Introduction and Summary
Herbicides account for nearly 70 percent of the total volume of pesticides applied in American agriculture. Reductions in herbicide use are needed in several intensively farmed regions to reduce cash production costs and well-documented adverse environmental effects, while also improving food safety and lowering risks incurred by farmers, applicators, farm-workers and rural neighbors.
Since the early 1990's several countries have adopted and achieved ambitious pesticide use reduction goals based on volume of active ingredients applied. Some farm organizations, environmentalists and public health specialists in the United States are urging adoption of similar initiatives. The Clinton Administration's 1993 pledge to gain adoption of IPM on 75 percent of harvested acreage by 2000 is fostering change in agency programs and priorities. Subsequent policy statements by Administration officials have included pesticide use reduction among additional important goals, but have stopped short of setting a numeric target or time-table.
In Sweden, Denmark, and the Netherlands, reductions in the volume of herbicide active ingredients (a.i.) applied have contributed decisively to achieving 50 percent or greater reductions in overall pesticide use. The reductions have been achieved in large part by shifting acre-treatments to new low-dose herbicides applied at 0.2 pound a.i. per acre or less, coupled with reducing acre- treatments with triazine and acetanilide products applied between 1 and 2 pounds a.i. per acre. Many question the net environmental benefits associated with such a shift and have called for deeper reflection by government, farmers, and crop protection specialists both on how to set goals for pesticide use and risk reduction, and how to establish measurable targets that can be monitored over time, triggering shifts in R+D priorities and policies when warranted.
While reducing the volume of pesticides applied is desirable, the preferred goal for public policy and challenge for researchers, pest management specialists, industry and farmers is reducing reliance on pesticides. The best way to reduce reliance on pesticides is adoption of bio-intensive integrated pest management systems.
This paper reports results of an empirical application of a new methodology that links adoption of integrated pest management (IPM) systems – in this case, integrated weed management (IWM) -- to indicators of reliance on pesticides and pesticide use. The focus herein is weed management in corn and soybeans in the mid-west during 1994. Farm-level data on herbicide use patterns and prevention-oriented weed management practices in the USDA's cropping practices survey are used to divide surveyed soybean and corn acreage into four zones along what is called the IWM continuum – "No", "Low", "Medium", and "High" levels of integrated weed management.
These levels of adoption correlate roughly to those studied by the Economic Research Service in its 1994 report "Adoption of IPM in U.S. Agriculture" (Vanderman et al); the "High" level of IWM systems studied herein correspond to those that fall within the bio-intensive, or "High" level zone along the IPM continuum (for further details on the IWM continuum and the results of the analysis reported herein, see the World Wildlife Fund Technical Monograph "Reliance on Herbicides and Adoption of Integrated Weed Management Systems in the Midwest" [Benbrook, in press]; for a detailed discussion of the IPM continuum and the distinguishing characteristics of bio-intensive IPM, see "Pest Management at the Crossroads", [Consumers Union, in press]).
While the focus of this discussion is methodological, preliminary empirical results are reported, emphasizing differences in herbicide reliance and use in the four zones along the IWM continuum. Results lend insight into the potential for reducing herbicide application rates, pounds applied, and acre-treatments as more farmers progress along the IWM continuum toward more biologically and ecologically based multi-tactic weed management systems.
A. RELIANCE ON HERBICIDES AND MEASURES OF HERBICIDE USE
Weed management systems strive to keep weeds from becoming a limiting factor in determining crop yields in a given season (Wyse; Swanton and Murphy). A second goal is to prevent the buildup over time of weed seedbanks, which would lead to tougher weed management challenges in future years (Forcella et al). A wide range of decisions in the selection and integration of crop and livestock enterprises, and in choosing tillage, planting, fertility, and related management practices each production season influences weed species and pressure (Liebman and Dyck). Weeds are known to co-evolve with farming systems (Ghersa et al; Jordan; Buhler et al). In general, weed pressure tends to rise in farming systems lacking diversity over time and space in tillage systems, rotations, planting methods and density, cultivation and herbicide use patterns.
Data presented herein suggests that farmers have already gone a long way in reducing herbicide application rates below maximum rates recommended on product labels. Further progress in reducing reliance on and use of herbicides will, on most farms, need to come from multi-tactic preventative approaches that draw on knowledge of weed ecology and biology. The key challenge in the field is managing diversity to constantly place problem-weed species at a competitive disadvantage, through what has been called the "many little hammers" approach (Liebman and Gallandt).
Mortensen has addressed linkages between preventative practices and weed thresholds, making the point that a higher degree of weed survival then typically acceptable on most farms today can be tolerated if other preventative practices limit growth in seedbanks by reducing the number of set seeds that remain viable and survive beyond germination (Mortensen et al). Ongoing research in several states has shown that farmers and weed managers have the option of incorporating a broader range of preventative approaches and tools if and when it becomes necessary or desirable to reduce reliance and use of herbicides (Swanton and Murphy; Wyse; Liebman and Jaenke; Forcella and Burnside). Jordan reports that a 20 percent reduction in over- winter seed survival has a significant positive impact on the costs of integrated weed management systems incorporating higher thresholds (Jordan, in press) and concludes that "an integrated preventive system can be highly effective, even when system components -- individual preventive practices -- are only of low to moderate effectiveness."
