Abstract
Powdery mildew (PM) is the most expensive grape disease to control and causes the greatest losses in quality and yield worldwide. Grape varieties resistant to PM are being developed, but the value of such varieties to growers in different industry segments is not yet determined. Barriers to adoption of new varieties vary by industry segment and are likely to affect the total potential benefit realized by growers. The first step in estimating the potential value of PM-resistant varieties is to establish the costs of PM management that these varieties will mitigate for each major segment of the grapegrowing industry. We used Pesticide Use Reports from the California Department of Pesticide Regulation, data on pesticide application costs, and measures of environmental impact to evaluate the pecuniary and nonpecuniary costs of managing PM in California grape production. We estimate the statewide cost of PM management in 2015 at ~$239 million, of which $176 million was borne by growers of winegrapes. In addition, PM management accounted for 89% of restricted material (pesticide) applications by grapegrowers, so eliminating PM would reduce the environmental burden from disease management in grapes significantly. Using choice experiments, we evaluated the preferences of individual growers for specific varietal traits and found that winegrape growers place a high value on both the varietal name and the savings in costs from reduced application of fungicides. We conclude by discussing possible adoption scenarios for resistant varieties and the resulting benefits to the industry.
- disease resistance
- grape production
- grapevine breeding
- powdery mildew
- raisin
- table grapes
- varietal
- vineyard management
- wine
Powdery mildew (PM) is a disease caused by several fungal species that affects a variety of plants. Grape PM (Erysiphe necator, syn. Uncinula necator) affects grape crops globally. In California, it causes the greatest losses in quality and yield, and the highest costs for control of any grapevine disease (Bettiga et al. 2013). Efforts are underway to develop grape varieties resistant to PM that can be used by growers of all types of grapes, including wine, table, and raisin grapes (e.g., VitisGen, www.vitisgen.org/). A first formal evaluation of the likely value of new, PM-resistant varieties to growers in different segments of the industry has been published (Fuller et al. 2014). Synthetic budget models for representative growers of raisin, table, and winegrapes were used to calculate potential reductions in operating costs/acre from the adoption of new varieties. Annual savings/acre ranged from $186 for growers of raisin grapes to $280 for growers of Chardonnay in the Central Coast region, and $287 for growers of Crimson Seedless table grapes. Because each budget was specific to a particular production system, total benefits from the adoption of resistant varieties were calculated only for specific subsets of the industry. The analysis also showed that the net benefits from adoption could vary by hundreds of millions of dollars depending on the rate and timing of adoption. In the same paper, the most common barriers to adoption of improved varieties bred using conventional marker-assisted breeding were discussed, particularly the premium growers of winegrapes place on Vitis vinifera varietal names.
Other work on the economics of PM management has focused primarily on the behavioral aspects of pesticide use by growers in response to disease pressure forecasting. The Gubler-Thomas Powdery Mildew Index (PMI) is the leading tool for forecasting PM and provides information about disease pressure and recommendations for adjusting spraying intervals for various groups of pesticides. In field trials, growers using the PMI to guide the timing of their treatment saved two to four sprays/year, a significant reduction in both pesticide application costs and the environmental burden from PM control (Thomas et al. 1994, Weber et al. 1996). Recent work on the use of the PMI by grapegrowers suggests that growers adjust the choice and dosage of pesticide products in addition to spraying intervals, and may eventually use more sprays or higher dosage over the course of the year than the field trials suggested (Sambucci 2015, Lybbert et al. 2016, Sambucci and Lybbert 2016).
In this study, we calculated the total costs of managing PM using statewide data on pesticide use, illness caused by pesticide exposure, and two measures of environmental impact. We also extended the evaluation of the likely benefits from PM resistance to the entire grapegrowing industry. We use stated preference experiments to evaluate growers’ willingness-to-pay for specific varietal traits, including the varietal name. We discuss potential barriers to adoption of varieties without a V. vinifera name in the winegrape industry, and the extent to which existing barriers to adoption could influence the realization of benefits from improved varieties.
