Field measurements of urban roughness-sublayer turbulence Andreas - - PowerPoint PPT Presentation

NCAS Urban Meteorology Workshop

• A. Christen / 30-31 March 2009

Field measurements of urban roughness-sublayer turbulence

Andreas Christen Department of Geography, Atmospheric Science Program University of British Columbia

NCAS Urban Meteorology Workshop 30-31 March 2009, University of Reading, UK

NCAS Urban Meteorology Workshop

• A. Christen / 30-31 March 2009

Outline

• Turbulence measurements in the urban roughness

sublayer (URSL) - historical developments, approaches, technology.

• Vertical profiles of mean flow and integral turbulence

statistics - towards a conceptual division of the URSL.

• The nature of exchange processes in the URSL - scales,

coherent structures, and the dissimilarity of exchange.

NCAS Urban Meteorology Workshop

• A. Christen / 30-31 March 2009

Turbulence measurements in the URSL Historical developments, approaches, technology

NCAS Urban Meteorology Workshop

• A. Christen / 30-31 March 2009

The challenge of measuring turbulence in cities

• No control on flow and geometry.
• The large variety of three-dimensional realizations of the

surface-atmosphere interface (morphometry, source distribution, human activity cycles, irradiance and shadowing cycles).

• The resulting micro-scale inhomogeneity of the flow field

around groups of obstacles including flow caused by moving obstacles. Can we sample the flow representatively?

• Technical restrictions and safety concerns of operating

towers and instrumentation in densely populated areas.

NCAS Urban Meteorology Workshop

• A. Christen / 30-31 March 2009

Single point or array measurements e.g. Ultrasonic anemometers Line or volume averages e.g. Scintillometry, mini-LIDAR Lagrangian sensors e.g. micro- tetroons Dispersion experiments e.g. tracer gas releases

Experimental approaches in URSL field studies

NCAS Urban Meteorology Workshop

• A. Christen / 30-31 March 2009

Photo: D. Steyn Vancouver-Sunset, 1978

NCAS Urban Meteorology Workshop

• A. Christen / 30-31 March 2009

Photo: M. Roth Vancouver-Sunset, 1989

NCAS Urban Meteorology Workshop

• A. Christen / 30-31 March 2009

Photo: M. W. Rotach Zürich, 1986

NCAS Urban Meteorology Workshop

• A. Christen / 30-31 March 2009

Basel, 1993-2002 Basta / BUBBLE - Photo: A. Christen (2001)

NCAS Urban Meteorology Workshop

• A. Christen / 30-31 March 2009

Photo from: Eliasson et al. 2006 Goteborg, Sweden, 2003

NCAS Urban Meteorology Workshop

• A. Christen / 30-31 March 2009

Photo from: Klein and Clark, 2007 Oklahoma City, USA, 2003

NCAS Urban Meteorology Workshop

• A. Christen / 30-31 March 2009

The horizontal averaging approach

Spatial averaging horizontal layers in URSL

z = 1 z = 2 z = 3 z = 4

From an experimental view, the horizontal averaging concept forms a significant challenge, and information

• n possibly resulting

dispersive terms is even more difficult to obtain in the field.

a(x, t) = a + a′′(x) + a′(x, t)

Raupach and Shaw (1982)

NCAS Urban Meteorology Workshop

• A. Christen / 30-31 March 2009

Most studies use ultrasonic anemometers either at a single location

• r arrayed vertically (barely horizontally) - in street canyons and

above roof level - mounted on towers, masts or tripods. Problem: Single point measurements do not provide horizontally averaged statistics, nor can they be used directly to infer the 3D flow field or quantify dispersive fluxes. Even for a single point, it requires long measurement periods to retrieve statistically representative measurements.

Restrictions of field studies

NCAS Urban Meteorology Workshop

• A. Christen / 30-31 March 2009

Rooftop C a n y

• n

Tower

NCAS Urban Meteorology Workshop

• A. Christen / 30-31 March 2009

Scintillometry and remote sensing?

Tokyo 2000 (Photo: M. Roth)

Three experiments used scintillometers in the urban surface layer (one experiment within the URSL) and tried to overcome the problem of spatial averaging. Problem: the underlying theory (MOST) is not fulfilled in the URSL (see discussions in Roth et al., 2006), dispersive fluxes not measured. Remote sensing (mini-LIDAR) might be promising, but no technology / studies yet at the scale needed.