1. Measures of Herbicide Use
Herbicide use is typically measured by the volume of active ingredient applied per acre per year, or per year in a geographic area. Both volume-based and spatial measures are needed in monitoring herbicide use and reliance. Changes in herbicide technology since the mid-1980's have led to the need for new indicators to supplement volume-based measures, which alone are miss- leading indicators of reliance and use, especially in soybean weed management. Several measures and/or indicators of herbicide use are needed in combination to fully understand current levels and trends in herbicide use and reliance --
Several different measures are encountered in state and federal government reports, and in private surveys presenting data on herbicide application rates per acre, acre-treatments, and pounds applied. Different tables in various government reports on herbicide use in corn production might include several different measures of application rates per acre --
In reporting estimates of application rates for individual herbicide active ingredients on a given crop, there are other options -- application rates of herbicide x on corn acres treated with herbicide x, treated with any herbicide, or all planted acres. In some tables, rates of application per acre are reported as the pounds of a herbicide a.i. applied on a crop in a region divided by the acres treated with any herbicide in the region, a misleading calculation.
Other Measures Simple measures of use and reliance fail to take into account the highly variable design and consequences of different pest management interventions. Additional concepts, data and methods will be needed to fully capture this diversity in pest management systems (see for example Kovach; Pease and Landy; Benbrook, 1995). Other measures of reliance useful in certain circumstances include --
In insect and plant disease control, where reliance on synthetic pesticides is high, reducing use of broadly toxic products is generally a necessary first step in the transition toward more biologically-based pest management systems because of the need to enhance biodiversity and build up the levels of beneficial organisms. Combinations of factors -- weather, resistance, emergence of a new strain of a weed, plant disease or insect -- will arise occasionally and cause pest populations to swell; reliance on pesticides and/or other control measures will increase for a season or two. Where and when reliance on pesticides rises steadily over several years it is a sign of adverse developments in the ecological interactions among organisms and the need for farmers and pest management specialists to become more creative in incorporating preventative practices into farming systems.
2. Measuring Reliance
Reliance on chemical, mechanical, genetic and biological pest management interventions is most accurately measured within a reasonably homogenous production region for a given crop and/or crop rotation. Reliance on pesticides is a sub-set of reliance on all pest management interventions, and reliance on synthetic pesticides, in contrast to biopesticides, is yet another important sub-indicator of the performance and nature of IPM systems. Reliance is inherently crop and field-specific and therefor can be most accurately measured at the field level. County, state, and national level indicators and measures of reliance can be estimated by aggregating indicators and measures from field level data.
Reliance should be monitored according to major categories of pests. A composite measure of reliance can be constructed, and would entail aggregating separate measures of reliance on interventions in controlling weeds, insects, and plant diseases. The importance of these three distinct measures of reliance could be weighted in several different ways: as a function of the share of pesticides applied (the weight for reliance on herbicides in weed management could equal total pounds of herbicides applied [or, dose adjusted acre treatments] divided by the total pounds of all pesticides applied [or, all acre treatments]); the share of total pest management expenditures, or expenditures on pesticides; or, relative to regional water quality and/or food safety challenges.
Reliance and Risk To respond to society's concerns and needs, measures of pesticide reliance and use need to also be assessed in terms of net changes in risk outcomes. Reductions in pesticide reliance over time will reduce the volume of use, and hence risks, from what they otherwise would be. But when use reduction -- measured by volume -- is accomplished by a switch from a herbicide typically used at 2 pounds per acre to a much more biologically active product applied at one-tenth pound per acre, the linkage between use and risk reduction requires further consideration. Despite a 95 percent reduction in use, the low rate pesticide is likely to be both more toxic, at least to certain species of plants and possibly soil microorganisms, as well as more persistent. Risks might rise as a result, especially from the perspective of the farmer.
In order to more fully assess changes in herbicide reliance, use and risk, it is essential to study changes in dose-adjusted acre-treatments, rates of application and pounds applied, as well as changes in the relative toxicity of products applied and in exposure levels and pathways, as effected by timing and methods of application. Necessary and sufficient conditions to conclude that pesticide reliance, use, and risks have fallen require monitoring trends in the volume of pesticides applied (and which pesticide active ingredients are applied), the number of times a given acre was treated and the number of different active ingredients applied (and how), rates of application and timing of applications, and the relatively toxicity of treatments.
B. ADOPTION OF INTEGRATED WEED MANAGEMENT SYSTEMS
IPM systems are dynamic and exist in limitless variety. It is hopeless to try to codify a set of practices that represent IPM because farming systems, technology, and pest complexes and levels of pressure are constantly changing (NAS, 1995). As public programs and agencies seek to promote IPM, a way is needed to distinguish those IPM systems that warrant explicit support through expenditure of public dollars and/or through policy reforms, in contrast to those that do not.
One way to study the consequences of pest management system choices is to divide IPM systems into categories or zones along a continuum, or spectrum, and then compare and contrast pest management practices and pesticide use in various zones along the continuum (for early applications of this method, see the World Wildlife Fund report Reducing Reliance on Pesticides in Great Lakes Basin Agriculture [Hoppin et al]; USDA's analysis of IPM adoption [Vanderman et al]; Benbrook, 1995; and Shifting Gears: The Transition to Bio-Intensive IPM [Consumers Union]. In this analysis, the organizing principle defining the zones is reliance on treatment-oriented interventions, particularly application of synthetic chemicals, in contrast to prevention-based practices, particularly those that are genetic, biological and/or ecological in nature. This organizing principle both reflects the ecological roots of IPM (Hinkle and Cates; NAS, 1995), and establishes a useful empirical linkage between adoption of IPM systems along a continuum and changes in average expected pesticide use levels and estimates of risk outcomes.
