Loss of parasitoid diversity in China’s corn agro-ecosystem over a 30-year time period

Globally, insect abundance and diversity are experiencing a rapid decline. Despite important inter-taxa and geographical variability, this can entail an extinction of ecological interactions and a decay of ecosystem functions. In this study, we compared the spatial distribution, abundance and species composition of Trichogramma spp. egg parasitoids (Hymenoptera: Trichogrammatidae) over a 30-year time period in China. During the 1980s and in 2016–2018, egg masses of the Asian corn borer (ACB), Ostrinia furnacalis, were systematically sampled from corn fields across the country. In 2018, five species were identified with Trichogramma ostriniae representing 90% of the species complex. Since the 1980s, two new species have made their appearance while nine (out of 12; i.e., 75%) species disappeared. These include comparative specialists but also generalists such as T. evanescens and T. exiguum. Across sites, species richness (R) and diversity (Shannon-Weiner index) have declined by a respective 25–86% and 56–100% (except for Heilongjiang province) over this time frame. We hypothesize that this is attributed to land use change, pesticide use and plant diversity decline in agro-landscapes. Conversely, no negative impacts were detected of augmentative biological control. Given the drastic reduction in ACB parasitoid richness, agro-ecological measures and diversification strategies should be deployed to restore the ecological resilience of local farming systems. Our work carries major implications for food security and helps to muster support for more nature-friendly, pest-resilient farming systems in China and abroad.


Introduction
Over the past decade, several scientific studies have reported a progressive decline of insects in many parts of the globe (Hallmann et al. 2017;Lister and Garcia 2018;Seibold et al. 2019;Sánchez-Bayo and Wyckhuys 2019;Wagner et al. 2021;Cardoso et al. 2020). Multiple context-and taxa-specific drivers have been identified, with landuse change, agrochemical pollution, biological invasions and climate change the most prominent factors (Seibold et al. 2019;Sánchez-Bayo and Wyckhuys 2019;Wagner et al. 2021). As insect biodiversity underpins the overall stability and functioning of natural and man-made ecosystems alike (Cardinale et al. 2012;Soliveres et al. 2016;Dainese et al. 2019), even subtle insect losses can have major, long-lasting impacts (Janzen 1971;Kehoe et al. 2021;Crossley et al. 2020). Though extensive cross-taxa studies are missing from Asia, initial surveys show that 18% of migratory insect species in China are becoming less abundant (Guo et al. 2020). Insectivorous dragonflies (Odonata) i.e., taxa that underpin ecosystem services such as natural biological control, are subject to 14% annual rates of decline in Japan and China (Kadoya et al. 2009;Guo et al. 2020). It remains however to be seen whether this loss in beneficial insects is widespread.
Insect-mediated biological control underpins the resilience of agricultural and natural ecosystems and is conservatively valued at US $ 4.5-17 billion per year in USA alone (Pimentel et al. 1997;Losey and Vaughan 2006). In China, beneficial insects (i.e., parasitoids, predators) play a key role in regulating populations of the Asian corn borer (ACB), Ostrinia furnacalis (Guenée 1854) (Lepidoptera: Crambidae). This widelydistributed species is a key pest of corn (Zea mays) in the Asia Pacific Region (Nafus and Schreiner 1991), and annually causes 6-9 million tons of yield loss in China (Zhou et al. 1988;He et al. 2003). Aside from direct damage to corn stalks and kernels, ACBinduced fungal infection leads to mycotoxin contamination and quality reduction of harvested grains (Song et al. 2009). As China is a globally-important producer and consumer of corn, ACB attack thus directly relates to the country's food security and rural development. Parasitic wasps of the genus Trichogramma are key natural enemies of corn borers in China and across the world (Consoli et al. 2010;Zang et al. 2021). In addition to the action of naturally-occurring Trichogramma spp., ACB biological control is provided by augmentative releases of mass-produced parasitoids (Wang et al. 2014;Zhan and Liang 1999;Zang et al. 2021).