Scintillometer

NCAS Urban Meteorology Workshop

• A. Christen / 30-31 March 2009

Several typical trajectories of small balloons released in a Chicago core canyon. So far, only applied to the study of mean flow.

Lagrangian approaches

De Paul, F.T., Shieh C.M., 1986, Atmos. Environ., 20, 455-459.

NCAS Urban Meteorology Workshop

• A. Christen / 30-31 March 2009

Code City (Project) UCZ1 Comments Selected References

Zu86 Zürich, Switzerland 2 Tower with 1D / 3D sonic at different locations above and upper canyon. Rotach (1991a,b, 1995) Va89 Vancouver, Canada 5 Tower with 1D and 2D sonic above suburban surface. Roth (1991, 1993), Roth and Oke (1993) Sa92 Sapporo, Japan 3 Three 3D Sonics on masts and 45m crane with residential surface upwind. Oikawa and Meng (1995) Hv94 Hannover, Germany Single 3D Sonic and air pollutant concentrations in 4-lane street canyon Schatzmann et al. (1999) Kastner-Klein et al. (2003) Ba95 Basel, Switzerland (Basta) 2 Tower with 3D sonics above roof level (1.5 < z/zh < 3.2) Feigenwinter et al. (1999, 2005) Rf97 A farm, UK

• Urban-like canyon between farm buildings,

3D sonic at different heights Louka et al. (2000) Na99 Nantes, France (URBCAP) 2 Instrumented street canyon with 4 3D sonics, focus: air pollution and traffic TKE. Vachon et al. (1999), Louka et al. (2002) Sl00 Salt Lake City, USA (URBAN 2000) 1 Urban tracer experiment, with near-field releases and 9 2D and 2 3D sonics. Allwine et al. (2002)

1Urban Climate Zones (Oke, 2004, WMO)

URSL turbulence experiments 1985 - 2000

NCAS Urban Meteorology Workshop

• A. Christen / 30-31 March 2009

Code City (Project) UCZ1 Comments Selected References

Ms01 Marseille, France (ESCOMPTE) 2 Tower with two 3D sonics at different positions. CO2 turbulence, scintillometry. Grimmond et al (2004), Salmond et al. (2006) Tk01 Tokyo, Japan 3 Long-term tower with one to four 3D sonic above roof level (2 < z/zh < 4) Kanda et al. (2006), Moriwaki and Kanda (2005, 2006) Mc01 Manchester, UK (SCAR) 2 3D sonic on telescopic mast with varying heights in asymmetric canyon. Longley et al. (2004) Ba02 Basel, Switzerland (BUBBLE) 2 Three towers with 3-6 3D sonics each (1.5 < z/zh < 3.2). Tracer release above roof. Scintillometry. Christen et al. (2007, 2009), Roth et al. (2006) Go03 Goteborg, Sweden 2 Horizontal and vertical array of 14 3D sonics in a street canyon and above roof. Eliasson et al. (2006) Oc03 Oklahoma City, USA (JU2003) 1 Instrumented canyon with >40 sonics in total, at canyon floor or roof level. Tracer releases. Nelson et al. (2007a,b), Hanna et al. (2007) Lo04 London, UK (DAPPLE 2004) 2 Instrumented intersection with 7 3D sonics, at street or roof level. Tracer release experiments. Arnold et al. (2004), Dobre et al. (2005) Ny05 New York, USA (MSG05 / MID 05) 1 Mainly tracer experiments. Instrumented area with 16 3D sonic sites, most at 3m or roof level. Hanna et al. (2007), Allwine and Flaherty (2007) Lo07 London, UK (DAPPLE 2007) 2 Short range tracer releases (<500 m), 8 sonics in UCL, 1 reference sonic at rooftop. Wood et al. (2009, accept.)

1Urban Climate Zones (Oke, 2004, WMO)

URSL turbulence experiments 2001 - 2009

NCAS Urban Meteorology Workshop

• A. Christen / 30-31 March 2009

0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 1 10 100

UCZ 1 (’highrise core’) UCZ 2 (’compact core’) UCZ 3 (’dense residential’) UCZ 5 (’suburban’)

20 30 2 3 5 50

high density low density Plan area ratio of buildings λP Average building height (m)

