Quality assessment of the National Ceiling and Visibility (NCV) forecasts is somewhat simpler than assessment for many other products. For a given time, the NCV forecasts are essentially two-dimensional; many other avaition forecasts are three-dimensional. Further, the ceiling and visibility observations (METARs) are relatively stationary and consistent over time; many other types of observations move or vary over time, e.g. rawinsonde, pilots' reports, satellite, radar.

Nonetheless, METAR observations of ceiling and visibility suffer from some biases. In particular, observations tend to be very dense in populated areas and quite sparse in outlying or rural areas (see graphic). Spatial statistics are being used to examine the relationship between neighboring stations. In particular, a network-design strategy is employed to identify stations that contain redundant ceiling and visibility information. For further information, see Gilleland and Fowler (2006).

Gilleland, E. and T. Fowler, 2006: Network design for verification of ceiling and visibility forecasts. Environmetrics, 17, 575-589.