- 20th International TOVS Study Conference, 28 Oct.-3 Nov., Lake Geneva, WI, USA -
High resolution IR cloudy radiances
simulations: comparison between
RTTOV-11.3 and VLIDORT
Jérôme Vidot (CMS)
Laurent C. Labonnote (LOA)
Objective and Method
Objective:
Evaluate the scattering approximation of RTTOV in IR for ice cloud
2 RTMs:
RTTOV v11.3 : « Chou scaling » approximation, Baran ice clouds model
(Vidot et al., JGR 2015)
VLIDORT (Spurr, JQSRT 2006) : Full scattering model, 16 streams
Configuration:
VLIDORT inputs: RTTOV transmittances, surface emissivities, cloud optical
properties
Lambertian surface, pseudo-spherical geometry and cloud fraction = 1
Two datasets :
RTTOV 83 training profiles @101L (for sanity check)
NWP SAF profiles dataset @137L (for the study)
2
Sanity check
RTTOV 83 training profiles @ 101L, surface emissivity = 1, nadir view
MEAN BIAS
VLIDORT − RTTOV [K]
0.020
0.015
0.010
0.005
0.000
−0.005
500
1000
1500
2000
−1
WAVENUMBER [cm ]
2500
3000
2500
3000
RMS
VLIDORT − RTTOV [K]
0.020
0.015
0.010
0.005
0.000
500
1000
1500
2000
−1
WAVENUMBER [cm ]
Mean Bias and RMS < 0.02 K
3
NWP SAF 137L profiles dataset
Use of cloud condensate sampling (5000 profiles): Land/sea mask, T, H, O3,
IWC, month
4
CLEAR-SKY over sea = 92 profiles
CLEAR-SKY over land = 251 profiles
CLOUDY-SKY over sea = 1467 profiles
CLOUDY-SKY over land = 456 profiles
Clear-sky cases over sea (92 profiles)
Surface emissivities from ISEM
VLIDORT − RTTOV [K]
MEAN BIAS
0.04
0.02
0.00
−0.02
−0.04
500
1000
1500
2000
−1
WAVENUMBER [cm ]
2500
3000
NADIR
o
VZA=30
o
VZA=60
VLIDORT − RTTOV [K]
RMS
0.05
0.04
0.03
0.02
0.01
0.00
500
5
1000
1500
2000
−1
WAVENUMBER [cm ]
2500
3000
Mean Bias and RMS < 0.05 K and error increase with VZA
Clear-sky cases over land (251 profiles)
Surface emissivities from UWIREMIS atlas
VLIDORT − RTTOV [K]
MEAN BIAS
0.04
0.02
0.00
−0.02
−0.04
−0.06
−0.08
500
1000
1500
2000
−1
WAVENUMBER [cm ]
2500
3000
NADIR
o
VZA=30
o
VZA=60
VLIDORT − RTTOV [K]
RMS
0.10
0.08
0.06
0.04
0.02
0.00
500
6
1000
1500
2000
−1
WAVENUMBER [cm ]
2500
3000
Mean Bias and RMS < 0.1 K and error decrease with VZA
Cloudy-sky cases over sea (1467 profiles)
Surface emissivities from ISEM
VLIDORT − RTTOV [K]
MEAN BIAS
0.1
0.0
−0.1
−0.2
−0.3
−0.4
−0.5
500
1000
1500
2000
−1
WAVENUMBER [cm ]
2500
3000
NADIR
o
VZA=30
o
VZA=60
VLIDORT − RTTOV [K]
RMS
0.6
0.5
0.4
0.3
0.2
0.1
0.0
500
7
1000
1500
2000
−1
WAVENUMBER [cm ]
2500
3000
Mean Bias and RMS < 0.6 K and error increase for large VZA
Cloudy-sky cases over land (456 profiles)
Surface emissivities from UWIREMIS atlas
VLIDORT − RTTOV [K]
MEAN BIAS
0.2
0.1
0.0
−0.1
−0.2
−0.3
−0.4
−0.5
500
1000
1500
2000
−1
WAVENUMBER [cm ]
2500
3000
NADIR
o
VZA=30
o
VZA=60
VLIDORT − RTTOV [K]
RMS
0.8
0.6
0.4
0.2
0.0
500
8
1000
1500
2000
−1
WAVENUMBER [cm ]
2500
3000
Mean Bias and RMS < 0.8 K and error increase for large VZA
Error vs cloud optical depth
Based on cloud channel selection (Martinet et al., QJRMS 2014)
CLOUDY SKY LAND MEAN BIAS
0.5
0.0
−0.5
−1.0 −6
10
1.5
VLIDORT − RTTOV [K]
VLIDORT − RTTOV [K]
CLOUDY SKY SEA MEAN BIAS
1.0
10
−4
−2
10
CLOUD OPTICAL DEPTH [−]
0
10
10
1.0
0.5
0.0
−0.5
−1.0 −6
10
2
10
−4
−2
10
CLOUD OPTICAL DEPTH [−]
0
10
NADIR
o
VZA=30
VZA=30
o
o
VZA=60
VZA=60
RMS
1.4
VLIDORT − RTTOV [K]
VLIDORT − RTTOV [K]
RMS
1.2
1.0
0.8
0.6
0.4
0.2
10
−4
−2
10
CLOUD OPTICAL DEPTH [−]
0
10
10
2
2.0
1.5
1.0
0.5
0.0 −6
10
10
−4
−2
10
CLOUD OPTICAL DEPTH [−]
0
10
Even for large COD, error of cloud channel selection is below 0.5 K
9
2
NADIR
o
0.0 −6
10
10
10
2
Conclusions & perspectives
Conclusions:
The error of the « Chou scaling » scattering approximation of RTTOV for ice
cloud simulations is < 0.8 K for all channels and is < 0.5 K for Martinet’s
cloud channel selection (overestimation from RTTOV)
Error is ~2 times lower for low viewing zenith angles.
Results are consistent in absolute values with Matricardi (2005)
VLIDORT is well suited for IR RTTOV validation (w-w/o scattering)
Perspectives:
Apply the method for liquid cloud (easy with NWP SAF profiles dataset)
Do the comparison for Jacobians (feasible with VLIDORT)
Simulations vs observations (complex for good cloudy profile collocations)
Comparison vs other RTMs : CRTM, 4A/OP, Sigma-IASI, kCARTA,… (hard
to organize – RTSP Working Group)
10
Thank you for your attention
11