Remote sensing from POLDERPARASOL Earth polarorbiting satellite enhanced

Remote sensing from POLDER/PARASOL Earth polarorbiting satellite: enhanced Surface/Aerosol properties retrieval P. Litvinov, O. Dubovik, T. Lapyonok, F. Ducos, D. Tanre Laboratoire d'Optique Atmosphérique, CNRS, Universite Lille-1, FRANCE

Enhanced space-borne Aerosol/Surface retrieval Multi-spectral, multi-angle I, Q, U measurements Advanced retrieval algorithm Enhanced retrieval of Aerosol/Surface properties AOD, Re(m), Im(m), Size distribution, fraction of spherical particles, surface parameters etc. Aerosol model Helsinki, August 19 -21, 2013 Surface reflection model

Reflection matrix for coupled atmospheresurface system Aerosol and surface properties should be Helsinki, simultaneously! August 19 -21, 2013 retrieved

“independent” POLDER/PARASOL measurements : VIEWS: NQ= 16: (800 ≤ Q ≤ 1800) INTENSITY (I): Ntl=6 (for aerosol): (0. 44, 0. 49, 0. 56, 0. 67, 0. 865, 1. 02 mm) Ntl=3 (for gas absorption): (0. 763, 0. 765, 0. 910 mm) POLARIZATION (Q, U): NPl=3: (0. 49, 0. 67, 0. 865 mm) SINGLE OBSERVATION: a lot !!! – as much as AERONET (Ntl+NPl)×NQ = (6+3)× 16= 144 Helsinki, August 19 -21, 2013

PARASOL daily coverage image, March 3, 2013 Swath: about 1600 km cross-track Global coverage: every 2 days Helsinki, August 19 -21, 2013 1 pixel spatial resolution: 5. 3 km × 6. 2 km

New POLDER/PARASOL algorithm (GRASP) (Dubovik et al, AMT, 2011) • The new algorithm uses complete set of PARASOL angular measurements in all spectral bands including both radiance and linear polarization measurements. • Continuous space of aerosol and surface properties is used. • The algorithm is based on statistically optimized fitting. The core of the new PARASOL algorithm is based on the same concept as AERONET aerosol retrieval (O. Dubovik and M. King, 2000; O. Dubovik, 2004; O. Dubovik et al, 2006). Helsinki, August 19 -21, 2013

The concept of the GRASP algorithm Two scenarios of retrieval (Dubovik et al. , AMT, 2011): - Conventional: single-pixel retrieval (each single pixel is inverted independently) - New concept: multiple-pixel retrieval (group of pixels are inverted simultaneously) Bern, July 15 -19, 2013

Aerosol model is the same as in AERONET retrieval (Mixing of particle shapes (Dubovik et al. , 2006)) retrieved C + (1 -C) Helsinki, August 19 -21, 2013

Surface reflection model • Semi-empirical models for surface total reflectance description: - RPV model (Rahman et al. , (1993)) - Ross-Li (Ross, (1981); Li, X. , Strahler (1992)) - Ross-Roujean model (Roujean et al. , (1992)) • Semi-empirical models for surface polarized reflectance description: - Nadal-Breon model (Nadal and Bréon, (1999)) - Maignan model (Maignan et al. , (2009)) - Fresnel facet model for Gaussian surfaces (Litvinov et al. , 2011) • Physically based models for reflection matrix for surfaces: - Cox-Munk model (for aerosol retrieval over ocean) - Physical models for land surface reflection matrix (Litvinov et al. , 2012) Helsinki, August 19 -21, 2013

Perfect retrieval of synthetic measurements!!! Bern, July 15 -19, 2013 10

Comparison with AERONET Banizoumbou Beiging Kanpur Mongu Banizoumbou: January, February, 2008 Surface: Grassland. Aerosol: Coarse mode is dominated. Mongu: August, September, 2008 Surface: Savanna. Aerosol: Fine mode is dominated. Beijing: April, December, 2008 Surface: Urban. Aerosol: Fine and Coarse modes. Kanpur: October-December, 2008 Surface: Urban. Aerosol: Fine and Coarse modes. The IGBP (International Geosphere Biosphere Programme) land type specification Bern, July 15 -19, 2013 11 was used

POLDER/AERONET. AOD. St. Dev. =0. 160 Bern, July 15 -19, 2013 St. Dev. =0. 128

POLDER/AERONET. Angstrom Exponent. St. Dev. =0. 262 Bern, July 15 -19, 2013

POLDER/AERONET. SSA. St. Dev. =0. 045 St. Dev. =0. 051 Bern, July 15 -19, 2013

DHR retrieval, Banizoumbou, 2007, 2008. Black curves and markers: DHR in 2007. Color curves and markers: DHR in 2008 There are possibilities for surface climatology and surface changes Bern, July 15 -19, 2013 15 monitoring!!!

