Abstract
Background Winter color changes in boxwood occur when foliage shifts from green to shades of yellow, orange, or red. The degree of winter color change in boxwood may influence the aesthetic value of ornamental plants and raise concerns for plant health care professionals.
Methods In 2021 and 2022, 40 cultivars of boxwood, including the species Buxus sempervirens, B. microphylla, B. sinica var. insularis, B. harlandii, B. balearica, and hybrids were evaluated for winter color change. In late summer of 2021, 29 plants representing 8 cultivars with intense color change were covered with 75% shade cloth tents or left in full sun from August 2021 through March 2022. In a separate experiment in fall 2021, B. sempervirens ‘American’, B. sempervirens ‘Rotundifolia’, and B. microphylla ‘Winter Gem’ were fertilized or not to determine if slow-release fertilizer could lessen the winter color change. The study was conducted on the transitional area between USDA Hardiness Zones 7b and 8a.
Results In 2021, B. sempervirens (61.0%) and hybrid (65.8%) cultivars had the highest percentage of visible color change, followed by B. microphylla (26.1%) and B. sinica var. insularis (19.1%). Buxus harlandii and B. balearica had negligible color change. Shade significantly (P < 0.001) reduced the winter color change observed in March 2022. Fertilization did not influence the visible color change (P = 0.2401).
Conclusions The differences in winter color change across boxwood species, cultivars, and under shaded and fertilized conditions will help horticulture practitioners better select cultivars and sites in the landscape relative to winter color change.
Introduction
Boxwood (Buxus spp.) is one of the most common woody landscape species found in ornamental plantings throughout much of temperate North America, as well as worldwide. Various species of boxwood are native to Europe, northern Africa, and Asia; however, no Buxus species are native to temperate North America (Daughtrey 2019). Despite their non-native status, boxwood are widely planted throughout the United States and have been a feature of formal North American gardens since the 17th century (Batdorf 2005). Boxwood are the highest-selling evergreen shrub in the United States. In 2019, the total value of boxwood sales in the USA exceeded $140 million (Hall et al. 2021).
Diseases and arthropod pests of boxwood are numerous, as are abiotic problems, which can lead to reduced ornamental value or plant mortality. Abiotic problems are most often related to growing conditions and include factors such as soil nutrient deficiencies, inappropriate soil pH, excess or deficient soil moisture, and exposure, i.e., planting in inappropriate locations (Batdorf 2005; Saunders Brothers 2018). Unfortunately, many abiotic factors that negatively impact health or appearance of boxwood are commonly misdiagnosed and “treated” with inappropriate methods. This unnecessarily increases the chemical input and potential negative ecological impacts of landscape management and ignores key components of integrated pest management (IPM), e.g., scouting or monitoring and accurate diagnosis of plant health problems (Kogan 1998).
One common abiotic “problem” observed on boxwood is the change in foliage color from typical green to yellow, red, orange, or bronze, occurring during the cool winter months and often referred to as “winter bronzing.” This phenomenon has been shown in boxwood to be highly plastic and will reverse without intervention with warmer temperatures or reduced light intensity during cool periods (Hormaetxe et al. 2007; García-Plazaola et al. 2008; Muñoz et al. 2021).
Despite abundant documentation to the contrary, it is very common for arborists and other landscape professionals to attribute winter color change, or winter bronzing, to nutrient deficiency and recommend fertilization without performing soil or foliar nutrient assessment (authors’ personal observation). This common practice increases chemical input, as well as costs to landscape managers or owners, and is contradictory to industry standards, which specify soil assessment should occur before specific soil management practices are recommended (TCIA 2018).
There were two primary objectives to this study: first, to document the degree of color change across 40 Buxus cultivars from 6 Buxus species or hybrids growing in full sun in a common garden with no apparent physiological stress factors; and second, to demonstrate that winter bronzing in apparently healthy Buxus spp. under appropriate growing conditions and ample nutrient availability is influenced by taxa and sun exposure, not by increased soil fertilization.