Together with the former USSR, Brazil and Mexico, China is globally renowned for its use of Trichogramma spp. in pest control (Van Lenteren et al. 2018). Since the 1970s, augmentative releases of Trichogramma spp. have been widely adopted in Northeast China i.e., in Heilongjiang, Jilin, Liaoning provinces (Wang et al. 2014;Zang et al. 2021). At present, augmentative biological control is annually performed on nearly four million hectares of corn i.e., roughly 10% of China's national corn acreage (Wang et al. 2014;Huang et al. 2020). These yearly releases of millions of Trichogramma dendrolimi have proven exceptionally successful, resulting in ACB parasitism rates above 70% (Feng 1996). In the remaining areas and throughout Eastern Asia, ACB is primarily managed through recurrent spray applications of synthetic pesticides (e.g., Yang et al. 2021). Meanwhile, China's corn production has been subject to important changes over the past decades. While local corn production is increasingly typified by inputintensive, mechanized farming schemes and mono-cropping (Ely et al. 2016), a fraction of corn growers have adopted more sustainable practices and reduced chemical fertilizer 1 3 use by 15-18% (Cui et al. 2018). These field-, farm-and landscape-level changes undoubtedly have impacted resident insect populations including biological control agents such as Trichogramma spp.
During the 1980s, Zhang et al. (1990) recorded a total of 12 different ACB-associated Trichogramma spp. in China's main corn-growing regions. ACB parasitism varied over space and time; certain Trichogramma spp. attained higher parasitism levels than others, as equally observed following assisted releases (Tan 1999;Feng et al. 1999;Wu et al. 2001;Xu et al. 2001;Guo et al. 2005). Though Zhang et al (1990) did not survey several of today's main corn-growing regions, follow-up parasitoid censuses can help to assess how the resident Trichogramma spp. complex has responded to changes in the farming system over the past three decades. It can help to evaluate the impact of T. dendrolimi augmentative releases, such as in Northeast China and Beijing's Miyun district, on the relative abundance and species composition of resident Trichogramma spp. Lastly, follow-up surveys provide an opportunity to refine integrative taxonomic approaches for the identification of minute parasitic wasps (Schlick-Steiner et al. 2010;Heethoff et al. 2011). Specifically, sole reliance upon morphological features is challenging given the high levels of phenological plasticity among Trichogramma spp. and pronounced female-biased sex ratios in field populations (Pinto et al. 1989). Advances in molecular biology enable an accurate identification of resident Trichogramma spp. and a more robust assessment of shifts in egg parasitoid community structure.
In this study, we report findings of a nationwide ACB parasitoid census that was carried out during 2016-2018 in all of China's main corn-growing regions. We compare current Trichogramma spp. community composition, relative abundance and geographical distribution with historical data from the 1980s (i.e., Zhang et al. 1990). Also, we contrast the above metrics for areas that were or were not subject to augmentation biological control. Our study uncovers temporal shifts in on-farm insect biodiversity over a 30-year time frame, and helps to make inferences regarding the strength of parasitoid-mediated biological control in one of Asia's prime food, feed and fiber crops.

Historic parasitoid censuses
Data were collated from successive sampling events that were carried out by Zhang et al. (1990). Sampling was conducted in 13 corn-growing provinces and counties between 1977 and 1987 i.e., over a 10-year period (Table S1). ACB egg masses were sampled from different field sites, kept in the laboratory until parasitoid emergence and parasitoids were identified based upon morphological characteristics. No further details are available on the exact sampling methodology or sample sizes i.e., number of (collected, parasitized) ACB egg masses. For each sampling site, parasitoid species identity and diversity measures were obtained.

