Discussion
Given that recent evidence indicates that increasing agriculture to meet the growing global food demand could simultaneously increase human infectious diseases (Rohr et al., 2019), such as human schistosomiasis, we conducted a study to identify agrochemicals that could sustainably improve food production without increasing schistosomiasis. Expanding upon previous work on predators in isolation (Halstead, Civitello, & Rohr, 2015), our study confirmed that most insecticides pose a high toxicity risk to snail predators at environmental concentrations in simulated aquatic communities. We identified that, within their pesticide class, the organophosphate malathion and the pyrethroid λ-cyhalothrin had the lowest risks of increasing Bu. truncatussnails that transmit urinary schistosomiasis in Africa. Bottom-up effects of fertilizer on snail biomass in natural communities occurred as fertilizers increased both snail food (periphyton) and habitat (aquatic vegetation), whereas herbicides decreased snail habitat (experiment one). Thus, fertilizer bottom-up effects are expected to be generally stronger than those for herbicides in natural communities. Positive bottom-up effects on snails for the herbicides atrazine and acetochlor were found in the absence of aquatic macrophytes and snail predators (experiment two), whereas some herbicides, such as alachlor, were negatively associated with snails under the same conditions. We found that snail biomass positively predicted their cercarial production, suggesting that changes in snail biomass associated with agrochemical pollution could have strong potential to influence human exposure to schistosome cercariae. Our findings highlight that maximizing the benefits of agricultural intensification, while minimizing indirect costs on human health, will require identifying bottom-up and top-down effects of agrochemicals within natural aquatic communities where the snails that transmit schistosome parasites are found.
Initial (24hr) mortality of the crayfish Procambarus alleni in our mesocosm experiment generally matched expectations from previous lab experiments, with the exception of the toxicity of some pyrethroids being lower in outdoor mesocosms. We confirmed previous findings that malathion is a low risk organophosphate, potentially because of its very high LC50 values for P. alleni (Halstead et al., 2015). Halstead et al. (2015) reported that all three of the pyrethroids that we tested were highly toxic to P. alleni . However, we found that λ-cyhalothrin, unlike other pyrethroids, did not lead to top-down effects on snails in our natural communities when applied at concentrations approximating the 96h LC50 for P. alleni (Halstead et al., 2015). One potential reason for this result may be that submerged vegetation, such as the H. verticillataused in our study, has been linked to reduced toxicity of pH-sensitive insecticides, including malathion and λ-cyhalothrin (Brogan & Relyea, 2014; Leistra et al., 2004). While the organophosphate chlorpyrifos and the pyrethroid esfenvalerate also exhibit pH-sensitive hydrolysis (Hertfordshire, 2013), they were likely applied at sufficiently high concentrations in our study (2x and 10x their respective 96h LC50 estimates) to preclude any increased degradation from mitigating acute toxicity to P. alleni . Thus, our findings suggest that, of the insecticides we tested, malathion and λ-cyhalothrin should represent the weakest top-down effects on snails in natural communities with aquatic vegetation, and the lowest risk of increasing snail biomass.
Fertilizers can increase submerged vegetation that acts as snail habitat, potentially making fertilizers more likely to have positive bottom-up effects on snails than herbicides and even modifying top-down effects. Unlike other agrochemicals, fertilizers are both commercially available and locally sourced via manure products, leading to their wide application even by rural subsistence farmers. Consistent with previous mesocosm experiments, we found that fertilizers can increase snails by increasing attached algae that snails eat (Halstead et al., 2018). However, we also documented that stronger bottom-up effects of fertilizer exist by increasing the submerged vegetation that serves as snail habitat. By increasing H. verticillata , fertilizers indirectly benefited Bu. truncatus biomass by increasing both snail reproduction and growth. In contrast, the same associations were negative for Bi. glabrata , likely because of resource depletion in some tanks given that Bi. glabrata infection rates and biomass were both an order of magnitude greater than for Bu. truncatus . In early weeks, algal food resources for snails fueled by fertilizer would be high and available for snail reproduction, growth, and parasite production by infected snails (Civitello et al., 2018). As Bi. glabrata populations quickly grew, per-capita resources in tanks decline such that snail populations crash as periphyton resources become limiting in microcosms (Civitello et al., 2018; Rohr, Halstead, & Raffel, 2012). Thus, aquatic vegetation likely fueled snail reproduction for both snail species, but Bi. glabrata biomass outpacedBu. truncatus biomass, likely increasing the likelihood of population crashes (negative associations) toward the end of our experiment. We know of no available studies documenting that such resource limitations occur in aquatic systems where schistosomiasis is transmitted to humans. In fact, available field studies suggest that aquatic vegetation is highly beneficial to both Bu. truncatus andBi. glabrata populations in natural systems and to their parasite production (Haggerty, Bakhoum, Civitello, et al., 2020).
