EPIDEMIOLOGY AND MANAGEMENT OF WALNUT BLIGHTwalnutresearch.ucdavis.edu/2003/2003_357.pdf · walnut...

EPIDEMIOLOGY AND MANAGEMENT OF WALNUT BLIGHT J. E. Adaskaveg, H. Förster, D. Thompson, C. Thomas, R. Buchner, B. Olson, C. Pickel, T. Prichard, and J. Grant ABSTRACT Field studies focused on evaluating the efficacy of new bactericides for reducing the incidence of walnut blight, verification of XanthoCast as an effective model for forecasting the disease, and on evaluating host resistance among walnut genotypes under ambient and simulated-rain conditions. In studies with new bactericides in plots in Butte, Tehama, Solano, and Fresno Co., no material evaluated was as effective as copper-Manex under low- and high-rainfall conditions for management of copper-resistant populations of the walnut blight pathogen in California. The non-registered DOW bactericide formulation 918-48C was effective at low rainfall by significantly reducing the disease incidence from that of the control in the Solano and Fresno Co. plots, but was not effective in plots where high rainfall occurred. Improved agricultural formulations of the DOW bactericide have to be developed with increased persistence on plants before any additional evaluation. The bactericide peroxyacetic acid (e.g., Zerotol) showed promise at the low rate evaluated in each trial. Significant reductions in disease from the untreated control with efficacy similar to copper-Manex treatments occurred in Fresno, Solano, and Butte Co. In Tehema Co., the Zerotol low-rate treatment showed a statistical trend for reducing disease by 50%. High rates were less effective in each of the plots. Possibly high rates or high application volumes of this material were phytotoxic to plant tissue. In the last three years, we demonstrated that simulated rainfall studies were very effective in developing efficacy data. In the Solano Co. trials, single applications of an inoculum-reduction treatment (Copper-Manex-0.5% Breakthru) under ambient or simulated rain-conditions failed for a third year, demonstrating that multiple, in-season applications of bactericides are needed under favorable environments. Thus, with a high reproduction potential of the pathogen under high rainfall environments (simulated rain treatments), this strategy is not part of our management guidelines. XanthoCast, the internet-based walnut blight-forecasting program, was available commercially for a third year for growers. The 7-day and cumulative indices followed the actual disease progress in each of 3 locations. Correlations of observed and predicted XanthoCast values that were based on 3- and 5-day weather forecasts provided by Fox Weather gave coefficients of determination or R2 values mostly between 0.72 and 0.88 for 12 locations. Thus, disease was accurately predicted for different specific microclimates recorded for each weather station for three years and the model was experimentally verified in simulated rain studies in Fresno, Solano, Butte, and Tehama Co. Using a 17-day temporal “yard-stick” to track environmental conditions back in time from actual disease progress values, three distinct infection periods were identified in Butte and Tehama Co. between mid-April and early May. Furthermore, the model and the forecast system allow for targeted bactericide applications based on walnut blight infection periods. Disease was high throughout the walnut growing regions of the state this year and XanthoCast called for two to four bactericide applications. In Butte Co., no significant differences were observed in disease incidence between a 9 half-row application program (0.35%) and the XanthoCast program where 3 applications of Kocide-Manex were made on every row (4%). Both programs significantly reduced the disease incidence compared to the

untreated check at 17%. In Tehama Co., one mid-April application was missed before a critical infection period, but still the management program based on XanthoCast significantly reduced disease as compared to the untreated control. Simulated-rain studies were also very important in evaluating walnut genotypes at KAC. In these studies, cv. Payne was the most susceptible with 15% disease incidence, whereas Serr, Franquette, Cheinovo, and 90-28-30 had no detectable incidence of disease on fruit. Other varieties were statistically intermediate between Payne and the four varieties that had no diseased fruit. INTRODUCTION Walnut blight, caused by Xanthomonas juglandis, is a major disease of walnut in central and northern California. The pathogen attacks catkins, female blossoms, green shoots, leaves, buds, and fruit of English walnut. Fruit infections account for most of the economic loss in California. These infections commonly occur in the spring under wet conditions. The bacterium survives from one year to the next in buds (healthy and diseased), diseased fruit that remain on the tree, and possibly in twig lesions (Miller and Bollen 1946; Mulrean and Schroth 1982; Teviotdale et al. 1985, Ogawa and English 1991). Chemical treatments have been the most commonly used management practices for walnut blight. Copper-based compounds have historically been the most efficacious and the most widely used. Failures in blight control are related to lack of protection during conducive environments for disease development, as well as to the development of populations of the bacterial pathogen that are less sensitive to copper-based bactericides (Lee et al. 1993). In any disease management program, dependence on any one chemical treatment potentially can lead to the loss of efficacy of the treatment due to the development of resistance in a pathogen population to that chemical (Adaskaveg et al. 2001). Because of the development of copper-resistant populations, management of the disease is dependent on applications with higher rates of copper, addition of other bactericidal compounds to copper treatments (e.g., Kocide®-Manex®), or on new bactericidal treatments with different modes of action. Potential alternatives that have been evaluated by us include natural bactericidal products (Serenade), bactericidal sanitation treatments (e.g., DBNPA, Zerotol, Oxidate, etc.), systemic acquired host resistance (SAR) compounds (e.g., Actigard®, Milsana®), and antibiotics (Starner). Starner is currently registered in Japan for control of bacterial diseases including those caused by Xanthomonas species. It was, however, no longer evaluated in our trials. Starner is known to develop resistance in pathogen populations and this class of antibiotics is not being developed in agriculture because of the importance in medicine. Based on our previous research, the DOW bactericide was selected as the most likely compound to be developed by its manufacturer. The formulations that were supplied previously (DOW 117-1, 117-2) were very water-soluble and did not persist well on the plant surface. Phytotoxicity was observed in some trials with these and an additional formulation (DOW-EC). No phytotoxicity, however, was observed with the 2002 formulations DB 918-48C and DB 918-54A. The DB 918-48C formulation was more effective in 2002 and thus, it was evaluated again in 2003. In 2003 we also evaluated the acidified hydrogen dioxide formulation Zerotol (BioSafe, Inc., Glastonbury, CT) that is registered for suppression of fungal diseases on a wide range of crops and had showed promise in 2002. Peroxyacetic acid is the active ingredient in Zerotol and other products such as Oxidate® and Storox®. These treatments were compared to a commercial pre-mix of copper and calcium hydroxide (Cuprofix 20DF® and 40DF®, Cerexagri

Inc.) and to Kocide (Griffin LLC), both mixed with Manex. A formulation of copper in gelatin (Bioacumen-Business Generation, North Sydney, Australia) was also evaluated. In 1998, severe epidemics of walnut blight caused substantial yield losses in northern California (Adaskaveg et al., 1998). In 1999 to 2002, however, low rainfall and cool temperatures were not conducive for the development of walnut blight (Adaskaveg et al. 2001). In each low-disease year, the incidence of disease has decreased in most orchards where trials were conducted regardless of bud populations of the pathogen observed during the dormant period. The concept of the disease triangle is essential for the development of the disease. A microclimate model to predict walnut blight in a forecasting system that is developed with all components of the disease triangle (host, pathogen, and environment) will help in the management of this potentially destructive disease of walnut. Inoculum appears to be the most predicable parameter to estimate in commercial orchards. This is because the pathogen is endemic in walnut orchards throughout California, the previous year’s disease incidence is an excellent indicator of inoculum potential in the orchard for the next season, and the pathogen has a high reproductive potential under conducive environments (as we have shown in our irrigation studies).

