Discussion
The immediate impact of rotenone treatments in August 2011 and 2012 was minor for benthic invertebrates in the rotenone treated River Fusta. Only a few taxa showed a marked decrease in abundance in the treated river compared to the untreated River Drevjaelva, including the Ephemeroptera Baetis rhodani and the Plecoptera generaDiura and Isoperla . Because the upstream Lake Fustvatnet was treated two months after the last river treatment, River Fusta also received rotenone from the lake. Still, the immediate effect of the lake rotenone treatment on benthic invertebrate community in the downstream river was minor. One exception was the Trichoptera Rhyacophila nubila, which occurred at high abundances in both rivers, but which disappeared from the River Fusta subsequently to the treatment of the upstream lake. All the above-mentioned taxa are known to be rotenone sensitive (Engstrom-Heg et al., 1978, Arnekleiv et al., 2001, Gladsø & Raddum, 2002) and negative impacts were expected.
The rotenone treatment of Lake Fustvatnet took place in October, just before the lake ice formed. Breakdown of rotenone was thus slowed down due to low temperatures and limited light. Consequently, the treated river was exposed to rotenone from the lake for several months. Even though the toxic effect of rotenone decreases with decreasing water temperatures (Meadows, 1973, Kjærstad et al. 2015), the long exposure apparently had a pronounced negative effect. In June 2013, eight months after the lake treatment, we recorded the lowest abundance of benthic invertebrates in the treated river. A similar decrease did not occur in the untreated river. The benthic invertebrate community in the treated river at this time were dominated by Chironomidae, Oligochaeta, Nematoda and elmid beetles. Other taxa occurred at very low abundances and several taxa were missing, including rotenone tolerant species like the Ephemeroptera Ephemerella . Moreover, despite that the freshwater pearl mussel (Margaritifera margaritifera ) is known to be rotenone tolerant (Dolmen et al., 1995), all known specimens in the treated river died during the spring of 2013 (Larsen, 2015). The temporal beta diversity confirms that the changes in taxa composition were by far highest, and significant, in the treated river from October 2012 to June 2013. This decline in benthic invertebrates was mainly due to losses both in terms of taxa abundances and presence-absence. This indicates that rivers may receive rotenone for several months following treatment of upstream lakes, which can have major impacts. Most taxa in the treated river, including rotenone tolerant taxa, decreased to very low abundances and some disappeared, like the freshwater pearl mussel.
Despite the strong negative impacts recorded in the treated river eight months after the lake rotenone treatment, recolonization was relatively fast. The highest gains in terms of species occurrence in the treated river occurred in August 2013. Already in August 2014 the total abundances exceeded pre-treatment levels. In August and October 2016, the change in taxa composition, in terms of abundance and presence-absence of taxa between sampling occasions, were rather similar for the treated and the untreated river. This indicates that the benthic fauna of the treated river had recovered to a great extent. Several taxa, like the Ephemeroptera Heptagenia dalecarlica andEphemerella , the Plecoptera Isoperla and Amphinemura borealis and Oligochaeta reached much higher abundances in 2013, as compared to pre-treatment levels. Ephemeroptera and Plecoptera are known to be among the most abundant groups in the drift of swift-flowing temperate streams (Brittain & Eikeland, 1988). These groups likely colonized rapidly from drift in untreated tributary streams, although some may have survived as eggs. Oligochaeta may have avoided rotenone exposure by digging deeper into the substratum. It took almost four years after the last rotenone treatment before most of the taxa had recolonized. The Plecoptera Taeniopteryx nebulosa were among the slowest to recolonize and appeared in 2016. Also, two rotenone tolerant taxa, the Ephemeroptera Ephemerella and elmid beetles did not reached pre-treatment abundances in the treated river four years after the treatments.
According to Vinson et al. (2010), the overall invertebrate abundances generally return to pre-treatment levels quicker than the biodiversity and taxonomic composition measures. This is in accordance to our findings. However, the recolonization will greatly depend on the taxonomic resolution of the data. Biodiversity measured at family and genera level will recolonize fast, while it may take longer before all species to recolonize. We have identified some taxa to species level, such as Ephemeroptera, Plecoptera and Trichoptera, and some taxa to a higher taxonomic level, such as the specious Chironomidae and Hydrachnidia. Thus, changes in abundance and recolonization patterns of many species in this study remain unknown.
