Published: 2008-08-03 10:52:34
Updated: 2008-08-03 10:52:34
Posted August 3, 2008
By Randy Dunson
MIKE MOSS SAYS: Randy, So long as there are enough surrounding observations and there are not excessive variations in terrain, elevation and the like between those stations and a location you would like to forecast for, a reasonable forecast value for temperature can be interpolated to that location, including interpolation from gridded model output fields. When developing automated forecasts for specific sites (like an airport) it is of course helpful to have a database of observations for that site, and over time the accuracy for a location like that may be slightly higher than for a nearby location without regular observations, but for a location like Hillsborough the difference would be pretty small.
As for the question about satellite views, we have an IR view called the "regional" projection that is fairly close in to North Carolina, enough to include counties on the map. While the time resolution isn't as high as it could be on the loop, we do provide a loop that runs 8 hours in length, longer than most other readily accesible satellite loops. If we zoomed in much tighter with the IR imagery, which seems to be what you're asking about, the data would begin to look very "blocky" because of the relatively low resolution of the IR sensors.
The visible data is a different issue and one that we may be able to address. Resolution is higher with visible data (a function of the shorter wavelengths involved) and we might be able to make an image that is zoomed in a bit more than the others. We did not do this when our online satellite imagery was developed because at the time the visible imagery we received was not in a format that translated well to a detailed image. We've been receiving the data in a more usable format for some time now, though, one that should lend itself to adding visible imagery to our site. Can't guarantee anything right away, but we'll certainly look into it.