Boundary Layer and Large-Scale Circulation
Gunilla Svensson
1985-95: Bachelor and PhD education at Uppsala University, Sweden
Thesis: Numerical modeling of the air quality in Athens, Greece
1996-97: Postdoc at Caltech, Pasadena, CA, USA
Project: Numerical modeling of marine stratocumulus clouds
1997-98: Junior Researcher, Uppsala University, Sweden
1998 – 2003: Junior Researcher, Stockholm University, Sweden
2005: Senior Lecturer, Stockholm University, Sweden
2008: Promoted to Professor of meteorology, Stockholm University, Sweden
2021: Guest professor (50%) of fluid mechanics with specialization in climate modeling at the KTH Royal Technical University, Stockholm
Visiting scientist in US in total more than 5 years: DRI, Reno, NV; Caltech, Pasadena, CA; NRL Monterey, CA; University of Colorado at Boulder, CO; NCAR, Boulder, CO
My academic background
2011 – 2022 Member of the Science Steering Group for the WMO/WWRP Polar Prediction Project, leader of the Process Task Team
2024 – Proposed member of the Science Steering Group for the WMO/WWRP Polar Polar Coupled Analysis and Prediction Services (PCAPS)
2016 – Affiliated scientist at NCAR
2017 – Member of the Science Advisory Council for the ECMWF - European Centre for Medium-range Weather Forecasts
2017 – 2022 Member of the Research Council's Council for Research Infrastructures (RFI), advisory group B: Observatories and measuring platforms for astronomy, the climate, the environment and earth sciences and Chair of the e-infrastructure committee
2023 - Member of the Board for RFI
Some things I find extra rewarding
– besides working with my students and doing research…
Choose challenging projects that you are interested in and try to have fun!
Take opportunities when they arise!
Be a bit selective – try to spend time on things that will lead you forward, but not too picky as it is hard to know exactly what opportunities engagement will lead to
Taking the opportunity to participate in a NCAR Colloquium is a good one – I did in 1991!
My former graduate students work at universities or as researchers/developers within defense, environment agency, wind energy industry, insurance, safety and security
My best advise…
Gunilla Svensson, MISU
5
Spatial and temporal scales in the atmosphere
Decomposition
Build on the concept of separated scales, split all variables into mean and fluctuating parts. For example, horizontal wind speed:
U=U + u’
Decomposition
Waves
Reynolds averaging
Averaging time 15-30 min
Space and time conversion based on Taylor’s hypothesis on frozen turbulence
Equations for atmospheric turbulent flow
Wind
Temperature
Continuity
Rotation and buoyancy
Equations for atmospheric turbulent flow
Five equations but 5+4 dependent variables
⇒ the ”Closure”-problem!
Effect of turbulence on the mean flow
Wind
Temperature
Continuity
Gunilla Svensson, MISU
11
Turbulent Kinetic Energy – TKE (e)
I: Time rate of change of e (or the TKE)
II: Shear production (always positive) of e, ”mechanical production”.