In our epidemiological studies, we developed critical information on field environmental conditions required for infection and disease development. The interaction of daily wetness period duration and temperature during the wetness period is important. The optimum temperature for in vitro bacterial growth is 28 to 32 C with a minimum of about 5 C and a maximum of 37 C, whereas infection of walnut tissues can occur between 5 and 27 C (Miller and Bollen 1946). Miller and Bollen (1946) also concluded that rainfall was involved in dissemination, and wetness periods of only 5 min were sufficient for fruit infection of very young tissues that were “water-congested”. We have been clearly demonstrating in our research in 1994-95 and 1997-2002, that extended wetness periods and temperature are critical for the development of walnut blight epidemics. Other parameters such as wind and relative humidity are weather parameters that need to be more critically evaluated in the field under natural inoculum levels. Our research has yielded several years of environmental and disease data for describing the seasonal development of walnut blight. In 2000, we initiated XanthoCast™ as a forecasting model for walnut blight in cooperation with FieldWise Inc., a commercial company specializing in disease forecasting. The accumulation model utilizes wetness period duration and temperature for calculating the risk of disease based on current ambient conditions for each field weather station. In 2002, a five-day-advanced weather forecast and prediction of XanthoCast for individual weather stations was also provided on the website with satellite rain analysis. In comparative field studies the XanthoCast model reduced the number of bactericide sprays as compared to calendar-based applications and disease control was similar for both application timings. The model, however, needs to be further tested to evaluate its performance under different climatic conditions, e.g. highly conducive and less favorable disease conditions. Additional studies are also needed to investigate the function of the fruit stomata in the bacterial infection process of walnut. In general, stomata are known to be important entry points for initiating bacterial diseases of plants. Previous studies have been conducted to map the occurrence of stomata on walnut fruit but the kinetics of stomatal behavior of walnut in relation to rain, wetness events, wind, and other environmental factors (Turner, 1991) have not been evaluated. Thus, investigations were initiated for this research objective in 2003. Research is also needed to

obtain information on the differential resistance among walnut cultivars and accessions against walnut blight that could be used in breeding programs. Thus, we initiated studies in an existing variety plot at Kearney Agricultural Center. Irrigation treatments were used to increase the natural incidence of disease and data could be obtained in this first year’s evaluations. OBJECTIVES I. Evaluate the toxicity of alternative, non-copper based chemicals to X. campestris pv.

juglandis. Compare, in laboratory, greenhouse, and small-scale field tests, the toxicity and efficacy of protective treatments including commercially available natural products (Milsana, Serenade), biocidal oxidation agents (Oxydate, Zerotol) and bactericidal treatments (e.g., re-formulated DOW-01, DBNPA) for control of walnut blight as compared to zinc-containing and fixed-copper compounds.

A) Comparative efficacy of new bactericides using handgun and air-blast spray application methods in field trials on walnut using simulated rain systems at the Kearney AgCenter, UC Davis-Plant Pathology Field Station, and in commercial orchards in Butte and possibly Tehama Co.

B) Greenhouse evaluations of new materials using pepper/walnut plants inoculated with X. vesicatoria/X. juglandis and incubated on misting benches or in growth chambers.

II. Measure residues of the DOW material on fruit over time. A) Develop in collaboration with DOW Chemical Company a residue

procedure on walnuts for the DOW compound. B) Evaluate residues of the DOW material on walnuts at selected time

intervals after application.

III. Continue to evaluate disease development throughout the spring and monitor environmental parameters (e.g., wetness periods, temperatures, and relative humidity) that are conducive to bacterial infection of walnut tissues using dataloggers.

A) Continue to determine the reproduction potential of pathogen on plant surface using spiral plating technology for potential incorporation in the existing model that is based on leaf wetness and temperature.

B) Continued evaluation of an improved detection method of the bacterial pathogen using commercially available ELISA systems from Agdia and Adgen, and specific and non-specific PCR primers for DNA amplifications. These methods will be used for accurate diagnosis and detection of the pathogen in our research on population dynamics and bacterial isolation.

C) Investigate the kinetics of stomatal behavior and the importance of stomata as infection sites for the bacterial pathogen on walnut fruit.

IV. Continue to develop and evaluate Xanthocast as a model for forecasting the incidence

of walnut blight. A) Evaluate the automated model of Xanthocast with up to a 5-day forecast

included in the latest version.

B) Apply bactericide treatments based on the forecasting model to determine if the total number of applications can be reduced as compared to a weekly calendar-based program.

C) Continue experimental validation of the model in irrigation plots at Kearney AgCenter (KAC) and at UC Davis.

D) Evaluate walnut genotypes in simulated-rain field studies at KAC for natural host resistance to walnut blight.

PROCEDURES Evaluation of alternative bactericides for management of walnut blight - field studies. Trials were established in two experimental orchards in Solano (UC Davis) and Fresno Co. (Kearney Agricultural Center) and in one commercial orchard in Butte Co. In the Solano Co. trial on Hartley walnuts, applications were done on April 23, April 30, May 9, May 16, May 23, and May 30. In the Fresno Co. trial on Chico walnuts applications were done on April 9, April 18, April 22, May 1, May 7, May 14, and May 22. In the Butte Co. trial on Vina walnuts applications were done on April 9, April 17, April 23, May 1, May 7, May 15, and May 23. Treatments were applied using an air-blast sprayer (100 gal/A). Simulated rain from sprinkler irrigation was applied to all plots. In the Solano Co. trial, trees were irrigated with high-angle sprinklers for 4 h on May 16, May 23, May 23, and June 6. In the Fresno Co. trial, trees were irrigated with high-angle sprinklers for 4 h the following day after each treatment application. In the third trial in Butte Co. the natural rainfall during the experimental period of 3.12 inches was complemented by an additional total of 1.56 inches that was provided in three simulated rain treatments on April 1, May 16, and May 27. Incidence of disease was based on the number of infected fruit in a sample of 50-200 fruit for each of four single-tree replications. Phytotoxicity evaluations were based on a scale of 0 (healthy) to 4 (extensive injury). Data were evaluated using analysis of variance and least significant difference mean separation procedures or general linear model and LSD mean separation procedures of SAS 8.2. Evaluation of a single bud-break application with Kocide-Manex plus 0.5% BreakThru. A trial was set up in Solano Co. using two walnut varieties, Hartley and Chico, to evaluate a single bud-break treatment of Kocide-Manex (8 lb-58 fl oz) plus 0.5% BreakThru at 100 gal/A for walnut blight control. In addition, sub-plots of the orchard were established: disease in one sub-plot was allowed to develop under the natural local weather conditions, whereas the other subplot was irrigated for 4 h on May 16, May 23, May 30, and June 6. These irrigations were done to increase disease development. The bud-break treatment was applied on April 10, 2003 using a tractor-pulled air-blast sprayer. Fruit evaluations for disease on June 12 were based on 100 fruit for each of the four single-tree replications. Data were evaluated using analysis of variance and least significant difference mean separation procedures and LSD mean separation procedures of SAS 8.2.

