Download - Areal rainfall statistics based on radar observations - ESSL

Areal rainfall statisticsbased on radar observations

Edouard Goudenhoofdt and Laurent DelobbeRoyal Meteorological Institute of Belgium (RMIB)

June 3, 2013, European Conference on Severe Storms

RMIB operates a C-band radar since 2001

I Single-polarisation

I Doppler filtering (clutter)

I Located 600 m asl,range of 240 km

I 5-elevation every 5 min(during 2 min)

I Resolution : 1° in azimuth,250 m in range

Quantitative precipitation estimates��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��I PCAPPI 800 m above

radar level

I Z = 200R1.6

I Hail:Z > 53 dBZ → 53 dBZ(75 mm/h)

I Cartesian grid 500 mresolution.

I Accumulation by linearinterpolation.

To be validated : clutter mitigation and profilecorrection.

Merging and verification with denseraingauge networks.

50 km

100 km

Radar and rain gauge networks

SPW (69)RMI(84)Radar

I hourly automaticraingauge network (blue)

I 1E 6 scale difference !I mean field bias : simple

and robustI max range 120 kmI min value 0.1 mmI min 10 valid pairs

A 8-year verification reveals relatively goodaccuracy.

Merging and verification with denseraingauge networks.

� � � � � �

��

�

��

��

��

��

��

��

I hourly automaticraingauge network (blue)

I 1E 6 scale difference !I mean field bias : simple

and robustI max range 120 kmI min value 0.1 mmI min 10 valid pairs

A 8-year verification reveals relatively goodaccuracy.

Mean hourly rainfall depth 2005-2012

150 100 50 0 50 100 150

150

100

50

0

50

100

150

0.06

0.07

0.08

0.09

0.10

0.11

0.12

0.13

0.14

['A

CR

R']

['bewid - 2012-01-01 00:00:00 - STAT(scan1_cap_aclPT1H_mfb_stmeanP8Y)']

I unconditional mean(dry periods are included)

I minimum 0.07 mm(600 mm/year) in theplains

I maximum 0.14 mm(1200 mm/year) in thehills

I clear correlation withtopography (cluttereffect?)

Those results are consistent with raingaugeclimatology.


180 km

0 200 400 600 800

Elevation [m]







150 100 50 0 50 100 150

150

100

50

0

50

100

150

0.06

0.07

0.08

0.09

0.10

0.11

0.12

0.13

0.14

['A

CR

R']

['bewid - 2012-01-01 00:00:00 - STAT(scan1_cap_aclPT1H_mfb_stmeanP8Y)']






Max hourly rainfall depth 2005-2012

200 100 0 100 200

200

100

0

100

200

10

15

20

25

30

35

40

45

50

55

60

['A

CR

R']

['bewid - 2013-01-01 00:00:00 - STAT(scan1_cap_aclPT1H_mfb_stmaxP8Y)']

I High small scalevariations.

I No significant large scaletrend.

I Slightly more max inSouth-East.

Highest values are due to stationary cells and/orhail.

Probability of hourly rainfall (1 mm).

150 100 50 0 50 100 150

150

100

50

0

50

100

150

1.2

1.6

2.0

2.4

2.8

3.2

3.6

4.0

%

['bewid - 2013-01-01 00:00:00 - STAT(scan1_cap_aclPT1H_mfb_spe1P8Y)']

I ranges from 2 % to4 %

I positive effect oftopography

Highly correlated with mean hourly rainfall

Probability of rainfall exceeding 10 mm.

200 100 0 100 200

200

100

0

100

200

0.016

0.024

0.032

0.040

0.048

0.056

0.064

0.072

%

['bewid - 2013-01-01 00:00:00 - STAT(scan1_cap_aclPT1H_mfb_spe10P8Y)']

I ranges from 0.02 %to 0.06 %

I less effect oftopography

I higher probabilitySouth-East of radar

Partially correlated with max hourly rainfall

Exceedance probability of four differentriver catchment.

��

��

��

�

��

��

��

��

��

��

��

I catchment ofdifferent size

I smooth andlogarithmicbehavior

I smallest catchment: max 20 mm

I largest catchment :max 8 mm

Computation of return periods is limited (8 years).

Exceedance probability of four differentriver catchment.

��

��

��

��

��

��

��

��

��

��

��

��

��

��

I catchment ofdifferent size

I smooth andlogarithmicbehavior

I smallest catchment: max 20 mm

I largest catchment :max 8 mm

Computation of return periods is limited (8 years).

Exceedance probability of adjacentequal-area squares.

0.1 0.2 0.5 1 2 5 10 20 50Mean rainfall depth [mm]

10-8

10-7

10-6

10-5

10-4

10-3

10-2

10-1

100

Pro

babili

ty

Area = 25 km2

I simpleapproximation of acatchment

I distance less than100 km for bestaccuracy

I space and timestationarity

I independencebetween windows?

Possibility to compute longer return periods?(theoretically 8 years x number of windows)



10-8

10-7

10-6

10-5

10-4

10-3

10-2

10-1

100

Pro

babili

ty

Area = 100 km2








10-7

10-6

10-5

10-4

10-3

10-2

10-1

100

Pro

babili

ty

Area = 400 km2






Conclusions

I Weather radar provide good areal rainfall estimates.

I Areal rainfall exceedance probability can be computed.

I Important application to river catchment.

I Longer return periods could be computed using a largerdomain.

Outlook

I Best single radar QPE reanalysis (almost ready)

I Radar composite to mitigate attenuation and beambroadening

I Effect of rainfall depth duration

I Proof using a proper theoretical framework