The twentieth century reanalysis varies in quality in both time and space, because it depends on the limited and variable supply of historical surface pressure observations. This video shows monthly mean-sea-level pressure from the reanalysis for 140 years (1871-2010) masked by fog where there are too few observations to allow a precise reconstruction.

All data are taken from the 20CR monthly composites ( The necessary climatologies are just the means of 20CR monthlies over 1981-2010.

The fog is based on the relative entropy ( of the analysis MSLP field with respect to climatology. Essentialy this is a metric of the information lost by using the climatology instead of 20CR - where RE is large 20CR is much more informative than climatology.

The fog masks regions where re is small (20CR only slightly better than climatology). Fog thickness is proportional to min(0,1-re/threshold) where the threshold is arbitrarily set to exp(1).

Loading more stuff…

Hmm…it looks like things are taking a while to load. Try again?

Loading videos…