Factors that influence measurements of sweet cherry (Prunus avium) flower bud cold hardiness obtained using differential thermal analysis

Abstract Differential thermal analysis (DTA) is a technique commonly used to evaluate the cold hardiness of plant organs that supercool as a means for cold survival. The aim of this study was to evaluate the effect of different pretest bud storage conditions, cooling rates, and bud excision techniques on dormant sweet cherry flower bud low temperature exotherms (LTEs) measured using DTA. Furthermore, this study compared cold hardiness estimates made using DTA and controlled freezing tests. We determined that buds stored at warmer temperatures (12.5 °C and room temperature) for 2–6 h prior to DTA or transported to the lab in a moist environment underwent biologically relevant changes in their apparent sensitivity to cold, as indicated by LTEs produced at warmer temperatures. The DTA cooling rate also significantly affected LTEs, with faster cooling resulting in the production of LTEs at warmer temperatures. Overall, LTEs were comparable among buds with varying amounts of plant material remaining attached to the bud base. It is important to note that the region directly subtending the primordia was always left intact on the buds being compared. This study demonstrated that overall, DTA and controlled freezing tests resulted in comparable measures of cold hardiness. The findings presented in this study are pertinent to researchers interested in conducting cold hardiness measurements in sweet cherry and highlight that consistency in DTA pretest conditions and bud preparation are required to achieve reliable LTE results that can be compared among studies.


