Landscape effects on spatial distribution and movement of brown marmorated stink bug in peach orchards

Final Report for GNE12-038

Project Type: Graduate Student
Funds awarded in 2012: $14,179.00
Projected End Date: 12/31/2014
Grant Recipient: Rutgers University
Region: Northeast
State: New Jersey
Graduate Student:
Faculty Advisor:
Dr. Cesar Rodriguez-Saona
Rutgers University
Faculty Advisor:
George Hamilton
Rutgers University
Expand All

Project Information


The spatial patterns of brown marmorated stink bug populations (BMSB), Halyomorpha halys (Stål), in peach orchards were investigated during the summers of 2013 and 2014. Orchards at two commercial farms and two research stations were monitored weekly for H. halys from May through September. These locations had been sampled in previous years and were confirmed to have H. halys through visual sampling. The visual sampling was mainly in peach orchards but included some adjacent habitat including lilacs and small apple trees. Assessments for damage were conducted as peaches ripened. These consisted of visual assessments for stings and cat facing damage. Close to fruit ripening, peaches were peeled to assess and compare damage on the inside and outside of each peach. ArcMap was used to visually display H. halys populations and damage levels in each orchard. Data was analyzed with a Ripley’s K test to determine clustering and conditional autoregressive model to look at the influence of landscape features on the number of H. halys in the orchard. We found that for certain dates, variety was linked to the clustering of stink bugs in the field. Data varied depending on date and location, so further analysis is being conducted to examine this.


The brown marmorated stink bug, Halyomorpha halys Stål, an insect native to China, Japan, and Korea, that is an invasive pest of agricultural crops in the mid-Atlantic United States (Hoebeke and Carter 2003, Gonzales 2012). It was introduced around 1996 into Allentown, PA and since then has caused significant crop losses in multiple states and has expanded its range to 42 states (Leskey and Hamilton 2014). On peach farms, exact numbers of brown marmorated stink bug damage have not been assessed, but growers in PA and MD have estimated upwards of 50%-60% crop loss in recent years. As a highly polyphagous pest, it is able to feed on a variety of non-agricultural plants in addition to numerous agricultural crops. H. halys can capitalize on this by using the diverse landscape of the mid-Atlantic to move from crop to crop and forested to crops.

Because of high levels of stink bugs and losses of fruit, growers have increased the frequency of insecticide applications to their orchards. Knowing where and when the stink bugs come from could help focus the control of the stink bugs. This approach would likely be cheaper and be friendlier towards the environment by targeting only parts of orchards that have higher stink bug populations at the right time of the season when fruit is most vulnerable.

Brown marmorated stink bug are known to travel into tree fruit orchards from outside of the orchard throughout the growing season to lay eggs and feed. They typically spend the winter time in environments such as between the bark of dead trees or in human-made structures (Lee et al. 2014).

GIS (Geographical information systems) is a useful tool that can be used to graphically visualize data, allowing us to approach a question from a different angle. By looking at maps of H. halys populations through time, we can get a sense of how these insects move through a field. Utilizing mapping technology will also help us look at how landscape features affect population clustering and densities.

The goal of this research was to monitor the movement of H. halys in peach orchards, determine if populations were clustered or non-clustered, and determine the landscape factors that may be contributing to this.

Project Objectives:

Objective 1. Determine if BMSB are more prevalent in peach orchards in North Jersey or South Jersey, when they increase in population, and what landscape factors contribute to this.

Sampling for this objective was conducted in 2012. This was difficult to assess because numbers of H. halys were very low across all farms. However, we were able to still see when high population levels occurred. Due to the low populations, we focused more on fine-scale surveys of peach orchards in areas with established high populations of H. halys.

Objective 2. Determine where BMSB populations are located within peach orchards, and what landscape factors contribute to this.

This objective was conducted throughout the three years at multiple orchards and research farms. Year by year, there were fluctuations in infestation levels of H. halys depending on the orchard sampled. In addition, certain rows of trees were removed due to age in the winter, so the landscape changed. However, sampling was still conducted, and sufficient numbers of H. halys continued to be found.

Objective 3. Determine what landscape factors contribute to high levels of fruit damage by BMSB.

This objective was conducted in 2013 and 2014 using fruit inspections and peeling of fruit.

Objective 4. Determine what locations/habitats near a peach orchard might contribute to BMSB immigration into orchards.