1. The Integrated Weed Management Continuum
In general the four zones of the integrated weed management (IWM) continuum can be described as --
Field Level Database The 1994 USDA cropping practices and chemical use surveys include some 4,713 samples of corn producers representing 62.5 million planted acres and 3,024 soybean producers covering 43.75 million acres (both over 90 percent of total planted acreage) (Padgitt). A statistical bulletin drawing on the cropping practices survey database has recently been released summarizing pesticide use and pest management practices on major field crops (Natural Resources and Environment Division, Economic Research Service, AREI Update Number 19).
The computer disc data product containing the 1994 cropping practices survey data was obtained and used in this study. Individual farm samples were divided into four zones in accord with values for the variable "IPM Ratio". IPM Ratio reflects the degree of reliance on herbicide dose-adjusted acre-treatments in contrast to prevention-based practices. It is calculated using two other variables: proxy dose-equivalent acre-treatments (PDE) divided by the number of preventive IWM practices (IWM Practices). The distribution of the variable IPM ratio was used to select ranges corresponding to each of the four zones along the IWM continuum (method described below). In all tables that follow, results are reported for all samples, as well as for each of the four zones along the IWM continuum, and for the "Not Classified" category (explained below).
2. Calculating the Variable Proxy Dose-Equivalent Acre-treatments
Because of the great differences in herbicide rates of application and the trend toward rising numbers of active ingredients applied on a given acre, volume based measures of herbicide use are a deficient indicator of reliance on herbicides. A better measure is the number of dose-adjusted acre-treatments made with distinct active ingredients. In order to estimate dose-adjusted acre- treatments across different farms applying different rates though, some way is needed to determine what constitutes a single "normal" dose, so that fractions of this dose rate can be calculated on each farm as a function of actual rates of application. Such a method is essential to encompass the many ways farmers are reducing use and reliance on herbicides – reducing rates of application, switching to low-dose products, increasing the number of different active ingredients applied (often through tank mixes), and splitting applications to better match the timing of application with weed populations and vulnerable periods.
One option in calculating PDE acre-treatments would be to set dose rates at full label rates, or some percent of full label rates. This option was rejected because of the degree to which label rates, or some arbitrary fraction of them applied to all products, would skew the analysis. There is a wide range across different active ingredients between minimum and maximum label rates, and often large differences between actual average rates of application and label rates. Variance in these ranges is a function of families of chemistry, formulations, soil characteristics, climate, and cropping system factors. Some herbicides need to be applied at much higher rates in high organic matter soils (or much lower rates in sandy soils), or face water quality-motivated restrictions in some states and regions. As a result, labels might call for rather high or very low rates for special circumstances that do not reflect typical conditions and which vary markedly from "normal" average rates of application.
For active ingredients with high maximum label rates relative to commonly applied rates, use of the label rate as the basis for estimating PDE acre-treatments would led to inappropriately low PDE values relative to other active ingredients with a narrower range between minimum and maximum label rates and actual average rates of application. Examples include glyphosate, trifluralin, cyanazine, and 2,4-D. Conversely, for several low-dose herbicides and most very-low dose herbicides, there is relatively little difference between maximum label rates and actual average rates of application. The full label rate for the most widely used soybean herbicide in 1994 – imazethapyr (Pursuit) – is 0.063 pond a.i. per acre and the average rate of application was 0.6 pounds a.i. per acre; for imazaquin (Scepter), 0.125 full label rate, 0.10 actual average rate; quizalofop-ethyl (Assure), 0.068 full label rate, actual average rate 0.05.
Lacking an objective way to pick a label rate, or fraction thereof from the many possibilities, other options were explored to estimate a typical dose rate for use in calculating dose- adjusted acre-treatments. A method was developed to calculate a proxy dose that falls between the actual mean rate and the full labeled rate. The proxy dose for each active ingredient is calculated as mean rate of application in pounds a.i. per acre (across one or more applications) plus two standard deviations in the rate of application. For example, the mean rate of application of atrazine on corn in 1994 was 1.07 pounds a.i. per acre; the full proxy dose was 1.88. The variable proxy dose equivalent acre-treatments for atrazine is the ratio of the actual rate of application to the proxy: PDE acre-treatments equals 0.57 [1.07 pounds a.i. per acre divided by 1.88 pounds a.i. per acre].
In the case of some low-dose herbicides, the calculated proxy dose exceeds the maximum labeled rate of application (see Table 3 for examples). Assuming all applications are made at or below maximum label rates, proxy doses in excess of label rates would seemingly skew the results, since it would under-estimate dose adjusted acre-treatments with these products. A further review of the data, however, shows that for these products, the proxy dose is still lower than the maximum rate applied.
