Atmospheric Boundary Layers:

Transcription

1 Atmospheric Boundary Layers: An introduction and model intercomparisons Bert Holtslag Lecture for Summer school on Land-Atmosphere Interactions, Valsavarenche, Valle d'aosta (Italy), 22 June, 2015 Meteorology and Air Quality Department

2 Atmospheric Boundary layer (ABL) (Stull, 1988) Daytime ABL: m Nighttime ABL: m The lower layer of the Atmosphere influenced by the presence of the earth s surface due to: Friction, Surface Heating (Convection) or Cooling Important characteristics are Turbulence and Diurnal Cycle! 2

3 Why is the boundary layer important? We live in this layer! Weather Forecasting: Surface temperatures, Wind, Fog,... Climate and Earth System Studies: Impact of changing conditions: Land use, Greenhouse gases? Relevant for Agriculture, Energy use, Health, Traffic, Urbanization, Dispersion of pollutants, et cetera 3

4 The interaction of the land-surface with the atmospheric boundary layer includes many processes and complex feedback mechanisms (Figure by Mike Ek, see also Ek and Holtslag, 2004)

5 History of day and night time temperature errors Monthly averages over Europe Standard deviation Mean error New land surface scheme Increased turbulent diffusion + soil moisture freezing Reduced outer layer diffusion in stable situations Courstesy Anton Beljaars, ECMWF ECMWF/GABLS workshop 7-10 November 2011 (5)

6 height [m] Daytime Clear Convective Boundary Layer over Land Why does an inversion form at ABL top? free atmosphere 500 thermal inv ersion conv ectiv e boundary layer Convection over land temperature [K] 6

7 2900 m Intermezzo 7 Potential temperature Temperature that a parcel will have if it is moved (dry) adiabatically to a reference level (P 0 =1000 hpa or z=0) T P 0 P R d / C p Poisson s equation For dry adiabatic process the potential temperature is constant! Actual temperature decrease (increase) for rising (sinking) parcel due to air expansion (compression)

8 Example: Cabauw tower observations 4 November 2006 (diurnal cycle of potential temperatures at 2 to 200 m) 8 Intense daytime turbulent mixing gives same potential temperature, but nighttime surface cooling leads to strong vertical gradients!

9 9 Temperature structure of standard atmosphere with turbulent ABL and capping inversion (Much more background on Atmospheric Boundary Layers is given by Roland Stull in Chapter 9 of the Atmospheric Science book by Wallace and Hobbs, 2006)

10 Convective Boundary Layer (CBL) or Mixing Layer Turbulence is dominated by buoyancy Entrainment Strong Mixing Surface Heat Strong mixing leads to rather uniform profiles, in particular for potential temperature (Figure by John Wyngaard, 1985) 10

11 The typical 3 phases of boundary layer evolution under fair weather over land (Fig 9.24) 11

12 Cumulus clouds How does the land surface impact on the onset of Cumulus? Role of soil moisture and atmospheric processes? weaker inversion stronger inversion dry soil wet soil Courtesy Michael Ek (NCEP) dry soil

13 How to quantify? (Ek and Mahrt, 1994; Ek&Holtslag 2004) Consider Relative Humidity (RH) tendency at boundary layer top Δθ γ θ γ q Δq Use mixed-layer bulk model for the tendency terms θ q

14 Analytical result (Ek & Holtslag 2004, JHM) Evaporative fraction ne: non-evaporative processes, including entrainment and boundary layer growth Confirmed with Cabauw (wet soils) and AMMA observations (dry soils) (Westra et al, 2012, JHM)

15 Boundary layer over heterogeneous terrain 15

16 Stable boundary layer Turbulence by friction, surface cooling stabilizes Large gradients in temperature and wind (Figure by John Wyngaard, 1985) 16

17 At night, besides of turbulence also radiative cooling Vincent van Gogh and waves 17

18 Sketch of ABL processes depending on stability dust devils (After Holtslag and Nieuwstadt 1986 and Nieuwstadt and Hunt 2003) 18

19 Diurnal cycle over land Free Atmosphere Convective Mixed Layer Residual Layer Stable (nocturnal) Boundary Layer Noon Sunset Midnight Sunrise Noon unstable stable 19

20 Mean profiles U Free Atmosphere Convective Mixed Layer U geo Residual Layer U geo U Low Level Jet Stable (nocturnal) Boundary Layer Noon Sunset Midnight Sunrise Noon unstable stable 20

