mistralanditssiblingsinrcms - goethe university frankfurt · 2014-07-01 ·...

1
Mistral and its Siblings in RCMs A NIKA O BERMANN ,B ODO A HRENS IAU, Goethe University Frankfurt Presented at COSMO/CLM/ART User Seminar 2014 in Offenbach D AILY M EAN W IND S PEED D ATA Buoy data daily mean wind speed from hourly measurements at Côte d’Azur and Gulf of Lion [1]. QuikSCAT gridded 0.25 scatterometer data of wind speed over sea (measurements 6 a.m and 6 p.m. local time) [2] fitted to buoy measurements at Gulf of Lion [3]. SAFRAN gridded daily mean wind speed over France using optimal interpolation method [4]. ERA-Interim daily mean windspeed [5]. M ODELS ERA-Interim driven runs on Med-CORDEX domain: COSMO-CLM [6] with 0.44 and 0.088 resolution. ALADIN-climate [7] with 0.44 and 0.11 resolution. E STIMATING F ETCH Distance the wind travelled over water before reaching grid cell estimated on QuikSCAT grid: 1. Distance to coastal cells in each direction (black cir- cles). 2. Line of Sight: Shading of far away cells by closer cells (red line). 3. Avaraging due to uncer- tainties in wind direction: Fetch in φ ± 90 interval weighted with a cos 2 func- tion (blue line). Fetch at Gulf of Lion buoy. M ISTRAL &S IBLINGS Mistral: Regional cold and dry north to northwest- erly wind following the lower Rhône Valley. It causes deep water forma- tion in the Mediterranean Sea. Therefore it is im- portant for modeling of the circulation in the Mediter- ranean Sea. Tramontane: In valley be- tween Pyrenees and Massif Central. Cierzo: Along Ebro Valley south of Pyrenees. Terrain height (m) of COSMO-CLM 0.088 Med-CORDEX run. Circles: buoys, triangles: surface stations. W IND T IME S ERIES Gusts 16 m s and gust direction at 13 stations in France [8]. Mistral (19%): At least one station in the valley, one at its exit, and one close to sea show gusts. Tramontane (43%): At least one inland station and one station close to sea show gusts. F ETCH D EPENDENCE OF E RRORS Fetch (0.25 pixels) on 09.02.2000. Mistral Days Tramontane Days Other Days Rel. Bias RMSE Red: COSMO-CLM 0.44 , orange: COSMO-CLM 0.088 , blue: ALADIN 0.44 , light blue: ALADIN 0.11 , yellow: ERA-Interim. Relative bias and RMSE 2000- 2008 in Mistral and Tramontane area (black square). Negative relative bias de- creases with increasing fetch (except for COSMO- CLM 0.088 ). RMSE decreases with in- creasing fetch and shows a drop at about 12 pixels on windy days. W IND S PEED Mean wind speed 2000-2008 fol- lowing Mistral and Tramontane tracks (black lines in right fig- ure). Rhône Valley (blue ar- rows): Increase in wind speed during Mistral days. Coast (black arrows): Lo- cation, shape and amount of wind speed gain in mod- els differ from observations for Mistral and Tramon- tane days. Observed wind speed (m/s) on 09.02.2000. Mistral Days Tramontane Days Other Days Mistral Tramontane Black: Observations, red: COSMO-CLM 0.44 , orange: COSMO-CLM 0.088 , blue: ALADIN 0.44 , light blue: ALADIN 0.11 , yellow: ERA-Interim. O UTLOOK Models work fine for non- windy days but underestimate wind speed in the Gulf of Lion on windy days. Possible modifications of Charnock formula for high winds and fetch? Error propagation from land to sea? Funneling in valley? Deceleration at valley exit? Other effects? Effects in other winds: Bora, Cierzo, ... R EFERENCES [1] provided by S. Belamari, Meteo France [2] Physical Oceanography DAAC (2001): SeaWinds on QuikSCAT Level 3 Daily, Gridded Ocean Wind Vectors (JPL SeaWinds Project): Guide Document [3] P. M. Ruti et Al. (2008): Comparison of analyzed and mea- sured wind speeds in the perspective of oceanic simulations over the Mediterranean basin: Analyses, QuikSCAT and buoy data, Journal of Marine Systems 70: 33-48 [4] J. P. Vidal et Al. (2010): A 50-Year High-Resolution Atmo- spheric Reanalysis Over France with the Safran System, INT J CLIMATOL 30 (11): 1627-1644 [5] D. P. Dee et Al. (2011): The ERA-Interim reanalysis: config- uration and performance of the data assimilation system, Q J ROY METEOR SOC 137 (656): 553-597, April 2011 Part A [6] G. Doms et Al. (2011): A Description of the Nonhydrostatic Regional COSMO Model, Part II: Physical Parameterization [7] M. Herrmann et Al. (2011): Representation of daily wind speed spatial and temporal variability and intense wind events over the Mediterranean Sea using dynamical down- scaling: impact of the regional climate model configuration, Nat. Hazards Earth Syst. Sci. 11: 1983-2001 [8] V. Jacq (2011): personal communication

Upload: others

Post on 25-Jul-2020

3 views

Category:

Documents


0 download

TRANSCRIPT

Page 1: MistralanditsSiblingsinRCMs - Goethe University Frankfurt · 2014-07-01 · MistralanditsSiblingsinRCMs ANIKA OBERMANN, BODO AHRENS IAU, Goethe University Frankfurt Presented at COSMO/CLM/ART

Mistral and its Siblings in RCMsANIKA OBERMANN, BODO AHRENS

IAU, Goethe University Frankfurt

Presented at COSMO/CLM/ART User Seminar 2014 in Offenbach

DAILY MEAN WIND SPEED DATA

• Buoy data daily mean wind speed from hourly measurementsat Côte d’Azur and Gulf of Lion [1].

