| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
a Dep. of Environmental Biology, Univ. of Guelph, Guelph, ON, Canada N1G 2W1
b IPM Program, Selby Hall, Ohio State Univ., Wooster, Ohio, 44691
Corresponding author (oaft{at}sentex.net)
Received for publication June 28, 1999.
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
Abbreviations: EI, Environmental Impact • EIQ, Environmental Impact Quotient • IPM, integrated pest management
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
Food Systems 2002 used the amount of active ingredient as the measure of pesticide use. This measure was chosen because a census of pesticide use by kilogram of active ingredient has been conducted in the Province of Ontario every five years since 1973 (Hunter and McGee, 1994, 1999; McGee, 1984; Moxley, 1989; Roller, 1975, 1979). The amount of active ingredient used in 1983, the last year of data available in 1987, was used as the baseline to measure the success of Food System 2002 in meeting its objective of a 50% reduction in pesticide use.
A reduction in pesticide use can be achieved by various means and may occur against a rapidly changing background. Thus, the reduction in pesticide use needs to be placed within context. Most government programs, such as Food Systems 2002, describe changes in a historical context. This type of reduction can be misleading. Pesticide use in agriculture could be reduced by 50% by eliminating 50% of the area farmed and importing the food produced from other jurisdictions, thereby transferring the pesticide use. Alternatively, pesticide use could be reduced by 50% by reducing the number of applications and/or the amount applied in each application. However, without the introduction of other methods of pest control this practice could result in a decrease in productivity. How would we rate the success of a program in which total pesticide use did not change, but productivity doubled and pesticide use per tonne of output declined 50%? Most farm organizations in the Province of Ontario felt that agricultural productivity should be enhanced to feed an expanding population and remain globally competitive. The human population of the Province of Ontario increased by approximately 30% from 1983 to 1998 while the area farmed decreased by 7.5% (Ontario Ministry of Agriculture and Food, 1984; Ontario Ministry of Agriculture, Food and Rural Affairs, 1999; Statistics Canada, 2000c).
Cropping patterns may also have a large effect on pesticide use. Corn and soybean accounted for 47% of the pesticide use and 25% of the crop area in the Province of Ontario in 1983 (McGee, 1984), while tobacco accounted for 20.5% of the pesticide use and less than 1% of the crop area. Hay and pasture accounted for 41% of the crop area, but less than 0.5% of the pesticide use. Because of the variation in the mean application rate among crops, changing cropping patterns over time may cause changes in pesticide use independent of other factors. Also, because of the large area, a small reduction in the pesticide use on corn would have the same effect as a large reduction in pesticide use on tobacco. However, the value per hectare of tobacco in 1983 was approximately 10 times the value per hectare of corn (Ontario Ministry of Agriculture and Food, 1984). Does the higher value of tobacco justify the greater pesticide use?
Finally, quantitative reductions in pesticide use measured solely in terms of kilograms of active ingredient have a major problem. The risk to nontarget organisms (i.e., farmworkers, consumers, birds, bees, fish, etc.) varies widely among pesticides. A common criticism is that the reduction in the amount of pesticide used results from the replacement of "high volume–low risk" pesticides with "low volume–high risk" pesticides, and there is little or no reduction in the environmental risks associated with pesticide use. Most would agree that the objective of any program reducing pesticide use should be to reduce the risk to nontarget organisms, and that reductions in the amount of pesticide used should be accompanied by equivalent or greater reductions in the environmental risk. However, as with changes in pesticide use, changes in risk need to be placed in context. In this paper we describe the changes in pesticide use and risk in the Province of Ontario from 1973 to 1998 in the historical context, and in the context of changes in area, yield, and unit crop value.
|
METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
Pesticide risk is the probability of an adverse outcome related to the exposure to that pesiticide. It is a function of the toxicity of the pesticide and the level of exposure. In discussing environmental pesticide risk an immediate question arises, "risk to whom?" Obviously the risk is to nontarget organisms but this encompasses a broad range, including applicators, fieldworkers, consumers, birds, bees, fish, beneficial insects, etc. The potential adverse outcomes and risk of individual pesticides to each of these groups may also vary widely. In attempting to assess the changes in risk associated with the agricultural use of pesticides in the Province of Ontario from 1973 to 1998, the problem was defining the risk to nontarget organisms posed by more than 180 pesticides used during a 25-yr period. We chose the Environmental Impact Quotient (EIQ), a scoring system for pesticide risk developed by the Integrated Pest Management Program at Cornell University (Kovach et al., 1992), as a measure of the risk of individual pesticides.
The EIQ scores the potential risk for a pesticide based on measures of toxicity such as the LD50 (dose at which 50% of the treatment group dies within the specified time period) or LC50 (concentration at which 50% of the treatment group dies within the specified time period), and measures of potential exposure such as the half-life, runoff or leaching potential, and pattern of use. The toxicity and indices of exposure are scored on a 1, 3, 5 scale. The overall EIQ for each pesticide is the average of three general risk categories, the farmworker, the consumer, and the ecological component (Fig. 1). The farmworker category includes potential effects to applicators and fieldworkers; the consumer category includes the potential effects of residues on the consumer and of ground water contamination; and the ecological category includes the potential effects on aquatic organisms, bees, birds, and beneficial arthropods.
