Time sequence of geopotential height differences; box is 10 km high.
The yellow are +7 gpm differences; the heights are higher using the MODIS winds. The green are -7 gpm; the heights are lower using the MODIS winds.
The sequence begins on 10 Aug 04 and continues for 10 days. These differences are from the analyses every 6 hours, with and without the MODIS winds.
As expected, and with few exceptions, the biggest difference features are in data sparse regions: Antarctica and surrounding region, the Arctic, and the Pacific. Some of the more interesting exceptions:
* Central Africa
* Tibet
There are indications that the Central African difference extends, at times, into the Indian Ocean along where the ITCZ should be. From this view we can't see the Atlantic, but there could be similar differences.
The Tibetan spike is probably an artifact because the plateau averages about 5km in height, which is 1/2 the depth of this box. The elevation of Antarctica is also a problem.
It takes a couple of days for the differences to really kick in over the Pacific. But, there are times, for example 00Z on the 18 & 19 Aug, where the differences over the Pacific are small.