Materials and Methods
Boxwood Diversity Garden Planting
A total of 142 boxwood representing 40 cultivars/selections were sourced from commercial nurseries in 2019 (n = 107) and 2020 (n = 35) and planted in full sun in Charlotte, North Carolina, USA. This planting was set up as a demonstration garden where cultivar placement was not randomized across all individuals; rather, all replicates of a given cultivar were grouped together. The number of replicates of each cultivar ranged from 2 to 4 (Table 1). Of these 40 boxwood cultivars, 15 were Buxus sempervirens, 11 were B. microphylla, 8 were hybrids between B. sempervirens and B. microphylla, 3 were B. sinica var. insularis, 2 were B. harlandii, and 1 was B. balearica. At planting, cultivars varied in size due to nursery availability. Most of the plants were received as small, containerized stock (n = 90), while some were field-dug and much larger planting stock (n = 52). Root balls ranged from 4.5 L (1 gal) to large, 1.5-m tall balled-and-burlapped field-grown boxwood. Root balls were disturbed prior to planting by removing any circling roots with hand tools, and some nursery substrate was still intact. Plants were fertilized at planting with a slow-release 30-0-12 granular fertilizer at a rate of 7.3 kg/100 m2. Plants were irrigated during the summer for establishment when rainfall was insufficient, and supplemental irrigation equated to approximately 25.4 mm per week. All plants were mulched for moisture retention with a 5- to 8-cm layer of noncomposted arborist wood chips.
Winter Color Ratings
On 2021 February 22, April 2, May 11, and July 23, 142 individuals representing 40 boxwood cultivars were visually assessed for the percent of outer canopy with leaves showing nongreen coloration. At each of these periods, digital photographs of each individual cultivar were taken with a Canon Rebel T3i camera (Canon, Tokyo, Japan). In 2022, visual ratings and photographs were captured similarly on 2022 March 9. Of the boxwood rated in 2021, 16 individuals of 8 cultivars were used in the experiment investigating the effect of shade on winter color change. Ratings of these 16 plants were included in the 2021 but not the 2022 analysis of percent outer canopy showing color change. Means of the percent of outer canopy leaves with winter color change were analyzed by analysis of variance (ANOVA) for boxwood cultivar and species and separated by Tukey’s Honest Significant Difference (HSD) test in JMP 16 (SAS Institute, Cary, North Carolina, USA). Each visual rating date was analyzed independently.
Soil and Foliar Nutrient Analyses
On 2021 March 22, one to two soil cores to a depth of 15 cm were collected from each plant and submitted to Waypoint Analytical Laboratories (Memphis, Tennessee, USA) for nutrient analysis. The S3M Mehlich 3 soil extraction was used by Waypoint Analytical Laboratories to calculate general soil properties (pH, organic matter, cation exchange capacity, nitrogen, phosphorus, potassium, magnesium, calcium, boron, sulfur, iron, manganese, copper, zinc, and sodium). Additionally, on 2021 March 2 and July 26, foliar samples were collected by cutting 5 to 10 13-cm twigs from the exterior foliage of each plant. The samples were submitted to Waypoint Analytical Laboratories for the micronutrient and macronutrient (PT2) analysis. The foliar nutrient sample collection dates were chosen based on peak winter color change (2021 March 2) and complete color reversion (2021 July 26).
Multivariate analysis was conducted in JMP 16, and correlations were reported between each soil nutrient and percent canopy with winter color change as recorded on 2021 February 21 (peak color change) and the associated nonparametric P-value from a Spearman’s P-test. Similarly, foliar nutrients were analyzed by multivariate analysis with default settings in JMP 16, and correlations between foliar nutrient and percent canopy with winter color change as recorded on 2021 February 21 were reported with the associated nonparametric P-value from a Spearman’s P-test. For both soil and foliar nutrient analyses, the P-values were adjusted using Bonferroni correction to reduce the chance of spurious significance from the multiple comparison analyses.
Shade Effects on Winter Color Change
In 2021 August, 29 boxwood with greater than 40% winter color change in the outer canopy representing 8 cultivars/selections were chosen for use in the shading experiment. The selected cultivars had a 79% mean and 75% median color change in outer canopy in the 2021 ratings. These cultivars/selections were ‘American,’ ‘Buddy,’ ‘Dee Runk,’ ‘Fastigiata,’ ‘Green Mound,’ ‘Green Velvet,’ ‘Piney Mountain,’ and ‘Vardar Valley.’
