Arthropods in aquatic ecosystems include a variety of insects in different trophic levels: grazers like chironomids (Diptera); scrapers and detritivorous larvae of mayflies (Ephemeroptera), stoneflies (Plecoptera) and caddisflies (Trichoptera); scavenger beetles (Coleoptera: Hydrophilidae); and predators such as damselflies and dragonflies (Odonata), alderflies and dobsonflies (Megaloptera), crane flies (Diptera: Tipulidae), backswimmers (Hemiptera: Notonectidae), water scorpions (Hemiptera: Nepidae), water striders (Hemiptera: Gerridae) and diving beetles (Coleoptera: Dytiscidae).
3.1. Insecticides
Insecticides are very toxic to all aquatic arthropods, particularly to zooplankton and insect larvae
[124][125]. Residues of insecticides in streams are at higher levels and more common near agricultural fields than in rivers and lakes, as these chemicals dissipate rather quickly from water. The exception are systemic insecticides, which are water-soluble and persistent in all these media
[126].
Impacts of insecticides on aquatic communities are usually due to direct effects on zooplankton crustaceans
[127][128][129] and insect larvae
[130][131], all of which are very susceptible to neurotoxic chemicals. However, the reduction in grazer arthropods indirectly boost the growth of producers such as algae and periphyton
[132][133][134], and also allows tolerant herbivore species of copepods, worms and molluscs to thrive
[135][136]. Moreover, direct reductions in predatory insects such as water bugs and dragonfly nymphs typically lead to increases in the abundance of their prey, i.e., tadpoles and snails that benefit from this indirect effect
[11]. Indeed, dragonfly nymphs pose a significant threat to amphibian larvae in aquatic communities and are capable of reducing tadpole biomass by >80% in a two-week period, so when the dragonflies were eliminated by the insecticide endosulfan in controlled mesocosms the tadpoles increased in numbers
[137].
Mixtures of insecticides can have devastating effects on susceptible benthic arthropods and zooplankton species. For example, a mixture of permethrin, λ-cyhalothrin and chlorpyrifos reduced dramatically the density of the waterflea
Daphnia magna (Crustacea: Cladocera) and the amphipod
Hyalella azteca (Crustacea: Amphipoda) within 24 h after application
[135]. The lethal effect of four insecticides (chlorpyrifos, diazinon, endosulfan and malathion) applied at 10 or 40 μg/L on waterfleas and copepods induced trophic cascades that facilitated algal blooms and abiotic changes. While the effect of the OPs was swift, endosulfan produced a lag effect by reducing the abundance of waterfleas and amphibians. The mixture treatment had lethal effects throughout the community that led to long-term effects on amphibian mass and unique indirect consequences on phytoplankton and water quality variables such as pH, dissolved oxygen and turbidity
[138].
The magnitude of the impacts depends largely on the properties of the individual chemicals
[139]. For example, malathion and carbaryl have the same effects on aquatic communities, but effects of the former are more persistent than those of the latter because the toxicity of OP insecticides last longer than that of carbamates
[140]. Concentrations of chlorpyrifos as low as 1 μg/L can significantly reduce the zooplankton communities, although the impacts are lower under higher temperatures due to the dissipation of the insecticide
[141]. Structural characteristics of the aquatic bodies can also modify the overall impacts of insecticides. For example, using a factorial design in mesocosms, Brogan et al.
[142] showed that macrophyte density protected zooplankton and other animal taxa against malathion spraying effects. Thus, species richness and abundance of spiders along Romanian streams were negatively associated not only with in-stream pesticide toxicity but also with the shading of the stream bank due to vegetation cover, a proxy for the quality of the habitat
[143].
Combined direct and indirect effects often result in long-term changes at the community and ecosystem level of organisation even after a single pulse of non-persistent insecticides
[136], but more severe indirect effects and longer recovery periods of the affected populations occur under repeated applications
[144][145]. Such effects reduce energy transfer efficiency, elongate the food chain and sometimes increase species richness
[127]. Eventually, the continuous disturbance of the aquatic food webs can ultimately lead to the collapse of entire fisheries that depend on invertebrate food sources, as it occurred in Lake Shinji (Japan) when imidacloprid residues that drained into the lake over the years eliminated the insects and zooplankton
[146].
