* Why are there data voids in parts of the plots?
* Why aren't the surface properties masked out under clouds?
* Why are there sometimes pockets of colder or warmer surface temperatures?
* Where can I find documentation on the algorithms?
Browser Issues:* Things don't always work right when I resize the window. What's going on?
* The web page doesn't work correctly with the Konqueror browser.
Other:* Are the actual data available rather than just the plots?
* What recent changes have there been?
Why are there data voids in parts of the plots?
A single satellite doesn't give complete coverage of the Arctic or Antarctic over an eight-hour period.Why aren't the surface properties masked out under clouds?
The surface albedo and temperature are estimated under cloud cover. These should be considered experimental. Having said that, the algorithms are generally sound but problems can arise as a result of the implementation. In particular, when the clear sky values are interpolated to cloudy locations for use in the adjustment for cloud cover, problems with cloud detection can be magnified in the clear sky surface properties. See the CASPR Reference Guide for details.Why are there areas indicated by the cloud mask as cloudy but without values for cloud properties and radiative fluxes?
There are a number of conditions for which retrievals are not done, so not all pixels identified as cloudy in the cloud detection step will have retrievals for other parameters. For example, with a very low sun and a bright surface cloud optical depth retrievals often fail and are set to the missing value. Similarly, the uncertainty in the surface albedo retrieval for solar zenith angles between 80 and 90 degrees is very high, so no retrieval is done. Of course, when any cloud or surface properties are missing, radiative fluxes cannot be calculated so they will also be missing.Why are there sometimes pockets of colder or warmer surface temperatures?
This occurs in extensively cloud areas when there are only a few clear pixels, some of which are misidentified as clear. When the clear sky pixels are interpolated to cloudy areas, which necessary for some parameter retrievals, some very cloudy areas will have incorrect surface temperatures. This is one of the most difficult problems in single image processing.When a given location could be filled by values from multiple overpasses, how do you decide which to use?
First of all, every overpass that falls within X hours of the target time, currently 4 hours either side of 12Z, is considered for inclusion in the composite. The every pixel is evaluated in terms of its time difference from the target time and its scan angle. The goal is to use pixels that are as close to nadir and as close to the target time as possible. As score is calculated asWhere can I find documentation on the algorithms?
The best source of information is the CASPR Reference Guide. The CASPR User's Guide might also be of some use. See the CASPR web page.Things don't always work right when I resize the window. What's going on?
This problem is being investigated. For now the best solution is to simply reload the page and display the plots again.The web page doesn't work correctly with the Konqueror browser.
It may be that the Javascript implementation is incomplete in Konqueror. No solution is available at this time.Are the actual data available rather than just the plots?
Not at present. If you're interested, send us a message.Are the data archived?
Old plots are available via the "Archive" link on the left. The actual parameter data are not saved. However, the original images from which CASPR generated the parameter results are archived. Let us know if you need them.What recent changes have there been?
See the revision history for a complete list of changes.