Abstract
Background and goals In plant disease epidemics, consistent visual symptoms are cues for identifying infections that, when monitored over time, reveal spatial patterns of disease. In the absence of diagnostic visual symptoms, disease may remain unnoticed, delaying necessary interventions. We studied the grapevine red blotch virus (GRBV)–Vitis vinifera pathosystem in three Napa Valley vineyard blocks. Due to cultivar (Sauvignon blanc) or viral co-infections, none of the blocks had visual symptoms characteristic of grapevine red blotch disease. In the absence of these visual cues, our aim was to reveal spatial patterns of viral infection by strategically selecting vines for diagnostic testing with loop-mediated isothermal amplification (LAMP).
Methods and key findings Adaptive cluster sampling was used to select vines for LAMP diagnostics and explore the extent of virus aggregation. In each study year (2021 to 2023), 3 to 4% of vines in each block were randomly selected and tested. A GRBV-positive (GRBV+) result triggered the selection of neighboring vines for diagnostic testing (i.e., the condition to adapt). Cluster outlier analysis and optimized hotspot analysis identified significant aggregations of GRBV+ vines in each block. We discerned block-specific spatiotemporal variability, which we principally attribute to local inoculum pressure and secondary spread.
Conclusions and significance Selecting vines for LAMP diagnostics with an adaptive sampling approach is effective for revealing spatial aggregations of GRBV+ vines in the absence of characteristic visual symptoms. This strategy can be employed to assess the risk of secondary spread and mitigate the negative effects of red blotch disease.
- adaptive cluster sampling
- confounding GRBD symptoms
- grapevine red blotch disease (GRBD)
- grapevine red blotch virus (GRBV)
- loop-mediated isothermal amplification (LAMP)
- spatiotemporal disease patterns
Introduction
Grapevine red blotch virus (GRBV), the causative agent of grapevine red blotch disease (GRBD), has emerged over the last decade as a consequential viral disease of winegrapes in North America (Sudarshana et al. 2015, Cieniewicz et al. 2020). GRBV is a member of the genus Grablovirus in the family Geminiviridae (Varsani et al. 2017), and all Vitis species, both cultivated and wild vines, are susceptible hosts (Bahder et al. 2016, Perry et al. 2016, Cieniewicz et al. 2018b, 2019, Hoyle et al. 2022). GRBV likely originated in North America (Reynard et al. 2022, Thompson 2022), where extensive propagation and unintended transport of infected plant material resulted in the spread of the virus throughout the continent (Krenz et al. 2014, Rumbaugh et al. 2021) and around the world (Krenz et al. 2023). Most recently, short-distance, secondary spread has been documented in California and Oregon, where spatial patterns of aggregated symptomatic vines combined with random, isolated infections suggests transmission by a mobile insect vector (Cieniewicz et al. 2017a, 2018a, Dalton et al. 2019, Flasco et al. 2023a). Notably, Spissistilus festinus [Say, 1830] (Hemiptera; Membracidae) has been shown to transmit GRBV and plays an epidemiological role in the spread of the virus (Flasco et al. 2021, 2023b, Hoyle et al. 2022).
GRBD poses multiple economic challenges to the wine industry, including suboptimal grape quality, decreased marketability (Hobbs et al. 2023), and vineyard removal at threshold levels of disease incidence. GRBV infection results in the disruption of berry ripening processes (Blanco-Ulate et al. 2017, Martínez-Lüscher et al. 2019), leading to poor wine quality (Pereira et al. 2021, Girardello et al. 2024) in both white (Girardello et al. 2020) and red wines (Girardello et al. 2019, 2024). These wines may exhibit vegetative flavors and aromas, lack black fruit flavors (Bowen et al. 2020), and have poor color and lower ethanol concentration (Girardello et al. 2019, 2020). Depending on vintage, some enological (Rumbaugh et al. 2024) and harvesting (Girardello et al. 2024) techniques may ameliorate the impact of GRBD on wine composition; however, the most widespread GRBD management strategy is the reduction of vineyard inoculum by removing infected vines or entire blocks of vines (Sudarshana et al. 2015, Cieniewicz et al. 2017b, Ricketts et al. 2017).
The biological context for GRBD management programs, including disease thresholds for vineyard removal, has been based on multiyear spatial studies tracking visual symptoms in black-fruited cultivars (Cieniewicz et al. 2017a, 2020, KC et al. 2022). Characteristic symptoms used for visual diagnosis include diffuse, red blotches on the leaf blade (Al Rwahnih et al. 2013, Sudarshana et al. 2015) that develop from veraison until leaf fall (Rohrs et al. 2023). If GRBV is the only viral infection present in a black-fruited cultivar, symptom monitoring can be an accurate and dependable technique to identify infections and determine the appropriate management response (Rohrs et al. 2023). However, visual symptom monitoring has limitations. In black-fruited cultivars with multiple viral infections, GRBD symptoms are visually indistinguishable from those of other viral diseases (Adiputra et al. 2018). White-berried cultivars often have less conspicuous symptoms, such as irregular chlorosis, necrotic areas, or leaf cupping (Sudarshana et al. 2015, Adiptura et al. 2018, Girardello et al. 2020), which can be mistaken for potassium or magnesium deficiency (Cieniewicz et al. 2017b, 2019) or grapevine leafroll disease (GLD) (Sudarshana et al. 2015). In some cases, symptoms may be entirely absent, as observed in Vitis vinifera Sauvignon blanc (Rohrs et al. 2024). Thus, accurate diagnosis of GRBV in white-berried cultivars and black-fruited cultivars with co-infections relies on diagnostic assays (Adiputra et al. 2018). Although not available for GRBV detection, enzyme-linked immunosorbent assay (ELISA) has been used to screen asymptomatic white cultivars for grapevine leafroll-associated viruses (GLRaV) as part of nursery or foundation plantings (Pietersen et al. 2013). Both ELISA and PCR have been widely used in the California Grapevine Registration and Certification program to screen large numbers of vines for regulated pathogens (Arnold et al. 2019). However, no studies provide strategies for GRBD monitoring or mitigation in commercial vineyards when visual assessments are unreliable. Consequently, in the absence of visual cues, assessments of disease incidence and spatial patterns may elude vineyard managers and delay tactics for intervention (Rimbaud et al. 2015).
