|
 Accomplishments
Causal Discovery
Hundreds of gigabytes of satellite data, model output
data, and meteorological data have been mined to produce
predictive models for forest fire risk that - using
existing data sources - outperform standard fire danger
ratings. These models are currently being integrated
in the Ecocast architecture for further testing and
validation.
Distributed Data Mining
TOPS and the IMAGEbotPlanner
have been integrated to create a distributed architecture
for automated data retrieval and analysis using optimized
data processing plans. An API has also been developed
and implemented to support rapid integration of novel
models discovered by Tetrad and other causal analysis
and discovery systems. Beginning in 2003, the integrated
system has and used to produce a range of ecocasts,
including the automated generation of daily 8 km maps
of Gross Primary Productivity (GPP),
soil
moisture, and evapo-transpiration
for the entire continental U.S.
|
