Evaluating invasion risk and population dynamics of the brown marmorated stink bug across the contiguous United States

Javier Gutiérrez Illán; Gengping Zhu; James F Walgenbach; Angel Acebes‐Doria; Arthur M Agnello; Diane G Alston; Heather Andrews; Elisabeth H Beers; J Christopher Bergh; Ricardo T Bessin; Brett R Blaauw; G David Buntin; Erik C Burkness; John P Cullum; Kent M Daane; Lauren E Fann; Joanna Fisher; Pierre Girod; Larry J Gut; George C Hamilton; James R Hepler; Richard Hilton; Kim A Hoelmer; William D Hutchison; Peter J Jentsch; Shimat V Joseph; George G Kennedy; Grzegorz Krawczyk; Thomas P Kuhar; Jana C Lee; Tracy C Leskey; Adrian T Marshal; Joshua M Milnes; Anne L Nielsen; Dilani K Patel; Hillary D Peterson; Dominic D Reisig; Jhalendra P Rijal; Ashfaq A Sial; Lori R Spears; Judith M Stahl; Kathy M Tatman; Sally V Taylor; Glynn Tillman; Michael D Toews; Raul T Villanueva; Celeste Welty; Nik G Wiman; Julianna K Wilson; Frank G Zalom; David W Crowder

doi:10.1002/ps.7113

. 2022 Aug 23;78(11):4929–4938. doi: 10.1002/ps.7113

Evaluating invasion risk and population dynamics of the brown marmorated stink bug across the contiguous United States

Javier Gutiérrez Illán ^1,^✉, Gengping Zhu ¹, James F Walgenbach ², Angel Acebes‐Doria ³, Arthur M Agnello ⁴, Diane G Alston ⁵, Heather Andrews ⁶, Elisabeth H Beers ⁷, J Christopher Bergh ⁸, Ricardo T Bessin ⁹, Brett R Blaauw ³, G David Buntin ³, Erik C Burkness ¹⁰, John P Cullum ¹¹, Kent M Daane ¹², Lauren E Fann ⁹, Joanna Fisher ¹³, Pierre Girod ¹⁴, Larry J Gut ¹⁵, George C Hamilton ¹⁴, James R Hepler ⁷, Richard Hilton ¹⁶, Kim A Hoelmer ¹⁷, William D Hutchison ¹⁰, Peter J Jentsch ¹⁸, Shimat V Joseph ³, George G Kennedy ¹⁹, Grzegorz Krawczyk ²⁰, Thomas P Kuhar ²¹, Jana C Lee ²², Tracy C Leskey ¹¹, Adrian T Marshal ⁷, Joshua M Milnes ⁷, Anne L Nielsen ¹⁴, Dilani K Patel ³, Hillary D Peterson ²⁰, Dominic D Reisig ²³, Jhalendra P Rijal ²⁴, Ashfaq A Sial ³, Lori R Spears ⁵, Judith M Stahl ¹², Kathy M Tatman ¹⁷, Sally V Taylor ²⁵, Glynn Tillman ²⁶, Michael D Toews ³, Raul T Villanueva ²⁷, Celeste Welty ²⁸, Nik G Wiman ⁶, Julianna K Wilson ¹⁵, Frank G Zalom ¹³, David W Crowder ¹

^¹Department of Entomology, Washington State University, Pullman, WA, USA

^²Department of Entomology and Plant Pathology, North Carolina State University, Mills River, NC, USA

^³Department of Entomology, University of Georgia, Athens, GA, USA

^⁴Department of Entomology, Cornell University, Geneva, NY, USA

^⁵Department of Biology, Utah State University, Logan, UT, USA

^⁶North Willamette Research and Extension Center, Oregon State University, Aurora, OR, USA

^⁷Department of Entomology, Tree Fruit Research and Extension Center, Washington State University, Wenatchee, WA, USA

^⁸Alison H. Smith, Jr, Agricultural Research and Extension Center, Virginia Tech, Winchester, VA, USA

^⁹Department of Entomology, University of Kentucky, Lexington, KY, USA

^¹⁰Department of Entomology, University of Minnesota, St Paul, MN, USA

^¹¹Appalachian Fruit Research Station, USDA‐ARS, Kearneysville, WV, USA

^¹²Department of Environmental Science, Policy and Management, University of California, Berkeley, CA, USA

^¹³Department of Entomology and Nematology, University of California, Davis, CA, USA

^¹⁴Department of Entomology, Rutgers University, New Brunswick, NJ, USA

^¹⁵Department of Entomology, Michigan State University, East Lansing, MI, USA

^¹⁶Southern Oregon Research and Extension Center, Oregon State University, Central Point, OR, USA

^¹⁷Beneficial Insects Introduction Research Unit, USDA‐ARS, Newark, DE, USA

^¹⁸Hudson Valley Research Lab, Cornell University, Highland, NY, USA

^¹⁹Department of Entomology and Plant Pathology, North Carolina State University, Raleigh, NC, USA

^²⁰Department of Entomology, Pennsylvania State University, Biglerville, PA, USA

^²¹Department of Entomology, Virginia Tech, Blacksburg, VA, USA

^²²Horticultural Crops Research, USDA‐ARS, Corvallis, OR, USA

^²³Department of Entomology and Plant Pathology, North Carolina State University, Plymouth, NC, USA

^²⁴Agriculture and Natural Resources and Statewide IPM Program, University of California, Davis, CA, USA

^²⁵Tidewater Agricultural Research and Extension Center, Virginia Tech University, Suffolk, VA, USA

^²⁶Southeast Watershed Research, USDA‐ARS, Tifton, GA, USA

^²⁷Department of Entomology, University of Kentucky, Princeton, KY, USA

^²⁸Department of Entomology, Ohio State University, Columbus, OH, USA

Correspondence to: JG Illán, Department of Entomology, Washington State University, Pullman, WA 28759, USA. E‐mail: javier.illan@wsu.edu

^✉

Corresponding author.

PMCID: PMC9804287 PMID: 36054536

Abstract

BACKGROUND

Invasive species threaten the productivity and stability of natural and managed ecosystems. Predicting the spread of invaders, which can aid in early mitigation efforts, is a major challenge, especially in the face of climate change. While ecological niche models are effective tools to assess habitat suitability for invaders, such models have rarely been created for invasive pest species with rapidly expanding ranges. Here, we leveraged a national monitoring effort from 543 sites over 3 years to assess factors mediating the occurrence and abundance of brown marmorated stink bug (BMSB, Halyomorpha halys), an invasive insect pest that has readily established throughout much of the United States.

RESULTS

We used maximum entropy models to estimate the suitable habitat of BMSB under several climate scenarios, and generalized boosted models to assess environmental factors that regulated BMSB abundance. Our models captured BMSB distribution and abundance with high accuracy, and predicted a 70% increase in suitable habitat under future climate scenarios. However, environmental factors that mediated the geographical distribution of BMSB were different from those driving abundance. While BMSB occurrence was most affected by winter precipitation and proximity to populated areas, BMSB abundance was influenced most strongly by evapotranspiration and solar photoperiod.

CONCLUSION

Our results suggest that linking models of establishment (occurrence) and population dynamics (abundance) offers a more effective way to forecast the spread and impact of BMSB and other invasive species than simply occurrence‐based models, allowing for targeted mitigation efforts. Implications of distribution shifts under climate change are discussed. © 2022 The Authors. Pest Management Science published by John Wiley & Sons Ltd on behalf of Society of Chemical Industry.

Keywords: climate change, distribution change, ecological niche models, invasive species, species distribution models

Our study is significant for the readers of Pest Management Science because it identifies and analyzes the relative importance of climatic and landscape factors driving the current and potential future population of a key insect pest in the agroecosystems across the contiguous United States.

graphic file with name PS-78-4929-g001.jpg

1. INTRODUCTION

The spread of invasive pest species is a major threat to global biodiversity and agricultural production. ¹ , ² Understanding the factors that shape the geographical distributions and population dynamics of invasive species has been a key issue in agroecology in recent decades, particularly in the current context of global change. ³ , ⁴ , ⁵ Niche‐based species distribution models are one of the most effective tools to assess dynamics of invasive species facing current environmental changes ⁶ , ⁷ , ⁸ and their use has become more popular as new powerful statistical techniques and geographical information system (GIS) tools become available. However, application of species distribution models has primarily focused on endangered species or biodiversity indicators, and seldom on insect pests in agroecosystems, particularly for pests with rapidly expanding distributions.

Many invasive insect species have wide host ranges and broad environmental tolerances, allowing them to rapidly expand their range in invaded regions. Moreover, empirical evidence has shown that climate change has caused many species, including insects, to shift their native distributions or realized niche. ⁹ , ¹⁰ , ¹¹ , ¹² Climate change also plays a key role in altering the distribution and population dynamics of invaders that are expanding their range, thus potentially exacerbating the damage caused by certain invasive agricultural pests. ¹³ , ¹⁴ , ¹⁵ However, few studies have assessed whether the environmental factors that mediate the occurrence of invasive pest species also drive population dynamics of invaders. Understanding these dynamics is critical to better assess factors that determine where invasive species are likely to be distributed, and areas where they are likely to have the greatest ecological impacts (i.e. outbreaks).

Here we assessed facs of the brown marmorated stink bug (BMSB), [Halyomorpha halys (Stål), Hemiptera: Pentatomidae], a native Asian species that was introduced to the United States in the mid‐20^th century and has since spread throughout much of North America and Europe. ¹⁶ BMSB is a generalist herbivore that feeds on nearly 170 described hosts, including many important crops that have been severely impacted across its introduced range. ¹⁶ , ¹⁷ BMSB is also known for its dispersal capacity, which allows it to move among crops, unmanaged forests and between urban and agricultural settings. ¹⁸ , ¹⁹ BMSB has proven challenging to manage, because in addition to its broad host range, it associates with humans by inhabiting dwellings in the winter, ²⁰ , ²¹ , ²² which in turn makes it more difficult to identify limiting predictors if the species is able to mitigate exposure through behavior. In recent years, scientists have focused on controlling the spread of BMSB and managing damage to crops, but populations have been detected in 46 states, it is considered to be established in 32 states across the United States, and its range continues to expand. Moreover, BMSB is considered to be both an urban nuisance pest and an agricultural pest in at least 15 states. ¹⁶ , ²³ Far from representing a problem only in the United States, it has also more recently invaded regions in Europe, where severe damage has been reported, particularly in pear and hazelnut orchards, ²⁴ , ²⁵ Chile ²⁶ and regions of the Caucasus. ²⁷

We leveraged a widespread monitoring network for BMSB that encompassed 543 sites over 3 years (2017–2019) across the contiguous United States ²⁸ to assess abiotic factors that mediated establishment (occurrence) and impact (abundance) of BMSB. We used niche‐based ecological models to assess whether the environmental factors that determine BMSB establishment were the same as, or different from, factors that mediated BMSB abundance. Such analyses could aid in mitigating the spread of BMSB into new regions (preventing new occurrences) and regulating BMSB abundance where populations are already established (managing existing populations). We also assessed how the range of BMSB may shift under future climate scenarios, given that studies have shown that BMSB populations in North America are adapting to changing climate conditions. ²⁹ , ³⁰ Overall, our study shows how widespread monitoring networks can be linked with niche‐based models to comprehensively explore factors shaping the dynamics of invasive pest species.