Materials and Methods
Pesticide use reports database
We used Pesticide Use Reports (PUR) collected by the California Department of Pesticide Regulation (DPR) to calculate the volume of fungicides applied to manage PM and the acreage treated, by category of grape and major region of production (www.cdpr.ca.gov/docs/pur/purmain.htm). Pesticides used for PM control are used almost exclusively for that purpose. The choice of product, timing, and amount of application allows us to identify PM treatments with a high degree of accuracy. The three most commonly used groups of fungicides for PM control are strobilurins, sterol inhibitors, and sulfur. These are used typically as preventive measures and can be applied during periods of low, moderate, or high disease pressure. Other types of pesticides such as biological, systemic acquired resistance products, or cell-signaling inhibitors are typically recommended when disease pressure is low to moderate. Contact materials such as horticultural oils are mainly used for eradication, but can also be applied to prevent an outbreak and as an alternative to sulfur (UC IPM 2015). Total amounts of pesticides applied by chemical category in pounds of active ingredient are summarized (Supplemental Table 1). The calculations include only pesticides applied during the “preferred disease management season” for each growing region, as described in the 2013 University of California Grape Pest Management Handbook (Bettiga 2013), but these totals accounted for 94% of total fungicide applications in 2015. Some of the remaining 6% of total fungicide applications were probably made for PM control; thus, 94% understates the total used for PM control. Some spraying specifically for PM may still be required to maintain resistance to PM in improved varieties, likely not more than one to two sprays/year and maybe every other year, rather than five to 15 sprays annually. This makes our estimates of cost savings optimistic. However, we do exclude some sprays made outside of the prescribed critical disease management season so the resulting estimates, which account for 94% of fungicides applied, might therefore allow a reasonable estimate of the total savings by adopting resistant varieties.
The quantity of pesticides applied gives some indication of the magnitude of the potential savings, both in vineyard management costs to the grower and in environmental costs from pesticide applications, if the danger of PM infections were significantly reduced by the introduction of resistant grape varieties. Pesticide use for PM management varies substantially by geographic location and according to the value of the crop. Geography and grape crop value in California are highly correlated, so it is feasible to analyze the two together. The next section describes our approach to analyzing the trends in the use of pesticides at the county and regional level, making use of the PURs.
County crop reports and major regions of grape production in California
We combined County Crop Reports obtained from the office of the Agricultural Commissioner for each county in California and viticultural information about growing regions to assemble data on grape production by major region. California includes five major grapegrowing regions: the North Coast, Central Coast, South Coast, Northern Central Valley, and Southern Central Valley (Table 1). This allows discussion of labor and application costs specific to a particular region. Each region has a slightly different growing season for grapes because of differences in climate (Bettiga et al. 2013). Thus, the timing of the critical disease control season also varies.
Grapegrowing regions can be classified further by PM disease pressure. PM thrives in moderate climates, so at any given time during the growing season, coastal regions and the foothills typically have greater disease pressure than the Northern and Southern Central Valley regions, where temperatures are higher. The bearing acreage, average price/ton, and total farm gate value of grape production for each region were also determined (Table 1).
Costs of products and application from cost and returns studies
The pecuniary costs of PM management were calculated, including the dollar value of pesticides applied and the costs of labor and equipment involved in the application process. University of California Cooperative Extension (UCCE) Cost and Return Studies (UCCE 2008–2016, see Supplemental References) indicate that PM management costs—including materials, labor, and costs of running tractors and other equipment involved in fungicide application—average ~3 to 7% of revenue for grapegrowers across different years and production systems. Given the information on revenue in Table 1, we can expect these costs to be between $180 and $420 million/yr. This estimate is comparable to those found in previous studies. The costs of PM were calculated to be $48 million/year for a subset of the industry covering 235,000 acres, about a quarter of total grape acreage in California (Fuller et al. 2014). Assuming that the average costs/acre are about the same for the rest of the industry, the costs for the entire state would then be roughly $200 million/year.
Several hundred products are registered with the State of California for use on PM, and specific pricing information is not readily available for all of them. Consequently, estimates of the monetary value of all products applied in a particular year are approximate. The products were grouped into categories according to their content of active ingredients, and prices/pound of chemical for the most-used products in each category were used to calculate the dollar value of products applied.