This study aimed to enhance the ability to locate spatial aggregations of infected vines and guide GRBV mitigation in vineyards when accurate visual assessment is not possible. Given the tendency of GRBV to have an aggregated spatial distribution (Cieniewicz et al. 2017a, 2019, Dalton et al. 2019), we employed adaptive sampling, which is based on the increased likelihood of finding an infected individual in the vicinity of a known infected individual within a clustered population (Ojiambo and Scherm 2010). The loop-mediated isothermal amplication (LAMP)-GRBV assay was selected as the primary diagnostic tool, as it is less expensive than PCR (Romero Romero et al. 2019), more appropriate for routine monitoring of GRBV infections (DeShields and KC 2023), and has been adopted as an “in-house” testing tool in the Napa Valley (Rohrs et al. 2024). With adaptive cluster sampling and LAMP assay diagnostics, our goal was to capture the occurrence and location of random and aggregated infections. To evaluate the effectiveness of this approach in identifying spatial aggregations of GRBV-positive (GRBV+) vines, we employed cluster outlier and optimized hotspot analyses. Discerning spatial patterns of infected vines in vineyards lacking clear visual symptoms will enable timely and effective management of GRBV at the local and regional scale, ultimately safeguarding grape quality and the wine industry’s economic viability.
Materials and Methods
Study sites
Three commercial vineyard blocks were selected in Napa Valley, CA, each from a distinct American Viticultural Area (AVA). Two blocks were V. vinifera cv. Sauvignon blanc, located in the Oak Knoll (“Oak Knoll”) and Rutherford (“Rutherford”) AVAs, respectively, and the third block was V. vinifera cv. Cabernet Sauvignon, located in the St. Helena AVA (“St. Helena”). The vineyard study blocks were planted between 1994 and 2018 (Table 1) and were cordon-trained and spur-pruned. The block sizes were similar at 0.74 (Oak Knoll), 1.1 (Rutherford), and 1.2 (St. Helena) ha. Disease symptoms could not be reliably distinguished in the study blocks: there were no visible symptoms in the Sauvignon blanc blocks (Figure 1), and in the Cabernet Sauvignon block, the presence of single and mixed infections of GRBV and GLRaV-3 confounded symptom assessment (Figure 2). Virus status in each of the vineyards was unknown at the outset of the study, however, vineyard managers were concerned about potential secondary spread of GRBV from nearby blocks with known infections. During the study period, there was no management of GRBV-infected vines (i.e., vine removal), nor were insecticides applied to target the vector, S. festinus. The removal of free-living Vitis species from vineyard-adjacent habitats may reduce GRBV inoculum pressure (Hoyle et al. 2025), however, there are currently no further recommendations for habitat management practices that target S. festinus. The vector has a diversity of feeding hosts, including up to 60 different plant families and over 171 unique plant genera, found in or near vineyard sites in the Napa Valley (Hoyle et al. 2025). All other vineyard management protocols followed conventional practices established for the Napa Valley region.
Planting details of study blocks, named according to their location within an American Viticultural Area (AVA) of the Napa Valley.
Canopies of grapevine red blotch virus (GRBV)-positive (+) vines at the Rutherford study block (Sauvignon blanc) on 20 Sept 2021. The foliage appeared healthy and did not display any disease symptoms (i.e., interveinal chlorosis or necrosis) typically reported for white-berried cultivars infected with GRBV. All vines (A to D) tested GRBV+ with the loop-mediated isothermal amplification (LAMP) assay, and the canopies could not be visually distinguished from GRBV-negative vines.
Incidence of grapevine leafroll disease (GLD) in the St. Helena study block confounded the identification of symptoms associated with grapevine red blotch disease (GRBD). Less severe red leaf symptoms were difficult to attribute to either GLD or GRBD (A, B, D, E; 20 Aug 2021). Severe symptoms appeared to be more characteristic of GLD (C and F; 23 Sept 2022), but may have been masking GRBD symptoms.
Vine selection: An adaptive approach
A georeferenced map of individual vine locations was generated for each study site using ArcGIS Pro (ver. 3.03, ESRI, Inc.). From September 2021 to 2023, a single-stage adaptive cluster sampling approach was employed, allowing us to identify the extent of GRBV+ vines at the block level, to then sample from the neighborhood of the GRBV+ vines to assess aggregation patterns (Thompson 1990). First, all vines within a block were binned into discrete areas (Thiessen proximal polygons) based on their proximity to a yellow panel trap used to monitor S. festinus (Figure 3). Based on the total number of traps per block, this resulted in six areas per block in 2021, and eight in subsequent years. Next, eight vines within each Thiessen proximal polygon were randomly selected, in addition to one vine at a fixed location that corresponded to the yellow panel trap. The total number of traps per block and random vines selected per polygon was based on feasibility. Petioles were collected from each vine (experimental unit) and tested with the LAMP assay. A GRBV+ test result triggered the adaptive stage of sampling, where adjacent vines within and across the row, i.e., “nearest neighbors”, of the GRBV+ vines were also tested with the LAMP assay. In 2021, the petioles of first (immediately adjacent) and second (next-to-adjacent) neighbors were tested, whereas in subsequent years, only the four immediately adjacent neighbors were tested. This resulted in a variable number of vines sampled across the blocks, influenced by the number of GRBV+ vines detected in the first round of sampling. For all vines selected, once a vine tested GRBV+, it was excluded from subsequent sampling. Vines that previously tested negative were sampled with replacement and could be tested again if they fit the selection criteria. In the St. Helena block, only random sampling was conducted. In 2024, only the Oak Knoll block was sampled to further explore emerging spatial patterns.
