Comparison of LM Verification against Multi Level Aircraft Measurements (MLAs)

Comparison

ofLM Verification against Multi Level Aircraft Measurements (MLAs)

with LM Verification against Temps

Ulrich Pflüger, Deutscher Wetterdienst

Comparisonof

LM Verification against Multi Level Aircraft Measurements (MLAs)with LM Verification against Temps

• by means of Time Series (June 2001 until May 2003) of monthly average of – bias and rmse of 00h, 24h, 48h Forecasts based on all MLAs/Temps in

LM area

– Variables are• temperature

• wind velocity

• wind direction

Multi Level Aircraft Measurementsare

• Several (not less than 4) single Aircraft Measurements during the climb or descent flight

• which are piecewise combined– beginning with the lowest

– next < 20 km horizontal distance of the first

– vertical distance < 55 hPa

– etc.

• and then used like a Temp Measurement

LM Area

All MLAs at 12 UTC

2003/05/01-2003/05/19

LM Area

All Temps Stations

12 UTC

2003/05/01-2003/05/19

Bias of LM-Temperatureagainst

MLA Temps

00 UTC

Analysis Analysis

24 h Forecast 24 h Forecast

48 h Forecast 48 h Forecast

small bias, similar structures in both cases

Forecasts have annual cycle, in upper and middle tropospherepositive bias in Summer, negative in Winter, near ground predominantly negativ bias

RMSE of LM-Temperatureagainst

MLA Temps

00 UTC

1.41.6

1.61.4

2.22.01.8 2.0

similar structures in both cases

0.8 0.8

similar structures in both cases, rmse(Temps) overall about 0.2 K greater than rmse(MLA)

Bias of LM-Temperatureagainst

MLA Temps

12 UTC

Negative bias> -0.3 K

positiv bias< 0.4 K

below 800 hPApredominantly negativ,

below 800 hPApredominantly negativ

below 800 hPASummer negativ,Winter positiv

below 800 hPASummer negativ,Winter positiv

RMSE of LM-Temperatureagainst

MLA Temps

12 UTC

1.41.6

1.61.4

2.22.01.6 1.8

Same as 00 UTC

Differences at Temperature Verification

• Relative few differences between Verification based on MLAs and Temps

• When bias is small there is predominantly a negative bias against MLAs and a positive bias against Temps

• Overall rmse of 24/48 h Forecast is about 0.2 K greater in Verification based on Temps

Bias of LM-Wind Velocityagainst

MLA Temps

00 UTC

Below 800 hPa, annual cycle, in winter strong positive bias Max.: 1.4 m/s

Below 800 hPa , annual cycle, in winter strong positive bias Max.: 1.8 m/s

Below 800 hPa , annual cycle, in winter strong positive bias Max.: 1.8 m/s

Below 900 hPa predominantly negative bias Min.: -1.1 m/s, Winter less

Below 900 hPa predominantly negative bias Min.: -0.8 m/s, Winter partly positive Max 0.2 m/s

Below 900 hPa predominantly negative bias Min.: -0.8 m/s, Winter partly positiv Max 0.2 m/s

RMSE of LM-Wind Velocityagainst

MLA Temps

00 UTC

Overall rmse(MLA) is about 0.5-1 m/sgreater than rmse(Temps)

Overall rmse(MLA) is equalto rmse(Temps)

Above 400 hPa rmse(Temps) is greaterthan rmse(MLA)

3.5 3 3.5 3

Bias of LM-Wind Velocityagainst

MLA Temps

12 UTC

Below 800 hPa, annual cycle, in winter strong positive bias, Max.: 1.9 m/s

Below 800 hPa , annual cycle, in winter strong positive bias, Max.: 2.4 m/s

Below 800 hPa , annual cycle, in winter strong positive bias, Max.: 2.3 m/s

Below 900 hPa predominantly negative bias Min.: -1.1 m/s,Winter partly positive, Max.: 0.2 m/s

Below 800 hPa Summer negative bias Min.: -0.6 m/s, Winter positive, Max 0.7 m/s

Below 800 hPa Summer predominantly negative bias Min.: -0.6 m/s, Winter predominantly positive, Max 0.8 m/s

RMSE of LM-Wind Velocityagainst

MLA Temps

12 UTC

3 2

4 3 4 3

Same as 00 UTC

Differences at Wind Velocity Verification

• bias with annual cycle– below 800 hPa

• positive against MLA

• negativ(Winter) to slight positive(Summer) 00 UTC

• negativ(Winter) to positive(Summer) 12 UTC

• rmse with annual cycle – at Analysis overall 0.5 -1 m/s higher against MLA

(wind is nudged to Temps)

– at Forecasts higher rmse against Temps above 400 hPa

Bias of LM-Wind Directionagainst

MLA Temps

00 UTC

Below 800 hPa, strong positive bias Max.: 7.4 degree, smaller in spring and autumn

Below 800 hPa, strong positive bias Max.: 10.3 degree

Below 800 hPa, strong positive bias Max.: 13.5 degree

Below 900 hPa negative bias Min.: -4.9 degree, Winter less

Below 900 hPa predominantly negative bias Min.: -7 degrees, Winter partly positive Max.: 6 degrees

Below 900 hPa predominantly negative bias Min.: -10 degrees, Winter partly positiv Max 6 degrees

RMSE of LM-Wind Directionagainst

MLA Temps

00 UTC

20

30 30

40 40

10

Overall rmse(MLA) is about 5-10 degreesgreater than rmse(Temps)

Overall rmse(MLA) is equalto rmse(Temps), except Summer

Above 400 hPa rmse(Temps) is partly greater than rmse(MLA)

Bias of LM-Wind Directionagainst

MLA Temps

12 UTC

Below 800 hPa, strong positive bias Max.: 6.1 degree, smaller as 00 UTC

Below 800 hPa, strong positive bias Max.: 10.1 degree

Below 800 hPa, strong positive bias Max.: 10.6 degree, smaller than 00 UTC

Below 900 hPa Winter negative bias Min.: -4.1 degree, Summer positive

Below 900 hPa predominantly positive Max.: 8 degrees, Winter smaller or negative

Below 900 hPa predominantly positive, Max.: 9 degrees, single months in winter positive

RMSE of LM-Wind Directionagainst

MLA Temps

12 UTC

Overall rmse(MLA) is about 5-10 degreesgreater than rmse(Temps)

Below 400 hPa rmse(MLA) is equalto rmse(Temps) except Summer

Above 400 hPa rmse(Temps) is partly greater than rmse(MLA)

Differences at Wind Direction Verification

• bias with annual cycle– below 800 hPa

• Analysis and Forecast against MLAS strong positive

• Analysis against Temps slight negative below 900 hPa, slight positve above against Temps

• Forecasts against Temps partly slight negativ(Winter) to slight positive(Summer) 12 UTC

• rmse with annual cycle – at Analysis overall 5 -10 degrees higher against MLA (wind is

nudged to Temps)– at Forecasts higher rmse against Temps above 400 hPa– at Forecasts higher rmse against MLA in Summer

Summary

• Temperature Verification shows only few differences between the two measuring systems

• Wind Verification shows below 800 hPa great differences between the two measuring systems

• Regarding the Wind Verification results up to now (Synop, Temp) LM is more likely to underestimate wind velocity

• Regarding the verification against aircraft measurements it seems that aircraft measurements below 800 hPa underestimate wind velocity (up to 2 m/s in the boundary layer).