Introduction
Temperature is one of the major limiting environmental factors affecting plant growth, development, and yield (Luo 2011).Low temperature conditions are one of the main abiotic stresses that limit plant distribution and crop production across the landscape (Charrier et al. 2015) and influence the suitability of locations for orchard establishment.Freezing events can be responsible for substantial economic losses (Snyder and De Melo-Abreu 2005;Snyder et al. 2005).Deciduous fruit trees, such as sweet cherry (Prunus avium), are particularly sensitive to this risk when grown in regions with welldifferentiated seasonal climates and can experience significant crop losses caused by freezing events (Chmielewski et al. 2018;Salazar-Gutiérrez and Chaves-Cordoba 2020).Due to the potential impact of cold damage on commercial sweet cherry production, research has aimed to: improve our understanding of sweet cherry flower bud cold hardiness (e.g., Andrews and Proebsting 1986;Kappel 2010;Salazar-Gutiérrez et al. 2014;Kose and Kaya 2022); generate methods to model and estimate flower bud cold hardiness (e.g., Andrews et al. 1987;Salazar-Gutiérrez and Chaves-Cordoba 2020;Houghton et al. 2023b); and evaluate potential ways to protect buds from freezing events (e.g., Snyder and De Melo-Abreu 2005;Arnoldussen et al. 2022).
Several methods have been developed to better understand the cold hardiness of plants.One of these is differential thermal analysis (DTA), which can be used to estimate lethally damaging temperatures in plant organs that have evolved mechanisms of supercooling to survive freezing conditions.The cold hardiness of several Prunus species that supercool has been measured using DTA including apricot (Prunus armeniaca) (Kaya et al. 2018;Kaya and Kose 2019;Kovaleski 2022), peach (Prunus persica) (Quamme 1974;Liu et al. 2019), Canadian plum (Prunus nigra) (Kovaleski 2022), black cherry (Prunus serotina) (Kader and Proebsting 1992), sour cherry (Prunus cerasus) (Mathers 2004), and sweet cherry (Salazar-Gutiérrez et al. 2014;Kaya and Kose 2021;Houghton et al. 2023a).To conduct DTA, flower buds are placed on thermoelectric modules (TEMs) and subjected to cooling temperatures (Salazar-Gutiérrez et al. 2014).When the internal supercooled water freezes, heat is released and is detected across the TEMs as a low temperature exotherm (LTE).The LTE is assumed to correspond to the temperature at which the buds become lethally injured (Quamme 1974;Palonen and Buszard 1997).High temperature exotherms (HTE) can also be produced during DTA and correspond with the formation of nonlethal extracellular ice (Kang et al. 1998).Controlled freezing tests are an alternative method to estimate flower bud lethal temperature.This method is commonly used during the later stages of flower bud deacclimation because DTA has been observed to be less reliable during this period (Salazar-Gutiérrez et al. 2014).Controlled freezing tests provide an alternative to DTA; DTA and controlled freezing tests are believed to provide comparable measures of the lethal temperature that causes 50% cold damage to sweet cherry flower buds prior to rapid bud deacclimation (Quamme 1974;Andrews and Proebsting 1987).
Previous studies have shown that the use of DTA to estimate lethal temperatures for supercooling-capable Prunus species flower buds can be influenced by sample preparation and test conditions, including pre-test sample storage temperature and moisture (Proebsting and Mills 1972;Proebsting and Sakai 1979;Andrews et al. 1983;Liu et al. 2019), sample size (Andrews et al. 1986), the presence of ice nucleators or ice nucleation (Ashworth 1985;Rajashekar 1989;Quamme et al. 1995), bud excision method (Quamme 1978;Proebsting and Sakai 1979), and the rate of cooling during analysis (Proebsting and Sakai 1979).A cooling rate of −4 • C h −1 is commonly used to complete DTA on supercooling Prunus species (Kader and Proebsting 1992;Salazar-Gutiérrez et al. 2014;Kaya et al. 2018;Liu et al. 2019;Kaya and Kose 2021;Kose and Kaya 2022).However, other cooling rates ranging from −1 to −6 • C h −1 have also been used during DTA or controlled freezing tests to estimate Prunus species cold hardiness (Quamme et al. 1982;Andrews et al. 1983Andrews et al. , 1986;;Ashworth and Davis 1987;Rajashekar 1989;Kader and Proebsting 1992;Mathers 2004;Dumanoglu et al. 2019;Hillmann et al. 2021;Szalay et al. 2021;Kovaleski 2022).Only limited work observing how pretest conditions (e.g., the effects of temperature, moisture, and duration) influence DTA measurements over shorter durations has been completed.Investigating the influence of pretest conditions over shorter durations may provide insight on how different sample collection and preparation protocols for DTA can influence cold hardiness estimates.
The influence of several aspects of DTA testing methodology on grapevine (Vitis vinifera) bud lethal temperature measurements has been researched (Kaya and Kose 2020;Londo et al. 2023).Analogous research for supercooling Prunus species, such as sweet cherry, does not appear to have been conducted.To our knowledge, no standardized DTA method has been developed for sweet cherry that takes into consideration bud pretest storage temperature and moisture conditions, the influence of variability in the rate of cooling, and best practices for bud excision from adjacent tissues.Therefore, there is a need for research to better understand dormant flower bud susceptibility to changes in ambient conditions over a time-scale that may be expected during sample transport as well as the effect of the bud preparation practices and cooling rates on observed LTE temperatures obtained using DTA.Furthermore, a better understanding of how lethal temperature measurements made using DTA compare with measurements made via controlled freezing tests using currently available equipment is needed.
The objectives of this study were to: (i) evaluate how pretest storage temperature, moisture, and duration impacts measured sweet cherry flower bud cold hardiness over a 0-25 h time period; (ii) assess how differences in DTA cooling rates impact measured cold hardiness; (iii) investigate how different methods of removing the flower bud from the bud spur influence measured cold hardiness; and (iv) compare cold hardiness measurements made using DTA and controlled freezing tests.

Plant material
Whole flower bud spurs were randomly selected from "Sweetheart"/Mazzard sweet cherry trees at a commercial orchard that was established in 2006 in Summerland, BC, Canada.Bud spurs were collected at random from the lowest branches to approximately 2 m in height from trees located in a plot of approximately 330 trees.In the following experiments, the standard method for transporting the flower buds from the field to the lab at Agriculture and Agri-Food Canada's Summerland Research and Development Centre (AAFC SuRDC) was to immediately transport the buds in a dry, sealed plastic bag at ambient outdoor air temperature.Transporting flower buds from the field after collection back to the lab took approximately 25 min.

