Spatial pattern of infestation risk and management of the invasive brown marmorated stink bug in soybeans

Final Report for GNE12-047

Project Type: Graduate Student
Funds awarded in 2012: $14,956.00
Projected End Date: 12/31/2014
Grant Recipient: University of Maryland
Region: Northeast
State: Maryland
Graduate Student:
Faculty Advisor:
Galen Dively
University of Maryland College Park
Faculty Advisor:
William Lamp
University of Maryland, College Park
Expand All

Project Information


Knowledge on spatial patterns of insect pest populations and the ecological processes influencing them can be directly applied to manage them in agricultural systems. Recent increases in stink bug populations, including that of the invasive brown marmorated stink bug (BMSB) Halyomorpha halys (Stål 1855), has caused unprecedented economic losses in the mid-Atlantic United States. To inform stink bug management programs, I quantified the spatial pattern of infestation risk of BMSB in field crops at multiple spatial scales (field edge, entire fields and regional), and identified the associated environmental and landscape factors. At corn and soybean field edges, highest density of stink bugs was limited to the first few crop rows. At some study sites, fields adjacent to woods and buildings harbored higher density of stink bugs than those adjacent to open areas. 

In entire fields of corn-soybean, H. halys was found in very low density or absent beyond 25m from the field edge. At study sites with high stink bug populations, interpolated density values showed dispersal of H. halys, particularly adults and large nymphs, from corn into soybean, coinciding with the end of dough stage in corn and beginning of seed soybean growth stage. Temperature and developed areas, and proportion of forest and crop areas were important predictors of regional patterns in H. halys abundance. Adjacent habitat influence on stink bug density was limited to field level, and differences between sites were driven primarily by temperature gradient. These results directly inform field level stink bug management strategies through planting date and orientation of fields in the landscape, and for timing and intensity of treatments, as well as area-wide management strategies. This research also identified role of climate and landscape in facilitating or impeding invasive pest populations.


Since its introduction into the United States near Allentown, PA during late 1990s, the brown marmorated stink bug (BMSB) H. halys has steadily expanded in population number and distributional range. While at present they are just a nuisance in many states, the recent explosion in H. halys populations has led to significant economic impacts as agricultural pests. H. halys caused unprecedented crop damage and economic losses particularly in the Northeast U. S. since 2009-2010 across a broad range of cropping systems (fruit, grain, vegetable, field and tree) in the Mid-Atlantic region (Leskey et al., 2012) . H. halys outbreak has also disrupted the existing IPM practices that are in place in many of these systems. The unequivocal response by farmers to H. halys outbreaks across many of these cropping systems in the Northeast (particularly mid-Atlantic) has been the regular and repeated use of broad spectrum insecticides (e.g. pyrethroids) through the growing season. Consequently, costs associated with pest control have risen along with concerns on food safety and negative impacts on beneficial insects.

The research/extension community has responded swiftly to the H. halys outbreak and over the past 3 years efforts to understand H. halys invasion and develop management strategies have increased with multi-institutional and multi-investigator collaborations and funding availability also has increased. However, these efforts primarily focus on specialty crops (e.g. $5.7 M USDA NIFA Specialty Crop Research Initiative Grant), and research and extension efforts in field crops, which are not within the purview of such grant initiatives, remains sparse. Particularly, despite field crops constituting very high proportion of overall crop area studies concentrating on developing strategies for H. halys are few. My research project filled these lacuna in information by quantifying the heterogeneity in H. halys population at multiple spatial scales and relating it to the multi-scale landscape and environmental influences.