Yellow panel traps used to monitor Spissitilus festinus were placed at eight discrete locations within each study block. From 2021 to 2023, the number of S. festinus per trap was recorded weekly from May through October. The total number of S. festinus captured from 2021 to 2023 is displayed next to each panel trap location. Based on their proximity to a yellow panel trap, vines were binned into discrete sampling areas (Thiessen proximal polygons). Each year of the study, eight vines within each polygon were randomly selected for loop-mediated isothermal amplification (LAMP) assay diagnostics, and a grapevine red blotch virus (GRBV)-positive (+) test result triggered the sampling of nearest vine neighbors. Vines that tested GRBV-negative (−) were sampled with replacement, along with the vines at the panel trap locations, which allowed for the repeated testing of a small subset of vines across multiple years. The locations of these vines are depicted for each study block. At the Oak Knoll and St. Helena study blocks, 12 and three vines, respectively, tested positive in Year 2 or 3 of the study after testing negative in Year 1. All vines that were tested across multiple years at Rutherford tested GRBV-.
Monitoring vector populations
Populations of S. festinus were monitored using six (2021) or eight (2022 to 2024) yellow panel traps per site. The traps (14 × 23 cm; Seabright LTD) were mounted on plastic supports and deployed at cordon height. The number of S. festinus per trap was recorded weekly from May through October.
LAMP-GRBV diagnostics in the Oak Knoll neighborhood
On the basis of the spatial patterns that emerged in the Oak Knoll block in 2021, a network of yellow panel traps (as described previously) was used to monitor S. festinus populations throughout the 2022 growing season in a neighboring Chardonnay block located on the western edge of the Oak Knoll block. On 9 Sept 2022, we investigated the presence of GRBV in the Chardonnay block. Vines were systematically selected for petiole collection and LAMP assay diagnostics. Every two to three rows, two vines were selected for diagnostics: one in the middle of the first half of the block and one in the middle of the second half of the block. This resulted in a total of 10 selected vines. Due to the high incidence of GRBD and associated impacts to fruit quality and vine health, the neighboring block of Chardonnay was removed following the 2022 growing season, halting any further monitoring of GRBV and S. festinus.
When additional spatial aggregations emerged in the Oak Knoll study block in 2024, GRBV presence was investigated in two neighboring blocks along the southern edge of the study block. One block, to the southeast, was a younger block (~5-yr-old) of Merlot with a low incidence of visual symptoms consistent with GRBD. The second block, to the southwest, was an older block (over 20-yr-old) with a high incidence of symptoms consistent with GRBD. On 13 Sept 2024, we walked the first three rows in each of these blocks that were closest to the edge of the study block, arbitrarily selected 10 symptomatic vines throughout the rows, and collected petioles from each vine for LAMP assay diagnostics.
Tissue collection for the LAMP assay and qPCR analysis
For all selected vines, six basal petioles were collected across the length of the cordon. Three petioles were analyzed with the LAMP assay and the remaining three were either kept for retesting, if needed, or submitted for real-time (RT) quantitative PCR (qPCR) analysis. Petioles were kept in a cooler during field collection, transferred to a refrigerator (4°C), and processed within 24 hr at the University of California Cooperative Extension, Napa. Petioles were collected in September of each year, surrounding harvest (Eichhorn-Lorenz 38 to 41) (Coombe 1995).
In 2021 and 2022, petiole samples from a subset of vines were sent to a cooperating lab at the University of California, Berkeley for RT-qPCR diagnostics of GRBV and GLRaV-3. The purpose of this testing was to compare GRBV diagnostic methods, therefore, the samples sent for RT-qPCR analysis were arbitrarily chosen from the total number of sampled vines. A higher incidence of GLRaV-3 was anticipated in the St. Helena block, therefore, 48 vines in 2021 and 43 vines in 2022 were selected for RT-qPCR analysis. At Oak Knoll and Rutherford, petiole samples from a subset of 50 or 51 vines, respectively, were analyzed with RT-qPCR. Samples from the adjacent blocks in the Oak Knoll neighborhood were not analyzed with RT-qPCR.
LAMP assay
Templates for the LAMP assay were prepared via the “pin-prick” method of tissue acquisition within 24 hr of field collection (Romero Romero et al. 2019). Each sample consisted of three basal petioles (Setiono et al. 2018, DeShields and KC 2023), and a sterile razor blade was used to expose the cross section of each petiole, which was lightly pricked three times with a sterile pipette tip (10 μL). For DNA elution, the pipette tip with the petiole tissue was placed in a prepared 5-mL centrifuge tube containing 10 μL of distilled water for at least 10 min. DNA templates were frozen (−20°C) until analyzed with the LAMP assay within 1 wk of sample collection.
Methods for the LAMP assay followed the standard protocols in Romero Romero et al. (2019). Primers were custom synthesized by IDT DNA Technologies, and the Warm Start Colorimetric LAMP 2X was sourced from New England BioLabs. Positive controls consisted of diluted GRBV DNA obtained from AL&L Crop Solutions, and negative controls consisted of leaf petiole material that had previously tested negative for GRBV. LAMP reactions were completed at 65°C for 35 min, and GRBV+ results were indicated by a pink-to-yellow color change in the colorimetric LAMP-reaction mix.
RT-qPCR
For tissue preparation, 0.05 g of petiole tissue was homogenized in a Precellys tissue homogenizer. Total nucleic acid (TNA) extraction was performed using the MagMax Plant RNA Isolation Kit (Thermo Fisher Scientific), omitting the DNase treatment. For the GRBV RT-qPCR, a reaction mix was prepared with 0.2 μL CXR dye, 10 μL GoTaq Master Mix (Promega), 0.6 μL primers, 5.2 μL nuclease-free water, and 4 μL of TNA (diluted 10×). The pair of 3V/4V primers from Setiono et al. (2018) were used, synthesized by IDT DNA Technologies. The cycling conditions were 95°C for 5 min, followed by 40 cycles of 95°C for 30 sec and 60°C for 30 sec. The cutoff Ct value for determining positivity was 36, and a melt curve analysis was done to confirm positive results.