We used two sources of information to estimate the average costs/acre to apply PM treatments, including costs of labor and equipment. PM fungicides can be applied either as dry dust (sulfur dust) or liquid spray (wettable sulfur and all other fungicides). The applications are made with a tractor and appropriate spraying equipment: a duster for dry applications or a large capacity vineyard sprayer (100 gallons or more) for the other fungicides. The most recent UCCE Cost and Return Studies for Napa, Sonoma, and Lake counties (which are in the North Coast region), the Northern and Southern San Joaquin Valley region, and the Sacramento Valley region, were used to estimate application costs/acre (Supplemental Tables 2 and 3). Assumptions about the types of labor and machinery used for fungicide applications, with allowances for setup and removal of equipment and payroll overhead, are based on the UCCE Cost and Return Studies. Northern San Joaquin Valley and Sacramento Valley are part of the Northern Central Valley grapegrowing region. The Southern San Joaquin Valley is located in the Southern Central Valley region. We used similar cost and return budgets for a selection of growers typical of the Central Coast and San Joaquin Valley Regions (Fuller et al. 2014) to update some estimates of application costs and practices that have changed since the most recent UCCE Cost and Return Studies (Supplemental Table 4).
Analysis of application costs for counties not included in the UCCE Cost and Return Studies requires a set of assumptions about appropriate work rates (hours/acre) for each type of application. Drawing on estimates from the available UCCE Cost and Return Studies for San Joaquin and Sacramento Valley regions, we assumed the work rate to be 0.3 hr/acre for dry dust, and 0.5 hr/acre for liquid spray. Tractor use time includes an additional 10%, and equipment operator time includes an additional 20% allowance for equipment set up and removal, maintenance, field repair, and work breaks. Operating costs are based on costs for a 60HP tractor. These rates are applied to all grapegrowing counties in the San Joaquin and Sacramento Valleys. The assumed work rate for the North Coast and Central Coast regions was greater than for the Central Valley: between 0.7 and 1.0 hr/acre for liquid sprays, and 0.5 hr/acre for dry dust, plus the allowances for additional tractor use time and operator time as above. Tractor use costs assume a 60 horsepower tractor and the same operator time allowance as above.
Environmental impact measures
The toxicity of each pesticide product used in agriculture is evaluated both with respect to humans and the environment. Common health hazards include skin or eye irritation and inhalation potential. Environmental hazards include groundwater contamination; toxicity to fish and other aquatic organisms, domestic animals, and livestock; and drinking water contamination. According to the United States Environmental Protection Agency (EPA) Pesticide Product Label System (PPLS), most products used for PM management are of low toxicity, with the exception of some synthetic products in the sterol inhibitor category. The PPLS provides a collection of pesticide product labels with three potential hazard levels: caution, warning, and danger. The hazard levels are based on exposure to undiluted chemicals, so they are especially relevant for handling during mixing and preparation for application. Labels are available on the EPA website (http://iaspub.epa.gov/apex/pesticides/f?p=PPLS:1).
Pesticide risk indicators typically use a ranking or an index based on toxicological and physiochemical properties of the pesticides, as well as site-specific environmental conditions (Bockstaller et al. 2009). Many pesticide risk indicators are developed with a specific purpose or user in mind and it is not always possible to apply them outside the intended scope (Labite et al. 2011). We quantify the environmental effect of PM control on human health and the environment using two risk measures: the environmental impact quotient (EIQ) and the pesticide use risk evaluation (PURE) system.
The EIQ
The EIQ is an aggregate measure of environmental impact, which combines the pesticide hazard posed to farm workers (applicator and harvester exposure), consumers (consumer exposure and groundwater contamination), and the environment (toxicity to aquatic and terrestrial organisms and bees) (Kovach et al. 1992).
Data on toxicity of individual chemicals was collected from sources such as the Extension Toxicology Network (Cornell University Pesticide Managment Education Program/ExToxNet), CHEM-NEWS (Cornell Cooperative Extension Network), the SELCTV database (Oregon State), studies by the USDA Economic Research Service and the EPA, and material safety data sheets from chemical manufacturers.
The EIQ is a simple average of three components: farm worker EIQ, consumer EIQ, and ecological EIQ. The basic principle behind EIQ is that the impact of the chemical is equal to a measure of toxicity multiplied by the potential time of exposure. Therefore, while the total EIQ is a simple average of three components, the elements within each component are weighted based on toxicity and exposure potential. Details on relevant aspects of chemical toxicity and calculating the EIQ are provided in Kovach et al. (1992).
