GRASP sensitivity ISTINA WP WP B. Torres and the rest of TESTS OBSERVATION CONDITIONS RETRIEVAL ASSUMPTIONS AERO.

Transcription

1 TESTS RETRIEVAL ASSUMPTIONS GRASP sensitivity ISTINA OBSERVATION CONDITIONS WP David Fuertes CTO Benjamín Torres Président GRASP Pavel Lyvynov CSO GRASP Oleg Dubovik Consultant WP AERO. MODELS B. Torres and the rest of

2 Summary 1.- Reminder of WP (ISTINA) 2.- Introduc>on to WP (ISTINA- ext) 3.- Applica>ons A.- Sun- photometer retrievals B.- Satellite retrievals C.- Other examples

3 Reminder of WP ISTINA - WP3380-2: Assessments of SensiIvity Tendencies in Aerosol Remote Sensing Experimental data. To establish fundamental limits for the different types of aerosol remote sensing approaches. To study the variability of retrieval accuracy limits for observaaons with different set ups. Key quesions: Best approximaaon for modelling size distribuaon. Can the refracave index of fine and coarse modes be disanguished and using which observaaons. How inevitable assumpaons in forward model limit the retrievals from highly accurate observaaons.

4 Reminder of WP ISTINA - WP3380-2: Focus on single scasering measurements: Aerosol remote sensing algorithms rely on the manifestaaon of angular and spectral features in aerosol scajering which are mostly determined by aerosol single scasering properies such as phase matrix, exancaon and absorpaon. We expect to clarify some fundamental tendencies and limitaions in informaaon content, and benefit from them in the follow- on more complex studies (WP3380-4) for outlining and establishing clear limits in the retrieval capaciaes.... having in mind the limita>ons of real ground- based and satellites observa>ons.

5 Reminder of WP ISTINA - WP3380-2: Volc anic

6 Objectives for WP (ISTINA-ext) Defining pracacal recommendaaons for improving accuracy of aerosol characterizaion in real remote sensing applicaaons (using the inputs from ISTINA - WP3380-2). SensiAvity analysis on surface reflectance modelling and defining recommendaaons for improving accuracy surface reflectance characterizaaon. Cartographie

7 ISTINA-Ext A.-AOD extinction GRASP- AOD inversion INPUT Ø 8 AOD measurements between nm Ø We assume opacal properaes (Refr. Ind. 16 parameters) & 1 sphericity. OUTPUT ² Bi- lognormal approximaaon (6 parameters): R Vf, s Vf, C Vf +R Vc, σ Vc, C Vc ² Derived parameters: τ f (500), R eff, C T, Advanced characteriza/on of aerosol proper/es from measurements of spectral op/cal thickness of the atmosphere. (Torres et al. 2016) Public available: hjp:// meas- tech- discuss.net/ amt /

8 ISTINA-Ext A.-AOD extinction a) SensiIvity tests 1) Self- consistency analysis 2) IniIal guess variaion 3) SimulaIons of AOD errors 4) SensiIvity to the refracive index dv/dlnr (µ m 3 / µ m 2 ) GSFC1 GSFC2 GSFC3 MEXI1 MEXI2 MEXI3 ZAMB1 ZAMB2 ZAMB3 dv/dlnr (µ m 3 / µ m 2 ) GSFC2 Reference Tests for different i.c. dv/dlnr (µ m 3 / µ m 2 ) BAHR1 Reference Non error All channels +ε(λ) All channels ε(λ) All error combinations dv/dlnr (µ m 3 / µ m 2 ) GSFC Reference n =n n =n+σ n n =n+2σ n n =n σ n n =n 2σ n Radius (µ m) Radius (µ m) Radius (µ m) Radius (µ m) b) ValidaIon and applicaions 1) ValidaIon AERONET (+650 inv.) 2) Moon Photometry 3) ApplicaIon to Airborne PLASMA 0.03 r vf daily mean bias (µ m) CASE A CASE B CASE C CASE D dv/dlnr (µ m 3 / µ m 2 ) AOD 08/04 16:11 AOD 08/04 17:03 AOD 09/04 06:48 ALM 09/04 06:48 AOD MOON dv/dlnr (µ m 3 / µ m 2 ) ALM Photo. #349 AOD Photo. #349 AOD PLASMA (500m) AOD PLASMA (1000m) AOD PLASMA (1500m) AOD PLASMA (2000m) AOD PLASMA (2500m) AOD PLASMA (3000m) AOD PLASMA (3500m) AOD PLASMA (4000m) 0 GSFC MEXI ZAMB SOLV BAHR LANA Radius (µ m) Radius (µ m)

9 ISTINA-Ext A.-AOD extinction Conclusions: The use of only spectral AOD measurements allows the discriminaion between the exincion of fine and coarse modes of aerosol (i.e. the characterizaion of τ f (500nm), is not dependent on any assumpion. Spectral AOD measurements are sufficient for a reliable characterizaion of aerosol fine mode opical properies. However, the retrieval accuracy of parameters such as R Vf, σ Vf and C Vf depends on available informaaon about the real part of refracave index and the quality of the AOD measurements. The characterizaaon of the coarse mode using only AOD measurements is less accurate, but can be significantly improved using moderate a priori assumpaons.