For example, the full label rate of fenoxaprop-ethyl (Acclaim) is 0.133 pounds a.i. per acre (Meister); the average rate of application on soybean acres in 1994 was 0.10; the proxy dose was 0.18, but the maximum rate applied was 0.25 pounds a.i. per acre. The case of the very-low dose herbicide thifensulfuron (Pinnacle) is even more striking. This product's label calls for application of 0.0039 pounds a.i. per acre on soybeans. Over 6 million acres were treated at an average rate of 0.0031 pounds a.i.; the proxy dose (the mean rate plus two standard deviations) was calculated as 0.01 pounds, over twice the full label rate, but only half the maximum rate applied – 0.20 pounds a.i. per acre. The maximum rate of application exceeds the full label rate more than six-fold. Clearly re-sprays can not explain the full difference between label rates and maximum rates of application. With the analysis completed to date, it is not possible to determine what portion of acreage with application rates over full label rates reflects re-treatments versus over-applications. This is clearly an issue which warrants further analysis since the consequences of over-application of low-dose products can be severe, both for farmers (cost, phyto-toxicity and carry-over) and the environment (adverse impacts on soil microbial activity and P uptake; aquatic ecosystems; non- target species of plants and wildlife).
For a given sample, the value of the variable PDE acre-treatments used in calculating IPM Ratio is the sum of proxy dose equivalent acre-treatment values across all active ingredients applied on a given sample point, and across all applications reported for that sample. USDA enumerators take considerable care in obtaining accurate information on exactly what formulated products and tank-mixes were applied, and at what rates per acre, so that the products applied can be converted accurately to pounds a.i. applied per acre for each distinct active ingredients.
On average across all corn samples, there were 2.55 different herbicide active ingredients applied on each acre. Most were applied well below label rates and on many farmers, significantly reduced rates were applied, as evident in the value of PDE acre-treatments, 1.44. If full label rates were used to calculate dose equivalents instead of the proxy dose rates described above, the number would be much lower, on the order of 0.7-0.8. Tables 1-5 report herbicide use patterns in 1994, including acres treated, average rates of application, proxy doses and the PDE acre-treatments for each major active ingredient, and contrasts average rates of application to full label rates: tables 1-3 cover high, medium and low-rate soybean herbicides and tables 4-5 high and medium rate corn herbicides. Full label rates are taken from the Weed Control Manual: Volume 30, 1996 (Meister).
3. Calculating Values for the Variable IPM Practices
A second key variable, IPM Practices, reflects the extent of prevention-oriented practices reported at a given sample. A simple point system is used to assign and then sum points for rotations and cover crops, cultivation for weed control, and scouting. Clearly, additional preventive practices and strategies were no doubt in place on several farms and should appropriately be taken into account in calculating values for this variable, but the structure of the variable IPM Practices and its component parts, are limited to those IWM practices and strategies that USDA covered in its cropping practices survey instrument.
Points were assigned as follows –
* Cultivation: 1 point for each post-plant pass with a tillage tool for weed control
Rotation Points Tables 6 and 7 present the distribution of soybean and corn planted acres according to three-year rotation. Rotation points reflect land use in the fall and spring/summer of 1993 and 1992. In the case of soybeans, over 85 percent of planted acres was either in S-C-S-C (or other row crop) rotations, continuous soybeans, or double-cropped soybeans. Just under 15 percent of the planted acres were in a rotation with potentially significant weed management benefits. Essentially the same patterns apply to acreage planted to corn in 1994.
In order to assess the importance of just rotations on herbicide use patterns, herbicide use patterns were calculated for all acreage in a three or more year rotation with at least one year of small grain, hay or pasture, or fallow. In both corn and soybeans, farmers using rotations are comparably reliant on herbicides as farmers using other rotations. In soybeans, the average acre of soybeans in a three or more year rotation were treated with an average of 1.29 pounds a.i. per acre, compared to 1.11 on land in row-crop only rotations, and 1.14 on all land. In corn, rotated land received on average 2.45 pounds of herbicide a.i. per acre, compared to 2.86 on row-crop land and 2.78 on all land. Across all other indicators of reliance on herbicides and adoption of IPM practices, only modest differences are evident on land in three year rotations compared to land in continuous row-crop rotations.
Findings reported herein on major differences in herbicide reliance and use across the four zones of the IWM continuum, coupled with surprisingly modest differences as a function of rotations alone, lends support to the hypothesis that farmers who are successfully reducing reliance on herbicides are doing so through adoption of multi-tactic IWM systems. Despite the compelling experimental evidence of the importance of rotations in reducing the need for herbicides (Wyse; Liebman and Dyck; NRC, 1989; Zoschke; Buhler et al; Crookston et al), the results presented herein show clearly that rotations are a necessary, but not sufficient condition to significantly reduce reliance on herbicides.
Cultivation Points Table 8 and 9 present the distribution of planted acres according to prevention-oriented practices as measured in the variable IPM Practices. These tables include the distribution of acreage for soybeans and corn as a function of the number of times cultivated for weed management, both for all acreage and in the four zones along the integrated weed management continuum.
Note that less than 17 percent of soybean planted acres is cultivated two or more time, and another 26.8 percent is cultivated one time. Cultivation is relied upon somewhat more heavily by corn growers: 47.9 percent of planted acreage cultivated once, and 16.2 percent cultivated twice or more. Over 56 percent of soybean acreage and over 36 percent of corn acres were not cultivated.
Clearly many growers are not fully utilizing the potential of cultivation to lessen weed, insect and plant pathogen pressure in the root-zone. It is equally clear that cultivation plays a key role in progress toward more prevention-oriented IWM systems. While less than 4 percent of the soybean fields in the "No" and "Low" zones were cultivated twice or more (4.4 percent in case of corn), almost 48 percent of soybean fields in the "High" zone were cultivated twice or more (over 44 percent in case of corn fields in "High" zone). Over a quarter of both corn and soybean fields in the "Medium" zone cultivated twice or more.