Overall, the toxicity of fungicides used for PM control is relatively low. Most have an EIQ of 40 or less, with the exception of lime sulfur. Lime sulfur has the greatest EIQ (67), but is not widely used. Even though the EIQ values for PM pesticides are relatively low, the application volume is high and consequently, the environmental impact of PM treatments is significant, relative to the total environmental impact of grape production. Supplemental Table 5 provides a summary of EIQ values for each fungicide product commonly used to control PM, and Supplemental Table 6 summarizes EIQ values for other chemicals used on grapes.
We used our calculations for the volume of each PM fungicide applied to grapes, by region, to calculate the total EIQ from PM management and the distribution of EIQ among regions and between wine and non-winegrapes.
The PURE system
An alternative method for evaluating the environmental impact of PM control is pesticide use risk scores. The PURE system uses information on pesticide properties (toxicity) and environmental conditions to evaluate the risk from pesticide use on a specific field with respect to five dimensions of the environment: groundwater, surface water, soil, air, and bees (Zhan and Zhang 2012). The model was developed as a decision support system for growers to help with evaluating the potential pesticide-use risk for a specific field. PURE is linked directly to the PUR and provides a risk score for each pesticide application and a total annual risk score for a particular field. The environmental conditions incorporated include soil properties, meteorological conditions at the time of application (precipitation and temperature), groundwater depth, ground slope, distance to surface water, and soil properties specific to the area of the site.
Pesticide illness surveillance program (PISP)
In California, physicians have been required to report any illness caused by pesticide exposure since 1971. Records of illness reports are available from the PISP of the DPR.
Stated preference surveys
An alternative to the budgetary approach to determining the value of reduced pesticide applications is a preference-based approach. We constructed discrete choice experiments to elicit producers’ willingness-to-pay (WTP), which encompasses not only the monetary costs but also any additional psychic disutility producers might derive from pesticide applications. Choice experiments have been used in a variety of studies to elicit consumers’ WTP for specific attributes of agricultural products, including organic versus milk produced from cows developed from a potential cloning process (Brooks and Lusk 2010), color of salmon (Alfnes et al. 2006), and wine labeling (Mueller and Umberger 2010). While hypothetical bias is often a concern with surveys, choice experiments can have high levels of external validity (Chang et al. 2009, Brooks and Lusk 2010) and produce marginal WTP values for different attributes that are on par with those derived from non-hypothetical approaches (Carlsson and Martinsson 2001, Lusk and Schroeder 2004).
To our knowledge, no prior studies have used choice experiments to elicit measures of producer WTP for varietal traits of grapes. Some previous studies have used choice experiments or similar methods to elicit consumer preferences for fruit quality traits in table grapes and their WTP for wine attributes (Birol et al. 2009, Gustafson et al. 2016). Studies of other perennial crops have used choice experiments to elicit growers’ and intermediaries’ preferences for quality traits in apples, sweet cherries, peaches, and strawberries (e.g., Gallardo et al. 2015, Choi et al. 2017), but none of these studies focused on agronomic traits such as disease resistance specifically; producers were found to value fruit quality attributes that made the fruit most attractive to consumers, such as crispness, juiciness, and color. Other studies have examined the adoption of new varieties of field crops, such as rice and maize, and a subset of that literature used choice experiments to elicit measures of producer valuations of specific traits, including agronomic traits such as pest and disease resistance, which have been a focus of innovation in field crops (e.g., Kassie et al. 2017). However, the large differences in crop production systems and production environment limit the relevance of this subset of literature to our research.
We designed and administered an online choice experiment survey to elicit valuations of specific varietal traits from a sample of grapegrowers. The survey was developed through a series of interviews and several rounds of pretests with viticulturists, farm advisers, and grapegrowers to determine the appropriate ranges of attributes and to test the choice experiment questions. The resulting online survey was then administered via UCCE and industry association mailing lists to ~900 respondents. A paper version of the survey was also distributed at several UCCE grapegrower events with a prepaid return envelope (see Grape Grower Survey in Supplemental Materials). Participants could opt out from taking the survey, so the resulting pool of respondents was self-selected. We treat the responses from a self-selected sample with caution, since these responses may not be valid for grapegrowers who do not share the same unobservable characteristics as the pool of the respondents. In particular, we are careful to not interpret the WTP values obtained from the analysis as representative of average WTP values for all growers.