21 Turbulence in boundary layer 1) Wind shear (S) 2) Buoyancy (B) + U + Day: S>0 B>0 - Day Night Night: S>0 B<0 21

22 Observations of temperature fluctuations at several heights over flat terrain 22

23 Spectrum of turbulence kinetic energy and the turbulence cascade 23 Big whirls have little whirls, That feed on their velocity; And little whirls have smaller whirls, and so on to viscosity (L.F. Richardson, 1928) Eddies size of boundary layer depth to smallest scale of about 1 mm! (Kolmogorov scale)

24 24 Turbulent heat flux and high frequency data 0 ' 0, ' 0 ' 0, ' ' ' w w and w thus w w w ' Reynolds decomposition ' ' ' ' ' ' ' ' w w w w w w w w w Flux is a co-variance w w' w

25 The origin of turbulent vertical heat flux 25

26 The origin of turbulent vertical heat flux 26

27 27 Turbulent heat flux ' ' ' ' 0 ' ' ' ' sm J K s m kgk J m kg m W w C H w w w w w w w w p near surface, thus: Turbulent heat flux

28 Break 28

29 Progress in the Atmospheric Sciences: Connecting scales Meteosat observations versus ECMWF predictions (T1279 ~ 15 km) (Courtesy MeteoSat and ECMWF)

30 Atmospheric Budget Equations Sub-grid processes Discretization

31 Budget equation for potential temperature after Reynolds decomposition 31 d dt U t V x W y z w' '... z w' ' K z... K: Eddy diffusivity for heat Similar equations for wind and any conserved quantity C (specific humidity, tracers, ) Can be solved for given initial and surface conditions, and if the diffusivities (K) are known

32 How to deal with turbulent mixing and eddy diffusivity? K f ( height, stability, ABL depth,...) Strategy Distinguish between stable and unstable conditions Use theoretical concepts, observations and numerical simulations (LES and DNS) as a guidance Compare model results with independent data Many turbulent mixing descriptions in the literature! 32

33 Non-local mixing in convective boundary layers Use profile function for eddy-diffusivity and non-local ( counter-gradient ) corrections w K z Deardorff (1966, 1972) Troen and Mahrt (1986) Holtslag and Boville (1993) Large et al (1994) Hong and Pan (1996) Beljaars and Viterbo (1998) Lock et al (2000) et cetera This approach is now widely adopted (e.g., NCAR, NCEP, UKMO, ECMWF,...), but details differ...

34 Flux-Gradient Theory (first order closure) C w' c' K z U 2 K l F z m, h l : length scale g U Ri z z ( Ri) 2 Turbulent Flux Diffusivity K depends on characteristics of turbulent flow Richardson number (measure for local stability) 34

35 Stable boundary layer mixing Enhanced friction needed in models F m LTG Three alternatives for stability functions of heat and momentum MO Revised LTG LTG: Louis et al (1982) MO: Extended Monin- Obukhov type, in agreement with Cabauw tower observations (Beljaars and Holtslag, 1991) F h MO Revised LTG LTG 35

36 Mean model difference in 2 meter temperature for January 1996 using two different stability functions in ECMWF model (Courtesy A. Beljaars) 36

37 Turbulent Kinetic Energy (TKE) closure scheme (applied in many atmospheric models ) TKE t Adv Shear Buoy Transport Dissipation K l TKE TKE represents the major characteristics of turbulence How to do the length scale? Is local gradient assumption valid? Many proposals have been made for various boundary layers 37

38 Does more complexity help? 38

39 Modeling Atmospheric Boundary Layers: It is still a challenge! Atmospheric models do have problems in representing the stable boundary layer and the diurnal cycle Sensitivity to details in mixing formulation Strategy Enhance understanding by benchmark studies over land and ice in comparison with observations and fine scale numerical model results So far focus on clear skies!

