The Importance of Subgrid Scale Winds on Climate

1
The Importance of Subgrid Scale Winds on Climate
2
GOAL To Understand the Climatic Effects of
Unresolved Features

• Approach
• Improve representation of subgrid winds
• Weibull PDF
• High resolution observations
• Stability dependent variance (shape)

3
GOAL To Understand the Climatic Effects of
Unresolved Features

• Approach
• Improve representation of subgrid winds
• Weibull PDF
• High resolution observations
• Stability dependent variance (shape)

• Importance of Winds
• Influence non-linear processes
• CO2 gas exchange
• Momentum flux
• Sensible heat flux
• Latent heat flux
• Aerosol flux

4
Sensitivity Studies Single Mean vs. PDF

• Weibull PDF is a good fit to observed surface
  winds
• Justus et al. (1977,1979)
• Monahan (2005)
• Pavia and O'Brien (1986)

• LSM Algorithm
• Four bin Weibull PDF
• Positively skewed
• 2 m/s mean wind speed

5
Sensitivity Studies Single Mean vs. PDF

• LSM Algorithm
• Four bin Weibull PDF
• Positively skewed
• 2 m/s mean wind speed

Bin Speeds 0.47 m/s 1.24 m/s 2.17 m/s 4.13 m/s
6
Sensitivity Studies Single Mean vs. PDF
Atmospheric RH10 Net Solar800 W/m2 Rough.
Length0.01 m Volumetric Water Content0.20 m3 m-3
Atmospheric RH50 Net Solar1000 W/m2 Rough.
Length0.10 m Volumetric Water Content0.0 m3 m-3
7
Sensitivity to Shape
Shape2.925
Shape2.312
8
Sensitivity to Shape
Shape2.925
Shape2.312
9

• QuikSCAT Observations
• 1999 - Current
• 10m surface wind speeds (m/s)
• Empirically derived relationship between ocean
  roughness and wind speed
• Asc/desc pass 600am/600pm

QuikSCAT 10m Wind Speeds (m s-1) July 2, 2006 at
6pm local time
QuikSCAT Grid Cell (0.25oX0.25o)
T42 Grid Cell
10

• QuikSCAT Observations
• 1999 - Current
• 10m surface wind speeds (m/s)
• Empirically derived relationship between ocean
  roughness and wind speed
• Asc/desc pass 600am/600pm

QuikSCAT 10m Wind Speeds (m s-1) July 2, 2006 at
6pm local time
QuikSCAT Grid Cell (0.25oX0.25o)
T42 Grid Cell

• Known Issues
• Rain cells discarded
• Underestimates wind speeds gt 25 m/s
• Directional errors at low wind speeds
• In situ vs. QuikSCAT RMS error differences lt1 m/s
  and 15o (Bourassa et al. 2003)

11
QuikSCAT Surface Winds 2005
A
Shape2.16 Scale10.67 m s-1
A
A
A
12
QuikSCAT Surface Winds 2005
A
Shape2.16 Scale10.67 m s-1
A
B
A
B
B
Shape7.01 Scale7.94 m s-1
A
B
13
Proposed Research I Characterizing QuikSCAT
Winds and Evaluating Model Representation
Sub-sample from CCSM output

• QuikSCAT
• 2/day x 365 days/year x 6 years 4,380
• CCSM
• 72/day x 365 days/year x 10 years 262,800

• QuikSCAT
• 2/day x 365 days/year x 6 years 4,380
• Spatial average of the temporal PDF
• CCSM
• 2/day x 365 days/year x 10 years 7,300
• Temporal PDF

14
Proposed Research I Characterizing QuikSCAT
Winds and Evaluating Model Representation

• Statistical Analysis
• Monthly, Seasonal and Annual timescales
• Dynamic Climate Regions
• Warm pool areas
• Storm tracks
• Monsoons
• ENSO
• Differences in diurnal, seasonal and other
  temporal cycles
• Statistics
• Skewness
• Scale/Mean
• Shape/Standard Deviation/Variance
• Extremes

15
Proposed Research I Characterizing QuikSCAT
Winds and Evaluating Model Representation

• Statistical Analysis
• Monthly, Seasonal and Annual timescales
• Dynamic Climate Regions
• Warm pool areas
• Storm tracks
• Monsoons
• ENSO
• Differences in diurnal, seasonal and other
  temporal cycles
• Statistics
• Skewness
• Scale/Mean
• Shape/Standard Deviation/Variance
• Extremes

Evaluation of surface flux biases resulting from
these errors in surface wind speed Are these
biases similar to differences in TOGA TAO vs.
CCSM surface fluxes?
16
Preliminary Comparisons Mean Wind Speed

• CCSM winds gt QuikSCAT
• Over strong currents
• North Equatorial
• Antarctic Circumpolar
• Kuroshio
• North Atlantic Drift
• South Equatorial
• CCSM winds lt Observed
• Western Pacific warm pool
• Throughout the ITCZ

CCSM Data NOT SUBSAMPLED
17
Preliminary Comparisons

• CCSM shape lt QuikSCAT
• Antarctic Circumpolar
• N. H. Mid-latitude Regions
• Equatorial Pacific
• CCSM Shape gt QuikSCAT
• West of Australia
• Equatorial Atlantic

CCSM Data NOT SUBSAMPLED
18
Proposed Research II Implementing the
Stability Dependent PDF
Velocity Scales (Cakmur et al. 2004)
Turbulence Kinetic Energy
Dry (Free) Convection
Turbulence
Gust Fronts
Richardson Number
R. V. Cakmur and R. L. Miller Incorporating the
effect of small-scale circulations upon dust
emission in an atmospheric general circulation
model, Journal of Geophys. Research, 109 2004
19
Proposed Research II Implementing the
Stability Dependent PDF

• 2-Parameter Weibull PDF
• Equal probability bins
• Shape is a function of
• Mean wind speed
• Atmospheric stability

Surface Winds
20
Proposed Research II Implementing the
Stability Dependent PDF

• Relationship Between Turbulence and Shape
• For regions where CCSM mean wind agrees with
  observations
• Evaluate shape/scale differences
• Determine relation between these differences and
  turbulence velocity scales
• TOGA TAO
• Verify CCSM and observation differences
• Does physical PDF bring CCSM fluxes closer to
  observations?

TOGA TAO
21
Preliminary Results PDF (not influenced by
stability)
22
(No Transcript)
23
Proposed Research III Roughness Length
Enhancements

• Improving CCSM Bare Ground Roughness Length
• Roughness length (z0) is a function of soil type
• Real World - Drag is partitioned between
  vegetation and ground
• Employ Raupach (1992 and 1994) models
• Bare ground z0 will account for intermixed
  vegetation

24
Proposed Research III Roughness Length
Enhancements

• Improving CCSM Representation of Vegetated
  Surface Roughness Lengths
• Vegetation z0 is only a function of height
• We will incorporate a more realistic z0 using
• Raupach's models
• Remote sensing studies
• Collaboration with plant physiological experts

25
Proposed Timeline
2006 - CCSM vs. Observed Surface Wind Comparison
2007 - Dynamic Wind Speed PDF Implementation and
Feedback Analysis
2008 Drag Partitioning Implementation and
Climate Sensitivity Studies
2009 - Improvement of Vegetation Structure
Representation