In this context, knowledge on the spatial distribution of BMSB abundance in soybean and the factors that influence this invasive stink bug, directly applies to developing integrated pest management (IPM) strategies. As part of both site–specific / precision farming and area-wide IPM strategies, measurement of the spatial variability and pattern of BMSB abundances within a field or across a larger landscape is required. This would allow us to develop BMSB management strategies that can be incorporated into existing IPM practices. The ability of farmers to efficiently control BMSB by adjusting local farm practices requires knowledge of its spatial distribution within fields and across large landscapes. Also, prevailing environmental factors and regional land use may inhibit or support outbreaks by impeding or facilitating connectivity of BMSB among fields. However, studies investigating the spatial distribution of BMSB in field crops, throughout the season and at multiple spatial scales, are not available. Also, the influence of land uses, crop characteristics and environmental conditions on BMSB distribution patterns is not well understood. Employing a combination of field sampling, Geographic Information Systems (GIS) and spatial analysis techniques, the proposed project informed field and landscape level pest management strategies and threat potentials of BMSB in soybean crops of Mid-Atlantic region in the United States.

Project Objectives:

To assess the influence of adjacent habitats on spatio-temporal differences in densities of BMSB in soybean field edges.

To assess the influence of field level environmental and crop characteristics on the spatio-temporal differences in densities of BMSB within adjacent corn-soybean fields

To assess the influence of land use at multiple scales and environmental factors, on the landscape scale differences in relative abundances of BMSB in soybean fields

To provide a table of infestation risk factors and in-field distribution patterns associated with specific farmscapes surrounding soybean fields and identify soybean production areas with higher vulnerability to outbreak of BMSB.


Click linked name(s) to expand/collapse or show everyone's info
  • Dr. Galen Dively
  • Dr. William Lamp


Materials and methods:

1. Influence of adjacent habitat on BMSB abundances at the field edge

Field selection and sampling stragegy

The study was conducted at the USDA Beltsville Agricultural Research Center at Beltsville, MD and University of Maryland agricultural research and education facilities (study sites) at Beltsville, Clarksville and Keedysville, MD. Study sites represent increasing stink bug population pressures, with highest pressure observed in Keedysville over the past 4 years. At these sites, soybean fields with a portion of their perimeter directly adjacent to wooded areas, man-made structures (buildings, houses and barns; henceforth buildings), corn and open, non-crop, non-vegetated areas (henceforth open) were selected for sampling in 2013. In each field, four transect lines were marked at distances 0, 1.5, 3, 4.5, 6, 9, 12, and 15 meters from the edge to field interior (henceforth distance) and stink bugs were enumerated using direct visual counts based on active search (4 lines X 8 distances = 32 plots / field). Stink bug adults, nymphs, and egg masses of BMSB and other stink bugs were enumerated on all plants within a half meter radius (1.55 m2), semicircular plots in soybean(Fig.1.). Fields were repeatedly sampled between mid Aug – late Sept and coincided with the seed development stages of soybean (R4-R7) which are associated with high stink bug abundance.

Statistical Analyses

The influences of adjacent habitat and distance, on the abundance and distribution of stink bugs in corn and soybean were analyzed through Generalized Linear Mix Models (GLMMs) based on Laplace approximation, with a lognormal-Poisson error distribution and log link function (Bolker et al., 2009). GLMMs were run with the overall data pooled across species and sites and on data from each site. to identify optimum fixed effect, an initial full model analysis (individual and interactive effects of adjacent habitat with 4 levels (woods, buildings, corn and open) and distance from edge with 8 levels (0-15m) ) was performed. If a significant interaction was found then model means were compared between all levels of adjacent habitat, at each level of the distance. If there was no significant interaction, then adjacent habitat and distance were independently used as fixed effects. The significance of the fixed effects was determined through Wald c2 tests.

2. Spatio-temporal differences in densities of BMSB within adjacent corn-soybean fields

The sequence of crop planting dates in Mid-Atlantic region in which soybean is planted later than corn in adjacent plots makes soybean a suitable host (at seed filling stages R5 & R6) for BMSB to transfer from corn. Hence, quantifying the spatial population dynamics of stink bugs in corn adjacent to soybean would aid in decisions for management of BMSB in the region. In this study we sampled BMSB in 2 separate grids of adjacent corn-soybean fields at Keedysville, MD and 1 grid in Beltsville, MD during July 20 – Sep 24 at the grain fill stages in corn till physiological maturity of soybean. Three soybean fields (2 in Keedysville and 1 in Beltsville) of approximately 1.2 acres each were chosen for sampling. BMSB abundances were enumerated 10 times, at 60 equally spaced locations throughout each field. This study aimed at understanding pattern of BMSB distribution throughout the field when fields had both corn and woods as one of the adjacent sides.

