Determination of the Thermal Roughness Length for a Built - PowerPoint PPT Presentation

Determination of the Thermal Roughness Length for a Built Environment using High Resolution Weather Stations Daniel Nadeau E. Bou-Zeid, M. B. Parlange, G. Barrenetxea, M. Vetterli Stockholm, 11 June 2008

Motivations Urban population is increasing � now: 3.3 billion � 2030 prediction: 5 billion (UN, 2008) Larger stress of built-up areas on the atmosphere Need to model land-atmosphere interactions in urban areas � surface roughness z 0 � thermal roughness length z 0 h Two methods to do so: morphometric or micrometeorological 2

Motivations Morphometric approaches � � use building geometry to calculate roughness parameters b ildi t t l l t h t different models often lead to widely varying estimates of roughness characteristics Micrometeorological approaches Stockholm area as seen by Modis � surface temperature typically inferred from satellite measurements � MODIS:1-km spatial resolution for T sfc Resolution too low to account for spatial heterogeneities Source: NASA, 2008 3

Motivations Morphometric approaches � use building geometry to calculate roughness parameters � b ildi t t l l t h t different models often lead to widely varying estimates of roughness characteristics Micrometeorological approaches Stockholm area as seen by Modis � surface temperature typically inferred from satellite measurements � MODIS:1-km spatial resolution for T sfc Resolution too low to account for spatial heterogeneities Source: NASA, 2008 Need for high resolution measurements of urban surfaces 4

Research Objectives • better understand the impacts of spatial heterogeneities on l land-atmosphere interactions over complex urban terrain d t h i t ti l b t i • calculate roughness lengths for momentum ( z 0 ) and for heat ( z 0 h ) using in situ measurements Source: S. Mortier, 2007 5

Background Thermal roughness length z 0 h z z 0 + d 0 z 0 h + d 0 z = 0 Source:Voogt and Grimmond, JAM , 2000 • also referred to as radiometric roughness length or scalar roughness for heat • intercept of the logarithmic profile for potential temperature in the inertial sublayer 6

Background Thermal roughness length z 0 h over heterogeneous surfaces vegetated small hills Sugita and Brutsaert ( WRR , 1990) flat grassland with low hedges Hignett ( BLM , 1994) sparsely vegetated surfaces Mahli ( QJRMS , 1996) light industrial area g Voogt and Grimmond ( JAM , 2000) 10 -17 10 -16 10 -15 10 -14 10 -13 10 -12 10 -11 10 -10 10 -9 10 -7 10 -6 10 0 10 -8 10 -5 10 -4 10 -3 10 -2 10 -1 z 0 h (m) 7

Background Monin-Obukhov Similarity Theory in the ABL WIND SPEED ⎡ ⎤ ⎛ ⎞ − − ⎛ ⎞ ⎛ ⎞ u * ln z d z d z = = − Ψ Ψ + Ψ + Ψ 0 0 0 ln ⎢ ⎢ ⎥ ⎥ ⎜ ⎜ ⎟ ⎟ ⎜ ⎜ ⎟ ⎟ ⎜ ⎜ ⎟ ⎟ u u m m ⎝ ⎠ ⎝ ⎠ ⎢ ⎥ ⎝ ⎠ k z L L ⎣ ⎦ 0 z : surface roughness (m) z 0 : surface roughness (m) AIR TEMPERATURE ⎡ ⎡ ⎤ ⎤ ⎛ ⎛ ⎞ ⎞ − − ⎛ ⎞ ⎛ ⎞ H z d z d z θ − θ = − Ψ + Ψ 0 0 0 ln ⎢ ⎥ ⎜ ⎟ h ⎜ ⎟ ⎜ ⎟ ρ s h h ⎝ ⎠ ⎝ ⎠ ⎢ ⎥ ⎝ ⎠ ku c z L L ⎣ ⎦ * 0 p h z 0 h : thermal roughness length (m) 8

The EPFL Campus a 750 x 500 m campus essentially consisting of buildings, vegetation, roads, and parking lots S Source: EPFL, 2008 EPFL 2008 9

10

The Experimental Setup Sensorscope stations • 92 wireless weather stations • operating from Nov. 06 to May 07 operating from Nov. 06 to May 07 • sampling time of 2 min, but we use 30 min averages • parameters measured: skin t temperature, air temperature, wind t i t t i d speed, relative humidity, etc. 11

12 SODAR SODAR The Experimental Setup N

The Experimental Setup SODAR / RASS system • operating from Jul. 06 to May 07 • wind and temperature profiles i d d fil measured from 40 to 400 m • averaging time of 30 min 13

Finding Neutral Profiles Range of interest • Buildings range from 5 to 30 m g g • Assume blending height ≈ 2 x h 0 • z max / z min > 2 (Bottema, AE , 1997) ⇒ ⇒ = = = [ [ 40 40, 100] m 100] m z z z z z z min max ⇒ = 20 m d (estimated) 0 Adapted from Britter and Hanna, ARFM , 2003 Criteria for near-neutral conditions 1) consistent wind direction with height u > 5 m/s > 5 / 2) 2) ∂ θ g v 3) Least-square fitting between u and ln( z - d 0 ) ∂ θ z = ⎡ v Ri yields R² ≥ 0.5 ⎤ g 2 2 ⎛ ⎞ ⎛ ⎞ ∂ ∂ U V ⎢ ⎢ ⎥ ⎥ + + 4) 4) | Ri | ≤ 0 1 | Ri g | ≤ 0.1 ⎜ ⎜ ⎟ ⎟ ⎜ ⎜ ⎟ ⎟ ∂ ∂ ⎢ ⎥ ⎝ z ⎠ ⎝ z ⎠ ⎣ ⎦ 14