Surface modeling effect on aerosol retrieval. AOD retrieval over Banizoumbou, Mongu. Bern, July 15 -19, 2013 16

Surface modeling effect on aerosol retrieval. SSA retrieval over Banizoumbou, Mongu 17 Physical models are required to. Bern, increase accuracy of aerosol retrieval. July 15 -19, 2013

I, Q, U retrieval vs I-retrieval: total reflectance fits Helsinki, August 19 -21, 2013

I, Q, U retrieval vs I-retrieval: polarized reflectance fits Helsinki, August 19 -21, 2013

I, Q, U retrieval vs I-retrieval: AOD (Banizoumbou) I-retrieval I, Q, U Helsinki, August 19 -21, 2013

I, Q, U retrieval vs I-retrieval: SSA (Banizoumbou) I, Q, U I-retrieval Helsinki, August 19 -21, 2013

Multi-pixel vs Single pixel retrieval: AOD (Banizoumbou) Multi-pixel Helsinki, August 19 -21, 2013 Single pixel

Multi-pixel vs Single pixel retrieval: SSA (Banizoumbou) Multi-pixel Single pixel Helsinki, August 19 -21, 2013

Regional maps (1800 x 1800 km). Banizoumbou, AOD 670 nm Strong spatial and temporal variation of AOD Helsinki, August 19 -21, 2013

Regional maps (1800 x 1800 km). Banizoumbou, SSA 670 nm Essential temporal variation of SSA Helsinki, August 19 -21, 2013

Daily variation of AOD and SSA at 670 nm. Banizoumbou (Jan. , Febr. 2008). 21. 02. 2008 23. 02. 2008 Helsinki, August 19 -21, 2013

Regional maps (1800 x 1800 km). Banizoumbou, SALB 670 nm Bern, July 15 -19, 2013 Jan. and Febr. ! Surface is very stable for

Regional maps (1800 x 1800 km). Mongu, SSA 670 nm NASA Global Fire Maps 28. 08. 2008 – 06. 09. 2008 (detected by MODIS) Small SSA correspond to biomass burning Helsinki, August 19 -21, 2013

Resume (space-borne Aerosol/Surface retrieval) Multi-spectral, Advanced retrieval multi-angle I, Q, U Algorithm - LOA (Lille, France) GRASP algorithm measurements (O. Dubovik et al, AMT, 2011). PARASOL will end up in September 2013. There may be 7 years gap in polarimetric space-borne measurements!!! - SRON (Utrecht, The Netherlands) algorithm (O. Hasekamp et al, JGR, 2012) Enhanced retrieval of Aerosol/Surface properties Aerosol model Surface reflection model At present time the mixture of spherical - Physical models are required to increase and spheroidal particles is the most accuracy of aerosol retrieval over land. comprehensive existent aerosol model - Physical models provide possibilities for Helsinki, August 19 -21, 2013 characterization (Litvinov et al. , 2012). (Dubovik et al. , 2006). surface

• Multi-spectral, multi-angle I, Q, U spaceborne measurements: - PARASOL will end up in September 2013! - 3 MI will be launched in 2020 (? ). - Ukrainian version of GISS NASA APS (Aerosol Polarimetry Sensor) instrument may be launched in 3 years (the possibility to set up Dutch SPEX instrument on the same satellite is being discussed now). There may be 7 years gap in polarimetric space-borne measurements!!! • Advanced retrieval algorithm: - LOA (Lille, France) GRASP algorithm (O. Dubovik et al, AMT, 2011) - SRON (Utrecht, The Netherlands) algorithm Helsinki, August 19 -21, 2013 (O. Hasekamp et al, JGR, 2012)

• Aerosol model: At present time the mixture of spherical and spheroidal particles is the most comprehensive existent aerosol model (Dubovik et al. , 2006). • Surface reflection model - Physical models are required to increase accuracy of aerosol retrieval over land (Litvinov et al. , 2011; Litvinov et al. , 2012). - Physical models provide possibilities for surface characterization (Litvinov et al. , 2012). Helsinki, August 19 -21, 2013