Evaluation of an improved detection and identification method for isolates of Xanthomonas juglandis, X. fragariae, and X. vesicatoria using Xanthomonas-specific primers in DNA amplifications. Published PCR primers for identification of plant pathogenic strains of Xanthomonas (Leite et al., 1994) were evaluated using DNA from X. juglandis, X. vesicatoria,

and X. fragariae. PCR products were separated in agarose gels and visualized after ethidium-bromide-staining. Disease evaluations and environmental monitoring using dataloggers in commercial walnut orchards and weather data from Fieldwise and UCIPM-CIMIS. In commercial orchards located in Butte and Tehama Co., as well as the PMA sites in Sutter/Yuba, and San Joaquin Co. approximately 100 fruit in each of two to three single-tree replications were tagged and monitored periodically (every 7-10 days) for the development of walnut blight from mid-April to late June 2003. Fruit were carefully examined for lesions and positive evaluations were re-checked in subsequent evaluations and in isolations of sub-samples of infected fruit as described previously. Disease incidence was determined as the number of infected fruit per total fruit sample minus the missing fruit. Percent new walnut blight per day (PNB) was calculated as: PNB = [D/t*(C-M)]*100 where D = the number of newly diseased fruit, t = time expressed as the number of days since the previous evaluation, C = number of healthy fruit at the beginning of the evaluation, and M = the number of fruit missing in the evaluation period not previously recorded as blighted. Disease incidence as an accumulation of PNB was plotted with Xanthocast indices (1-day, 7-day, and seasonal accumulation) for the calendar dates of evaluation. In general in Butte and Tehama Co., physiological time (days after pistillate flower emergence) occurred approximately on April 1 for cv. Vina and Ashley. In over 50 sites of the Fieldwise Inc. network in the Sacramento valley, electronic sensors and dataloggers (Campbell Scientific or Adcon Telemetry) were used to monitor leaf wetness, temperature, relative humidity, and rainfall. Dataloggers were programmed to make readings every minute and to calculate quarter-hourly, hourly, and daily averages for each micro-environmental parameter throughout the evaluation period. Environmental data were downloaded and summarized as hourly and daily summaries. CIMIS environmental data was also downloaded from Butte and Tehama county stations for the same time period. Disease progress curves for 2003 were developed and compared to weekly and seasonal disease index values obtained from our model - XanthoCast V.481. A disease index for the season or for a re-setting weekly index was determined using parameters of the model and plotted over time (Gregorian date). XanthoCast V.481 develops a forecasting index based on duration of leaf wetness for three temperature scales. The forecast is 14 to 21 days in advance of actual disease based on a latent period for disease expression after infection has occurred. This is the accumulation model described in previous reports. A 17-day “yardstick” was used to measure time distances between onset and termination of disease infection periods and actual observation of symptoms of the disease. Kinetics of stomatal behavior and role of stomata as infection sites for X. juglandis on walnut fruit. Studies were initiated on stomatal behavior of walnut fruits. To determine the spatial distribution of stomata on the fruit surface, thin surface cuts were made with a razor blade from the stem end, middle, and blossom end of cv. Chico fruits. Tissue samples were examined under a microscope with a 40x objective and the number of stomata was determined for ca. 15 fields of view for each of 6 samples per fruit location. To investigate the stomatal opening status during dry and wet conditions, fruit samples were taken in an orchard before, during, and after

simulated rain events. Samples were fixed immediately in glutaraldehyde and were processed for scanning electron microscopy.

Evaluation and utilization of XanthoCast for forecasting and management of walnut blight. In orchards in Butte, Tehama, and Yuba Co. bactericide treatments (i.e. Kocide-Manex) were applied as calendar-based or XanthoCast-based programs. For the XanthoCast program, as infection periods occurred, bactericide treatments were applied and accumulation was delayed for 7 days. If no infection periods occurred, bactericides were not applied. The treatments were designed to evaluate and improve timing and to reduce the total number of applications of bactericide treatments. A dilute, handgun (400 gal/A) or an air-blast sprayer (100 gal/A) application of copper-maneb was used for treatments in the XanthoCast and calendar-based programs. Disease management results for this program are presented in the PMA (summarized in the PMA Report) and BMP projects (summarized in the Buchner et al. Annual Report) and in part in this report.

Development of 1- to 5-day forecasts using XanthoCast parameters. Forecasts using the XanthoCast model for predicting walnut blight were developed in conjunction with Fieldwise Inc. and Fox Weather. The goal of this research is to automate this process into one XanthoCast program by showing actual and forecasted indices up to five days ahead of time. One proprietary, microclimate-forecasting model was used by Fox Weather and was evaluated using the XanthoCast parameters to generate daily and seven-day XanthoCast indices for 1-, 2-, 3-, 4-, and 5-day forecasts. The forecasts were qualitatively and quantitatively compared to the actual XanthoCast daily indices for accuracy and precision in predicting the occurrence and magnitude of infection events during the spring season. Regression statistics were used to compare observed with predicted values. Forecasts were done for all of the Fieldwise, Inc. weather station sites. Six sites in Butte, Yolo/Solano, Tehama, and Yuba/Sutter Co. are presented in the results below and are compared to observed 7-day XanthoCast indices obtained during the walnut growing season.

Evaluate walnut genotypes in simulated-rain field studies at KAC for natural host resistance to walnut blight. For these studies, an existing walnut orchard at KAC was used. Dr. McGranahan’s program selected the genotypes and retains information on the phylogeny of each genotype. An overhead sprinkler irrigation system was installed between each row and irrigations were conducted weekly from mid-April to the end of May (4/10, 4/18, 4/24, 5/2, 5/8, 5/16, 5/24/03). This is the second year of the study. Seventy-five to one hundred fruit on each of four to five single-tree replications were evaluated for blight on July 1, 2003. Data were analyzed using ANOVA and LSD mean separation procedures of SAS version 8.2. RESULTS AND DISCUSSION

Evaluation of alternative bactericides for management of walnut blight - field studies. Incidence of walnut blight in 2003 was higher as compared to the previous few years. Still, simulated rainfall that was applied in three test sites caused disease to further increase and treatments could be tested under severe disease pressure. In the Fresno Co. trial where a mixture of copper-sensitive and copper-resistant strains of the pathogen was present disease incidence on trees treated with bactericides ranged from 2% to 11% as compared to the control at 19.3% incidence (Fig. 1). The two Cuprofix DF® treatments and the low rate of Zerotol significantly

reduced blight as compared to the control. All other treatments including DB 918-48C and Bioacumen were statistically intermediate in their performance.

In the second trial in Solano Co. with only copper-sensitive strains, Kocide 50DF/Manex, the two rates of Zerotol, and the high rate of DB 918-48C had a significantly lower disease incidence (<8.5%) than the control (25.5%) (Fig. 2). The low rate of DB 918-48C was not significantly different from the control. Among the treatments, Kocide 50DF/Manex and the two Cuprofix/Manex treatments had the lowest incidence of disease, ranging from 1.8% to 3.5%. The low rate of Zerotol resulted in 4%, whereas there was 8.5% disease using the high rate.

In the third trial in Butte Co., disease incidence at the first evaluation date on 5/28/03 in the untreated control was 40.8% in the sub-plot with only natural rain and 73.5% in the subplot with additional simulated rain (Fig. 3). Due to nut fall these values dropped to 30.9% and 57.3%, respectively, by the second evaluation date on 6/12/03. The efficacy of seven calendar-based treatments with DB 918-48C (high rate) or Zerotol (high rate) and of three Kocide-Manex treatments that were applied using the XanthoCast model was evaluated in the sub-plot with natural rain and compared to the untreated control. Disease was significantly reduced by DB 918-48C and by XanthoCast treatments with Kocide-Manex. In the sub-plot with additional simulated rain, seven calendar-based treatments of DB 918-48C and Zerotol (both at two rates) were compared to Cuprofix-Manex. At the first evaluation date on May 28, disease incidence was significantly reduced by the DB 918-48C or Zerotol treatments (20 to 28.3% disease as compared to 73.5% in the control). For Cuprofix-Manex, 8.1% disease was observed. At the second evaluation date on June 18, disease incidence was significantly reduced by the low-rate treatment of Zerotol (23.5%) or Cuprofix-Manex (8.7%) as compared to the control (57.3%) (Fig. 3). Thus, the alternatives were more effective under less conducive environments (natural incidence of rain) than under more conducive environments (simulated rain). Furthermore, they were only moderately effective against the disease as compared to the copper (Cuprofix or Kocide)-Manex treatments in any environment.