|
The risk associated with pesticide use was termed the Environmental Impact (EI), and was calculated by multiplying the amount of pesticide used on each crop in a given year by the EIQ, then summing the values such that:
 | [1] |
The total EI was summed across all crops. The changes in pesticide use and EI were then calculated as a percentage of the 1983 value, the baseline year for Food Systems 2002.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
|
|
|
|
|
|
|
Pesticide use on corn declined 48% from 1983 to 1998 because of a 17% decrease in area and a 37% decrease in the mean application rate, but corn still accounted for 40% of the total pesticide use in 1998. A reduction in the use of triazine herbicides accounted for 84.5% of the decrease in pesticide use on corn. Atrazine use declined 67%, from 1.72 x 106 kg in 1983 to 0.57 x 106 kg in 1998, and cyanazine use declined 89%, from 0.43 x 106 kg to 0.047 x 106 kg. Alachlor (2-chloro-2',6'-diethyl-N-(methoxymethyl)acetanilide), which accounted for 16% of the pesticide use on corn in 1983, was discontinued after 1988, and there was a 97.5% decrease in the use of butylate (S-ethyl di-isobutylthiocarbamate) from 1983 to 1998. These herbicides were replaced by metolachlor (2-chloro-N-(2-ethyl-6-methylphenyl)-N-(2-methoxy-1methylethyl) acetamide), dicamba (3,6-dichloro-2-methoxybenzoic acid), pendimethalin (N-(1-ethylpropyl)-3,4-dimethyl-2,6-dinitrobenzenamine), and rimsulfuron (N-[[(4,6-dimethoxy-2-pyrimidinyl)amino]carbonyl]-3-(ethylsulfonyl)-2-pyridinesulfonamide).
The EI for corn was less than the EI for tobacco in 1973, but corn had the highest EI in the other surveys. The EI for corn only increased 84.5% from 1973 to 1983, less than the increase in pesticide use, because the mean EIQ declined by 8%. In 1983 corn accounted for 41% of the total EI versus 47% of the total pesticide use because the mean EIQ for corn was below the mean EIQ for all crops. The 51.5% decrease in the EI for corn from 1983 to 1998 exceeded the decrease in pesticide use because of a 7% decrease in the mean EIQ. The partial replacement of atrazine, which has an EIQ of 33.2 (Kovach et al., 1992), by metolachlor with an EIQ of 18, and the 94.5% reduction in the use of insecticides that have high EIQ values (Kovach et al., 1992) contributed to the reduction in the mean EIQ. Even though insecticides accounted for <4% of the pesticides applied to corn in 1983, the reduced use of insecticides on corn accounted for 49% of the total reduction in insecticide use on agricultural crops from 1983 to 1998.
Soybean
There was 4.5-fold increase in the area of soybean from 1973 to 1998. Pesticide use increased from 0.36 x 106 kg in 1973 to 1.70 x 106 kg in 1988 when soybean replaced tobacco as the second-largest pesticide use after corn. Pesticide use on soybean declined to 1.14 x 106 kg in 1993, then increased to 1.29 x 106 kg in 1998. The increase in pesticide use on soybean from 1973 to 1983 resulted from 90% increases in the area and mean application rate. The mean application rate declined 57% from 1983 to 1998, and total pesticide use on soybean was similar in the two years despite a 133.5% increase in area. Herbicides accounted for 
99% of the pesticides applied to soybean.
The changes in EI from 1973 to 1998 were similar to the changes in pesticide use, and soybean had the second-highest EI from 1988 to 1998. The mean EIQ varied from 23.3 in 1973 to 28.0 in 1993. This reflected changes in the most common herbicide. Chloramben (3-amino-2,5-dichlorobenzoic acid), with an EIQ of 15.7, was the most common herbicide in 1973. It was replaced by alachlor (EIQ = 21.3) and metolachlor (EIQ = 18) in 1983. Alachlor was discontinued in 1988 and replaced by metolachlor and metribuzin (4-amino-6-(1,1-dimethylethyl)-3-(methylthio)-1,2,4-triazin-5(4H)-one), with an EIQ of 35.3. In 1998 metribuzin was replaced by glyphosate (N-(phosphonomethyl)glycine), with an EIQ of 32.4. Glyphosate use on soybean increased from 28.5 x 103 kg in 1983 to 164.8 x 103 kg in 1993 and 381.7 x 103 in 1998. This accounted for 61% of the total increase in the use of glyphosate on agricultural crops from 1983 to 1998, and 90% from 1993 to 1998.
Field Bean
Field bean covered 0.8 to 1.6% of the crop area. There were no obvious patterns in the changes in area and pesticide use among years. The area and pesticide use were relatively low in 1983. Pesticide use increased 275.5% from 1983 to 1988 because of a 67% increase in area and a 125% increase in the mean application rate. Both area and pesticide use declined from 1988 to 1998. The area in 1998 was only 5% less than in 1983, but pesticide use was 37% less because of a 34% decrease in the mean application rate. The EI mirrored the changes in pesticide use; however, the mean EIQ increased from 17.0 in 1973 to 25.8 in 1993, before declining to 24.4 in 1998. Herbicides accounted for 97% of the pesticide use. EPTC (S-ethyl dipropylthiocarbamate) and metobromuron (3-(4-bromophenyl)-1-methoxy-1-methylurea), which have EIQ scores <14 (Kovach et al., 1992), accounted for >80% of the herbicides used in 1973 and 1978. They were replaced by metolachlor, bentazon (3-isopropyl-1H-2,1,3-benzothiadiazin-4(3H)-one 2,2-dioxide), ethalfluralin (N-ethyl-N-(2-methyl-2-propenyl)-2,6-dinitro-4-(trifluromethyl)benzenamine), glyphosate, linuron (3-(3,4-dichlorophenyl)-1-methoxy-1-methylurea) and trifluralin (
,,-trifluoro-2,6-dinitro-N,N-dipropyl-p-toluidine)), which have much higher EIQ scores.