It appears that release from competition among species with different sensitivity is the major indirect effect of insecticides. For example, the elimination of sensitive grazing waterfleas (Cladocera) often results in increases in copepods and Rotifera species due to a release from competition with waterfleas
[127][128][129]. Similarly, the elimination of various predatory arthropods after treatment with fenvalerate allowed the worm
Stylaria lacustris (Oligochaeta) to increase in abundance
[136]. Dinotefuran applied to experimental rice mesocosms at the recommended rates (10 kg/ha) increased the abundance of some insects, particularly chironomid larvae and nymphs of the dragonfly
Crocothemis servilia (Odonata: Libellulidae) due to a combined indirect effect from the lack of competition with other dragonfly species and increased chironomid prey
[147]. In this context, it is important to know the individual species sensitivities for predicting the possible indirect effects of insecticidal compounds. Some indicator taxa are water striders,
Gerris lacustris (Hemiptera: Gerridae), dragonfly nymphs of
Orthetrum albistylum (Odonata: Libellulidae) and diving beetles such as
Hydroglyphus japonicus (Coleoptera: Dytiscidae) for exposures to clothianidin, fipronil or chlorantraniliprole, respectively
[148].
Compensatory effects occur due to the differential toxicity and tolerance of individual species in the ecosystem. For example, the decrease in numbers of
Streblocerus pygmaeus (Crustacea: Cladocera) by chlorpyrifos was compensated by increases in a more tolerant congener such as
Dunhevedia crassa (Crustacea: Cladocera)
[145]. The influence of intraspecific competition on a detritivorous caddisfly,
Limnephilus lunatus (Trichoptera: Limnephilidae) was demonstrated in mesocosms treated with the pyrethroid fenvalerate. The compensation of direct effects by the chemical at high concentrations was due to a reduction of intraspecific pressure in the population, whereas at low concentrations (≤0.1 μg/L) the effects of the toxicant were not compensated
[149]. Fenvalerate also caused high mortality in populations of
D. magna in a microcosm, but the waterfleas recovered quickly in both numbers and biomass within 2 weeks; nevertheless, the recovery was faster and achieved higher biomass under low intraspecific competition than under high competition
[150].
Leaves and other vegetable matter that fall into streams and rivers are consumed by a large array of detritivorous arthropods, including larvae of mayflies, stoneflies and caddisflies as well as amphipods. Senescent leaves of trees that had been treated with the systemic insecticide imidacloprid to control the Asian longhorn beetle,
Anoplophora glabripennis (Coleoptera: Cerambycidae), contained residue levels sufficient to reduce significantly the natural decomposition processes carried out by the stonefly
Pteronarcys dorsata (Plecoptera: Pteronarcyidae) and nymphs and the crane fly
Tipula sp. (Diptera: Tipulidae) in aquatic microcosms, even if the insects did not die
[151]. However, higher concentrations of imidacloprid in water (15 μg/L) led to the starvation of the amphipod
Gammarus pulex (Crustacea: Amphipoda) by directly inhibiting its feeding on the litter
[152]. In microcosms treated with thiacloprid,
G. pulex was able to feed on leaf litter and prey on nymphs of the mayfly
Baetis rhodani (Ephemeroptera: Baetidae) up to 1 μg/L; however, at higher concentrations (4 μg/L) the amphipod reduced its leaf consumption and stopped predation on the mayflies
[153]. Thus, sublethal impacts of such insecticides on detritivorous arthropods can result in impairment of trophic relationships and reductions in the decomposition and nutrient recycling processes in aquatic systems
[154].
