Managing field borders for weed seed predators

Final Report for GS10-091

Project Type: Graduate Student
Funds awarded in 2010: $9,856.00
Projected End Date: 12/31/2011
Grant Recipient: North Carolina State University
Region: Southern
State: North Carolina
Graduate Student:
Major Professor:
Dr. Chris Reberg-Horton
North Carolina State University
Expand All

Project Information


Despite great potential for weed seed eating organisms (i.e. 'weed seed predators') to contribute to ecological weed management programs, there are very few dependable strategies that farmers can reliable use to conserve and enhance these organisms. We conducted a large, field-scale experiment to determine if increasing vegetative diversity along the borders of crop fields increases weed seed predator abundance and enhances the ecosystem service these organisms provide. Our results definitively show that manipulating field border vegetation along the borders of crop fields is not a viable strategy to conserve weed seed predators. However, crops that provided greater overhead cover did see increased weed seed predation and greater numbers of native weed seed predators. Approaches to accomplish this denser cover in the fall, when summer annual weeds are being shed, may enhance weed seed predation services and contribute to an ecological weed management program.


Preventing weed seeds from entering the soil is an important weed management strategy because the weed seedbank is the main source of new weeds in agricultural fields. Increases in the weed seedbank lead to greater management costs in subsequent seasons. Weed seed predation accounts for greater losses to seedbanks than aging, microbial decay or even disturbances like cultivation. Estimates of annual weed seed losses due to granivory by epigeal vertebrates and invertebrates typically range from 33 to 90%.

Despite the potential benefits of weed seed predation, the literature is not conclusive on how to best conserve weed seed predators to create a consistent and dependable weed management strategy. Increasing vegetative diversity surrounding crop fields through set aside programs may enhance this ecosystem service.

The conditions in the humid subtropical climate of the Southeast United States (i.e. the warm temperate zone) raise particular issues for how weed seed predators will respond to increasing vegetative diversity in the agricultural landscape. Although an earlier study confirmed carabid beetles are the predominant weed seed eating invertebrate in the Southeast (Brust & House 1988) a more recent study found the invasive fire ant, Solenopsis invicta Buren, is now the dominant weed seed predator in the region (Pullaro et al. 2006). These non-native, diurnal, social ants will not respond to conservation strategies the same way as native, nocturnal, solitary ground beetles.

This study was undertaken to examine how field border management in the southeastern U.S. affects seed predator abundance and field type. Managed habitats along fields are increasingly present in this region due to cost share programs designed to enhance wildlife habitat in the region, particularly for quail. Multiple field border types were tested, varying in vegetative diversity and management practices. This project was part of a multidisciplinary effort to find a crop field border conservation strategy that maximizes ecosystems services.

Project Objectives:

1.) Determine the effects four different field border vegetation communities have on levels of weed seed predation in adjacent crop fields by invertebrates and vertebrates.

2.) Identify the specific predators responsible for this weed seed predation.

3.) Disseminate the results of this project to organic growers.


Materials and methods:
Objective 1

Research was conducted at the Center for Environmental Farming System’s (CEFS) Organic Research Unit (ORU) in Goldsboro, NC. Nine fields (6-10 acres each) in the ORU were used for this study (see Field Map).

The crops in these fields followed a typical organic rotation for the southeastern United States:, soybeans followed by maize followed by hay. Each of these crops was planted in three of the nine fields every year. The first year the hay crop consisted of orchard grass and white clover. The second year the hay crop consisted of sorghum Sudan grass, cowpea, and forage soybean. The maize was harvested prior to the start of the experiment each year, and the soybeans were harvested after the last trial each year.

Four experimental habitats were established around the borders of each field prior to our study. The first was a frequently mowed grassy border (Mowed). The other three represented various levels of vegetative diversity, 1) unmanaged vegetation (Fallow), 2) a mix of planted native prairie flowers (Flowers Only), and 3) planted native prairie flowers plus native-warm season grasses (NWSG/Flowers). The flower species used for the Flowers Only and NWSG/Flowers borders were lance-leaved coreopsis (Coreopsis lanceolata L.), purple coneflower (Echinacea purpurea (L.) Moench), black-eyed Susan (Rudbeckia hirta L.), butterfly milkweed (Asclepias tuberosa L.), common milkweed (Asclepias syriaca L.), swamp sunflower (Helianthus angustifolius L.), white heath aster (Symphyotrichum pilosum (Willd.) G.L.Nesom), and showy goldenrod (Solidago speciosa Nutt.). The perennial bunch grasses planted in the NWSG/Flowers borders were indiangrass (Sorghastrum nutans (L.) Nash) and little bluestem (Schizachyrium scoparium (Michx.) Nash).

