The concept of thresholds has many applications in weed science, depending on the response being measured. The most common adjectives used to describe thresholds are damage, economic, period, and action. Damage threshold is the term used to define the weed population at which a negative crop yield response is detected. An economic threshold is the weed population at which the cost of control is equal to the crop value increase from control of the weeds present. Economic threshold may be used to describe short-term effects of weed interference occurring in a single growing season, or multiple-season effects including some cost associated with seed produced by uncontrolled plants. The term period threshold implies that there are times during the crop cycle in which weeds are more or less damaging than at others. Action threshold is the point at which some control action is initiated, and usually includes economic considerations along with other less tangible factors such as aesthetics, risk aversion, or sociological pressures. Regardless of the type, thresholds imply that weed effects are population dependent, and as such, allow some type of prediction to be made relative to the consequences of control decisions. One successful approach to the implementation of thresholds has been through the development of computerized decision-aid software. These programs allow users to compare economic and environmental consequences of potential control actions before committing to one particular decision.

Computer models designed to simulate the fate and transport of pesticide residues in surface runoff and water that recharges shallow aquifers are used at several stages in the regulation of pesticides. Models can be a powerful tool to identify environments in which pesticides are more mobile and are more likely to impact ground-water quality. Models can also be used to highlight pesticides that can contaminate surface or ground water at an early stage of the registration process. At present, models are utilized primarily as comparative tools to identify which compounds should be examined on a priority basis. Results may be used to calibrate numerical models and estimate the long term impact on the environment from the use of a particular product or its alternatives. To date, field monitoring studies have been required for over 40 chemicals by the Office of Pesticide Programs (OPP) to evaluate the leaching potential of pesticides under realistic usage conditions. As the capabilities of models improve and linkages are developed between databases and the models, computer simulation models will play an increasingly important role in the Environmental Protection Agency's (EPA's) efforts to manage pesticide use to prevent adverse effects on the environment.

Two introduced windgrass species have become crop weeds in North America. Common windgrass is a major weed of winter cereals in Europe and was first documented in North America in the early 1800s. It is a weed of roadsides and waste areas in the northeastern United States and in winter grain fields of southern Ontario and Michigan. Interrupted windgrass was first reported in North America approximately 90 yr ago; it is adapted to more arid sites than common windgrass and is distributed predominantly in the northwestern U.S.A. During the past 10 to 15 yr, interrupted windgrass has adversely affected winter grain and grass seed producers in the Pacific Northwest due to additional control costs.

Field studies were conducted in 1989 and 1990 to determine the potential interaction of terbufos and nicosulfuron on corn. Terbufos applied in-furrow at planting interacted with nicosulfuron applied POST to cause significant injury 25, 30, 40, 50, and 60 d after planting in 1989 and 1990. Injury decreased significantly when nicosulfuron was applied at later growth stages of corn. Plant population was not affected by terbufos and nicosulfuron regardless of the growth stage of corn at application of nicosulfuron in either year. Corn ear number and grain yield in 1989 and 1990 were lower in plots treated with terbufos at planting and nicosulfuron applied at the three leaf stage than in plots treated with only nicosulfuron. There were no differences in the height of corn, ear number, or grain yield when nicosulfuron was applied at the seven-leaf stage regardless of terbufos application in 1989 or 1990.

A survey conducted to determine the distribution of red horned poppy in western Oklahoma identified infestations in 14 Oklahoma and 3 Texas counties. Soil pH from infested fields ranged from 7.0 to 8.5. Red horned poppy grew as an annual, germinating in fall, spring, or summer. Plants that emerged in the fall survived −25 C, flowered in April and produced mature seed in June. Plants that emerged in April, flowered in June and produced mature seed in July. Corolla color varied from red to yellow with time of anthesis.

A screening procedure has been developed to evaluate the relative loss of pesticides from soils, to assist the Soil Conservation Service in implementing water quality for conservation planning. This screening procedure is a first-tier evaluation of the impact of using a particular pesticide on a specified soil. The screening procedure was developed from evaluating pesticide loss from over 40 thousand runs of the GLEAMS (Groundwater Loading Efects of Agricultural Management Systems) model. The model input data were a combination of soils and pesticides with a wide range of properties. The estimated pesticide losses were categorized into losses by leaching, losses adsorbed on sediment in runoff, and losses in the solution phase of runoff. Algorithms using soil properties were developed to rate soils into four loss potential classes for leaching and three loss potential classes for the two categories of runoff. Also, algorithms using pesticide properties were developed to rate pesticides into four loss potential classes for leaching and three loss potential classes for runoff. The soil and pesticide groupings are combined in a matrix to give an overall loss potential rating. Statistics of the overall loss potential indicate the low loss potential class is pure, that is, it does not contain occurrences that have medium or high losses. The medium loss potential classes does not contain occurrences of high loss, but does contain many occurrences of low loss. The high loss potential class contains incidences of medium and low loss.

