Conservation of biological control agents, such as parasitoid wasps, in agricultural areas can aid in integrated pest management schemes. However, to successfully integrate chemical and biological control it is necessary to determine the effects of pesticides on natural enemies and balance insecticide use with conservation of natural populations of predators and parasitoids (Janssen and van Rijn 2021). Parasitoid wasps in particular can serve as good indicators of effects of pesticides, since they are particularly sensitive to pesticides compared to other natural enemies (Mates et al., 2012).

Most tests of pesticide non-target effects are done in laboratory settings, focusing on the effect of a specific pesticide on a specific species of parasitic wasp. The ecological realism of such studies is often low. Laboratory experiments generally lack potential sheltered locations where wasps could avoid direct pesticide exposure and the wasps lack the ability to escape pesticides in space (Macfadyen et al., 2014). Also, the wasp populations tested may lack genetic variability which could be protective and may be affected by conditions of laboratory rearing (Macfadyen et al., 2014). On the other hand, abiotic stressors (such as suboptimal temperature or humidity), which may interact with pesticides to exacerbate parasitoid mortality, are usually eliminated from laboratory tests. Further, toxicology assays under laboratory conditions generally test one species at a time, and do not consider community-level effects of pesticides. Field studies can serve as a more realistic model of the effects of pesticides on entire parasitoid communities. For example, a study of the effects of Deltamethrin, a pyrethroid, in wheat fields indicated a decrease in density of wasps in the center of fields compared to edges, with recovery within 12 days (Longley et al., 1997). Another study of wheat fields sprayed with Dimethoate, an organophosphate, showed a decline in wasps of the genus Aphidius in sprayed and neighboring unsprayed areas of the fields, up to 20 days after insecticide application (Holland et al., 2000). In contrast, another wheat field study where Fenvalerate, a pyrethroid, and Dimethoate were sprayed found no effect on aphid parasitoids throughout the season following pesticide applications (Giller et al., 1995). An additional study in various field crops found negative effects of the insecticides Demeton-S-methyl, an organothiophosphate, and Cypermethrin, a pyrethroid, on abundance of aphid parasitizing wasps, and less so of the insecticide Pirimicarb, a carbamate (Smart et al., 1989). While these examples suggest some potential harmful side effects of pesticides to parasitoid assemblages, communities-level field studies remain few and scattered. In particular, how parasitoid communities recover after spraying remains poorly explored.

To address this gap, here we focus on Indoxacarb use in vineyards and its impacts on the resident parasitoid assemblages. Indoxacarb is an oxadiazine insecticide used against the European grapevine moth, Lobesia botrana Denis and Schiffermüller, 1775. It is applied as a spray to control lepidopterans at the larval stage and works by blocking voltage-gated slow-inactivated sodium channels, disrupting the nervous system (von Stein et al., 2013). It is one of the most commonly applied pesticides in vineyards (Nash et al., 2010). Laboratory experiments have been used to test the effects of Indoxacarb on specific parasitoid wasp species, and it has been shown to be harmful to Cotesia plutellae Kurdjumov, 1834, a larval endoparasitoid, in the adult stage (Haseeb et al., 2004), adult Trichogramma pretiosum Riley, 1879, an egg parasitoid (Scholz and Zalucki 2000), and Aphidius colemani Viereck, 1912 adults, which parasitize aphids (Bostanian and Akalach 2004). Fecundity and longevity of females were also reduced in C. plutellae and species in the genus Trichogramma (Haseeb et al., 2004; de Paiva et al., 2018; Gallego et al., 2019). Effects on emergence after insecticide application on immature stages varied among studies. There was a reduction in emergence of Trichogramma achaeae Nagaraja and Nagarkatti, 1970 (Gallego et al., 2019) and Trichogramma cacoeciae Marchal, 1927 from host eggs (Asma et al., 2018), but no effect on Trichogramma pretiosum (de Paiva et al., 2018) and Aphidius colemani emergence from eggs and aphid mummies, respectively (Bostanian and Akalach 2004). Altogether, the effect of Indoxacarb in laboratory experiments was highly dependent on the species and stage affected, and the dose the individual receives. Laboratory tests of Indoxacarb’s effects on predatory mites and bugs also yielded mixed results (Bostanian and Akalach 2006). The consequences of Indoxacarb applications for parasitoid communities under field conditions have not been evaluated yet.