Differential thermal analysis
To measure flower bud lethal temperature, DTA was completed following modified methods by Mills et al. (2006).The standard flower bud preparation and DTA methodology used in this study are as follows: (1) Individual flower buds were excised from the bud spur at the leaf scar.(2) Excised flower buds were placed on TEMs, with six buds per TEM, in trays that each contained nine TEMs.The buds were covered with foam pads to ensure good contact between the buds and the TEM plates and improve heat transfer.Six TEM trays were then loaded into a programmable Tenney Freezer Unit (Thermal Product Solutions, New Columbia, PA, USA) as depicted in Fig. 1.It took approximately 45 min to prepare buds and TEM trays for analysis; this work was done at room temperature.(3) The buds were held at an initial temperature of 3 • C until the TEMs reached 3 • C and were then subjected to cooling temperatures at a rate of −4 • C h −1 until −36 • C was reached, unless indicated otherwise.(4) The peak identification software Bud Processor (v.1.8.0, Brock University, St. Catharines, ON, CA) was used to identify the temperature at which the LTEs occurred.An example temperature voltage profile produced from DTA with the HTE and LTEs identified is presented in the supplementary material (Fig. S1).
A variety of factors were investigated to determine if they significantly affected the LTEs measured using DTA.The following section will outline what factors were explored.The standard plant material transportation, flower bud preparation, and DTA methodology previously described were used to conduct the following experiments, unless described otherwise.Additionally, the number of flower buds subjected to DTA in each treatment for the following experiments was a Fig. 1.Example thermoelectric modules (TEM) tray set up for differential thermal analysis (A) tray with nine TEMs with six flower buds per TEM cell, (B) buds covered with the first layer of flexible foam, (C) buds covered with the first layer and second layer of foam to fill up remaining space in the TEM cells, (D) trays fully prepared and closed with plastic cover, (E) six trays placed in the Tenney programmable freezer.result of the availability of programmable freezers on sampling dates, the number of TEMs that could be placed in each freezer (54 TEMs per freezer), and the limit of six buds per TEM that we chose for conducting DTA.This resulted in a total maximum number of 324 buds per freezer following the above protocols.
Investigating factors that may influence LTE measured using DTA

Experiment 1: Pretest storage duration and moisture
The effects of pretest sample storage duration and moisture conditions on DTA results were evaluated by subjecting flower buds to six different pretest storage treatments, applied in a factorial design: storage at 12.5 • C for 0, 2, or 5 h × storage in dry or moist conditions.Storage at 12.5 • C was selected because this was the average temperature measured in a cooler with ice packs that had been stored at room temperature for 5 h.This experiment aimed to replicate conditions that may be experienced if buds were transported within a room temperature vehicle in a cooler with ice packs for a varying period of time.This is a possible scenario for plant material sampled from orchards at varying distances from the research facility where the measurements will be taken.
To complete this experiment, a composite sample of flower bud spurs was randomly collected as described above; while still in the field, 72 flower bud spurs were randomly assigned to each of the six treatments.For each time interval (0, 2, and 5 h), bud spurs were either stored in a dry sealed plastic bag (dry conditions) or in a sealed plastic bag with paper towel moistened with 15 mL of water (moist conditions).The same surface area of paper towel was used for every replicate and on every sampling date.The bags of bud spurs were immediately transported to the lab in a cooler with ice (∼25 min) so buds were exposed to temperatures near 12.5 • C while they were in transit.One hundred and sixty-two individual flower buds for the 0 h × dry and the 0 h × moist treatment were then immediately prepared for DTA.The 2 h × dry and 2 h × moist as well as the 5 h × dry and 5 h × moist treatments buds were stored in a programmable freezer set to 12.5 • C for 2 and 5 h, respectively, prior to completing DTA.This experiment was completed three times over the 2021-2022 dormant season (7 November 2021, 19 December 2021, and 30 January 2022).