Seed predation assays were conducted during the first two weeks of October and November both years; a September trial was added the second year. Seed predation was measured with point estimates using weed seeds glued to 10×15 cm cards. Twenty-five seeds of three prevalent weed species, redroot pigweed (Amaranthus retroflexus L.), broadleaf signalgrass (Urochloa platyphylla (Munro ex C. Wright) R.D. Webster), or sicklepod (Senna obtusifolia (L.) H.S.Irwin & Barneby) were adhered to each card. Seeds were bought from Azlin Seed Service (Leland, MS). Weed free soil was sprinkled on each card to remove any stickiness. Three cards, each carrying a different weed species, were placed in every exclosure cage.

Each exclosure cage consisted of a coarse wire mesh cylinder (8 in. tall by 8 in. in diameter) with a square base and top (16in. by 16in.) made of the same material. The coarse wire mesh had 4in. by 4in. openings. The tops were covered with aluminum window screen (mesh size < 0.1in.) and had a 8in. by 8in. opening in their centers covered with a square piece ( 9in.) of removable aluminum window screen secured with Velcro ® strips around its perimeter. Three types of cages were used for this experiment. The first (ALL) included only the basic 4 by 4 inch coarse wire mesh which allowed all seed predators, including vertebrates, inside to feed on the seeds . The sides and base of the second cage type (INV) were covered with 0.5 by 0.5 inch hardware cloth, which excluded vertebrates, such as mice and birds. However, the cage could easily be entered by invertebrates. The third type of cage (NONE) served as a control and used aluminum window screen (mesh size < 0.1in.) to exclude all seed predators.

Each of the three cage types were placed at measured intervals along transects extending from the middle of each experimental field border into adjacent crop fields. The first set of cages was placed at the interface of the field border habitat and the crop field (20 feet from the habitat center). The second set of cages was placed 40ft from the habitat center, and the third set 115ft from the habitat center. The third set was placed so the cage was >130ft from a neighboring field border habitat. In total there were 324 total cages per trial each with three seed cards (972 total seed cards per trial).

After two weeks of exposure in the field all of the cards were collected, placed individually into labeled plastic bags, and returned to the laboratory. The number of seeds remaining on each card was counted.

Objective 2

Two studies were conducted to determine what weed seed predators were prevalent.

The first was conducted in during the same time as the seed cards were deployed in the Organic Research Unit at CEFS. Catches from pitfall traps (4in. diameter) were used to measure the relative abundance (i.e. activity-density) of ground dwelling invertebrate weed seed predators. Traps were placed along the transect extending from the middle of each experimental field border. One trap was placed in each of the following locations: in the center of the border habitat, at the interface of the border habitat and crop field (20ft from the habitat center), and 40, 65, and 115 feet from the habitat center. Traps were made of two nested 450 mL plastic containers.

Traps were opened at the beginning of each month’s predation assay, filled with 50% ethylene glycol solution to a depth of 1 inch, and resealed after 96 hours. Invertebrate specimens from each trap were preserved in alcohol and brought back to lab to be identified and counted. Ground beetles (Carabidae) were identified to species while crickets (Gryllidae) were identified to genus. Ants were counted as one group because the vast majority of the ants in these fields are red imported fire ants. Identifications were confirmed by David Stephan of the North Carolina State University Plant Disease and Insect Clinic (Raleigh, NC).

THE SECOND STUDY conducted to identify the specific predators responsible for weed seed predation in the Southeast utilized a monitoring study using night vision cameras. Research was conducted in a 10 acre conventionally managed soybean field in 2011. 6 plots, each 100ft by 100ft and each at least 100ft away from each other were used. 3 of these plots were treated with hydramethylnon to remove the insect weed seed predators, including the red imported fire ants.