A reduction of total benthic invertebrate abundance was observed both in the treated lake and untreated lake immediately after the lake treatment, but to a higher extent in the treated lake. This indicates both natural changes and a negative effect of rotenone. Taxa with clearly decreasing abundances shortly after the treatment in Lake Fustvatnet, but not in the untreated Lake Drevvatnet, included the Ephemeroptera Centroptilum luteolum and the PlecopteraCapnia sp. While Capnia is known to be rotenone sensitive,C. luteolum is quite tolerant (Kjærstad & Arnekleiv, 2011). The reason for the seemingly high sensitivity of C. luteolum to rotenone in the treated lake is unknown.
The recolonization in terms of abundance of benthic invertebrates in the treated lake was fast, reaching pre-treatment levels eight months after the treatment. A few increased in abundance after the treatment in the treated lake, but not in the untreated lake. This was particularly evident for the diving beetles (Dytiscidae). Beetles are generally known to be very rotenone tolerant (Engstrom-Heg et al., 1978). Some will also benefit from increased prey availability because of the general increase in invertebrate abundance after the recolonization. Even though most taxa increased in abundance during the years following the treatments, the Plecoptera Capnia , the Trichoptera Polycentropodidae and the Gastropoda Gyraulus acronicus decreased and had not reached their pre-treatment levels in the treated lake four years after the treatment. Considering that Gastropoda are rotenone tolerant (Holocombe et al., 1987, Arnekleiv et al., 1997), the decline of G. acronicus in the treated lake was surprising. Mangun & Madrigal (1999) found that of several taxa missing five years after rotenone treatments, most belonged to rotenone sensitive groups, like Ephemeroptera, Plecoptera and Trichoptera. Taxa that were tolerant to rotenone were identified as” non-missing“ or ”briefly missing“ (Mangun & Madrigal, 1999).
Fish may play an important role in structuring benthic invertebrate communities (e.g. Weyl et al., 2010, Alexiades & Kraft, 2017). Stocking of Atlantic salmon (S. salar ) took place in the treated river between 2013 and 2016, whereas Arctic charr (S. alpinus ) and brown trout (S. trutta ) were stocked in the treated lake in 2014. A few benthic invertebrate taxa did not reach pre-treatment abundance. However, if fish was the main driver, we would expect a decrease in abundance of other common taxa that are eaten by fish in the years after the treatments. Instead, many taxa increased in abundances and exceeded pre-treatment levels, indicating a relatively low predation pressure from fish in the treated river. In the treated lake, contrary to the untreated lake, we observed a strong decrease in total abundances of benthic invertebrates in August 2015. This was most likely due to predation from fish stocked during the previous year. The decrease was particularly high for Chironomidae and Ephemeroptera. Abundance of large sized and important fish prey, like the Ephemeroptera Ephemera vulgata , increased to a high abundance in 2013-2014, but decreased drastically in 2015 in the treated lake, probably due to predation from fish. The species was not found before the treatment, but occurred in all sampling events after the treatment, indicating a lower predation pressure from fish after the treatment. Despite the rotenone treatments and the fish removal/fish stocking, the abundance of many taxa varied to a similar extent in the impact and control sites. As an exception, the effect was more prominent in the impacted treated river than in the untreated river after the lake treatment and subsequent months of rotenone exposure. This illustrates the importance of using control sites when performing time-series experiments.
It is striking to observe the large fluctuations in the abundance and in the presence-absence of benthic invertebrates over time of both treated and untreated localities. Even before the lake treatment, the invertebrate abundances in both lakes were reduced by approximately 50% from August 2011 to August 2012. According to Vinson et al. (2010), the accumulation curve for benthic invertebrate genera showed little inclination for flattening out during a ten-year monthly sampling of a Utah river. We did not see the same pattern in our data, perhaps due to fewer taxa present. Even though it will vary greatly with habitat type, latitude etc., a rising taxa accumulating curve would certainly be found elsewhere and shows how random the community composition may be registered at any given time.
Our study found minor short-term (i.e. < one month) effects of the river rotenone treatments, whereas the lake treatment caused major temporary negative effects on the lotic fauna eight months after the treatment. No major short- or long-term (i.e. four years) effects on lake benthic invertebrate communities where detected. Recovery time of invertebrate abundance after the lake treatment was fast whereas the recovery in terms of presence-absence of species were comparably much longer, lasting for several years. The current study underpins the importance of performing long-terms studies during investigation of the effects of rotenone on benthic invertebrates.