3. Influence of land use at multiple scales and environmental factors on large scale BMSB abundance

BSMB abundances in soybean fields across a large landscape were recorded through field surveys during growing season of 2013. The study was conducted within soybean fields in a 3000 X 3000 mile study area in Washington, Fredrick, Montgomery, and Carroll Counties in Maryland; Jefferson County in West Virginia and all counties in Delaware. The study area chosen represents the invasion front of BMSB as they move south from PA, encompassing landscapes with varying proportions of forests, habitations, and crop area (Cropscape 2012). The study area was gridded and 83 soybean fields in MD and 40 fields in DE, each at least 5 km apart, were visited and number of BMSB and native stink bugs were sampled and recorded by taking 25 sweeps with a standard sweep net at 3 different sides of the field edge. The soybean growth stage and edge habitat were identified for all three edges. The spatial coordinates for the fields were recorded using a hand-held global positioning system. Sampling was carried out once during Sept 3 – Sept 15 2013, coinciding with the pod filling stage of soybean (R4 – R6), which is associated with higher abundances of stink bugs.

The landuse surrounding each of the sampled field was quantified using the spatilly explicit land use data layer CROPSCAPE. Also, the environmental factors (mainly temperature) and topographical factors (altitude, slope and aspect) were derived from NOAA-climate data online and USGS-National Elevation Dataset, repectively

4. Fields associated with higher risk of stink bug invasion and management

Results from all the three previous objectives were compiled to identify fields with high risk of infestation and appropriate management strategies have been suggested.


Research results and discussion:

1. Influence of adjacent habitat on BMSB abundances at the field edge

GLMM showed significant interactive influences of adjacent habitat and distance on stink bug abundance (Wald c2= 49.1, df = 21, P < 0.001). Multiple comparisons of means showed that abundance of stink bugs was significantly higher along woods than open areas, at all distances (Fig. 2). Abundance along woods were consistently higher than MMS and corn, but were overall not significantly different at multiple distances. Across all habitat types, maximum abundance was observed at the immediate field edge and declined considerably by 9m and were lowest at 15m.

GLMMs for site level data showed significant interactive influence of adjacent habitat and distance on abundance at Keedysville (Wald c2= 47.8, df = 21, P < 0.001), but not at Clarksville (Wald c2= 9.2, df = 14, P = 0.818) and Beltsville (Wald c2= 26.6, df = 21, P = 0.315). Multiple means comparisons for Keedysville data showed significantly higher abundance along woods than open at all distances, and significantly higher abundance along corn than open till 9m from field edge (Fig. 2). Abundance along corn was higher than that of buildings in general, though not statistically significant. Separate GLMMs with adjacent habitat and distance as individual explanatory variables for Beltsville data showed that abundance along woods and MMS were significantly different than corn and open. Stink bug abundance at Beltsville was significantly higher along immediate field edge than all other distances and abundance was similar between 9-15m. Separate GLMMs with adjacent habitat and distance as individual explanatory variables for Clarksville data showed that abundance along buildings and corn were significantly different than open. Stink bug abundance at Clarksville was significantly higher along immediate field edge (0m) than all other distances and abundance was similar between 9-15m.