Another indirect effect is the enhancement of insecticidal effects by predation. For instance, the insecticide chlorpyrifos at 1 μg/L directly reduced the biomass of herbivorous plankton (4 waterflea species) by 7–12% in microcosms, with the smallest decrease when all species were present. The introduction of a predatory glassworm,
Chaoborus obscuripes (Diptera: Chaoboridae) increased the total impact of the insecticide on the waterfleas
[18]. Apart from the direct effects of imidacloprid on benthic macroinvertebrate assemblages
[155], exposure of the caddisfly
Sericostoma vittatum (Trichoptera: Sericostomatidae) and the midge
Chironomus riparius (Diptera: Chironomidae) to sublethal levels of this insecticide also compromised antipredator behavioural responses in both insect species
[156]. While chlorantraniliprole reduced the decomposition of leaves carried out by the shredder caddisfly
S. vittatum, as well as the growth of midge
C. riparius, these effects were enhanced in the presence of predation by nymphs of the golden-ring dragonfly,
Cordulegaster boltonii (Odonata: Cordulegastridae)
[157]. The combined direct and indirect effects of sublethal concentrations of pesticides can have negative consequences in terms of mortality from predation in benthic insect populations, and also induce maladaptive responses among zooplankton species which may reduce their long-term viability in the field
[158].
3.2. Fungicides
Fungicides are very toxic to arthropods and other aquatic organisms. In European surface waters, fungicides dominate the residue levels (median 0.96 μg/L) compared to herbicides (median 0.063 μg/L) and insecticides (median 0.034 μg/L). A comprehensive review of the risks that such contamination poses to aquatic ecosystems can be found in Zubrod et al.
[159].
A commonly reported indirect effect of fungicides is an increase in phytoplankton abundance or biofilm biomass. This could be due to (a) reduced grazing pressure by affected invertebrate consumers, mainly zooplankton species and chironomids or (b) microorganisms benefiting from the metabolization of organic material set free by dying organisms. In some studies, interactions within microbial communities result in increasing diatoms and decreasing protozoans
[160].
For example, mesocosms treated with pentachlorophenol at 121 μg/L increased the abundance of the alga
Cryptomonas sp. (Cryptophyceae) due to reduced grazing pressure, reduced competition, or increased decomposition of the fungicide. These effects were evident also at lower treatment levels in autumn but not in winter
[161]. In Thailand, mesocosms treated with carbendazim produced blooms of the floating macrophyte
Wolffia sp. (Alismatales: Araceaae) due to the reduction or elimination of zooplankton rotifers (
Keratella tropica), waterfleas (
Moina micrura, Ceriodaphnia cornuta and
Diaphanosoma sp.) and cyclopoid copepods. These changes resulted also in altered water conditions, which became anoxic during the last three weeks of the experiment
[162]. The dithiocarbamate fungicide metiram applied to microcosms up to 324 μg/L also produced increases in phytoplankton due to the reduction in densities of grazing rotifers and copepods, as these were the most sensitive taxa towards the fungicide
[163].
Another common indirect effect is the increase in abundance of tolerant macroinvertebrates, which results from either reduced predation pressure or release of competition with species in the same trophic guild that are more susceptible to the fungicide, or both. In microcosms treated with the dithiocarbamate fungicide metiram at various concentrations (4 to 324 μg/L), population densities of a few macroinvertebrates increased in the short-term, although they were not consistent with the concentrations used. This unexpected outcome was likely due to shifts in species interactions as a result of direct toxic effects of the fungicide on susceptible species such as predatory beetles (Dytiscidae) and
Caenis sp. mayflies (Ephemeroptera: Caenidae)
[163]. Similarly, a concentration of carbendazim at 1000 µg/L applied in microcosms resulted in increases of the snail
Lymnaea sp., as the flatworm predator
Girardia tigrina was reduced in numbers together with macroinvertebrate herbivores that exploited the same niche (i.e., oligochaetes, the crustaceans
Gammarus juvenile and
G. pulex and the molluscs
Bithynia tentaculata and
B. leachi)
[164]. The same indirect effect was observed in tropical mesocosm treated with this fungicide
[162].
