Satellite-flown instruments have been acquiring data that pertain to the Earth's
atmosphere, cryosphere, land, and oceans since the early 1970's. These data include
measurements over time from a variety of instruments with different spatial and temporal
resolutions. The data can provide answers to earth-science questions, for example, What
changes in radiation budgets are associated with global climate change? Each
instrument alone may provide answers to specific questions that relate to the instrument's
spatial and temporal coverage.  To address broader questions, we integrate data from
multiple instruments to obtain spatial/temporal coverage and resolution that are greater
than coverage and resolution of individual instruments.  We utilize instrument-
measurements and human observations available from surface weather stations to
validate and complement data from satellite-flown instruments. In a preliminary
investigation we automatically constructed a decision-tree model from AVHRR data
labeled with surface observations. The decision tree detected clouds from test AVHRR
data with accuracy that was 25-percent better than the accuracy of human-generated
decision trees.

TRL: 3-4 tested on cloud detection from AVHRR data with good results