We described the seasonal distribution of Geographic Positioning System (GPS)-collared mountain goats (Oreamnos americanus) in Mount Rainier, North Cascades, and Olympic National Parks to evaluate aerial survey sampling designs and provide general information for park managers. This work complemented a companion study published elsewhere of aerial detection biases of mountain goat surveys in western Washington. Specific objectives reported here were to determine seasonal and altitudinal movements, home range distributions, and temporal dynamics of mountain goat movements in and out of aerial survey sampling frames established within each park. We captured 25 mountain goats in Mount Rainier (9), North Cascades (5), and Olympic (11) National Parks, and fitted them with GPS-collars programmed to obtain 6-8 locations daily. We obtained location data on 23 mountain goats for a range of 39-751 days from 2003 to 2008. Altitudinal distributions of GPS-collared mountain goats varied individually and seasonally, but median altitudes used by individual goats during winter ranged from 817 to 1,541 meters in Olympic and North Cascades National Parks, and 1,215 to 1,787 meters in Mount Rainier National Park. Median altitudes used by GPS-collared goats during summer ranged from 1,312 to 1,819 meters in Olympic and North Cascades National Parks, and 1,780 to 2,061 meters in Mount Rainier National Park. GPS-collared mountain goats generally moved from low-altitude winter ranges to high-altitude summer ranges between June 11 and June 19 (range April 24-July 3) and from summer to winter ranges between October 26 and November 9 (range September 11-December 23). Seasonal home ranges (95 percent of adaptive kernel utilization distribution) of males and female mountain goats were highly variable, ranging from 1.6 to 37.0 kilometers during summers and 0.7 to 9.5 kilometers during winters. Locations of GPS-collared mountain goats were almost 100 percent within the sampling frame used for mountain goat surveys in Mount Rainier National Park, whereas generally greater than 80 and greater than 60 percent of locations were within sampling units delineated in North Cascades and Olympic National Parks, respectively. Presence of GPS-collared mountain goats within the sampling frame of Olympic National Park varied by diurnal period (midday versus crepuscular), survey season (July versus September), and the interaction of diurnal period and survey season. Aerial surveys conducted in developing a sightability model for mountain goat aerial surveys indicated mean detection probabilities of 0.69, 0.76, and 0.87 in North Cascades, Olympic, and Mount Rainier National Parks, respectively. Higher detection probabilities in Mount Rainier likely reflected larger group sizes and more open habitat conditions than in North Cascades and Olympic National Parks. Use of sightability models will reduce biases of population estimates in each park, but resulting population estimates must still be considered minimum population estimates in Olympic and North Cascades National Parks because the current sampling frames do not encompass those populations completely. Because mountain goats were reliably present within the sampling frame in Mount Rainier National Park, we found no compelling need to adjust mountain goat survey boundaries in that park. Expanding survey coverage in North Cascades and Olympic National Parks to more reliably encompass the altitudinal distribution of mountain goats during summer would enhance population estimation accuracy in the future. Lowering the altitude boundary of mountain goat survey units by as little as 100 meters to 1,425 meters in Olympic National Park would increase mountain goat presence within the survey and reduce variation in counts related to movements of mountain goats outside the survey boundaries.