Collection of ACB egg masses
There are six main corn-growing regions in China i.e., the North Spring (NS), Huang-Huai-Hai Summer (HS), Southwest Hills (SWH), South Hills (SH), Northwest Inland Irrigation (NWI), and the Qinghai-Tibet Plateau Corn Region. During 2016-2018, ACB egg masses were collected from five regions i.e., NS, HS, SWH, SH, NWI throughout the ACB 1 3 oviposition period. Throughout its distribution range, ACB completes between one and seven generations per year and sampling was conducted during the corn tasseling period i.e., mid-July to late October. Egg masses were collected from successive ACB generations in the different regions i.e., 2nd generation (NS, NWI region), 3rd generation (HS) or 4th to 5th generation (SH, SWH). In total, 49 sampling sites were selected and 20-30 ACB egg masses were collected at min. 5 m distances per site. Only one sampling event was conducted per site, for which local cropping patterns and agronomic practices were recorded (Table S1). Upon field collection, ACB egg masses were individualized in sterilized 10 ml micro tubes, covered with a cotton plug and kept at 8 °C. Tubes were then transferred to the laboratory and kept at 25 °C until the emergence of adult Trichogramma or Telenomus spp. parasitoids. Parasitoids that emerged from a single egg mass were treated as an individual population and were kept at the Institute of Plant Protection, Chinese Academy of Agricultural Sciences (IPP-CAAS), Beijing, China. Each single population was maintained on ACB egg masses, and adults were fed with 20% honey-water solution. Dead specimens were kept in sterilized 1.5 ml tube at − 20 °C until further morphological and molecular analyses.

Molecular identification
For each parasitoid population, genomic DNA was extracted from several specimens using Tris-HCL lytic buffer as per Li (2007). Individual Trichogramma sp. or Telenomus sp. specimens were crushed in 10 µl of Tris-HCL buffer on parafilm with a sterilized micropestle. Next, 25 µl of the homogenate was transferred to a sterilized 200 µl PCR tube and centrifuged. The homogenate was then incubated in a Techne TC-5000 Thermocycler (Techne, Minneapolis, MN) at 65 °C for 30 min, and 96 °C for 10 min and a final hold at − 20 °C. Two microliters of DNA was used for the PCR template. The PCR was performed in a total volume of 25 µl using the above thermocycler. For each reaction, we used 2 µl of DNA template, 13 µl mixTaq DNA polymerase, 8 µl DD H 2 O, and 1 µl of each forward and reverse primers. To amplify Trichogramma spp. COII regions, we used 5′-ATT GGA CAT CAA TGA TAT TGA-3′ (forward) and 5′-CCA CCA ATT TCT GAA CAT TGA CCA -3′ (reverse) primers (Stouthamer et al. 1999). For Telenomus spp., the following COI regions were used: 5′-GGT CAA CAA ATC ATA AAG ATA TTG G-3′ (forward) and 5′-TAA ACT TCA GGG TGA CCA AAA AAT CA-3′ (reverse). PCR amplification was achieved by initially denaturing DNA at 94 °C for 3 min, followed by 35 amplification cycles (i.e., 30 s at 94 °C, 30 s at 55 °C, and 1min at 72 °C), an extension at 72 °C for 8 min and subsequent storage at − 20 °C. Five microliters of the PCR products were subject to gel electrophoresis on a 1.5% agarose gel that was stained with Ethidium bromide along with a size ladder (Trans 2K® plus DNA Marker100-bp). Following a 25 min electrophoresis run at constant 120 V, DNA bands were visualized under ultraviolet (UV) light and amplification was confirmed based upon the length of the amplification fragment (Fig. S1). Each PCR product was then sequenced in an automatic sequencer (Sangon Biotech, Beijing) to obtain the matching sequences in both directions of DNA strands. The resulting forward and reverse reads were manually aligned and checked for agreement in DNAMAN and DNASTAR (Laser gene v.7). Species were identified by submitting the nucleotide sequences in Gen Bank Database, NCBI using BLAST to search for similarity with nucleotides present in the database. Sequence matches of 99-100% were used to identify each specimen (or population); confirmed sequences were equally deposited in GenBank.