One factor which may protect from harmful effects of pesticides could be the availability of field margins. A study testing for effects of pesticides on parasitism in field margins found none (Bakker et al., 2021), suggesting that the margins may serve as a source for recolonization of fields. The possibility of migration from non-sprayed habitats has been raised by several studies (Longley et al., 1997; Mates et al., 2012; Bakker et al., 2021), and this could allow recovery of parasitoid communities in agricultural habitats. However, it is necessary to examine the dynamics of the parasitoid community over space and time to further understanding of the effects of pesticides and mechanisms of recovery.

The aims of this study are to characterize dynamics of populations of common parasitoid wasps after insecticide application in vineyards and the distribution of parasitoid wasps in crop and non-crop habitat. Specifically, we tested the effect of Indoxacarb application on parasitoid abundance and diversity, and abundance of dominant species. We asked: 1) What is the effect in vineyard center, vineyard border, and natural habitat? 2) What is the effect in herbaceous and vine vegetation in the vineyard? 3) What is the effect over the course of 2 weeks after pesticide application? 4) Do wasps migrate between the vineyard and the natural habitat in response to insecticide application?

We sampled the parasitoid communities in wine-producing vineyards in Binyamina, Israel (32°30′14″N, 34°56′30″E-32°32′12″N, 34°57′24″E) during the 2019 and 2020 grape-growing seasons. The climate at the study area is Mediterranean, with cool rainy winters and hot dry summers. Pre-spray samples were taken 0–24 days (median 9 days) before Indoxacarb application. In the first year (2019), post-spray samples were taken twice or three times in the week after spraying, and in the second year (2020) samples were taken five to six times in the 2 weeks after spraying (Supplementary Table S1). For analysis, we chose sets of samples that were taken before and after application of Indoxacarb, without other insecticides, and with sampling for at least a week after spraying, with no other insecticides applied during that time. In 2019, five vineyards were sampled, one was sprayed twice, in mid-June and late July, and was included as two replicates. In 2020, four vineyards were included. Altogether, there were ten replicates across 2 years. Further information on the vineyards’ characteristics and on additional pesticides applied to them is provided in Supplementary Tables S1, S2.

Insects were sampled using vacuuming and sticky traps. A vacuum sampler (Vortis, United Kingdom) was passed through vegetation for 15 seconds. Sampling was done in ten locations at each vineyard: At the edge of the vineyard in the vine and groundcover herbaceous vegetation, 25 m into the vineyard, also in the vine and herb, and in a natural area near the vineyard, 10 m from the vineyard. Sampling was done in two such transects (Figure 1). The dominant species in the groundcover vegetation are listed in the Supplementary Table S3; most were summer flowering annuals, about half were weeds, and the most common family was Asteraceae. Sticky traps were hung at the edge and in the center of the vine at a height of ∼1.2 m (Figure 1), for two-three days at a time. Additionally, to explore effects of spraying at a greater distance from the vineyard, sampling was done at distances of 10–50 m from one vineyard, in 10 m intervals, in the second year.

Vacuum samples were stored in 75% ethanol. Parasitoid wasps were sorted out from each sample and identified to morphospecies, to obtain species richness per sample. The most abundant species were identified to genus or species. The keys used for identification of dominant species were Doutt and Viggiani 1968; Gibson et al., 1997; Goulet and Huber 1993; Huber et al., 2009; Masner 1980.

