Long-term monitoring of western aspen--lessons learned.

Aspen woodland is an important ecosystem in the western United States. Aspen is currently declining in western mountains; stressors include conifer expansion due to fire suppression, drought, disease, heavy wildlife and livestock use, and human development. Forecasting of tree species distributions under future climate scenarios predicts severe losses of western aspen within the next 50 years. As a result, aspen has been selected as one of 14 vital signs for long-term monitoring by the National Park Service Upper Columbia Basin Network. This article describes the development of a monitoring protocol for aspen including inventory mapping, selection of sampling locations, statistical considerations, a method for accounting for spatial dependence, field sampling strategies, and data management. We emphasize the importance of collecting pilot data for use in statistical power analysis and semi-variogram analysis prior to protocol implementation. Given the spatial and temporal variability within aspen stem size classes, we recommend implementing permanent plots that are distributed spatially within and among stands. Because of our careful statistical design, we were able to detect change between sampling periods with desired confidence and power. Engaging a protocol development and implementation team with necessary and complementary knowledge and skills is critical for success. Besides the project leader, we engaged field sampling personnel, GIS specialists, statisticians, and a data management specialist. We underline the importance of frequent communication with park personnel and network coordinators.