Observed Updraft & Mass Flux in Shallow Cumulus

at ARM Southern Great Plains sitePreliminary results

Yunyan Zhang, Steve Klein & Pavlos Kollias

CFMIP/GCSS Boundary Layer WGJune 11, 2009, Vancouver, Canada

MotivationThere are decade-long comprehensive observations at ARMSouthern Great Plains site To document diurnal cycle of different convection regimes

e.g. fair-weather non-precipitating shallow cumulus e.g. late-afternoon precipitating deep convection

To assess convection theories To construct composite case for LES or CRM studies To feedback on improvement of GCM parameterization

Outline Diurnal cycle study (2 slides) The vertical velocity observational study (11 slides)

Diurnal Cycles in Warm SeasonsABRFC Precipitation CMBE ARSCL Cloud Fraction

Clear-sky

Fair-weatherShallow Cumulus

Late afternoonDeep Convection

On the preconditioning of free troposphere humidity for deep convection

The impact of boundary layer inhomogeneity on deep convection

Use ARM data to assess convection theories

Data from LSSONDE Data from SMOS and OK Mesonet

Vertical Velocity Observation

• Vertical pointing Doppler Radar

• Measure the movement ofscattering targets. In non-precipitating shallow cumulus, the target is the liquid water cloud droplet

• Usually the terminal velocity of liquid cloud droplet is about ~cm/s, this is much smaller compared to air motion velocity ~ m/s Thus the vertical velocity of cloud droplet is representative of air motion

ARM SGP Millimeter Wavelength Cloud Radar (MMCR)

Pavlos K

ollias

Vertical Velocity Observation

Pavlos K

ollias

vertical: 45 mhorizontal: 10 mfrequency: 10 s

• There is no retrieval in clear air nor the precipitating part of the cloud; it is particularly good for non-precipitating liquid-phase shallow cumulus

• Pavlos has retrieved data and made hourly averages, from 1999 to present

Data detail:• Hydrometeor fraction / low-dBZ

fraction• In-cloud Updraft / downdraft• In-cloud updraft / downdraft fraction

Updraft fraction = cloud fraction * in-cloud updraft fractionUpdraft mass flux = updraft fraction * updraft velocity

• Make diurnal composites for different convection regimes– Fair-weather non-precipitating shallow cumulus – Shallow cumulus before late-afternoon deep convection

• How to average clouds with different cloud base heights and preserve the intrinsic shape of the vertical profiles?

Methodology

Average cloud base height

e.g. Mass Flux

• What do we know from LES? (BOMEX, an ocean case)

– Cloud fraction, updraft fraction updraft mass flux peak above cloud base and then decrease with height

– Updraft velocity increases with height – Updraft fraction dominates in cloud fraction– Updraft mass flux dominates in net mass flux

• Brown et al, 2002 (SGP, 06/21/1997, a land case) – Similar profile shapes are shown for updraft fraction and updraft mass flux– This one day simulation will serve as a qualitative comparison in the following

Non-Precipitating Shallow Cumulus

Figures from Siebesma and Cuijpers 1995; Siebesma et al, 2003

Updraft mass flux = updraft fraction * updraft velocity

Updraft Mass Flux

• vertical shape• diurnal variation• comparable to LES study of Brown et al, 2002

Updraft fraction = cloud fraction * in-cloud updraft fraction

Updraft Fraction

In-Cloud Updraft

Updraft velocity decreases with height. Such behavior is different from that of updraft in shallow cumulus over ocean.

Remarks on updraft velocityBuoyancy diurnal cycle based on composite sounding for non-precipitating shallow cu

Composite average CIN = 67 J/kg

Downdraft & Net Mass Flux

The symmetry between updraft and downdraft might be explained by an overshooting of non-entraining plume. But this is not consistent with highly entraining plume for shallow convection.

Remarks on symmetry

• the radar might only see the undilute convective core, but not cloud edges

• the contribution to mass flux from eddies, whose scale is smaller than the radar observational resolution, is not included in the data

• to resolve these issues, we will need higher resolution radar data (e.g. WCAR data) and radar simulator studies combined with large eddy simulations

Remarks on Shallow Cumulus• Observed composite updraft fraction and updraft mass flux

are comparable to LES results in both magnitudes and the shapes of vertical profile.

• Observed composite updraft has magnitude below 1m/s and decreases with height. The composite updraft mass flux does not dominate in the composite net mass flux; this is because observed downdraft and updraft have very similar statistics. – These observations are different from we knew according to BOMEX,

an ocean case. Some tentative explanation provided.– Will LES help if we try similar sampling and compositing as

observations?

Fair-weather vs. before Deep Convection