This objective was conducted through the sampling from 2012-2014. Originally, the plan was to mark H. halys in situ and track their movement. However, this was difficult, as the numbers of adults were low. Most adults that were collected in the field were used to supplement our colony, which required frequent input from field-collected specimens.


Click linked name(s) to expand/collapse or show everyone's info
  • Dr. George Hamilton
  • Dr. Cesar Rodriguez-Saona


Materials and methods:

Objective 1.

In 2012, thirteen peach orchards in farms in northern New Jersey and 10 peach orchards in farms in southern New Jersey were selected for sampling. These were surveyed weekly using visual timed counts of a minute and a half in addition to beat sampling with 5 beats from a nerf bat and a beat sheet. 20 randomly selected trees were sampled from each orchard each week, and the number of eggs, nymphs and adults was counted. This was done from May through August. Due to low numbers of H. halys in almost all of the peach orchards, a different approach was taken in which high intensity sampling was undertaken at a smaller number of sites.

In 2013 and 2014, specific orchards were chosen for fine-scale sampling. Focus was at two orchards at commercial farms and two orchards in agricultural research stations. These orchards had between 100-400 trees that were sampled on a weekly basis using visual timed counts.

Objective 2.

Four orchards were selected for sampling. These was one orchard each at the two agricultural research stations (Cream Ridge and RAREC) and at two commercial farms (Farm 1 and Farm 2). Landscapes immediately around each of the orchards differed in their context. Two of the orchards had one forested border, one of the orchards was bordered by forest, and one was in the middle of an agricultural field. The landscape in the area around each farm was different as well. Some had more agriculture or forested areas around them. At each farm, a large number of trees (>100) using visual sampling and beat sampling. The number of eggs, nymphs and adults was counted at each orchard. The number of trees sampled at farm 1, farm 2, Cream Ridge, and RAREC was ~350, ~500, ~250, and ~400. The numbers of trees fluctuated each year because of tree death and removal.

Objective 3.

This was conducted at the sampled orchard at the Cream Ridge agricultural research station. In 2013, 10 peaches on every tree were sampled for signs of H. halys damage. This was assessed by peeling each fruit and counting the number of stings on the outside and the inside of the fruit. In 2014, fruit was visually assessed for catfacing once every two weeks at each site that was sampled in objective 2. This was done by counting all of the fruit on each tree that had severe catfacing and those that did not have any obvious blemishes. Only those with catfacing and sap from puncture wounds were counted as damaged. Fruit was peeled and assessed for damage at the Cream Ridge agricultural research station again, but the number of fruit peeled was increased to 20 per tree.

Objective 4.

Although the original plan of marking H. halys in situ was not completed, we were able to gather enough information from the sampling in objective 2 to analyze where the insects might be moving into the field from.

Research results and discussion:

In 2012, the number of H. halys found in the northern and southern farms was low. However, it was still enough to see that the majority of adults were found in the months of July and August. All of the stink bugs were found in northern New Jersey farms. All but 4 of the stink bugs found were at one orchard, so it is difficult to draw any conclusions of the prevalence of H. halys. Maps I have created of black light trap captures in New Jersey indicate populations throughout the state each year, although there seems to have been a decrease in populations since 2011. Higher populations seemed to be clustered around agricultural sites around the Philadelphia area as well as west-central New Jersey. This may have been an artifact of the fact that the black light traps were in agricultural areas nearby. This has been supported by the number of crowd sourced reports of H. halys home invasion in those areas as well. This low number of sampled stink bugs helped us make the decision to proceed with a fine-scale sampling approach, sampling most or all of the trees in a single orchard. We had realized that part of the problem with our low counts may be the low sampling effort.

Midway through the 2012 growing season, we began sampling in an orchard that was confirmed to have H. halys. The number of stink bugs decreased rapidly after one week of insecticide application. However, reinfestation of the orchard took place along with the harvesting schedule. As ripening peaches were removed from trees, H. halys began invading and feeding on the hanging ripe fruit (Figure 2).

In 2013 and 2014, we switched to a focus on fine-scale sampling and chose four farms in which to concentrate our efforts. The two research farms and two commercial farms exhibited populations of H. halys, as displayed in this map, an example of populations at the Cream Ridge and RAREC research farms in 2013 (Figure 3). The two commercial farms had very low populations of H. halys throughout the summers of 2013 and 2014, possibly due to heavy insecticide input in previous years. At one of the commercial farms in 2014, after sampling ~550 trees on a weekly basis from May through August, only 7 egg masses and 6 adults were found. Even though there were low populations in orchards, communication with the growers in 2014 indicated infrequent insecticide application, hinting that populations of H. halys were lower in the sampled areas than in previous years.