Thus, this year's studies indicated that under highly conducive environments the only consistent treatment was copper-Manex in controlling walnut blight in California. Potential alternatives such as the non-registered bactericide product DOW-DB 918-48C and Zerotol (peroxyacetic acid from acidified hydrogen dioxide) significantly reduced disease as compared to the control. Still, these treatments were only moderately effective as compared to copper-Manex under simulated rainfall. Our research has led to the identification of moderately effective bactericides for agricultural uses under favorable environments but still none can replace copper-Manex treatments with equivalent levels of disease management. Thus, no other material is immediately available for management of copper-resistant populations of the walnut blight pathogen in California. Alternative materials including oxidants (e.g., Zerotol) still may have promise and additional testing will be required to increase persistence of the materials under conducive environments for disease. Consistent results in multiple trials are needed. Trials done in 2003 will help in prioritizing treatments in trials planned in 2004 that will be again conducted under simulated rainfall. Research with the DOW bactericide will not proceed until a registrant is found (see ‘Measurements…’ section below for details).

Evaluation of a single bud-break application with Kocide-Manex plus 0.5% BreakThru. For a third year, a trial was set up in Solano Co. using two walnut varieties, Hartley and Chico, to evaluate a single bud-break treatment with Kocide-Manex plus 0.5% BreakThru at 100 gal/A for walnut blight control. In addition, sub-plots of the orchard were established: disease in one sub-plot was allowed to develop under the ambient weather conditions, whereas the other subplot was treated once a week for 3 weeks with high-angle sprinkler irrigation to simulate rainfall and to increase disease development. Each simulated rain event lasted 4 hr. In a factorial analysis, the variable variety was not significant, however the irrigation treatment was a significant factor (Table 1). The bud-break treatment significantly reduced the incidence of disease in the non-irrigated plot from 24.9% in the control to 12.1%. In the irrigation plot, however, there was no significant difference between the control and the bactericide-surfactant treatment. These data indicate that overwintering bacteria in buds cannot be eradicated with a single Kocide-Manex-BreakThru application at 100 gal/A and that under favorable conditions the pathogen populations and the disease will increase after the treatment. Thus, walnut blight control can only be effective if treatments are repeatedly applied during favorable environments and host susceptibility periods.

Measurement of residues of the DOW material on fruit over time. The DOW DB 918-48C bactericide was effective against walnut blight under low rainfall conditions and performed similarly to copper compounds, but neither could prevent the disease from developing. Thus, the DOW material does not eradicate the disease and is probably not very persistent. Under high-rainfall, the persistence of residues of the DOW bactericide was presumably too low to be effective. The non-persistence of the chemical, however, can also be considered a positive characteristic. The chemical quickly dissipates, potentially leaving no residues in the developing pear fruit. This aspect of the bactericide possibly could facilitate its registration by EPA if pursued by the manufacturer. At this time, however, animal feeding, environmental metabolite, and residue studies will need to be conducted to fully characterize the DOW biocide before any food use registration is permitted. Because walnut hulls are not used for human or animal food, state and federal regulators indicated that a Section 24C (Special Local Need Registration) would be permitted if a two-year metabolite study ($275,000) was done. As indicated in the introduction, the active ingredient of DB 918-48C is registered with the US-EPA in California as a water treatment. The additional studies, however, are required for use on an agricultural crop. At this time, DOW Chemical is not proceeding with the registration because of the estimated costs ($2-4 million) and time required (5-7 years) to complete the required studies. We are currently working with third-party pesticide companies that are interested in registering bactericides, however, no commitments have been made towards agricultural registration.

Evaluation of an improved detection and identification method for isolates of Xanthomonas spp. using Xanthomonas-specific primers in DNA amplifications. Two published DNA primer pairs were evaluated in PCR amplifications with DNA from isolates of X. juglandis, X. vesicatoria, and isolates of Xanthomonas from strawberry. Our results indicated that only DNA from X. vesicatoria yielded a specific and intense DNA band after amplification. Thus, none of the molecular methods evaluated to date, including the antibody-based ELISA assays that were tested previously, was able to detect the walnut pathogen. A reliable detection and identification method for X. juglandis would be important not only for accurate disease diagnosis, but also for epidemiological studies, where pathogen populations need to be accurately identified. Therefore,

a method stills needs to be developed, because cultural characteristics are generally difficult to assess in short time periods. Kinetics of stomatal behavior and role of stomata as infection sites for X. juglandis on walnut fruit. Investigating the kinetics of stomatal opening is important because stomata in general are known to be important entry points for initiating bacterial diseases of plants. No information, however, is available for walnut blight. Furthermore, there is little information on stomatal kinetics around rain events. In our studies, we first evaluated the spatial distribution of stomata on the walnut fruit. Using light microscopy, we found that the density of stomata was significantly higher at the blossom end than on other areas of the fruit. There were 0.5, 1.1, and 1.4 stomata per field of view at the stem end, middle section, and blossom end of the fruit, respectively. Subsequently, samples were taken in the field before, during, and after simulated rain events. These samples are being examined using scanning electron microscopy, and results are pending. Information from these studies may ultimately help in the optimal timing of bactericide-growth regulator treatments. Thus, if stomata are closed throughout rain events, infections periods may be avoided. This strategy needs to be evaluated experimentally in field trials.

Evaluation of the XanthoCast model and the 3- and 5-day forecasts from Fox Weather. The walnut blight-forecasting model XanthoCast was launched as a commercial product in 2001 in collaboration with FieldWise, Inc. The model is one of two that has been developed. The XanthoCast model is a host phenology, temperature mediated-wetness accumulation model that has been incorporated into the FieldWise website (www.fieldwise.com) for evaluation. This system uses on-line microclimate telemetry data to update the XanthoCast index daily throughout the Sacramento valley. Several base stations are located in selected counties. XanthoCast was available for a third year free service on the website and growers utilized it during the growing season. In 2003, we again conducted field trials to evaluate the accuracy of the model. In these studies with high disease pressure, no difference was observed in the efficacy of the calendar-based and XanthoCast-based programs in providing disease control in all sites where XanthoCast was followed correctly. Both programs significantly lowered disease as compared to the untreated control. Still, XanthoCast called for fewer spray applications (2-3) than the calendar-based programs (6 to 9 half sprays). Thus, the XanthoCast model provided current conditions, as well as up to five-day forecasts of walnut blight based on microclimate weather forecasts and satellite rain analysis from Fox Weather service. Table 2 shows the R-square values (correlation coefficients) for regressions of 1-day, 3-day, and 5-day forecasts on XanthoCast observed 7-day indices. In general, 1-day correlation coefficients were higher than 3-day and these were higher than 5-day values. Overall, Fox Weather was extremely accurate in all forecasts. A correlation coefficient of 1 is 100% accurate, whereas coefficients below 0.5 would demonstrate a poor relationship. Correlation coefficients for the 1-day forecast were between 0.70 and 0.92, for the 3-day forecast they were between 0.60 and 0.88, whereas values for the 5-day forecast they were between 0.63 and 0.87 for all twelve sites evaluated (Table 2). XanthoCast 1-day, 7-day, and cumulative indices, as well as the 3- and 5-day forecasts from Fox Weather are graphically shown for Davis (Solano Co.), Marysville (Yuba Co.), Durham and Cana Highway (Butte Co.), and Gerber and East Red Bluff (Tehama Co.) in Figs. 5-10A-C. XanthoCast was very accurate in predicting the cumulative incidence of blight at 32% with 33% observed on cv. Hartley at the

http://www.fieldwise.com/

Davis site on June 13 (Fig. 5). In the Marysville site, 20% blight was observed and 32% was predicted on June 13 (Fig. 6). In the Durham site, 58% blight was predicted and 16.8% was observed on June 12. In the simulated rain plot at the Durham site, however, 57% disease incidence was observed (Fig. 7) and thus, the Adcon datalogger of FieldWise that was located in this plot was affected by the simulated rain and accurately predicted the disease occurrence within 1% on cv. Vina. XanthoCast predicted 48% disease incidence in Tehama Co. where 35% incidence was observed on June 12th. In summary, XanthoCast did well in forecasting cumulative incidence of walnut blight in a wet year such as 2003, as well as in the previous dry years (2001 and 2002).