The same TNA extracts were used to test for GLRaV-3. For the GLRaV-3 RT-qPCR, a 10-μL reaction was prepared with 2.5 μL TaqMan Fast Virus 1-Step Master Mix (Thermo Fisher Scientific), 1 μL primers and probes (Diaz-Lara et al. 2018), 4.5 μL nuclease free H2O, and 2 μL of TNA extract (undiluted). Primers and probes were synthesized by IDT DNA Technologies. Cycling conditions were 50°C for 5 min, 95°C for 20 sec, then 40 cycles of 95°C for 30 sec and 60°C for 30 sec. Plant samples with Ct values below 25 were considered positive.
Cluster outlier analysis
ArcGIS Pro (ver. 3.03, ESRI, Inc.) was used to run a cluster outlier analysis that employed the Anselin Local Moran’s I statistic to identify clusters of GRBV+ vines. This method effectively detects spatial autocorrelation by comparing the attribute values of each feature to those of its neighboring features. To run the analysis, GRBV+ vines were assigned a value of 1 and GRBV-negative (GRBV-) vines were assigned a value of 0. For each site, a separate analysis was performed for each year of non-cumulative test results. The spatial relationship was set as inverse distance, and each function was run for 499 permutations. The inverse distance is a weighted parameter and was selected because of the known spatial nature of GRBV spread; it is useful in cases where the spatial effect decreases with distance (i.e., virus transmission is more likely between neighboring vines than between distant vines). Combined with row standardization, the inverse distance parameter addresses the potential bias introduced by adaptive cluster sampling. Areas with statistically significant Local Moran’s I values were classified as either high-high clusters (aggregations of GRBV+ vines) or low-low clusters (aggregations of GRBV- vines). Additionally, outlier features (negative vines surrounded by a statistically significant number of GRBV+ vines or single GRBV+ vines surrounded by a statistically significant number of GRBV- vines) were identified, highlighting atypical observations. Outlying negative vines surrounded by GRBV+ vines indicated vines that were at a higher risk of becoming infected. Outlying positive vines indicated areas where the GRBV distribution was random, and there was lower risk of aggregated spread.
Optimized hot spot analysis
The optimized hot spot analysis tool in ArcGIS Pro was used to assess the spatial distribution of GRBV within vineyard blocks. The analysis was conducted cumulatively, starting with data from Year 1, then combining Year 1 and Year 2, and so forth, to capture changes in the number of GRBV+ vines over time. First, the cumulative data from tested vines were aggregated into a standardized grid for each vineyard, summarizing the counts of GRBV+ and GRBV- vines. These counts were then normalized by the total number of vines tested within each grid area. The grid size was determined by the vine density at each site: St. Helena (1.8 × 3.0 m) and Rutherford (1.8 × 3.4 m) were divided into 0.4 ha grids, while Oak Knoll (1.2 × 2.4 m) used 0.02 ha grids, resulting in ~56 to 70 vines per grid cell. Using these aggregated grid areas, an optimized hot spot analysis with ArcGIS Pro’s Getis-Ord Gi* statistic was performed to identify statistically significant areas of high (hot spots) and low (cold spots) virus incidence. This analysis was repeated for each of the three years in which the vines were tested.
Additional vines were tested at Oak Knoll in 2024, which allowed for a fourth year of analysis at that study site. Initially the hotspot analysis was run for the entirety of the grid at that site, resulting in a smaller but significant hotspot at the southeastern section of the block. However, a considerable increase in GRBV detections in the final year obscured the detection of lower-intensity outbreaks, therefore a second hotspot analysis was run on a subset of the larger grid, isolating a subset of the data in the middle portion of the block.
Results
GRBV detections and vector occurrence
GRBV+ vines were detected at all sites and we observed site-specific and interannual variability (Table 2). In the Oak Knoll block, the proportion of total GRBV+ vines detected annually increased from 14% in 2021 to 42% in 2024. In the St. Helena block, 19% of tested vines were GRBV+ in 2021 and 2022, and 29% in 2023. Conversely, in the Rutherford block, the total annual detections of GRBV varied between 4, 8, and 10% of tested vines, with no annual increase. Cumulatively, from 2021 to 2023, the proportion of GRBV detections was similar at Oak Knoll (26%) and St. Helena (25%), even though adaptive sampling was not conducted in the St. Helena block. Overall, substantially fewer GRBV-infected vines were detected in the Rutherford block from 2021 to 2023, both on an annual and cumulative basis. By 2023, 8% of the vines tested (cumulative) at Rutherford were GRBV+.
Proportion of randomly selected vines and neighbor vines that tested positive for grapevine red blotch virus (GRBV), using loop-mediated isothermal amplification (LAMP) at vineyard study blocks named according to their location within an American Viticultural Area of the Napa Valley: Oak Knoll (OK), Rutherford (RD), and St. Helena (SH). The proportion of GRBV-positive (+) vines are reported for each study year (2021 to 2024) as a count and percent (in parentheses) of the total in each category.
GRBV was detected in a greater proportion of neighbor vines than random vines during all study years at Oak Knoll and during two years (2021 and 2023) at Rutherford. For example, in the Oak Knoll block, the proportion of neighbor GRBV+ vines was 30 to 34% in 2021 to 2023, increasing to 47% in 2024. In comparison, the proportion of random GRBV+ vines at Oak Knoll increased from 6 to 28% in 2021 to 2023, and up to 35% in 2024. At Rutherford, the proportion of neighbor GRBV+ vines was 15% and 19% in 2021 and 2023, respectively, whereas the proportion of random GRBV+ vines during those two years was 7% and 6%, respectively (Table 2).
Vines that tested GRBV- were sampled with replacement unless they tested GRBV+, which allowed for the repeated testing of a small subset of vines across multiple years. We identified 12 vines at Oak Knoll, zero at Rutherford, and three at St. Helena that tested GRBV- initially, and GRBV+ in a subsequent growing season (Figure 3). Of the 12 vines at Oak Knoll, there were two newly positive in 2022, four in 2023, and six in 2024. In the St. Helena block, all three vines were newly positive in 2023.
