The severity of Bot. infection was calculated as the isolation frequency per site

Total canopy cover was thus the sum of percent GV and NGV, and dieback was calculated as the total percent NGV to reflect the severity of canopy-level symptoms across each site.Ten individuals within each of the 30 sites were randomly selected for sampling using the random points generator feature in ArcMap , for a total of 300 shrubs. Individuals were located in the field using a combination of a 1m resolution NAIP imagery base map , a GPS device, a laser range finder, and transect tape. For stands not located within a polygon, individuals were selected either using a transect tape and a point intercept method , or haphazardly selected within the accessible confines of the stand to provide an even distribution of sampling throughout the stand. Whenever possible, individuals located more than 2m from trails and fuel breaks were selected to avoid any edge effects. Individuals with any signs of pruning or other human damage caused by humans were not selected. All individuals were sampled once between April and September 2019. Two branchlets , each containing necrotic lesions and adjacent asymptomatic wood tissue, plant pots round were clipped per individual using sterile techniques, for a total of 600 samples. Samples were retrieved from approximately breast height and opposite sides of the shrub, whenever possible.

All individuals had at least two necrotic lesions, even if no significant dieback was observed, allowing these methods to be carried out across all 300 individuals. Samples were then placed in labeled plastic bags, stored on ice, and brought back to the lab and placed in a 4o C refrigerator.Samples were rinsed of dirt and debris and surfaced sterilized using 100% ethyl alcohol, 0.5% bleach, and a 70% ethyl alcohol rinse. Cross sections between 1-2 mm were isolated from the advancing canker margin and plated onto half-strength potato dextrose agar amended with streptomycin antibiotic. Cultures were incubated at room temperature until fungal colonies developed , and isolates of hyphal tips near the advancing margin were then re-plated onto half-strength PDA-strep to obtain pure cultures. From pure culture, any samples identified to have morphological characteristics consistent with those of Bot. fungi  were selected for PCR. A few isolates from cultures inconsistent with Bot. characteristics were randomly selected from each site to amplify and sequence to verify our morphotyping method. The internal transcribed spacer region 1 and alpha-elongation factor-1 genes were amplified using PCR primer pairs ITS1F and ITS4, and EF1-728F and EF1-986R, respectively, using methods modified from White et al., and Slippers et al . Successfully amplified samples were sequenced using the UC Berkeley Sequencing Facility .

Data were square-root transformed when necessary to meet the assumptions of normality. Differences in mean Bot. infection severity between elevation categories were calculated using one-way ANOVA with Tukey’s HSD for post-hoc analysis with R Statistical Software . Correlations between actual elevation and Bot. infection severity were assessed using simple linear regression and ANOVA to test for significance . Generalized linear models were developed to identify patterns of dieback, with dieback severity values as the response variable, and elevation , Bot. infection severity , and aspect as possible explanatory variables. If multiple models received substantial support , the best model was confirmed by calculating the relative importance of each term based on the sum of their Akaike weights . The proportion of variance explained by the models was calculated by measuring the adjusted D2 value .This study provides definitive support for the hypothesis that shrub dieback, during a recent drought, and pathogen infection are strongly related in a wild shrubland setting. This is the first known quantitative support for the hypothesis that in A. glauca, an ecologically important shrub species in the study region, dieback is related to pathogen infection occurring along an elevational gradient.

As expected, N. australe and B. dothidea were the two most frequently retrieved pathogens across all sites, however N. australe, the introduced pathogen, had almost twice the abundance of B. dothidea. N. australe is driving the correlation between elevation and Bot. infection, as the frequency was greater at lower elevations compared to upper elevations, while B. dothidea abundance did not change significantly across elevations. Level of Bot. infection was confirmed to be a significant predictor of stand-level dieback severity. The data also confirm that stand dieback severity is generally greater at lower elevations, which in this region experience higher temperatures and lower annual rainfall than the higher elevations sampled.While the presence of Bot. species has been reported previously in Santa Barbara County, this study represents the first effort to understand the abundance and distribution of Bots occurring in natural shrublands, and the first wild land shrub survey of Bots across a climate gradient. The high frequency and wide distribution of Bots retrieved from our study sites support the hypothesis that Bot. species are widespread across a natural landscape, and likely contributing to the extensive dieback resulting from the recent drought. Bot. fungi were retrieved from nearly every site in this study . We could not determine Bots. presence from three sites due to contamination issues. The broad extent of the study area suggests that infection is widespread in the region, and likely extends beyond the range of our study. While both N. australe and B. dothidea together made up the most frequently retrieved pathogens, our data show that N. australe has a larger distribution and occurs in greater abundance across the study region than B. dothidea . This trend was consistent across all elevations, but particularly at lower elevations . One possible explanation for this is that N. australe, being a recently introduced pathogen, spreads more rapidly as an exotic species in A. glauca compared to B. dothidea, which has been established in California for over 150 years . This hypothesis is consistent with previous studies that have shown variations in Bot. species abundance and virulence in Myrtaceous hosts occurring in native versus introduced ranges . However, it is difficult to evaluate the incidence of B. dothidea and N. australe in the present study in relation to historical documentation since many species in the Bot. complex have, until recently, been mischaracterized . Only with the recent development of molecular tools have researchers begun to accurately trace phylogenetic and geographic origins of Bot. species. Such studies are beginning to elucidate the complex existence of Bot. fungi as both endophytes and pathogens around the world, and much more research is needed to understand their pathogenicity in various hosts under different conditions. Nevertheless, large round garden pots it remains clear from our study that Bot. species, particularly N. australe, are both abundant and widely distributed in this region, and are important pathogens in A. glauca shrubs.Because Bot. taxa were the most frequently retrieved pathogens and were significantly correlated with dieback, we believe that they drive A. glauca dieback. Further, stand dieback severity increased significantly with Bot. infection. This is not to say that other pathogens do not also contribute to disease symptoms, but we found no evidence of any other pathogens occurring in such high incidence as Bot. species. While Brooks and Ferrin identified B. dothidea as a likely contributor to disease and dieback in dozens of native chaparral species during an earlier drought event in southern California, and Swiecki and Bernhardt found B. dothidea in association with a dieback event in stands of Arctostaphylos myrtifolia in northern California, our study yields the most extensive results of Bot. infection and related dieback in a chaparral shrub species across a landscape. Further, our study resolves species identity within the Bot. clade and highlights the role of the recently introduced pathogen, N. australe.A significant finding in this study was the relationship of Bot. infection and dieback with elevation. Bot. abundance and dieback were both found to be greatest at lower elevations, which was driven mostly by the high frequency of N. australe retrieved at these sites.