Morphological identification
Individual specimens of Trichogramma and Telenomus spp. wasps were slide-mounted following procedures adapted from Noyes (1982). Dried male specimens were incubated in 10% KOH for 24 h at room temperature until body parts became transparent. After pipetting the KOH, each specimen was rinsed three times in distilled water and then washed with ethanol for another three times (35%, 70% and 95% concentration; 15-20 min each). Next, dehydrated specimens were soaked in high-quality clove oil overnight, and subsequently transferred onto the glass slide in a drop of clove oil and Canada Balsam (1:1). Specific organs were dissected with a No. 1 needle and properly positioned. After positioning the cover slip on each specimen, slides were air-dried for 3-4 days at room temperature. From each parasitized ACB egg mass, 5 male Trichogramma individuals were randomly selected for species identification. In total, 6000 male individual specimens were thus identified. Pictures were taken using an Olympus SZ61 light microscope (Olympus Corporation, Tokyo, Japan) and the following genital structures were examined: gonoforceps (GF), chelate structures (CS), median ventral projection (MVP), dorsal expansion of gonobase (DEG), lateral lobes (LL) and central ridge (CR). Ratios were defined as per Pinto (1999) (Fig. S2). For each sampling event and geographical location (i.e., sampling site, province, corn-growing region), we calculated the Shannon-Wiener diversity index (H) and richness index (R), in which R equals the number of Trichogramma species. Over 1977Over -1987 male Trichogramma individuals were identified to species-level and 12 congeneric Trichogramma spp. were recorded from ACB eggs collected in China's main corn-growing regions (Zhang et al. 1990).  (Table 1). In NWI, T. brassicae and T. ostriniae emerged from field-collected egg masses; the former species was solely recorded from this region. In HS, only T. ostriniae was recovered from most sites except for those in Beijing where the species comprised 92.0% of the entire complex. In SWH, 99.5% of emerged parasitoids were T. ostriniae while the remainder were T. chilonis. Lastly, in SH, T. ostriniae accounted for 66.3% all individuals while a further three species were recorded.

Long-term shifts in occurrence and abundance
Over a 30-year time frame, Trichogramma spp. community structure and geographic distribution changed considerably (Table 2). While T. ostriniae continues to be the most widely distributed species, other parasitoids experienced a more restricted distribution (T. dendrolimi) or were entirely absent from 2018 censuses (e.g., T. evanescens, T. leucaniae). On a province-level, Trichogramma species richness (R) declined from 2-7 during the 1980s to 1-3 over 2016-2018. While richness remained unchanged in Heilongjiang province, it dropped by 25.0-85.7% in the remaining 7 provinces. Province-level changes in species richness (R) were paralleled by shifts in the overall species composition, e.g., with T. ostriniae and T. dendrolimi switching in their relative abundance in Heilongjiang province (Fig. 1).
On a province level, Trichogramma species Shannon-Weiner index ranged from 0.12-0.94 and 0. 00-0.58 during 1977-1987 and 2016-2018, respectively. Compared to the 1980s, overall Trichogramma species Shannon-Weiner index declined by 56.2-100% except for Heilongjiang province, where it increased by 314.3%. In the three northeastern provinces where T. dendrolimi mass-releases were conducted, species richness (R) ranged from 1-3 (as compared to 1-2 for provinces where no augmentative releases were done). Furthermore, H indices in provinces with or without augmentative biological control ranged from a respective 0-0.58 to 0-0.09 during 2018 (Table 3).

Discussion
Drawing upon historic census data, extensive field surveys across China's main corn-growing regions and integrative taxonomy approaches, we unveil temporal changes in the ACB egg parasitoid community. Out of the 12 different Trichogramma spp. that were identified during the 1980s, only three were recovered over 2016-2018 (while two new species were found in certain areas). In addition to the declines in parasitoid richness (R) and diversity, several Trichogramma spp. exhibited a more restricted geographical distribution. We 50% or more since the 1980s (Zhang et al. 1990). In Liaoning, richness (R) even declined by 85.7% over this time frame (Table 3). This could be ascribed to inaccuracies in earlier morphology-based identification, which are common for minute parasitic wasps such as Trichogramma spp. For example, Trichogramma species that proved absent during recent censuses may have been incorrectly identified during the 1980s. Also, parasitoid community composition can vary considerably throughout the cropping season, with T. dendrolimi and T. ostriniae being the main parasitoids of a respective 1st vs. 2nd generation ACB in the 1980s (Zhang et al. 1990). Furthermore, changes in landscape-level crop diversity or farming methods can lead to parasitoid diversity loss (Laliberté and Tylianakis 2010;Jonsson et al. 2012). Though the disappearance of comparative specialists such as T. forcipiformis, T. tielingensis, T. brassicae, T. leucaniae or T. poliae is directly tied to the loss of natural habitat or lepidopteran hosts (Habel et al. 2019a), other factors drive the disappearance of generalists such as T. evanescens, T. exiguum, T. pintoi or T. closterae. Other facets of China's agricultural intensification e.g., the extensive use of synthetic pesticides and an overall deterioration of habitat quality likely explain the latter pattern (Sanchez-Bayo and Wyckhuys 2019; Raven and Wagner. 2021;Habel et al. 2022). Similar phenomena have been linked to the dramatic decline in European butterflies over the past decades (Raven and Wagner. 2021;Habel 2019b;Seibold et al. 2019;Habel et al. 2022). Shifting cultivation patterns can also be a key contributor: planting area of corn and soybean in Northeast China increased by a respective 160% and 47% over the study period. Meanwhile, in the Huang-Huai-Hai Summer Corn Region, corn acreage increased by 94% and geneticallyuniform corn crops replaced more diverse cropping systems. Conversely, soybean acreage increased by 116% in Heilongjiang province. Lastly, the overuse of synthetic pesticides on China's farmland undoubtedly contributes to biodiversity loss in agricultural settings and beyond (Wu et al. 2018;Tang et al. 2021). While pesticide-based crop protection can lead to short-term declines in pest pressure (but see Janssen and van Rijn 2021), these effects are likely outweighed by a steady decay of ecosystem functionalities and ecological resilience.
In several sites e.g., Beijing, Shandong, T. dendrolimi has been released against O. furnacalis for extended periods of time (Yu et al. 1982;Feng 1996, Feng et al. 1999). Our findings do not show that these augmentative releases negatively affected the ACB egg 1 3 parasitoid community in corn fields. As compared to experiences in Switzerland with T. brassicae Babendreier et al. 2003), laboratory-reared T. dendrolimi possibly do not readily establish in local corn fields where T. ostriniae remains the dominant species. This could relate to differences in competitive advantage (Liu 2019), host range and host acceptance behavior of both species, with T. ostriniae preferably oviposits in small host eggs e.g., Corcyra cephalonica, Sitotroga cerealella, as compared to T. dendrolimi which accepts larger eggs such as those from Antheraea pernyi. Aside from its clear advantages for laboratory-based mass-rearing (Wang et al. 2014), a preference for large-size hosts might lower risks for negative environmental impacts in agricultural settings. As comparatively few T. dendrolimi were recorded in the field samples, the species potentially prefers temporally-stable habitats such as natural areas, fruit orchards or forest plots (Oztemiz 2007;Wang et al. 2014) where it associates with non-pest lepidopterans e.g., Pandemis heparana, Gastropacha populifolia, Dictyoploca japonica, Dendrolimus punctatus, Cerura menciana (Zhang et al. 1979). Hence, further censuses in non-agricultural settings are essential to ascertain whether inundative releases of T. dendrolimi do not negatively impact on resident (non-pest) biota e.g., species belonging to the Saturniidae. Evidently, any future ecological impact assessment of augmentative biological control needs to be conducted at ecologically-relevant spatial scales and encompass both agricultural and natural habitats. Changes in host community composition can also bring about shifts in parasitoid species identity and incidence. In areas such as the Northwest Inland Irrigated Corn Region, corn crops are either colonized by O. furnacalis or by the European corn borer Ostrinia nubilalis (Wang et al 2017). Both species only co-occur in Yili district of Western Xinjiang, where O. furnacalis has been gradually displacing O. nubilalis in local cropping systems (Yang et al. 20082011;Wang et al. 2017). In areas where O. nubilalis is endemic, such as Iran or Turkey, T. brassicae is the main Trichogramma species (Poorjavad et al. 2012;Koca et al. 2018). Similarly, our survey in Xinjiang only yielded T. brassicae though other species e.g., T. pintoi occur at background abundance levels (Wu et al., 2008). This low species diversity can thus possibly be attributed to the exact identity of local lepidopteran hosts, but could also be ascribed to Xinjiang's geographical isolation and climatic conditions. The latter may explain local absence of T. dendrolimi and T. chilonis, despite being frequently released for augmentation biological control (Xu et al. 2001). The above two species possibly cannot compete with the locally dominant T. brassicae and fail to persist in the harsh and ecologically fragile (desert) settings. Overall, temporal shifts in host and parasitoid community carry implications for crop protection, and habitat management schemes will need to be attuned to the locally prevailing parasitoid species and (a-) biotic conditions (Gurr et al. 2017;Perović et al. 2018). Applied ecological research is thus essential to ensure that (biodiversity-based) management strategies are well-suited to local farming contexts. Similarly, much can be learned from inundative biological control efforts with T. brassicae in areas where O. nubalilis is endemic (Burgio and Maini 1995;Hawlitzky et al. 1994;Suverkropp 1994;1997;).