15 90° 90 120° 60° 120 60 Wind Sectors 150° 30° 180° 180 0° 0 330° 210° 240° 300° 270°

Finding Neutral Profiles 0° 330° 30° 0 330 30 300° 60° 60 60 300 300 270° 90° 270 90 100 200 240 240 120° 240° 120 300 400 210 150 150° 210° 180 180° Number of cases of wind speed exceeding 5 m/s Number of cases of wind speed exceeding 5 m/s. Measurements at 50 m from July 06 to May 07. 16

Momentum Surface Roughness no data 0° 330° 30° 0 330 30 300° 60° 60 60 300 300 270° 90° 270 90 100 200 240 240 120° 240° 120 300 400 210 150 150° 210° 180 180° Number of cases of wind speed exceeding 5 m/s Number of cases of wind speed exceeding 5 m/s. 2.00 2 00 Measurements at 50 m from July 06 to May 07. 1.75 1.50 1.25 m) z 0 (m Center of large towns and cities (Stull, 1988) Center of large towns and cities (Stull 1988) 1.00 0.75 0.50 Total number of profiles: 16 128 0.25 Near-neutral profiles: 108 (0.7 %) 0 270° 300° 330° 0° 30° 60° wind sector Median of surface roughness distribution for wind 17 sectors with sufficient data.

Thermal Roughness Length Regression for near-neutral potential temperature profiles 3 Nov. 2006 at 7:30 pm 5.5 5 4.5 d 0 ) (m) 4 ln(z-d 3.5 3 2 5 2.5 2 ln(z 0h ) 1.5 1 -3.5 -3 -2.5 -2 -1.5 -1 -0.5 0 θ s - θ (K) 18

Thermal Roughness Length Preliminary results for z 0 h For bluff-rough surfaces 0 10 ) ( ) ( ⎡ ⎤ = − − 0.247 exp 4.31Re 5 z z k ⎣ ⎦ 0 0 * h (Cahill et al., WRR , 1997) -5 10 u z = * 0 0 Re Re z 0h (m) * ν -10 10 bluff rough sfcs bluff-rough sfcs using θ sfc 0 330 30 median for θ sfc cases 60 300 using θ air 270 90 240 median for θ air cases 120 210 -15 15 150 150 180 10 0 5 10 15 Re * 4 x 10 19

Thermal Roughness Length Considering surface type SKIN TEMPERATURE AIR TEMPERATURE 30 30 Stations over VEGETATION Stations over VEGETATION Stations over GRAVEL Stations over GRAVEL Stations over GRAVEL Stations over GRAVEL 25 25 Stations over CONCRETE perature [°C] Stations over CONCRETE rature [°C] 20 20 15 15 15 15 Skin Tem Air Tempe 10 10 5 5 0 0 00:00 06:00 12:00 18:00 00:00 00:00 06:00 12:00 18:00 00:00 14 Mar 2007 14 Mar 2007 20

Thermal Roughness Length θ = θ + θ a b Using a weighted average: ,vegetation ,urban s s s For a 10 km fetch 0 10 a : estimated fractional cover of vegetation b : estimated fractional cover of built-up areas -5 10 z 0h (m) -10 10 bluff-rough sfcs g using θ sfc 0 330 30 median for θ sfc cases 60 300 using θ air 270 90 median for θ air cases 240 using weighted θ sfc 120 210 using weighted θ sfc cases i ht d θ 150 150 i 180 -15 10 0 5 10 15 Re * 4 x 10 21

Conclusions and Future Work Conclusions • momentum surface roughness obtained by regressing near neutral profiles • momentum surface roughness obtained by regressing near-neutral profiles • large values of z 0 h found (compared to literature): from 10 -6 to 1 m • z 0 h very far from approximation for bluff-rough surfaces Future work • study convective cases for z 0 h study convective cases for z 0 h • footprint analysis • compare with morphometric models • perform LES simulations p Source: E. Ouyang, E. Bou-Zeid, 2007 22

Future Work Modeling shaded areas • shaded areas can greatly shaded areas can greatly influence skin temperature and in turn the spatially averaged heat flux (Sun and Mahrt, BLM , 1995) • dependance of z 0 h on the sun angle (Kustas et al., AFM , 1989) • heat flux dominated by sunlit areas (Voogt and Grimmond, JAM , 2000) 23

24 Thank you !

Thermal Roughness Length Dependance of z 0 / z 0 h on the flow Source: Brutsaert, Evaporation into the Atmosphere , 1982 25

Instruments SODAR/RASS accuracy u horizontal: 0 1 - 0 3 m/s u horizontal: 0.1 0.3 m/s u vertical: 0.03 - 0.1 m/s wind direction: 2 - 3 ° thickness of vertical layers: 5 – 100 m y range: 200 – 500 m Scintec Flat Array SFAS temperature: 0.2 ° C Sensorscope accuracy surface temperature: 0.6 ° C air temperature: 0.3 ° C p Sensorscope station