Results from the overall GLMMs showed that adjacent habitats, particularly woods influenced high abundance of BMSB and other stink bugs. Fields adjoining woods, across all sites and distances, consistently harbored significantly higher abundance of stink bugs than open areas. Also, the observed mean abundance along woods was consistently higher than MMS and corn at various distances, albeit not being statistically significant in all cases. These results indicate the role of the wooded regions, with wide range of BMSB wild host plants (up to 170 species of cultivated and wild hosts, in supporting stink bug populations that colonize crops. During the study particularly high abundance of BMSB was observed in soybean fields bordering woods with tree of heaven (Ailanthus altissima Swingle), princess tree (Paulownia tomentosa Baill.), and black cherry (Prunus serotina Ehrhart), all which support high densities of BMSB. Based on our results of overall data, there was a clear edge effect observed in abundance of BMSB and other stink bugs. Across all adjacent habitats, BMSB and other stink bugs abundance was generally highest immediately adjacent to field edge (0m) and the first few meters from the edge (0-3m). Stink bug abundance reduced gradually till 9m, low generally between 9-15, and least at 15m. The strong edge effect exhibited by BMSB with highest abundance of bugs at 0-5m from edge, is similar to the abundance pattern commonly reported for other native stink bug species in US crops.

Site level analyses showed differences in stink bug abundance patterns between sites both in field corn and soybean edges. Particularly at Keedysville, overall higher mean stink bug abundance was observed soybean than at Beltsville. These results, in part, highlight the role of other adjacent cultivated crops as source of high stink bug abundance. For e.g. similar to our observations for stink bugs (BMSB) at Keedysville, adjacent field corn has been identified to influence high densities of BSB and SGSB in cotton and peanut fields. Our results, with high abundance in soybean adjacent to field corn only at Keedysville, indicate that the role of adjacent cultivated habitat as source of high stink bug abundance could be related to differing stink bug population pressure between sites. Also, the influence of the landscape on stink bug abundance could extend to spatial scales beyond habitats just immediately adjacent, to a larger, regional scale. Since insect population dynamics and distributions are affected by regional landscape context and species traits such as dispersal ability, the community structure, species distribution and abundance of BMSB may depend on habitat characteristics at spatial scales greater than the local agricultural field. Differences in landscape structure between our study sites at regional scale and their influence on local stink bug abundance could be a possible reason for overall higher abundance of stink bugs at Keedysville and role of adjacent crops as source of stink bugs in field corn and soybean.

2. Spatio-temporal differences in densities of BMSB within adjacent corn-soybean fields

Results showed that at Keedysville, BMSB density in corn was highest during milk-dough stage and during seed growth stages (R5-R6) in soybean. Also, stink bug densities initially built up in corn and subsequently increased in soybean along the corn-soy interface, once corn kernel started hardening and maturing (see the video file attached). Stink bug density, including that of nymphs increased in soy that was adjacent to corn, implying dispersal of adults and nymphs from corn. However, in Beltsville this was not observed. These results identify the site level differences in the role of corn as source of BMSB that invade soybean, during the pod filling stages. Only in sites with high overall population pressure does the BMSB population buildup in corn and eventually disperses into soybean (see attached video). Stink bug activity at the interface of corn and soybean reveal the potential dispersal of BMSB, particularly adults and large nymps from corn into soybean. Differences in landscape structure and environmental characteristics between our study sites at regional scale and their influence on local stink bug abundance could be a possible reason for overall higher abundance of stink bugs at Keedysville and role of corn as source of stink bugs in soybean.

3. Influence of land use at multiple scales and environmental factors on large scale BMSB abundance

Spatial analysis of abundance data and associated multi-scale land use and environmental factors showed that temperature was one of the most important predictor of stink bug abundance across the mid-Atlantic region. Particularly, BMSB abundance increased with the average minimum temperature in June, and sites with temperature well above lower threshold, supported higher populations of BMSB. Similarly, average minimum temperature in July negatively influenced abundance. The higher temperatures in July reduced the population of BMSB at soybean fields. Finally, stink bug abundances were also high in areas that were in proximity to developed regions with high housing density (less than 1km; Fig 3.). Analysis also helped identify that while adjacent habitat is important, temperature was the primary driver of patterns in the abundance across the region.