URSL turbulence studies Morphometry of the urban surface

Va89 Oc03 Z86 Ba02u2 Tk01 Ba02s Sa92 Ba02u1 Go03 Ms01 Ba95 Na99 Ny05 Lo07

NCAS Urban Meteorology Workshop

• A. Christen / 30-31 March 2009

0.5 1 1.5 2 2.5 3 1 10 100

Trees in canyon No trees in canyon

20 30 40

narrow canyon wide canyon Street canyon aspect ratio zh/W

Turbulence studies in street canyons Canyon aspect ratio zh/W

Average building height (m) deep canyon shallow canyon

Oc03 Ba02u2 Oc03 Z86 Ba02u2 Ba02u1 Go03 Na99 Mc01 Rf97

Skimming flow

2 3 4

Hv94

WIF

NCAS Urban Meteorology Workshop

• A. Christen / 30-31 March 2009

Measured vertical profiles of mean flow and integral turbulence statistics Towards a conceptual division of the URSL

NCAS Urban Meteorology Workshop

• A. Christen / 30-31 March 2009

Basel-Sperrstrasse, 2002 Ba02u1

NCAS Urban Meteorology Workshop

• A. Christen / 30-31 March 2009

Christen et al., 2009

Variability of wind profiles with direction

NCAS Urban Meteorology Workshop

• A. Christen / 30-31 March 2009

Zu86 (Rotach 1991) Ba02u1 (Christen 2005) Ba02u2 (Christen 2005) Ba02s (Christen 2005)

0.5 1 1.5 2 2.5 3 3.5 4 0.5 1 1.5

‘Directionally’ averaged wind profiles

neutral conditions only

z/zh u(z)/u(2zh)

NCAS Urban Meteorology Workshop

• A. Christen / 30-31 March 2009

0.5 1 1.5 2 2.5 3 3.5 4

• 1.2
• 1
• 0.8
• 0.6
• 0.4
• 0.2

0.2

Zu86 (Rotach 1991) Sa92 (Oikawa and Meng, 1995) Ba95 (Feigenwinter et al. 1999) Ba02u1 (Christen et al. 2009) Ba02u2 (Christen et al. 2009) Ba02s (Christen 2005)

Vertical profiles of Reynolds stress

SL

z/zh u′w′/u∗

Air Buildings

zh

Building volume fraction (m3 m-3) at Ba02U1

NCAS Urban Meteorology Workshop

• A. Christen / 30-31 March 2009

0.5 1 1.5 2 2.5 3 3.5 4

• 0.5
• 0.375
• 0.25
• 0.125

0.125

Sl76 (Clarke at al. 1976) Zu86 (Rotach 1991) Va89 (Roth and Oke, 1991) Sa92 (Oikawa and Meng, 1995) Ba95 (Feigenwinter et al. 1999) Ba02u1 (Christen et al. 2009) Ba02u2 (Christen et al. 2009) Ba02s (Christen 2005)

Exchange efficiency for Reynolds stress

SL

z/zh ruw

ruw = u′w′ σuσw

NCAS Urban Meteorology Workshop

• A. Christen / 30-31 March 2009

Directional variability of ruw

NCAS Urban Meteorology Workshop

• A. Christen / 30-31 March 2009

Directional variability of ruw

NCAS Urban Meteorology Workshop

• A. Christen / 30-31 March 2009

Dispersive stress

Logistically nearly impossible to measure. Indirectly estimated through directional variation of mean horizontal and vertical wind at one location.

Ba02u1 (Christen et al. 2009)

NCAS Urban Meteorology Workshop

• A. Christen / 30-31 March 2009

The concept of ‘local scaling’

Local scaling (Högström et al., 1982) assumes a local equilibrium between local turbulence production and local dissipation and has been shown to work reasonably well above roof.

Roth (2000)

local friction velocity local Obukhov Length

NCAS Urban Meteorology Workshop

• A. Christen / 30-31 March 2009

0.5 1 1.5 2 2.5 3 3.5 4 0.5 1 1.5 2 2.5

Sl76 (Clarke at al. 1976) Zu86 (Rotach 1991) Va89 (Roth and Oke, 1991) Sa92 (Oikawa and Meng, 1995) Ba95 (Feigenwinter et al. 1999) Ba02u1 (Christen et al. 2009) Ba02u2 (Christen et al. 2009) Ba02s (Christen 2005) Ok03 (Hanna et al. 2007) Ny05 (Hanna et al. 2007)

Locally scaled velocity variances

SL neutral limit

z/zh Aw

Aw = σw/(u′w′2 + v′w′2)1/4

Systematic

• verestimation

NCAS Urban Meteorology Workshop

• A. Christen / 30-31 March 2009

Sl76 (Clarke at al. 1976) Zu86 (Rotach 1991) Va89 (Roth and Oke, 1991) Sa92 (Oikawa and Meng, 1995) Ba95 (Feigenwinter et al. 1999) Ba02u1 (Christen et al. 2009) Ba02u2 (Christen et al. 2009) Ba02s (Christen 2005) Oc03 (Hanna et al. 2007) Ny05 (Hanna et al. 2007)