40 GEWEX Atmospheric Boundary Layer Studies (GABLS) provides platform for model intercomparison and development to benefit studies of Climate, Weather, Air Quality and Wind Energy GABLS1 GABLS2 GABLS3 LES as reference Data (CASES99) Data (CABAUW) Academic set up Idealized forcings Realistic forcings Prescribed T s Prescribed T s Full coupling (SCM) Prescribed T s (LES) No Radiation No Radiation Radiation included Turbulent mixing Diurnal cycle Low levet jet + transitions LES: Large Eddy Simulation; SCM: Single Column Model

41 GABLS3: SCM and LES model studies GABLS3-SCM 12 Z 01 July Z 02 Jul GABLS3-LES 09 Z 02 Jul 12 Z 02 Jul Initialization Profiles Cabauw tower, Profiler, De Bilt Sounding Geostrophic Wind (time-height dependent) Similar for both SCM and LES Large-scale Advection (time-height dependent) Similar for both SCM and LES Cabauw tower (KNMI, NL) Surface Boundary Conditions Cabauw tower 41

42 GABLS3 intercomparison of Single Column versions (SCM) of operational and research models (Coordinated by Fred Bosveld, KNMI) Note: Each SCM uses its own radiation and land surface scheme interacting with the boundary layer scheme on usual resolution! (Nlev is number of vertical levels in whole atmosphere) Name Institute Nlev BL.Scheme Skin ALADIN Meteo France 41 TKE No AROME Meteo France 41 TKE No GLBL38 Met Office 38 K (long tail) Yes UK4L70 Met Office 70 K (short tail) Yes D91 WUR 91 K Yes GEM Env. Cananda 89 TKE-l No ACM2 NOAA 155 K+non-local No WRF YSU NOAA 61 K No WRF MYJ NOAA 61 TKE-l No WRFTEMF NOAA 61 Total E No COSMO DWD 41 GFS NCEP 57 K Yes WRF MYJ NCEP 57 TKE-l Yes WRF YSU NCEP 57 K Yes MIUU MISU 65 2nd order MUSC KNMI 41 TKE-l No RACMO KNMI 80 TKE Yes C31R1 ECMWF 80 K Yes CLUBB UWM 250 Higher order No

43 Surface fluxes noon midnight noon

44 Surface fluxes noon midnight noon

45 Surface fluxes noon midnight noon

46 Surface fluxes noon midnight noon

47 noon midnight noon

48 Surface parameters noon midnight noon

49 Low level jet noon midnight noon

50 Novelty: Process diagrams, here for the influence of mixing Sensitivity runs with RACMO-SCM using various parameters and forcings

51 GABLS3 Large Eddy Simulation intercomparison (coordinated by Sukanta Basu, NC State Univ) Initialized at midnight (02-jul :00 UTC) and run for 9 hours (11 LES models) Prescribed temperature at lowest model level from observations! Wind speed magnitude Potential Temperature Mean profiles UTC (Red dots: Tower; Blue squares: Wind Profiler) Green dashed line: 1m LES run (Courtesy Siegfried Raasch) ISARS-2010, Paris, June 2010

52 GABLS3 Large Eddy Simulation intercomparison (coordinated by Sukanta Basu, NC State Univ) Sensible Heat flux (W/m2) Momentum flux (N/m2) Flux profiles UTC (Red dots: Cabauw Tower observations) Green dashed line: 1m LES run (Courtesy Siegfried Raasch) ISARS-2010, Paris, June 2010

53 Temporal evolution 53 53

54 Holtslag et al, 2013, BAMS

55 Holtslag et al, 2013, BAMS

56 Diurnal cycles of temperature and wind A challenge for weather and climate models! Significant variation are seen in models which can be related to relevant atmospheric and land surface processes Sensitivity to details in mixing formulation, interaction with the land surface, the representation of radiation (divergence), et cetera Overview of results and citations in Holtslag et al, 2013, Bulletin American Meteorological Society (BAMS, online) 56

57 At present GABLS4: Antarctic summer case (Dome-C) (Eric Bazile, Fleur Couvreux et al) Helsinki meeting 14/15 Feb 2013 GABLS4/GABLS4.html

58 GABLS basic publications (plus many conference and invited presentations) GABLS1: Special issue Feb 2006, Boundary Layer Meteorology (7 papers) Svensson and Holtslag, 2009, BLM (wind turning issue) GABLS2: Steeneveld et al, 2006, JAS (SCM) and 2008, JAMC (Mesoscale study) Holtslag et al, 2007, BLM (Coupling to land surface) Kumar et al, 2010, JAMC (LES study) Svensson et al, 2011, BLM (SCM intercomparison) GABLS3: Baas et al 2010, QJRMS (set up case and SCM tests) Special issue of BLM 2014, including intercomparison papers by Bosveld et al (2014, SCM), Edwards et al (2014, LES + Radiation)... GABLS overview paper in 2013 (Holtslag et al, BAMS, online) 58