Historical changes in pool habitats in the Columbia River Basin

An historical stream survey (1934-1945) was compared with current surveys ( 1987-1997) to assess changes in pool frequencies in the Columbia River Basi n. We surveyed 2267 km of 122 streams across the basin, representing a wide range of lithologies, stream sizes, land use histories, ownerships, and ec oregions. Based on pool classes inherited from the historical surveys, the frequencies of large (greater than or equal to 20 m(2) and greater than or equal to 0.9 m depth) and deep (greater than or equal to 20 m(2) and greate r than or equal to 1.8 m depth) pools have decreased significantly (P < 0.0 1) since the 1930s. We classified streams as natural or commodity based on their watershed management and land use histories. Natural streams were in watersheds minimally affected by human activities (e.g., wilderness or road less designation, limited entry), with only 12% having roads in riparian ar eas. Commodity streams were defined as having watersheds managed predominan tly for extraction of resources via timber harvest, livestock grazing, and other human activities. Ninety percent of these streams had roads in the ri parian areas. In natural streams, large-pool frequencies increased or remai ned the same in 96% of the streams (88% for deep pools). In commodity strea ms. large- and deep-pool frequencies decreased in 52% and 54% of the stream s, respectively. Despite differences in stream size and the level of human activities, the magnitude and direction of these changes were consistent. L and ownership did not influence trends; pools decreased significantly on bo th private and public lands. Only where entire watersheds or headwaters wer e designated as wilderness or roadless areas did pools consistently remain unchanged or increase. Pool frequencies decreased in all ecoregions except the North Cascades ecoregion. We developed regional histories of human acti vities for the Columbia River Basin. Human activity histories were typicall y of low spatial resolution and available for broad geographic areas only; we rarely were able to obtain information at the scale of individual waters heds. Consequently, we were unable to test the relationship between tempora l and spatial patterns in human activities and their influence on site-spec ific trends in pools. Despite our inability to isolate causal mechanisms, m anagement emphasis and human activities clearly influenced trends in pools. We conclude that the persistent effects of human activities have simplifie d stream channels and reduced large- and deep-pool frequencies in watershed s outside of designated wilderness and roadless areas in the Columbia River Basin.