Western U.S. rivers are currently influenced by legacy effects of reduced large wood (LW) loading and retention that has led to reduced in-stream habitat complexity. Historical land use practices such as tie-driving, snag removal, and beaver trapping have all contributed to declines in either input or retention of LW in mountain streams. Effects from these practices have persisted over a century after the activities ceased; suggesting streams have entered an alternative stable state. One that is characterized by loss of multi-thread stream reaches, reduced sediment and nutrient retention, steeper channel gradient, reduced pool volume and altered pool geometry, creating narrower and shallower streams with higher width to depth ratios, and overall decreased stream and valley complexity.
Using one of the last remaining patches of old-growth forest on the Front Range of Colorado (USA) we compared population densities, individual growth rates, diet compositions, and annual prey consumption demand of brook trout Salvelinus fontinalis in streams across a gradient of wood volumes. Brook trout populations were sampled via multi-pass electrofishing events, while approximately 30 individuals at each site were sacrificed for growth and diet studies. Individual growth was back-calculated from otolith sections, and diet composition was estimated from stomach samples collected at each site. Average growth rates and diet compositions were then combined in a bioenergetics model that provided an estimate of the amount of prey resources needed to support observed growth and population densities.
Trout population densities appear positively related to increasing number of pools, and standing stock biomass of aquatic insects. Yet, at different scales (square-meter vs valley-length) different variables were better at predicting trout biomass. At both the local (square-meter) and landscape (valley length) scales, standing stock biomass of aquatic insects was the best predictor variable chosen by model selection to predict trout biomass. However, at the valley scale, number of pools became more important in predicting trout biomass, as a combination of standing stock biomass of aquatic insects and number of pools held the most predictive power.
Annual individual prey consumption demand did not differ between two sites of high and low wood volume and trout biomass. Therefore, total prey demand at a site is determined by population density, and less so by physical habitat. Between the two intensive sites, diet compositions were significantly different during the summer season. Fish at the high wood site were consuming more small aquatic insect larvae (e.g., Family Chironomidae and Simullidae) to support growth and fish biomass, whereas diet composition at the low wood site consisted of terrestrial insects (ants) and larger aquatic insect larvae.
Individual growth rates of age-1 brook trout were negatively affected by increasing density. However, this pattern only held true for fish that were not the largest individuals, nor the smallest individuals, at each site. The largest fish at each site are presumably dominant individuals, and can out compete all other individuals for optimum foraging positions regardless of density, and are therefore not affected by density. While the smallest fish, presumably the most subordinate individuals, are outcompeted for resources at all densities, and are likewise also not affected by increasing density.
Our results suggest that LW, in the correct geomorphic context, can drastically alter stream and valley habitat complexity. The creation of multi-thread stream reaches can greatly increase aquatic animal production by creating pool habitat preferred by trout and by prey resources that dramatically increase trout populations. The negative effects of historical land uses have persisted >100 years suggesting a resilience of streams to return to pre-perturbation states, and a loss of trout production in mountain watersheds due to land use over a century ago.