The model was also very responsive to predicting disease infection periods using the 7-day index. Using arbitrary 7-day index threshold values of 5 to 7 used in combination with “rising” or “falling” indices, Fox Weather 1- to 5-day forecasts of XanthoCast, and forecasted rain events, walnut blight infection periods (and application of bactericides) could be determined (Figs. 5-10). In the Davis site, three infection periods were identified in the weeks of April 21, May 1, and May 14 when 7-day index values were at 5, 12, and 6, respectively (Fig. 5). In the Marysville site, two infections periods were identified in the weeks May 1 and May 14 when 7-day index values were at 12 and 6, respectively (Fig. 5). In the Durham location, the Adcon and Campbell datalogger systems were in the same orchard. XanthoCast indices were very similar but not exactly the same. To illustrate similarities and differences between the equipment, Fig. 7 shows XanthoCast data from the Adcon system, whereas Fig. 11 shows data from the Campbell system. In Figs. 7 and 11, infections periods occurred from April 13/15 through May14/May 15 when 7-day index values were above 6. Differences were considered minor and were probably due to drying rates of wetness sensors and air movement in different locations in the tree and in the orchard between the two systems.

An example of using a spreadsheet for utilizing XanthoCast data for a specific weather station is shown in Table 3. Daily and 7-day Xanthocast indices were recorded for each day starting on April 1, 2003. A critical value of 7 was selected as a threshold value for initiating management practices. In 2003, a bactericide (Kocide-Manex) application was made on April 1 by the farm advisor (B. Olson) in Butte Co. XanthoCast data were not available at his date because traditionally we started the model at pistillate flower emergence. Based on this aspect of the model, an application was not needed. A pistillate flower emergence starting date for the model is supported by the lack of stomata on early developing fruit (i.e., lack of bacterial entry points into fruit) and, as I will describe later, the development of disease in the disease progress curves was beyond known latency periods for an April 1 infection period. Starting date is still a critical factor and the initial date for collecting microclimate data should be evaluated in more detail. The first application should have been applied when the 7-day index reached the threshold value of 6-7 on April 13 (Table 3). The next application should have been applied on April 26 (Table 3) when the Re-set 7-day index following the 7th day after a bactericide application reached 7. The actual application however was on April 21 because of impending rain (an excellent ‘common sense’ decision strongly supported by the 5-day Xanthocast forecast that projected an index of 13, Fig. 7C). The next application was not called for until May 8. The actual third application was made on May 1 probably because of the high 7-day index (Table 3 and Fig. 11). Overall, the three bactericide applications based on XanthoCast were in general very reasonable for April 1 to May 30. This program resulted in only 4.1% disease and was similar to the weekly

calendar spray program where 9 half-row applications were made in the spring. Both programs were significantly different from the 16.7 % and 57% disease observed in the ambient and simulated rain untreated controls, respectively.

In Tehama Co., the Gerber weather station was used during the season (Fig. 9) and the microclimate data from the orchard are shown in Fig. 12. Again the data were similar between the two stations. The XanthoCast 7-day index did not reach 6 until April 24 because of cold weather during early April rains. No application, however, was made by the farm advisor (R. Buchner) until May 1 despite the high forecasted XanthoCast 7-day index and a strong chance of rain forecasted in the Satellite Analysis supplied by Fox Weather. The May 1 application was done in between the two large rain events and another application was done a week later. Still, this two-spray program did remarkably well, resulting in approximately 50% disease control. The XanthoCast program resulted in 17% disease incidence, whereas the untreated control had a 34% disease incidence. The grower management program of 9.5 sprays resulted in 4% incidence of disease. In San Joaquin Co. the first two applications were considered optional and done based purely on forecasted rain (T. Pritcher, IPM Specialist). The XanthoCast 7-day index was low until April 20 when it reached 6. In this case, an application was made and a second one was made on May 7. Again the second application could have been made on May 1, 8 days after the first application if forecasted data was used. Still, this program with a reduced number of applications as compared to the calendar program resulted in 100% disease control and was similar to the grower program. In this plot, the untreated control resulted in approximately 10% disease.

In summary, XanthoCast is not completely being used correctly in combination with the 5-day forecasted XanthoCast 7-day index and satellite rain analysis that was provided by Fox Weather Service. Thus, this analysis was described in detail to help improve a potential user’s understanding and is not intended to be derogative toward any of the participants or users. The Butte Co. trial of XanthoCast came the closest to the intended idea of usage and had the best results regarding number of applications and disease control. Table 3 illustrated this example and indicated improvements in the decision making process.

In the Butte, Tehama, and San Joaquin plots, the early rains in late March and the first week of April did not extensively contribute to disease development. The rain events on April 12 & 13 did contribute to the initial epidemic with a slow increase in disease from late April to May 15. Using a 17-day latent period, the rain events that started on April 21 and extended to May 5 contributed to the logarithmic increase of the disease epidemic observed in the May 15 and May 28 evaluations on untreated trees in 2003 (Figs. 11-13). The importance of rainfall for disease development was again experimentally verified in 2003 in our simulated rain trials in Fresno, Solano, Butte, and Tehama Co. In Butte Co., natural rainfall during the experimental period was 3.12 inches and an additional total of 1.56 inches was provided in three simulated rain treatments on April 1, May 16, and May 27. As discussed above and as presented before in our 2001 and 2002 reports, increased rainfall resulted in a significant increase in disease incidence. This further validated the role of wetness in the model and also the role of moderate temperatures (12-18 C) as being most conducive. Temperatures outside of this range are markedly less conducive to disease even with high rainfall. In both the Butte and Tehama plots, three or two applications of simulated rain, respectively increased the incidence of disease by 5-6% above the incidence in

trees under ambient conditions by June 2003 (Fig. 14). The high levels of rainfall that occurred in late April and early May also increased incidence of disease in both ambient and simulated rainfall plots. Differences would be more dramatic in drier years (Adaskaveg - Annual reports of 2001 and 2002). Still, the start dates of simulated rain were different in both plots. The early-simulated rain in Butte Co. resulted in no increase in disease after a 14-21 day latent period after application (Fig. 14). The late applications of simulated rain in both plots were applied when temperatures were less conducive (e.g., in Zone 3 as shown in Figs. 11-12). In summary, simulated rain or rainfall early in the season or late in the season may not be conducive if temperatures are too cold (e.g., zone 1) or too warm (e.g., zone 3), respectively. Thus, determination of infection periods will be dependent on phenological stage of walnut fruit development and environmental conditions with the assumption that inoculum in older orchards is omnipresent. A strategy of applying two to three bactericides at bud break without taking into account wetness and temperature may be effective in wet years but would indiscriminately apply bactericides in dry years. Use of the XanthoCast model in combination with the five-day XanthoCast forecast from Fox Weather allowed discrimination between wet and dry years and for responding to seasonal wetness under ideal temperatures for disease development. In dry years, only one to two (e.g., 2001-02), whereas in wet years, only two to three (four) (e.g., 2003) bactericide applications were required for managing the disease effectively. In conclusion, walnut blight can be effectively managed with properly timed in-season applications prior to favorable environments that are dependent on host phenology, wetness and temperature following the XanthoCast model.