In our first experiment, presence of H. verticillata in the tank provided an alternative resource for crayfish (Lodge, Kershner, Aloi, & Covich, 1994), in addition to providing potential refugia and increased resource availability for snails (Bronmark, 1989). Crayfish of the genera Procambarus and Oronectes have been associated with declines in both submerged macrophytes and snails (Lodge et al., 1994; Rodríguez, Bécares, & Fernández-Aláez, 2003). Reductions in macrophytes by crayfish occur through both consumptive and non-consumptive removal of plant biomass (Lodge et al., 1994; Rodríguez et al., 2003), while reductions in snails are primarily consumptive (Hobbs, Jass, & Huner, 1989). We found that crayfish were negatively associated with H. verticillata cover, consistent with crayfish consuming this macrophyte. However, submerged plants actually were associated with reduced final crayfish counts, possibly by providing refugia for snails thus reducing crayfish predation rates on this primary protein source. In addition to providing potential refugia from snail predators, macrophytes provide increased surface area for epiphytic periphyton growth, and grazing of epiphytes by snails can, in turn, increase macrophyte growth (Bronmark, 1989; Lodge et al., 1994; Schmidt, Koehler, & Alfermann, 2011; Underwood, 1991). Nutrients, particularly ammonia, released by snails can also facilitate macrophyte growth (Underwood, 1991). Thus, aquatic vegetation can increase snails by both providing resources to snails and reducing predation rates on snails. In contrast to the potential for fertilizers to foster the mutualism between snails and vegetation, we found that effects of herbicides on snails in natural communities were generally weaker, potentially due to a negative association between herbicides and submerged plants in experiment one.
Herbicides generally had negative effects on snails by reducing aquatic vegetation, with positive bottom-up effects of atrazine and acetochlor only in the absence of aquatic vegetation and snail predators. The positive association that we found between some herbicides, such as atrazine, one of the most common herbicides globally, and snails in our second experiment is consistent with previous studies (Halstead et al., 2018; Rohr et al., 2008). Halstead et al. (2018) actually detected positive atrazine effects while controlling for crayfish density. However, Halstead et al. (2018) did not include submerged aquatic vegetation, which likely increased the chances of detecting bottom-up herbicide effects because it omits the negative herbicide effects on snail habitat that we found in experiment one. Atrazine and other herbicides can increase snails by killing phytoplankton that competes with periphyton algae eaten by snails (Halstead et al., 2018). Instead of such mechanisms leading to consistent positive effects on snails among herbicides, we found positive and negative effects in our study for acetochlor and alachlor, respectively. As algae absorb available nutrients, both periphyton and snails that consume these nutrients, eventually hit a carrying capacity after which their populations can decline or even crash (Rohr et al., 2012). Herbicide effects can be negative overall when considering only data points at the end of the experiment after snail populations exceed carrying capacity and populations subsequently crash (Rohr, 2017). Thus, we think that variation in snail population growth among tanks likely influenced alachlor and acetochlor associations with dead and total snails, respectively, when measured at the end of our second experiment. Nonetheless, our results suggest that fertilizers are far more likely than herbicides to have positive bottom-up effects on snails, and especially in communities where snail predators are present. Our finding that fertilizers can increase snail habitat and resources is consistent with field observations in sub-Saharan Africa that revealed positive associations between fertilizer use and snail habitat, snail abundance, and human schistosomiasis infections (Haggerty, Bakhoum, Chamberlin, et al., 2020).
The importance of other functional groups, such as aquatic vegetation, in mediating the indirect effects of contaminants on snails that transmit the parasites causing schistosomiasis in humans underscores the need to include these relationships when assessing environmental risk of contaminants. This becomes even more critical when assessing the risk of contaminant mixtures. The combined indirect effects of multiple contaminants can either mitigate or exacerbate the predicted effects of a single-contaminant scenario (Halstead et al., 2014). Our study inferences are restricted to observing chemicals individually and we could not simulate all biotic components of the aquatic communities that occur in transmission sites that might influence snail responses to agrochemicals. For example, eutrophication from long-term fertilizer inputs could cause unpredictable changes to snail populations or species richness, potentially eliminating important functional redundancies that can alleviate losses in functionally-important sensitive species (Hooper et al., 2005). Thus, the extent to which changes in biodiversity and abiotic conditions can mediate indirect effects of contaminants (either individually or as mixtures) on Biomphalaria and Bulinussnails warrants further study. This is especially true given that changes in snail biomass associated with agrochemicals that we found appear very likely to change parasite production. Further, human co-infection by parasites released by Biomphalaria andBulinus snails is not uncommon in Africa.
Overall, our findings suggest the need to identify low risk insecticides, such as malathion, that could reduce crop pests without increasing snails. For example, our findings recommend against pesticides that harm crayfish, because these predators reduce transmission directly by consuming snails and indirectly by consuming snail habitat. The more consistent and stronger positive impact of fertilizers on snails than herbicides, by increasing both their food and habitat, highlights that the removal of aquatic vegetation could offer important public health benefits by limiting its bottom-up effects on schistosomiasis risk.