2. MATERIALS AND METHODS

2.1. Study system

Our study assessed occurrence and abundance of BMSB populations across the contiguous United States (Fig. 1). BMSB were sampled from 2017 to 2019 across both recently invaded regions and regions where BMSB is a well‐established agricultural pest (Fig. 1). At each site, we deployed three, clear sticky panel traps baited with the commercial BMSB Dual lure (Trécé, Inc., Adair, OK, USA). This lure contains its aggregation pheromone, a combination of the stereoisomers (3S,6S,7R,10S)‐10,11‐epoxy‐1‐bisabolen‐3‐ol and (3R,6S,7R,10S)‐10,11‐epoxy‐1 bisabolen‐3‐ol, and the pheromone synergist, methyl (E,E,Z)‐2,4,6‐decatrienoate, ³¹ and has been shown to be highly attractive to BMSB adults and nymphs. ³² , ³³ , ³⁴ , ³⁵ Baited traps were fastened on to wooden stakes spaced at 50 m intervals (after cases in the literature ²⁸ , ³³ ). Traps were placed at interfaces between natural habitat patches and crop fields that were suitable hosts for BMSB. Traps were deployed in early spring and collected once every 2 weeks until late autumn at each site to ensure we captured most of the BMSB activity period. Our network included 276 sites in 18 states in 2017, 268 sites in 17 states in 2018, and 280 sites in 17 states in 2019; from these sites we collected 24 476, 28 902, and 38 131 BMSB adult individuals, respectively, per year. We standardized the counts for modeling analyses as average BMSB adults per trap per site per week. To account for misidentification or vagrant individuals not from an established population, we excluded from analyses the sites which only one BMSB in total was recorded.

Map showing the geographical extent of the study. Sites monitored in each of three study years (2017–2019) are shown in different colors (points overlap across years). The diameter of the circles in each year represents the average weekly captures of adult *Halyomorpha halys* per three pheromone‐baited traps throughout each season at each site.

2.2. Environmental variables

To assess factors affecting BMSB occurrence and abundance, we collected climate data for all sites (Table 1). For most metrics, we acquired gridded variables describing climatic conditions from the Parameter Regression of Independent Slope Model (PRISM) ³⁶ that links meteorological data, digital elevation models, and spatial data to calculate spatially‐explicit climate. These variables were calculated per month (averaged over metrics for each day in the month) using terrain variables within a regression framework to interpolate observations from a dense network of weather stations to 800 m resolution grids. ³⁷ , ³⁸ We included minimum temperature of the coldest month (January), maximum temperature of the warmest month (July), and total precipitation in January and July We used accumulated growing degree days per year (base temperature 10°C) obtained from the US National Phenology Network (www.usanpn.org). Annual evapotranspiration was obtained as a ~1 km cell sized raster from the MOD16 Global Evapotranspiration Product. ³⁹

Table 1.

Climate and landscape variables included in the models, with associated means and ranges (in parentheses), original data resolution, data sources, and units

Variable	Code	Mean (minimum–maximum)	Original resolution	Source	Units/classes
Climatic
Maximum temperature summer	tmax_jul	30.5 (23.2–36.7)	30 arc‐seconds	PRISM	°C
Minimum temperature January	tmin_jan	−3.3 (−16.3–7.4)	30 arc‐seconds	PRISM	°C
Precipitation summer	ppt_jul	88.7 (0.1–316.4)	30 arc‐seconds	PRISM	mm
Precipitation January	ppt_jan	511.6 (11.4–2614.9)	30 arc‐seconds	PRISM	mm
Evapotranspiration	evapot	48.6 (16.6–72.0)	927 m	MODIS	mm/yr
Growing degree days	gdd	996 (547–1415)	2500 m	USA‐NPN	NA
Landscape
Distance to water bodies	hidro	19.8 (0.2–191.7)	30 arc‐seconds	GLWD	m
Distance to urban areas	urban	400.5 (0–4110.6)	30 m	USGS	m
Land cover class	cdl	categorical	30 m	Cropscape	82‐classes
Soil pH	soilph	5.9 (5.0–8.1)	1000 m	FAO	NA
Photoperiod/solar radiation	sol_pho	12 515 (1.6–255 618)	30 arc‐seconds	PRISM	kJ h m⁻²
Elevation	dem	350.1 (1.1–1702)	3 arc‐seconds	USGS	m

Open in a new tab

Note: NA, not available.

We also selected six biologically meaningful landscape variables that may affect BMSB populations (Table 1). We used a digital elevation model at 30 m spatial resolution to estimate the elevation of each site, ⁴⁰ and the land cover class each year as a categorical predictor obtained from the US Geological Survey (USGS) Cropland Data Layers. ⁴¹ , ⁴² Based on land cover data, we also calculated (i) distance to water (Euclidean distance to nearest permanent water) and (ii) distance to urban areas (Euclidean distance to nearest population area with more than 2500 residents). We included soil pH (10 km spatial resolution) as a proxy for potentially different plant communities that host BMSB. Our final landscape variable was based on the combination of photoperiod and solar radiation, calculated using the solar radiation analysis tool in ArcGIS 10.5 (ESRI, Redlands, CA, USA) and the SRTM digital elevation model.

Before analyses, all environmental layers were re‐sampled to 30 arc‐seconds grid cell size ⁴³ and variance inflation factors (VIFs) were calculated ⁴⁴ (Supporting Information, Table S1). VIFs showed no significant multicollinearity between our selected variables except maximum temperature in summer, where the VIF was slightly above a recommended threshold of 5 (5.14). However, we retained this variable in the analyses as it has been deemed ecologically important for BMSB distributions. ⁴⁵ , ⁴⁶ , ⁴⁷

2.3. Model development

Factors affecting BMSB occurrence were assessed using maximum entropy (MAXENT) models and factors affecting abundance with generalized boosted models (GBMs). MAXENT models were used as they often outperform other algorithms (i.e. GARP, CART) with presence‐only data ⁴⁸ , ⁴⁹ and are far less sensitive to sample size. ⁵⁰ MAXENT models compare presence locations with randomly selected points within a potential habitat to find the largest spread (maximum entropy) of current conditions relative to a ‘background’ of environmental variables. These analyses create maps of habitat suitability and determine contributions of specific variables on patterns of occurrence.

GBMs were created using the ‘gbm’ package ⁵¹ in R. ⁵² GBMs employ machine‐learnicbine the strengths of regression trees and boosting to fit a parsimonious model from the outputs of individual trees. ⁵³ By combining many simple models, GBMs can include different types of predictor variables and can accommodate missing data, both of which improve predictive performance while minimizing the risks of overfitting. ⁴⁸ In addition, GBMs are sufficiently flexible to include non‐linear relationships and interactions between predictors and responses. ⁵⁴ , ⁵⁵ One of the reasons we chose GBM is that, although training generally takes longer, GBM performs generally better than other similar algorithms, such as Random Forests, if parameters are tuned correctly. ⁵⁶ , ⁵⁷ This is mainly because GBMs try to add new trees that complement the already built ones, thereby providing better accuracy with fewer trees. ⁵⁵

We generated MAXENT models (with default settings) and GBMs with six climate and six landscape variables as predictors and observed BMSB occurrence and abundance at each site as responses (Fig. 2). Both occurrence and abundance models can suffer from imperfect detection, ⁵⁸ , ⁵⁹ but we did not account for detection in models for three reasons. First, data were collected by trained personnel accustomed to identifying BMSB. Second, no machine learning methods currently exist to account for imperfect detection, and machine learning methods such as GBMs enable fitting of complex structures (non‐linearities, interactions) that would be computationally challenging in an occupancy framework. Third, ‘occupancy’, after accounting for imperfect detection, is a latent variable that cannot be validated on independent data because the true state of independent data is unknown. ⁶⁰ As our primary objectives were model validation and prediction, we considered our approach to be the most appropriate. Finally, we acknowledge that the introduced models would potentially estimate a truncated niche, as they do not account for native Asian populations, and therefore might underestimate BMSB potential distribution in the United States. ⁶¹ , ⁶² To control for this effect, we calculated the realized niche occupied by native and introduced populations, as well as the predicted habitat suitability based on these two models (Supporting Information, Fig. S1). Nevertheless, because the purpose of this study was to explore and compare factors driving BMSB distribution and abundance in the contiguous United States, we argue that the introduced US BMSB populations represent the best source of information to calculate their realized niche in the United States.

The relative contribution of each climate and landscape metric in MAXENT (occurrence) models and GBMs (abundance). Values shown are the result of the ten‐fold model averaging. Scale bars show the proportion of variance explained by each factor in the combined models with both climate and landscape factors. (i) cdl: landcover class; (ii) dem: elevation; (iii) evapot: evapotranspiration; (iv) gdd: growing degree days; (v) hidro: distance to water bodies; (vi) ppt_jan: precipitation in January; (vii) ppt_jul: precipitation in July; (viii) tmax_jul: maximum temperature in July; (ix) tmin_jan: minimum temperature in January; (x) soilph: soil pH; (xi): sol_pho: interaction solar radiation and photoperiod; (xii): urban: distance to urban areas.

2.4. Model performance and evaluation

We assessed MAXENT models using area under the receiver operating characteristic curve (AUC) of plots and binary omission rates, ⁶³ , ⁶⁴ which are widely recognized as reliable performance metrics for species distribution models, ⁶⁵ , ⁶⁶ , ⁶⁷ although we also acknowledge potential limitations that have been identified (see Lobo et al. ⁶⁸ ). AUC is a threshold‐independent metric that weights omission error (predicted absence in areas of observed presence) and commission error (predicted presence in areas of observed absence) equally. AUC juxtaposes correct and incorrect predictions over a range of thresholds, and values of model fit range from 0.5 (no better than random association) to 1 (perfect fit). Useful models produce AUC values of 0.7 to 0.9 and ‘good discriminating models’ have values above 0.8. ⁶⁶ Final results of MAXENT models were the average of ten model runs. We assessed GBMs using Spearman's non‐parametric rank correlation coefficients (rho, ρ) between predicted and observed abundance values. We used rank correlation coefficients because observed BMSB counts were low at many sites (leading to deviations from normality) and due to potential non‐linear relationships between predicted and observed abundance. To account for data independence, we randomly selected 70% (573) of sites to train the models and 30% (245) for validation. ⁶⁹

2.5. Predicting BMSB distribution under future climate scenarios

Our final analysis involved predicting changes in habitat suitability for BMSB under future climate change scenarios. We assessed these scenarios to project how BMSB populations might expand in the near and long‐term future. We developed climate‐only distribution MAXENT models and calculated the environmental niche suitability under two climate change scenarios for 2080. ⁷⁰ Climate data for the IPCC (Intergovernmental Panel on Climate Change) scenarios considered were generated with ClimateNA version 5.10 software. ⁷¹ We chose the A1B and B1 scenarios, which are moderate scenarios often used in recent global change studies to project future climatic conditions. ⁷² , ⁷³ , ⁷⁴ , ⁷⁵ Scenario A1B assumed a balanced global energy portfolio, generating moderate emission levels and describes a future world of very rapid economic growth. Scenario B1 described a world with similar global population than A1B but transitioning to sustainable energy, generating lower greenhouse gas emissions. ⁷⁶ , ⁷⁷

3. RESULTS

3.1. Model performance and validation

Model validation showed both MAXENT (occurrence) models and GBMs (abundance) accurately captured the current distribution of BMSB in the United States. The three MAXENT models with climate variables only, landscape variables only, or climate and landscape variables combined had an AUC of 0.93, 0.89, and 0.95, respectively. We also built a MAXENT model with the top three climate and top three landscape variables, which had an AUC of 0.94. The climate‐only, landscape‐only, and combined GBMs had a ρ of 0.44, 0.48, and 0.44, respectively. The reduced combined model had a ρ of 0.52 (Table S2).

3.2. Relative contribution of variables

Although both MAXENT models and GBMs effectively predicted BMSB occurrence and abundance, the relative contribution of factors across these models differed considerably (Figs 2, S2 and S3). This held whether we considered the single top variable in each model, or whether a variable was one of the three top predictors. In the MAXENT models, the distance to urban areas and January precipitation were by far the most important factors mediating BMSB occurrence, accounting for more than 80% of the variance, followed by land cover and evapotranspiration (Figs 2 and S4). In the GBMs, in contrast, BMSB abundance was primarily mediated by evapotranspiration and the interaction between photoperiod and solar radiation, followed by soil pH and land cover (Fig. 2).

3.3. Predicted effects of climate change on the distribution of BMSB

Our models showed that habitat suitability for BMSB will likely change considerably under future climate scenarios. Under both moderate scenarios considered, areas of high suitability habitat are expected to increase in future decades (Fig. 3). In general, we found that habitat suitability will increase moving northward over time, especially under the A1B scenario (Fig. 3(B)). On the west coast, we also showed that habitat would become more suitable moving inland in future years, particularly in the northwest. However, there were some areas that are expected to decline in suitability by the 2080s, such as some parts of the Sacramento Valley of California (Fig. 3). Overall, our models suggest that currently 2.3% of the contiguous United States has a BMSB habitat suitability above 0.5, and less than 0.01% has suitability above 0.7. Those values are predicted to increase to 13% and 7%, respectively, under the A1B and B1 scenarios (Table S3).