• QuikSCAT gridded 0.25◦ scatterometer data of wind speedover sea (measurements ≈ 6 a.m and 6 p.m. local time) [2] fittedto buoy measurements at Gulf of Lion [3].

• SAFRAN gridded daily mean wind speed over France usingoptimal interpolation method [4].

• ERA-Interim daily mean windspeed [5].

MODELS

ERA-Interim driven runs on Med-CORDEX domain:

• COSMO-CLM [6] with 0.44◦ and 0.088◦ resolution.• ALADIN-climate [7] with 0.44◦ and 0.11◦ resolution.

ESTIMATING FETCHDistance the wind travelled overwater before reaching grid cellestimated on QuikSCAT grid:

1. Distance to coastal cells ineach direction (black cir-cles).

2. Line of Sight: Shading offar away cells by closercells (red line).

3. Avaraging due to uncer-tainties in wind direction:Fetch in φ ± 90◦ intervalweighted with a cos2 func-tion (blue line).

Fetch at Gulf of Lion buoy.

MISTRAL & SIBLINGS

• Mistral: Regional cold anddry north to northwest-erly wind following thelower Rhône Valley. Itcauses deep water forma-tion in the MediterraneanSea. Therefore it is im-portant for modeling of thecirculation in the Mediter-ranean Sea.

• Tramontane: In valley be-tween Pyrenees and MassifCentral.

• Cierzo: Along Ebro Valleysouth of Pyrenees.

Terrain height (m) of COSMO-CLM0.088◦ Med-CORDEX run. Circles:buoys, triangles: surface stations.

WIND TIME SERIESGusts ≥ 16m

s and gust directionat 13 stations in France [8].

• Mistral (≈ 19%): At leastone station in the valley,one at its exit, and oneclose to sea show gusts.

• Tramontane (≈ 43%): Atleast one inland stationand one station close to seashow gusts.

FETCH DEPENDENCE OF ERRORS

Fetch (0.25◦ pixels) on 09.02.2000.

Mistral Days Tramontane Days Other Days

Rel

.Bia

sR

MSE

Red: COSMO-CLM 0.44◦, orange: COSMO-CLM 0.088◦, blue: ALADIN0.44◦, light blue: ALADIN 0.11◦, yellow: ERA-Interim.

Relative bias and RMSE 2000-2008 in Mistral and Tramontanearea (black square).

• Negative relative bias de-creases with increasingfetch (except for COSMO-CLM 0.088◦).

• RMSE decreases with in-creasing fetch and shows adrop at about 12 pixels onwindy days.

WIND SPEED

Mean wind speed 2000-2008 fol-lowing Mistral and Tramontanetracks (black lines in right fig-ure).

• Rhône Valley (blue ar-rows): Increase in windspeed during Mistral days.

• Coast (black arrows): Lo-cation, shape and amountof wind speed gain in mod-els differ from observationsfor Mistral and Tramon-tane days.

Observed wind speed (m/s) on09.02.2000.

Mistral Days Tramontane Days Other Days

Mis

tral

Tram

onta

ne

Black: Observations, red: COSMO-CLM 0.44◦, orange: COSMO-CLM 0.088◦,blue: ALADIN 0.44◦, light blue: ALADIN 0.11◦, yellow: ERA-Interim.

OUTLOOKModels work fine for non-windy days but underestimatewind speed in the Gulf of Lionon windy days.

• Possible modifications ofCharnock formula forhigh winds and fetch?

• Error propagation fromland to sea?

– Funneling in valley?– Deceleration at valley

exit?– Other effects?

• Effects in other winds:Bora, Cierzo, ...

REFERENCES[1] provided by S. Belamari, Meteo France

[2] Physical Oceanography DAAC (2001): SeaWinds onQuikSCAT Level 3 Daily, Gridded Ocean Wind Vectors (JPLSeaWinds Project): Guide Document

[3] P. M. Ruti et Al. (2008): Comparison of analyzed and mea-sured wind speeds in the perspective of oceanic simulationsover the Mediterranean basin: Analyses, QuikSCAT andbuoy data, Journal of Marine Systems 70: 33-48

[4] J. P. Vidal et Al. (2010): A 50-Year High-Resolution Atmo-spheric Reanalysis Over France with the Safran System, INTJ CLIMATOL 30 (11): 1627-1644

[5] D. P. Dee et Al. (2011): The ERA-Interim reanalysis: config-uration and performance of the data assimilation system, QJ ROY METEOR SOC 137 (656): 553-597, April 2011 Part A

[6] G. Doms et Al. (2011): A Description of the NonhydrostaticRegional COSMO Model, Part II: Physical Parameterization

[7] M. Herrmann et Al. (2011): Representation of daily windspeed spatial and temporal variability and intense windevents over the Mediterranean Sea using dynamical down-scaling: impact of the regional climate model configuration,Nat. Hazards Earth Syst. Sci. 11: 1983-2001

[8] V. Jacq (2011): personal communication