Within each cultivar, individuals were randomly selected to receive the shade treatment (n = 16) or full-sun control (n = 13). Shade tents were constructed using tee posts and 70% shade cloth (Joepen, Guangzhou City, China) rated for 65% to 75% shade. Shade cloth was cut to size to drape the entire height of each plant, and the cloth was affixed to tee posts with zip-ties, ensuring the cloth was not in contact with the foliage. The nonshaded treatments were left as is in full sun. On September 13 at 3:36 p.m. and September 14 at 1:30 p.m., light intensity was measured with a Dr. Meter LX1330B Digital Illuminance Light Meter (Dr. Meter, Shenzhen, China) on the south-facing side of each plant to gauge the difference in light intensity measured in lux.
Shade tents were removed on 2022 March 9, and plants were rated as previously for percent outer canopy leaves with winter color change. These were assessed visually, and a representative plant from each cultivar-treatment combination was photographed. To capture the difference in color change between sun exposure levels and between years, the average difference in winter color change between 2021 and 2022 ratings for each individual were analyzed with a mixed model ANOVA in JMP 16, where shade treatment was a fixed effect and cultivar was a random effect in the model. Further, a regression analysis between the differences in winter color change between 2021 and 2022 and light intensity was analyzed in JMP 16.
Fertilization Experiment
On 2021 April 21, 12 of each Buxus sempervirens ‘American,’ B. sempervirens ‘Rotundifolia,’ and B. microphylla ‘Winter Gem’ were planted as described previously in a full-sun field in a randomized block design. These plants were sourced as 3.8-L containerized boxwood from a commercial nursery. Half of the plants for each cultivar (n = 6) were fertilized with a slow-release liquid 20-0-6 fertilizer by soil injecting 1 L on the east and west side of each plant (approximately 10 g N/m2), and the other half (n = 6) were left as nontreated controls. The fertilization treatment was applied on 2021 September 13 prior to any change in color. Color was visually rated as a percentage of the outer canopy with nongreen foliage on 2022 January 5 and March 9, and the average percent canopy with color change was used as the response variable. Due to non-normal data, the average percent canopy with color change was transformed with log (y + 1) for statistical analysis. The transformed data were analyzed with an ANOVA in JMP 16 with cultivar, treatment, and an interaction fixed effect to determine the response of fertilization as a preventive measure to lessen the degree of winter color change.
Similar to the common garden study, soil samples were taken on 2022 March 29 from all 36 boxwood to determine if soil nutrients varied across the fertilizer treatments as it related to winter color change. Soil nutrients and properties were analyzed by multivariate analysis with default settings in JMP 16, and correlations between soil property and average percent canopy with winter color change data were reported using the nonparametric Spearman’s correlation and P-value.
Weather Data
Cumulative measurements were compiled for the 2021 and 2022 growing seasons for the following parameters: solar radiation (W/m2), solar energy (kWh/m2), minimum temperature (°C), maximum temperature (°C), mean temperature (°C), cloud cover (%), visibility (km), and precipitation (mm). These data were queried from Visual Crossing (Hamburg, Germany), a commercially available, web-based database which uses a weighted average of daily data from up to 8 weather stations from within a < 50-km radius of the experimental location. Definitions of these metrics as defined by Visual Crossing are provided in Table S1. These daily averages are weighted based on proximity to the address zip code (Charlotte, North Carolina, USA 28278) with closer weather stations having more weight in the calculation. The query was performed with default settings on www.visualcrossing.com (data pulled on 2022 March 10). A table of weather stations and locations are provided in Table S2. The summation of each average daily weather parameter was used to compare the overwintering seasons (October 1 through March 10) in 2021 and 2022.