Predicted change in habitat suitability for *Halyomorpha halys* in the 2080s compared to current conditions under the climate scenarios (A) B1 (emphasis on global solutions to economic, social, and environmental sustainability, but without additional climate initiatives) and (B) A1B (a future world of very rapid economic growth, low population growth and rapid introduction of new and more efficient technology) (both rated as moderate). Both panels show changes in climatic habitat suitability from cold (decrease in suitability) to warm colors (increase in suitability). Green tonalities denote no change in habitat suitability for BMSB.

4. DISCUSSION

We show that ecological niche models, created with data from a nationwide monitoring effort, accurately predicted the occurrence and abundance of BMSB populations throughout the United States. These results conform with those from studies showing high predictive ability of ecological niche models that are trained and tested over short time periods. ⁷⁸ , ⁷⁹ , ⁸⁰ While both types of models were accurate, the most important factors mediating BMSB occurrence differed from those regulating BMSB abundance (Fig. 2), possibly reflecting different processes involved in the establishment of an invasive species, as opposed to those driving regional population dynamics. ⁸¹ , ⁸² For example, the most important predictor for BMSB occurrence was proximity to populated areas, which may reflect human‐assisted transport of BMSB to urban locations, the availability of human structures that can serve as overwintering sites for BMSB, and the diversity of ornamental plants (e.g. Ailanthus altissima) acting as additional resources to BMSB in the spring. However, it is important to note that the different nature of the two model approaches and the interaction of the presence and abundance data with the predictors can also influence the ranking of the variables. Urban areas might thus provide better availability of hosts for BMSB feeding and oviposition in spring, which may in turn increase its abundance and contribute to its subsequent spread in nearby areas. However, for populations that are already established, climate factors had the greatest impacts on abundance. ⁸³ Overall, these results suggest that predicting the distribution and population dynamics of an invasive species requires complementary modeling approaches to assess both establishment and population regulation.

BMSB has been spreading rapidly throughout the United States, and its population status still has not reached complete equilibrium, as it has in the native Asian range. ⁸⁴ Our models showed that the most suitable areas for BMSB establishment are in the Mid‐Atlantic region and surrounding the Great Lakes, with additional favorable regions in south‐eastern states and valleys of the West Coast (Fig. S4). These regions include highly productive croplands that are vulnerable to significant economic losses due to BMSB each year. Importantly, our models predicted regions of high suitability where BMSB has not yet been detected, suggesting that its further spread and greater impacts on agriculture may be anticipated (e.g. southern California, Treasure Valley, Idaho). However, our models also suggest that BMSB is unlikely to establish a continuous distribution throughout the contiguous United States, given the lack of suitable habitat throughout much of the central and southern United States (Figs 3 and S4). It is worth noting that a previous model by Kriticos et al. ⁶ had identified certain regions of Florida and the Gulf coast as favorable for BMSB colonization whereas our models do not show those areas at high risk of invasion. Our predictions seem to be at least partially supported by observations from collaborators in Florida, who have detected adults of BMSB but very rarely found nymphs or eggs, suggesting BMSB have not established their population in Florida and might not be a climatically suitable region. ⁸⁵

When analyzing factors affecting BMSB occurrence, we showed that January precipitation was the most important climatic factor, followed by evapotranspiration and minimum temperature. Interestingly, BMSB are chill‐intolerant, with rapidly increasing mortality at temperatures below 10 °C. ²⁹ , ⁸⁶ The greater role of precipitation, compared with temperature, however, has been reported for multiple other insect taxa in North America ⁴ , ⁸⁷ and for some non‐insect taxa. ⁹ The role of higher winter precipitation and humidity potentially leading to a lower mortality of BMSB has been reported to be important factors for other insect species. ⁸⁸ , ⁸⁹ Indeed, cold, dry air is known to exacerbate desiccation stress for insect species, ⁹⁰ and as BMSB overwinters in aggregations, a mechanism considered to help reduce desiccation stress, ⁹¹ winter precipitation is likely another factor that is important to survivorship. Increased precipitation during January likely increases the relative humidity and reduces evapotranspiration from overwintering bugs. However, BMSB have only been found overwintering in dry, tight locations in nature in location such as beneath the bark and deep within the wood of dead, standing trees, but never in moist locations such as leaf litter or damp, downed trees lying on the ground based on extensive surveys by detector canine and human surveyors. ⁸⁴ , ⁹¹ Overwintering BMSB also choose to shelter in dry locations over moist sites in significantly greater numbers. ⁹² Studies of aggregating insect species, specific abiotic conditions found within overwintering sites and their impact on subsequent insect survivorship are rare as articulated by Susset et al. ⁹³ ; their study of the ladybird beetle Hippodamia undecimnotata (Schneider) revealed abiotic conditions within aggregation sites did not provide optimal conditions to combat unfavorable temperatures or the potential for fungal pathogen infection. Finally, higher humidity during this time of the year might make natural hosts more abundant and more suitable, increasing BMSB abundance and their potential to establish in crops the following season. It is clear that more detailed studies of factors influencing overwintering survivorship success of BMSB are warranted.

Distance to urban areas was the most important landscape variable affecting BMSB occurrence, accounting for almost 41% of the variance, making it the most important variable overall. Factors associated with human‐dominated landscapes, including roads and urban areas, similarly affect other invasive pests ⁹⁴ , ⁹⁵ and non‐pest insects. ⁹⁶ Our results suggest that BMSB is most likely to establish in agricultural regions where it already represents a nuisance pest, ⁹⁷ because individuals that overwinter in human structures may disperse into crop fields each spring, and also because higher diversity of crops provide more options for optimal BMSB development. ⁹⁸ However, the proximity to populated areas was not an important variable predicting BMSB abundance. This suggests that once populations establish, they may not need to rely on humans for overwintering or transport, and population dynamics are primarily driven by climate. In other words, if a BMSB population has already established, the proximity of human structures is no longer a critical factor for BMSB abundance, which become more influenced by the regional climate.

In contrast to MAXENT models, GBMs revealed that evapotranspiration had the strongest impacts on BMSB abundance, followed by the interaction of solar radiation and photoperiod. Together, these two predictors accounted for more than 50% of the variance in abundance. This suggests water availability, and sunlight exposure, are limiting factors regulating BMSB populations. Other studies have similarly shown that moisture can strongly affect sap‐feeding insects, both in their adult stage and especially during nymphal development. ⁴ , ⁹⁹ , ¹⁰⁰ , ¹⁰¹ Reduced evapotranspiration and the associated temperature increase can impact survival of insect herbivores that are unable to move to more suitable sites, ¹⁰² and could prove to be a factor in BMSB nymphal survival in hotter and drier locations. Moisture is also a factor determining the availability and suitability of host plants, increasing BMSB success after they break diapause and start feeding. This in turn enhances the potential for population increase and ultimately to establish in crop hosts the following season. Policymakers and growers might be able to use potential evapotranspiration to better predict the risk of BMSB impact in the following year, allowing them to make more proactive management decisions.

By analyzing future climate scenarios, we showed that BMSB populations have a large potential to continue to expand within the United States, and particularly northward. Similar expected shifts have been reported for other insect species in the Northern Hemisphere. ¹⁰³ , ¹⁰⁴ , ¹⁰⁵ The change in future habitat suitability was similar under both climate scenarios, and both reflected steady increases in the proportion of highly suitable habitat for BMSB by 2080 compared with 2020 (Fig. 3). According to our models, BMSB may expand its range considerably in future decades, threatening crops in regions where it has not yet been detected. Nevertheless, our models can be used to estimate areas at highest risk of potential spread so that mitigation and eradication efforts might be implemented in new regions. As reported by Stoeckli et al., ¹⁴ BMSB is also expected to expand its current distribution in Europe, where extensive range expansion into higher altitudes and an increased number of generations per year is predicted under several climate change scenarios.

While our models accurately described BMSB occurrence and abundance, we note that substantial variance remained unexplained. We suggest that biotic factors, such as the presence and abundance of competitors and natural enemies, also likely affect BMSB distributions. ⁸⁰ , ¹⁰⁶ For example, recent studies show that parasitoid wasps can help suppress BMSB populations. ¹⁰⁷ , ¹⁰⁸ , ¹⁰⁹ Similarly, other native stink bug species may compete with BMSB for resources and limit their spread. ¹¹⁰ , ¹¹¹ Another source of variance arises from the scale of the predictors, as most variables included in ecological niche models do not reflect microclimates that can provide refugia for insects ¹¹² ; however, such datasets are not yet available. However, the capacity of BMSB to adapt physiologically to new environmental conditions may affect the interpretation of predictions. Finally, variation in management strategies for BMSB, such as the intensity of pesticide use, might affect regional distributions. ¹¹³ Despite these confounding sources of variance, we showed that our models could accurately capture variability in BMSB occurrence and abundance at the national scale of the BMSB invasion. Finally, while it is common to consider climate and landscape factors independently, they may interact in some contexts. For example, over longer timescales, insect populations that are invading new areas due to anthropogenic land‐use change may be more likely to behaviorally adapt to climate change. However, the connectivity of landscapes may be reduced due to landscape fragmentation, decreasing the potential for population establishment and range shifts due to climate change. Of course, a major step forward will be to efficiently assess the interactive effects of climate and landscape factors, but this has proven too challenging to date, at least for most observational studies.

Our study highlights the applicability of ecological niche models for assessing the potential spread and impacts of an invasive pest species in agroecosystems. Such models can in turn be used to guide mitigation and eradication strategies before invaders establish in new regions, while maximizing the effectiveness of limited resources available for tracking invasive spread. Moreover, we addressed the complementarity of two modeling approaches, occurrence and abundance models, that offer different information related to managing an invasive pest insect. While the distance to populated areas seems to be key for BMSB occurrence, climatic factors, particularly those related to water availability, were the most important drivers of BMSB abundance. This study leveraged a rare, large‐scale monitoring effort across most of the contiguous United States that included sampling sites across a wide range of environmental conditions. Thus, results presented here are robust, and predictions made at such a continental level are likely more reliable than could be achieved with a smaller scale study. Our models thus represent a powerful tool for ecologists and policymakers to better understand expansion of an invasive insect pest across diverse agroecosystems.

Supporting information

Figure S1. (a) Realized climate niche spaces occupied by native (green volume) and introduced (red volume) brown marmorated stink bug (BMSB) populations. Realized niches were presented by minimum ellipsoid volumes in reduced dimensions (i.e. principal components 1–3). (b, c) Comparison of habitat suitability based on native (b) and US introduced (c) population of BMSB. Spearman's rank correlation between predicted habitat suitability also shown.

Figure S2. Relative contribution of each predictor in climatic (A) and landscape models (B). Results shown for MAXENT (occurrence) models and Generalized Boosted (abundance) models are the average of the ten‐fold model. Scale bars show the proportion of variance explained by each factor in each model type. Abbreviations are described in Table 1.

Figure S3. Response curves for the top three climate factors in the MAXENT models of brown marmorated stink bug: (a) precipitation in January; (b) evapotranspiration; (c) minimum temperature in January. The top three landscape variables: (d) distance to urban areas; (e) landscape cover class; (f) soil pH. Solid red lines represent the average of the ten‐fold model runs and blue lines show the range of predictions from the individual model runs.

Figure S4. Map showing the predicted current habitat suitability for Halyomorpha halys in the contiguous United States based on the combined model that included the three most important climatic variables and three most important landscape variables. Higher suitability values (in warmer colors) represent habitats more likely to be occupied by brown marmorated stink bug (BMSB) populations.

Table S1. Variance inflation factors (VIFs) were calculated to assess multicollinearity among variables in all models. VIF > 5 are generally considered problematic.