Scouting Tables 8 and 9 also show the distribution of acres according to whether fields were scouted: by whom, whether a systematic pattern was used, and average scouting fee per acre. Scouting by farmers, family-members and employees accounts for about 70 percent of all scouting. Note that corn producers in the "Medium" zone along the IWM continuum are almost 4 times more likely to use the services of a paid crop consultant than producers in the "Low" IWM zone. While 3.56 percent of corn farmers in the "No" IWM zone used a commercial scouting service, they paid about one-third the per acre fee of farmers in the "Low" zone, suggesting a narrower range of services provided – most likely focusing on fertility management.
In producing soybeans, farmers are clearly less inclined to seek crop consultant advice in moving along the IWM continuum – the percent of farmers paying a fee for consulting services, while low overall, was higher in the "No" IWM zone than in the "High" zone. One possible explanation is some farmers heavily reliant on herbicides are seeking outside technical assistance to avoid crop injury and control failures, given that margins for error are sometimes very thin with new low-dose, post-emerge weed management systems, especially relative to herbicide-dependent systems in the past.
4. Delineating Four Zones Along the IWM Continuum
For each sample, a value was calculated for the variable IPM ratio. Frequency distributions were then produced, showing the extent of acreage in 10 percent increments of total planted soybean or corn acreage, by average values of IPM ratio. Figures 1 and 2 (not included) show these distributions. Based on the distribution of the data, distinct ranges were selected for the four zones along the IWM continuum. For soybeans, the "No" IPM zone had values greater than 1.4; "Low," less than 1.4 but greater than .95; "Medium," less than .95 but greater than .15; and "High," less than 0.15. In the case of corn, the "No" zone had values greater than 1.05; "Low," less than 1.05 and greater than .65; "Medium,", less than .15 and greater than .65; and, "High,", less than .15.
Planted acreage sample points were then assigned to the four zones along the IWM continuum as a function of IPM Ratio values. Any sample point with less than 0.25 PDE acre- treatments per acre was assigned to the "High" zone, assuming yields were not markedly lower than average. Any sample reflecting low yields (more than one standard deviation below the mean yield in the state) was placed in the category "Not classified". This was done to eliminate from the sample farms and regions with unusual weather, under poor management, or otherwise affected by some problem that could lead to spurious values. A review of the values in the column "Not Classified" suggests that most farms with low yields are more nearly similar to those in the "Low" and "No" IWM zones than the two other zones along the IWM continuum.
Tables 10 and 11 present the distribution of acres planted to soybeans and corn across the four zones of the IWM continuum, and in the "Not classified" category (samples with low yields) for all states in the 1994 cropping practices survey (not including soybeans in Delaware because of inadequate sample size).
IWM Adoption The results show that 6 percent and 7 percent of soybean and corn planted acreage is managed under a bio-intensive IWM system, while 58 percent and 48 percent is managed in the "No" and "Low" IWM zones. In contrast in its 1994 report on IPM adoption, USDA estimated that 36 percent and 28 percent of soybean and corn acreage was managed under "High" level IPM, another 21 percent and 23 percent under "Medium" level IPM, and 41 percent and 47 percent under "Low" and "No" (only 2 percent of each crop was judged to fall in the "Low" category.
Differences between USDA and BCS estimates of adoption by zones reflect markedly different methodologies. USDA based its estimates on a simple decision-rule -- acres that were scouted and on which herbicide applications were based on a threshold were considered under "Low" level IPM; acres under "Medium" level IPM were also treated with 1 of 7 additional practice (rotations; cultivation; alternating herbicide a.i.'s; post-emergent applications only; banding; reduced rates; and weed spot treatments) and 2 additional practices were required in the case of "High" level IPM. But most of the practices used by USDA to judge level of IPM adoption are appropriate only for assessing whether herbicides have been relied upon cost- effectively in the context of chemical-intensive IPM. According to USDA, 5 out of the 7 weed management practices "deal with pesticide management rather than weed management." (page 23, Vanderman et al).
The methodology described herein reflects a more ecologically-grounded notion of IPM. Progress along the continuum occurs only when herbicide dose adjusted acre-treatments are reduced, and/or when practices are adopted with documented potential to lessen weed pressure or yield loses through non-chemical or reduced chemical approaches. The methodology also differences significantly from USDA's in that it implicitly accounts for the interactions and linkages between reliance on herbicides, the number and extent of preventative practices needed, and weed pressure. In USDA's system, the adoption of practices alone determines the zone along the IPM continuum, irrespective of pest pressure or herbicide use. Here, the interaction of the variables PDE and IPM Practices determines where a given field falls along the continuum. Farmers facing high levels of pest pressure will generally rely more heavily on pesticides, but still might be operating in the bio-intensive zone of the IPM continuum if multiple preventative practices are also in place.
In applying this methodology in a given production region or watershed, or to a set of experimental farming system studies, farmers, scientists and pest management specialists will need to collectively determine the best way to calculate dose adjusted acre-treatments. The method suggested herein is one of several plausible ways. They will also need to review and revise the list of practices accorded points toward the variable IPM Practices, while also adjusting the number of points assigned to a given practice, reflecting the relative importance of a practice as a preventative weed management intervention. The variable IPM Practices in this analysis is clearly constrained by the availability of data in the cropping practices survey. Field level studies should appropriately encompass a broader range of practices. Last, in regional and field studies, analysts will be able to draw on more detailed data and experience in determining the ranges in the variable IPM ratio corresponding to the four zones along the IPM continuum.