Grains
Grains, including rye (Secale cereale L.), winter and spring wheat (Triticum aestivum L.), oat (Avena sativa L.), barley (Hordeum vulgare L.), and mixed grains, covered the second-largest area after hay and pasture in 1973. The area decreased by 29% from 1973 to 1998. Pesticide use peaked in 1988, then declined. In 1998 it was 5% less than in 1983. Despite the large area, grains accounted for <8% of the pesticide used because of the low mean application rate. However, grains contributed a higher proportion of the total EI than expected on the basis of pesticide use. Grains had the highest mean EIQ (45.0) in 1978. The mean EIQ declined to the fourth highest (35.8) in 1998, but still exceeded the mean EIQ for all crops.
Herbicides accounted for >99% of the pesticides used on grain, with phenoxy herbicides, particularly 2,4-D (2,4-dichlorophenoxyacetic acid) and MCP (4-chloro-o-tolyoxyacetic acid), accounting for >95% of the pesticide used in 1973 and 1978. Phenoxy herbicides declined as a percentage of the pesticide used after 1978 because of the increased use of glyphosate and bromoxynil (3,5-dibromo-4-hydroxybenzonitrile and 2,6-dibromo-4-cyanophenyl octanoate). In 1998 phenoxy herbicides accounted for 61% of the pesticide used. The decline in the mean EIQ resulted from a shift from 2,4-D amine (EIQ = 56.3) to 2,4-D butyl (EIQ = 26.2) and the increased use of bromoxynil (EIQ = 23).
Hay and Pasture
Hay and pasture covered the largest area but had the lowest pesticide use from 1973 to 1983. The area declined 15.9% from 1973 to 1983 and 25% from 1983 to 1998. Pesticide use increased 275% from 1983 to 1998, from 24.9 x 103 kg to 93.4 x 103 kg. Nevertheless, hay and pasture still accounted for <2% of the pesticide applied to agricultural crops in 1998. The increased pesticide use on hay and pasture resulted from an increase in the percentage of the area treated and the increased use of glyphosate, which has a higher application rate than other products (Ontario Ministry of Agriculture, Food and Rural Affairs, 1998b). Pesticides were applied to <2% of the area of hay and pasture in 1983 and >4% in 1988, and glyphosate accounted for >99% of the increased pesticide use from 1983 to 1998. The EI for hay and pasture increased 360% from 1983 to 1998, exceeding the increase in pesticide use. The mean EIQ increased by 22.7% because 2,4-D amine replaced 2,4-D butyl.
Tobacco
Tobacco covered approximately 1% of the area farmed from 1973 to 1983, but had the second-highest pesticide use because the mean application rate was 9.5 to 16.7 times the mean application rate of corn. Pesticide use on tobacco declined 81.5% from 1983 to 1998, from 1.67 x 106 kg to 0.31 x 106 kg. The area declined by 33% and the mean application rate declined by 72%, but was still five times the mean application rate of corn. Tobacco consistently had the highest pesticide use per tonne of yield (Fig. 8), and despite the decrease in pesticide use from 1983 to 1998, the pesticide use per tonne in 1998 was still 2.7 times the pesticide use per tonne of fruits, which had the second-highest pesticide use per tonne.
|
The EI for tobacco declined 82.5% from 1983 to 1998 as the mean EIQ declined 6%. The reduction in the use of dichloropropenes (EIQ = 35.7) accounted for most of the reduction in the mean EIQ, but there was also a reduction in the use of insecticides with EIQ scores >50. The reduced use of dichloropropenes, which have a farmworker risk score of 78, accounted for 68% of the reduction in farmworker risk in the province and all of the reduction in the farmworker risk per kilogram of pesticide. However, because they are applied prior to planting and have a relatively short half-life, dichloropropenes only have an ecological risk score of 21.7. Their reduced use accounted for 90% of the increase in the ecological risk per kilogram of pesticide in the province.
Fruits
Fruits covered 0.7 to 0.8% of the crop area, but had the fourth highest pesticide use until 1988, and the third highest pesticide use in 1993 and 1998. Fruits had the second-highest mean application rate after tobacco from 1978 to 1988, and the highest mean application rate from 1993 to 1998 because of the decline in the mean application rate on tobacco. Fruits had the second-highest pesticide use per tonne of yield after tobacco except in 1978 and 1988, when grains had the second-highest use.
Pesticide use on fruits increased 80% from 1978 to 1993, from 0.35 x 106 kg to 0.63 x 106 kg, then declined to 0.54 x 106 kg in 1998. The increase in pesticide use from 1978 to 1983 resulted from a 61% increase in the mean application rate. From 1983 to 1993 both the area and mean application rate increased slightly. In 1998 the area was 6% less than in 1983, but the mean application rate was 3% higher.
In 1983 fungicides accounted for 73% of the pesticides applied to fruit and insecticides accounted for 25.5%. Fungicide use increased 12% from 1983 to 1998, when they accounted for 85% of the pesticide use. The amount of insecticide did not change significantly from 1978 to 1993, but declined 50% from 1993 to 1998, when they accounted for 13% of the pesticide use.