Experiment 2: Pretest storage temperature and duration
The influence of pretest storage temperature and duration conditions on DTA results was evaluated by subjecting flower buds to three temperatures--room temperature (approximately 23 • C), 4 • C, or outdoor ambient air temperatures-for 0, 2, 6, and 25 h prior to conducting DTA.The 0 h conditions acted as a control where buds were transported from the field to the lab at ambient air temperatures and then immediately prepared for DTA without being exposed to differing temperatures.These storage temperatures were selected to replicate conditions that sampled buds may endure prior to DTA (e.g., processing buds immediately, leaving buds on a counter at room temperature, placing buds in a fridge [4 • C], or storing buds outside until they are processed).
To complete this experiment, a composite sample of flower bud spurs was randomly collected as described above; 36 flower bud spurs were randomly assigned to each of the 10 treatments.Buds were transported back to the lab at ambient air temperature (∼25 mins) and were then either immediately prepared for DTA (0 h) or subjected to their respective pre-treatment temperatures (room temperature, 4 • C, or outdoor ambient air temperature) for 2, 6, or 25 h prior to being prepared for DTA.One hundred and eight individual flower buds were subjected to DTA for each treatment.To determine the approximate temperatures to which the outdoor ambient air temperature treatment buds were exposed, weather data from the nearest weather station to AAFC SuRDC (Summerland CS) were obtained from the Government of Canada's online historical weather and climate database (Government of Canada 2023) (https://climate.weather.gc.ca).This experiment was completed three times over the 2022-2023 dormant season (17 November 2022, 12 December 2022, and 11 January 2023).

Experiment 3: Cooling rate
The effects of DTA cooling rate on flower bud LTE were evaluated by comparing measurements made when DTA was conducted using three different cooling rates: −1, −4, and −8 • C h −1 .To complete this experiment, a composite sample of approximately 300 flower bud spurs was randomly collected in the field, transported to the lab, and then 972 flower buds were prepared for DTA, as described above.Three Tenney programmable freezers were run simultaneously to conduct DTA using different cooling rates for this experiment.Three hundred and twenty-four excised flower buds were placed in each freezer.The freezers were initially held at 3 • C, and the buds were subjected to cooling at a rate of −1, −4, or −8 • C h −1 until −36 • C was reached.This experiment was completed two times over the 2022-2023 dormant season (13 December 2022 and 9 January 2023).
To determine typical cooling rates in the Okanagan Valley, BC, over the past ten years (2012)(2013)(2014)(2015)(2016)(2017)(2018)(2019)(2020)(2021)(2022)(2023), hourly weather data from the nearest weather station to AAFC SuRDC (Summerland CS) were obtained from the Government of Canada's online historical weather and climate database (Government of Canada 2023).The rate of hourly temperature change was calculated ( • C h −1 ) and the average and maximum cooling rates for the months of November-March were determined (from November 2022 to March 2023) for all rates less than zero.

