Recommendation: proceed with the data analysis using the geographic subdivisions listed in the plan, but stress the Tri-State and Canadian segments along with the overall RMP (the latter being the population); emphasize a 5 year time frame.

Geographic subdivisions (e.g., Red Rock Lakes NWR, Yellowstone NP, etc.) do not represent populations, but rather nesting/wintering distributions.

The analysis will be used to gauge whether changes in numbers are in a positive direction or not. If the trend does not indicate that objectives as delineated in the plan are going to be met, then managers might take some actions for change.

Managers do not, and should not, make management recommendations based on one year changes/trend in numbers. It is suspected that variability in the data due to collection techniques is great enough that there would be distrust in analyses based on one year's change. A five year time horizon was suggested.

FWS is interested in trend analysis as a "report card" (descriptive statistics) to assess what has occurred rather as a "diagnostic" (predictive statistics) to simulate effects of future management.

Red Rock Lakes NWR should have all data that exists for Tri-State areas. Standard, Mid-Winter Waterfowl Population Census data should not be used. The special winter swan aerial survey data should be, and that data goes back to the early 1970's. Bob Trost will send Rick Sojda the summaries he has of the official breeding survey done at five-year intervals. It was not known when the current breeding pair surveys were started.