Validation of a weather forecast model at radiance level against satellite observations allowing quantification of temperature, humidity, and cloud-related biases Academic Article uri icon


  • 2016. The Authors. An established radiative transfer model (RTM) is adapted for simulating all-sky infrared radiance spectra from the Canadian Global Environmental Multiscale (GEM) model in order to validate its forecasts at the radiance level against Atmospheric InfraRed Sounder (AIRS) observations. Synthetic spectra are generated for 2 months from short-term (39 h) GEM forecasts. The RTM uses a monthly climatological land surface emissivity/reflectivity atlas. An updated ice particle optical property library was introduced for cloudy radiance calculations. Forward model brightness temperature (BT) biases are assessed to be of the order of 1 K for both clear-sky and overcast conditions. To quantify GEM forecast meteorological variables biases, spectral sensitivity kernels are generated and used to attribute radiance biases to surface and atmospheric temperatures, atmospheric humidity, and clouds biases. The kernel method, supplemented with retrieved profiles based on AIRS observations in collocation with a microwave sounder, achieves good closure in explaining clear-sky radiance biases, which are attributed mostly to surface temperature and upper tropospheric water vapor biases. Cloudy-sky radiance biases are dominated by cloud-induced radiance biases. Prominent GEM biases are identified as: (1) too low surface temperature over land, causing about 5 K bias in the atmospheric window region; (2) too high upper tropospheric water vapor, inducing about 3 K bias in the water vapor absorption band; (3) too few high clouds in the convective regions, generating about +10 K bias in window band and about +6 K bias in the water vapor band.

published proceedings


altmetric score

  • 2.25

author list (cited authors)

  • Shahabadi, M. B., Huang, Y. i., Garand, L., Heilliette, S., & Yang, P.

citation count

  • 7

complete list of authors

  • Shahabadi, Maziar Bani||Huang, Yi||Garand, Louis||Heilliette, Sylvain||Yang, Ping

publication date

  • September 2016