Effects of resource limitation on a detrital‐based ecosystem

We examined the importance of terrestrial detrital inputs to secondary productivity of a headwater stream. Following a year of pretreatment studies on two headwater streams, we excluded terrestrial litter inputs (=treatment) to one stream while using the other as a reference. We excluded litter for 3 yr followed by 1 yr of small woody debris (≤10 cm diameter) removal and litter exclusion. Monthly benthic samples were collected from dominant mixed substrate (cobble, pebble, and sand‐silt) as well as from moss‐covered bedrock outcrop substrates. We used randomized intervention analysis (RIA) to test the null hypotheses that no change in abundance or biomass of functional feeding groups or specific taxa occurred in the treatment stream relative to the reference stream. Benthic organic matter was significantly lower in mixed substrate habitats of the treatment stream; however, small woody debris did not show a significant reduction prior to manual removal during year 4. At the end of the treatment period, total benthic invertebrate abundance in mixed substrates in the treatment stream was less than one‐tenth of that in the reference stream, and biomass in the treatment stream was one‐sixth of that in the reference stream. Biomass and abundance of shredders, gatherers, total primary consumers, and predators displayed significant treatment effects (P < 0.012–0.00001, RIA) in the mixed substrate habitats during the 4‐yr treatment. Only scraper and filterer functional groups failed to show differences between streams during the 4‐yr treatment. Twenty of the 30 taxa comprising >90% of total secondary production displayed a significant decrease in abundance, biomass, or both, in the treatment stream relative to the reference stream (P < 0.05, RIA). Total secondary production in mixed substrate habitats declined to 22% of pretreatment values by the fourth year of treatment and is among the lowest reported for streams. Removal of small woody debris resulted in an additional 47–50% decrease in abundance, biomass, and production of the benthic fauna compared to the third year of litter exclusion. In contrast, fauna of moss‐covered bedrock substrates displayed no significant differences between streams in terms of any functional group, or in terms of abundance and biomass of individual taxa. Furthermore, secondary production on bedrock outcrops remained similar between streams. Results strongly suggest that food webs of bedrock habitats are not as closely linked to immediate allochthonous inputs from the surrounding forest as those of mixed substrates, although they may be linked in the long term. Using a 9‐yr record, we determined the relationship between organic matter standing crop and invertebrate abundance, biomass, and production in the treatment stream. There is a strong relationship between leaf litter standing crops and secondary productivity in mixed substrate habitats. In contrast, bedrock outcrop fauna showed a stronger relationship to fine benthic organic matter (FBOM) standing crop than to leaf litter. Despite a large residual mass of organic matter in the treatment stream, as well as other ongoing sources of input (dissolved organic matter, soil organic matter, and throughfall), litter exclusion clearly shows strong bottom‐up effects extending from primary consumers to predators. Abundance, biomass, and production of predators in the treatment stream were also strongly related to that of their prey throughout all pretreatment and treatment years. Our results suggest that, owing to simple bioenergetic efficiencies, production of predators is constrained by productivity of their prey. Although our results clearly show strong bottom‐up effects, the importance of predators (top‐down) can not be ignored, as available data suggest that predators consume most benthic invertebrate production. Our study underscores the importance of terrestrial–aquatic linkages in maintaining productivity of headwater streams draining forested catchments.