A third indirect effect of fungicides on aquatic ecosystems is the alteration of saprophytic function in the aquatic ecosystem, carried out mainly by detritivorous crustaceans and larvae of mayflies, caddisflies and stoneflies, as well as by bacteria consortia
[165][166]. As with insecticides, this effect is the consequence of the direct reduction of populations of these benthic invertebrates due to the toxicity of the fungicides. Measuring this functional disturbance can only be done in mesocosms and microcosms. For example, the fungicide thiram at 35 and 170 μg/L applied to stream mesocosms and ponds reduced the overall litter break-down a few weeks after the treatment, as the detritivorous asellids (Isopoda) and gammarids (Amphipoda) were eliminated in large numbers over that period
[167]. Similarly, the reduced decomposition of banana leaves observed eight weeks after application of carbendazim in tropical mesocosms was considered to be the indirect effect of a decreased microbial activity that resulted from the anoxic water conditions created by algae blooms
[162]. This effect may be also due to avoidance of the litter that contains fungicide residues, as several other mesocosm studies indicate. In choice experiments with leaves of black alder (
Alnus glutinosa) treated or not with tebuconazole at 50 and 500 μg/L,
Gammarus pulex (Crustacea: Amphipoda) significantly preferred untreated leaves over those treated with the fungicide. It appears that the fungicide eliminated fungal species such as
Alatospora acumunata, Clavariopsis aquatica or
Flagellospora curvula, which are preferred by the amphipod
[168]. Indeed, the nutritional quality of the litter leaves matters to the shredders to the extent that in another experiment, that amphipod and the caddisfly
Halesus radiatus (Trichoptera: Limnephilidae) consumed significantly more leaves that had lower microbial biomass as a result of having been treated with the fungicide propiconazole, in order to compensate for the reduced nutritional quality of the litter
[169]. In another experiment,
G. pulex showed a preference for black alder leaves that were treated with a mixture of fungicides (azoxystrobin, cyprodinil, quinoxyfen and tebuconazole at recommended field rates), probably because this treatment reduced pathogenic fungi while allowing the growth of other microbial and fungal communities that were more palatable to the amphipods. This shift in fungal community composition and increased nutritional quality and palatability of leaf material ultimately resulted in a higher gammarid growth (up to 300% increase) during a 24-day long-term feeding assay
[170].
3.3. Herbicides
Herbicides are the most common pesticides found in freshwaters
[171]. While they are not as toxic to animals as insecticides and fungicides are, the presence of a large cocktail of plant killers in aquatic ecosystems cannot be overlooked, as algae and communities of aquatic macrophytes can be seriously affected.
Two main indirect effects of herbicides have been observed in experimental ecosystems. The first is a decrease in the abundance of grazing invertebrates as a consequence of a reduction in phytoplankton and/or periphyton by the direct herbicidal toxicity. For example, the herbicide terbutryn applied to microcosms up to 6 μg/L eliminated the periphyton food source of the grazer mayfly
Rhithrogena semicolorate (Ephemeroptera: Heptageniidae), which larvae decreased significantly in numbers when compared to untreated controls
[172]. Atrazine in experimental ponds at various concentrations decreased the abundance of chironomids and other herbivorous insects presumably through reduction of the periphyton food source and, to some extent, their habitat. The abundance of detritivorous insects such as the caddisfly
Oxyethira pallida (Trichoptera: Hydroptilidae) was not affected, although early emergence was observed, and populations of predatory insects were not affected. However, the species richness and evenness of the pond was reduced
[173]. Microcosms treated with the herbicide linuron at 100 μg/L reduced the biomass of five species of algae by 17%, and a further reduction of 42% was observed when four species of herbivorous zooplankton were introduced into the system. The highest effects occurred after 6 days and then declined as the herbicide dissipated and the algae recovered
[18]. Interestingly, the total impact of the herbicide in the system was mitigated with the introduction of the predatory glassworm
Chaoborus obscuripes (Diptera: Chaoboridae), which alleviated the grazing pressure on the phytoplankton and allowed it a fast recovery. This effect has also been observed in field situations. Typical combinations of herbicides applied routinely to rice fields in Japan have been reported as reducing the abundance of grazing worms (Oligochaeta) and other herbivorous invertebrate taxa, as their periphyton food source is almost eliminated by the direct action of the herbicides. The only taxon that does not appear to be affected by the herbicides is midges (Chironomidae), perhaps because they can also feed on periphyton. However, as in the microcosm experiments, further reductions in both worms and chironomids occur when predatory insects are present in the fields
[174].
