Status of Mountain Goats in Washington (2008)

Approach

I developed an estimate of the number of mountain goats in Washington based on aerial surveys conducted 2004-2007 and expert guessing for areas not surveyed. To determine geographic units for estimation I started with 2007 WDFW hunting units, added areas not covered in 2007 units from 2002 units, and then added ad hoc polygons for areas not included by either. For surveys, I adjusted counts and calculated 90% confidence intervals based on our sightability model and averaged across years for units with multiple surveys. For non-surveyed locales, I solicited opinion (guestimates) as to best guess, likely minimum and likely maximum for designated ad hoc units.

In some cases units were partially surveyed, where I generally added the survey estimate and the guestimate for the remaining areas, but in some cases the guestimate included the surveyed area, in which case I used the guestimate for that unit. For the total estimate range, I took the simple expedient of adding the upper and lower confidence intervals from the surveys with the likely minimum and likely maximum guestimates.

I summarized unit level estimates in 2 hierarchical groupings, Area and Zone. Survey estimates and guestimates were added separately within each Area and Zone. For surveys, I added the variances and recalculated confidence intervals based on the total variance. For guestimates, I simply added likely minimum and likely maximum guestimates. Hence, precision increased with broad scale estimation for surveys, but not for guestimates.