Abstract
The Multi-functional Transport Satellite (MTSAT)-1R is a geostationary imager located at 140°E over the tropical western pacific operated by the Japanese Meteorological Agency and was launched on February 26, 2005. It has been operational from June 2005 to June 2010 and now serves as the backup instrument when the MTSAT-2 ground segment is maintained annually during November and December. The Clouds and the Earth's Radiant Energy System (CERES) project utilizes geostationary (GEO) derived broadband fluxes to infer the regional diurnal cycle between CERES observed broadband observations. The GEO visible channels are first calibrated against Aqua-MODIS using ray-matched coincident 0.5° gridded radiance pairs, which are regressed monthly to determine the GEO calibration coefficients. Unlike other GEOs the MTSAT-1R visible channel exhibited a nonlinear sensor response. In order to ensure that the ray-matching algorithm is not introducing any systematic biases, the navigation, the MTSAT-1R and MODIS spectral band differences, and MTSAT-1R space offset are carefully examined and were determined not to be the cause. Also Terra-MODIS or TRMM-VIRS and MTSAT-1R ray-matched radiance pairs also showed similar nonlinear behavior. However, VIRS onboard the TRMM precessing satellite, revealed that the nonlinear behavior was dependent on solar zenith angle or dynamic range. No further progress could be achieved until coincident MTSAT-2 and MTSAT-1R images taken in December 2010 became available. Comparing the coincident imagery the image blurring effect was noticeable. Dark regions neighboring bright clouds within 500-km were brightened. This effect was not noticed in the IR imagery, even though both visible and IR optical paths are shared. The slight blurring effect is attributed to the mirror surface by either flawed polishing or by a dust contaminant. The dispersed light of the signal was assumed to be small and randomly distributed around the optical axis allowing the image to be deconvolved using an inverted point spread function. The PSF removed ~80% of the blurring effect and the MTSAT-1R sensor observed a linear response when ray-matched with Aqua-MODIS radiance pairs. This chain of events, emphasizes the need for coincident imagery when replacing operational GEO imagers, as part of the commissioning process. If MTSAT-2, which was launched in February 2006, provided coincident imagery with MTSAT-1R within a few months after launch, retrieval quality imagery could have been available 5-years earlier.
The Diagnosis, Derivation and Validation of a Point Spread Function to Mitigate the Slight Blurring Manifested in the MTSAT-1R Visible Channel Imagery
The Multi-functional Transport Satellite (MTSAT)-1R is a geostationary imager located at 140°E over the tropical western pacific operated by the Japanese Meteorological Agency and was launched on February 26, 2005. It has been operational from June 2005 to June 2010 and now serves as the backup instrument when the MTSAT-2 ground segment is maintained annually during November and December. The Clouds and the Earth's Radiant Energy System (CERES) project utilizes geostationary (GEO) derived broadband fluxes to infer the regional diurnal cycle between CERES observed broadband observations. The GEO visible channels are first calibrated against Aqua-MODIS using ray-matched coincident 0.5° gridded radiance pairs, which are regressed monthly to determine the GEO calibration coefficients. Unlike other GEOs the MTSAT-1R visible channel exhibited a nonlinear sensor response. In order to ensure that the ray-matching algorithm is not introducing any systematic biases, the navigation, the MTSAT-1R and MODIS spectral band differences, and MTSAT-1R space offset are carefully examined and were determined not to be the cause. Also Terra-MODIS or TRMM-VIRS and MTSAT-1R ray-matched radiance pairs also showed similar nonlinear behavior. However, VIRS onboard the TRMM precessing satellite, revealed that the nonlinear behavior was dependent on solar zenith angle or dynamic range. No further progress could be achieved until coincident MTSAT-2 and MTSAT-1R images taken in December 2010 became available. Comparing the coincident imagery the image blurring effect was noticeable. Dark regions neighboring bright clouds within 500-km were brightened. This effect was not noticed in the IR imagery, even though both visible and IR optical paths are shared. The slight blurring effect is attributed to the mirror surface by either flawed polishing or by a dust contaminant. The dispersed light of the signal was assumed to be small and randomly distributed around the optical axis allowing the image to be deconvolved using an inverted point spread function. The PSF removed ~80% of the blurring effect and the MTSAT-1R sensor observed a linear response when ray-matched with Aqua-MODIS radiance pairs. This chain of events, emphasizes the need for coincident imagery when replacing operational GEO imagers, as part of the commissioning process. If MTSAT-2, which was launched in February 2006, provided coincident imagery with MTSAT-1R within a few months after launch, retrieval quality imagery could have been available 5-years earlier.