In 2004, we plan to continue to provide XanthoCast data through Ag Vise Inc., a Chico, CA-based company that purchased FieldWise in 2003, and 5-day forecasts by Fox Weather. XanthoCast is a very robust program and regional forecasts are planned based on a few strategically placed data loggers in the Sacramento Valley including Sutter/Yuba, Butte, and Tehama Co.

Evaluation of natural host resistance to walnut blight among walnut genotypes in simulated-rain field studies at KAC. Host resistance among walnut genotypes in simulated rain studies was evaluated at the KAC. As a further demonstration that repeated rain events over a range of ambient temperatures during the spring will lead to disease in any walnut orchard in California, we started simulated rain studies in 2002 and continued in 2003 in the walnut genotype plot at KAC. Overall incidence of disease was very low in 2002 when the study was initiated (Adaskaveg, Annual Report 2002) but in 2003 the disease increased on most genotypes (Fig. 4). A wide range of susceptibility was found among the genotypes. Payne was the most susceptible with 15% disease incidence, whereas Serr, Franquette, Cheinovo, and 90-28-30 have no detectable incidence of disease on fruit (Fig. 4). Other varieties such as Adams 10, Chandler, Chase D9, Hartley, Sinensis, Tulare, 91-28-1, and 76-80 were statistically intermediate between Payne and the four varieties that had no diseased fruit. We plan to continue to conduct simulated rain studies to provide wetness for repeated host-pathogen interactions (the disease triangle) under predictable increased levels of inoculum to evaluate genotype susceptibility.

ACKNOWLEDGMENTS Special thanks to the growers in Butte, Tehama, Sutter-Yuba, and San Joaquin Co. who allowed us to conduct our research in their orchards, to Janine Hasey, Carolyn Pickel, Terry Prichard, and Sara Goldman for evaluating the Butte, Sutter-Yuba, and San Joaquin PMA sites where XanthoCast was tested, Joe Grant for his evaluations of XanthoCast in San Joaquin Co., and to the chemical industry representatives who cooperated with us with during this research.

REFERENCES 1. Adaskaveg, J. E, et al. 1998. Annual Walnut Reports - 1998 Walnut Marketing Board of

California. Sacramento, CA 2. Adaskaveg, J. E., et al. 2001. Annual Walnut Reports - 2001 Walnut Marketing Board of

California. Sacramento, CA 3. Buchner, R. et al. 2001. Annual Walnut Report - 2001 Walnut Marketing Board of

California. Sacramento, CA 4. Lee, Y. -A., M. N. Schroth, M. Hendson, S. E. Lindow, X. -L. Wang, B. Olson, R. P.

Buchner, and B. Teviotdale. 1993. Phytopathology 83: 1460-1465. 5. Miller, P. W. and W. B. Bollen. 1946. Oregon Agric. Exp. Stn. Tech. Bull. 9. 107 pp. 6. Mulrean, E. N. and M. N. Schroth. 1981. Phytopathology 71: 336-339. 7. Mulrean, E. N. and M. N. Schroth. 1982. Phytopathology 72: 434-438. 8. Ogawa, J. M. and H. English. 1991. Diseases of Temperate Zone Tree Fruit and Nut Crops.

University of California, Division of Agriculture and Natural Resources, Oakland, CA. Publ. No. 3345. 461 pp.

9. Teviotdale, B. L., M. N. Schroth, and E. N. Mulrean. 1985. In: Walnut Orchard Management. Ed. by D. E. Ramos. Univ. of Calif. Coop. Ext. Publ. 21410.

Table 2. Summary of R-square values of correlations between XanthoCast observed 7-day indices and 1-, 3-, and 5-day predictions.

- Forecasts by Fox Weather -

Table 1. Evaluation of a single bud-break application with Kocide-Manex + 0.5% BreakThru for management of walnut blight in high-angle irrigated and non-irrigated

sub-plots in a Hartley and Chico walnut orchard in Solano Co. 2003.

Trees were sprayed with Kocide-Manex (8 lb-58 fl oz) + 0.5% BreakThru at 100 gal/A at bud break (4-10-03). Trees were irrigated with high-angle sprinklers for 4 h on 5/16, 5/23, 5/30, and 6/6/03. Fruit were evaluated for disease on June 13, 2003. Incidence is based on evaluation of 100 fruit for each tree. No significant difference was observed in disease incidence between varieties or in the interaction of varieties and irrigation treatment.

Treatment Irrigated Incidence (%) LSD

Control - 24.9 aKocide-Manex +

Breakthru - 12.1 b

Control + 38.4 aKocide-Manex +

Breakthru + 24.5 a

Treatment Incidence (%) LSD

Irrigated 31.7 a

Non-Irrigated 18.5 b

A.

B..

Location Forecasted 1 day Forecasted 3 day Forecasted 5 day

Arbuckle 0.7 0.6 0.63Artois 0.8 0.76 0.72

Cana Highway 0.86 0.79 0.74Davis 0.87 0.81 0.83

Dunnigan Hills (S) 0.91 0.88 0.87Durham 0.87 0.78 0.76Gerber 0.86 0.81 0.77

Marysville (N) 0.86 0.79 0.81Red Bluff (E) 0.79 0.74 0.71

Rio Oso 0.78 0.68 0.68Robbins (N) 0.92 0.8 0.77

Wilson Landing 0.8 0.75 0.69

Based on correlation statisitics from regression analyses between observed and 1-, 3-, and 5-day predicitions of XanthoCast 7-day indices.

Table 3. Spreadsheet method for utilizing XanthoCast data and determining timing of bactericide applications (ex. Butte Co. trial 2003).

* - Application done based on rain forecasts and not XanthoCast data.** - Application based on impending rain from weather forecast.

Date Daily XanthoCast

XanthoCast 7-day index

Recommended application &

7-day duration of activity

Re-Setting of threshold index

Actual applications

1-Apr Spray*2-Apr 1 1 23-Apr 1 2 34-Apr 1 3 45-Apr 1 4 56-Apr 0 4 67-Apr 1 5 78-Apr 1 69-Apr 1 6

10-Apr 1 611-Apr 1 612-Apr 1 613-Apr 1 7 Spray14-Apr 1 7 215-Apr 1 7 316-Apr 1 7 417-Apr 1 7 518-Apr 2 8 619-Apr 1 8 720-Apr 1 8 121-Apr 1 8 2 Spray**22-Apr 1 8 3 223-Apr 1 8 4 324-Apr 1 8 5 425-Apr 1 7 6 526-Apr 1 7 Spray 7 627-Apr 1 7 2 728-Apr 1 7 329-Apr 1 7 430-Apr 0 6 51-May 1 6 6 Spray**2-May 3 8 7 23-May 4 11 4 34-May 1 11 5 45-May 0 10 5 56-May 0 9 5 67-May 1 10 6 78-May 1 10 Spray 79-May 1 8 2

10-May 1 5 311-May 1 5 412-May 0 5 513-May 1 6 614-May 0 5 715-May 0 4 016-May 0 3 017-May 0 2 018-May 0 1 019-May 0 1 020-May 1 1 121-May 0 1 122-May 0 1 123-May 0 1 124-May 0 1 125-May 0 1 126-May 0 1 127-May 0 0 028-May 1 1 129-May 1 2 230-May 1 3 331-May 4 7 Spray 71-Jun 0 7 22-Jun 0 7 33-Jun 0 7 44-Jun 0 6 55-Jun 0 5 66-Jun 0 4 7

ControlDB 918-48C 128 fl oz + Bond 24 fl ozDB 918-48C 64 fl oz + Bond 24 fl oz

Zerotol 2700 ppmZerotol 1350 ppm

Cuprofix 20DF 15 lb + Manex 58 fl ozCuprofix 40DF 7.5 lb + Manex 58 fl oz

Bioacumen (1:19 dilution)

0 5 10 15 20Disease incidence (%)

Fig. 1. Efficacy of new bactericides and antibiotics for management of walnut blight on Chico walnuts in Fresno Co. 2003.

