Climate Change and Artificial Intelligence Techniques

1
Climate Change and Artificial Intelligence
Techniques
The Issues Uncertainties related to 1. Data
(due to inadequate description of their
nature) 2. Model structure and parameters (due
to inadequate understanding of phenomena
under study) Climate change only increases those
uncertainties Usual Solution Based on
probabilities and statistics (scenario analysis,
data described by distributions,
adaptive modeling)
Nicolas Lauzon, C-Ciarn-BC Meeting, UBC, October
30, 2002
2
The Case of Artificial Intelligence Techniques
1. Applicability (what you can do with other
methods, you can do with artificial
intelligence techniques) 2. Adaptable
(stand-alone or combined with other methods for
enhanced results) 3. Implementation (easy and
flexible) 4. Description of data (patterns and
knowledge maps) 5. Handle both hard
(observations) and soft (expert judgment) data
3
Application Cases

• 1. Hydrologic Inflow Modeling Adaptive
  Parameters
• - Improvement overall compared with standard
  model
• - Adaptability to changing conditions on the
  watersheds
• 2. Detection of Shifts and Trends in Hydrometric
  Data
• - Compare favorably with standard methods of
  detection
• - Multivariate cases