Experiment 4: Flower bud excision
We evaluated how flower bud excision from the adjacent bud spur tissue influenced DTA by conducting DTA on flower buds with varying levels of material left on the bud base.To complete this experiment, a composite sample of approximately 200 freshly collected flower bud spurs was randomly collected and transported to the lab, as described above in the standard experiment protocol.Sixty to sixty-six flower buds were then excised from the spur above the base of the bud (just above the physical tapering of the bud), at the base of the bud, at the leaf scar, or below the leaf scar using a scalpel, or buds were manually pulled off the bud spur without using a scalpel; bud preparation was conducted at room temperature.DTA was then conducted on 60-66 flower buds from each excision type, as described in the standard DTA protocol above.This experiment was completed twice in the 2022-2023 dormant season (17 January 2023 and 2 February 2023).Images were taken of each bud under a dissecting scope to determine the approximate distance between the excision point and the base of bud primordia (Fig. 2).
After conducting DTA and identifying the LTEs for Experiments 1-4, the percentage of LTEs produced per bud was also calculated.It is important to highlight that DTA was completed with six flower buds per TEM; it is possible that LTEs from multiple buds occurred at the same temperature.Multiple LTEs at the same temperature would appear as a single peak in the computer output, resulting in a lower number of LTEs identified per TEM and, ultimately, an underestimate of the percentage of LTEs produced per bud.LTE 50 measured using DTA versus LT 50 measured using controlled freezing tests The median low temperature exotherm (LTE 50 ) measured using DTA and the lethal temperature that caused 50% bud damage (LT 50 ) measured using controlled freezing tests were compared.To conduct this experiment, three separate plots of 12 trees (at the same commercial orchard mentioned above) were sampled, with each plot randomly assigned to one of three replicates for this experiment.Six bud spurs per tree in each plot were randomly sampled, for a total of 72 bud spurs per plot.Buds were then transported back to the lab and DTA was conducted on 54 excised flower buds for each replicate following the standard protocol described previously.On the same day in a separate Tenney unit freezer, controlled freezing tests were conducted on three replicates of 120-130 excised flower buds (taken from the samples of 72 bud spurs described above and processed at the same time as the buds used for DTA); buds were placed in the freezer in sealed plastic bags with 10 buds per bag.The freezer temperature was initially held at 1 • C and was then lowered by 1 • C every 15 min (−4 • C h −1 ).An internal bud reference temperature was measured by inserting a thermocouple into one bud which was placed in the freezer during the controlled freezing tests.One bag of 10 buds was removed from the freezer after the reference bud reached each target temperature.Target temperatures were chosen to best select a range of temperatures that would cause 0%-100% mortality in the samples of buds, based on Fig. 2. The remaining material left on the base of five different flower buds following excision from the bud spur (A) using a scalpel cut just above the base of the bud, (B) using a scalpel cut at the bud base, (C) by pulling the bud off the bud spur without the using a scalpel, (D) using a scalpel cut on the leaf scar, (E) using a scalpel cut just below the leaf scar.Photographs were taken on 24 February 2023, using a dissecting scope.
recently conducted DTA analyses.The buds were then refrigerated for a minimum of 12 h and exposed to room temperature for a minimum of 2 h before being cut open to assess cold damage, indicated by the presence of browning primordia tissue.Flower bud cold hardiness measurements using both DTA and controlled freezing tests were completed three times (6 November 2021, 18 December 2021, and 29 January 2022).
To calculate the LTE 50 from measurements obtained using DTA, the median of the identified LTEs was determined.To calculate LT 50 from controlled freezing test measurements, the proportion of new bud damage at each target temperature was assumed to be binomially distributed and was modelled using a logit link (Zuur et al. 2009).The parameter estimates from these models were then used to calculate LT 50 , which was assumed to occur at a probability of 0.5.

Statistical analysis
Treatment significance of the flower bud mean LTE (mLTE) was evaluated on each sampling date using a two-way Analysis of variance (ANOVA) (Experiment 1 -storage duration × moisture) or a one-way ANOVA (Experiment 3 -cooling rate, Experiment 4 -bud excision).One-way ANOVAs were also completed on each treatment factor separately if interactions were significant.Assumptions of normality and homoscedasticity were visually verified using residual plots and treatment means were compared at p ≤ 0.05 using Tukey's HSD (α ≤ 0.05).A linear regression was completed using the LTE 50 and LT 50 values measured using DTA and controlled freezing test.Statistical analyses were completed using RStudio (v1.3.1093;R Core Team 2022).Statistical analysis was not conducted on bud mLTE when observing the effects of storage duration and temperature (Experiment 2) due to the drastic variation among treatments in the number of LTEs produced per bud, which ultimately influenced the resulting LTE sample size.