Apples (Malus domestica Borkh.), which covered 34% of the area in 1978 and 43% in 1998, accounted for approximately 50% of the pesticide use on fruits from 1978 to 1993, and 85% in 1998. Fungicide use on apples increased 35% from 1993 to 1998, but declined on all other fruit crops. Apples accounted for 56% of the insecticides used on fruits in 1993. Insecticide use on apples declined 36.5% from 1993 to 1998, but they still accounted for 71% of the insecticides used on fruit. The decreased insecticide use on most fruit crops was associated with a switch to newer insecticides such as imidacloprid (1-(6-chloro-3-pyridin-3-ylmethyl)-N-nitroimidazolidin-2-ylidenamine), pyridaben (2-tert-butyl-5-(4-tert-butylbenzylthio)-4-chloropyridazin-3(2H)-one), tebufenozide (N-tert-butyl-N'-(4-ethylbenzoyl)-3,5-dimethylbenzohydrazide), cyhalothrin-lambda (-cyano-3-phenoxybenzyl 3-(2-chloro-3,3,3-trifluoroprop-1-enyl)-2,2-dimethylcyclopropanecarboxylate), and other synthetic pyrethroids that are applied at lower rates than the older products (Ontario Ministry of Agriculture, Food and Rural Affairs, 1998a).
The EI for fruits increased from 1978 to 1988, then declined, but was still higher in 1998 than in 1983. The mean EIQ for fruits varied from 39.5 to 46.1, and fruits had the highest mean EIQ in 1983 and 1998. From 1983 to 1998 fruits had second-highest EI per tonne of yield because of the high mean EIQ. The year-to-year changes in the mean EIQ were caused by changes in the fungicides used. Mancozeb ([[1,2-ethanediylbis[carbamodithioato]](2-)] manganese mixture with [[1,2-ethanediylbis[carbamodithioato]](2-)]zinc), with an EIQ of 62.3, and metiram (tris[ammine-[ethylen bis(dithiocarbamato)]zinc(II)][tetrahydro-1,2,4,7-dithiadiazocine-3,8-dithione] polymer), with an EIQ of 55.9, accounted for 66% of the fungicide used in 1988 and 57% in 1998, the years with the highest mean EIQ. In the other years these two products accounted for less than 50% of the fungicide used and were replaced by captan (N-trichloromethylthio-4-cyclohexene-1,2-dicarboximide), with an EIQ of 28.6, sulfur (EIQ = 45.5), and folpet (N-[(trichloromethyl)thio]phthalimide), with an EIQ of 22.2.
Vegetables
Vegetables covered 1.7 to 1.8% of crop area. They ranked sixth in pesticide use from 1983 to 1993 and fifth in 1998. The mean application rate was the third highest among the crops from 1983 to 1998, but was only 23 to 28% of the mean application rate of fruits. Pesticide use on vegetables increased 14% from 1983 to 1993, from 0.35 x 106 kg to 0.41 x 106 kg, because of a 21% increase in the mean application rate. In 1998 pesticide use was 0.33 x 106 kg, 7% lower than in 1983 because of a 3.5% decrease in the mean application rate and a 4% decrease in area.
Fungicides and insecticides accounted for approximately 70% of the pesticides used on vegetables from 1978 to 1988, 82.5% in 1993, and 56.5% in 1998. In 1998 fungicide use was similar to the amount used in 1983, but insecticide use was 67% less. Potatoes (Solanum tuberosum L.) and tomatoes (Lycopersicon sp.) received 82% of the fungicides used on vegetables in 1978, 78% in 1983, 85% in 1988, 59% in 1993, and 85% in 1998 even though they only covered 35 to 45% of the crop area. They also accounted for 68.5% of the insecticides used on vegetables in 1978, 45% in 1983 and 1988, 57% in 1993 and 17% in 1998. Insecticide use on potatoes was 50% higher in 1993 than in 1983, and 98% higher on tomatoes. From 1993 to 1998 insecticide use declined 90% on potatoes and 96.5% on tomatoes. The marked decline in insecticide use on potatoes resulted from the introduction of imidacloprid, which was applied at much lower rates than the insecticides used previously (Ontario Ministry of Agriculture, Food and Rural Affairs, 1998c).
The EI for vegetables decreased 18% from 1978 to 1983, then increased 33.5% from 1983 to 1993. In 1998 the EI was 13.5% lower than in 1983. The mean EIQ did not change from 1978 to 1983, but increased 17% from 1983 to 1993, then declined to 93% of the 1983 value in 1998. Vegetables had the highest mean EIQ in 1988 and 1993, and the second highest in 1978, 1983 and 1998. Part of the increase in the mean EIQ from 1983 to 1993 resulted from the replacement of the fungicide captafol (N-[(1,1,2,2-tetrachloroethyl)thio]-4-cyclohexene-1,2-dicarboximide), with an EIQ of 17.3, by chlorothalonil (tetrachloroisophthalonitrile) with an EIQ of 46, maneb ([1,2-ethanediylbis[carbamodithioato](2-)]manganese), with an EIQ of 64.1, and mancozeb. The use of mancozeb and maneb declined from 1993 to 1998 when they were partially replaced by copper hydroxide (EIQ = 33.3) and copper sulfate (cupric sulfate pentahydroxide), with an EIQ of 47.8. The mean EIQ for insecticides also increased from 1983 to 1993 as the percentage of insecticides with an EIQ score >40 increased from 67% in 1983 to 85.5% in 1993. In 1998 only 55% of the insecticides had an EIQ >40 and the mean EIQ for insecticides declined.