C. WEEED MANAGEMENT AND HERBICIDE USE PATTERNS ALONG
THE IWM CONTINUUM
The USDA cropping practices survey includes detailed questions on rotations, tillage systems, planting methods and rates, fertility programs and pesticide applications. Further insight into the differences across farming systems along the IWM continuum are accessible through a review of data on various practices and system components.
1. Tillage Practices
Tables 12 and 13 report data on pre-plant tillage systems. As expected, farmers in the "High" zone along the IWM continuum were more likely to use conventional tillage and much less likely to use no-till.
2. Reliance on Herbicides
Tables 14 and 15 present a basic set of indicators of reliance on herbicides. The average rate of application on soybean acreage in the "High" zone is 0.99 pounds a.i. per acre, 36 percent less than in the "No" IWM zone. The reduction in average rates on corn farms was even greater -- 51 percent. Likewise, all other measures of reliance on herbicides decline markedly in the "Medium" and "High" zones, including the proxy dose equivalent acre-treatments which show most dramatically the lessened level of reliance on herbicides on farms adopting IWM systems with reduced rates and/or several prevention-oriented practices. On corn fields, dose adjusted acre-treatments are less than half on fields in the "High" zone in contrast to the "No" zone. The reduction is slightly less dramatic in the case of soybeans – from 2.22 to 1.12 dose adjusted acre- treatments.
It is interesting to note that on both corn and soybean fields, by far the largest percentage reductions in herbicide dose adjusted acre-treatments and pounds applied per acre occur when farmers move from the "No" IWM zone into the "Low" zone, and then from the "Low" zone to the "Medium" zone. In fact, there are only modest differences between herbicide reliance and use patterns on fields in the "High" and "Medium" zones. The policy implication is obvious -- to the extent pesticide use and risk reduction are primary goals in the setting and implementation of pest management policies, effort should focus initially on over-coming barriers to progress along the IWM continuum in the "No" and "Low" zones.
Equally important, it is clear that much progress can be made without the need for farmers to adopt three year or more rotations including one or more year in a crop other than corn, soybeans, or another row crop. About 44 percent of corn farmers have moved into the "Medium" or "High" zones of the IWM spectrum. In contrast to corn producers still not using IWM, these farmers have cut herbicide reliance and use in half. Comparable reductions are within each of the other 56 percent of corn producers, suggesting that over all herbicide use could be markedly reduced in the next decade without any significant technological breakthroughs or changes in policy. The key appears to be adoption of multi-tactic IWM systems that include both steps to lessen weed pressure and deal with weeds through cultivation, as well steps to reduce per acre herbicide application rates through banding, spot spraying and other practices.
Data in Tables 14 and 15 provide a method to estimate the extent of acreage on which no herbicides were applied. In the case of soybeans, about 662,000 acres -- or about 1.5 percent of total soybean acres were not treated with a herbicide; all this acreage falls within the "High" IWM zone. In the case of corn, about 1.04 million acres were not treated with a herbicide, or about 1.7 percent of total corn acreage.
3. Timing and Methods of Application
Significant differences also exist in the timing and methods of application, as evident in Tables 16 and 17. Producers less dependent on herbicides are more likely to apply herbicides in bands, or make directed or spot spray applications. Note that over 22 percent of soybean herbicides were applied in a band on fields in the "Medium" and "High" zones, compared to just 3.2 percent of "No" zone soybean fields and 1.7 percent of corn fields.
Broadcast applications are by far the dominate choice on fields in the "No" and "low" zones, while about 42 percent of the applications of herbicides were broadcast on soybean fields in the "High" IWM zone.
BIBLIOGRAPHY
Angers, D. A., A. N'dayegamiye, and D. Cote. 1993. Tillage-induced differences in organic matter of particle-size fractions and microbial biomass. Soil Science Society of America Journal 57(March-April): 512-516.
Ashton, F. M. and A. S. Crafts. 1981. Mode of Action of Herbicides, 2nd ed. New York: John Wiley & Sons.
Benbrook, C. M. 1996. Reliance on Herbicides and Adoption of Integrated Weed Management Systems in the Midwest. World Wildlife Fund [in press].
Benbrook, C. M. and D. J. Marquart. 1993. Challenge and Change: A Progressive Approach to Pesticide Regulation in California. Sacramento, Calif.: California Environmental Protection Agency.
Bezdicek, D. F., and D. Granatstein. 1989. Crop rotation efficiencies and biological diversity in farming systems. American Journal of Alternative Agriculture 4(3-4):111-119.
Bohlen, P.J., and C. A. Edwards. 1995. Earthworm effects on N dynamics and soil respiration in microcosms receiving organic and inorganic nutrients. Soil Biol. Biochem. 27:341-348.
Bottrell, D. G., and R. R. Weil. 1995. Protecting crops and the environment: Striving for Durability. Pp. 55-73 in Agriculture and the Environment: Bridging Food Production and Environmental Protection in Developing Countries, A. S. R. Juo, and R. D. Freed, eds. Madison, Wisc.: American Society of Agronomy, Inc., Crop Science Society of America, Inc., and Soil Science Society of America, Inc.
