Proc. SPIE. 9259, Remote Sensing of the Atmosphere, Clouds, and Precipitation V
KEYWORDS: Clouds, Satellites, Reflectivity, Detection and tracking algorithms, Remote sensing, Solar radiation, Meteorological satellites, Visible radiation, Temperature metrology, Algorithm development
Cloud detection is a key work for the estimation of solar radiation from remote sensing. Particularly, the
detection of thin cirrus cloud and the edges of thicker cloud is critical and difficult. To obtain accurate
estimates of cloud cover of MTSAT-1R image, we propose an effective cloud detection algorithm for
improving the detection of thin cirrus cloud and the edges of thicker cloud. Using the brightness
temperature difference (BTD) and lookup table to identify cloud-free and cloud-filled pixels is not
sufficient for MTSAT-1R data on the region of China. Therefore, a new lookup table (LUT) is made by
extending the original one. On the basis of the exiting method, in order to apply to the MTSAT-1R
satellite data in China region, we expand the scope of the latitude and extend the applicable scope of
satellite zenith angle. We change the interpolation method from linear mode to nonlinear mode. The
evaluation results indicate that our proposed method is effective for the cirrus and the edges of thicker
cloud detection of MTSAT-1R in China region.
The solar radiation incidence on the horizontal plane is not the true solar radiation (also called surface solar radiation) on the earth surface, it not take the influence of the rugged terrain into account. Topographic correction process is established which necessarily take integrated consideration of the geographic factors and the local topographic factors (i.e. slope, aspect, terrain inter-shielding effect). Based on the high resolution Digital Elevation Model (DEM) data and the horizontal solar radiation as the input data of topographic correction process, using the mountain solar radiation correction model to simulate the topographic correction process and to present the spatial distribution of surface solar radiation in China Ganzi region on June 30, 2010. Because of the influence of the rugged terrain, the spatial distribution of surface solar radiation is accompanied by the strong spatial heterogeneity, and the spatial representativeness of the observed data of meteorological station is limited. By use of the variogram model to calculate the spatial representativeness and to associate the strength of spatial representativeness with the distance. The results indicated that: 1) rugged terrain mainly makes the solar radiation the redistribution effect significantly on sunny/shady slope of local region, and the increase of slope has a subduction effect on radiation. The terrain factor is essential on determining the solar radiation over the complex terrain. 2) The spatial representativeness of Ganzi meteorological station is approximately 350 meters, the strength of spatial representativeness has the negatively correlation with the distance. There is a necessary to consider the spatial representativeness when verifying the retrieved data.
Downward longwave radiation (DLR) at the earth’s surface is a major component of surface radiation budget and thus the climate, and remote sensing provides the most effective method to get surface DLR on a large scale. This paper presents a comparison of several DLR algorithms for both clear-sky and cloudy-sky conditions. These algorithms were applied to MODIS Terra data and extensively validated using one year's ground data at 13 stations around globe. For clear sky conditions, two algorithms using atmospheric parameters, two algorithms using satellite thermal radiances, and an algorithm that combined using satellite thermal data and atmospheric parameters were compared. The validation result indicated that the first type of algorithms often underestimated DLRs over high altitude regions, while the second type of algorithms performed well over these regions but had significant positive errors over arid regions. The third type of algorithm had acceptable results over all types of regions. Furthermore, the study found that using NCEP derived atmospheric parameters could effectively improve the performance of the first algorithms over high altitude regions, compared with MODIS atmospheric product. For cloudy conditions, three parametric algorithms that determined cloud radiative effect by cloud base temperature, and an empirical algorithm that employed cloud water path and cloud ice path were compared. The validation results indicated that the empirical algorithm had best results in most of the sites, while the three parametric algorithms were greatly influenced by the uncertainties of cloud parameters.
Downward shortwave radiation (DSR) receipt at the Earth’s surface is an important parameter in models of ecosystem dynamics and climate change. This paper presents a methodology to estimate DSR using hourly geostationary satellite (MTSAT-1R) and MODIS BRDF albedo parameter product (MCD43C1). The proposed algorithm retrieves atmospheric parameters directly from MTSAT-1R images by searching and interpolating look-up tables (LUT), which are created by the SBDART. The derived cloud optical thickness together with surface albedo and DEM are used to calculate the instantaneous downward shortwave radiation under cloudy sky. Hourly and daily DSR is calculated by the diurnal cycle integration of hourly instantaneous downward flux. The retrieved daily DSR is compared with ground-based measurements at 96 stations from China Meteorological Administration (CMA). The results show that the estimated DSR is in good agreement with ground measurements over China with a correlation coefficient of 0.93 and a mean bias of 5.8%. Root-mean square differences in the daily DSR are 20.7% for all sky conditions. The daily DSR is also compared with observations on Tibetan Plateau and the results shows a correlation coefficient of 0.91 and a mean bias of 1.53%. Root-mean square differences are 17.5%. The differences between the satellite derived estimates and ground observations may be attributed to calibration uncertainty of the satellite sensor and the ground instruments, undetected cloud shadows, steps of the LUT parameters, uncertainty in determining surface reflectance, and errors in ground observations.