Toshiaki Izumiya

The estimation of aerosol optical parameters from ADEOS/POLDER data

A cross calibration analysis between POLDER and OCTS sensors on board ADEOS satellite in terms of the space reflectance was made. Space reflectance values, computed from the POLDER and OCTS data (acquired simultaneously on 26 April 1997) for the same ocean target, were examined and we found that they are in good agreement in all bands, when the new in-flight calibration gain factors for each sensor are assumed. Furthermore, we retrieved aerosol optical parameters (i.e., the optical thickness, Angstrom exponent, and refractive index) by using several retrieving algorithms from using ADEOS/POLDERs directional space reflectance and polarization data in infrared bands simultaneously. Retrieved results were presented by the form of distribution maps. In addition, good and bad points of each retrieval algorithm were discussed. Finally, we proposed an improved method for retrieving aerosol optical parameters, which utilizes theoretical P-SR (polarization-space reflectance) diagrams at given viewing angles in an 865 nm band. The estimated aerosol optical parameters by our proposed method were compared with the simultaneously measured sky observation data. We found an excellent agreement between them.

New estimation algorithm of aerosol optical thickness from space remote sensing data and its validation

Abstract In the atmospheric correction for space remote sensing images, the aerosol optical thickness at the satellite observation time is critically needed. In this study, we proposed a new algorithm for estimating the aerosol optical thickness from satellite image data alone. In this new algorithm, the meteorological range, V, was used as a free parameter to compute the surface albedo distribution from satellite image data. The aerosol optical thickness can be easily computed from the meteorological range. The analytical approximation method for atmospheric correction and Modtran code were also utilized in our algorithm. As for the algorithm validation, the simultaneous ground and sky measurement experiments were conducted in 1996 and 1997 with ADEOS/AVNIR and LANDSAT/TM, respectively. We found that the estimation error by the proposed algorithm is less than 0.04 in terms of aerosol optical thickness. The surface albedo distribution image computed from ADEOS/AVNIR data (taken on April 24, 1997) was also presented, by using the estimated meteorological range.

Retrieval algorithm for aerosol optical parameters from POLDER's directional reflectance and polarization data

We examined three different methods, namely, Method-I, -II, and -III for retrieving local aerosols optical parameters over the Japan Sea using the ADEOS/POLDER data by comparing with the validation data. Method-I, and Method-II use parameterized directional reflectance and polarization diagrams in two infrared bands, respectively. On the other hand, Method-III uses parameterized directional polarization- reflectance diagrams in a single infrared band. We found that Method-III gives the best agreement with the measured sky validation data. We also presented retrieved distribution maps for the aerosol optical thickness, and Angstrom exponent by Method-III and some discussions on further improvement for Method-III were given.

Reflectance analysis of POLDER data in Mongolia

Abstract POLDER data obtained over land on June 13, 1997 is analyzed and reported in this paper. Aerosol optical parameters for desert area in Mongolia are estimated both from reflectance and polarization data at 443 nm. The POLDER image data is corrected into surface image data using the atmospheric correction scheme based on aerosol optical parameters. The BRDF parameters for land surfaces are estimated for the study region in Mongolia from the surface reflectance images at 443, 670 and 865 nm wave length.

Aerosol optical parameter retrieval from POLDER data

We proposed a new method for retrieving local aerosol optical parameters, such as the real part of refractive index, optical thickness, and /spl Aring/ngstro/spl uml/m exponent, using the ADEOS/POLDER data. The retrieved results agreed well with the simultaneous sky observation data. The real part of the refractive index of aerosol particles was retrieved.

Distribution maps of aerosol's three optical parameters from POLDER's multi-directional data

It is possible to retrieve three aerosol optical parameters, namely, its optical thickness /spl tau//sub n/(500) at the wavelength of 500 nm, /spl Aring/ngstro/spl uml/m exponent a, and real part of refractive index Nr, from ADEOS/POLDER data sets by analyzing their directional reflectance and polarization information. In this retrieval we applied the directional R-P (Reflectance-Polarization) algorithm to the POLDER 865-nm band data and the average distribution maps of the three aerosol optical parameters were presented over the oceans from the Indian Ocean to the western Pacific Ocean during April 1-30, 1997.

Comparison of aerosol parameter retrieval results from the analysis of ADEOS/POLDER and ADEOS/OCTS data

Abstract In this study we presented a new multi-directional R-P(Reflectance-Polarization) algorithm which retrieves aerosols three optical parameters, namely, the real part of refractive index, optical thickness and Angstrom exponent, by using ADEOS/ POLDERs single near-infrared band data. We validated our retrieved results, by comparing with the simultaneously measured sky observation data at the target site. We found a good agreement between them. Furthermore, the retrieval results by the proposed algorithm were compared with those by a multi-directional R-R(Reflectance-Reflectance) algorithm using the reflectance data of POLDER in a single infrared band and by a mono-directional R-R algorithm using the reflectance data of ADEOS/OCTS in two infrared bands. Although we had a good agreement in the aerosol optical thickness among these algorithms, we found that the retrieval results of Angstrom exponent are not reliable in the cases of the latter two algorithms using the reflectance data alone.

Estimation of the BRDF of land surfaces from ADEOS/POLDER image data

The authors have analyzed the POLDER data over lands obtained in June 13, 1997. The atmospheric effects are removed from observed reflectance images by using their atmospheric correction scheme. In the atmospheric correction, their needed a realistic atmospheric model. The POLDERs unique sensor has an ability to measure both the directional reflectance and polarization in the visible and near-infrared bands reflected by the Earth-atmosphere system. They estimated aerosol optical parameters from both the reflectance and polarization data at 443 nm for the study area. They converted the POLDER image data into the surface reflectance image data using the atmospheric correction scheme based on retrieval aerosol optical parameters. Then they got estimates of the Roujeans BRDF parameters for land surfaces in Mongolia from their surface reflectance images at 443, 670, and 865 nm.

Analysis of ADEOS/POLDER data over land surfaces

Abstract In this paper, we made multiple scattering simulations by assuming a single atmospheric layer model with ground surface reflection models. This paper shows the estimated results of surface reflectance A using the observed reflectance and polarization data at 443 nm, 670 nm and 865 nm over land surfaces measured by the ADEOS/POLDER. POLDER level-1 products over the Sahara Desert in West Africa taken on November 18, 1996, were investigated. The main results of this study can be summarized as follows: 1) We found that the theoretical model can satisfy both the observed reflectance and the observed linear polarization data against different observed zenith viewing angles. 2) An assumption of Lambertian reflection for “Desert” seems to be valid. We use the Junge model with ν = 3 and refractive index m=1.55−0.005 as typical desert aerosol in this computation.