At the Oak Knoll site, the locations of infected vines suggested sources of inoculum external to the study block, and diagnostic testing confirmed this observation. In the Chardonnay block bordering the western edge of Oak Knoll, 100% of the samples (10/10) tested GRBV+ in 2022. In the Merlot blocks bordering the southern edge of Oak Knoll, 80% of the samples (16/20) tested positive for GRBV in 2024.
From 2021 to 2024, populations of S. festinus were monitored weekly from May through October. Most S. festinus were captured in traps from June to August, with the greatest abundance detected in July. S. festinus was detected in yellow panel traps in the Oak Knoll and St. Helena blocks, but not in the Rutherford block (Table 3). The greatest number of S. festinus were collected on an annual and cumulative basis at the Oak Knoll site, where the vector was detected in both the study block and adjacent Chardonnay block.
Total number of Spissistilus festinus reported for each study year (2021 to 2024), captured annually in yellow panel traps deployed at study blocks named according to their location within an American Viticultural Area of the Napa Valley: Oak Knoll (OK), Rutherford (RD), and St. Helena (SH).
LAMP and qPCR diagnostic results
During the first two years of the study, LAMP and qPCR diagnostic results were compared for 25% and 32% of the total samples taken at Oak Knoll and Rutherford, respectively. The level of agreement between the two diagnostic methods for GRBV was high for both sites (Table 4). At the Rutherford site, the diagnostic methods were perfectly aligned at 20% GRBV+, whereas slightly more vines were detected with qPCR (46%) compared to LAMP (42%) at Oak Knoll. At St. Helena, 75% of the samples taken during the first two years of the study were analyzed with both diagnostic methods, and LAMP assay diagnostics detected 10% more GRBV+ vines than did qPCR testing. In the St. Helena block, 41% of the samples tested positive for GLRaV-3, and 10% of the samples tested positive for both GLRaV-3 and GRBV (Table 4). There were no GLRaV-3 detections at Rutherford, and only one GLRaV-3 detection at Oak Knoll.
Loop-mediated isothermal amplification (LAMP) assay diagnostic results compared to qPCR testing results for grapevine red blotch virus (GRBV) at study blocks named according to their location within an American Viticultural Area of the Napa Valley: Oak Knoll (OK), Rutherford (RD), and St. Helena (SH). The qPCR diagnostic results are also reported for grapevine leafroll-associated virus 3 (GLRaV-3) and co-infections of GRBV and GLRaV-3 (GRBV + GLRaV-3). For each diagnostic method, the number of positive detections followed by the percent of positive detections (in parentheses) is reported.
Cluster outlier analysis
At Oak Knoll, a significant GRBV+ cluster was first detected on the western end of the block in 2021, and clusters were repeatedly identified in this location each subsequent year. GRBV spread on the western end of the block was evident in 2022 and 2023, when the area encompassing the GRBV- clustered vines expanded from 0.04 to 0.08 ha, respectively (Figure 4A). Spread was also reflected by the number of vines within a GRBV+ cluster, which increased from six to 13 from 2021 to 2022, and up to 22 vines in 2023. The western end GRBV cluster was detected again in 2024, and although the cluster encompassed a smaller area (0.02 ha), virus spread was identified in different sections of the block. Specifically, two new GRBV+ clusters were identified in 2024: one larger cluster in the middle section of the block, and a smaller cluster at the eastern end of the block (Figure 4A). The two largest GRBV+ clusters (west end and middle) were spatially separated by a significant GRBV- cluster, which reinforced the distinct locations of the new GRBV+ clusters.
A cluster outlier analysis using Anselin Local Moran’s I statistic identified significant spatial clusters of grapevine red blotch virus (GRBV)- positive (+) and negative (−) vines, and associated outliers across study blocks (Oak Knoll, Rutherford, and St. Helena). GRBV+ outliers indicate infected vines surrounded by more healthy vines (GRBV-) than expected given a random distribution. GRBV- outliers indicate healthy vines surrounded by more GRBV+ vines than expected given a random distribution. Data were analyzed non-cumulatively. Adaptive cluster sampling was used to select vines for loop-mediated isothermal amplification (LAMP) assay diagnostics carried out each year at Oak Knoll and Rutherford; vines were selected at St. Helena using a random approach. Grid size for Oak Knoll (A) was 0.02 ha (66 vines per grid). Grid sizes for Rutherford (B) and St. Helena (C) were 0.04 ha, with 56 and 70 vines per grid, respectively.
At Rutherford, a significant GRBV+ cluster was consistently identified within the same 0.04 ha grid from 2021 to 2023 (Figure 4B). The location of this aggregation in the northwest corner of the block was reinforced by the detection of a GRBV- cluster in the middle and southwest part of the block (2022, 2023). In contrast to the other sites, cluster outlier analysis did not indicate GRBV spread. The number of GRBV+ vines that constituted a cluster did not increase over time, and the area encompassed by the vines within a cluster also did not expand over time.
At St. Helena, cluster outlier analysis detected significant GRBV+ clusters each year of the study, even though only one round of vine selection (random) was conducted. In 2021, a GRBV+ cluster was detected in the southeastern corner of the block, and in subsequent years, significant clusters were detected in the same location with additional spread into other grids (Figure 4C). For instance, from 2021 to 2023, the area encompassing a GRBV cluster increased from 0.04 to 0.2 ha. GRBV spread was also reflected by the number of vines that constituted a significant cluster, which increased from two in 2021 to eight in 2023. The location of the GRBV+ clusters in the southeastern corner of the block was reinforced by the 2023 detection of a GRBV- cluster in the northern half of the block.
In summary, significant clusters of GRBV+ vines were detected each year of the study at each of the study blocks, regardless of vine selection approach (random only or adaptive cluster sampling). The GRBV+ clusters were denser and more defined at Oak Knoll (Figure 4A) and Rutherford (Figure 4B), where the nearest neighbors were sampled during the adaptive stage of vine selection.
