Publications

Coll, C; Galve, JM; Niclos, R; Valor, E; Barbera, MJ (2019). Angular variations of brightness surface temperatures derived from dual-view measurements of the Advanced Along-Track Scanning Radiometer using a new single band atmospheric correction method. REMOTE SENSING OF ENVIRONMENT, 223, 274-290.

Abstract
Surface temperatures derived from remote sensing data over heterogeneous, non-isothermal land surfaces depend on the viewing and solar angles mainly due to variations in sunlit and shaded fractions of the different elements in the field of view. The near-simultaneous dual-view capability of the Advanced Along-Track Scanning Radiometer (AATSR) can be used to estimate differences in brightness surface temperatures (BSTs) between the nadir (satellite zenith angle of 0 degrees-21.7 degrees) and forward views (53 degrees-55.6 degrees) in the 11 and 12 mu M bands. BST is defined as the black-body temperature corresponding to the radiance at surface level (that is, corrected for atmospheric absorption and emission). We developed an automated, pixel-by-pixel atmospheric correction method to obtain the BST for the 11 and 12 pm bands at nadir and forward views. It uses atmospheric profiles from the NCEP reanalysis product provided at 1 degrees x 1 degrees spatial resolution every 6 h, the MODTRAN 5 radiative transfer model, and a digital elevation model with spatial resolution of 1 km. The method provides the atmospheric transmittance and upwelling radiance at the AATSR resolution (1 km) for the two bands and views, taking into account the geographical coordinates, altitude and zenith observation angle of each pixel, and the AATSR overpass time. The method was applied to eight daytime and nighttime AATSR scenes over the Iberian Peninsula on the four seasons, for which the nadir-forward BST differences were obtained. For sea surfaces, the angular BST difference was uniform and independent on the time and season with an average about +0.7 K. This is expected since the sea surface is homogeneous and flat, the only angular effect being due to the known angular variation of emissivity. Nadir-forward BST differences were usually positive for daytime data over land surfaces, since AATSR observes from the North so there is a larger fraction of sunlit elements in the nadir view in the northern hemisphere. We found a clear seasonal dependence as expected, the largest (smallest) angular BST differences being obtained in summer (winter). The spatial variability of the daytime angular BST differences over land was also highest (lowest) in summer (winter). For the nighttime data, the angular variation was much smaller and uniform for all seasons. In order to analyze the daytime angular BST differences over land, we used concurrent Leaf Area Index (LAI) data obtained from the MODIS product MOD15A2. For a given date, the largest differences were found for surfaces with LAI values ranging from 0.5 to 2, due to the large change in the proportions of sunlit and shaded elements observed by the two views. For such cases in summer, differences up to 8 K were found, with average values of 4 K and high variability. For LAI values larger than 3, the angular differences decreased to about 2 K and were more uniform. We also found a positive correlation between the angular BST difference and the nadir BST.

DOI:
10.1016/j.rse.2019.01.021

ISSN:
0034-4257