Results
Winter Color Ratings
At the peak of winter acclimation color change in our 2 years of visual ratings (2021 February and 2022 March) there were significant differences between boxwood cultivars and species in the intensity of winter color change when planted in full sun (Tables 1 and 2). In 2021, cultivars of B. sempervirens (61.0 ± 37.4%) and Buxus interspecific hybrids (65.8 ± 32.1%) had the highest degree of winter color change, followed by cultivars of B. microphylla (26.1 ± 27.6%) and B. sinica var. insularis (19.1 ± 24.8%). Selections of B. balearica (6.7 ± 2.9%) and B. harlandii (0.6 ± 1.8%) had very little to no winter color change in 2021. In 2022, there was a similar trend where cultivars of B. sempervirens (16.0 ± 27.2%) and Buxus interspecific hybrids (8.5 ± 13.8%) had the highest degree of winter color change, followed by cultivars of B. microphylla (3.0 ± 4.8%) and B. sinica var. insularis (0.9 ± 3.0%). Selections of B. balearica and B. harlandii had no winter color change in 2022. There was significantly less intense winter color change in 2022 compared to 2021.
Of the B. sempervirens cultivars, ‘American,’ ‘Aurea-Pendula,’ ‘Buddy,’ and ‘Longwood’ showed the highest level of winter color change, while ‘Elegantissima,’ ‘Variegata,’ and ‘Rotundifolia’ showed the lowest (Table 1). Of the Buxus interspecific hybrids, ‘Green Mound’ and ‘Green Velvet’ consistently had the highest level of winter color change, while ‘Green Mountain’ and ‘Independence’ had the lowest (Table 2). Of the B. microphylla cultivars, the only 2 cultivars that had higher than 50% of canopy with winter color change were ‘Golden Dream’ and ‘John Baldwin’; all others had less than 35% canopy with winter color change. Selections of B. balearica and B. harlandii had little to no winter color change in 2021 and 2022 (Table 2).
In 2021, color associated with winter acclimatization began to revert by April 2 for most of the boxwood (98%), and by May 11, winter-related colors were only observed in 12% of the 40 boxwood selections in this observational study (Figure 1). All boxwood with persistent winter color change on May 11 were cultivars of B. sempervirens, and the average percentage of canopy remaining with winter color was 9.5 ± 23.4%. Multiple ratings were not conducted in 2022.
Soil and Foliar Nutrient Analyses
The multivariate analysis of soil samples taken in 2021 comparing soil chemical properties to percent canopy with winter color change from the 2021 February 21 measurements yielded no significant correlations after adjusting the P-value using the Bonferroni correction (Table 3).
Similarly, most foliar nutrients were not correlated with percent canopy with winter color change from 2021 February 21 (Table 3). However, nitrogen (rho = −0.5, P = 0.002) had a weak negative correlation that was statistically significant in the multivariate analysis after adjusting the P-value with the Bonferroni correction (Table 3). After the foliar pigments reverted mostly back to green on 2021 July 23, there were no longer any significant correlations between any nutrient and the percent canopy with winter color change as recorded on 2021 February 21, indicating that the winter reduction in foliar nitrogen was transient and returned to growing season levels without intervention. Raw soil and foliar nutrient analysis can be found in Tables S3 and S4.
Shade Effects on Winter Color Change
Mean light intensity was 66,792 (nonshaded) and 3,949 (shaded) lux on 2021 September 13 taken at 3:30 p.m., and 114,431 (nonshaded) and 10,819 (shaded) on 2021 September 14 taken at 1:30 p.m. The mean change in percent canopy with winter color change between 2021 ratings and 2022 ratings was 31.9% in the full-sun treatment, which is indicative of year-to-year differences in color change, and 80.3% in the shaded treatment, a difference of 48.2% (P = 0.00007) that is indicative of the impact of shade and year-to-year difference on color change. In 2022, the mean percent difference between shade and full-sun treatments was 41% (P < 0.0001)(Figure 2), which is indicative of the impact of shade alone. There was a significant positive correlation between winter color change and light intensity in 2022 visual ratings, where color change as rated in 2022 March increased as light intensity measured the previous autumn (2021 September 13–14, 2021) increased (R 2 = 0.43, P = 0.0001). While there was a significant difference in winter color change intensity from 2021 and 2022, the difference between the visual ratings from each year were not significantly different (P > 0.05) within each of the boxwood cultivars.
Fertilization to Prevent Winter Color Change
There was no significant effect of fertilization (P = 0.24) nor was there an interaction between fertilization and cultivar (P = 0.79) on winter color change in B. sempervirens ‘American,’ B. sempervirens ‘Rotundifolia,’ and B. microphylla ‘Winter Gem’ (Figure 3a). There were significant differences across the two species in this trial, where selections of B. sempervirens (‘American’ and ‘Rotundifolia’) had significantly higher percent canopy with winter color change (P = 0.0025) compared to B. microphylla ‘Winter Gem’ (Figure 3b).