A theoretical model was developed to assess the relative importance of weed competition and seed dispersal in determining long-term crop yield reductions imposed by weeds. Weed seed dispersal by cultivation, natural means, and harvest machines was incorporated into a weed and crop population model. Model simulations were conducted with green foxtail and spring wheat demographic data from the literature. Simulations indicated that seed dispersal from a weed species invading a field may have a more important influence on crop yield than the relative competitive ability of the weed with respect to the crop. However, if a weed species has a uniform or a high frequency random distribution, there is less difference in the relative importance of dispersal versus competitive ability in determining crop yield reductions. Harvest machines may be an important mechanism for dispersing weed seeds that are harvested with the crop. The practical implication is to make harvesters weed seed ‘predators’ rather than dispersal agents and thereby reduce herbicide and mechanical weed control inputs.

As an extension of weed threshold models in which crop losses are based on weed density, an alternative model for grass weeds in cereal crops is proposed that incorporates the theoretical importance of selection for herbicide resistance, initial weed population frequency, and weed seed dispersal. Simulations suggest optimum weed population levels (thresholds) for maintaining genotypes that are susceptible to control practices and which minimize crop yield reductions. Weed population frequency, in combination with dispersal and competitive traits may determine optimum weed management strategies/Model simulations indicate that understanding how agricultural practices select for “weedy” traits (e.g. herbicide resistance, competitive ability, dispersal potential) may be important in determining weed density thresholds.

The environment for agricultural research and development, technology transfer, and production is marked by conflict among persons with diverse ideas and goals for agriculture. The objective of this analysis was to identify and compare models for researching and problem solving that can provide a conceptual framework for understanding and improving complex situations marked by conflict. The research activities of scientists involved in development of genetically-engineered-herbicide resistance were modeled as reductionist science, technology development, and optimizing systems. An analysis of these models of goal-seeking research indicated that values and assumptions implicit in goals such as greater productivity were not evaluated or questioned. Views of experts influenced development and application of technologies and systems more than concerns of producers and society. A soft systems methodology and research system is proposed to involve more diverse ideas or views of the world, to shift the role of the researcher from expert to facilitator, and to move toward consensus concerning research and technology development in agriculture.

Field experiments were conducted in 1990 and 1991 near Artesia and Las Cruces, NM to evaluate directed POST treatments of oxyfluorfen for the control of broadleaf weeds that emerge after thinning chile peppers. At Artesia, Wright groundcherry and prostrate pigweed reduced red pepper yield by 33% in untreated plots in 1990 and clusterflower reduced yield by 19% in 1991. At Las Cruces, Wright groundcherry, Palmer amaranth and common lambsquarters reduced yield by 61% in 1990 and by 76% in 1991. A sequential treatment of oxyfluorfen at 0.14 kg ai ha-1 2 wk after thinning, when the peppers were 15 cm tall, followed by 0.14 kg ha-1 4 wk after thinning controlled all species except common lambsquarters without injuring the peppers.

Do quasi-physically based models with more detail perform better than regression or other empirical models? This is a question that was raised many years ago and still remains. In an effort to respond to this question, the author reviews the needs and concerns of users and then divides the large number of models into three classes: (1) screening, (2) research, and (3) planning, monitoring, and assessment. Empirical and causal (physically based) models are contrasted and the advantages and disadvantages of each described. Sources of model uncertainty (properties of data bases, model structure, parameter estimation methods, algorithmic implementation, verification and validation, and future users) that lead to skepticism about models' performance are investigated. Simulation scale and spatial variability are also important considerations. The leading question is then discussed from the perspective of screening, research, and planning, monitoring, and regulatory models.

The strategy to optimize weed management systems with a minimum use of herbicides includes both the adaptation of crop management practices and well designed decision making systems, based on postemergence observations of weed infestations. Both strategies require thorough quantitative insight into the crop weed ecosystem, which can be provided by systems analysis, using process based models. These models also can be applied to similar systems like intercropping. For practical application, however, a simple measure of weed infestation and a simple model which relates weed infestation to yield loss are required. Recent progress in model development is discussed. An alternative empirical model that predicts yield loss from the relative leaf area of the weeds shortly after crop emergence, seems to be a useful approach for prediction of yield loss shortly after crop emergence. The use of systems approaches at different levels of detail for bridging the gap between research and practical application is discussed.