Locally scaled velocity variances

SL neutral conditions only

Au z/zh

Au = σu/(u′w′2 + v′w′2)1/4

0.5 1 1.5 2 2.5 3 3.5 4 1.5 2 2.5 3 3.5 4

Systematic underestimation

NCAS Urban Meteorology Workshop

• A. Christen / 30-31 March 2009

Normalized third order moments

3 Sites during BUBBLE Ba02u1, Ba02u2, Ba02s (from Christen et al. 2006, AMS Urb Env.)

NCAS Urban Meteorology Workshop

• A. Christen / 30-31 March 2009

ΔS0 denotes the difference of the flux fractions of sweeps - ejections.

Ba02u1 (from Christen et al. 2007)

Quadrant analysis

Momentum flux Sensible heat flux

NCAS Urban Meteorology Workshop

• A. Christen / 30-31 March 2009

0.5 1 1.5 2 2.5 3 3.5 4

• 0.5
• 0.25

0.25 0.5

Zu86 (Rotach 1991) Sa92 (Oikawa and Meng, 1995) Ba95 (Feigenwinter et al. 1999) Ba02u1 (Christen et al. 2007) Ba02u2 (Christen et al. 2007) Ba02s (Christen 2005)

ΔS0 for Reynolds stress

neutral cases

z/zh

∆S0 = S4 − S2

∆S0

ejections dominate sweeps dominate

NCAS Urban Meteorology Workshop

• A. Christen / 30-31 March 2009

0.5 1 1.5 2 2.5 3 3.5 4

• 1.5
• 1
• 0.5

Zu86 (Rotach 1991) Ba95 (Feigenwinter et al. 1999) Ba02u1 (Christen et al. 2007) Ba02u2 (Christen et al. 2007) Ba02s (Christen 2005)

Exuberance of momentum exchange

neutral cases

z/zh

efficiency increases

Ex

Ex = (S1 + S3)/(S2 + S4)

NCAS Urban Meteorology Workshop

• A. Christen / 30-31 March 2009

2.17 1.53 1.23 1.01 0.77 0.25

• 4
• 3
• 2
• 1

1 2 3 4 Shear Production Wake Production Dissipation Turbulent transport Residual

Create TKE Remove TKE z / zh k h / u3 *

The neutral TKE budget

Christen et al., 2009 Ba02u1, directionally averaged

NCAS Urban Meteorology Workshop

• A. Christen / 30-31 March 2009

A conceptual division of the URSL based on field data

Blending layer 3 to 5 > z/zh > 1.5

• Wakes of individual buildings interact and mix more and more with height until

horizontal homogeneity is restored (‘blending height’).

• On the spatial average, vertical profiles of turbulent fluxes are constant with height,
• Velocity variances are approaching intertial sublayer predictions.
• Exchange of heat and momentum is dominated by ejections.

Shear layer 1.5 > z/zh > 0.7

• Profiles of mean wind, Reynolds stress, and higher order moments are all

characterized by strong vertical gradients.

• The wind profile – on average – shows an inflection point
• Mechanical production of turbulence is extraordinarily strong.
• A notable amount of this turbulence is exported through sweeps into the street

canyon and by ejections into higher layers.

• Neutral limits of velocity variances are significantly lower than predicted by local

scaling. Deep canopy layer 0.7 > z/zh >

• Flow is characterized by quasi-stationary patterns and dispersive fluxes might be

important.

• Local turbulence production is of minor importance
• Turbulence is mainly imported sweeps from rooftop.

NCAS Urban Meteorology Workshop

• A. Christen / 30-31 March 2009

The nature of exchange processes in the URSL Scales, coherent structures, and the dissimilarity of exchange.

NCAS Urban Meteorology Workshop

• A. Christen / 30-31 March 2009

Our ‘family portrait’ from urban canopies reveals many similarities to the ‘portrait’ of plant canopies (e.g. Finnigan, 2000) But be aware - the density (volume occupied by obstacles), the non-permeability, clustering and stiffness of buildings that characterize most obstacles in a real urban canopy compared to the flexible, porous and highly fractal structures that are present in vegetation canopies do not justify a physical analogy between vegetation canopy studies and urban environments a priori.