This represents the first quantitative evidence supporting that A. glauca vulnerability to fungal infection is influenced by stress levels along an elevation gradient. A similar pattern was observed in northern California by Swiecki and Bernhardt , who suggested that dieback in Ione manzanita infected with B. dothidea was greater in drier sites compared to more mesic ones, although no comparison of infection rates between sites was conducted in their study. The elevation gradient in our study was used a proxy for stress levels because annual precipitation decreases with decreasing elevation within our study region . Higher temperatures, which are associated with lower elevations, are also known to play an important role in drought-related mortality, as water loss from evapotranspiration is increased . Furthermore, unpublished data for dry season predawn xylem pressure potentials on a subset of sites along the same elevational gradient revealed more negative water potentials in A. glauca at lower elevations compared to upper elevations as spring and summer drought sets in . Thus, there is evidence that shrubs at low elevations indeed experienced the greatest water stress during the 2011-2018 drought, which predisposed them to higher levels of Bot. infection and enhanced dieback compared to upper elevation sites. More in-depth studies on the microbial communities and fungal loads of healthy and diseased shrubs throughout the region would help elucidate such trends. Another possibility for the higher incidence of Bot. infection at lower elevations is that the lower ranges of A. glauca populations in Santa Barbara are often located adjacent orin close proximity to agricultural orchards, ranches, and urban settings, which are common sources of plant pathogens, including Bots . Eucalyptus, avocado, and grapevines, which are abundant in these areas, are particularly well-known Bot. hosts and potential facilitators of Bot. introduction . Therefore, sources of inoculum from nearby populations of agricultural and horticultural hosts could be responsible for continual transmission Bots in wildland A. glauca populations, and would likely result in greater rates of infection at lower elevations. Furthermore, many of the lower sites in the survey were located near roads and/or trails, which are often subjected to additional stress from human activity like pruning and trail clearing; activities that are known to spread and promote infection by Bot. pathogens . While we avoided sites that showed signs of such activities in our survey, we cannot rule out the potential contributions of proximity to human encroachment to the overall higher rates of Bot. infection across the lower elevation zone. It is worth noting that while our study revealed a trend of increased dieback in lower elevations, some upper elevation sites also exhibited high levels of dieback, and Bot. fungi were retrieved from many of these sites. Upper elevations also experienced significant stress during the 2011-2018 drought, and water-related microsite variables outside the scope of this study like slope, solar incidence, soil composition, and summer fog patterns factors likely contributed to increased stress and subsequent dieback. Additionally, N. luteum, N. parvum, and D. sarmentorum were isolated primarily from upper sites. Host plants in these sites may serve as potential reservoirs for disease because the milder climate conditions promote greater host survival and thus pathogen persistence asendophytes. This serves as an important reminder that continued global change-type drought may eventually jeopardize susceptible species populations even at the upper boundary of their range.Our results are consistent with well-known theoretical models describing the relationship between environmental stress and biotic infection, which generally ascribe extreme drought stress as a mechanism for plant predisposition to disease . These frameworks illustrate dynamic interactions between environmental stress, plant hydraulic functioning and carbon balance, and biotic attack, and a growing body of research has focused on understanding the roles of these factors in driving plant mortality, especially during extreme drought . While the data collected in this study do not directly address the specific mechanisms leading to Bot. infection and dieback in A. glauca, our results can be discussed in the context of how life histories and physiological adaptions elicit differential responses to drought in woody plants, particularly in chaparral shrubs. For example, shallow-rooted, obligate seeder shrubs like A. glauca have been shown to be more susceptible to drought-induced mortality during acute, high intensity drought than deep-rooted, resprouter shrubs . This supports our observations of pronounced A. glauca decline during an historic California drought compared to nearby resprouter species like chamise , and laurel sumac .