A monitoring arena was set-up in the center of each of these six plots. The arena consisted of a 4inch diameter Petri dish covered in weed-seed-free soil encased in an exclosure cage that allowed all weed seed predators to enter. On top of the exclosure cage a night vision camera was positioned so that it focused on the entire area of the covered Petri dish. Five seeds of three prevalent weed species, redroot pigweed, broadleaf signalgrass, and sicklepod were placed on top of the covered Petri dish and were filmed for three hours during the day (12:00-15:00) and three hours during the night (20:00-23:00). The study was conducted over three consecutive days in October 2011. Video tape was analyzed to find what weed seed predators were present and what organisms were removing or eating the weed seeds.

Research results and discussion:

Increasing vegetative diversity along field borders did not enhance weed seed predation rates in crop fields nor did it increase the abundance of prevalent invertebrate weed seed predators. None of the experimental habitats influenced seed predation rates for any of the three weed species. Although we found three taxa of invertebrate weed seed predators, both native and non-native, none of the field border types increased any of these organisms’ activity-densities.

We found three prevalent invertebrate weed seed predators: the red imported fire ant (Solenopsis invicta), field crickets (Gryllus sp.), and the omnivorous ground beetle, Harpalus pensylvanicus.

Unlike field border vegetation, crop vegetation had a large effect on weed seed predation. The reduced weed seed removal rates and lower activity-density of native invertebrate weed seed predators in the harvested maize fields support the hypothesis that increased cover enhances weed seed predation. (See Figure 1)

While the red imported fire ant was the most common invertebrate found in our fields, their high numbers did not correlate with high weed seed predation. Seed predation was high in the hay fields, higher the first year, and highest in October both years. Fire ant numbers, on the other hand, were lower the first year and seemed to be the lowest in the hay fields (Figure 1). Fire ants had the highest numbers in September rather than October, most likely because of the higher soil temperatures (Figure 2). Results from our seed predation assays match better with the activity-densities of the native invertebrate seed predators, especially field crickets. The activity-densities of native seed predators were high in October (Figure 2), and the especially high seed predation in first year's hay fields corresponded with elevated numbers of field crickets (Figure 1).

Our video study confirmed these results --- showing that even when the invasive red imported fire ant is present in high numbers they do not contribute to weed seed predation services. Instead, native organisms, such as mice and field crickets, are the ones destroying weed seeds in Southeast crop fields.


-Two year study completed on the impact of field border vegetation on weed seed predation services
-Manuscript being submitted
-Results presented at three scientific meetings
-Poster presented at Center for Environmental Farming Systems field day
-Poster will be presented at the Carolina Farm Stewardship Association's annual conference

-Video monitoring study completed to determine the predominant weed seed predators in the area
-Manuscript being prepared
-Results presented at a scientific meeting and available on the Entomology Society of America's YouTube channel;

-Results presented at the Entomological Society of America's annual meetings and the Ecological Society of America's annual meeting in 2012
-I led a half-day extension workshop that was attended by over 50 people
-An extension publication on weed seed predation will be published next spring

Participation Summary

Project Outcomes

Project outcomes:

Although weed seed predation holds great promise, this ecosystem service cannot be depended on without management strategies that consistently conserve these organisms. Manipulating vegetative diversity along the edges of crop fields is not a viable conservation strategy for these organisms in the southeastern U.S. However, our results showing a large effect of vegetative cover on seed predation is in accordance with other research and should be promoted as a viable conservation approach. Efforts to accomplish this denser cover, though, such as delaying harvest or delaying hay cutting, are not likely to be adopted due to economic constraints. Other suggested approaches, such as eliminating or delaying fall tillage or using hay species that provide denser cover may be more tenable approaches that work with growers’ rotations and economic restraints.


Areas needing additional study

Manipulating vegetative diversity along the borders of our crop fields may not have influenced weed seed predation for a number of reasons specific to our region and study site, including (1) the high diversity of the surrounding landscape, (2) the high resource abundance in organic cropping systems, (3) the removal of native weed seed predators by an invasive species.

Future studies need to determine if these factors, not present in all regions and fields, are in fact reducing the impact field border manipulations have on weed seed predation.

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.