Pesticides versus Production
The 46% increase in pesticide use from 1973 to 1983 was accompanied by an 18% increase in total yield and a 16% increase in the average yield per hectare. This occurred primarily because corn, which had a higher pesticide use and yield per hectare, replaced hay and pasture. From 1983 to 1998 pesticide use declined by 38.5%, but the total yield increased by 6% and the average yield per hectare increased by 14.5%. This represented a 72% increase in yield per kilogram of pesticide. Combining the periods from 1973 to 1983 and 1983 to 1998 (Fig. 9), there was not a significant correlation between the mean yield per hectare and mean application rate (r = -0.003; p = 0.99) or mean EI per hectare (r = -0.20; p = 0.71). The yield per hectare appeared to increase with increasing pesticide use on fruits, grains, and hay and pasture, but increased with a decrease in pesticide use on corn, soybean, field bean, tobacco, and vegetables. However, none of the correlations for individual crops were statistically significant (p > 0.05).
|
|
Changes in Pesticide Risk Based on the Environmental Impact Quotient and Environmental Impact
The EIQ measures the potential risk of a pesticide while the EI measures the risk associated with pesticide use. Pesticide risk, as defined by the EI, peaked in 1983, then declined. However, there was little change in the mean EIQ, or EI per kilogram of pesticide. Nevertheless, there was a shift to lower risk pesticides. The median EIQ declined from 34.6 in 1983 to 30.5 in 1998, and the average EIQ (
EIQi/n) declined from 39.1 in 1983 to 36.0 in 1998. This occurred because of the introduction of newer pesticides with lower EIQs and the elimination of pesticides with high EIQs. Five of the 10 pesticides with the highest EIQ values in 1983 were no longer used in 1998 (Table 2) and there were marked reductions in the use of two other insecticides, parathion (O,O-diethyl O-(4-nitrophenyl) phosphorothioate) and oxydemeton methyl (S-[2-(ethylsulfinyl)ethyl] O,O-dimethyl phosphorothioate). Only one new insecticide, propoxur (2-(1-methylethoxy)phenyl methylcarbamate), was among the 10 highest EIQ values in 1998. Mancozeb, which had the 10th highest EIQ in 1998, had the 16th highest EIQ in 1983. Insecticides accounted for 8 of the 10 highest EIQs in 1983 and 5 in 1998.
|
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
The reductions in pesticide use and risk from 1983 to 1998 were not equal for all crops. The greatest reductions in pesticide use and risk were on corn and tobacco, which accounted for 67% of the pesticide use in 1983. There were significant reductions in pesticide use and risk per hectare and per tonne of yield on both crops. The pesticide use and risk for soybean did not change significantly from 1983 to 1998, but the pesticide use and risk per hectare, per tonne of yield, and per dollar value declined significantly because of the 133.5% increase in area and 39% increase in yield per hectare.
The reduction in risk was also not equal for all stakeholders. The greatest reduction in risk was to the farmworker, but there were also significant reductions in the consumer and ecological risk.
Factors Affecting the Reduction in Pesticide Use
A priori, one would like to assume that the reduction in pesticide use from 1983 to 1998 resulted from initiatives (grower education, IPM programs, development of new technologies) taken under the Food Systems 2002 program. However, while these initiatives may have had an impact, other factors also contributed to the reduction in pesticide use.
First, there was a 7.5% decrease in crop area. However, most of the reduction in crop area occurred because of the reduction in the area of hay and pasture, which had a low application rate despite a 400% increase in the mean application rate from 1983 to 1998. Also, the reduction in the area of hay and pasture exceeded the total reduction in crop area by 28%, indicating that some of the hay and pasture was replaced by other crops that have higher mean application rates. Thus, the reduction in crop area had little impact on the total pesticide use.
A second major factor in the reduction in pesticide use was the change in cropping patterns driven by market forces and new varieties. From 1983 to 1998 there were significant reductions in the area of corn, grains, tobacco, and hay and pasture, and an increase in the area of soybean. The shift from corn and grains to soybean was facilitated by the introduction of new varieties of soybean that have a shorter growing season and are better adapted to the Ontario climate. The increase in soybean production led to the rotational cropping of corn, soybean, and winter wheat–clover (Trifolium spp.), which reduced the need for pesticides and forced a change in the pesticides used. For example, the use of atrazine on corn declined by 67% from 1983 to 1998 because of the switch to alternate herbicides. Atrazine accumulates in soils and ground water, and may damage subsequent crops. The increase in rotational cropping was also a major factor in the 94% reduction in the use of insecticides to control corn rootworm (Diabrotica spp.), which is not a significant pest of first-year corn.
The decrease in the area of tobacco, which had the highest mean application rate in 1983, resulted from social pressures on tobacco use. The reduction in area was also accompanied by a 72% reduction in mean application rate, primarily because of the reduced use of nematocides, which have a high risk to the applicator.
