Midge subsidy has long-lasting effect on terrestrial plant communities

Background/Question/Methods

Aquatic insects can be an important conduit of resources moving from freshwater to riparian habitats that subsidize terrestrial food webs.  However little is known about the cumulative effects of several years of aquatic insect deposition to land.  Lake Mývatn in northern Iceland has extraordinary midge (Chironomidae) emergences that result in large inputs of biomass and nutrients to terrestrial communities.  In two complementary experiments using paired plots, we examined the effects of midges on riparian plants over 5 years by either (1) adding midge carcasses or (2) blocking midge inputs.  1.  We added midges (150 g (dw), equivalent to ~15 g of N) to 1 m² plots at a site with low natural midge deposition.  2. We blocked midges in 4 m² plots at a site with high natural midge deposition, effectively eliminating midge deposition within exclusion plots.  We measured plant chemistry (C:N, %N, stable isotopes: δ¹³C, δ¹⁵N), primary productivity (Leaf Area Index, LAI) and soil nutrients (using resin bags), in paired control and midge-exclusion/addition plots.

Results/Conclusions

At our midge-addition experiment, ion exchange resin bags confirmed that plant available N in the soil was >5x higher in midge-addition than control plots.  We also observed significant changes in the plant community.  Grass was present in all plots at low density, but grass biomass increased dramatically following midge addition.  LAI increased by 50% in the year following midge addition.  In the years following a one-time midge addition, midge-addition plots were visually distinguishable from paired controls based on accumulated litter and a shift toward grass dominance, though woody heathland species were still present in all plots.  The response of the plant community to midge addition is consistent with natural gradients observed along a midge-deposition gradient at various distances from high-midge lakes.

While the addition of midges resulted in a visual shift from a heathland to a grassland in our midge addition plots, the removal of midge inputs had no discernible effect on the plant community at our midge removal experiment.  Plant chemistry and soil nutrients did not differ between midge exclusion and control plots.  This may be due to the history of midge deposition at the site of our removal experiment.  Annual midge inputs for hundreds of years have resulted in accumulated soil and plant litter that can effectively store nutrients from past years.  Furthermore, well established root systems may have enabled the transport of midge-derived nutrients from outside exclusion plots.