Canopy equal canopy?

NCAS Urban Meteorology Workshop

• A. Christen / 30-31 March 2009

Small scale local turbulence Large coherent structures

Christen et al. (2007)

One-point length scales

Ba02u1

NCAS Urban Meteorology Workshop

• A. Christen / 30-31 March 2009

Two-point correlations (temperature)

Ba02u1 Christen et al. (2007)

Isotropic turbulence Surface layer Shear layer

ze zh

NCAS Urban Meteorology Workshop

• A. Christen / 30-31 March 2009

Streamwise vorticity spacing Shear length scale

Is the plane mixing layer analogy adequate?

ze is height of inflection point (which is close to zh)

Ba02u1 Christen et al. (2006), AMS Symp. Urb Env.

NCAS Urban Meteorology Workshop

• A. Christen / 30-31 March 2009
• A number of recent urban turbulence studies demonstrated

that coherent structures are of importance in the URSL (Feigenwinter and Vogt, 2005; Salmond et al., 2005, Christen et al. 2007, Arms and Klein, 2009).

• It is still not clear how they are generated (single obstacles,

inflection point, outer-layer dynamics?)

Feigenwinter and Vogt (2005) Ba95

The role of coherent structures?

NCAS Urban Meteorology Workshop

• A. Christen / 30-31 March 2009

Strong ejection weak sweep ejection sweep

Conditional sampling - UCL and above roof

zh

Ba02u1

NCAS Urban Meteorology Workshop

• A. Christen / 30-31 March 2009

Sensible heat exchange

Ba02u1

NCAS Urban Meteorology Workshop

• A. Christen / 30-31 March 2009

Contribution of coherent structures to total exchange

3.2 (76m) 0.57 0.71 2.1 (50m) 0.69 0.90 1.5 (36m) 0.75 0.92

Feigenwinter and Vogt (2005) Average for afternoon situation (13-19, Tab. 3)

z/zh

{u′w′}/u′w′

{w′T ′}/w′T ′

NCAS Urban Meteorology Workshop

• A. Christen / 30-31 March 2009

Carbon dioxide exchange

Ms01 Marseile, France Salmond et al. (2005)

Carbon dioxide Vertical wind

NCAS Urban Meteorology Workshop

• A. Christen / 30-31 March 2009
• Small scale and local turbulence at roof-top (strongest

shear production) superimposed large coherent structures that span through the whole URSL (and above).

• Vertical wind velocity is related to the timing of large

scale temperature and horizontal variation and microfront-type events can be detected that are significant for total turbulent exchange.

• Intermittent exchange during calm nights.

Scales and structures - Key findings

NCAS Urban Meteorology Workshop

• A. Christen / 30-31 March 2009

Photo: M. Roth Vancouver-Sunset, 1978 / 1989

NCAS Urban Meteorology Workshop

• A. Christen / 30-31 March 2009

increasingly unstable neutral

sensible heat more efficient than water vapour water vapour more efficient than sensible heat

Roth and Oke, 1993

Scalar dissimilarity - Vancouver 1989

Va89

NCAS Urban Meteorology Workshop

• A. Christen / 30-31 March 2009

Time of day Time of day Ba02s Ba02u1 Ba02u1 Rural reference

Patchiness of sources?

rural: same efficiency during day urban: higher sensible heat lower water,

Urban and rural transfer efficiencies

Christen, 2005

Sensible heat Water vapour

NCAS Urban Meteorology Workshop

• A. Christen / 30-31 March 2009

0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 1 10 100

Plan area ratio of buildings λP Average building height (m)

Expanding the family - Covering a more diverse range of surfaces?

Blocks

Not studied at all

Modern core

Large efforts, but remains very challenging

Houses and Garden

Not studied in UCL

Old (compact) core

Well studied

Compact housing

Reasonably well studied

Shantytown

Not studied at all

Extensive Lowrise LCZ classificaion by Stewart and Oke (2009)

NCAS Urban Meteorology Workshop

• A. Christen / 30-31 March 2009

Future challenges?

• Need to research spatial statistics in the urban roughness

sublayer (e.g. applicability of Taylor’s hypothesis, dispersive fluxes, two-point statistics).

• The role of pressure fluctuations / pressure terms

(technically challenging).

• New mobile technology (UAVs, Largangian sensors) or

remote sensing (mini-LIDAR) might be promising, and could provide opportunities sample spatially.

• Thinking - experimental design is the key (not

technology, logistics, money, or computing power)