One year's seeding: a seedbank approach to sustainable weed management

To fulfill the lack of information available on sustainable weed management the extension bulletin “Integrated Weed Management ‘One Year’s Seeding…’” (IWM) was released in February 2005 as the result of a collaborative effort of researchers, extension educators, and producers. Since then, over 2,000 copies have been sold, several workshops have been held to present the information to North Central Region farmers, and fifteen on-farm trials have been conducted to test some of the weed management methods published. Thus far, information we have collected from growers indicates that these activities have increased their knowledge of the diversity of weed management techniques.

Introduction:

An exclusive focus on killing seedlings is a major obstacle to sustainable weed management in the North Central Region (NCR), and is reflected in adoption rates of Roundup Ready™ crops as high as 80% in the NCR (Owen and Zelaya, 2002). Dependence on herbicide-resistant crops threatens sustainability through weed species shifts, herbicide resistance, loss of weed management knowledge, chemical dependence, technology fees, and restricted grain markets. Over-use of physical control of weed seedlings in alternative farming systems drives up fuel consumption, destroys soil tilth, and neglects the weed seedbank, which controls long-term weed management success (Liebman and Davis, 2000).

Models of cropping system effects on annual weed populations (Jordan et al., 1995; Davis et al., 2003) show that reducing weed seed numbers in the soil, especially through decreased seed survival, is the key to keeping weed populations small and manageable. For example, underseeding small grains with forage legumes can increase insect consumption of weed seeds and reduce weed populations, compared to sole crops of small grains (Davis et al., 2003). To help farmers implement a seedbank approach to weed management, decision aids are needed.

Michigan State University has made a commitment to providing high-quality, user-friendly agroecology information to producers in Michigan and throughout the North Central Region (Cavigelli et al., 1998; Cavigelli et al., 2000). Through an EPA Pesticide Environmental Stewardship Program grant (June 2003), our team assembled a diverse working group (the Ecological Seedbank Management Working Group) to produce a draft of an easy to use weed seedbank management guide and train MI farmers in its use. In this project we proposed to 1) create a finished version of the manual for publication, 2) train farmers throughout the NCR to use the manual, 3) field test the manual throughout the NCR, and 4) conduct and evaluate on-farm trials to explore new seedbank management methods.

The rationale for this project is that producers looking for alternatives to conventional production practices tend to look to other producers as their first source of information (Walz, 1998), yet this information is not always shared widely, nor does it affect the research agendas of most land-grant weed scientists. We aimed to make public the decision making process of outstanding producers who are already trying to manage the whole weed life cycle, and then extend this information to other producers and to researchers. In doing so, we increased adoption of existing ecological weed management practices and identified gaps in existing weed management knowledge to direct future research. A key assumption in this project was that lack of relevant, reliable information was holding back adoption of ecological weed management practices. The high priority placed on non-chemical approaches to weed management by farmers in the 3rd biennial national survey of organic farmers (Walz, 1998) suggests that this assumption was well-founded.

Annual weeds spend most of their lives as seeds in the soil seedbank (Cousens and Mortimer, 1995). Simulation models of management effects on the population dynamics of annual weeds show that management practices that kill seeds when they are in the seedbank, or prevent new seeds from entering the seedbank, will have a much larger proportional effect on weed population growth and size than management practices aimed solely at killing weed seedlings (Jordan et al., 1995; Davis et al., 2003). Most weed control in the U.S. is targeted at killing weed seedlings, either with herbicides or with cultivation. Ninety percent of all corn and soybean acres in the U.S. are treated with herbicides, totaling more than 90.6 million tons of active ingredients applied annually (Pike et al., 1995). The literature is likewise focused on ways of controlling the weed seedling stage. A search of the agricultural database AGRICOLA in October, 2003, returned 19,490 entries for the search term “herbicides”, 704 entries for the search term “cultivation & weeds”, and 284 entries for the search term “weed seed”, and 32 entries for the search term “weed seedbank.”

Techniques for eliminating weed seedlings are more common than those aimed at weed seeds for an obvious reason: seedlings are much easier to find and destroy than seeds are, especially seeds that are already in the soil seedbank. We are not suggesting that weed seedling control be abandoned; after all, it is a very effective way of reducing potential inputs to the weed seedbank. Rather, seedling control should be complemented with methods that target other stages in the weed life cycle (Mohler, 1996), with the overall goal of reducing weed seedbank density. Inevitably, some seedlings are going to escape early-season control efforts and survive to produce seeds. Can farmers reduce the damage done by these weeds? Can this be done without increasing herbicide applications? And if the weed seedbank is reduced, can this aid future weed control efforts? We believe that the answer to each of these questions is “yes.”

Although most research has focused on weed seedling control, there have been significant findings over the years that suggest good opportunities for seedbank management (Renner, 2000; Buhler, 2002; Elstein and Suszkiw, 2003). Crop competition with weeds is one of the strongest factors affecting the number of seeds that each weed produces, and can be manipulated to minimize weed seed inputs to the soil seedbank (Perera and Hartwig, 1980; Jordan, 1993; Mohler, 2001). Pathogenic fungi and bacteria are important causes of seed death in the soil seedbank (Kremer, 1993; Kennedy and Kremer, 1996), yet little is known about the factors controlling pathogenic attack of weed seeds. Stale seedbed techniques are regularly practiced in low-external-input cropping systems in Europe, greatly aiding other weed management practices (Hatcher and Melander, 2003). Weed seed destruction by insects, rodents and birds can be very important in reducing the weed seed rain (Marino et al., 1997; Menalled et al., 2002), and rates of predation may be manipulated by cropping system characteristics (Davis and Liebman, 2003). Mechanical destruction of weed seeds by harvesting equipment, once a common practice, has also been gaining in popularity (Gossen et al., 1998), and is now commercially available for small grain production in Canada. Interest in weed seedbank management does appear to be growing among weed scientists. A recent international conference on weed seedbank biology and management in Reading, UK, was attended by over 40 delegates from 18 countries (Bakker et al., 2003).

A number of projects in the SARE database were particularly relevant to the extension and research activities that we propose to conduct. The project “Controlling cheat and annual ryegrass in small grains using novel crop harvesting technologies” (Peeper, AS96-025) showed that roller or hammer mills added as aftermarket attachments to combine grain harvesters could dramatically reduce survival of cheat grass seed in small grains. Other projects, including “Cover Cropping and Residue Management for Weed Suppression, Soil Fertility and Organic Crop Production” (LS02-132, Baldwin), and “Diversity & Intensity of Cover Crop Systems: Managing Weed Seed Bank & Soil Health” (LNE01-141, Gallandt) have examined the relationship between cropping system diversification with cover crops and weed seedbank dynamics. One recently funded proposal, “Microbial processes underlying the natural weed suppressiveness of soils” (LNC03-225, Hallett) represents an important advance in seedbank management research—mechanistic as well as descriptive studies of seedbank processes.

Our proposal was complementary with the above proposals, but is also unique. The above projects were primarily driven by the research team, with some on-farm sites and extension of results. Our project was based on the premise that farmer involvement in the question asking stage (e.g. LNC97-112, Mutch: “Enhancing farmer adoption of sustainable agriculture practices via farmer-driven research”) is critical to generating knowledge that farmers want to use. Our work synthesized current farmer and researcher insights into seedbank management strategies, presented this information to producers throughout the NCR, obtained feedback from these farmers on areas that need more work, initiated on-farm research to fill some of these knowledge gaps, and lead to the creation of a sequel to the decision guide. By involving farmers at each stage of the process, our intent was to keep the information relevant and useful to them.