Fig. 9: The analysis increment of relative humidity resulting from introducing a perturbation of 1 K to the clear observation (pentagram). | npj Climate and Atmospheric Science

Fig. 9: The analysis increment of relative humidity resulting from introducing a perturbation of 1 K to the clear observation (pentagram).

From: FuXi-DA: a generalized deep learning data assimilation framework for assimilating satellite observations

Fig. 9

Horizontal distribution of R300 increments resulting from introducing a perturbation to channel 9 at the beginning (a), middle (b), and end (c) of the assimilation time window. Vertical distribution of relative humidity increments resulting from introducing a perturbation to channel 9 at the beginning (d), middle (e), and end (f) of the assimilation time window, obtained from a cross-section delineated by the dashed lines in (a). Horizontal distribution of R300 increments resulting from introducing a perturbation to channel 11 at the beginning (h), middle (i), and end (j) of the assimilation time window. Vertical distribution of relative humidity increments resulting from introducing a perturbation to channel 11 at the beginning (k), middle (l), and end (m) of the assimilation time window, obtained from a cross-section delineated by the dashed lines in (a). Humidity Jacobian functions for channel 9 (g) and channel 11 (n). The humidity Jacobian is computed using the rapid radiative transfer model RTTOV version 13.2 under clear-sky condition, with the atmospheric profile input to RTTOV derived from the US standard atmosphere.

Back to article page