Optimized hotspot analysis
Statistically significant hotspots were detected at all three vineyard sites, although the year of detection, hotspot statistic, and hotspot size varied across sites. Where hotspots were identified in consecutive years, they were detected in the same location. Oak Knoll was the only site where a hotspot emerged in a new location during the course of the study.
A significant hotspot was identified at the Oak Knoll block by 2022, the second year of the study, detected after testing 190 cumulative vines, or 8% of the block (Table 2). In 2022, the Oak Knoll hotspot encompassed 0.12 ha, with the statistically strongest (99%) aggregation of vines in a 0.04 ha section and a less significant (90 to 95%) aggregation in a 0.08 ha section (Figure 5A). After vine selection and testing in 2023, an increased number of GRBV+ vines was detected, which included more GRBV+ vines within the hotspot area. Consequently, the section of vineyard with the statistically strongest (99%) aggregation of GRBV+ vines increased from 0.04 (2022) to 0.08 ha (2023) (Figure 5A). The 2023 hotspot included a total of 49 GRBV+ vines.
Cumulative spatial locations of grapevine red blotch virus (GRBV)-negative (−) and positive (+) vines, and GRBV hotspots for each study block (Oak Knoll, Rutherford, and St. Helena). The optimized hotspot analysis identified statistically significant areas of high disease incidence using the Getis-Ord Gi* statistic (hotspot confidence). After 2021, the analysis was based on cumulative data and includes testing results from previous years. Adaptive cluster sampling was used to select vines for loop-mediated isothermal amplification (LAMP) assay diagnostics carried out each year at Oak Knoll and Rutherford; vines were selected at St. Helena using a random approach. Grid size for Oak Knoll (A) was 0.02 ha (66 vines per grid). The additional 2024 Oak Knoll hotspot analysis was limited to the middle 30 grids, as delineated by the black rectangle. Grid sizes for Rutherford (B) and St. Helena (C) were 0.04 ha, with 56 and 70 vines per grid, respectively.
Following the 2024 sampling at Oak Knoll, a new GRBV hotspot was detected that encompassed 0.12 ha in the middle of the block (Figure 5A). This aggregation of infected vines emerged over several years. For example, we first detected two GRBV+ vines in the middle section of the block in 2022, which increased to eight and 25 in 2023 and 2024, respectively.
In the Rutherford block, a GRBV hotspot was not detected until the third year of the study (Figure 5B). Relative to the other sites, a greater cumulative number of vines (223) representing a larger proportion of the block (12%) was tested before a significant (90%) hotspot was detected (Table 2). The GRBV hotspot encompassed the smallest area (0.04 ha) compared to the other sites and contained the fewest number (eight) of GRBV+ vines.
At St. Helena, we first identified a statistically significant hotspot in 2022 that encompassed a 0.16-ha section of the vineyard. The hotspot was detected after sampling 110 vines, or 5% of the block (Table 2 and Figure 5C). The entire hotspot, encompassing areas of 90 to 99% confidence, increased from 0.16 ha (2022) to 0.2 ha (2023). Additionally, the area with the highest hotspot confidence (99%) increased from 0.04 ha (2022) to 0.08 ha (2023) (Figure 5C). The 2023 hotspot included a total of 25 GRBV-infected vines.
The hotspot analysis was conducted on cumulative data. Accordingly, for a given year, the analysis included the current and all previous years’ location data for GRBV+ and GRBV- vines. Since vines that tested GRBV- were sampled with replacement, a small subset of vines was repeatedly tested across multiple years. However, the majority of vines that tested GRBV- were not retested in subsequent years, and the virus status remained GRBV- in the cumulative analyses. While this likely created a bias toward a higher incidence of GRBV- vines, significant hotspots of GRBV+ vines were still detected at each study block.
Discussion
Despite numerous studies documenting the spatiotemporal dynamics of GRBD based on visual symptoms in black-fruited cultivars (Cieniewicz et al. 2017a, 2019, Flasco et al. 2023a, Rohrs et al. 2023), there are no studies recommending methods to detect spatial patterns in commercial vineyards when symptom-based assessments are not possible. In blocks without characteristic visual symptoms, adaptive cluster sampling can be used to indicate GRBV levels and spread, locate and delineate aggregations of GRBV+ vines, and detect newly emerging aggregations of infected vines. Discerning spatial GRBV patterns is crucial for selecting mitigation efforts that are evidence-based and economically viable. Identifying the temporal and spatial patterns of GRBV+ vines can provide insights into the primary source of GRBV inoculum (Flasco et al. 2023a), define target areas to employ spatial roguing strategies (Atallah et al. 2015, Hesler et al. 2022), and help focus future observations (Rohrs et al. 2023). Spatial patterns can also alert vineyard managers to blocks that may be an inoculum source to surrounding vineyards (Flasco et al. 2023a).
Oak Knoll
Oak Knoll was the only site where multiple GRBV+ clusters and hotspots developed, offering the unique opportunity to confirm that adaptive cluster sampling can be used to identify emerging aggregations of infected vines and provide the impetus to investigate adjacent inoculum sources. A GRBV+ cluster was first detected on the west end of the block in 2021, and each consecutive year, a significant cluster was consistently detected in this location. The locations of the clusters overlapped with the 2022 and 2023 GRBV hotspots. Interestingly, the GRBV+ cluster and hotspot at the west end of the block were immediately adjacent to a block of Chardonnay where we confirmed a high incidence of GRBV. Although the Chardonnay block was removed in 2022, this proximal inoculum source likely initiated and fueled the spread of GRBV within the west-end hotspot. Vector-mediated transmission of GRBV between adjacent blocks has been documented by analyzing the distribution of phylogenic clades of GRBV in neighboring vineyards (Flasco et al. 2023a). Furthermore, populations of S. festinus were consistently detected at this site, and in greater numbers than at the other study sites. These conditions resulted in a strong likelihood of secondary spread (Cieniewicz et al. 2017a, 2018a, 2019, Flasco et al. 2023a), which was captured in our diagnostic testing and spatial analyses.
