High resolution simulations from the Unified Model
I have been using the Met Office Unified Model (UM) to study the evolution of convective storms in the Southern UK. I have run the current version of the operational UKV (1.5 km grid length over the whole UK, see figure) on some of the DYMECS cases, I have also been doing some sensitivity studies changing aspects of the microphysics and sub-grid turbulence schemes. To test the dependence of resolution I have run a 500 m resolution model covering a 400 x 400 km region centred on Chilbolton (red box). I have also run a 200 m resolution model covering a 300 x 225 km region within the 500 m domain (green box). The UKV has 70 vertical levels and I have been running the 500m and 200m models with 70 and 140 levels.
To see animations of different DYMECS cases click on the appropriate link below. All rainrates have been smoothed to 1.5 km.
To see skew-T plots of different DYMECS cases click on the appropriate link below.
I have used a storm tracking algorithm to track the cells in the model data and the Nimrod data. Cells are tracked every 5 minutes. To see some cell statistics for these cases click on the links below. All data has been regridded to the 1 km Nimrod grid before calculating statistics.
To investigate the effect of horizontal resolution on storm size and updraught size I have produced scatter plots of vertical velocity against maximum rainrate and convective updraught area. To see these click on the links below. All data has been regridded to the 1 km Nimrod grid.
I have also been looking at vertical cross-sections of reflectivity, cloud ice water, vertical velocity and horizontal wind through the most intense storms.
UKV with prognostic graupel
I have run the UKV with a prognostic graupel scheme. Here are some rainfall animations and vertical cross-sections through the storms, comparing the standard UKV and the UKV with prognostic graupel. As before all rainrates have been smoothed to 1.5 km.
Varying the Smagorinsky mixing length
I have also been investigating the effect of varying the mixing length used in the Smagorinsky subgrid mixing scheme on the size of the cells. Here are some animations of rainrate from the UKV and 500 m models with the mixing length varying from 40 m to 500 m. As before all rainrates have been smoothed to 1.5 km. (Note the standard UKV shown above does not use the 3D Smagorinsky subgrid mixing scheme.)
You can see some statistics for these runs here:
3D radar scans from Chilbolton and UKV operational forecasts for the DYMECS cases are available on the DYMECS website
Predictability of convection in COPS: high-resolution ensemble forecasts from the Unified Model
Accurate forecasting of convective precipitation over complex terrain is important because of the substantial flooding that can be caused by the intense precipitation. In principle, the predictability of convection should be increased over orography, where localized vertical motions owing to specific topographic features can remove the convective inhibition and trigger cells. At present however, the forecasting skill for heavy convective showers and thunderstorms over orography is low. This is partly due to poor understanding of thermally-driven orographic flow of moisture and aerosols which feed the storm. Additional uncertainties arise from parameterised sub-gridscale processes within the model as many of the small-scale features that initiate convection are either completely unresolved (e.g. boundary layer thermals) or highly uncertain (e.g. surface fluxes), even in high resolution models.
The availability of data collected during COPS provides a unique opportunity to study the role of orography in determining the predictability of convection. Accurate prediction of the location of convective initiation requires high-resolution grids which can represent the fine-scale terrain details. Currently I am performing high-resolution (1 km) simulations using the UK Met Office Unified Model over the COPS region to quantify the predictability of severe convective storms encountered during COPS. An ensemble approach is taken using initial and boundary data from the Met Office Global and Regional Ensemble Prediction System (MOGREPS). I have been looking at IOP 8b (15th July 2007), where convection was triggered by a thermally-induced low-level convergence zone and IOP 9c (20th July 2007), where a synoptically-forced squall line passed through the region.