The survey was completed by 106 growers of table, raisin, and winegrapes, with 15 responses arriving by mail and the rest submitted online. We disregarded surveys with incomplete responses and used the remaining 72 surveys for analysis. The demographic characteristics of the respondents were summarized (Supplemental Table 7). The survey consisted of two parts. The first part included a series of simple importance-scale questions. The exact list of attributes and the layout of the scale questions from the survey are shown (Supplemental Table 8). For each attribute on the list, the respondents were asked to choose a category to designate the attribute as “very unimportant,” “somewhat unimportant,” “somewhat important,” or “very important.” The rating of the varietal traits provides an overview of what growers consider as important characteristics of grapevines, but does not allow us to understand how growers would rank a combination of specific traits or choose among varieties, each with a set of desirable traits, or how growers might make tradeoffs between attributes. The second part of the survey included the choice experiment, which we designed to elicit grower preferences for varietal traits such as PM resistance, whether the variety was genetically modified, whether the variety was able to retain the vinifera name, the number of fungicide applications required/season for managing PM, and the environmental impact of those applications. A summary of the experimental design and the options offered is shown (Supplemental Table 9).
Overall, there were 108 possible combinations of characteristics for Option A (a conventional variety) and 16 possible combinations for Option B (a variety with resistance to PM). This produces 1728 possible choices. From this full factorial, we selected 16 choices that minimized the D-efficiency score; essentially minimizing the correlation between attributes both within and across options. The 16 choices were blocked into two sets of eight, and one of the two blocks was presented to the growers at random. An example of a choice question presented to a grower of winegrapes is provided (Table 2).
We used a random utility model to derive a growers’ WTP. For each grower i, the utility from choosing the jth grape variety can be expressed as: Eq. 1where Vij is the systematic portion of the utility derived from planting variety j, and εij is the random element. We defined an empirical form for the deterministic part of Equation 1, describing the utility derived by grower i from planting variety j, as follows:
Eq. 2
In Equation 2,PMj, GMj, Chardj, NumFungAppsj, Envj, and WHealthj are variables for a particular variety, j, representing whether the variety is resistant to PM or genetically modified, the varietal name, and variables representing the number of fungicide applications, environmental impact, and impact on worker health for that variety, implied by its other attributes. Variable Pij is the price at which variety j is offered to grower i, and ξij is the error term. Coefficients β1 through β7 represent the added utility from planting a variety that has the corresponding characteristics. WTP is determined as the price difference between two options (or attribute levels) that makes a grower indifferent between the two (the difference in utility of attributes expressed relative to the marginal utility of money). WTP for PM resistance, for example, is computed as: –β1/β7. Of specific interest is WTP for a genetically engineered variety that is PM resistant, computed as: –(β1 + β2)/β7.
We estimate the parameters in Equation 2 by maximizing the log-likelihood function:
Eq. 3
In Equation 3,Cij = 1 if grower i chooses variety j, and Cij is equal to 0 otherwise. In the same equation, Prij is the probability of grower i choosing variety j. Because the choices in this model are binary, a simple logit model can be estimated where the variables are specified as difference in attributes between options A and B. We estimated the model only for growers of winegrapes because the survey data did not include enough observations for statistical analysis of responses by raisin and table grapegrowers.
Results
As a first step, we established the statewide costs of PM management, which resistant varieties would serve to mitigate. The total statewide costs of PM management in California include pecuniary costs, such as the costs of purchasing and applying fungicides, nonpecuniary costs such as environmental and worker health effects, and the inconvenience to the grower from having to worry about the appropriate methods of disease control or the potential damage to the crop.
Pecuniary costs
We grouped the products used to control PM into categories according to their content of active ingredients, and used an estimate of the average price/pound of chemical for each category to calculate the cost of products applied (Table 3). Total expenditure on PM pesticide products in California in 2015 was ~$78 million. Over 50% was spent on synthetic fungicides such as sterol inhibitors and strobilurins. Sulfur was ~22% of total expenditure, but >90% of the total pounds of pesticide products applied.
In addition to the costs of pesticide products, we estimated the application costs/acre and used PUR data to calculate the total application costs, including labor and equipment, for each region (Table 4). The differences in costs result from higher prices for labor and equipment in North Coast counties and also longer operation times. Longer operation times may occur because of terrain or applicators taking greater care with a crop of higher value. Additionally, Napa County wage rates include a payroll overhead of 45%, which is significantly larger than the overhead for other counties, which have payroll overhead of 33 or 34%.
