Evaluation of reflective groundcovers for pest and fruit quality management in commercial pear orchards

Reflective groundcovers can improve crop yield and quality and deter pests by increasing light penetration into plant canopies. In pear (Pyrus communis L.) orchards, this could increase profitability and reduce reliance on insecticides to manage pear psylla (Cacopsylla pyricola [Foerster]). Two types of reflective groundcovers, metallized polyethylene film (metallized film) and white geotextile fabric (reflective fabric), are known to increase pear yield and quality, and metallized film was recently reported to suppress pear psylla. Here, we tested reflective fabric for pear psylla suppression at a research orchard in 2019, then we partnered with five pear orchard managers to test metallized film and reflective fabric for pear psylla suppression at five conventional commercial orchards in 2020 and one organic commercial orchard in 2021. There was evidence of pear psylla suppression by both metallized film and reflective fabric, but effects varied among reflective materials, pear psylla life stages, season and year of study. Reflective groundcovers did not affect fruit size, weight, sunburn or pest mite abundance, and they were compatible with pear psylla biocontrol. Low perceived benefit relative to cost was a barrier to continued trialling and adoption. Whereas orchardists in our study used reflective groundcovers in addition to normal pesticide regimes, previous research studies suggest that reflective groundcovers can profitably replace some insecticides. Future studies are needed to test the outcomes of replacing pesticides with reflective groundcovers in commercial orchards and understand under what circumstances they improve pear quality and yield.

Pear psylla, a pest with three to four generations in the Northwest United States, reduces the value of pears when honeydew produced by leaf-feeding nymphs drips onto fruit, causing black marks from sooty mould growth . Suppressing firstgeneration pear psylla is important to reduce the size of subsequent generations, and management of first-generation pear psylla is practically always warranted according to regional understanding (VanBuskirk et al., 1999). In orchards practising integrated pest management, there is a risk of pear psylla damage during their second generation, but pear psylla's natural enemies usually suppress the third pear psylla generation (DuPont et al., 2021). In insecticidedependent orchards, third-generation pear psylla nymph populations can flare due to disruption of biocontrol (DuPont et al., 2021), often coinciding with harvest of 'Anjou', the most prevalent pear cultivar in Washington (USDA NASS, 2017). Pear psylla and related Cacopsylla species cause similar problems for pear production across Eurasia and northern Africa (Dong et al., 2019;Esmaeily et al., 2021;Sanchez & Ortín-Angulo, 2012), and there is a need to reduce insecticide reliance across these systems.
Pest management is not the only challenge to pear orchardists, and integrated pest management approaches need to consider the context of the whole farm, including horticultural and economic factors (Granatstein & Kupferman, 2008). Reflective groundcovers are a tool that could reduce insecticide reliance while increasing fruit quality and yield. Reflected light can suppress many pest insects by interfering with their orientation behaviour (Croxton & Stansly, 2014;Nottingham & Kuhar, 2016;Shimoda & Honda, 2013;Simmons et al., 2010). By increasing photosynthetically active radiation in canopies, reflective groundcovers can also increase pear fruit set, size or weight in some years and locations (Bertelsen, 2005;Einhorn et al., 2012;Hanrahan et al., 2011), but not others (Vangdal et al., 2007;Yamamoto & Miyamoto, 2005). As disadvantages, increased light could increase the risk of pear sunburn (Spera et al., 2023), and a previous pear orchard study reports that a reflective groundcover increased density of a secondary pest, spider mite (Tetranychus spp) (Nottingham et al., 2022). Despite the potential benefits and research studies going back decades, reflective groundcovers are scarcely used in Washington pear orchards. This calls for an investigation of what factors may be limiting their use. In addition, the potential of reflective groundcovers to suppress pear psylla has not been previously evaluated on commercial farms.
In this 3-year study, we quantified the effects of reflective groundcovers on pear pest populations and fruit characteristics on commercial farms, and we interviewed the growers to understand the perceived advantages and disadvantages of reflective groundcovers during the project. Materials tested included a metallized film and a reflective fabric, comprising the two main options for reflective groundcovers in orchards (Bertelsen, 2005;Einhorn et al., 2012;Hanrahan et al., 2011;Vangdal et al., 2007;Yamamoto & Miyamoto, 2005). Previous studies show that reflective groundcovers in pear can reduce the spring generation of pear psylla eggs and nymphs by over 50% compared with control trees (Nottingham et al., 2022;, but these studies only investigated metallized film on university research orchards. In the first year of our study, we tested for effects of reflective fabric on pear psylla in a research pear orchard. In years 2 and 3, we tested reflective fabric and metallized film on plots in commercial pear orchards to account for commercial pest and horticultural management context. were about 2 m in diameter. There were two treatments (reflective fabric and non-manipulated control) and three replications. Sampling plots were 2 tree rows (one 'Anjou' and one 'Bartlett') × 2 trees.