The third factor responsible for the decline in pesticide use has been the introduction of new technologies. Several of the newer herbicides, such as fenoxaprop-ethyl ((±)-ethyl 2-[4-[(6-chloro-2-benzoxazolyl)oxy]phenoxy]propanoate), imazethapyr ((±)-2-[4,5-dihydro-4-methyl-4-(1-methylethyl)-5-oxo-1H-imidazol-2-yl]-5-ethyl-3-pyridinecarboxylic acid), and the sulfonyl-ureas, are applied at rates of grams per hectare rather than kilograms per hectare (Ahrens, 1994; Hatzios, 1998), as are many of the newer insecticides (Thomson, 1998). Improvements in the manufacture of existing pesticides have also increased the proportion of active ingredient per kilogram of pesticide. For example, fenoxaprop-P-ethyl ((+)-ethyl 2-[4-[(6-chloro-2-benzoxazolyl)oxy]phenoxy]propanoate), the active form of fenoxaprop-ethyl, requires 50% less pesticide, and S-metolachlor (2-chloro-N-(2-ethyl-6-methylphenyl)-N-(2-methoxy-1-methylethyl) acetamide, S-enantiomer), the active form of metolachlor, requires 35% less pesticide (Hatzios, 1998). Unfortunately, S-metolachlor was not licenced for use in Canada until the 1999 growing season. Metolachlor accounted for 27.5% of the total amount of pesticide used in 1998, and 40% of the pesticide used on corn and soybean. Had S-metolachlor been available in 1998, pesticide use on corn would have been reduced by an additional 7%, and pesticide use on soybean would have been reduced by an additional 15%. Total pesticide use would have been reduced by an additional 6%, to 56% of the amount used in 1983. The 44% reduction in pesticide use would have achieved 88% of the goal of Food Systems 2002.
There have also been improvements in pesticide application, such as electrostatic sprayers, improved sprayer calibration, and banding of pesticides, which have contributed to the reduction in pesticide use. The introduction and implementation of these new technologies can be attributed in part to the grower education and research components of Food Systems 2002. The introduction of IPM programs, particularly in the fruit and vegetable industries, may also have contributed to the reduction in pesticide use. Pesticide use on fruits and vegetables increased from 1983 to 1993, but declined from 1993 to 1998. Integrated pest management programs can reduce the pesticide use on individual crops by 23 to 61% (Table 4).
|
The changes in pesticide risk usually mirrored the changes in pesticide use on individual crops, but the magnitude often differed because of changes in the mean EIQ caused by changes in the relative amounts of pesticides used. For example, the mean EIQ for corn declined because of the replacement of older herbicides (atrazine, alachlor, butylate, and cyanazine) by newer herbicides (metolachlor, dimethenamid (2-chloro-N-(2,4-dimethyl-3-thienyl)-N-(2-methoxy-1-methylethyl)acetamide), and the sulfonylureas) with lower EIQ values. Conversely, much of the increase in the mean EIQ of soybean resulted from the replacement of alachlor by glyphosate and bentazon, which have higher EIQ values. However, because glyphosate and bentazon were applied at lower rates than alachlor (McGee, 1984; Moxley, 1989), the switch to lesser amounts of pesticides with higher EIQs actually reduced the EI.
The reduction in risk was also not equal for all categories of the EIQ calculation. The major reduction in risk was to the farmworker, primarily from the reduced use of nematocides on tobacco. There were also significant decreases in the consumer and ecological risk, but they resulted primarily from the decrease in pesticide use as the mean EI per kilogram of pesticide increased for both categories.
While the results of this analysis indicate little overall change in the pesticide risk, and even an increase for some crops and categories, the results may be misleading and the reduction in risk may be underestimated. First, many of the newer pesticides, which are applied at grams per hectare rather than kilograms per hectare, have low EIQ values. Switching from "high use–higher risk" pesticides (e.g., atrazine with an EIQ of 33.2 applied at 2 kg ha-1) to "low use–lower risk" pesticides (e.g., rimsulfuron with an EIQ of 14.7 applied at 5–10 g ha-1) may substantially reduce pesticide use and EI, but the reduction in the mean EIQ primarily reflects the weight assigned to the "high use" pesticide. For example, replacing 50% of the atrazine used in Ontario with rimsulfuron reduces the pesticide use and EI by almost 50%, but the mean EIQ decreases <1% because we would still use 200 times more atrazine than rimsulfuron.
A second factor that may cause an underestimation of the reduction in risk is the calculation of the EIQ. The EIQ is one of the few composite environmental impact rating systems for pesticides and incorporates the the widest range of environmental impacts of the published scoring systems (Levitan et al., 1995), but has limitations as a measure of pesticide risk (Dushoff et al., 1994; Levitan et al., 1995). One of the limitations is the scoring system. In calculating the EIQ, the toxicity of pesticides and potential exposure are scored on a 1, 3, 5 scale. This reduces the sensitivity of the EIQ to differences among pesticides. For example, the scores for fish toxicity are assigned to logarithmic differences in the 96-h LC50 (e.g., 1 = >10 mg L-1, 3 = 1–10 mg L-1, and 5 = <1 mg L-1). Thus, thousand-fold differences in toxicity may only differ by a factor of 5, and hundred-fold differences may have the same score. Because of the low sensitivity of the EIQ to differences among pesticides, highly toxic pesticides may appear to have a lower environmental impact than less toxic pesticides when the EIQ is combined with volume data (Dushoff et al., 1994). However, this is also a problem with many other toxicity rating systems (Hertwich et al., 1997).
In defense of the EIQ, defining toxicity is not without problems. The toxicity of a pesticide may vary widely among species. For example, the 96-h LC50 of the insecticide azinophos-methyl (O,O-dimethyl S-[(4-oxo-1,2,3-benzotriazin-3(4H)-yl)methyl]phosphorodithioate), was 0.0011 mg L-1 for bluegill (Lepomis macrochirus) and 4.27 mg L-1 for goldfish (Carassius auratus) (Mayer and Ellersieck, 1986). Even within a species, factors such as size and stage of development, and water temperature, pH, and hardness can cause several-fold differences in toxicity.