The effects of Indoxacarb spraying on parasitoid wasp abundance and species richness were analyzed in a generalized linear model with a Poisson distribution, including location of sample, vineyard, time in relation to spraying, year, and the interaction of location and time from spraying as factors. The abundances of the dominant species were also analyzed in generalized linear models with the same explanatory variables. We used two models for each species. The first model compared all vineyard locations to nature locations, and the second model included only vineyard locations with vegetation type and location in vineyard as factors. Thus, it was possible to determine what aspect of location affected populations. Both models also included year, time in relation to spraying, and interactions as factors. Wasp abundance per day in sticky traps was log transformed to fit a normal distribution and was analyzed in a general linear model with time in relation to spraying, location, year, and interactions as factors. The correlation between distance from the vineyard and wasp abundance, species richness, and Telenomus sp. abundance was tested for each date in relation to spraying, using a Pearson correlation. The analysis of the transect into the natural area focused on Telenomus sp. because the abundance of other species was low in this area.

To avoid confounding effects between date of sampling and time after spraying, samples included different vineyards which were sprayed at different dates. This might affect the results if there are seasonal trends in parasitoid abundance. To test for this possibility, we correlated the dates of pre-spraying sampling and the abundance of all wasps and of the dominant species using Spearman correlations. All analyses were done using IBM SPSS version 24 software.

Parasitoid richness and abundance were highest in natural habitat, higher in the herbaceous vegetation than in the vines, and higher in the vineyard margins than at their center (Figure 2), i.e., vineyard habitats that were close or similar in vegetation to the natural areas were richer in parasitoids. Similar patterns of parasitoid abundance and diversity were observed in previous surveys of vineyards in Israel’s Mediterranean region (Shapira et al., 2018), but not in desert vineyards (Segoli et al., 2020). The difference between the natural and the vineyard habitats became more marked in the first 2 days after Indoxacarb application, as parasitoid abundance and richness declined in the vineyards but not in the natural habitat. Yet, all areas in the vineyards also showed a recovery in total abundance and in species richness within 2 weeks. These findings may result from a combination of demographic processes: mortality and emigration of adult parasitoids due to insecticide exposure can lead to population declines. The subsequent recovery in their abundance may reflect the emergence of new adult cohorts within the vineyard and recolonization from nearby natural areas. A study of beetles, for example, found that recovery patterns differed among families, with some recovering from within fields and some from boundaries (Jepson and Thacker 1990). Our sampling design did not allow assessment of in-situ recovery within the vineyards. We did, however, find evidence for parasitoid migration out of the vineyard following Indoxacarb applications.

The increases in wasp capture on sticky traps suggest that there was migration to the natural habitats and subsequent migration back to the vineyard. The trend towards increasing abundance that was observed in the natural habitat transect in days 3–10, particularly at greater distances from the vineyard, provides additional suggestive evidence of migration to and recovery from the natural habitat. The distance of 50 m from the vineyard which was sampled represents a reasonable distance that the wasps could have traveled since a previous study has indicated that Anagrus erythroneurae Trjapitzin and Chiappini, 1994 wasps, which are similar in size to the common species in this study, were observed traveling 30 m in 6 days (Irvin et al., 2018). The mechanism driving adult wasp migration from the vineyard to natural habitats could be a repellent effect of the insecticide, which has been shown to increase parasitoid mobility (Desneux et al., 2007). These results provide support for the suggestion in previous studies that parasitoid recovery occurs from non-sprayed areas that contain potential habitat for parasitoids (Mates et al., 2012). Spatial analysis of recovery of predatory beetles after insecticide application has also shown gradual increase from unsprayed edges towards the sprayed center of fields over time (Duffield and Aebischer 1994). Further, in a study in corn fields, the presence of strips of natural vegetation reduced the negative effect of pesticide applications on predatory beetles (Lee et al., 2001). Since the vineyards in this study were small, the center of the vineyards was close to the margins, potentially allowing rapid recovery (Kattwinkel et al., 2015).