Buhler, D. D., J. L. Gunsolus, and D. F. Ralston. 1992. Integrated weed management techniques to reduce herbicide inputs in soybean. Agronomy Journal 84: 973-978.
Cate, J.R. and M.K. Hinkle. 1993. Integrated Pest Management: The Path of a Paradigm. Washington, D.C.: National Audubon Society.
Coble, H. D., and D. A. Mortensen. 1992. The threshold concept and its application to weed science. Weed Technology 61(1): 191-195.
Cohen, B., R. Wiles, E. Bondoc. 1995. Weed Killers by the Glass: A Citizen's Tap Water Monitoring Project in 29 Cities. Washington, D.C.: Environmental Working Group.
Cook, R. J., and R. J. Veseth. 1991. Wheat Health Management. St. Paul, Minn.: American Phytopathological Society.
Crookston, R. K., J. E. Kurle, P. J. Copeland, J. H. Ford, and W. E. Lueschen. 1991. Rotational cropping sequence affects yield of corn and soybean. Agronomy Journal 8:108-113.
Curl, E. A. 1982. The rhizosphere: Relation to pathogen behavior and root disease. Plant Disease 66(7): 624-630.
Dushoff, J., B. Caldwell, and C. L. Mohler. 1994. Evaluating the environmental effect of pesticides: A critique of the environmental impact quotient. American Entomologist (Fall): 180- 184.
Forcella, F., R.G. Wilson, K.A. Renner, J. Dekker, R.G. Harvey, D.A. Alm, D.D. Buhler, and J. Cardina. 1992. Weed Seedbanks of the U.S. Corn Belt: Magnitude, Variation, Emergence, and Application. Weed Science, Volume 40:636-644.
Ghersa, C.M., M.L. Roush, S.R. Radosevich, and S.M. Cordray. 1994. Co-evolution of Agroecosystems and Weed Management: Weed-management practices have become closely linked to social and economic, rather than biological factors. BioScience Vol. 44, No.2: 85-94.
Gianessi, L. P. 1992. U. S. Pesticide Use Trends: 1966-1989. Washington, D.C.: Resources for the Future.
Gianessi, L. P. and J. E. Anderson. 1995. Pesticide Use in U.S. Crop Production: National Data Report. Washington, D.C.: National Center for Food and Agricultural Policy.
Goolsby, D.A., W. A. Battaglin, and E. M. Thurman. Occurrence and Transport of Agricultural Chemicals in the Mississippi River Basin, July through August, 1993. Circular No. 1120-C. Denver, Col.: U. S. Geological Survey.
Grier, N., E. Clough, and A. Clewell. 1994. Toxic Water: A Report on the Adverse Effects of Pesticides on the Pacific Coho Salmon and the Prevalence of Pesticides in Coho Habitat. Eugene, OR: Northwest Coalition for Alternatives to Pesticides.
Hatfield, J.L. and D.L. Karlen. 1994. Sustainable Agriculture Systems. Ann Arbor, Michigan: Lewis Publishers.
Higley, L. G., and W. K. Wintersteen. 1992. A novel approach to environmental risk assessment of pesticides as a basis for incorporating environmental costs into economic injury levels. American Entomologist (Spring):34-39.
Hoppin, P., Lirhoff, R., and Michelle Miller. 1996. Reducing Reliance on Pesticides in Great Lakes Basin Agriculture. Washington, D.C. World Wildlife Fund [in press].
Jawson, M. D., A. J. Franzluebbers, D. K. Galusha, and R. M Aiken. 1993. Soil Fumigation within monoculture and rotations: Response of corn and mycorrhizae. Agronomy Journal 851174-1180.
Jordan, N. 1996. Weed Prevention: Priority Research for Alternative Weed Management. Journal of Production Agriculture [in press].
Kirchner, M. J., A. G. Wollum II, and L. D. King. 1993. Soil microbial populations and activities in reduced chemical input agroecosystems. Soil Science Society of America Journal 57: 1289-1295.
Kovach, J., C. Petzoldt, J. Degnil, and J. Tette. 1992. A method to measure the environmental impacts of pesticides. New York's Food and Live Sciences Bulletin 139:1-4.
Landy, D. 1995(May). Multiattribute Ranking Systems for Pesticides, Masters of Science Thesis. Berkeley, Calif.: Energy and Resources Department, University of California, Berkeley.
Leng, M. L., E. M. K. Leovey, and P. L. Zubkoff, eds. 1995. Agrochemical Environmental Fate: State of the Art. New York: Lewis Publishers.
Liebman, M., and E. Dyck. 1993. Weed management: A need to develop ecological approaches. Ecological Applications 3(1):39-41.
Liebman, M. and R. Janke. 1990. Sustainable Weed Management Practices. Sustainable Agriculture in Temperature Zones. John Wiley, N.Y.: 111-143.
Lin, B.H., M. Padgitt, L. Bull, H. Delvo, D. Shank, and H. Taylor. 1995. Pesticide and Fertilizer Use and Trends in U.S. Agriculture. Agricultural Economic Report No. 717. Economic Research Service. Washington, D.C.: U. S. Department of Agriculture.
Meister Publishing Company. 1996. Weed Control Manual: Volume 30. Wiloughby, OH.
