Jean-Marcel PiriouCentre National de Recherches MétéorologiquesGroupe de Modélisation pour l’Assimilation et la Prévision

Evolution of convective cloud top height:entrainment and humidifying processesEvolutionEvolution of convectiveof convective cloudcloud toptop heightheight::entrainment and humidifying processesentrainment and humidifying processes

EUROCS Workshop, Madrid, 16-19/12/2002

SummarySummary

Questions:

1. (1D EUROCS idealized humidity case, 3D) What impact on diurnal cycle if the ARPEGE convective scheme sensitivity to humidity is removed?

2. (1D EUROCS diurnal cycle of deep convection case, 3D) Cloud topheight evolutions: link with saturation deficit & its diurnal phase

EUROCS idealized humidity case

ARPEGE operational version

MassMass fluxesfluxes forforthethe 44environmental environmental humiditieshumidities

Precipitations vs thePrecipitations vs the44 humiditieshumidities..ContinuousContinuous: CRMs: CRMsDottedDotted: SCM: SCM

ARPEGE NWP ModelARPEGE NWP ModelJ.M.J.M. PiriouPiriou 20022002

IFS NWP ModelIFS NWP ModelBeljaarsBeljaars 20022002

ObservationsObservationsYangYang and Slingoand Slingo MWR 2001MWR 2001

Unified ClimateUnified Climate ModelModelYangYang and Slingoand Slingo MWR 2001MWR 2001

Diurnal cycleof convection / JJA

EUROCS idealized humidity case

ARPEGE operational version Operational + reduced entrainment

ObservationsObservationsYangYang and Slingoand Slingo MWR 2001MWR 2001

Diurnal cycleof convection

CONTROL = ARPEGE NWPCONTROL = ARPEGE NWPoperationaloperational model, 10model, 10 two two days preddays pred..décdéc 2002

CONTROLCONTROL++ Reduced entrainment

2002

Reduced entrainment

EUROCS diurnal cycle ofEUROCS diurnal cycle of deepdeep convectionconvection

SCMSCM runrun;; operationaloperational ARPEGEARPEGE physicsphysics

ConvectiveConvective precipprecip.. Cloud topCloud top heightheight

Sensitivity to humidity: top of clouds

TOGARed. 2002

EUROCSId. qv25%

TOGATOGA--COARE / J.L. Redelsperger, D. Parsons, F.COARE / J.L. Redelsperger, D. Parsons, F. GuichardGuichard, JAS 2002, JAS 2002

CRM dataCRM data fromfrom FrançoiseFrançoiseGuichard andGuichard and JeanJean--PierrePierreChaboureauChaboureau, CNRS / , CNRS / MétéoMétéo--FranceFrance

CRM MESOCRM MESO--NHNH

SCM ARPEGE convectionSCM ARPEGE convection switchedswitched offoff

SCM ARPEGE convectionSCM ARPEGE convection switchedswitched onon

EUROCS diurnal cycle ofEUROCS diurnal cycle of deepdeep convectionconvection

Sensitivity to humidity: top of clouds

TOGARed. 2002

EUROCSId. qv25%

EUROCSDiurn. CycleDeep conv.

TOGATOGA--COARE / J.L. Redelsperger, D. Parsons, F.COARE / J.L. Redelsperger, D. Parsons, F. GuichardGuichard, JAS 2002, JAS 2002

Pentasat IR image

SAT. DEF. between 2.5 and 4.5km, in g/kg. ARPEGE +6h analysis first guess

Sensitivity to humidity: top of clouds

TOGARed. 2002

EUROCSId. qv25%

EUROCSDiurn. CycleDeep conv.