Heterogeneous vegetation within and around vineyards improves biodiversity and pest control (Paiola et al., 2020). Our study supports this idea by providing evidence that natural enemies migrate between vineyards and neighboring natural habitats after insecticide applications. Increased presence of natural habitat has been shown to increase biocontrol, though by predators and not parasitoids, when pesticide pressure was low, suggesting that if pesticides are sprayed more frequently, there is insufficient time for recovery from natural habitats (Ricci et al., 2019). Another study where low toxicity pesticides were used found no effect of agrochemical use, but a positive effect of plant diversity and complexity, on arthropod diversity, suggesting that the combination of low pesticide intensity and presence of appropriate habitat for arthropods promote higher arthropod diversity in agricultural habitats (Geldenhuys et al., 2021).

Of the four abundant species, Telenomus sp. and Trichogramma sp. were more affected by spraying than Oligosita sp. and Lymaenon litoralis. All the abundant species are egg parasitoids; Telenomus sp. parasitize Lepidoptera and Hemiptera, Trichogramma sp. parasitize Lepidoptera, Oligosita sp. parasitize Hemiptera and Thysanoptera, and Lymaenon litoralis parasitize Hempitera. Telenomus sp. and Trichogramma sp. may be most affected because their hosts are targeted by Indoxacarb, in addition to the toxic effect by contact on adult wasps that could be experienced by all species. The decline within the first 2 days in total wasp abundance suggests a direct toxic effect, while the decline in species parasitizing Lepidoptera suggests an indirect effect on the parasitoids via their hosts, if the insecticide is toxic to the moth eggs, which are parasitized by the wasps. On the other hand, egg parasitizing species may be protected by the chorion of their hosts while they develop and survive if they emerge after the effect of pesticide residue subsides (Loch 2005; Bueno et al., 2017). For Telenomus sp. and Trichogramma sp., emergence of adults from eggs that were parasitized before spraying would occur up to 8–13 days later (Navasero and Oatman 1989; Reda Abd el Monsef 2004), within the sampling time in the second year of the study.

Interestingly, Trichogramma sp. was among the most abundant species inside the vineyards, though this is a genus that has been shown to be relatively sensitive to pesticides (Theiling and Croft 1988), and specifically to Indoxacarb (Scholz and Zalucki 2000; Asma et al., 2018; de Paiva et al., 2018; Gallego et al., 2019; Madhusudan and Bhushanam 2020). It is, however, a genus that has been previously found in high densities inside vineyards, reflecting abundance of lepidopteran hosts (Reda Abd el Monsef 2004).

Our study documented the consequences of Indoxacarb applications that were administered by the farmers, rather than our own controlled manipulations of the spraying schedules. This observational approach did not allow us to set up Indoxacarb-free plots as matched controls to the sprayed vineyards. Untreated control plots are important for detecting seasonal population trends that are independent of the insecticide’s effects (Longley et al., 1997). For example, population declines that occur both in insecticide-treated plots and in their matched control plots should not be attributed to the insecticide, but instead to the species’ seasonal phenology. In our study, we reduced the potential confound between Indoxacarb-induced and seasonal population changes by sampling Indoxacarb-treated vineyards at various dates along the whole grape-growing season (Supplementary Table S1). In addition, we found no seasonal trend in the pre-spray total abundance of all parasitoids combined, suggesting that their post-spray decline reflects mortality associated with Indoxacarb exposure.

Overall, the dominant species in the natural and vineyard habitat, as well as overall parasitoid abundance and species richness appeared to recover within 2 weeks. The natural habitat present near the vineyards could provide a non-sprayed shelter for parasitoids repelled by the pesticide and could be a source to replace vineyard wasp communities. Thus, preservation of some natural habitat near vineyards, along with infrequent pesticide applications, could contribute to conservation of natural parasitoid communities, as shown here. In fact, a retrospective analysis of a large pest control dataset from vineyards found that natural habitats around the vineyards reduce pest outbreaks, as well as the need to apply chemical insecticides (Paredes et al., 2021). A further question to be explored would be the dynamics of the community outside the vineyard in space and time after spraying, and how the frequency of sprayings affects the community over an entire season.