Liebmann, B., and C. A. Smith, 1996: Description of a complete (interpolated) outgoing longwave radiation dataset. Bull. Amer. Met. Soc., 77, 1275-1277.
Estimates of outgoing longwave radiation (OLR) from National Oceanic and Atmospheric Administration (NOAA) polar-orbiting satellite (Gruber and Winston 1978; Gruber and Krueger 1984) are often used to distinguish areas of deep tropical convection and to estimate the earth's radiation budget. The National Environmental Satellite Data and Information Service (NESDIS) of NOAA archives the data onto 2.5° lat × 2.5° long grids. The two grids per day correspond to the daytime and nighttime orbits. The local equatorial crossing times have changed over the years (Gruber and Krueger 1984).
One problem in using these data is that missing grids, and missing values within grids, are often present, presumably owing to satellite problems, archival problems, or incomplete global coverage. When a grid is incomplete, but not entirely missing, the missing values are completed by a nearest neighbor spatial interpolation. These interpolated values are flagged as negative. Missing values tend to occur in "swaths." Thus, when the number of missing values is large, the resultant errors are large as well, because of the large distance over which a value is interpolated. This problem is evident in Fig. 1a, which shows the NESDIS estimate of the OLR field for the daytime pass of 24 January 1979. On this date 6837 (out of 10,224) values were flagged as interpolated. Clearly, the field is too smooth east of 90°E. Figure 1b shows the field after missing values have been interpolated by the algorithm described below, in which missing data are primarily interpolated in time. This field is seen to appear more realistic and at least captures the low-frequency phenomena.
The point of this note is to describe a dataset that is available to the public. Missing values have been removed by temporal and spatial interpolation, except for the gap between 17 March and 31 December 1978.