4. Factors associated with potential high risk of infestation by H. halys and implications for management

Based on the results from each chapter, diffrent factors influence densities of stink bug at different spatial scales. Soybean fields in high population density area (For e.g counties along the tri-state border of WV, VA and MD and the PA counties closeby are at higher risk of infestation when in proximity to cities. Particularly, those fields with wooded regions, buildings are at particular risk of infestation. Farmers in this region also need to be aware of the role of corn as adjacent habitat, particularly when planted as strips of corn and soybean. Soybean fields adjacent to corn are at higher risk of infestation as H. halys could build up populations in corn and subsequently large nymphs and adults disperse to adjacent soybean field.

These results have huge implications for management of stink bugs in the region. High abundances in Eastern shores of DELMARVA (high crop acreage) are unlikely due to higher summer temperatures. Scouting and control efforts could be targeted towards crops in proximity (< 2 km) to urban areas and if needed, insecticide application can be targeted at field edges bordering woods & buildings. Better awareness and education to home owners in high abundance areas could help in keeping H. halys out during late Fall to reduce overwintering populations. Finally, temperatures in spring do not seem conducive to support high abundance in North Central US region (highest soy acreage).

My study results support the possibility that a reduced, site-specific treatment just along the perimeter will be sufficient for efficient H. halys control without jeopardizing yield and quality. Results from the edge and grid pattern studies clearly identify the edge-mediated distribution of stink bugs including H. halys and density beyond 50 feet is low for causing seed injury and losses. Hence, perimeter only treatment of insecticides will largely control H. halys and prevent reinvasion. Experimental trials in VA during 2011-2013 growing seasons have established that perimeter treatments (45 feet from edge) control 90-95 % of H. halys and out of 24 fields trials only one field had re-colonization (Ames Herbert, personal communication). So, applying 90 foot swaths from the field edge is probably not required in most infested field situations for control of H. halys in soybean. It may be possible to effectively treat less than the current spray swath width (90 in total; 45 feet on both sides) and avoid the yield loss due to sprayer wheel tracks. One swath of 45 feet treatment from the field edge will ensure adequate control of stink bugs. If field conditions prevent movement of the sprayer at the field edge, the first few rows (which do not typically yield as much as rows further infield) could be trampled and 45 foot from the edge could be treated. Also H. halys abundances are the highest along perimeters adjacent to woods, corn and buildings and lowest in perimeters adjacent to open areas. Hence, insecticide applications may be restricted to field edges along woods, corn and buildings and not applied in field edges bordering open areas. 

Research conclusions:

Based on interactions with farmers, crop advisors and other stakeholders, entire field treament of soybean fields have been largely reduced to just the perimeter treatment. This is the most signficant outcome of my project, with results depicting edge concentrated distribution of stink bugs as evidence that entire field treatments are not necessary and farmers are successfully employing this recommendation. The project findings, hence, have resulted in reduced costs in terms of both monetary and effort requirements for farmers and has also reduced the input of insecticides into the field.

Participation Summary

Education & Outreach Activities and Participation Summary

Participation Summary:

Education/outreach description:

Currently three separate manuscripts describing these results are under preparation for publishing in scientific journals.
Addendum: publication based on this research is currently out and available here

Results from this study have been presented at both professional symposia and extension meetings.

P. Dilip Venugopal, Galen P. Dively, Ames Herbert and William Lamp. Environmental and spatial factors influencing patterns in stink bug communities in soybean. Eastern Branch ESA. March 2013.

Hooks, C.R.R., P. Dilip Venugopal and G. Dively. Fifteenth Annual Maryland Commodity Classic.  Update on BMSB activity and control in grain crops and soybeans. Centreville MD, July 25, 2013.

Upcoming presentations

P. Dilip Venugopal, Galen Dively and William Lamp. Spatiotemporal dynamics and movement of Halyomorpha halys (Stål, 1855) (Hemiptera: Pentatomidae) in and between adjacent corn and soybean fields. BMSB working group meeting. Jun 16-18 Georgetown, DE.

Any opinions, findings, conclusions, or recommendations expressed in this publication are those of the author(s) and do not necessarily reflect the view of the U.S. Department of Agriculture or SARE.