Transcript Diapositivo 1 - Copernicus.org

Effects Of Different Model Lower Boundary Conditions In
The Simulation Of An Orographic Precipitation Extreme
Event
Physics Department – University of
Aveiro
J. Teixeira, A. C. Carvalho, T. Luna and A. Rocha
Correspond to: [email protected]
Introduction
Atmospheric Models are sensible to lower boundary conditions
Topography forced processes are difficult to simulate accurately
It is expected → Better description of the lower boundary → Better results
→ Topography driven precipitation
→ Wind flow paths
Introduction
5 different lower boundary datasets were used
Topography
Land use
– GTOPO 30
– USGS Land Use
Resolution = 30”
Year = 1996
– SRTM
Resolution = 3”
Year = 2005
– ASTER
Default in WRF
Resolution = 30”
Year = 1993
Categories = 25
– CORINE Land Cover
Resolution = 100 m
Year = 2006
Catgories = 44
Resolution = 1”
Year = 2006
Recategorisation according to Pineda et al. (2004)
in order to be compatible with WRF
Objectives
→ Study WRF model sensitivity to different lower boundary conditions in an
extreme orographic precipitation event
Case Study → Extreme precipitation over Madeira island – 20 de February de 2010
Method
Model Configuration
→ Triple domain with two-way nesting
d01
d02
d03
Horizonta Res. (km)
25
5
1
Time step (s)
150
30
6
Method
→ Observed data location
→ ● Portuguese Meteorological Institute
→ ○ Madeira's Regional Laboratory of Civil Engineering
Method
It is considered that the model has skill when:
– Modelled standard deviation
approximate to the observed
• S ~ Sobs
– Bias squared less than the error
squared
• Bias2 << E2
– Model root mean squared error smaller than the observed standard deviation
• E < Sobs
• EUB < Sobs
Method
Sinoptic Setting – 20 February at 1200 UTC
→ Quick transition from a hight to a low pressure system
→ Large amount of precipitable water available over Madeira – Atmospheric river
Sea level pressure (hPa)
Precipitable water (mm)
Results
Topography differences (SRTM – GTOPO30) – WRF 1 km (d03)
→ Higher summits and deeper valleys
→ GTOPO30 topography is smother
→ Better representation of areas with steep slopes (ex: Ponta do Parco – West)
→ Similar differences for ASTER – GTOPO30
GTOPO30
SRTM – GTOPO30
Results
10 m wind intensity difference (SRTM – CTL)
→ Main differences are located over the island
→ High correlation with topography differences (~ 0.6 – SRTM e ASTER)
→ Small differences at leeward
CTL
Mean
Difference
Results
Total accumulated precipitation difference (SRTM – CTL)
→ Large differences in Madeira's mountainous region
→ More precipitation in the summits
→ Less precipitation in the valleys
→ Correlation with the topography difference of 0.36 – SRTM and 0.46 – ASTER
→ Similar differences for ASTER simulation
CTL
SRTM – CTL
Results
USGS land use
CORINE land use
Results
CORINE – CTL differences
→ There are only small differences for this particular event – specially for precipitation
→ Topography gives greater differences
10 m mean wind intensity difference
Total accumulated precipitation
difference
Results
Taylor Diagrams – Wind
→ Low skill simulating wind
→ Better model performance when the new boundary is used for v wind component
→ Worse model performance when the new boundary is used for u wind component
Componente u
Componente v
Results
Skill Diagrams – Precipitation
→ There is skill in simulating precipitation
→ Similar skill results between different simulations
→ Worse skill when the new boundary condition is used
Taylor Diagram
Skill Diagram
Results
Skill Diagrams – Regions
→ 4 distinct regions have been defined:
– Mountainous
– Windward
– Coastal
– Leeward
→ Better model skill for the Coastal region – Wind and Precipitation.
Particular case for SRTM
Windward / Leeward
→ Worse skill result fot precipitation and better for wind at Leeward
→ Better skill result for precipitation and worse for wind at Windward
Precipitation – Leeward
v wind component – Windward
Concluding Remarks
→ Large differences between the new boundary and default model datasets
→ There is a change in modelled results – Precipitation and Wind
→ There is a local enhancement of model skill in simulating this extreme
precipitation event
However dependent on the representativeness of the
location of the observations