Using assumed average application rates for each region, the total application costs for chemicals used for PM treatment were calculated as ~$162 million statewide (Table 5). The total pecuniary costs of PM management for materials and application were $239 million in 2015, which was ~4% of total gross revenue for that year (Table 1).
Environmental and health effects of PM
We applied EIQ values to products used to control PM according to the PURs in 2015 to calculate the environmental impact of all PM fungicides applied during the 2015 growing season (Figure 1). Sulfur is by far the biggest contributor to the EIQ for grapegrowers because of the dosage/acre (10 pounds or more/application) and the frequency of application. A review of literature on the environmental and health impacts of sulfur suggests that sulfur is widely considered to be environmentally neutral (Cornell University Pesticide Management Education Program/ExToxNet 1995), and the main concern with sulfur applications in agriculture is the effect on human health. Specifically, sulfur may cause respiratory irritation and potentially illness (McGourty 2008), although the exact mechanism and type of exposure that causes illness is unknown (Lee et al. 2005). Therefore, we discount the ecological component of the sulfur EIQ and consider the human health component in more depth. For example, the exposure risk to workers is greatest during sulfur applications, because of the large/acre volume and frequent applications. In this case, the risk is additive because there is potential for exposure during each application. In contrast, some synthetic chemicals are applied at longer intervals (up to 21 days) and in very diluted form (the application rate for synthetics is less than one pound/acre, versus 10 or more pounds/acre for sulfur).
Appropriate equipment and safety precautions can reduce or eliminate worker exposure, but implementation and management of such equipment and precautions is a cost to the grower. The cost of enforcing the safety rules (e.g., worker training, supervision, and any necessary equipment) is a pecuniary cost, although it is not counted in our budgets, while the general inconvenience of having to implement these rules and the existing potential for unknown negative health effects is a nonpecuniary cost.
The PURE values/acre of products used for PM control was compared to that of other products applied on grapes (Figure 2). In contrast to calculations made with the EIQ, according to the PURE, the impact on the environment from products used for PM control is smaller than that of other products applied to grapes. While we did not calculate the total EIQ from non-PM pesticides, the EIQ is additive and the large volume of sulfur causes the EIQ from the use of sulfur (and consequently, from PM control) to exceed the EIQ from any other chemical. The annual PURE scores/acre (a total for all applications during the growing season) were calculated for each environmental component (Figure 2). While the effect of applications to control PM is generally less than that of non-PM applications, soil and air appear to be the components of the environment most affected by PM control. In addition, applications to control PM have a greater median score for bees, although the average annual values/acre are still greater for non-PM products. For other environmental components, the effect of fungicides used to control PM is relatively benign compared to other pesticides applied on grapes. The PURE scores are specific to California, so they may not be representative of the environmental effects from PM management in other states. However, California accounts for the vast majority of United States grape production.
We can use the EIQ or PURE scores to evaluate the environmental impact of PM pesticides relative to other chemicals and to make general comparison of applications/acre among growers in different regions or who grow grapes for different market segments or end-uses. While fungicides used for PM treatments are relatively nontoxic compared with some other classes of pesticides, the large volumes applied and frequency of application do matter, particularly in reference to safety measures to prevent worker exposure. Although the exact effect on human health is unknown, implementation of safety measures and the potential for human health effects factor into both pecuniary and nonpecuniary costs for the grower. The average share of illness reports for grapes in all illnesses from exposure of farm workers is summarized (Figure 3).
Between 2005 and 2015, there were ~2100 reports of illness from exposure to agricultural pesticides in a farm setting, or 210 reports/yr on average. Of those, 343, or ~34/yr, were from use of pesticides on grapes. About 11% of all illness reports from pesticide exposure in a farm setting are related to fungicide exposure in vineyards. Of all illness reports from pesticide exposure in vineyards, just over 67% are from fungicides and so are most likely related to management of PM. Reports of illness from exposure to sulfur account for 40% of illness reports related to the use of fungicides on vineyards, and for 60% of illness reports related to the use of fungicides on other crops.