Table S2. Performance metrics for each model type. Maxent model performance was measured via AUC (area under the receiver operating characteristic curve), and generalized boosted model (GBM) performance was calculated as a non‐parametric correlation (Spearman's rho) between predicted and observed Halyomorpha halys abundance. Models were validated with 30% of the data, and the remaining 70% was used for training. Values shown are the average of ten replicates of each model in all cases.

Table S3. Change in extension of suitable habitat for Halyomorpha halys under current conditions and under two IPCC climate change scenarios (B1 and A1B). Values shown are calculated as percentage of pixels that showed values for habitat suitability above each threshold.

Click here for additional data file.^{(2.5MB, docx)}

ACKNOWLEDGEMENTS

This work was supported by USDA‐NIFA‐SCRI grant no. 2016‐51181‐25409. The authors also thank the many students, colleagues and research assistants who helped collect BMSB data from the field and assist in the data analyses.

DATA AVAILABILITY STATEMENT

The data that support the findings of this study are available from the corresponding author upon reasonable request.

REFERENCES

1. Mack RN, Simberloff D, Lonsdale WM, Evans H, Clout M and Bazzaz FA, Biotic invasions: causes, epidemiology, global consequences, and control. Ecol Appl 10:689–710 (2000). [Google Scholar]
2. Venette RC and Hutchison WD, Invasive insect species: global challenges, strategies & opportunities. Front Insect Sci 1:650520 (2021). [DOI] [PMC free article] [PubMed] [Google Scholar]
3. Guisan A and Zimmermann NE, Predictive habitat distribution models in ecology. Ecol Model 135:147–186 (2000). [Google Scholar]
4. Gutiérrez Illán J, Bloom EH, Wohleb CH, Wenninger EJ, Rondon SI, Jensen AS et al., Landscape structure and climate drive population dynamics of an insect vector within intensely managed agroecosystems. Ecol Appl 30:e02109 (2020). [DOI] [PubMed] [Google Scholar]
5. Parmesan C, Ryrholm N, Stefanescu C, Hill TK, Thomas CD, Descimon H et al., Poleward shifts in geographical ranges of butterfly species associated with regional warming. Nature 399:579–583 (1999). [Google Scholar]
6. Kriticos DJ, Kean JM, Phillips CB, Senay SD, Acosta H and Haye T, The potential global distribution of the brown marmorated stink bug, Halyomorpha halys, a critical threat to plant biosecurity. J Pest Sci 90:1033–1043 (2017). [Google Scholar]
7. Mainali KP, Warren DL, Dhileepan K, McConnachie A, Strathie L, Hassan G et al., Projecting future expansion of invasive species: comparing and improving methodologies for species distribution modeling. Glob Chang Biol 21:4464–4480 (2015). [DOI] [PubMed] [Google Scholar]
8. Zhu G, Illan JG, Looney C and Crowder DW, Assessing the ecological niche and invasion potential of the Asian giant hornet. Proc Natl Acad Sci 117:14730–14737 (2020). [DOI] [PMC free article] [PubMed] [Google Scholar]
9. Gutiérrez Illán J, Thomas CD, Jones JA, Wong WK, Shirley SM and Betts MG, Precipitation and winter temperature predict long‐term range‐scale abundance changes in Western north American birds. Glob Chang Biol 20:3351–3364 (2014). [DOI] [PubMed] [Google Scholar]
10. Parmesan C and Yohe G, A globally coherent fingerprint of climate change impacts across natural systems. Nature 421:37–42 (2003). [DOI] [PubMed] [Google Scholar]
11. Thomas CD and Lennon JJ, Birds extend their ranges northwards. Nature 399:213–213 (1999). [Google Scholar]
12. Wilson RJ, Gutiérrez D, Gutiérrez Illan J, Martínez D, Agudo R and Monserrat VJ, Changes to the elevational limits and extent of species ranges associated with climate change. Ecol Lett 8:1138–1146 (2005). [DOI] [PubMed] [Google Scholar]
13. Bale JS, Masters GJ, Hodkinson ID, Awmack C, Bezemer TM, Brown VK et al., Herbivory in global climate change research: direct effects of rising temperature on insect herbivores. Glob Chang Biol 8:1–16 (2002). [Google Scholar]
14. Stoeckli S, Felber R and Haye T, Current distribution and voltinism of the brown marmorated stink bug, Halyomorpha halys, in Switzerland and its response to climate change using a high‐resolution CLIMEX model. Int J Biometeorol 64:2019–2032 (2020). [DOI] [PMC free article] [PubMed] [Google Scholar]
15. Van Der Wal J, Shoo LP, Johnson CN and Williams SE, Abundance and the environmental niche: environmental suitability estimated from niche models predicts the upper limit of local abundance. Am Nat 174:282–291 (2009). [DOI] [PubMed] [Google Scholar]
16. Leskey TC and Nielsen AL, Impact of the invasive brown marmorated stink bug in North America and Europe: history, biology, ecology, and management. Annu Rev Entomol 63:599–618 (2018). [DOI] [PubMed] [Google Scholar]
17. Rice KB, Bergh CJ, Bergmann EJ, Biddinger DJ, Dieckhoff C, Dively G et al., Biology, ecology, and management of brown marmorated stink bug (Hemiptera: Pentatomidae). J Integr Pest Manage 5:1–13 (2014). [Google Scholar]
18. Lee D‐H and Leskey TC, Flight behavior of foraging and overwintering brown marmorated stink bug, Halyomorpha halys (Hemiptera: Pentatomidae). Bull Entomol Res 105:566–573 (2015). [DOI] [PMC free article] [PubMed] [Google Scholar]
19. Wiman NG, Walton VM, Shearer PW, Rondon SI and Lee JC, Factors affecting flight capacity of brown marmorated stink bug, Halyomorpha halys (Hemiptera: Pentatomidae). J Pest Sci 88:37–47 (2014). [Google Scholar]
20. Hahn NG, Kaufman AJ, Rodriguez‐Saona C, Nielsen AL, LaForest J and Hamilton GC, Exploring the spread of brown marmorated stink bug in New Jersey through the use of crowdsourced reports. Am Entomol 62:36–45 (2016). [Google Scholar]
21. Inkley DB, Characteristics of home invasion by the brown marmorated stink bug (Hemiptera: Pentatomidae). J Entomol Sci 47:125–130 (2012). [Google Scholar]
22. Watanabe M, Arakawa R, Shinagawa Y and Okazawa T, Overwintering flight of brown‐marmorated stink bug, Halyomorpha mista, to the buildings. Med Entomol Zool 45:25–31 (1994). [Google Scholar]
23. Northeastern IPM Center . 2014. Stink bug detectives. http://www.stopbmsb.org/stink-bug-bulletin/stink-bug-detectives/ [May 2020].
24. Gariepy TD, Bruin A, Haye T, Milonas P and Vétek G, Occurrence and genetic diversity of new populations of Halyomorpha halys in Europe. J Pest Sci 88:451–460 (2015). [Google Scholar]
25. Haye T and Weber DC, Special issue on the brown marmorated stink bug, Halyomorpha halys: an emerging pest of global concern. J Pest Sci 90:987–988 (2017). [Google Scholar]
26. Faundez EI and Rider DA, The brown marmorated stink bug Halyomorpha halys (Stal, 1855) (Heteroptera: Pentatomidae) in Chile. Arq Entomol 17:305–307 (2017). [Google Scholar]
27. Gapon DA, First records of the brown marmorated stink bug Halyomorpha halys (Stål, 1855) (Heteroptera, Pentatomidae) in Russia, Abkhazia, and Georgia. Entomol Rev 96:1086–1088 (2016). [Google Scholar]
28. Acebes‐Doria AL, Agnello AM, Blaauw BR, Buntin GD, Alston DG, Beers EH et al., Season‐long monitoring of the brown marmorated stink bug throughout the United States using commercially available traps and lures. J Econ Entomol 112:159–171 (2020). [DOI] [PubMed] [Google Scholar]
29. Cira TM, Venette RC, Aigner J, Kuhar T, Mullins DE, Gabbert SE et al., Cold tolerance of Halyomorpha halys (Hemiptera: Pentatomidae) across geographic and temporal scales. Environ Entomol 45:484–491 (2016). [DOI] [PubMed] [Google Scholar]
30. Govindan BN and Hutchison WD, Influence of temperature on age‐stage, two‐sex life tables for a Minnesota‐acclimated population of the brown marmorated stink bug (Halyomorpha halys). Insects 11:108 (2020). [DOI] [PMC free article] [PubMed] [Google Scholar]
31. Weber DC, Leskey TC, Walsh GC and Khrimian A, Synergy of aggregation pheromone with methyl (E,E,Z)‐2,4,6‐decatrienoate in attraction of Halyomorpha halys (Hemiptera: Pentatomidae). J Econ Entomol 107:1061–1068 (2014). [DOI] [PubMed] [Google Scholar]
32. Khrimian A, Zhang A, Weber DC, Ho HY, Aldrich JR, Vermillion KE et al., Discovery of the aggregation pheromone of the brown marmorated stink bug (Halyomorpha halys) through the creation of stereoisomeric libraries of 1‐bisabolen‐3‐ols. J Nat Prod 77:1708–1717 (2014). [DOI] [PubMed] [Google Scholar]
33. Leskey TC, Agnello A, Bergh JC, Dively GP, Hamilton GC, Jentsch P et al., Attraction of the invasive Halyomorpha halys (Hemiptera: Pentatomidae) to traps baited with semiochemical stimuli across the United States. J Econ Entomol 44:746–756 (2015a). [DOI] [PubMed] [Google Scholar]
34. Leskey TC, Khrimian A, Weber DC, Aldrich JC, Short BD, Lee DH et al., Behavioral responses of the invasive Halyomorpha halys (Stål) to traps baited with stereoisomeric mixtures of 10, 11‐epoxy‐1‐bisabolen‐3‐ol. J Chem Ecol 41:418–429 (2015b). [DOI] [PMC free article] [PubMed] [Google Scholar]
35. Weber DC, Morrison WR, Khrimian A, Rice KB, Leskey TC, Rodriguez‐Saona C et al., Chemical ecology of Halyomorpha halys: discoveries and applications. J Pest Sci 90:989–1008 (2017). [Google Scholar]
36. Daly C, Gibson WP, Taylor GH, Johnson GL and Pasteris P, A knowledge‐based approach to the statistical mapping of climate. Clim Res 22:99–113 (2002). [Google Scholar]
37. Daly C, Taylor GH, Gibson WP, Parzybok TW, Johnson GL and Pasteris PA, High‐quality spatial climate data sets for the United States and beyond. Trans ASAE 43:1957–1962 (2000). [Google Scholar]
38. Daly C, Halbleib M, Smith JI, Gibson WP, Doggett MK, Taylor GH et al., Physiographically‐sensitive mapping of temperature and precipitation across the conterminous United States. Int J Climatol 28:2031–2064 (2008). [Google Scholar]
39. Mu, Q. , Zhao M., and Steven W.. 2013. Algorithm Theoretical Basis Document: MODIS Global Terrestrial Evapotranspiration (ET) Product (NASA MOD16A2/A3) Collection 5. NASA Headquarters 2013. http://files.ntsg.umt.edu/data/NTSG_Products/MOD16/.
40. NASA Shuttle Radar Topography Mission (SRTM) (2013). Shuttle Radar Topography Mission (SRTM) Global. Distributed by OpenTopography. 10.5069/G9445JDF. [7 January 2021]. [DOI]
41. Han W, Yang Z, Di L and Mueller R, CropScape: a web service based application for exploring and disseminating US continuous geospatial cropland data products for decision support. Comput Electron Agric 84:111–123 (2012). [Google Scholar]
42. NLCD (2010) National land cover database (2010). USEPA, Washington, DC. Available at: www.epa.gov/mrlc/nlcd-2001.html [03 March 2018].
43. Merow C, Smith MJ and Silander JA Jr, A practical guide to MaxEnt for modeling species' distributions: what it does, and why inputs and settings matter. Ecography 36:1058–1069 (2013). [Google Scholar]
44. Graham MH, Confronting multicollinearity in ecological multiple regression. Ecology 84:2809–2815 (2003). [Google Scholar]
45. Cobos ME, Peterson AT, Osorio‐Olvera L and Jiménez‐García D, An exhaustive analysis of heuristic methods for variable selection in ecological niche modeling and species distribution modeling. Eco Inform 53:100983 (2019). [Google Scholar]
46. González‐Suárez M and Revilla E, Variability in life‐history and ecological traits is a buffer against extinction in mammals. Ecol Lett 16:242–251 (2013). [DOI] [PubMed] [Google Scholar]
47. O'Brien R, A caution regarding rules of thumb for variance inflation factors. Qual Quant 41:673–690 (2007). [Google Scholar]
48. Elith J, Graham H, Anderson R, Dudík M, Ferrier S, Guisan A et al., Novel methods improve prediction of species' distributions from occurrence data. Ecography 29:129–151 (2006). [Google Scholar]
49. Phillips SJ and Dudík M, Modeling of species distributions with Maxent: new extensions and a comprehensive evaluation. Ecography 31:161–175 (2008). [Google Scholar]
50. Wisz MS, Hijmans RJ, Li J, Peterson AT, Graham CH, Guisan A et al., Effects of sample size on the performance of species distribution models. Divers Distrib 14:763–773 (2008). [Google Scholar]
51. Greenwell, B. , Boehmke B., and Cunningham J.. 2019. Developers (https://github.com/gbm‐developers). Generalized Boosted Regression Models. Package ‘gbm’ R package version, 2(5).
52. R Development Core Team , R: A Language and Environment for Statistical Computing, Vol. 2018. R Foundation for Statistical Computing, Vienna, Austria: (2018). [Google Scholar]
53. Hijmans, RJ , and Elith J. Species Distribution Modeling with R. R CRAN Project; (2013). [Google Scholar]
54. Death G, Boosted trees for ecological modeling and prediction. Ecology 88:243–225 (2007). [DOI] [PubMed] [Google Scholar]