Soil fertility was not associated with winter color change for most soil properties in this controlled experiment, especially nitrogen (rho = −0.079, P = 0.648). There was a significant (P = 0.007) weak positive correlation (rho = 0.4) between soil iron levels and the percent winter color change across all experimental units.
Weather Data
There were 161 daily recordings of weather data in each of the overwintering years of 2021 and 2022 from October 1 through March 10. There were differences between the 2 years from each of the 2 overwintering seasons that are summarized in Table 4. Notably, the cumulative solar radiation and cumulative solar energy were less in the 2022 season compared to the 2021 season (by 9,945.6 W/m2 and 170.1 kWh/m2, respectively). Lastly, the cumulative max temperature was 209.4 °C warmer over the course of the 2022 season compared to the 2021 season.
Discussion
Winter Color Change vs. Winter Damage
It is important to note that the evaluations of cultivars in this common garden study focused on a reversible winter color change, as opposed to winter damage (Hormaetxe et al. 2007; García-Plazaola et al. 2008). ‘Winter damage’ occurs due to cold temperatures below the plant-adapted USDA Hardiness Zone, often combined with drying winds, and involves permanent discoloration and branch tip dieback to whole plant mortality (Le Duc et al. 2000; Batdorf 2005; Saunders Brothers 2018). The current study was conducted in USDA Hardiness Zones 7b–8a, well within the recommended zones for all species and cultivars.
Winter Color Change Evaluations
To our knowledge, this is the most extensive common garden study focused on evaluating winter color change across boxwood taxa. Sources such as the Boxwood Handbook (Batdorf 2005) and Boxwood Guide (Saunders Brothers 2018) mention winter color and Hardiness Zone within some, but not all, species/cultivar descriptions. The data presented here can aid landscape designers and homeowners in cultivar selection and placement in the landscape in relation to sun exposure for boxwood species and a wide variety of boxwood cultivars.
Perhaps the most important finding of this portion of the study was the significant differences in intensity of winter color change between species (Table 1). When considered at the species level, cultivars of Buxus sempervirens and interspecific hybrid Buxus cultivars showed the highest degree of winter color change compared to cultivars of other species; however there were exceptions within species (Table 2). For example, B. sempervirens ‘Rotundifolia’ exhibited very little color change, while B. sempervirenscultivars ‘Buddy,’ ‘American,’ ‘Aurea-Pendula,’ and ‘Longwood’ exhibited extreme color change involving all or nearly all foliage. This could be due to ‘Rotundifolia’ being cold hardy to USDA Hardiness Zones 5, while ‘Buddy,’ ‘Aurea-Pendula,’ and ‘Longwood’ are cold hardy to USDA Hardiness Zone 6; however, ‘American’ boxwood is hardy to Zone 5 and was among the selections with the most winter color change in our assessment.
Cultivars of B. microphylla and B. sinica var. insularis exhibited significantly less winter color change in our study compared to the B. sempervirens and interspecific hybrid cultivars. This contrasts somewhat with findings by Hawke (1994) in a study of 10 boxwood cultivars growing at the Chicago Botanic Garden. In that 8-year study, winter color change was most notable in B. microphylla cultivar ‘Fiorii,’ B. microphylla var. japonica cultivar ‘Green Beauty,’ and B. sinica var. insularis cultivar ‘Winter Beauty.’ In cultivars included in both Hawke’s study and ours, interspecific hybrids ‘Green Mound,’ ‘Green Mountain,’ and ‘Green Velvet’ were rated as “typically green” in winter by Hawke, while we found ‘Green Mound’ and ‘Green Velvet’ to be among the cultivars exhibiting the most dramatic degree of winter color change (Table 1). Contradictory to our results, Le Duc et al. (2000) found that at southern exposures, ‘Green Mountain’ and ‘Green Velvet’ exhibited less bronzing in winter than ‘Winter Gem’ (B. microphylla), while we found that ‘Green Velvet’ exhibited significantly more winter color change compared to ‘Green Mountain’ and ‘Winter Gem,’ both of which displayed similarly moderate color change. Differences in color change ratings may be due to the plants growing in USDA Hardiness Zones 5 (Hawke 1994), or 5 and 6 (Le Duc et al. 2000), while our study was conducted in Zones 7b–8a. In addition, the differences in winter light intensity due to differences in latitudes of study locations could have influenced the degree of observed color change.