Low-spatial resolution satellite imagery from the NOAA-10 polar-orbiting meteorological satellite was analyzed to determine if central New Mexico grasslands infested by broom snakeweed could be discriminated from unaffected areas. Distinctive phenological characteristics of broom snakeweed, including an early season growth flush and late season flowering, enable moderate to heavily infested areas to be separated from grasslands having few or no weeds present. The procedure used shows promise as a tool for locating and monitoring brown snakeweed and other weeds growing on shortgrass prairie.

Two methods were used to obtain the sensitivity of chemical leaching depth to variations in the input parameters of the Pesticide Root Zone Model (PRZM). First a Plackett-Burman (PB) screening design was used to vary 35 PRZM inputs over seven ranges around a nominal value. Six of the seven ranges were approximately 0.1, 0.25, 0.5, 1.0, 5.0, and 15%, the seventh range was chosen to cover a range appropriate for a soybean herbicide applied preemergence in the Midwestern region defined by the USDA–SCS land resource region M. Next, Fourier amplitude sensitivity testing (FAST) was then used to vary from 19 to 25 parameters over four of the ranges previously tested. For the smaller parameter ranges the two methods typically gave equivalent results but the PB method required far fewer simulations. For the simulation of the Midwestern region where some parameter varied by larger amounts the relative magnitudes of the sensitivity coefficients obtained by the two methods were similar but the magnitude of the coefficients obtained using FAST were smaller than those obtained using PB.

According to Grime's model, velvetleaf can be classified as a competitive-ruderal species. Its biological and ecological characteristics make it particularly competitive and persistent in cultivated fields. Herbicide control is difficult and, once established, it is almost impossible to avoid seed production. The economic threshold, on a single year basis, varies between 0.3 and 1.7 plants m–2. When the competitive relationships and their effects are evaluated, it is very difficult to apply a weed management system based on a single-season threshold. When one considers the agronomic implications of the ecological characteristics of velvetleaf over a longer period, it can be concluded that the use of a single-season economic threshold is probably not the best management strategy.

Organizer's comments: Systems inquiry represents an array of research-based problem-solving approaches. European experience with the pesticide debate suggests several simple messages contained in this paper. As weed scientists, we have everything to gain by considering a global sphere of views, practices, logic, experience, and message. Read this paper for the essence, the message, the systems thinking, the research framework, and the interface between society and weed science. Read it because science and society must communicate.

Experiments were conducted to evaluate the effect of tillage, rye residue levels, and control strategies on weed suppression and potato yield. Reduced tillage (RT) yields in 1989 were lower than in conventional tillage (CT) due to high hairy galinsoga populations. However, in 1990, potato yields were greater in RT when early emerging weeds were suppressed. Increasing residue levels up to 2300 kg ha-1 had no effect on weeds or yields in either year. In a separate study, preemergence herbicides (linuron plus metolachlor, 1.7 kg ai ha-1 each) applied 10 d after planting (DAP) in reduced tillage controlled weeds for 6 wk, but efficacy decreased after hilling. Metribuzin applied postemergence (0.19 kg ai ha-1) 7 d after hilling controlled late germinating weeds. Potato yields were greatest when the preemergence herbicides were applied.

The GLEAMS model (Groundwater Loading Effects of Agricultural Management Systems) is used to illustrate model application in evaluating potential pesticide runoff of two similar pesticides from one soil. This limited application was chosen for simplicity in illustrating relationships between annual means and single events. When using annual totals of simulated pesticide runoff for comparing two pesticides or assessing risks, long-term 50-yr simulations are preferable to short 10-yr simulations. When short-term simulations are performed, care must be exercised in selecting representative climatic periods. For short half-life pesticides, as demonstrated in this study, initial rainfall events on or near the day of application will often contribute most to annual pesticide lost. In these cases, single event analysis may be required. Procedures are demonstrated for expressing annual total pesticide losses and single rainfall event losses probabilistically in terms of expected recurrence intervals.

Barnyardgrass at 10 or more plants per 1 m of row caused over 80% root yield loss in spring-sown sugarbeets in California. The yield loss decreased to between 5 and 20% when barnyardgrass densities were one plant ever 2 to 3 m of crop row. The relationship between crop yield and barnyardgrass density was described by a rectangular hyperbolic function. The economic threshold density was in the range of one barnyardgrass plant every 5 to 20 m or row. Barnyardgrass plants growing at densities of less than one plant per 1 m in sugarbeets experienced only interspecific competition. Such individual barnyardgrass plants were estimated to produce from 30 000 to over 200 000 seeds, with an average of 100 000 seeds per plant. This seed production equated to a seed rain of about 4000 to nearly 20 000 seeds per m from a density of one weed every 10 m of crop row. These data suggest that using single-year economic thresholds for barnyardgrass is not a sound management practice because even low densities can produce large quantities of seed.