Data Products (Beta release)
The multilayered cloud data products are available in addition to the operational MODIS Direct Broadcast data products that may be obtained through the CIMSS home page or by the above link. The operational MODIS DB products include the Level 1B radiances, geolocation, cloud clearing (MOD35) and cloud products (MOD06).
All the DB images have 1354 pixels per scan line, so divide the size of the images by 1354 to get the number of scan lines. All three products are the same size. To obtain the data products: http://cimss.ssec.wisc.edu/ol
Description of the Data Products
Cloud thermodynamic phase (1-km spatial resolution)While the MODIS operational cloud phase product is derived at 5-km spatial resolution (i.e., for a 5x5 pixel array of 1-km data where the radiances have been averaged), the cloud phase product derived as part of the MODIS DB effort is derived at full resolution (1 km). At 1-km resolution, one might notice some artifacts in the product, such as striping, in some areas. Cloud clearing is performed using the same set of multispectral tests that is used to produce the operational MODIS product MOD35. The IR-based product is derived from the brightness temperatures at 8.5 and 11 microns (MODIS bands 29 and 31, respectively). In the near future, we intend to augment the simple IR-based tests with visible/near-infrared bands. When this occurs, additional products will appear on the site.
The cloud phase product is simply a binary file, with 1 byte/pixel. The values may be interpreted as follows:
0: clear
3: mixed
4: ice
5: water
6: uncertain
7: no retrieval
Multilayered cloud detection product (1-km spatial resolution)
The multilayered cloud, or overlapping cloud, detection scheme is described in Baum and Spinhirne (2000) with additional information on the Detection Methodology page. A paper is now being prepared to describe this in more detail. The multilayered cloud product is also a 1-byte/pixel binary file, with the values ranging from 0 to 255. This number describes the number of times each pixel was analyzed as potentially containing multilayered cloud. The higher the number, the higher the certainty that the assumption of a single cloud layer in the pixel is false. Another way of saying this is that the higher the number, the higher the likelihood that the pixel contains more than one cloud layer. The number (N) can be converted into a multilayered cloud detection percentage by (N/256) * 100.
Merged cloud phase/multilayered cloud product (1-km spatial resolution)
As an experimental product, we have merged the cloud clearing, cloud phase, and cloud overlap products as follows. Again, this is a 1-byte/pixel product. Pixels designated as clear-sky or having insufficient information for cloud clearing (i.e., no retrieval) are combined into one category. The uncertain phase and mixed-phase categories obtained from the cloud thermodynamic phase process are likewise combined into one category. Pixels retrieved as either "uncertain phase" or "mixed phase" are combined into one category. Two multilayered cloud overlap categories are provided: one category for multilayered cloud detection percentage for between 1% and 20% (i.e., lower certainty), and the second category for multilayered cloud detection percentage greater than 20% (i.e., high certainty).
0: clear/no retrieval
1: ice phase
2: water phase
3: uncertain phase or mixed phase
4: 1% to 20% multilayered cloud detection
5: 20% to 100% multilayered cloud detection