At Oak Knoll, two new GRBV+ clusters and a new hotspot developed in 2024. The largest cluster developed in the middle of the block, overlapping with the location of the 2024 GRBV hotspot (0.12 ha). The 2024 hotspot was detected after conducting an analysis that encompassed a subset of 30 grids, isolating the data in the middle portion of the block. This enhanced the sensitivity of the analysis, allowing for the identification of a more nuanced hotspot that might otherwise remain undetected. The emerging area of aggregated infections in the Oak Knoll block alerted us to a potential inoculum source along the southern edge of the study block, where an 80% positivity outcome for GRBV was confirmed. This inoculum source may also be related to a third GRBV+ cluster identified in 2024 at the eastern end of the block. The detection of additional clusters and a hotspot confirmed that adaptive cluster sampling can be used to locate emerging aggregations of GRBV-infected vines and initiate larger scale investigations of neighborhood epidemiological dynamics (Arnold et al. 2017, Atallah et al. 2017, Flasco et al. 2023a).
Rutherford
The GRBV incidence and spatial dynamics at Rutherford were very different from Oak Knoll. This is likely attributed to differences in local inoculum pressure and lack of vector presence, as populations of S. festinus were not detected in any year of the study. Rutherford had the lowest number of GRBV detections compared to the other study blocks, and there were no upward trends in annual GRBV incidence. Although significant GRBV+ clusters were identified each year of the study, Rutherford had the fewest number of vines in a GRBV+ cluster relative to the other blocks. The cluster locations were encompassed by the GRBV hotspot detected in 2023, which was the smallest hotspot (0.04 ha) of the study. The hotspot was detected after a larger number of vines (223) and larger portion (12%) of the block was tested compared to Oak Knoll and St. Helena. This indicates that blocks with a lower GRBV incidence will require more testing to detect smaller aggregations of infected vines.
It is likely that a neighborhood inoculum source affected the Rutherford study block, as the location of the GRBV hotspot and clusters were adjacent to a previously infected block that was removed in 2018 (Flasco et al. 2023a). This action appeared to effectively limit GRBV spread into the study block, evidenced by low annual GRBV detections, no indication of GRBV spread over time, and that vines sampled with replacement continued to test negative. For vineyard blocks with low GRBV incidence and limited spread, both spatial analysis techniques confirmed that adaptive cluster sampling can successfully detect significant aggregations of GRBV+ vines. Since the adaptive sampling approach incorporates information about the distribution of infected vines, this results in increased sampling efficiency over conventional designs (i.e., simple random or stratified sampling), especially in scenarios where disease is at low levels and spatially aggregated. In conventional designs, the detection of infected vines can be more of a chance event (Ojiambo and Scherm 2010). Our results indicate that annually testing 50 to 70 vines within an adaptive sampling framework is adequate for continued monitoring of GRBV incidence and spatial trends in this 1.1-ha block.
St. Helena
This is an older Cabernet Sauvignon block, where GRBD symptoms were confounded by the presence of GLRaV-3 (Adiputra et al. 2018). Diagnostic results revealed what could not be visually detected: in each year of the study, we uncovered an increasing number of GRBV+ vines, resulting in a 6% increase in the cumulative GRBV+ vines over the course of the study. The presence of S. festinus, along with newly detected infections in previously GRBV- vines, reinforces the likelihood of secondary spread in this block (Cieniewicz et al. 2017a, 2018a, 2019, Flasco et al. 2023a).
The spatial analyses resulted in the detection of GRBV-infected clusters and a hotspot, consistent with the characteristic, aggregated distribution of infected vines previously reported in spatial studies of GRBV (Cieniewicz et al. 2017a, 2018a, Dalton et al. 2019, Flasco et al. 2023a). We identified GRBV+ clusters in each year of the study, and these aggregations were encompassed by the hotspots detected in 2022 and 2023. Spatial analyses also indicated GRBV spread: the number of vines constituting a GRBV+ cluster increased from 2021 to 2023, and the GRBV hotspot area expanded from 2022 to 2023. Therefore, a single, annual stage of random vine selection (54 to 72 vines per year) for LAMP diagnostics was sufficient to detect GRBV levels and significant aggregations of infections in this block. One limitation of implementing a random-only selection approach is that it resulted in GRBV+ clusters with a lower density and less definition compared to Oak Knoll, whereas the adaptive sampling approach resulted in further delineation of GRBV+ clusters.
Management implications
Rutherford, Oak Knoll, and St. Helena represent three different vineyard scenarios with varying virus status, local inoculum pressure, vineyard age, and location. Management approaches will need to be tailored for each block, and the data gathered in this study can be used to inform mitigation strategies within the context of disease thresholds, block age (Ricketts et al. 2017), and background inoculum levels (Flasco et al. 2023a).
At Oak Knoll, the young age of the block (3 yr) and the initial incidence of GRBV+ vines in 2021 would have prioritized this block for immediate virus mitigation (Ricketts et al. 2017). If there had been early intervention when virus incidence was low, the removal of individual infected vines may have delayed the accumulation of virus inoculum in the block (Arnold et al. 2017). However, the spatial arrangement of GRBV+ clusters and hotspots within the study block in relation to adjacent, highly GRBV-infected vineyards indicates that neighborhood inoculum levels are consequential for virus transmission to this block (Arnold et al. 2017, Atallah et al. 2017, Flasco et al. 2023a). The mounting levels of virus spread and aggregation in the study block over time necessitates cooperative mitigation efforts to manage the continuous virus transmission from neighboring blocks (Atallah et al. 2015).
