F Couvreux ihop juin04

Water vapour variability within the growing convective boundary layer of 14 June 2002 with large eddy simulations and observations: 

Water vapour variability within the growing convective boundary layer of 14 June 2002 with large eddy simulations and observations OUTLINE Methodology and Objectives Observational data: several scales of variability LES simulations: comparisons obs/model Conclusions & Perspectives Fleur Couvreux Françoise Guichard, Jean-Luc Redelsperger, Cyrille Flamant, Jean-Philippe Lafore, Valéry Masson rv (g/kg) W (m/s) ENE WSW 6 4 2 0 -2 -4 11 10 8 9 7 12 Time (UTC)

Slide2: 

Methodology & Objectives LES : Is such a high resolution model able to represent the observed water vapour variability ?

Slide3: 

Time 12h 13h 14h 15h 16h 17h 18h 19h  (K) 306 294 1 0 Zi (km) Time 12h 13h 14h 15h 16h 17h 18h 19h 2 morning Early afternoon Boundary layer mean value of  & rv Time in UTC=local time+5h 2 1 0 300 rv (g/kg) 7 12 10 A classic convective BL growth … but with large fluctuations of rv Zi (km)

Slide4: 

Main characteristics : « Relatively » simple case of a growing boundary layer: a well mixed boundary layer reaching 1.5 km High Pressure system, homogeneous temperature field Weak subsidence constant whole day Low shear and weak wind (< 5m/s) from N to NE Existence of thermals (cf Cloud radar) Small cumuli developed after 1500 UTC NE/SW moisture gradient Heavy precipitation the day before, coherent distribution with moisture gradient The 14 june 2002 case

Slide5: 

Different scales of variability: evidence in soundings 2 soundings separated from less than 10 km 1g.kg-1 1830 UTC rv 

Slide6: 

Different scales of variability : aircraft data and lidars + rv 3g.kg-1 rv’ 1.5 g.kg-1 Aircraft rv measurement 1710 Time (UTC) 1710 12 8 12 7 12 8 +0.7 -0.7 Time (UTC) Time (UTC) 1745 rv 1710 1745

Slide7: 

Modelling: LES with Méso-NH (Lafore et al. 1998) Simulation : x=y=100m, z streched (< 50m in BL) 3D turbulence scheme (Cuxart et al. 2000) early morning to early afternoon (duration 7h) a ‘realistic’ simulation: ISFF2 surface fluxes (prescribed) homogeneous initial sounding = composite of soundings at 1130 UTC large scale advection estimated from MM5 simulations and soundings

Slide8: 

Mean profiles Temporal evolution of mean profiles  rv

Slide9: 

Time variations of boundary layer characteristics Sensitivity to surface fluxes : +Bo -> +zi Sensitivity to large-scale advection : +ADV -> +zi +Bo -> +m +Bo -> + qm Cf Bo ie SSH et SLH +ADV -> + m +ADV -> + qm Several factors : Ws -> zi ->, q Adv  ->  -> zi Adv q -> q … Validated reference simulation, quantification of sensibility to different forcings rv  zi

Slide10: 

Horizontal cross sections Z/zi=0.3 Z/zi=0.3 Z/zi=0.8 Z/zi=1. Z/zi=0.8 Z/zi=1. 305 304.5 304. 303.5 W (m/s) v (K) 305 304.5 304. 303.5 306 304.5 304. 303.5 305.5 305 rv (g/kg) 10 km

Slide11: 

Characteristic length scale Los C()= Reference simulation at 17h rv length scale is larger than length scale of v, , w from Lohou et al. (2002) (m) z/zi Rv  v w

Slide12: 

LEANDRE and LES horizontal cross-sections Measurements from LEANDRE LES Simulations At 1600 UTC ~10 km ~10 km 1.2 3.5 1.2 3.5 1.2 3.5 1.2 3.5

Slide13: 

Vertical cross sections Evidence of dry downdrafts Several thermals in one humid zone LES rv & w DLR-DIAL rv

Slide14: 

Evaluation of histograms of , rv, w P3 aircraft KA aircraft Z=0.4zi model … max --- min w’ ’ rv’ Equivalent gaussian

Slide15: 

Boundary layer precipitable water

Slide16: 

Second order moments __ simulation __ lidar DIAL o P3 o KA Strong relation between The inversion strength and the variance Max(rv) (g.kg-1) q à zi (g.kg-1) 1.2 0.72 0.81 0.41 0.56 5.7 MNH 4.4 MNH 4.3 DIAL 3.9 DIAL 4.1 DIAL 3 lidar 10km-long Cross-section at 1730 UTC LES 1700 & 1800 UTC rv

Slide17: 

Joint probability distribution At z/zi=0.3 + z/zi + larger spectrum of w+ and q- At z/zi=0.8

Conclusions:: 

Conclusions: Evaluation of the LES Able to represent the variability observed at scales lower than 10 km (comparisons to soundings, lidars (DIAL et SRL), aircraft time-series) Quantification the impact of scales > 10 km on variability at scales < 10km At first order, the boundary layer dynamics explain the observed variability at scales lower than 10 km even without surface heterogeneities and variability in the initial atmospheric state Dry narrow downdrafts as a signature of the BL growth (via dynamics at the top) [Crum et al. (1987) and Weckwerth et al. (1996)] impact on length scale, skewness, vertical transport.. Negative skewness is common (cf other IHOP days) Observations from 14 juin 2002 during IHOP_2002 : Several scales of variability ( < 10 km et > 10 km)

Slide19: 

Perspectives: Systematic analysis of IHOP data : - objective : to identify pertinent parameters controlling the water vapour variability in the boundary layer (such as strength of inversion (,q), fluxes…) from more idealised simulations -> 1D parameterizations Understand the impact of such a variability on cloud formation and maintenance Quantify time scales concerned by mechanisms involved in the water vapour variability: dry intrusion life time, transient state

Slide20: 

F I N

Development of the CBL (courtesy of Bart Geerts): 

Development of the CBL (courtesy of Bart Geerts) aspect ratio: 1:1 1330 UTC 1415 UTC 1530 UTC 1630 UTC 1730 UTC

Slide22: 

Surface fluxes Sensible heat flux Latent heat flux Bo~1. Bo~0.5

Slide23: 

Large scale forcings (advection) Large-scale forcings Deduced from MM5 Horizontal advection of  Horizontal advection of rv Subsiding w