10 ISTINA-Ext A.-AOD extinction Advanced characteriza/on of aerosol proper/es from measurements of spectral op/cal thickness of the atmosphere. (Torres et al. 2016) Public available: hjp:// meas- tech- discuss.net/ amt / The main objectives of the paper are well described and discussed. GRASP is becoming a very powerful tool for aerosol characterization from remote sensing measurements and the inclusion of this new capability is of great interest for the scientific community. Anonymous Referee 1. The paper is extremely thorough and well written. I recommend that this paper be accepted for publication with only minor editorial changes. M. D. King (NASA) The approach is interesting and deserves to be explored for its potential for the retrieval of particle size distribution. The paper is well written and easy to read. Z. Li (Chinese Academy of Sciences) The paper is clearly organized and very well written. This new algorithm provides inversions that may be extremely useful in analysis of particle size related aerosol parameters, since these retrievals can be made much more frequently from AOD spectra alone than from the combined input of AOD spectra and sky radiance distributions, such as in almucantar inversions in AERONET. T. Eck (AERONET- NASA)

11 ISTINA-Ext A.-AOD extinction if the algorithm is to be operaaonal T. Eck (AERONET- NASA) Main tasks from the revision: - SensiAvity study - To include a much larger uncertainty/variaaon in the iniaal guess of the real refracave index for the fine mode dominaaon cases. To include some cases with very low (AOD[440 nm]=0.1) and very high (AOD[440 nm]=1.5). Considering the spectral dependece of AOD errors in the analysis. Analysis of the use of spherical and non- spherical model for desert dust cases. Analysis of the effects of reducing the spectral range to the 4 universal wavelengths of AERONET instruments (440 to 1020 nm) New task from this workshop: Redo the analysis using the standard spectral range of GAW- PFR network (368 to 862nm) - ValidaAon and applicaaons - To include 2 extra sites in the real data analysis

12 ISTINA-Ext B.- Quality assessments for satellite retrievals ObjecAve: To stablish quality levels for the GRASP retrievals using satellite data. Our strategy: To find/define the criteria for the different parameters/products. To evaluate their efficacy through comparisons with AERONET retrievals. Contrôle et prévision de la pollution Améliorer l agriculture Prévisions météorologiques Contrôle des ressources Cartographie Science

13 ISTINA-Ext B.- Quality assessments for satellite retrievals Criteria: 1) Geometry Applica>on to PARASOL/POLDER Θ< <Θ< <Θ< <Θ< <Θ< 180 At least 2 measurements in these sectors 2a) Residual filter Contrôle et prévision de la pollution Améliorer l agriculture Fit < 7% Prévisions météorologiques Contrôle des ressources Cartographie Aerosol optical depth 1 Science 2b) AOD filter AOD (440) > 0.4 AOD (870) > 0.2 Alpha Angstrom 2 AOD (440) > 0.4 Optical properties 3

14 ISTINA-Ext B.- Quality assessments for satellite retrievals 1 AOD (440nm) Retrievals Residual filter PARASOL vs AERONET Contrôle et prévision de la pollution (2008) Angular filter Prévisions météorologiques Angular filter + Residual filter Contrôle des ressources Cartographie Angular filter Science Residual filter

15 ISTINA-Ext B.- Quality assessments for satellite retrievals 2 α ( nm) Retrievals AOD filter PARASOL vs AERONET Contrôle et prévision de la pollution (2008) Angular filter Prévisions météorologiques Angular filter + AOD filter Contrôle des ressources Cartographie Angular filter Science AOD filter

16 ISTINA-Ext B.- Quality assessments for satellite retrievals 3 SSA (670nm) Retrievals AOD filter PARASOL vs AERONET Contrôle et prévision de la pollution (2008) Angular filter Prévisions météorologiques Angular filter + AOD filter Contrôle des ressources Cartographie Angular filter Science AOD filter

17 ISTINA-Ext B.- Satellite data: Analysis on surface models. Excellent correspondence of the models! Range covered by GRASP multi-pixel (2-3 months) Unreachable for PARASOL Rarely observed Typical PARASOL range for one day: Science

18 ISTINA-Ext C.- Other examples Example: Retrieval of moon photometry + lunar aureole Téléchargeable: downloads Remote sensing of lunar aureole with a sky camera: Adding informa/on in the nocturnal retrieval of aerosol proper/es with GRASP code R. Román, B. Torres, D. Fuertes et al. (Remote Sensing Env.) Normalized radiances from sky camera measurements Aerosol properies during the night Contrôle et prévision de la pollution Améliorer l agriculture Prévisions météorologiques Cartographie Science

19 Merci! Gracias! Danke! Thanks! спасибо! xiè xiè! Grazie!