At Rutherford, the cumulative GRBV detection (8%) is within the recommended range for removing and replanting vines (Cieniewicz et al. 2017a), and well below thresholds for vineyard removal (Ricketts et al. 2017). Given the moderate age of the block (10 yr) and low levels of local inoculum, it is likely that virus mitigation strategies would be economically beneficial (Ricketts et al. 2017). The locations of the GRBV+ cluster and hotspot detected within the block could be used to fine-tune mitigation approaches; for example, given the single area of virus aggregation, vineyard managers might consider a spatial roguing approach where groups of adjacent vines are removed (Atallah et al. 2015, Hesler et al. 2022). Upward trends in annual GRBV incidence or vector populations were not detected in our study, and adoption of an annual testing program would be valuable in subsequent years to monitor GRBV distribution and guide vine removal efforts.
St. Helena is a 21-yr-old block with confounding disease symptoms related to co-infections of GRBV and GLRaV-3, and it is unlikely that investment in disease mitigation strategies would be economically beneficial (Ricketts et al. 2017, Hobbs et al. 2023). Evidence of significant clustering of GRBV+ vines, the development of a hotspot, and annual GRBV spread are central for understanding the contribution of GRBV to the overall disease status of the block. Previous studies have recommended a full vineyard replacement if the incidence of GLD and GRBD is above 25% or 30%, respectively (Ricketts et al. 2015, 2017). Within the framework of high GLRaV-3 incidence and the advanced age of the block, vineyard managers may consider block removal. However, if vine removal efforts are warranted, testing the nearest neighbors of GRBV+ vines could be used to delineate the GRBV+ clusters and define an area of vine removal. With an extended lag time between vector-mediated inoculation and symptom development (Flasco et al. 2023a), and the ability of pre-symptomatic vines to serve as a source of inoculum for secondary spread (Flasco et al. 2023b), a spatial roguing approach may be more effective (Atallah et al. 2015, Hesler et al. 2022).
Feasibility for practitioners
In all three vineyard scenarios, spatial analyses confirmed that adaptive cluster sampling can be used to capture GRBV spread and spatial aggregation. The methods for vine selection are practical, do not require sophisticated technology, and can be implemented by growers. LAMP diagnostics are feasible for assaying large numbers of vines: for instance, 70% of LAMP users in Napa Valley tested 600 to 1000 vines annually (Rohrs et al. 2024). In the current study, both LAMP and qPCR diagnostic methods detected a similar number of GRBV+ vines in samples from the Sauvignon blanc blocks, although at St. Helena, the LAMP assay detected a higher number of GRBV+ vines. LAMP is a highly sensitive assay that can outperform PCR and qPCR in virus detection (Romero Romero et al. 2019), but petioles must be sampled in later phenological stages (after veraison), when the virus is more evenly distributed in the host (DeShields and KC 2023, Rohrs et al. 2024). Over the last several years, there has been an increase in the adoption of the LAMP assay by Napa Valley growers, who report that LAMP is a highly useful decision support tool to make vine removal decisions, verify visual assessments, and confirm infections in asymptomatic cultivars (Rohrs et al. 2024). This study serves as an example of how to expand the use of the LAMP assay to identify spatial aggregations of GRBV in vineyards where symptoms are not consistent and visual assessment is not possible.
Future work
This study demonstrated that GRBV+ vines are spatially clustered and that multiple clusters can emerge in blocks with more than one neighboring inoculum source (e.g., Oak Knoll). Future work should investigate the application of adaptive sampling designs across varying degrees of clustering to attain unbiased estimates of virus incidence. In an adaptive sampling design using a quadrat as the experimental unit, virus incidence can be estimated in a population that has a clustered virus distribution (Ojiambo and Scherm 2010). For example, if a field is divided into equally sized quadrats that are randomly selected for counts of diseased plants, the condition to switch to adaptive sampling will be triggered when the number of diseased plants exceeds a predetermined threshold. When this condition is met, neighboring quadrats are then systematically sampled. Inevitably, samples obtained through the adaptive approach are biased, as the locations containing the species of interest are represented disproportionally. Following this, unbiased statistical estimators of population mean and variance can be obtained with the Rao-Blackwell theorem (Thompson 1990, Dryer and Thompson 2005). Adaptive sampling strategies can provide an effective method to estimate disease incidence in a clustered population and can result in a lower variance and higher relative precision than conventional sampling strategies (Thompson 1990, Ojiambo and Scherm 2010). If GRBD management programs use virus incidence thresholds as a decision framework for the removal of vines or whole blocks, estimators can then be used to guide management decisions in blocks that lack characteristic visual symptoms.
Conclusion
Pattern identification, both spatial and temporal, is critical to the study of disease dynamics and their underlying biological mechanisms. Previous research has relied on GRBD symptom expression to track the evolution of spatiotemporal disease patterns in black-fruited winegrape cultivars. However, GRBD symptoms may be indistinguishable in black-fruited cultivars with multiple viral infections, and symptoms may be subtle or absent in white-berried cultivars. With a lack of clear symptom expression, the presence of disease may elude growers, delaying necessary interventions and preventing the assessment of disease levels, spread, and spatial patterns. Given these challenges, this study demonstrates that adaptive cluster sampling and LAMP assay diagnostics can be used to discern spatial patterns of infected vines when visual assessments are insufficient, resulting in actionable and informed GRBD management decisions.
Footnotes
This work was funded by a grant from the California Department of Food and Agriculture, Pierce’s disease and glassy-winged sharpshooter program, Agreement Number 21-0273-000-SA. We are grateful to the Napa Valley grapegrowing community for their continued curiosity and supportive engagement that makes our work possible and relevant.
Rohrs JK, MacDonald SL, Fendell-Hummel HG, Brown A, and Cooper ML. 2025. Discerning spatial patterns of grapevine red blotch virus-infected vines in the absence of visually diagnostic symptoms. Am J Enol Vitic 76:0760009. DOI: 10.5344/ajev.2025.24067
By downloading and/or receiving this article, you agree to the Disclaimer of Warranties and Liability. If you do not agree to the Disclaimers, do not download and/or accept this article.
All data underlying this study are included in the article.
- Received November 2024.
- Accepted February 2025.
- Published online April 2025
This is an open access article distributed under the CC BY 4.0 license.
References
Vol 76 Issue 1
Jump to section
Related Articles
Cited By...
More from this TOC section
Similar Articles
