DYMECS: Dynamical and DYMECS: Dynamical and Microphysical Evolution of Microphysical Evolution of

Convective StormsConvective Storms(NERC Standard Grant)(NERC Standard Grant)

University of Reading: University of Reading: Robin Hogan, Bob Plant, Robin Hogan, Bob Plant, Thorwald Stein, Kirsty Hanley, John NicolThorwald Stein, Kirsty Hanley, John Nicol

Met Office: Met Office: Humphrey Lean, Carol HalliwellHumphrey Lean, Carol Halliwell

The DYMECS approach: beyond case studiesThe DYMECS approach: beyond case studies

NIMROD radar network rainfall

Track storms in real time and

automatically scan Chilbolton radar

Derive properties of hundreds of storms on ~40 days:•Vertical velocity•3D structure•Rain & hail•Ice water content•TKE & dissipation rate

Evaluate these properties in model varying:•Resolution•Microphysics scheme•Sub-grid turbulence parametrization

Nimrod radar 1.5-km model

500-m model 200-m model

Kirsty HanleyKirsty Hanley

Nimrod radar 1.5-km model

500-m model 200-m model

Kirsty HanleyKirsty Hanley

Too many

Too few

Storm size Storm size distributiondistribution

• Smagorinsky mixing length plays a key role in determining number of small storms

1.5-km model

500-m model

Kirsty HanleyKirsty Hanley

20 April 201220 April 2012 25 Aug 201225 Aug 2012

200-m model best

500-m model best

200-m model best

1.5-km model

best

Kirsty HanleyKirsty Hanley

Vertical Vertical profileprofile

First 60% of storms by cloud-top height

Next 30%

Top 10%

Thorwald SteinThorwald Stein

Ice density too low?

Higher reflectivit

y core

Observations 1.5-km model 1.5-km + graupel

Vertical Vertical profileprofile

First 60% of storms by cloud-top height

Next 30%

Top 10%

Observations 200-m model 500-m model

Thorwald SteinThorwald Stein

Estimation of vertical velocities Estimation of vertical velocities from continuityfrom continuity

• Vertical cross-sections (RHIs) are typically made at low elevations (e.g. < 10°)

• Radial velocities provide accurate estimate of the horizontal winds

• Assume vertical winds are zero at the surface• Working upwards, changes in horizontal winds at a given level

increment the vertical wind up to that point• Must account for density change with height

John NicolJohn Nicol

• Key uncertainty in models is convective updraft intensity and spatial scale

• Can we estimate updrafts from Doppler wind sufficiently well to characterize the distribution of intensity and spatial scale?

Vertical wind (m/s)

Retrieved vertical wind (m/s)

Retrieval error (m/s)

Reflectivity (dBZ)

Horizontal wind (m/s)

Estimating retrieval errors Estimating retrieval errors from the Unified Modelfrom the Unified Model

John NicolJohn Nicol

dBZ

u (m/s)

w (m/s)

12:45 07 August 2011 16:37 07 August 2011

John NicolJohn Nicol

Scientific and modelling questionsScientific and modelling questions• What is magnitude and scale of convective updrafts? How do two

observational methods compare to model at various resolutions?• What model configurations lead to the best 3D storm structure

and evolution, and why?• How good are predictions of hail occurrence and turbulence?• How is boundary-layer grey zone best treated at high resolution,

and what is the role of the Smagorinsky length scale?• Does BL scheme “diffuse away” gust fronts necessary to capture

triggering of daughter cells and if so how can this be corrected?• Can models distinguish single cells, multi-cell storms & squall

lines, and the location of daughter cells formed by gust fronts?• What are the characteristics common to quasi-stationary storms in

the UK from the large DYMECS database?• Can we diagnose parameters that should be used in convection

schemes from observations?