Grower valuation of varietal traits
The grower survey indicated that the traits growers consider most important are savings from reduced pesticide sprays, drought tolerance, improved fruit flavor, and the ability to retain a V. vinifera varietal name (Figure 4). We estimated the model only for growers of winegrapes, because the survey data did not include enough observations for statistical analysis of responses by raisin and table grapegrowers (Table 6). The dependent variable is equal to 1 if a grower chooses Variety Option B (new variety resistant to PM), and equal to 0 if a grower chooses Option A (conventional variety not resistant to PM) (Table 6). The attributes are specified as differences between Option B and Option A. For example, the variable indicating price is calculated as Price B – Price A. Of 378 total choices, the model correctly predicts the outcome 59% of the time. Of these, 147 (39%) were for Option A (not resistant to PM) and 231 (61%) were for Option B (resistant to PM). In our scenario, a resistant variety can only retain the V. vinifera name if it is genetically modified. If resistance is introduced by conventionally crossing a desirable V. vinifera variety with a non-vinifera variety, or even with another vinifera variety, the valuable varietal name is lost.
Our analysis presumes that it may be possible to use genetic engineering or other modern biotechnology techniques to introduce resistance genes to desirable V. vinifera varieties and preserve the varietal name. As of the time of this writing, it has not been established whether a genetically engineered grape variety would be allowed to retain a V. vinifera name in the United States or other jurisdictions. Other tools of modern biotechnology (such as gene editing) make it possible to modify existing organisms without introducing foreign genes, and the United States government has opted not to regulate gene-edited crops as though they were GMOs. It seems reasonable, therefore, to speculate that Chardonnay modified using, for example, CRISPR to make it resistant to PM, might still be called Chardonnay. However, we do not know yet.
The only statistically significant coefficients were related to price, varietal name, and the number of fungicide applications. The intercept coefficient is the average utility of Option B compared to Option A, not explained by other attributes. In our analysis, it represents the value of PM resistance in a variety without a familiar V. vinifera name compared to a conventional, non-resistant Chardonnay, and it was not statistically significantly different from zero. The increased price of the resistant variety relative to a conventional variety decreased the likelihood of this variety being chosen. The ability of the variety to retain the vinifera name (Chardonnay) increased the likelihood of this variety being chosen. A failure of the resistant variety to decrease the number of fungicide applications decreased the likelihood of the variety being chosen. These results are consistent with the rankings of varietal traits (Figure 4).
We use the results of the logistic regression (Table 6) to calculate the WTP values for attributes of interest. The marginal WTP for the ability of a resistant variety to retain the name Chardonnay is $16.9/vine, while the WTP for a reduction in the number of fungicide applications/yr relative to a non-resistant variety is $1.72/application/vine. The regression results can be used to simulate the predicted market share of the two varieties with different traits. Results for several sample baseline simulations are shown (Table 7).
The baseline scenario was based on the average values of characteristics for a conventional variety and most likely expected characteristics of a PM-resistant variety (Table 7). Scenarios 1, 2, and 3 show how changes in price, varietal name, and the number of fungicide applications would affect the predicted market share for each variety. The biggest change in predicted market share happens if the resistant variety is able to retain the vinifera name (in this case, Chardonnay). Since this is not possible for a variety that is conventionally bred (but, perhaps, possible for a genetically modified variety), Scenarios 1 and 2 provide examples that are more realistic. An increase in the price of a conventional variety or a decrease in the number of fungicide applications required by a traditional variety both decrease the market share of variety B and increase the market share of variety A.
Discussion
Grapevines resistant to PM would potentially provide large economic benefits to grapegrowers in California. The potential cost savings for all types of raisin grapes, Crimson Seedless table grapes, and Central Coast Chardonnay was estimated at $48 million/yr (Fuller et al. 2014). We estimate pecuniary costs of managing PM at $239 million/yr for the entire state of California. It is difficult to place a monetary value on environmental costs, and different measures of the impact on the environment lead to different conclusions about the environmental impact from PM management on grapes. According to the EIQ, which emphasizes the volume of application of each chemical, depending on the location and type of grapes being grown, PM management can account for >90% of the environmental burden from pesticide applications on grapes. According to the PURE score, which emphasizes the toxicity of individual chemicals, dosage during application, and weather and soil conditions specific to California, PM management has a smaller impact on the environment than other chemicals applied on grapes, and the environmental impact from PM management is limited mostly to soil, air, and, in some cases, bees.