Treatments were applied on: 4-9, 4-18, 4-22, 5-1, 5-7, 5-14, and 5-22-03.Trees were irrigated with high-angle sprinklers for 4 h the following day after each treatment application. Disease was evaluated on 6-17-03. Disease incidence is the number of infected nutsper 50-100 nuts evaluated on each of four single-tree replications. No phytotoxicity was observed in any treatment.

aab

b

abab

ab

bb

Fig. 2. Efficacy of new bactericides and biological treatments for management walnut blight on Hartley walnuts in Solano Co. 2003.

ControlDB 918-48C 128 fl oz + Bond 24 fl ozDB 918-48C 64 fl oz + Bond 24 fl oz

Zerotol 2700 ppmZerotol 1350 ppm

Kocide 50DF 8 lb + Manex 58 fl ozCuprofix 20DF 15 lb + Manex 58 fl ozCuprofix 40DF 7.5 lb + Manex 58 fl oz

0 5 10 15 20 25 30Disease incidence (%)

Treatments were applied on: 4-23, 4-30, 5-9, 5-16, 5-23, and 5-30-03. Trees were irrigated with high-angle sprinklers for 4 h on 5/16, 5/23, 5/30, and 6/6/03. Disease was evaluated on 6-23-03. Disease incidence is the number of infected nuts per 100 nuts evaluated on each of four single-tree replications. No copper-resistant strains of Xanthomonas campestris pv. juglandis are present in this orchard.

a

abbc

bcc

bcbc

c

0 20 40 60 80

Control

Cuprofix 20DF 15 lb + Manex 58 fl oz

DB 918-48C 128 fl oz + Bond 24 fl oz

DB 918-48C 64 fl oz + Bond 24 fl oz

Zerotol 675 ppm

Zerotol 337.5 ppm0 20 40 60 80

Fig. 3. Efficacy of new bactericides and antibiotics for management of walnut blight on Vina walnuts in Butte Co. 2003

Treatments were applied on: 4-9, 4-17, 4-23, 5-1, 5-7, 5-15, and 5-23-03. Xanthocast model treatments were done on 3-31, 4-21, and 5-6-03. Natural rainfall during the experimental period was 3.12 inches and an additional total of 1.56 inches was provided in three simulated rain treatments on April 1, May 16, and May 27 in the sub-plot with simulated rain. Disease was evaluated on 5-28-03 and 6-12-03. No phytotoxicity was observed in any treatment.

a

ababc

cbc

bbc

bc

bc

ac

abc

0 5 10 15 20 25 30 35Disease incidence (%)

ControlDB 918-48C 128 fl oz + Bond 24 fl oz

Zerotol 675 ppmKocide 2000 6 lb/Manex 58 fl oz

0 10 20 30 40 50Disease incidence (%)

a

abb

Not evaluated

ab

ba

XanthoCast

A. With simulated rain

B. Without simulated rain

Fig. 4. Evaluation of walnut varieties for walnut blight susceptibility at Kearney Agricultural Center 2003.

76-8090-28-3091-28-1

Adams 10Chandler

Chase D9Cheinovo

FranquetteHartleyPayne

SerrSinensis 5

Tulare0 5 10 15

Disease incidence (%)Trees were irrigated weekly from mid-April to the end of May for 4-6 h with high-angle sprinklers. Natural disease incidence is the number of infected nuts per total nuts evaluated on each of four or five single-tree replications.

a

c

bc

bccbc

bc

cc

c

bc

bc

bc

5/28/03 6/12/03

A. XanthoCast indices for the Davis test site

!

14 Apr 28 Apr 12 May 26 May 9 Jun 23 Jun

Date

0

5

10

15

20

25

30

35

Xan

thoc

a st I

ndex

Xan_dayXan 7-dayXan-CumNatural incidence!


Date

0

5

10

15

20

25

Xant

hoca

st In

dex

7 Day Observed7 day forcasted 3 days

B. Davis XanthoCast 7-day indexforecasted 3 days in advance by Fox Weather

C. Davis XanthoCast 7-day indexforecasted 5 days in advance by Fox Weather


Date

0

5

10

15

20

25

Xant

hoca

s t In

dex


In Fig. A, actual disease incidence was 32.9%, whereas XanthoCast predicted 32%. In Figs. B and C , the gray line represents the algorithm used by Fox Weather to accurately and precisely predict XanthoCast values either 3 or 5 days prior to observed weather. R2 values describe the relationship between observed and predicted values.

R2 = 0.81

R2 = 0.83

Fig. 5. Percent blight, XanthoCast indices, and 3- and 5-day forecasts for

the Davis trial site (Solano Co.)- 2003 -


the Marysville trial site (Yuba Co.)- 2003 -

A. XanthoCast indices for the Marysville test site

!

!


Date

0

10

20

30

40

50

Xan

thoc

ast I

ndex

Xan_dayXan 7-dayXan-CumNatural incidenceIncidence-XanthoCast!

!


Date

0

5

10

15

20

25

Xant

hoca

st In

dex


B. Marysville XanthoCast 7-day indexforecasted 3 days in advance by Fox Weather

C. Marysville XanthoCast 7-day indexforecasted 5 days in advance by Fox Weather


Date

0

5

10

15

20

25

Xant

hoca

s t In

dex


In Fig. A, actual disease incidence was 20.3% (*), whereas XanthoCast predicted 39%. Management practices using XanthoCast resulted in 2.7% blight (x) (Courtesy of the PMA trials). In Figs. B and C , the gray line represents the algorithm used by Fox Weather to accurately and precisely predict XanthoCast values either 3 or 5 days prior to observed weather. R2 values describe the relationship between observed and predicted values.

R2 = 0.79

R2 = 0.81

A. XanthoCast indices for the Durham test site


Date

0

5

10

15

20

25

Xan

thoc

ast I

ndex


B. Durham XanthoCast 7-day indexforecasted 3 days in advance by Fox Weather

C. Durham XanthoCast 7-day indexforecasted 5 days in advance by Fox Weather


Date

0

5

10

15

20

25

Xant

hoca

st In

dex


R2 = 0.76

!

!

!

1 Apr 15 Apr 29 Apr 13 May 27 May 10 Jun 24 Jun

Date

0

10

20

30

40

50

60

Xan

tho c

ast I

ndex

Xan_day Xan 7-dayXan-Cum Natural incidenceIncidence-XanthoCast Incidence-simul. rain!

! !