In order to determine whether or not populations were clustered in  each field, I used a Ripley’s K test. This test shows whether the population exhibits clustering that deviates significantly from a theoretically randomly dispersed population. This was done for each individual sampling date, showing that populations on certain dates exhibited clustering while others showed random dispersion. For example, this is the Ripley’s K test for clustering of H. halys at the Cream Ridge location on August 27, 2013.(Figure 4).

To determine the effect of landscape on this clustering, I utilized a conditional autoregressive model, which can look at any factors entered into a model, such as the distance from overwintering sites or the varieties of peaches in the orchard. I decided to start by looking at these factors. I am focusing on the fine-scale distribution of H. halys in the orchards as opposed to broad-scale distribution in the state of New Jersey. Adam Wallner, a former colleague, as investigated data from black light trap captures and how they are related to land use in almost 50 categories. There are not many urban dwellings or houses nearby the sampled orchards, but distance from the nearest house will be integrated into my model in the future. The conditional autoregressive model can be created and tested for each sampling date. This is the list of AIC scores for the clustering found in Cream Ridge on 8/27/2013. Total ~ Distance: 340.53, Total ~ Variety: 339.04, Total ~ Distance + Variety: 340.95. This indicated that Variety seemed to have more of an influence on the clustering in the field than distance from overwintering sites on this date. Additional variables including distance from other neighboring crops and the number of fruit per tree will be integrated into the model in the future. The amount of damage to peaches has been inputted for each orchard, but has not yet been integrated into the conditional autoregressive model. The same procedure of testing the model for these different factors will be undertaken.

Research conclusions:

One of the main purposes for conducting this project was to lower the amount of pesticide applied to orchards, both to reduce the cost for growers and reduce the environmental input created while producing peaches.

It is my hope that growers will begin to investigate insect invasions more carefully and decide on more economical and environmentally friendly ways to manage their pest populations. If growers are able to reduce the number of insecticide applications in one or multiple fields by concentrating most of their efforts on areas harboring H. halys instead of blanketing all areas with pesticide, this work will have made its impact. This work may even extend to the public, as H. halys can be a severe nuisance pest to homeowners by overwintering in attics and crawlspaces. If areas near agricultural sites with large numbers of stink bugs are managed, the hope is that fewer stink bugs will make it to nearby houses to overwinter. I think growers would benefit from creating maps and examining their fields for hot spots of infestation so they know where and when H. halys will invade their farms.

Participation Summary

Education & Outreach Activities and Participation Summary

Participation Summary:

Education/outreach description:

The results of this work have been presented at the annual meeting of the eastern branch of the Entomological Society of America in Rehoboth Beach, DE (March 14-17, 2015). They will also be presented at the annual meeting of the Entomological Society of America in Minneapolis, MN (November 15-18, 2015) and at future brown marmorated stink bug working group meetings. A portion of this work was translated into an outreach event for middle school students to teach them the utility and interpretation of mapping. The lesson plan from the two outreach events (September 28, 2013 and May 9, 2015) is being formed into a publication that STEM educators will be able to use.

Project Outcomes

Project outcomes:

There were no strict numerical values of the economic repercussions of this research. However, the hope is that with targeted insecticide applications, growers will be able to save on costs related to insecticide use, sprayer operation, time and labor.

Farmer Adoption

Growers have been interested in the immigration of H. halys into their fields. In years with high levels of infestations, growers have resorted to frequent insecticide application. With information about where H. halys travels from, growers will be able to target insecticide applications in these areas, preventing any insects from coming into the field. Targeted application should lead to lower costs and inputs into the field. The results of this research will be shared with growers, extension agents, and researchers at the brown marmorated stink bug working group meetings. Many growers I worked with already conduct sprays of bordering crops as a preventative measure.

Assessment of Project Approach and Areas of Further Study:

Areas needing additional study

This study examines the potential effects of landscape on the immigration of H. halys into fields, and their movement within fields. The results show that numbers and clustering vary depending on the date, and that variety may be tied to these. Additional factors are being considered for the model, and damage levels in the field are being integrated into the model. Research on the actual movement of stink bugs would be an excellent supplement to this project. Knowing whether or not populations exhibit a random or directed walk would help us to finely tune our management practices. Researchers are already using a harmonic radar to see if they can track individuals in the field.

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.