Defining exposure is also difficult. The EIQ measures potential exposure as half-life, run-off potential, and preemergent versus postemergent use for herbicides. However, timing of application may be as important as half-life in determining exposure to farmworkers or other species, and soil type and temperature may influence pesticide adsorption, solubility, and half-life (Dushoff et al., 1994), the major components in the calculation of leaching and runoff potential (Kovach et al., 1992). Also, because the EIQ was developed primarily for horticultural crops, the measures of exposure may have limited value with modern production methods on large-field crops, which account for most of the pesticide use. Many changes in agricultural practice have reduced the potential for pesticide exposure to applicators, fieldworkers, consumers, and the environment.
Major changes in agricultural practice that have reduced applicator exposure include: mandatory pesticide education programs for all growers applying agricultural pesticides in Ontario, increased custom application by professionals using high-quality equipment, new pesticide formulations that do not require on-farm mixing of concentrates (e.g., dissolvable packets, lock and load formulations), and a greater number of enclosed tractor cabs that reduce exposure to pesticide drift. Modern field crop production using combines markedly reduces the probability of fieldworker exposure during harvest, particularly when most of the pesticide use is preemergent and postemergent herbicides. Increased mechanical harvesting of horticultural crops, such as cherries (Prunus spp.) and grapes (Vitis spp.), has also reduced the number of individuals picking crops and the potential for direct exposure. Conversely, increased numbers of pick-your-own operations for fruits and vegetables have increased the potential for picker exposure.
At the consumer level, Canadian residue studies demonstrate that the number of products that exceed acceptable residue levels continues to decline (Agriculture Canada, 1991, p. 202–219). There is no evidence of residue contamination in major field crops, which account for most of the area of production.
At the ecological level, approximately 16000 producers in the Province of Ontario have participated in the Environmental Farm Plan, with more than 8000 having completed their farm plans (Ontario Farm Environmental Coalition, unpublished observations, 1999). Modules on pest management, pesticide storage and handling, soil management, and stream protection are part of this program. Improved practices, such as no-till planting, buffer strips along waterways, new types of sprayers, improved sprayer calibration, and banding of pesticides can reduce pesticide use and runoff (Baker and Mickelson, 1994). Also, many farmers, particularly fruit and vegetable growers, have enrolled in IPM programs that have reduced pesticide use and the environmental impact of pesticides.
Despite its weaknesses as a measure of risk, the EIQ is the best available technique for identifiying high-risk pesticides and assessing changes in pesticide risk over large areas such as the Province of Ontario. Benbrook and coworkers (Benbrook et al., 1996; Groth et al., 1999) have proposed a toxicity index that overcomes one of the limitations of the EIQ by increasing the range of toxicity scores. However, they have only developed a toxicity index for mammals, and their measures of exposure have been the amount of pesticide applied (Benbrook et al., 1996) and pesticide residues in foods (Groth et al., 1999). Pesticide residues provide a reasonably accurate measure of exposure, but there may be little relationship between pesticide use and exposure because of differences in half-life, solubility, soil adsorption, and use pattern among pesticides. These factors are incorporated into the EIQ. Defining the exposure to a pesticide is difficult, and differences in toxicity among species complicate the calculation of risk. More complicated scoring systems with greater inputs might be preferred for calculating the potential risk of pesticides, but collecting detailed information on species composition and population sizes, the toxicity for each species and segment of each population, and the true level of exposure for each segment of the population, is probably beyond the capacity of any individual or organization, especially over a large area such as the Province of Ontario.
Policy Implications
The 38.5% reduction in pesticide use from 1983 to 1998 achieved 77% of the goal set by the Food System 2002 program. Part of the reduction resulted from the decrease in the area farmed and there was only a 33.5% reduction in the mean application rate, but pesticide use per tonne of production decreased by 42% because of the increase in productivity. The decrease in pesticide use occurred because of the switch to newer pesticides applied at lower rates, more efficient application of pesticides, and the introduction of IPM programs that reduce pesticide use. However, reducing pesticide use has not reduced the efficacy of pest control as measured by the increase in production.
Most of the reduction in pesticide use from 1983 to 1998 was on corn and tobacco, the crops with the largest pesticide use in 1983. The amount of pesticide used on soybean did not change from 1983 to 1998, but the mean application rate declined dramatically. Defining the goal of the Food Systems 2002 program solely in a historical context implies that limited resources should be targeted to further reductions in pesticide use on the large-field crops, corn and soybean, and on tobacco. Indeed, replacing "high use" herbicides with "low use" herbicides on corn and soybean can achieve dramatic reductions in pesticide use. Replacing metolachlor with S-metolachlor would have further reduced pesticide use on corn and soybean by 9%, and achieved 88% of the goal of the Food Systems 2002 program. However, even though corn and soybean accounted for 65.5% of the pesticide use in 1998, they only accounted for 55.5% of the EI. Both crops have relatively low pesticide use and risk per hectare and per tonne of production, and a relatively low mean EIQ.