55. Elith J, Leathwick JR and Hastie T, A working guide to boosted regression trees. J Anim Ecol 77:802–813 (2008). [DOI] [PubMed] [Google Scholar]
56. Friedman JH. Stochastic Gradient Boosting. Computational Statistics and Data Analysis 38(4):367–378 (2002). [Google Scholar]
57. Hastie T, Tibshirani R, and Friedman J. Random forests. In The elements of statistical learning (pp. 587–604). Springer, New York, NY; (2009). [Google Scholar]
58. Kéry M, Towards the modelling of true species distributions. J Biogeogr 38:617–618 (2011). [Google Scholar]
59. MacKenzie DI, Nichols JD, Hines JE, Knutson MG and Franklin AB, Estimating site occupancy, colonization, and local extinction when a species is detected imperfectly. Ecology 84:2200–2207 (2003). [Google Scholar]
60. Welsh AH, Lindenmayer DB and Donnelly CF, Fitting and interpreting occupancy models. PLoS One 8:e52015 (2013). [DOI] [PMC free article] [PubMed] [Google Scholar]
61. Guisan A and Thuiller W, Predicting species distribution: offering more than simple habitat models. Ecol Lett 10:435–435, 1009 (2005). [DOI] [PubMed] [Google Scholar]
62. Peterson AT, Soberón J, Pearson RG, Anderson RP, Martínez‐Meyer E, Nakamura M et al., Ecological Niches and Geographic Distributions: A Modeling Perspective. Princeton University Press, Princeton, New Jersey, USA: (2011). [Google Scholar]
63. Manel S, Williams HC and Ormerod SJ, Evaluating presence–absence models in ecology: the need to account for prevalence. J Appl Ecol 38:921–931 (2001). [Google Scholar]
64. McPherson JM, Jetz W and Rogers DJ, The effects of species' range sizes on the accuracy of distribution models: ecological phenomenon or statistical artefact? J Appl Ecol 41:811–823 (2004). [Google Scholar]
65. Elith J, Phillips TH, Dudík M, Chee YE and Yates CJ, A statistical explanation of MaxEnt for ecologists. Divers Distrib 17:43–57 (2011). [Google Scholar]
66. Fielding AH and Bell JF, A review of methods for the assessment of prediction errors in conservation presence/absence models. Environ Conserv 1:38–49 (1997). [Google Scholar]
67. Phillips SJ, Anderson RP and Schapire RE, Maximum entropy modeling of species geographic distributions. Ecol Model 190:231–259 (2006). [Google Scholar]
68. Lobo JM, Jiménez‐Valverde A and Real R, AUC: a misleading measure of the performance of predictive distribution models. Glob Ecol Biogeogr 17:145–151 (2008). [Google Scholar]
69. Bahn V and McGill BJ, Testing the predictive performance of distribution models. Oikos 122:321–331 (2013). [Google Scholar]
70. Knutti R and Sedláček J, Robustness and uncertainties in the new CMIP5 climate model projections. Nat Clim Change 3:369–373 (2013). [Google Scholar]
71. Wang T, Hamann A, Spittlehouse DL and Carroll C, Locally downscaled and spatially customizable climate data for historical and future periods for North America. PLoS One 11:e0156720 (2016). [DOI] [PMC free article] [PubMed] [Google Scholar]
72. Grundel R, Frohnapple KJ, Zaya DN, Glowacki GA, Weiskerger CJ, Patterson TA et al., Geographic coincidence of richness, mass, conservation value, and response to climate of US land birds. Ecol Appl 24:791–811 (2014). [DOI] [PubMed] [Google Scholar]
73. Sohl TL, Sayler KL, Bouchard MA, Reker RR, Friesz AM, Bennett SL et al., Spatially explicit modeling of 1992‐2100 land cover and forest stand age for the conterminous United States. Ecol Appl 24:1015–1036 (2014). [DOI] [PubMed] [Google Scholar]
74. Tabor K and Williams JW, Globally downscaled climate projections for assessing the conservation impacts of climate change. Ecol Appl 20:554–565 (2010). [DOI] [PubMed] [Google Scholar]
75. Ziter C, Robinson EA and Newman JA, Climate change and voltinism in Californian insect pest species: sensitivity to location, scenario and climate model choice. Glob Chang Biol 18:2771–2780 (2012). [DOI] [PubMed] [Google Scholar]
76. Intergovernmental panel on climate change (IPCC), in Climate Change 2014: IPCC. Climate Change 2014: Synthesis Report. Contribution of Working Groups I, II and III to the Fifth Assessment Report of the Intergovenmental Panel on Climate Change. Cambridge University Press, Cambridge: (2014). [Google Scholar]
77. Nakicenovic N, Alcamo J, Davis G, de Vries B, Fenhann J, Gaffin S et al., Special Report on Emissions Scenarios: A Special Report of Working Group III of the Intergovernmental Panel on Climate Change. Cambridge University Press, Cambridge, UK: (2000). [Google Scholar]
78. La Sorte FA and Jetz W, Tracking of climatic niche boundaries under recent climate change. J Anim Ecol 81:914–925 (2012). [DOI] [PubMed] [Google Scholar]
79. Renwick AR, Massimino D, Newson SE, Chamberlain DE, Pearce‐Higgins JW and Johnston A, Modelling changes in species' abundance in response to projected climate change. Divers Distrib 18:121–132 (2012). [Google Scholar]
80. Zhu G, Bu W, Gao Y and Liu G, Potential geographic distribution of brown marmorated stink bug invasion (Halyomorpha halys). PLoS One 7:31246 (2012). [DOI] [PMC free article] [PubMed] [Google Scholar]
81. Aguirre‐Gutiérrez J, Carvalheiro LG, Polce C, van Loon EE, Raes N, Reemer M et al., Fit‐for‐purpose: species distribution model performance depends on evaluation criteria–Dutch hoverflies as a case study. PLoS One 8:63708 (2013). [DOI] [PMC free article] [PubMed] [Google Scholar]
82. Roura‐Pascual N, Brotons L, Peterson AT and Thuiller W, Consensual predictions of potential distributional areas for invasive species: a case study of argentine ants in the Iberian Peninsula. Biol Invasions 11:1017–1031 (2009). [Google Scholar]
83. Nielsen AL, Chen S and Fleischer SJ, Coupling developmental physiology, photoperiod, and temperature to model phenology and dynamics of an invasive heteropteran, Halyomorpha halys . Front Physiol 7:165 (2016). [DOI] [PMC free article] [PubMed] [Google Scholar]
84. Lee D‐H, Short BD, Joseph SV, Bergh JC and Leskey TC, Review of the biology, ecology, and management of Halyomorpha halys (Hemiptera: Pentatomidae) in China, Japan, and the Republic of Korea. Environ Entomol 42:627–641 (2013). [DOI] [PubMed] [Google Scholar]
85. Khadka A, Hodges AC, Leppla NC and Tillman PG, Halyomorpha halys (Stål)(Hemiptera: Pentatomidae) nymph survival and adult feeding preferences for crop plants in Florida. Fla Entomol 104:136–139 (2021). [Google Scholar]
86. Scaccini D, Vanishvili L, Tirello P, Walton VM, Duso C and Pozzebon A, Lethal and sub‐lethal effects of low‐temperature exposures on Halyomorpha halys (Hemiptera: Pentatomidae) adults before and after overwintering. Sci Rep 10:1–9 (2020). [DOI] [PMC free article] [PubMed] [Google Scholar]
87. McLaughlin JF, Hellmann JJ, Boggs CL and Ehrlich PR, The route to extinction: population dynamics of a threatened butterfly. Oecologia 132:538–548 (2002). [DOI] [PubMed] [Google Scholar]
88. Moore MV and Lee RE Jr, Surviving the big chill: overwintering strategies of aquatic and terrestrial insects. Am Entomol 37:111–118 (1991). [Google Scholar]
89. Overgaard J and MacMillan HA, The integrative physiology of insect chill tolerance. Annu Rev Physiol 79:187–208 (2017). [DOI] [PubMed] [Google Scholar]
90. Danks HV, Dehydration in dormant insects. J Insect Physiol 46:837–852 (2000). [DOI] [PubMed] [Google Scholar]
91. Ciancio JJ, Turnbull KF, Gariepy TD and Sinclair BJ, Cold tolerance, water balance, energetics, gas exchange, and diapause in overwintering brown marmorated stink bugs. J Insect Physiol 128:104171 (2021). [DOI] [PubMed] [Google Scholar]
92. Cullum JP, Nixon LJ, Morrison WR, Raupp MJ, Shrewsbury PM, Venugopal PD et al., Influence of landscape factors and abiotic conditions on dispersal behavior and overwintering site selection by Halyomorpha halys (Hemiptera: Pentatomidae). J Econ Entomol 113:2016–2021 (2020). [DOI] [PubMed] [Google Scholar]
93. Susset EC, Hemptinne JL and Magro A, Overwintering sites might not be safe haven for Hippodamia undecimnotata (Schneider) (Coleoptera: Coccinellidae). Coleopterists Bull 71:556–564 (2017). [Google Scholar]
94. Prasad AM, Iverson LR, Peters MP, Bossenbroek JM, Matthews SN, Sydnor TD et al., Modeling the invasive emerald ash borer risk of spread using a spatially explicit cellular model. Landsc Ecol 25:353–369 (2010). [Google Scholar]
95. Tonini F, Hochmair HH, Scheffrahn RH and Deangelis DL, Simulating the spread of an invasive termite in an urban environment using a stochastic individual‐based model. Environ Entomol 42:412–423 (2013). [DOI] [PubMed] [Google Scholar]
96. Snep RPH, Opdam PFM, Baveco JM, Wallis DeVries MF, Timmermans W, Kwak RGM et al., How peri‐urban areas can strengthen animal populations within cities: a modeling approach. Biol Conserv 127:345–355 (2006). [Google Scholar]
97. Hancock TJ, Lee DH, Bergh JC, Morrison WR III and Leskey TC, Presence of the invasive brown marmorated stink bug Halyomorpha halys (Stål) (Hemiptera: Pentatomidae) on home exteriors during the autumn dispersal period: results generated by citizen scientists. Agric For Entomol 21:99–108 (2019). [Google Scholar]
98. Acebes‐Doria AL, Leskey TC and Bergh JC, Host plant effects on Halyomorpha halys (Hemiptera: Pentatomidae) nymphal development and survivorship. Environ Entomol 45:663–670 (2016). [DOI] [PubMed] [Google Scholar]
99. Fisher JJ, Rijal JP and Zalom FG, Temperature and humidity interact to influence brown marmorated stink bug (Hemiptera: Pentatomidae) survival. Environ Entomol 50:390–398 (2021). [DOI] [PubMed] [Google Scholar]
100. Huberty AF and Denno RF, Plant water stress and its consequences for herbivorous insects: a new synthesis. Ecology 85:1383–1398 (2004). [Google Scholar]
101. Mody K, Eichenberger D and Dorn S, Stress magnitude matters: different intensities of pulsed water stress produce non‐monotonic resistance responses of host plants to insect herbivores. Ecol Entomol 34:133–143 (2009). [Google Scholar]
102. Pincebourde S and Woods HA, Climate uncertainty on leaf surfaces: perspectives on the leaf microclimate from biophysical ecology. Funct Ecol 26:844–853 (2012). [Google Scholar]
103. Jepsen JU, Kapari L, Hagen SB, Schott T, Vindstad OPL, Nilssen AC et al., Rapid northwards expansion of a forest insect pest attributed to spring phenology matching with sub‐Arctic birch. Glob Chang Biol 17:2071–2083 (2011). [Google Scholar]
104. Porter JH, Parry ML and Carter TR, The potential effects of climatic change on agricultural insect pests. Agric For Meteorol 57:221–240 (1991). [Google Scholar]
105. Svobodová E, Trnka M, Dubrovský M, Semerádová D, Eitzinger J, Štěpánek P et al., Determination of areas with the most significant shift in persistence of pests in Europe under climate change. Pest Manage Sci 70:708–715 (2014). [DOI] [PubMed] [Google Scholar]
106. Gaba S, Bretagnolle F, Rigaud T and Philippot L, Managing biotic interactions for ecological intensification of agroecosystems. Front Ecol Evol 2:29 (2014). [Google Scholar]
107. Abram PK, Mills NJ and Beers EH, Classical biological control of invasive stink bugs with egg parasitoids – what does success look like? Pest Manage Sci 76:1980–1992 (2020). [DOI] [PubMed] [Google Scholar]
108. Tillman G, Toews M, Blaauw B, Sial A, Cottrell T, Talamas E et al., Parasitism and predation of sentinel eggs of the invasive brown marmorated stink bug, Halyomorpha halys (Stål) (Hemiptera: Pentatomidae), in the southeastern US. Biol Control 145:e104247 (2020). [Google Scholar]
109. Yonow T, Kriticos DJ, Ota N, Avila GA, Hoelmer KA, Chen H et al., Modelling the potential geographic distribution of two Trissolcus species for the Brown Marmorated stink bug Halyomorpha halys . Insects 12:491 (2021). [DOI] [PMC free article] [PubMed] [Google Scholar]
110. Song HW and Wang CM, Damage by Halyomorpha halys (Stal) and Erthesina fullo (Thunberg) to jujube trees and their control. Entomol Knowl 30(4):225–228 (1993). [Google Scholar]
111. Yu CL, Jin XF, Liu XQ, Zhao HY and Jin CC, Study on the damage of Halyomorpha halys and Dolycoris baccarum to the pear and its control strategy. China Fruits 2:5–7 (2002). [Google Scholar]
112. Moritz C and Agudo R, The future of species under climate change: resilience or decline? Science 341:504–508 (2013). [DOI] [PubMed] [Google Scholar]
113. Blois JL, Zarnetske PL, Fitzpatrick MC and Finnegan S, Climate change and the past, present, and future of biotic interactions. Science 341:499–504 (2013). [DOI] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