TOGATOGA--COARE / J.L. Redelsperger, D. Parsons, F.COARE / J.L. Redelsperger, D. Parsons, F. GuichardGuichard, JAS 2002, JAS 2002

Pentasat IR image

Composite cloudiness image. ARPEGE +6h analysis first guess

Pentasat IR image

Composite cloudiness image. ARPEGE +6h modified analysis first guess

EUROCS diurnal cycle ofEUROCS diurnal cycle of deepdeep convectionconvection

ConvectiveConvective precipprecip.. Cloud topCloud top heightheight

SCMSCM runrun;; modifiedmodified ARPEGEARPEGE physicsphysics: ad hoc TOC : ad hoc TOC

as aas a functionfunction of saturationof saturation deficitdeficit

Conclusions / perspectives Conclusions / perspectives (1/2)(1/2)

• ARPEGE convection scheme sensitivity to humidity seems OKonly for « wet » atmospheres, i.e. sat. defs (2.5 > 4.5km) < 2.5 g/kg, but likely underestimated for dry atmospheres

• Reducing entrainment rates to suppress the sensitivity in the idealized case � small impact on diurnal cycle. The 3h phase lagof precipitation of ARPEGE vs other models remains yet unexplained

Conclusions / perspectives Conclusions / perspectives (2/2)(2/2)

The sensitivity of cloud top height to environmental humidity seems to befor saturation deficits (2.5 > 4.5km) between 2 and 3 g/kg:

1. TOGA-COARE (Redelsperger, Parsons and Guichard, JAS 2002)2. CRM MESO-NH in EUROCS diurnal cycle of deep convection case3. 5sat image versus ARPEGE global analysis of saturation deficits

> Humidifying a dry atmosphere: a good candidate to explain large phase lags of precipitation vs surface forcings, from diurnal to 10days time scales?

Perspectives: (i) retune the entrainment in drier contexts than previously done, to get the right cloud tops / (ii) increase humidification ratebelow cloud top / (iii) see impact on diurnal cycle!

EUROCS idealized humidity case

Operational version Operational + « no relax towards und. »

Entrainment: undilute plume relaxation

cundccc und �������

�

�

�����

�

��

�

�

qvT��

Towards environment

Towards undilute: easy way to take into account that a ratherconstant fraction of ascents remain undiluted, without havingto use negative entrainments in the expression just on the left

TimeTime--heightheight crosscross--sections ofsections of the the difference between current difference between current humidityhumidity (g/kg)(g/kg) andand initial oneinitial one

SCMSCM operationaloperational ARPEGE ARPEGE

CRM MESOCRM MESO--NH (NH (referencereference))predictionprediction; data; data fromfrom FrançoiseFrançoiseGuichard andGuichard and JeanJean--PierrePierreChaboureauChaboureau, CNRS / Météo, CNRS / Météo--FranceFrance

TimeTime--heightheight crosscross--sections ofsections of the the difference between current difference between current humidityhumidity (g/kg)(g/kg) andand initial oneinitial one

CRM MESOCRM MESO--NH (NH (referencereference))predictionprediction; data; data fromfrom FrançoiseFrançoiseGuichard andGuichard and JeanJean--PierrePierreChaboureauChaboureau, CNRS / Météo, CNRS / Météo--FranceFrance

SCMSCM modifiedmodified ARPEGE: ad hoc TOC ARPEGE: ad hoc TOC as aas a functionfunction of saturationof saturation deficitdeficit

Trends, or under progress

Present operational schemes

1st order closure scheme afterLouis (1979), Louis and al. (1981), using a flux-gradient K-theory with Ri dependency, variable roughness lengths over sea (Charnock)

mass-flux scheme, CISK-typeclosure and triggering, watervapour budget using a Kuo-type closure, downdrafts,momentum flux

Diagnostic in ql/i, allsupersaturation removed,liquid/ice condensation � T,melting/ freezing/ evaporation/Kessler (1979), Clough and Franks (1991)

Geleyn and Hollingsworth(1979), Ritter and Geleyn(1992)

PrognosticTKE scheme,mixing « Betts »conservative variablesthetal and qt instead oftheta and qv

Turbulence

Entrainment & dry intrusions, modified triggerfunctions (use of TKE, CIN) / Revised shallowand deep conv.

Subgrid-scale cloud scheme(convection)

Prognostic ql/i and/or qr/sGrid-scale cloud scheme

More accurate infra-red exchanges betweensurface and layers

Radiation

ARPEGE NWP modelARPEGE NWP modelJJA 2002JJA 2002

ObservationsObservationsYangYang and Slingoand Slingo MWR 2001MWR 2001

Diurnal cycleof convection

ARPEGE NWP modelARPEGE NWP model1010 daysdays CAPECAPE decdec 20022002

Operational version Operational + reduced entrainment