While there may be a lesser impact on pesticide use, perhaps resources should be concentrated on "high risk" crops such as fruits and tobacco. Fruits accounted for 11% of the pesticide use in 1998, but had the highest mean application rate and pesticide use per dollar of crop value, and the second-highest pesticide use per tonne of production. They had the highest mean EIQ and accounted for 16.5% of the EI. When factors not included in the EIQ calculation are considered, fruits may pose an even greater risk. Much of the fruit production in Ontario is adjacent to urban centers, air blast sprayers increase the drift potential, and consumers often eat the commodity raw. Eliminating fruit production in 1998 would have reduced the total pesticide use by 45%, approximately the same amount as switching from metolachlor to S-metolachlor on field crops. But, the EI would have been reduced by 49.5%. The reduction in risk from eliminating fruit production is much greater than that of switching from metolachlor to S-metolachlor because the mean EIQ of fruits (46.1) is more than 2.5 times the EIQ of metolachlor (18). Eliminating fruit production may not be a viable policy option because of the economic and social value of the crop, but targeting resources toward reduced pesticide use and alternate pest management strategies on fruits and vegetables, the two crops with the highest mean EIQ, will achieve a much greater reduction in risk than further reductions in pesticide use on the large-field crops.
In evaluating changes in risk we also need to consider the category of risk. There has been a 54.5% reduction in risk to the farmworker, primarily because of the decreased use of nematocides on tobacco. However, nematocides have relatively low consumer and ecological risks. Perhaps we should concentrate on reducing ecological and consumer risk. Reducing herbicide use on corn and soybean will achieve substantial reductions in consumer and ecological risk, but reducing fungicide and insecticide use on fruits and vegetables will provide greater reductions in the consumer and ecological risk per hectare.
Another factor that may have a major impact on the trends in pesticide use and risk is the increased cultivation of genetically modified (GM) crops, particularly those producing insecticides or with herbicide resistance. Bt-corn, which produces its own insecticide, currently accounts for approximately 27% of the field corn planted in Ontario (Statistics Canada, 2000b). It is used to control the European corn borer (Ostrinia nubilalis), a pest not economically controllable previously (Ontario Ministry of Agriculture, Food and Rural Affairs, 1997). The amount of pesticide produced by plants is not measured in the pesticide surveys, but may be significant if the pesticide is produced for the life of the plant. The risk posed by plant-produced pesticides is unknown. Theoretically, there is negligible risk to the applicator, and because the pesticide is contained within the plant, there is limited potential for runoff or leaching. However, pesticide residues in the harvested plant material may pose a risk to fieldworkers and consumers, and to seed or fruit-eating mammals, birds, or insects. Bacillus thuringensis has a relatively low risk (EIQ = 13.5), but other pesticides may have a greater risk.
The adoption of herbicide-resistant crops poses a greater challenge. Glyphosate had the largest increase in use of any pesticide from 1983 to 1998. The increased use of glyphosate on soybean accounted for 61% of the increased use of glyphosate from 1983 to 1998, and 90% of the increase from 1993 to 1998. A large part of this increase may be attributed to the introduction of Roundup-Ready soybean. However, the increased use of glyphosate, which may be applied at 4 kg ha-1 for spot treatment of weeds (Ontario Ministry of Agriculture, Food and Rural Affairs, 1998a), may inhibit the use of newer herbicides such as pendimethalin, dimethenamid, and fomesafen (5-[2-chloro-4-(trifluoromethyl)phenoxy]-N-(methylsulfonyl)-2-nitrobenzamide), which have lower EIQs and may be applied at much lower rates. In essence, by retaining an "old technology", the advantages of reduced pesticide use and risk to be gained from the "new technology" will be lost.
Also, do we need to consider nonagricultural uses of pesticides? Nonagricultural pesticide use was not included in Food Sytems 2002, and prior to 1993 information on pesticide use other than on field crops was only collected for roadside use. This accounted for <1% of the total amount of pesticide used on crops in Ontario in 1983 (McGee, 1984). In 1993, Ontario Ministry of the Environment and Energy (MOEE)-licensed pesticide applicators were included in the pesticide use survey (Hunter and McGee, 1994). The MOEE-licensed applicators applied 1.3 x 106 kg of pesticides, making them the second-largest pesticide users in the Province of Ontario after corn growers. Residential lawns received 62.5% of the pesticide applied and 20% was applied to commercial or public lawns. Only corn and soybean, the two largest field crops, had more pesticides applied than lawns (Fig. 2). The EI for MOEE-licensed applicators was 49.6 x 106, which was second only to corn (Fig. 6), and the mean EIQ was 38.1, which was second only to the value for vegetables (Fig. 7).
Finally, in evaluating the changes in pesticide use in the Province of Ontario we should be aware that pesticide use surveys at 5-yr intervals only provide snapshots of a given year. Pesticide use may vary from year to year because of external factors such as market forces and weather, which influence the amount and type of pesticide used. Cold, wet springs and hot, dry summers, such as 1983, may delay crop development and reduce production (Ontario Ministry of Agriculture and Food, 1983) independent of other factors. Weather may also influence the timing and intensity of pest pressure and pesticide use. For example, the Agriphone system (Table 4) uses weather to predict pest pressure on fruits and vegetables. The potential effects of weather and market forces on both pesticide use and production may be difficult to quantify, but affect both the baseline and the end-point in any comparison.
Food Systems 2002 has a goal of a 50% reduction in agricultural pesticide use in the Province of Ontario. There was a definite reduction in pesticide use and risk from 1983 to 1998, and the Food Systems 2002 goal may be attained in the historical context. In other contexts, there may be a larger or smaller reduction in pesticide use and risk, and increasing nonagricultural use may offset the reduced agricultural use. The reductions in pesticide use and risk have also been confined to a few crops, and pesticide use and risk have not changed or increased on other crops. Appreciating the changing patterns of pesticide use and risk in different contexts over time should determine the directions for future research and interventions, and our interpretation of the success of the Food Systems 2002 program.
|
ACKNOWLEDGMENTS |
|---|
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
This article has been cited by other articles:
