Ecological Indicators: Software Development

1
Ecological Indicators Software Development

• Sergei Rodionov
• Joint Institute for the Study of the Atmosphere
  and Ocean, University of Washington, Seattle,
  Washington

2
Outline

• Part I. Red noise problem. Upgrade to the regime
  shift detection method (STARS).
• Part II. Knowledge management system for the
  Bering Sea.

3
True and spurious regime shifts
4
Entry Form for STARS
www.BeringClimate.noaa.gov
5
Estimates of AR1 for PDO
6
PDO Index Before and After Prewhitening
7
Part II. Dealing with Information Overload

• Dimensionality reduction (e.g., principal
  component analysis, singular value decomposition,
  multidimensional scaling)
• Knowledge management system

8
Components of the KMS

• Data Explorer
• Rule Explorer
• Inference Engine
• Graphical Interface
• Search Facility
• Reporting Facility

9
Data Explorer
10
List of rules (partial) with the IF variables
that affect walleye pollock recruitment
11
Part of the influence diagram for walleye pollock
recruitment
12
Rule Explorer
13
Rule Edit Form
14
Example of a rule
IF ENSO event El Niño, AND Aleutian low
circulation type W1, THEN SAT at St. Paul
above normal CF 10.
15
Running the project to find the value of the
target variable
16
Questions about the variable in the terminal nodes
17
Search Form
18
Confidence Factors
CF (P(C e) - P(C e)) 100,
CF Degree of Belief Degree of Disbelief
CFcomb CFold CFnew - (CFold CFnew)/100.
19
Custom Property Window
20
Report documenting the inference process
21
A spoon is valuable at the lunch time.
Russian proverb