The potential benefits from PM resistance for the industry would be substantial, even though not all of these costs could be avoided with resistant varieties. In particular, a major objection to the adoption of resistant varieties by growers of winegrapes stems from the fact that, no matter how similar they may be to their V. vinifera parents, varieties with improved resistance acquired through conventional breeding are not able to use valuable V. vinifera varietal names. This objection would be especially relevant for growers located in the North Coast and Central Coast regions, where the highest-priced winegrapes are grown and where varietal names are most valuable. The breakdown of pecuniary and nonpecuniary PM costs by region is shown (Table 8).
Growers of winegrapes in California incurred $176 million in pecuniary costs to control PM in 2015. Of these, growers in the North Coast region incurred $44 million and growers in the Central Coast region, $44 million. The remaining $88 million was incurred by growers of winegrapes located in the Northern and Southern Central Valley regions. Costs of managing PM for table and raisin grapegrowers (located in the Northern and Southern Central Valley regions) were $63 million in 2015. The results of choice experiments indicate that while growers of winegrapes do value the varietal name, they also value reduced costs of controlling PM. While we do not know exactly which growers will or will not adopt the new improved varieties, growers in the Central Valley regions—where yields are higher, prices/ton are lower, and the premium for varietal names is less significant—will be more likely to adopt resistant varieties. In addition, growers of raisin and table grapes do not have similar reservations about varietal names and are likely to be among the first adopters of resistant varieties. If the lack of a premium V. vinifera name prevented all growers of winegrapes from adopting improved varieties, the potential annual benefit from resistant varieties would fall from $239 million to $63 million/yr. However, if only growers of higher-valued grapes in the coastal valleys were hesitant to adopt, the potential benefits could be as high as $151 million, shared among the table, raisin, and winegrape growers in the Central Valley.
Conclusion
PM is responsible for the bulk of pecuniary disease management costs for the entire grape industry. In addition, the volume of pesticides applied creates an environmental burden that contributes to nonpecuniary costs from PM. The magnitude of the statewide costs of managing PM suggests a large potential value of PM-resistant varieties to growers, even considering the potential barriers to adoption discussed here and elsewhere (see Fuller et al. 2014).
For individual growers, the varietal name and cost savings from reduced fungicide applications appear to be the biggest drivers of adoption. With growers of winegrapes less likely to adopt improved varieties, potential annual benefits could be significantly reduced (since total potential benefits to the wine industry are large at $176 million/yr). However, even if it does not bear a valuable vinifera name, a PM-resistant variety may still command a significant market share if it reduces the bulk of the PM-related fungicide applications and its price is comparable to that of a traditional variety. Consumer acceptance of blended wines or wines without specific varietal names could increase the likelihood of acceptance of non-vinifera varieties by winegrape growers.
Acknowledgments
The work for this project was partially supported by Specialty Crop Research Initiative Competitive Grant Award No. 2011-51181-30635 of the USDA National Institute of Food and Agriculture (the VitisGen project), Specialty Crop Research Initiative Competitive Grant Award No. 2015-51181-24393 of the USDA National Institute of Food and Agriculture (the Efficient Vineyard project), by the Giannini Foundation of Agricultural Economics, and by USDA National Institute of Food and Agriculture Hatch/Multistate project 1011710. The authors are grateful for advice and comments provided by Mark Battany, Larry Bettiga, Monica Cooper, Nick Dokoozlian, Matthew Fidelibus, Franka Gabler, Doug Gubler, Ross Jones, Travis Lybbert, Mike Moriyama, Kathleen Nave, David Ramming, Rhonda Smith, Rick Stark, and Andrew Walker.
Footnotes
Supplemental data is freely available with the online version of this article at www.ajevonline.org.
By downloading and/or receiving this article, you agree to the Disclaimer of Warranties and Liability. The full statement of the Disclaimers is available at http://www.ajevonline.org/content/proprietary-rights-notice-ajev-online. If you do not agree to the Disclaimers, do not download and/or accept this article.
- Received March 2018.
- Revision received September 2018.
- Revision received December 2018.
- Accepted December 2018.
- Published online April 2019
- ©2019 by the American Society for Enology and Viticulture