Table S1. Variance inflation factors (VIFs) were calculated to assess multicollinearity among variables in all models. VIF > 5 are generally considered problematic.

Click here for additional data file.^{(2.5MB, docx)}

Data Availability Statement

The data that support the findings of this study are available from the corresponding author upon reasonable request.

[ps7113-bib-0001] 1. Mack RN, Simberloff D, Lonsdale WM, Evans H, Clout M and Bazzaz FA, Biotic invasions: causes, epidemiology, global consequences, and control. Ecol Appl 10:689–710 (2000). [Google Scholar]

[ps7113-bib-0002] 2. Venette RC and Hutchison WD, Invasive insect species: global challenges, strategies & opportunities. Front Insect Sci 1:650520 (2021). [DOI] [PMC free article] [PubMed] [Google Scholar]

[ps7113-bib-0003] 3. Guisan A and Zimmermann NE, Predictive habitat distribution models in ecology. Ecol Model 135:147–186 (2000). [Google Scholar]

[ps7113-bib-0004] 4. Gutiérrez Illán J, Bloom EH, Wohleb CH, Wenninger EJ, Rondon SI, Jensen AS et al., Landscape structure and climate drive population dynamics of an insect vector within intensely managed agroecosystems. Ecol Appl 30:e02109 (2020). [DOI] [PubMed] [Google Scholar]

[ps7113-bib-0005] 5. Parmesan C, Ryrholm N, Stefanescu C, Hill TK, Thomas CD, Descimon H et al., Poleward shifts in geographical ranges of butterfly species associated with regional warming. Nature 399:579–583 (1999). [Google Scholar]

[ps7113-bib-0006] 6. Kriticos DJ, Kean JM, Phillips CB, Senay SD, Acosta H and Haye T, The potential global distribution of the brown marmorated stink bug, Halyomorpha halys, a critical threat to plant biosecurity. J Pest Sci 90:1033–1043 (2017). [Google Scholar]

[ps7113-bib-0007] 7. Mainali KP, Warren DL, Dhileepan K, McConnachie A, Strathie L, Hassan G et al., Projecting future expansion of invasive species: comparing and improving methodologies for species distribution modeling. Glob Chang Biol 21:4464–4480 (2015). [DOI] [PubMed] [Google Scholar]

[ps7113-bib-0008] 8. Zhu G, Illan JG, Looney C and Crowder DW, Assessing the ecological niche and invasion potential of the Asian giant hornet. Proc Natl Acad Sci 117:14730–14737 (2020). [DOI] [PMC free article] [PubMed] [Google Scholar]

[ps7113-bib-0009] 9. Gutiérrez Illán J, Thomas CD, Jones JA, Wong WK, Shirley SM and Betts MG, Precipitation and winter temperature predict long‐term range‐scale abundance changes in Western north American birds. Glob Chang Biol 20:3351–3364 (2014). [DOI] [PubMed] [Google Scholar]

[ps7113-bib-0010] 10. Parmesan C and Yohe G, A globally coherent fingerprint of climate change impacts across natural systems. Nature 421:37–42 (2003). [DOI] [PubMed] [Google Scholar]

[ps7113-bib-0011] 11. Thomas CD and Lennon JJ, Birds extend their ranges northwards. Nature 399:213–213 (1999). [Google Scholar]

[ps7113-bib-0012] 12. Wilson RJ, Gutiérrez D, Gutiérrez Illan J, Martínez D, Agudo R and Monserrat VJ, Changes to the elevational limits and extent of species ranges associated with climate change. Ecol Lett 8:1138–1146 (2005). [DOI] [PubMed] [Google Scholar]

[ps7113-bib-0013] 13. Bale JS, Masters GJ, Hodkinson ID, Awmack C, Bezemer TM, Brown VK et al., Herbivory in global climate change research: direct effects of rising temperature on insect herbivores. Glob Chang Biol 8:1–16 (2002). [Google Scholar]

[ps7113-bib-0014] 14. Stoeckli S, Felber R and Haye T, Current distribution and voltinism of the brown marmorated stink bug, Halyomorpha halys, in Switzerland and its response to climate change using a high‐resolution CLIMEX model. Int J Biometeorol 64:2019–2032 (2020). [DOI] [PMC free article] [PubMed] [Google Scholar]

[ps7113-bib-0015] 15. Van Der Wal J, Shoo LP, Johnson CN and Williams SE, Abundance and the environmental niche: environmental suitability estimated from niche models predicts the upper limit of local abundance. Am Nat 174:282–291 (2009). [DOI] [PubMed] [Google Scholar]

[ps7113-bib-0016] 16. Leskey TC and Nielsen AL, Impact of the invasive brown marmorated stink bug in North America and Europe: history, biology, ecology, and management. Annu Rev Entomol 63:599–618 (2018). [DOI] [PubMed] [Google Scholar]

[ps7113-bib-0017] 17. Rice KB, Bergh CJ, Bergmann EJ, Biddinger DJ, Dieckhoff C, Dively G et al., Biology, ecology, and management of brown marmorated stink bug (Hemiptera: Pentatomidae). J Integr Pest Manage 5:1–13 (2014). [Google Scholar]

[ps7113-bib-0018] 18. Lee D‐H and Leskey TC, Flight behavior of foraging and overwintering brown marmorated stink bug, Halyomorpha halys (Hemiptera: Pentatomidae). Bull Entomol Res 105:566–573 (2015). [DOI] [PMC free article] [PubMed] [Google Scholar]

[ps7113-bib-0019] 19. Wiman NG, Walton VM, Shearer PW, Rondon SI and Lee JC, Factors affecting flight capacity of brown marmorated stink bug, Halyomorpha halys (Hemiptera: Pentatomidae). J Pest Sci 88:37–47 (2014). [Google Scholar]

[ps7113-bib-0020] 20. Hahn NG, Kaufman AJ, Rodriguez‐Saona C, Nielsen AL, LaForest J and Hamilton GC, Exploring the spread of brown marmorated stink bug in New Jersey through the use of crowdsourced reports. Am Entomol 62:36–45 (2016). [Google Scholar]

[ps7113-bib-0021] 21. Inkley DB, Characteristics of home invasion by the brown marmorated stink bug (Hemiptera: Pentatomidae). J Entomol Sci 47:125–130 (2012). [Google Scholar]

[ps7113-bib-0022] 22. Watanabe M, Arakawa R, Shinagawa Y and Okazawa T, Overwintering flight of brown‐marmorated stink bug, Halyomorpha mista, to the buildings. Med Entomol Zool 45:25–31 (1994). [Google Scholar]

[ps7113-bib-0023] 23. Northeastern IPM Center . 2014. Stink bug detectives. http://www.stopbmsb.org/stink-bug-bulletin/stink-bug-detectives/ [May 2020].

[ps7113-bib-0024] 24. Gariepy TD, Bruin A, Haye T, Milonas P and Vétek G, Occurrence and genetic diversity of new populations of Halyomorpha halys in Europe. J Pest Sci 88:451–460 (2015). [Google Scholar]

[ps7113-bib-0025] 25. Haye T and Weber DC, Special issue on the brown marmorated stink bug, Halyomorpha halys: an emerging pest of global concern. J Pest Sci 90:987–988 (2017). [Google Scholar]

[ps7113-bib-0026] 26. Faundez EI and Rider DA, The brown marmorated stink bug Halyomorpha halys (Stal, 1855) (Heteroptera: Pentatomidae) in Chile. Arq Entomol 17:305–307 (2017). [Google Scholar]

[ps7113-bib-0027] 27. Gapon DA, First records of the brown marmorated stink bug Halyomorpha halys (Stål, 1855) (Heteroptera, Pentatomidae) in Russia, Abkhazia, and Georgia. Entomol Rev 96:1086–1088 (2016). [Google Scholar]

[ps7113-bib-0028] 28. Acebes‐Doria AL, Agnello AM, Blaauw BR, Buntin GD, Alston DG, Beers EH et al., Season‐long monitoring of the brown marmorated stink bug throughout the United States using commercially available traps and lures. J Econ Entomol 112:159–171 (2020). [DOI] [PubMed] [Google Scholar]

[ps7113-bib-0029] 29. Cira TM, Venette RC, Aigner J, Kuhar T, Mullins DE, Gabbert SE et al., Cold tolerance of Halyomorpha halys (Hemiptera: Pentatomidae) across geographic and temporal scales. Environ Entomol 45:484–491 (2016). [DOI] [PubMed] [Google Scholar]

[ps7113-bib-0030] 30. Govindan BN and Hutchison WD, Influence of temperature on age‐stage, two‐sex life tables for a Minnesota‐acclimated population of the brown marmorated stink bug (Halyomorpha halys). Insects 11:108 (2020). [DOI] [PMC free article] [PubMed] [Google Scholar]

[ps7113-bib-0031] 31. Weber DC, Leskey TC, Walsh GC and Khrimian A, Synergy of aggregation pheromone with methyl (E,E,Z)‐2,4,6‐decatrienoate in attraction of Halyomorpha halys (Hemiptera: Pentatomidae). J Econ Entomol 107:1061–1068 (2014). [DOI] [PubMed] [Google Scholar]

[ps7113-bib-0032] 32. Khrimian A, Zhang A, Weber DC, Ho HY, Aldrich JR, Vermillion KE et al., Discovery of the aggregation pheromone of the brown marmorated stink bug (Halyomorpha halys) through the creation of stereoisomeric libraries of 1‐bisabolen‐3‐ols. J Nat Prod 77:1708–1717 (2014). [DOI] [PubMed] [Google Scholar]

[ps7113-bib-0033] 33. Leskey TC, Agnello A, Bergh JC, Dively GP, Hamilton GC, Jentsch P et al., Attraction of the invasive Halyomorpha halys (Hemiptera: Pentatomidae) to traps baited with semiochemical stimuli across the United States. J Econ Entomol 44:746–756 (2015a). [DOI] [PubMed] [Google Scholar]

[ps7113-bib-0034] 34. Leskey TC, Khrimian A, Weber DC, Aldrich JC, Short BD, Lee DH et al., Behavioral responses of the invasive Halyomorpha halys (Stål) to traps baited with stereoisomeric mixtures of 10, 11‐epoxy‐1‐bisabolen‐3‐ol. J Chem Ecol 41:418–429 (2015b). [DOI] [PMC free article] [PubMed] [Google Scholar]

[ps7113-bib-0035] 35. Weber DC, Morrison WR, Khrimian A, Rice KB, Leskey TC, Rodriguez‐Saona C et al., Chemical ecology of Halyomorpha halys: discoveries and applications. J Pest Sci 90:989–1008 (2017). [Google Scholar]

[ps7113-bib-0036] 36. Daly C, Gibson WP, Taylor GH, Johnson GL and Pasteris P, A knowledge‐based approach to the statistical mapping of climate. Clim Res 22:99–113 (2002). [Google Scholar]

[ps7113-bib-0037] 37. Daly C, Taylor GH, Gibson WP, Parzybok TW, Johnson GL and Pasteris PA, High‐quality spatial climate data sets for the United States and beyond. Trans ASAE 43:1957–1962 (2000). [Google Scholar]

[ps7113-bib-0038] 38. Daly C, Halbleib M, Smith JI, Gibson WP, Doggett MK, Taylor GH et al., Physiographically‐sensitive mapping of temperature and precipitation across the conterminous United States. Int J Climatol 28:2031–2064 (2008). [Google Scholar]

[ps7113-bib-0039] 39. Mu, Q. , Zhao M., and Steven W.. 2013. Algorithm Theoretical Basis Document: MODIS Global Terrestrial Evapotranspiration (ET) Product (NASA MOD16A2/A3) Collection 5. NASA Headquarters 2013. http://files.ntsg.umt.edu/data/NTSG_Products/MOD16/.

[ps7113-bib-0040] 40. NASA Shuttle Radar Topography Mission (SRTM) (2013). Shuttle Radar Topography Mission (SRTM) Global. Distributed by OpenTopography. 10.5069/G9445JDF. [7 January 2021]. [DOI]

[ps7113-bib-0041] 41. Han W, Yang Z, Di L and Mueller R, CropScape: a web service based application for exploring and disseminating US continuous geospatial cropland data products for decision support. Comput Electron Agric 84:111–123 (2012). [Google Scholar]

[ps7113-bib-0042] 42. NLCD (2010) National land cover database (2010). USEPA, Washington, DC. Available at: www.epa.gov/mrlc/nlcd-2001.html [03 March 2018].

[ps7113-bib-0043] 43. Merow C, Smith MJ and Silander JA Jr, A practical guide to MaxEnt for modeling species' distributions: what it does, and why inputs and settings matter. Ecography 36:1058–1069 (2013). [Google Scholar]

[ps7113-bib-0044] 44. Graham MH, Confronting multicollinearity in ecological multiple regression. Ecology 84:2809–2815 (2003). [Google Scholar]

[ps7113-bib-0045] 45. Cobos ME, Peterson AT, Osorio‐Olvera L and Jiménez‐García D, An exhaustive analysis of heuristic methods for variable selection in ecological niche modeling and species distribution modeling. Eco Inform 53:100983 (2019). [Google Scholar]

[ps7113-bib-0046] 46. González‐Suárez M and Revilla E, Variability in life‐history and ecological traits is a buffer against extinction in mammals. Ecol Lett 16:242–251 (2013). [DOI] [PubMed] [Google Scholar]

[ps7113-bib-0047] 47. O'Brien R, A caution regarding rules of thumb for variance inflation factors. Qual Quant 41:673–690 (2007). [Google Scholar]

[ps7113-bib-0048] 48. Elith J, Graham H, Anderson R, Dudík M, Ferrier S, Guisan A et al., Novel methods improve prediction of species' distributions from occurrence data. Ecography 29:129–151 (2006). [Google Scholar]

[ps7113-bib-0049] 49. Phillips SJ and Dudík M, Modeling of species distributions with Maxent: new extensions and a comprehensive evaluation. Ecography 31:161–175 (2008). [Google Scholar]

[ps7113-bib-0050] 50. Wisz MS, Hijmans RJ, Li J, Peterson AT, Graham CH, Guisan A et al., Effects of sample size on the performance of species distribution models. Divers Distrib 14:763–773 (2008). [Google Scholar]

[ps7113-bib-0051] 51. Greenwell, B. , Boehmke B., and Cunningham J.. 2019. Developers (https://github.com/gbm‐developers). Generalized Boosted Regression Models. Package ‘gbm’ R package version, 2(5).

[ps7113-bib-0052] 52. R Development Core Team , R: A Language and Environment for Statistical Computing, Vol. 2018. R Foundation for Statistical Computing, Vienna, Austria: (2018). [Google Scholar]

[ps7113-bib-0053] 53. Hijmans, RJ , and Elith J. Species Distribution Modeling with R. R CRAN Project; (2013). [Google Scholar]

[ps7113-bib-0054] 54. Death G, Boosted trees for ecological modeling and prediction. Ecology 88:243–225 (2007). [DOI] [PubMed] [Google Scholar]

[ps7113-bib-0055] 55. Elith J, Leathwick JR and Hastie T, A working guide to boosted regression trees. J Anim Ecol 77:802–813 (2008). [DOI] [PubMed] [Google Scholar]

[ps7113-bib-0056] 56. Friedman JH. Stochastic Gradient Boosting. Computational Statistics and Data Analysis 38(4):367–378 (2002). [Google Scholar]

[ps7113-bib-0057] 57. Hastie T, Tibshirani R, and Friedman J. Random forests. In The elements of statistical learning (pp. 587–604). Springer, New York, NY; (2009). [Google Scholar]

[ps7113-bib-0058] 58. Kéry M, Towards the modelling of true species distributions. J Biogeogr 38:617–618 (2011). [Google Scholar]

[ps7113-bib-0059] 59. MacKenzie DI, Nichols JD, Hines JE, Knutson MG and Franklin AB, Estimating site occupancy, colonization, and local extinction when a species is detected imperfectly. Ecology 84:2200–2207 (2003). [Google Scholar]

[ps7113-bib-0060] 60. Welsh AH, Lindenmayer DB and Donnelly CF, Fitting and interpreting occupancy models. PLoS One 8:e52015 (2013). [DOI] [PMC free article] [PubMed] [Google Scholar]

[ps7113-bib-0061] 61. Guisan A and Thuiller W, Predicting species distribution: offering more than simple habitat models. Ecol Lett 10:435–435, 1009 (2005). [DOI] [PubMed] [Google Scholar]

[ps7113-bib-0062] 62. Peterson AT, Soberón J, Pearson RG, Anderson RP, Martínez‐Meyer E, Nakamura M et al., Ecological Niches and Geographic Distributions: A Modeling Perspective. Princeton University Press, Princeton, New Jersey, USA: (2011). [Google Scholar]

[ps7113-bib-0063] 63. Manel S, Williams HC and Ormerod SJ, Evaluating presence–absence models in ecology: the need to account for prevalence. J Appl Ecol 38:921–931 (2001). [Google Scholar]

[ps7113-bib-0064] 64. McPherson JM, Jetz W and Rogers DJ, The effects of species' range sizes on the accuracy of distribution models: ecological phenomenon or statistical artefact? J Appl Ecol 41:811–823 (2004). [Google Scholar]

[ps7113-bib-0065] 65. Elith J, Phillips TH, Dudík M, Chee YE and Yates CJ, A statistical explanation of MaxEnt for ecologists. Divers Distrib 17:43–57 (2011). [Google Scholar]

[ps7113-bib-0066] 66. Fielding AH and Bell JF, A review of methods for the assessment of prediction errors in conservation presence/absence models. Environ Conserv 1:38–49 (1997). [Google Scholar]

[ps7113-bib-0067] 67. Phillips SJ, Anderson RP and Schapire RE, Maximum entropy modeling of species geographic distributions. Ecol Model 190:231–259 (2006). [Google Scholar]

[ps7113-bib-0068] 68. Lobo JM, Jiménez‐Valverde A and Real R, AUC: a misleading measure of the performance of predictive distribution models. Glob Ecol Biogeogr 17:145–151 (2008). [Google Scholar]

[ps7113-bib-0069] 69. Bahn V and McGill BJ, Testing the predictive performance of distribution models. Oikos 122:321–331 (2013). [Google Scholar]

[ps7113-bib-0070] 70. Knutti R and Sedláček J, Robustness and uncertainties in the new CMIP5 climate model projections. Nat Clim Change 3:369–373 (2013). [Google Scholar]

[ps7113-bib-0071] 71. Wang T, Hamann A, Spittlehouse DL and Carroll C, Locally downscaled and spatially customizable climate data for historical and future periods for North America. PLoS One 11:e0156720 (2016). [DOI] [PMC free article] [PubMed] [Google Scholar]

[ps7113-bib-0072] 72. Grundel R, Frohnapple KJ, Zaya DN, Glowacki GA, Weiskerger CJ, Patterson TA et al., Geographic coincidence of richness, mass, conservation value, and response to climate of US land birds. Ecol Appl 24:791–811 (2014). [DOI] [PubMed] [Google Scholar]

[ps7113-bib-0073] 73. Sohl TL, Sayler KL, Bouchard MA, Reker RR, Friesz AM, Bennett SL et al., Spatially explicit modeling of 1992‐2100 land cover and forest stand age for the conterminous United States. Ecol Appl 24:1015–1036 (2014). [DOI] [PubMed] [Google Scholar]

[ps7113-bib-0074] 74. Tabor K and Williams JW, Globally downscaled climate projections for assessing the conservation impacts of climate change. Ecol Appl 20:554–565 (2010). [DOI] [PubMed] [Google Scholar]

[ps7113-bib-0075] 75. Ziter C, Robinson EA and Newman JA, Climate change and voltinism in Californian insect pest species: sensitivity to location, scenario and climate model choice. Glob Chang Biol 18:2771–2780 (2012). [DOI] [PubMed] [Google Scholar]

[ps7113-bib-0076] 76. Intergovernmental panel on climate change (IPCC), in Climate Change 2014: IPCC. Climate Change 2014: Synthesis Report. Contribution of Working Groups I, II and III to the Fifth Assessment Report of the Intergovenmental Panel on Climate Change. Cambridge University Press, Cambridge: (2014). [Google Scholar]

[ps7113-bib-0077] 77. Nakicenovic N, Alcamo J, Davis G, de Vries B, Fenhann J, Gaffin S et al., Special Report on Emissions Scenarios: A Special Report of Working Group III of the Intergovernmental Panel on Climate Change. Cambridge University Press, Cambridge, UK: (2000). [Google Scholar]

[ps7113-bib-0078] 78. La Sorte FA and Jetz W, Tracking of climatic niche boundaries under recent climate change. J Anim Ecol 81:914–925 (2012). [DOI] [PubMed] [Google Scholar]

[ps7113-bib-0079] 79. Renwick AR, Massimino D, Newson SE, Chamberlain DE, Pearce‐Higgins JW and Johnston A, Modelling changes in species' abundance in response to projected climate change. Divers Distrib 18:121–132 (2012). [Google Scholar]

[ps7113-bib-0080] 80. Zhu G, Bu W, Gao Y and Liu G, Potential geographic distribution of brown marmorated stink bug invasion (Halyomorpha halys). PLoS One 7:31246 (2012). [DOI] [PMC free article] [PubMed] [Google Scholar]

[ps7113-bib-0081] 81. Aguirre‐Gutiérrez J, Carvalheiro LG, Polce C, van Loon EE, Raes N, Reemer M et al., Fit‐for‐purpose: species distribution model performance depends on evaluation criteria–Dutch hoverflies as a case study. PLoS One 8:63708 (2013). [DOI] [PMC free article] [PubMed] [Google Scholar]

[ps7113-bib-0082] 82. Roura‐Pascual N, Brotons L, Peterson AT and Thuiller W, Consensual predictions of potential distributional areas for invasive species: a case study of argentine ants in the Iberian Peninsula. Biol Invasions 11:1017–1031 (2009). [Google Scholar]

[ps7113-bib-0083] 83. Nielsen AL, Chen S and Fleischer SJ, Coupling developmental physiology, photoperiod, and temperature to model phenology and dynamics of an invasive heteropteran, Halyomorpha halys . Front Physiol 7:165 (2016). [DOI] [PMC free article] [PubMed] [Google Scholar]

[ps7113-bib-0084] 84. Lee D‐H, Short BD, Joseph SV, Bergh JC and Leskey TC, Review of the biology, ecology, and management of Halyomorpha halys (Hemiptera: Pentatomidae) in China, Japan, and the Republic of Korea. Environ Entomol 42:627–641 (2013). [DOI] [PubMed] [Google Scholar]

[ps7113-bib-0085] 85. Khadka A, Hodges AC, Leppla NC and Tillman PG, Halyomorpha halys (Stål)(Hemiptera: Pentatomidae) nymph survival and adult feeding preferences for crop plants in Florida. Fla Entomol 104:136–139 (2021). [Google Scholar]

[ps7113-bib-0086] 86. Scaccini D, Vanishvili L, Tirello P, Walton VM, Duso C and Pozzebon A, Lethal and sub‐lethal effects of low‐temperature exposures on Halyomorpha halys (Hemiptera: Pentatomidae) adults before and after overwintering. Sci Rep 10:1–9 (2020). [DOI] [PMC free article] [PubMed] [Google Scholar]

[ps7113-bib-0087] 87. McLaughlin JF, Hellmann JJ, Boggs CL and Ehrlich PR, The route to extinction: population dynamics of a threatened butterfly. Oecologia 132:538–548 (2002). [DOI] [PubMed] [Google Scholar]

[ps7113-bib-0088] 88. Moore MV and Lee RE Jr, Surviving the big chill: overwintering strategies of aquatic and terrestrial insects. Am Entomol 37:111–118 (1991). [Google Scholar]

[ps7113-bib-0089] 89. Overgaard J and MacMillan HA, The integrative physiology of insect chill tolerance. Annu Rev Physiol 79:187–208 (2017). [DOI] [PubMed] [Google Scholar]

[ps7113-bib-0090] 90. Danks HV, Dehydration in dormant insects. J Insect Physiol 46:837–852 (2000). [DOI] [PubMed] [Google Scholar]

[ps7113-bib-0091] 91. Ciancio JJ, Turnbull KF, Gariepy TD and Sinclair BJ, Cold tolerance, water balance, energetics, gas exchange, and diapause in overwintering brown marmorated stink bugs. J Insect Physiol 128:104171 (2021). [DOI] [PubMed] [Google Scholar]

[ps7113-bib-0092] 92. Cullum JP, Nixon LJ, Morrison WR, Raupp MJ, Shrewsbury PM, Venugopal PD et al., Influence of landscape factors and abiotic conditions on dispersal behavior and overwintering site selection by Halyomorpha halys (Hemiptera: Pentatomidae). J Econ Entomol 113:2016–2021 (2020). [DOI] [PubMed] [Google Scholar]

[ps7113-bib-0093] 93. Susset EC, Hemptinne JL and Magro A, Overwintering sites might not be safe haven for Hippodamia undecimnotata (Schneider) (Coleoptera: Coccinellidae). Coleopterists Bull 71:556–564 (2017). [Google Scholar]

[ps7113-bib-0094] 94. Prasad AM, Iverson LR, Peters MP, Bossenbroek JM, Matthews SN, Sydnor TD et al., Modeling the invasive emerald ash borer risk of spread using a spatially explicit cellular model. Landsc Ecol 25:353–369 (2010). [Google Scholar]

[ps7113-bib-0095] 95. Tonini F, Hochmair HH, Scheffrahn RH and Deangelis DL, Simulating the spread of an invasive termite in an urban environment using a stochastic individual‐based model. Environ Entomol 42:412–423 (2013). [DOI] [PubMed] [Google Scholar]

[ps7113-bib-0096] 96. Snep RPH, Opdam PFM, Baveco JM, Wallis DeVries MF, Timmermans W, Kwak RGM et al., How peri‐urban areas can strengthen animal populations within cities: a modeling approach. Biol Conserv 127:345–355 (2006). [Google Scholar]

[ps7113-bib-0097] 97. Hancock TJ, Lee DH, Bergh JC, Morrison WR III and Leskey TC, Presence of the invasive brown marmorated stink bug Halyomorpha halys (Stål) (Hemiptera: Pentatomidae) on home exteriors during the autumn dispersal period: results generated by citizen scientists. Agric For Entomol 21:99–108 (2019). [Google Scholar]

[ps7113-bib-0098] 98. Acebes‐Doria AL, Leskey TC and Bergh JC, Host plant effects on Halyomorpha halys (Hemiptera: Pentatomidae) nymphal development and survivorship. Environ Entomol 45:663–670 (2016). [DOI] [PubMed] [Google Scholar]

[ps7113-bib-0099] 99. Fisher JJ, Rijal JP and Zalom FG, Temperature and humidity interact to influence brown marmorated stink bug (Hemiptera: Pentatomidae) survival. Environ Entomol 50:390–398 (2021). [DOI] [PubMed] [Google Scholar]

[ps7113-bib-0100] 100. Huberty AF and Denno RF, Plant water stress and its consequences for herbivorous insects: a new synthesis. Ecology 85:1383–1398 (2004). [Google Scholar]

[ps7113-bib-0101] 101. Mody K, Eichenberger D and Dorn S, Stress magnitude matters: different intensities of pulsed water stress produce non‐monotonic resistance responses of host plants to insect herbivores. Ecol Entomol 34:133–143 (2009). [Google Scholar]

[ps7113-bib-0102] 102. Pincebourde S and Woods HA, Climate uncertainty on leaf surfaces: perspectives on the leaf microclimate from biophysical ecology. Funct Ecol 26:844–853 (2012). [Google Scholar]

[ps7113-bib-0103] 103. Jepsen JU, Kapari L, Hagen SB, Schott T, Vindstad OPL, Nilssen AC et al., Rapid northwards expansion of a forest insect pest attributed to spring phenology matching with sub‐Arctic birch. Glob Chang Biol 17:2071–2083 (2011). [Google Scholar]

[ps7113-bib-0104] 104. Porter JH, Parry ML and Carter TR, The potential effects of climatic change on agricultural insect pests. Agric For Meteorol 57:221–240 (1991). [Google Scholar]

[ps7113-bib-0105] 105. Svobodová E, Trnka M, Dubrovský M, Semerádová D, Eitzinger J, Štěpánek P et al., Determination of areas with the most significant shift in persistence of pests in Europe under climate change. Pest Manage Sci 70:708–715 (2014). [DOI] [PubMed] [Google Scholar]

[ps7113-bib-0106] 106. Gaba S, Bretagnolle F, Rigaud T and Philippot L, Managing biotic interactions for ecological intensification of agroecosystems. Front Ecol Evol 2:29 (2014). [Google Scholar]

[ps7113-bib-0107] 107. Abram PK, Mills NJ and Beers EH, Classical biological control of invasive stink bugs with egg parasitoids – what does success look like? Pest Manage Sci 76:1980–1992 (2020). [DOI] [PubMed] [Google Scholar]

[ps7113-bib-0108] 108. Tillman G, Toews M, Blaauw B, Sial A, Cottrell T, Talamas E et al., Parasitism and predation of sentinel eggs of the invasive brown marmorated stink bug, Halyomorpha halys (Stål) (Hemiptera: Pentatomidae), in the southeastern US. Biol Control 145:e104247 (2020). [Google Scholar]

[ps7113-bib-0109] 109. Yonow T, Kriticos DJ, Ota N, Avila GA, Hoelmer KA, Chen H et al., Modelling the potential geographic distribution of two Trissolcus species for the Brown Marmorated stink bug Halyomorpha halys . Insects 12:491 (2021). [DOI] [PMC free article] [PubMed] [Google Scholar]

[ps7113-bib-0110] 110. Song HW and Wang CM, Damage by Halyomorpha halys (Stal) and Erthesina fullo (Thunberg) to jujube trees and their control. Entomol Knowl 30(4):225–228 (1993). [Google Scholar]

[ps7113-bib-0111] 111. Yu CL, Jin XF, Liu XQ, Zhao HY and Jin CC, Study on the damage of Halyomorpha halys and Dolycoris baccarum to the pear and its control strategy. China Fruits 2:5–7 (2002). [Google Scholar]

[ps7113-bib-0112] 112. Moritz C and Agudo R, The future of species under climate change: resilience or decline? Science 341:504–508 (2013). [DOI] [PubMed] [Google Scholar]

[ps7113-bib-0113] 113. Blois JL, Zarnetske PL, Fitzpatrick MC and Finnegan S, Climate change and the past, present, and future of biotic interactions. Science 341:499–504 (2013). [DOI] [PubMed] [Google Scholar]

PERMALINK

Evaluating invasion risk and population dynamics of the brown marmorated stink bug across the contiguous United States

Javier Gutiérrez Illán

Gengping Zhu

James F Walgenbach

Angel Acebes‐Doria

Arthur M Agnello

Diane G Alston

Heather Andrews

Elisabeth H Beers

J Christopher Bergh

Ricardo T Bessin

Brett R Blaauw

G David Buntin

Erik C Burkness

John P Cullum

Kent M Daane

Lauren E Fann

Joanna Fisher

Pierre Girod

Larry J Gut

George C Hamilton

James R Hepler

Richard Hilton

Kim A Hoelmer

William D Hutchison

Peter J Jentsch

Shimat V Joseph

George G Kennedy

Grzegorz Krawczyk

Thomas P Kuhar

Jana C Lee

Tracy C Leskey

Adrian T Marshal

Joshua M Milnes

Anne L Nielsen

Dilani K Patel

Hillary D Peterson

Dominic D Reisig

Jhalendra P Rijal

Ashfaq A Sial

Lori R Spears

Judith M Stahl

Kathy M Tatman

Sally V Taylor

Glynn Tillman

Michael D Toews

Raul T Villanueva

Celeste Welty

Nik G Wiman

Julianna K Wilson

Frank G Zalom

David W Crowder

Abstract

BACKGROUND

RESULTS

CONCLUSION

1. INTRODUCTION

2. MATERIALS AND METHODS

2.1. Study system

Figure 1.

2.2. Environmental variables

Table 1.

2.3. Model development

Figure 2.

2.4. Model performance and evaluation

2.5. Predicting BMSB distribution under future climate scenarios

3. RESULTS

3.1. Model performance and validation

3.2. Relative contribution of variables

3.3. Predicted effects of climate change on the distribution of BMSB

Figure 3.

4. DISCUSSION

Supporting information

ACKNOWLEDGEMENTS

DATA AVAILABILITY STATEMENT

REFERENCES

Associated Data

Supplementary Materials

Data Availability Statement