In the 1980s, four Trichogramma species, T. ostriniae, T. dendrolimi, T. chilonis and T. evanescens, were widely distributed in various areas of China (Zhang et al. 1990). These patterns contrast markedly with those in 2018 (Tables 2, 3), where only few areas harbored three species and where one single species (T. ostriniae) made up the bulk of the egg parasitoid community in several areas e.g., Southwest Hilly Corn Region. The above compares to the markedly higher abundance of T. chilonis on ACB egg masses during the 1980s (Zhang et al. 1990). Meanwhile, two Trichogramma spp. are newly recorded from ACB egg masses (i.e., T. brassicae, T. bilingensis) and Telenomus remus is equally reported from Guangxi province. Considering how T. bilingensis represents 36% of the resident ACB egg parasitoid community at certain locations, work is urgently needed to examine the biology, ecology and biological control potential of these new species. On the other hand, T. remus is a well-studied generalist parasitoid that provides effective control of the newly-invasive fall armyworm, Spodoptera frugiperda (Kenis et al. 2019;Liao et al. 2019;Jiang et al. 2019;Paolo et al. 2020). An in-depth characterization of the emergence of these new species and overall species turnover within the O. furnacalis parasitoid community may prove worthwhile e.g., in view of climate-driven range expansion of target pests (Zeng et al. 2020). As such, ecological studies in corn agro-ecosystems within (sub-)tropical settings can help to anticipate community shifts (and associated changes in ecosystem services) under climate change scenarios.
As a follow-up to earlier censuses during the 1980s (Zhang et al. 1990), our survey for ACB egg parasitoids covered a broad range of agro-ecological contexts and farming systems in China. We record strong declines in Trichogramma species richness (R) and Shannon-Weiner index in local corn agro-ecosystems. Out of the 12 species that were recorded in the 1980s, only three were recovered during 2016-2018 i.e., T. ostriniae, T. dendrolimi and T. chilonis while two new species have made their appearance. This 75% loss in species richness (R) can partially be ascribed to sampling bias and variable sampling intensity between geographies during the 1980s, lack of a uniformity in survey methodology between both time periods, or faulty species identification. Limited sampling i.e., one single observation per site during 2016-2018 also represents a major weakness of our study. Despite these shortcomings in survey methodology, the observed trends are most certainly related to major changes in China's agriculture over the past decades e.g., shifting cultivation patterns, (chemical) intensification and a progressive loss of crop and non-crop plant diversity. In particular, the increased use of insecticides and herbicides likely exerted net negative impacts on resident parasitoid biota. While our analyses do not reveal negative impacts of augmentative T. dendrolimi releases on on-farm parasitoid diversity, follow-up assessments are needed to ascertain its ecological impacts in natural habitats. Overall, a precipitous decline in parasitoid biodiversity as recorded in our study can lead to loss of trophic regulation and a proliferation of crop pests, with potentially grave impacts for food security and societal wellbeing (Burra et al. 2021). Hence, agro-ecological and biodiversity-based measures should be prioritized to intensify China's corn production systems in a way that benefits farmers, society at large and the environment. Biological control e.g., inundative releases of laboratory-reared Trichogramma spp. can be a cornerstone of such ecological intensification strategy.