Mortensen, D.A., A.R. Martin, T.E. Harvill, and T.A. Bauer. 1993. The influence of rotation diversity on economic optimum thresholds in soybean. In 8th EWRS Symposium: Quantitative Approaches in Weed and Herbicide Research and Their Practical Application. pp. 815-824.
National Research Council. 1986. Pesticide Resistance: Strategies and Tactics for Management. Washington, D.C.: National Academy Press.
National Research Council. 1989. Alternative Agriculture. Washington, D.C.: National Academy Press. 448 pp.
National Research Council. 1993. Soil and Water Quality: An Agenda for Agriculture. Washington, D.C.: National Academy Press.
National Research Council. 1995. Ecologically Based Pest Management: New Solutions for a New Century. Washington, D.C. National Academy Press.
Osteen, C. D., and P. I. Szmedra. 1989 (September). Agricultural Pesticide Use Trends and Policy Issues. Agricultural Economic Report No. 622. Economic Research Service. Washington, D. C.: U. S. Department of Agriculture.
Padgitt, M. 1995. 1994 Cropping Practices Survey: Data File Specifications and Documentation. Resource Information System Section, Resources and Technology Division, Economic Research Service, U.S. Department of Agriculture, Washington, D.C., and accompanying data product.
Pease, W. S., R. A. Morello-Frosch, D.S. Albright, A. D. Kyle, and J. C. Robinson. 1993. Preventing Pesticide-related Illness in California Agriculture: Strategies and Priorities. Berkeley, Calif.: California Policy Seminar.
Pedigo, L. P. and L. G. Higley. 1992. The Economic injury level concept and environmental quality: A new perspective. American Entomologist: (September).
Pimentel, D., J. Acquay, M. Biltonen, P. Rice, M. Silva, J. Nelson, V. Lipner, S. Giordano, A. Horowitz, and M. D'Amore. 1992. Environmental and economic costs of pesticide use. BioScience 42(10): 750-760.
Reganold, J. P. 1995. Soil quality and profitability of biodynamic and conventional farming systems: A review. American Journal of Alternative Agriculture 10(1): 36-45.
Reuveni, R. (ed.) 1995. Novel Approaches to Integrated Pest Management. Ann Arbor, MI: Lewis Publishers.
Ribaudo, M. O., and A. Bouzaher. 1994 (September). Atrazine: Environmental Characteristics and Economics of Management. Agricultural Economic Report No. 699. Economic Research Service. Washington, D. C.: U. S. Department of Agriculture.
Rothrock, C. S., T. L. Kirkpatrick, R. E. Frans, and H. D. Scott. 1995. The influence of winter legume cover crops on soilborne plant pathogens and cotton seedling dieseases. Plant Disease 79(2): 167-171.
Sine, C., ed. Updated annually. Farm Chemicals Handbook. Willoughby, Oh.: Meister
Swanton, C.J. and S.D. Murphy. 1994. Weed science beyond the weeds: The role of integrated weed management (IWM) In agroecosystem health. In Press, Journal of Production Agriculture.
U. S. Department of Agriculture. 1994a. Agricultural Resources and Environmental Indicators. Agricultural Handbook No. 705. Economic Research Service. Washington, D.C.
U. S. Department of Agriculture. 1994b. Cropping Practices Survey 1993, Tables in Data Product. Economic Research Service. Washington, D.C.
U. S. Department of Agriculture. 1995 (March). Agricultural Chemical Usage: 1994 Field Crops Summary. Economic Research Service. Washington, D.C.
U. S. Department of Agriculture. 1995 (May). 1994 Cropping Practices Survey: Data File Specifications and Documentation. Economic Research Service. Washington, D. C.
U.S. Department of Agriculture. 1995. AREI Updates (No. 19): Pest Management on Field Crops. Economic Research Service. Washington, D.C.
University of California. 1992. Beyond Pesticides: Biological Approaches to Pest Management in California. Oakland, Calif.: Division of Agriculture and Natural Resources.
Vanderman, A., J. Fernandez-Cornejo, S. Jans, and B.H. Lin. 1994. Adoption of Integrated Pest Management in U.S. Agriculture. Agricultural Information Bulletin No. 707. Economic Research Service. Washington, D.C.: U. S. Department of Agriculture.
Walker, K., J. Liebman, and W. Pease. 1995. Pesticide-induced Disruptions of Agricultural Ecosystems. Berkeley, Calif.: California Policy Seminar.
Wauchope, R. D., T. M. Butler, A. G. Hornsby, P. W. M. Augustijn-Beckers, and J. P. Burt. 1992. The SCS/ARD/CES pesticides properties database for environmental decision-making. Rev. Environ. Contam. Toxicol. 123:1-164.
Wiles, R., B. Cohen, C. Campbell, and S. Elderkin. 1994. Tap Water Blues: Herbicides in Drinking Water. Washington, D.C.: Environmental Working Group and Physicians for Social Responsibility.
World Wide Fund for Nature. 1995. Pesticide Reduction: Economic Instruments. Gland, Switzerland.
Wyse, D. L. 1994. New technologies and approaches for weed management in sustainable agriculture systems. Weed Technology 8(2): 403-407.
Zimdahl, R.L. 1991 (March). Weed Science: A Plea for Thought. Washington, D. C.: U. S. Department of Agriculture, and Cooperative State Research Service.
Zoschke, A. 1994. Toward reduced herbicide rates and adapted weed management. Weed Technology 8(2): 376-386.