Other factors that may have influenced the degree of winter color change in our study were weather differences between evaluation years and establishment time. In this study, the winter of 2020–2021 had cooler temperatures combined with higher levels of solar radiation compared to the winter of 2021–2022, and degree of color change was significantly higher in most cultivars in early 2021 compared to early 2022. Establishment time may have contributed to the differences in color change between the winters of 2020–2021 and 2021–2022. All boxwoods in these studies were recently planted (1 to 3 years prior to study) at the time of data collection, giving plants an additional year of establishment time at the 2022 data collection date. However, an extra growing season of establishment time did not appear to have influenced our results, as the ranked trends across species in the 2 years were largely the same. Furthermore, analysis of planting stock (i.e., field-dug or container)(P = 0.92) and year of planting (P = 0.62) had no significant effect on winter color change, while species (P < 0.0001) was always highly significant when analyzed together independently (data not shown).
Shade Effects on Winter Color
Shading of boxwood plants dramatically reduced the degree of winter color change in our study. While specific pigments were not measured in our study, this visual result was expected based on previous research specific to Buxus species, which demonstrated that pigment changes that result in visible color differences occur in response to increasing light intensity at low temperatures (García-Plazaola et al. 2000; Hormaetxe et al. 2004; Hormaetxe et al. 2007; García-Plazaola et al. 2008; Muñoz et al. 2021). In the landscape, planting boxwood that are significantly affected by winter pigment change (e.g., B. sempervirens or intergeneric hybrid selections) in part shade to shade could reduce this overwintering phenomenon and should be incorporated into design plans where evergreen function over the winter is desired or winter bronzing is unacceptable.
Fertilization to Prevent Color Change
Supplemental fertilization had no impact on the degree of winter color change in B. sempervirens ‘American,’ B. sempervirens ‘Rotundifolia,’ or B. microphylla ‘Winter Gem’ (Figure 3). In our experiment, the nonfertilized boxwoods were growing in soil with adequate nutrients, and supplemental fertilization did not impact color change. Fertilization may help reduce winter color change if plants are deficient; however, if nutrient supply is adequate, additional input will not have any effect. Thus, as stated in the ANSI-A300 standards for soil management, nutrient analysis should be conducted prior to application of fertilizer products (TCIA 2018). Although this controlled fertilization experiment was not replicated, we suspect that supplemental fertilization prior to wintertime will not reduce the winter color change of foliage.
Conclusions
Under favorable growing conditions, winter color change in boxwoods is a normal physiological phenomenon related to temperature, sun exposure, and species/cultivar, and is unaffected by additional nutrient input. When plants experience nutrient deficiencies, pigment shifts do occur and should be addressed only after accurate diagnosis of the underlying cause(s), which may be related to soil factors such as nutrient content, pH, moisture availability, compaction, or root health. Before additional nutrients are applied, suspected nutrient deficiencies based on foliage color should be validated by soil or tissue testing for nutrient content and other potential causal factors. The goal of any landscape professional should be to maintain ecological and aesthetic function of the managed landscape in the most environmentally sustainable manner possible. When plant “problems” arise, misdiagnosis or treatment without diagnosis of the underlying cause(s) leads to inappropriate chemical input. Understanding the underlying mechanism and species/cultivar variability of a common aesthetic concern in Buxus selections, the most common evergreen shrub in North America, is an important step in practicing sustainable landscape management. Furthermore, the sun exposure within a location of a landscape can significantly impact the winter color change and should be considered when boxwood are installed in a garden.
Conflicts of Interest
The authors reported no conflicts of interest.
Acknowledgements
The authors are grateful to the F.A. Bartlett Tree Expert Company for funding this work and the field support from Erika Wright, Caitlin Littlejohn, Amber Stiller, and Matt Borden. We are also thankful to Tom Smiley for reviewing the manuscript prior to submission.
Appendix.
- © 2023 International Society of Arboriculture