R2 = 0.78


the Durham trial site (Butte Co.)- 2003 -

A. XanthoCast indices for the Cana Highway test site


Date

0

10

20

30

40

50

60

Xan

thoc

ast I

ndex

, Per

cent

Blig

ht Xan_dayXan 7-dayXan-Cum


Date

0

5

10

15

20

25

Xan

thoc

ast I

nde x


B. Cana HighwayXanthoCast 7-day indexforecasted 3 days in advance by Fox Weather

C. Cana Highway XanthoCast 7-day indexforecasted 5 days in advance by Fox Weather


Date

0

5

10

15

20

25

Xan

thoc

ast I

ndex


In Fig. A, daily, 7-day, and cumulative XanthoCast indices are shown. In Figs. B and C, the gray line represents the algorithm used by Fox Weather to accurately and precisely predict XanthoCast values either 3 or 5 days prior to observed weather. R2 values describe the relationship between observed and predicted values.

R2 = 0.79

R2 = 0.74

Fig. 8. XanthoCast indices, and 3- and 5-day forecasts for the Cana Highway

trial site (Butte Co.)- 2003 -

In Fig. A, actual disease incidence was 16.8% (*), whereas XanthoCast predicted 58% (influenced by simulated rain studies, see Fig. 11 where disease incidence was 57.4%). Management practices using XanthoCast resulted in 4.1% blight (x). In Figs. B and C , the gray line represents the algorithm used by Fox Weather to accurately and precisely predict XanthoCast values either 3 or 5 days prior to observed weather. R2 values describe the relationship between observed and predicted values.

A. XanthoCast indices for the Gerber test site

!

!


Date

0

10

20

30

40

50

60

70

Xant

hoc a

st In

dex

Xan_dayXan 7-dayXan-CumNatural incidenceIncidence-XanthoCast!!


Date

0

5

10

15

20

25

Xant

hoca

st In

dex


B. Gerber XanthoCast 7-day indexforecasted 3 days in advance by Fox Weather

C. Gerber XanthoCast 7-day indexforecasted 5 days in advance by Fox Weather


Date

0

5

10

15

20

25

Xant

hoca

st In

dex


R2 = 0.81

R2 = 0.77


the Gerber trial site (Tehama Co.)- 2003 -

A. XanthoCast indices for the East Red Bluff test site


Date

0

10

20

30

40

50

Xant

hoca

st In

dex

Xan_dayXan 7-dayXan-Cum


Date

0

5

10

15

20

25

Xant

hoca

st In

dex


B. East Red Bluff XanthoCast 7-day indexforecasted 3 days in advance by Fox Weather

C. East Red Bluff XanthoCast 7-day indexforecasted 5 days in advance by Fox Weather


Date

0

5

10

15

20

25

Xant

hoc a

st In

dex


R2 = 0.74

R2 = 0.71

Fig. 10. XanthoCast indices, and 3- and 5-day forecasts for the East Red Bluff

trial site (Tehama Co.)- 2003 -

In Fig. A, actual disease incidence was 34.6% (*), whereas XanthoCast predicted 64%. Management practices using XanthoCast resulted in 17.26% blight (x). In Figs. B and C , the gray line represents the algorithm used by Fox Weather to accurately and precisely predict XanthoCast values either 3 or 5 days prior to observed weather. R2 values describe the relationship between observed and predicted values.

In Fig. A, daily, 7-day, and cumulative XanthoCast indices are shown. In Figs. B and C , the gray line represents the algorithm used by Fox Weather to accurately and precisely predict XanthoCast values either 3 or 5 days prior to observed weather. R2 values describe the relationship between observed and predicted values.

Fig. 11. Natural and simulated rain, temperature, XanthoCast indices, and walnut blight disease progress for the Butte Co. trial

!

!

! !

!

!

! !

3/31 4/14 4/28 5/12 5/26 6/9

Date

0

15

30

45

60

75

90D

ise a

se in

cide

nce

(%)

0

5

10

15

20

25

30

Precipitation (mm

)/Temperatrue (C

)

Precip (mm) Avg. Air Temp. (C) Simulated rain (mm)XanthoCast daily XanthoCast 7-day indices Disease Inc. (%)!

Max. threshold

Min. threshold

Zone 1

Zone 2

Zone 317-day interval latent period XanthoC

ast index

Microclimate data from the Campbell (shown) and Adcon systems in the same orchard site were similar. The asterisks are additional applications of copper-Manex that XanthoCast called for but that were not done. A double asterisk is a better timing for the second actual application (arrows). Arrowheads indicate questionable bactericide applications that were applied because of rain forecasts at low temperatures.

Fig. 12. Natural and simulated rain, temperature, XanthoCast indices, and walnut blight disease progress for the Tehama Co. trial

!

!

! !! !

!

3/31 4/14 4/28 5/12 5/26

Date

0

15

30

45

60

75

90

Dis

ease

inci

denc

e (%

)

0

5

10

15

20

25

30 Precipitation (mm

)/Temperatrue (C

)

Precip (mm) Avg. Air Temp. (C) Simulated rain (mm)XanthoCast daily XanthoCast 7-day indices Disease Inc. (%)!

Max. threshold

Min. threshold

Zone 1

Zone 2

Zone 317-day interval latent period

XanthoCast index

*

Microclimate and XanthoCast data from the Gerber (Adcon) and the orchard (Campbell) (data shown) sites were similar. Arrows along the x-axis indicate actual applications of copper-Manex based on Xanthocast, whereas the *-symbol is an additional application of copper-Manex that XanthoCast called for but that was not done.

*

** *^

Fig. 13. Natural rainfall, temperature, XanthoCast indices, and walnut blight disease progress for the San Joaquin Co. trial

!

!

!!

!

3/31 4/14 4/28 5/12 5/26 6/9

Date

0

5

10

15

20

25

30

Dis

ease

inci

denc

e (%

)

0

5

10

15

20

25

30 Precipitation (m

m)/Tem

peratrue (C)

Precip (mm) Avg. Air Temp. (C) XanthoCast dailyXanthoCast 7-day indices Disease Inc. (%)!

Max. threshold

Min. threshold Zone 1

Zone 2

Zone 317-day interval latent period

XanthoC

ast index

Arrowheads, arrows, and the asterisk along the x-axis indicate actual applications of copper-Manex by the grower or based on XanthoCast (arrows and asterisk). The asterisk is an additional application of copper-Manex that XanthoCast called for but that was not done. Arrowheads indicate questionable bactericide applications that were applied because of rain forecasts at low temperatures and low 7-indices of XanthoCast. Courtesy of the PMA trials.

^^ *

Fig. 14. Natural and simulated rain and walnut blight disease progress for the Butte and Tehama Co. trials

3/31 4/14 4/28 5/12 5/26 6/9

Date

0

15

30

45

60

75

90

Dis

eas e

inci

denc

e (%

)

0

5

10

15

20

25

30

Precipitation (m

m)

Precip (mm)Simulated rain (mm)Ambient Disease Inc. (%)Simulated Rain Disease Inc. (%)

3/31 4/14 4/28 5/12 5/26 6/9

Date

0

10

20

30

40

50

60

Dis

ease

inci

d enc

e (%

)

0

5

10

15

20

25

30

Precipitation (m

m)

Precip (mm)Simulated rain (mm)Ambient Disease Inc. (%)Simulated Rain Disease Inc. (%)

Butte Co.

Tehama Co.

Arrows over selected precipitation bars indicate simulated rains. In Butte Co. simulated rains were applied on 4/1 (8.1 mm), 5/16 (17 mm), and 5/27 (14.5 mm). In Tehama Co., they were applied on 5/17 (6.1 mm), and 5/27 (9.1 mm).

EPIDEMIOLOGY AND MANAGEMENT OF WALNUT BLIGHTwalnutresearch.ucdavis.edu/2003/2003_357.pdf · walnut...

Documents

Transcript of EPIDEMIOLOGY AND MANAGEMENT OF WALNUT BLIGHTwalnutresearch.ucdavis.edu/2003/2003_357.pdf · walnut...