of convective mode evolution. Wea. Forecasting, 33, 783-796.
Gallus, W. A., Jr., J. Wolff, J. Halley Gotway, and M. Harrod, 2019: The impacts of using mixed physics in the Community Leveraged Unified Ensemble. Wea. Forecasting, 34, 849-867.
Jahn, D., E. S. Takle, and W. A. Gallus, Jr., 2017a: Wind ramp forecast sensitivity to boundary layer scheme closure parameters. Bound. Layer Met., 1-16. Doi:10.1007/s10546-017-0250-5.
Jahn, D., E. S. Takle, and W. A. Gallus, Jr., 2017b: Improving numerical wind forecasts of wind ramps at 100m height in the stable boundary layer. Bound. Layer Met., 163, 423-446, doi: 10.1007/s10546-017-0237-2.
Jahn, D. E. and W. A. Gallus, Jr., 2018: Impacts of modifications to a local planetary boundary layer scheme on Great Plains low-level jet forecasts. Wea. Forecasting, 33, 1109-1120.
Squitieri, B. J., and W. A. Gallus, Jr., 2020: On the forecast sensitivity of MCS cold pools and related features to horizontal grid spacing in convection-allowing WRF simulations. Wea. Forecasting, 32, 325-346.
Stelten, S. and W. A. Gallus, Jr., 2017: Pristine nocturnal convective initiation: A climatology and preliminary examination of predictability. Wea. Forecasting, 32, 1613-1635.
Thielen, J., and W. A. Gallus, Jr., 2019: Horizontal grid spacing influences on WRF forecasts of convective morphology evolution for nocturnal MCSs in weakly-forced environments. Wea.Forecasting, 34, 1495-1517.
Vertz, N., W. A. Gallus, Jr., and B. J. Squitieri, 2020: Relationship of MCS initiation errors to moisture errors in the inflow region. Wea. Forecasting (under revision for re-submission).
Weckwerth, T. M., J. Hanesiak, J. W. Wilson, X. Wang, R. D. Roberts, W. A. Gallus, Jr., and S. B. Trier, 2019: Nocturnal convective initiation during PECAN 2015. Bull. Amer. Meteor. Soc., 100, 2223-2239.
Note: An additional 19 conference presentations have been given on the results from this grant