SOLAR RADIATION VALIDATION Jean-Louis Roujean CNRM / Météo-France.

SOLAR RADIATIONSOLAR RADIATION

VALIDATIONVALIDATION

Jean-Louis RoujeanJean-Louis Roujean

CNRM / Météo-FranceCNRM / Météo-France

2

DSSF daily cycle DSSF daily cycle

3

DSSF Validation StrategyDSSF Validation Strategy

• Validation with in-situ data: Carpentras, Roissy

• Other ground stations: Data access to be coordinated with other SAFs

• Satellite product inter-comparison (SAFs,…)

• Comparison with model output (ECMWF)

4

Validation with in situ measurementsValidation with in situ measurements

5

DAILY TIME SERIES EXAMPLES

Clear Sky

Cloudy Sky

Clear Sky

Cloudy Sky

AOT440nmX10-3

6

7

Comparisons with in-situ measurements(period 2004-2006)

Stat. 2004 à 2006Stat. 2004 à 2006

Clear sky : bias : 5 W/m2 stdev : 40 W/m2

Cloudy sky :Bias : 5 W/m2 stdev : 115 W/m2

8

VALIDATION TIME SERIES: ALL SITES

bias=5W/m2 stdev=40W/m2

6 Stations in Europe (2004-2006)

ab

so

lute

4 Stations in Africa (2006)

bias=45W/m2 stdev=75W/m2a

bs

olu

te

bias=-2W/m2 stdev=125W/m2a

bs

olu

te

bias=5W/m2 stdev=115W/m2a

bs

olu

te

Overestimation of the radiation in clear sky situations (potentially due to aerosol effect)

9

Validation of DSSF over France with RADOME network (8 km)Validation of DSSF over France with RADOME network (8 km)

15/01/2006 12 h 15/06/2006 12 h 15/10/2006 12 h

10

Yearly profile (2006) of DSSF for 2 contrasted sitesYearly profile (2006) of DSSF for 2 contrasted sites

Carpentras (Provence) & Roissy (Paris airport)Carpentras (Provence) & Roissy (Paris airport)

ECMWF: -> under-estimate (clear) & over-estimate (cloudy)

DSSF Validation : NIAMEY DSSF Validation : NIAMEY (from RADAGAST project, ARM mobile station)(from RADAGAST project, ARM mobile station)

March, 2March, 2 May, 12May, 12 July, 1July, 1

DSSF Validation : AGOUFOU DSSF Validation : AGOUFOU (CESBIO data)(CESBIO data)

January, 13January, 13 March, 2March, 2 May, 12May, 12

Comparison of DSSF SAF-LAND / ECMWF – August 2005Comparison of DSSF SAF-LAND / ECMWF – August 2005

-> Tracking > Tracking of convective zones of convective zones

-> Less radiation-> Less radiation

14

Comparison between daily cumulative DSSF (DIDSSF) (left) and ECMWF forecast product (SSRD) over France at different dates during June 2009.

Daily DSSF (DIDSSF)

15

Comparison between daily cumulative DSSF (DIDSSF) (left) and ECMWF forecast product (SSRD) over France at different dates during June 2009.

16

COMPARISON WITH MODIS ALBEDO

[0.4µm,0.7µm]

[0.7µm,4µm]

[0.3µm,4µm]

over

Eur

ope

(fro

m 2

005-

06-0

1 to

200

6-10

- 01)

16-day average MSG Albedo –

MODIS albedo projected on SEVIRI grid

10-25 juin 2006

17

Comparison MSG with MODIS

MSGMSG MODISMODIS

18

June 10, 2005

MSG – VIS-DH SAF Land – Quality Flag

MSG – MODIS : absolute difference MSG – MODIS : relative difference

VISIBLE Broad-Band

19

June 10, 2005

MSG – NIR-DH SAF Land – Quality Flag

MSG – MODIS : absolute difference MSG – MODIS : relative difference

NIR Broad-Band

20

Statistical results (bias/stdev) betweenbroadband albedos MSG et MODIS over Europe

Bias is negligible exceptVIS = 0.015 (due to aerosols ?)

Standard deviation :0.015 for VIS0.030 for NIR

21

Albedo Time Series: EvoraAlbedo Time Series: Evora

0.6µm 0.8µm 1.6µm

[0.4µm,0.7µm] [0.3µm,4µm] [0.7µm,4µm]

22

0.6µm 0.8µm 1.6µm

[0.4µm,0.7µm] [0.3µm,4µm] [0.7µm,4µm]

Albedo Time Series: BarraxAlbedo Time Series: Barrax

23

12-27 juillet 2006

COMPARISON WITH MODIS ALBEDO

24

July 19, 2006

MSG – VIS-DH SAF Land – Quality Flag

MSG – MODIS : absolute difference MSG – MODIS : relative difference

VISIBLE Broad-Band

Bias = 0,047 ; RMSE = 0,026

25

X10 for AOT

SAF-Land albedo

Observed albedo

Rainfall

The temporal evolution of the albedo estimate is related to the evolution of surface properties (typically rainfall). The spurious fluctuations appear to be caused by aerosol effects.

Modis SW-DH albedo

ALBEDO TIME SERIES (Agoufou; Mali)

26

The temporal evolution of the albedo estimate is related to the evolution of surface properties (typically rainfall). The spurious fluctuations appear to be caused by aerosol effects.

ALBEDO TIME SERIES (Banizoumbou; Niger)

27

The temporal evolution of the albedo estimate is related to the evolution of surface properties (typically rainfall). The spurious fluctuations appear to be caused by aerosol effects.

ALBEDO TIME SERIES (Niamey; Niger)