III: Buoyancy production (can be both positive and negative) of e, ”thermal production”
IV: Turbulent transport of turbulent energy
V: Pressure transport of e
VI: Dissipation of e
I
II
III
IV
V
VI
V. Walfrid Ekman (1874 – 1954)
Assume stationary and horizontally homogeneous processes in the Reynolds equations, height constant pressure gradient represented by Ug and Vg:
Ekman Layer
Further assume that:
The equations become:
Define a complex wind W=(u + iv). Assume that KM is constant. Multiply the second equation with i = (-1)1/2 and add the equations:
Rearrange:
with the solution:
and after applying boundary
conditions:
Ekman Layer
Ekman Layer – Ekman spiral
u (ms-1)
LES
Operational
Research-Meso
Research
α
GABLS 1 – GEWEX Atmospheric Boundary Layer Study first experiment
Idealized case with constant pressure gradient and surface cooling
Operational
Research
LES
GABLS 1 – GEWEX Atmospheric Boundary Layer Study first experiment
Cuxart et al., 2006
Svensson and Holtslag, 2009
Ekman layer equations
Provides a relation between the surface angle and the ageostrophic flow
(Svensson and Holtslag 2009)
Assume steady-state and vg=0, then the cross-isobaric flow is given by:
Integrating over the atmospheric column,
note that =0 above the boundary layer:
α
The ageostrophic flow
Svensson and Holtslag, 2009
Operational
LES
L
Secondary circulation – spin down
Deeper PBL gives larger friction velocity and larger drag, the angle is important for the integrated mass flux across isobars
The ageostrophic flow in ERA5 visualized by trajectories
Trajectories (+4 days) calculated using Lagranto initiated from 70°N at different Δp above surface pressure
Use of balance
Beare and Cullen, 2013
Balance model
Beare and Cullen, 2013
Low-level frontal jet
Beare and Cullen, 2013
Baroclinic wave
Beare and Cullen, 2013
Baroclinic wave
Beare and Cullen, 2013
Large-scale circulation and surface friction
In idealized AGCMs, surface jet strength and latitude are highly sensitive to surface drag, via feedback on baroclinic eddies
Decreasing drag
Chen et al., 2007
u (ms-1)
LES
Operational
Research-Meso
Research
α
GABLS 1 – GEWEX Atmospheric Boundary Layer Study first experiment
Stable boundary layer mixing
NWP models need a long tail formulation i.e. more mixing to get the synoptic scale right
(Louis et al. 1982)
Observations follow the M-O type of functions (Beljaars and Holtslag, 1991)
By changing this functions you can easily change the modeled temperature significantly
Stability functions for momentum
MO
Enhanced friction
needed in models
and heat
Increasing damping by stratification
ECMWF IFS Courtesy A. Beljaars
Effect of MO-stability functions (reduced diffusion) instead of operational formulation, on 500hPa NH height scores
Global numerical weather prediction model�Stability functions affect the large scale forecast scores
0.5 day
Review on stable PBL and waves, Sun et al., 2015
Stable boundary layer mixing
Stable boundary layer mixing
weak wind conditions
Review on stable PBL and waves, Sun et al., 2015
Can we use observations to better constrain models?
Surface fluxes, momentum�Satellite-based observations�
Climatological wind turning
Bias in ERA INTERIM
Lindvall and Svensson, 2019
Climatological wind turning
Lindvall and Svensson, 2019
Pyykkö and Svensson, 2023
CMIP6 models
Wind turning in CMIP6 models over the PBL
Pyykkö and Svensson, 2023
Cross-isobaric mass flux over the PBL
Pyykkö and Svensson, 2023
(Bougeault 1990)
Subgrid-scale orographic drag is parameterized as an additional surface stress (TMS) – enhanced surface roughness
The size of the stress is dependent on the stability and the variance of orography in a gridbox
Some tests with NCAR CAM
Subgrid-scale orographic drag
Changing turbulent diffusion in stable conditions
Default CAM5 (CONTROL)
No turbulence when Ri > 0.19
CAM5 with enhanced diffusion in stable conditions (Longtail)
More turbulence in stable conditions
(°C)
Annual mean 2-m temperature bias
Lindvall et al. 2017
Mean sea-level pressure
Spring (MAM)
Lindvall et al. 2017
CTRL
LONGTAIL
ERA
DJF
MAM
JJA
SON
NoTMS
-15 -10 -5 0 5 10 15
[106 m2 s-1]
Zonal anomaly of the 500hPa streamfunction
PBL LTAIL
Lindvall et al. 2017
Atmospheric blocking frequency
All model
versions have too few blockings, specially for the Euro-Atlantic sector
Euro-Atlantic sector
Pacific sector
Lindvall et al. 2017
Control is closer to observations
No version captures the Atlantic blockings in winter
Lindvall et al. 2017
Atmospheric blocking frequency
More information