Hi David:thanks for trying out that tile. Yep, it's bad. Like I said - and shown below - merging it with my 250-m data yields the best results of any of my methods (not surprisingly) but notice those raised edges along some of the shorelines.One thing that neither Vlada's method nor yours can deal with is an area like that big black hole in the lower left area. What looks like a perfect meteor crater are in fact two lovely islands surrounded by ocean (they are more fuzzy in my image because they are based on 250-m data). No amount of Kriging will make these appear if there's not even a single height point. :-( The problem is that all these geostatistical methods are inherently dumb whereas generally available data for these areas, like GTOPO30, are extremely coarse (though GT30 does make the islands appear). Maybe there's a way of incorporating both procedures/data...Regarding the offset issue. Thanks for the clippings - I knew about the file format and my EXCEL work tool compensates for it. No, the offset appears to be more general in nature though it's less pronounced than with the USGS data for SE Alaska.Cheers, Holger

Hi Holger and David,here is also the tile N50W125 after the kriging in the Surfer:http://www.volny.cz/stoje/scen/N50W125.jpgWe can't expect miracles from any interpolation, but it can help. If you will wish to compare it in close view for some small areas, the modified file hgt is at http://www.volny.cz/stoje/scen/N50W125K.hgt.zipI am sorry that in the help for the Surfer there are no formulas for this method, only references. I believe that the inverse distance to power in the Microdem or the linear interpolation with the Delaunay triangulation in the David's utility can do similar thing, especially in the combination with the merging of the other sources of data and flattening the water with the land/water masks. Looking forward for the news on this field!Cheers