Main content area

Temporal Trends in Stream Habitat on Managed Forestlands in Coastal Southeast Alaska

Martin, Douglas J., Shelly, Alice
North American Journal of Fisheries Management 2017 v.37 no.4 pp. 882-902
fish, gravel, habitats, landslides, models, monitoring, particle size, riparian forests, spawning, streams, trees, wind, windthrow, wood, Alaska
We evaluated trends in fish habitat condition and their relationship to natural and management-associated disturbances in forests of coastal Southeast Alaska. Monitoring data spanning 9–19 years (1994–2012) from 14 stream reaches were examined using linear mixed-effects models to evaluate patterns and trends in fish habitat condition. Modeling detected multiple patterns of habitat change that included postharvest pulses in large wood (LW) recruitment and reductions in substrate particle size, both followed by long-term trends (average LW density increased by 0.8–0.9 pieces/year; median substrate particle size decreased by 0.6–1.3 mm/year) that were observed up to two decades after harvest. Trends in spawning gravel followed patterns of change in substrate except at higher-gradient, cobble-dominated reaches, where the reduction in substrate particle size improved spawning habitat. Trends in pool habitat were not detectable over the short term for most attributes but exhibited gradual changes after harvest that became strongly evident over the long term (average residual pool depth increased by 0.5 cm/year, wood-formed pools increased by 0.9% per year). The changes in habitat conditions and correlations with riparian forest attributes (buffer strip length and wind exposure) suggested that the habitat responses were related to increased tree windthrow after timber harvest. Furthermore, the degree to which LW recruits functioned in forming pools appeared to influence long-term trends in stream habitat. We observed that only 26% of new LW recruits, on average, were functioning to form habitat and that wood function increased over time, peaking (72% of recruits were functioning) about three decades after recruitment. Our monitoring over two decades enabled us to discern slowly evolving trends, showed how trends in pool habitat that are not initially evident may be associated with buffer disturbances (windthrow and landslides), and demonstrated the long-term consequences of management actions. Received August 2, 2016; accepted May 5, 2017 Published online July 12, 2017