Another indirect effect is the reduction of available refugia for predatory insects, as herbicides can reduce significantly the biomass of macrophytes
[173]. This effect is more subtle and not as evident as the former but has been demonstrated in mesocosms with macrophytes treated with the herbicide pentoxazone. The herbicide did not affect the phytoplankton and consequently, there were no clear negative impacts on zooplankton nor on herbivorous and detritivores insects. However, as the herbicide significantly reduced the biomass and surface cover of the aquatic plants, nymphs of the dragonfly
Orthetrum albistylum (Odonata: Libellulidae) were notoriously absent from the treated plots. Lack of protection by the aquatic plants resulted also in smaller decreases in abundance in other insect predators
[175]. In contrast, mesocosms with dense stands of aquatic plants reduced the abundance of periphyton and hence lowered the numbers of grazing snails and tadpoles
[142]. Differences in macrophyte density and associated invertebrate communities between channels treated or not with herbicides had already been reported in agricultural settings
[176].
3.4. Pesticide Mixtures
Mixtures of pesticides can have compensatory effects or else result in additive negative impacts. Ecological theory can help predict the direction of effects of multiple chemical mixtures by integrating information on each functional group’s (1) sensitivity to the chemicals (direct effects), (2) reproductive and recovery rates, (3) interaction strength with other functional groups (indirect effects) and (4) links to ecosystem properties
[177]. The initial composition of the community usually influences the direction of the combined pesticide effects.
An example of additive negative effects is when the herbicide atrazine is applied together with the insecticide terbufos in a microcosm. The herbicide at 15 μg/L reduced the algae density significantly and indirectly led to a reduction in waterfleas and chironomids abundance, whereas terbufos (at 0.1 or 10 μg/L) directly suppressed the latter two taxa due to its high toxicity
[178]. Equally, in mesocosms treated with atrazine (25 μg/L) and endosulfan (10 μg/L) in two pulses, atrazine directly decreased periphyton and this effect indirectly reduced chironomid abundance, while endosulfan reduced chironomid larvae dramatically, which resulted in indirect increases of snails and decreases in competing tadpoles
[12]. Effects may not be always noticeable when the pesticides in water are present at sublethal concentrations to both the algae and the grazers, as it often occurs with the herbicide diuron and the insecticide imidacloprid. The combination of these two pesticides at relevant environmental concentrations (5 μg/L) in microcosms, however, altered the nutritional behaviour of chironomid larvae, with imidacloprid leading to an inhibition of grazing activity, while diuron provoked a nutritional quality loss of algae which probably affected their palatability
[179].
Finally, a comprehensive mesocosm study on the impacts of 2 insecticides (carbaryl and malathion) mixed with 2 herbicides (glyphosate and 2,4-D) on non-target organisms was carried out after application at the recommended commercial rates. Among the 25 species in the mesocosms were six predatory insects: water bugs
Notonecta and
Belostoma (Hemiptera) and nymphs of
Anax and
Tramea dragonflies (Odonata) that prey on both tadpoles and snails, and larval predatory beetles (Coleoptera: Dytiscidae) such as
Dytiscus sp. that prey on tadpoles and
Acilius sp. that prey on zooplankton. As it could be expected, the insecticides reduced the diversity and biomass of zooplankton and predatory insects, which indirectly increased the abundance of several species of tadpoles due to a lack of predation. The two herbicides, however, did not reduce the periphyton biomass and had no effects on zooplankton, insect predators, nor snails
[129].