| MATERIAL S AND ME
Reflective fabric was deployed in the centre of all three drive rows adjacent to appropriate plot trees and extended an additional onetree buffer distance on each plot edge ( Figure S1). Hence, control plot edges were 13.8 m from any reflective fabric. Treatments were alternatingly interspersed in a complete block design. Reflective fabric was installed on 3rd April (1 week before bud burst) and removed on 13th May 2019 (2 weeks post-bloom). The orchard was treated with Surround CF (kaolin clay) at 56 kg ha −1 before the study to provide a realistic test environment, as kaolin is a highly effective pear psylla deterrent sprayed by most commercial orchards in our study region of Central Washington (DuPont et al., 2021;Nottingham et al., 2022).

| Pear psylla monitoring
Pear psylla adult monitoring was conducted with four beat tray samples per plot (one from each tree) per sample day. One tray sample consisted of three strikes of one limb by a stiff rubber hose to dislodge insects onto a 45 × 45 cm white cloth tray. Limbs chosen for sampling were ca. 1.5 m from the ground, ca. 5 cm in diameter and of average character. Sampling started on 1st April (pre-installation) and continued weekly until 13th May (immediately before the reflective fabric was removed on that day).
We monitored pear psylla eggs and nymphs from 9th April through 29th April by collecting 20 non-terminal flower buds among the trees in each plot. Under a stereoscope, we inspected woody areas up to 2.5 cm below buds for eggs and nymphs, and we also inspected green tissues once present. On 6th May and 13th May, we counted eggs and nymphs from samples of 50 leaves from each plot.
The leaves were collected randomly across plots from the youngest third of newly formed vegetative growth. CA) and one non-manipulated control. We supplied the materials to cooperators gratis. Installation was managed by the cooperator and their hired labourers following basic instructions to achieve deployment as follows. Reflective fabric was installed in the middle of drive rows and did not extend to herbicide strips ( Figure 1a). It was secured using a bungee cord and hook system provided by the manufacturer, although some cooperators in addition used soil to weigh down the fabric. Metallized film was installed on either side of trees in the herbicide strips (the area below tree canopies that is not mowed) ( Figure 1b). It was secured by weighing down edges with soil. We asked cooperators to instal materials as early as practical in spring (before bud burst) and remove them whenever they chose post-bloom (Table 1).

|
All orchards were managed with conventional spray programs.
Pre-bloom, this involved two to three sprays across March and April including one to two kaolin applications in conjunction with some combination of additional insecticides such as malathion, pyriproxyfen, tolfenpyrad, buprofezin, novaluron, acetamiprid and oil. Postbloom, management involved six to 11 sprays targeted against pear psylla, similar to conventional spray programs presented by DuPont et al. (2021).

| Pear psylla monitoring
We monitored pear psylla with the same methods described in the 2019 research orchard experiment with beat tray tap counts (20 taps per plot; 17th March to 18th August), bud sampling (20 buds per plot; 17th March to 16th April) and leaf sampling (50 leaves per plot; 17th April to 18th August). Trees for sampling were chosen arbitrarily on each visit in a Z-shaped walking pattern across each plot (VanBuskirk et al., 1999).

| Design, site and reflective materials
To test for the effects of reflective groundcovers on pear psylla and fruit quality in 2021, we used a randomized complete block design with four replications of three treatments (reflective fabric, metallized film and non-manipulated control) at a certified organic commercial pear orchard managed by the owner of site D from the 2020 experiments (Table 1). The site was planted in circa 1960, and cultivars were a mix of 'Anjou', 'Bartlett' and 'Bosc', but only 'Anjou' and 'Bartlett' trees were used for sampling. Experimental units were 25 m (five drive rows) by 24 m. We supplied reflective groundcover materials in January 2021, which the cooperator installed on 17th March (before bud burst) and removed it on 8th August (the week before 'Bartlett' harvest). The site was selected for study by the site D owner, who was the only cooperator interested in trialling the groundcovers for a second year. Pest management at the site consisted of two pre-bloom kaolin applications (one application including Paecilomyces fumosoroseus) and seven post-bloom applications including one of Paecilomyces fumosoroseus with mineral oil, one of mineral oil, four of mineral oil with azadirachtin and one of cinnamon oil. There was also one application of calcium carbonate (CaCO 3 ), which can deter pear psylla oviposition .

TA B L E 1
Pear orchard study sites used in the 2020 reflective groundcovers experiment, locations (nearest town), cultivars, row spacing and installation and removal dates for reflective groundcovers (metallized film and reflective fabric). Each site was managed by a different commercial grower.

| Fruit characteristics measurements
We randomly collected 80 'Bartlett' fruits from lower canopies (up to 2.5 m high) in each plot on 18th August, which was less than 1 week before commercial harvest of 'Bartlett' fruit in the study orchard. Fruit characteristics were measured using an AWETA (Aweta Americas) pome fruit sorting line with an optical Power Vision HS camera system for imaging external fruit characteristics. Data were acquired from individual fruit to calculate mean fruit weight, length, width and sunburn incidence.

| Interviews
We interviewed each of the five 2020 cooperators individually at their study site or by phone at the end of the 2020 season. We used a semi-structured interview with open-ended scripted questions and unscripted follow-up questions. Our purpose was to understand factors that would affect the potential for commercial adoption of reflective groundcovers in pear including labour costs, compatibility F I G U R E 1 Pictures of reflective groundcovers at commercial pear orchards: reflective fabric on day of installation, 2020 (a); metallized film on day of installation, 2020 (b); reflective fabric during summer, 2021 (c); metallized film during summer, showing plants overgrowing edges 2021 (d); wind-blown reflective fabric edge onto pear tree due to improper fastening or insufficient soil shovelled on edge to weight down, 2021 (e); wind-blown metallized film due to insufficient soil shovelled for weight, 2020 (f); plant growth around metallized film and ripping deterioration of the material in summer, 2021 (g); and shredded remnants of metallized film after attempted removal, 2021 (h). [Colour figure can be viewed at wileyonlinelibrary.com] with other orchard practices, ease of use and effects on factors other than pear psylla. In 2021, we interviewed the single cooperator following the same protocol as in 2020. The qualitative responses from both years were analysed separately by categorizing and summarizing themes presented by cooperators . We also obtained spray records for study sites. Interview protocols were determined exempt from federal regulation by the Washington State University Human Research Protection Program.

| Statistical analysis
All statistical analyses can be reproduced from publicly available raw data and code ([dataset] Orpet, 2022). Statistical tests were done in R version 4.0.0 (R Core Team, 2020) using the 'aov' function to construct models, the 'summary' function to assess significance and the 'emmeans' function of the 'emmeans' package (Lenth, 2022) for Tukey tests.
We used ANOVA to assess the effects of reflective groundcover treatments on response variables of: cumulative insect days (Ruppel, 1983)   of ANOVA. The results presented are non-transformed. We used Tukey's tests (α = 0.05) for multiple treatment comparisons in models with a significant treatment effect (p < 0.05).

| Research orchard experiment, 2019
Pear psylla adult abundance was similar among treatment plots before reflective fabric was installed at the research centre orchard (Table 2). Post-installation, there was a significant treatment effect on eggs per bud, which was 40% lower in reflective fabric plots (Table 2). There was a marginally significant effect of reflective fabric on adults per tray (p = 0.085), but the effect was small (15% fewer adults than the control) ( Table 2). Nymphs were abundant and there was no treatment effect (Table 2).

| Conventional commercial orchards experiment, 2020
In 2020, pear psylla adults were at their maximum observed density before treatments were applied to commercial orchards, egg laying had already begun and no pear psylla nymphs were present Eggs and nymphs were not present at the time of the pre-treatment count.
Spring insect sampling started a day after groundcover installation and 2 days after the first orchard spray (kaolin; Figure 3). During spring, there was one generation of pear psylla eggs ( Figure 3b) and nymphs ( Figure 3c). Across all plots, pear psylla adults decreased to a low level before the main period of egg increase (Figure 3a). Pear psylla eggs were found only on the woody area below buds until 1st April, when flower buds first opened. In spring, there were significantly fewer pear psylla adults in the metallized film and reflective fabric treatments relative to the control (F = 4; df = 2,6; p = 0.0003, During summer sampling, pear psylla adult counts remained low and never exceeded an average of one adult per tray for any treatment on any sample day (Figure 3a). Significantly fewer adult pear psylla were found in reflective fabric plots compared with metallized film and control treatments ((F = 13, df = 2,6; p = 0.006); Table 3), but there was no effect on eggs (F = 1.8; df = 2,6; p = 0.24) or nymphs

| Natural enemy monitoring
Campylomma verbasci adults were the most common natural enemy on beat tray samples. These were first observed on 9th June and were most abundant throughout July ( Figure 3D). There was no sig- On sticky cards, T. insidiosus and C. verbasci were the most common natural enemies (Table 4). Other natural enemy taxa were not found in meaningful numbers for analysis (Table 4). In summer, there was a significant treatment effect on both T. insidiosus (F = 8.1; df = 2,6; p = 0.020) and C. verbasci (F = 6.5; df = 2,6; p = 0.031). There were significantly fewer T. insidiosus in the reflective fabric relative to metallized film and control treatments, and there were significantly more C. verbasci in reflective fabric relative to the metallized film treatment (Table 4). Other natural enemy taxa were uncommon on sticky cards (

| Fruit characteristics measurements
Fruit characteristics did not differ between treatments (Table 5).

| Interviews
Average labour costs across the 2020 and 2021 experiments for installation and removal were similar between metallized film and reflective fabric ($1071 and $1029 ha −1 ). However, the re-usability of reflective fabric resulted in lower total per annum cost when materials expenses are included ($2371 vs. $1729 ha −1 year −1 ; Table 6). The material could not be fully removed by the grower or research team because attempts to do so resulted in shredding (Figure 1g).
Residual material was left to degrade and continue to be enveloped by vegetation (Figure 1h).
Cooperators found that reflective groundcovers had the potential for management purposes other than pear psylla. Before the trials, cooperators were interested in: altruism to help the pear industry via research (two people), increasing fruit yield or quality (two people) and suppressing pear psylla (four people). At the end of the 2020 trial, four of the growers were more interested in the potential for positive horticultural effects than in effects on pear psylla. One cooperator reported higher tree vigour in reflective groundcover plots, and one other reported larger fruit in their plot with reflective fabric (both subjective observations). One cooperator also reported lower mowing requirements in metallized film plots, requiring one less pass per row due to expanded weed-free zone underneath trees from film covering.
None of the cooperators considered the potential groundcover-

| DISCUSS ION
To stop chronic pear psylla outbreaks in central Washington State, it is necessary to integrate selective pesticide use with cultural management techniques and biological control (Alway, 2001(Alway, , 2003Brunner & Burts, 1981;Burts, 1983;DuPont et al., 2021;DuPont & Strohm, 2019;Westigard et al., 1979). A previous study suggests that reflective groundcovers could replace insecticides used for pre-bloom pear psylla management because metallized film deployed from dormant to petal fall (alone or in combination with one kaolin spray) suppressed first-generation pear psylla to the same extent as a program of multiple broad-spectrum insecticides or a program with multiple kaolin sprays (Nottingham et al., 2022). In the current study, we found that both reflective fabric and metallized film suppressed pear psylla in commercial orchards. However, the reflective groundcovers were used in addition to and not in place of conventional spray programs of the commercial orchards. The additional suppression from reflective groundcovers observed on commercial orchards was modest and TA B L E 5 Commercial orchard 2021 mean ± SEM of weight, length, width and percentage sunburn of 80 Bartlett fruit sampled randomly from each of N = 4 plots per treatment collected on 18th August. Metallized film and white reflective fabric were installed on 17th March and removed it on 8th August.  (Nottingham et al., 2022;. We also did not find increases in pear fruit size or weight, unlike horticulturally focused studies with reflective fabric in the study region (Einhorn et al., 2012;Hanrahan et al., 2011).
To build on research showing pear psylla suppression with metallized film (Nottingham et al., 2022;, we first tested whether reflective fabric had similar effects. Our 2019 experiment with reflective fabric was conducted in the same research pear orchard used by , who observed >50% fewer spring pear psylla adults, eggs and nymphs on pear trees with metallized film compared with control trees. In a follow-up study, Nottingham et al. (2022) observed similar effects in a different research pear orchard. However, reflective fabric significantly reduced only pear psylla eggs in our study. Differences in orchard insecticide sprays may explain this discrepancy. Previous research orchard trials included no pre-installation pear psylla sprays in metallized film treatments (Nottingham et al., 2022;, whereas our research orchard trial included a preinstallation spray of kaolin clay. Kaolin clay is a highly effective pear psylla adult and oviposition deterrent (Puterka et al., 2005), and suppression of pear psylla across plots may explain the limited effects observed in our 2019 study.
We next tested reflective fabric and metallized film in commercial orchards by collaborating with pear orchardists, which allowed us to collect economic data, learn what factors affect whether orchardists adopt or reject reflective groundcovers and adapt study goals based on stakeholder interests. Collaborative on-farm research has many advantages over top-down expert-to-layperson knowledge transfer Warner, 2008), and communication among farmers is an important way that knowledge spreads . Therefore, we thought that involving orchardists in research would help catalyse adoption and spread of reflective groundcovers if cooperators' experiences were positive.
Reflective groundcover effects on pear psylla were modest in our study, perhaps because their effects were dominated by those of commercial spray programs. In spring at commercial orchards, metal- Physical and horticultural factors may explain variation and timing of effects on pear psylla from different reflective groundcovers. In our 2021 study conducted in an organic orchard, metallized film became covered with spray residues (particularly kaolin) and partly overgrown with understory plants more so than reflective fabric, and the resulting reduction in reflectivity may explain why pear psylla adults were suppressed only in reflective fabric plots during summer.  observed less light penetration from metallized film into canopies after trees leafed out versus in early spring, resulting in lack of pear psylla adult deterrence during summer. Pear orchards in our study were old low-density trees with high amounts of structural wood (similar to Einhorn et al., 2012;. Reflective groundcovers might be expected to have greater effects in orchards with smaller canopies (as in Hanrahan et al., 2011;Nottingham et al., 2022). Horticultural effects could explain treatment effects on pear psylla post-removal of reflective groundcovers. Higher later-season pest abundance in reflective groundcover plots is observed with common bean (Phaseolus vulgaris L.) possibly due to increased plant attractiveness and suitability for insects TA B L E 6 Labour (installation and removal, assuming wages of $16 h −1 ) and materials costs ha −1 for reflective materials applied across 0.2 ha per site in 2020 and 2021 commercial orchard experiments. b Metallized film removal efforts were abandoned because vegetation overgrowth and ripping made removal untenable (Figure 1g,h). (Wells et al., 1984). Similarly,    (Nottingham et al., 2022; but similar or more natural enemies relative to insecticide-treated plots (Nottingham et al., 2022). Studies in other crops found either no effects of reflective groundcovers on biocontrol agents (Comeau et al., 2013;Nottingham & Kuhar, 2016;Simmons et al., 2010) or fewer biocontrol agents relative to buckwheat or white clover groundcover plots (Frank & Liburd, 2005).
In our 2021 study, strong biocontrol in the organic orchard could explain the lack of reflective groundcover effects on pear psylla through the summer, as predators kept pear psylla at low levels throughout plots in summer.
Pest Tetranychus mites were abundant during our 2021 experiment during summer through fall, when they typically move from orchard floor vegetation into pear trees (Flexner et al., 1991), but mite abundance in reflective groundcover plots did not differ from control plots. Tetranychus mites are important leaf-feeding pear pests in our study region (VanBuskirk et al., 1999). They are associated with hot and dry weather, and factors such as herbicide use, cultivation, species of weeds present and acaracide use can also mediate pest mite abundance and behaviour in pear orchards (Flexner et al., 1991). Hence, interactions between groundcovers and these factors could result in variable effects on mites in other contexts.
Horticultural effects and labour costs affect whether reflective groundcovers are profitable in pear. Reflective groundcovers deployed before full bloom and left out until harvest results in greater horticultural benefit than when reflective groundcovers are removed midway through the season (Einhorn et al., 2012). Although reflective groundcovers were deployed from pre-bloom to harvest in our 2021 experiment, we did not find an increase in fruit weight or size from reflective groundcovers. Other studies found reflective groundcovers increased pear yields in Washington (Hanrahan et al., 2011) and Oregon (Einhorn et al., 2012) in the first year of use. In Oregon, reflective fabric netted $3985 ha −1 in added profit based on increased returns from higher yield ($13,818 ha −1 ) minus labour to pick the extra fruit ($1311 ha −1 ), labour to instal and remove the fabric ($618 ha −1 ) and annual costs of the materials per 7- year lifespan (Einhorn et al., 2012). Labour costs in this study were estimated higher than Einhorn et al. (2012) and were closer to those of Meinhold et al. (2011), who report reflective fabric and metallized film both required 112 h ha −1 in labour to implement in an apple orchard in Germany. Studies in other regions found variable effects on pears, including higher fruit size and number per tree in Denmark (Bertelsen, 2005) or no practical effects in Norway (Vangdal et al., 2007) and Japan (Yamamoto & Miyamoto, 2005

CO N FLI C T O F I NTE R E S T S TATE M E NT
The authors have no conflict of interest.