Results
Factors that may influence LTE measured using DTA

Experiment 1: Pretest storage duration and moisture
The average temperatures for the 7 days prior to the sampling dates of 7 November 2021, 19 December 2021, and 30 January 2022 were 4.5, −2.2, and −2.3 • C, respectively.Storage duration significantly affected flower bud mLTE on all three measurements dates (Table 1).Overall, storing buds at 12.5 • C resulted in significantly higher mLTEs temperatures after 2-5 h.Additionally, buds stored in bags with a moist paper towel (moist) had significantly higher mLTE temperature than buds stored in a bag without a moist paper towel (dry) in the 0 h treatment on two of the three sample dates (19 December 2021 and 30 January 2022).After 2 and 5 h, no treatment difference between the dry and moist conditions was observed on these dates.Overall, the percentage of LTE produced per bud was observed to be similar among treatments on each measurement date.

Experiment 2: Pretest storage temperature and duration
The average temperatures for the 7 days prior to the sampling dates of 17 November 2022, 12 December 2022, and 11 January 2023 were −1.7, −3.1, and 1.5 • C, respectively.For the outdoor ambient temperature treatment, the average approximate air temperature, based on hourly temperature measured at the nearest weather station to AAFC SuRDC (Government of Canada 2023) was 0.9, 0.9, −1.1   spectively.When flower buds were stored at outdoor ambient temperatures, the mLTE and the percent LTE measured per flower bud were similar when measured up to 25 h after buds were sampled on all three measurement dates (Table 2;  2), though none of these trends could be analyzed statistically.

Experiment 3: Cooling rate
In the Okanagan Valley, BC over the past 10 years (2012-2023), the average cooling rates from November to March ranged from −0.5 to −0.8 • C h −1 (Table 3).Additionally, maximum cooling rates among these months ranged from −4.2 to −6.1 The DTA cooling rate (−1, −4, and −8 • C h −1 ) significantly affected measures of mLTE (Table 4).Slower cooling rates resulted in significantly lower mLTE temperatures on both measurement dates.Furthermore, a generally lower number of LTE produced per bud was observed as the cooling rate increased.

Experiment 4: Flower bud excision
The amount of material removed from the flower bud base prior to DTA significantly influenced mLTE on one of the two measurement dates (Table 5).On 17 January 2023, excision of the bud at the leaf scar resulted in significantly lower mLTE than excision of buds just above the taper of the bud base (above bud base) or removal of buds without excision (pulling buds off spurs).These differences were not observed on 2 February 2023.LTE 50 measured using DTA versus LT 50 measured using controlled freezing tests Linear regression completed on LTE 50 and LT 50 measurements from all three sample dates had a small, positive intercept that was not significantly different from zero (p = 0.16) and a significant slope (p < 0.05) close to one (Fig. 4).When compared to a one-to-one regression line, it is evident that the LTE 50 occurred at slightly warmer temperatures than the LT 50 values, overall.Differences between measured LTE 50 and LT 50 values ranged from 0.1 • C to 2.5 • C.

Discussion
The purpose of this research was to identify best practices for measuring sweet cherry flower bud cold hardiness, to facilitate meaningful comparisons across studies or sampling dates.The current research demonstrated that "Sweetheart"/Mazzard sweet cherry flower bud mLTE was affected by sample pretest temperature and moisture conditions, storage duration, DTA cooling rates, and methods of removing individual flower buds from bud spurs.This research also showed that LTE 50 measured using DTA and LT 50 measured using controlled freezing tests are comparable measures of flower bud cold hardiness.

Pretest temperature conditions and duration of storage
In the current study, we demonstrated that storing buds at 12.5 • C resulted in statistically higher mLTE temperatures after only 2-5 h (Experiment 1).Similarly, biologically relevant losses in cold hardiness, and the ratio of LTEs per bud, were observed after 6 h of bud storage at room temperature (∼23 • C) (Experiment 2).These increases in mLTE temperature appeared to be more pronounced later in the dormant season.Earlier in the season, the buds are in endodormancy, which is a period when the dormancy-inducing signal comes from within the affected structure.Later in the season, once chill requirements have been met, the buds enter an ecodormant state.This is a period when growth is limited by environmental factors, such as temperature and moisture conditions (Lang 1987).The more rapid loss in cold hardiness when exposed to warmer temperatures later in the season may be a result of the buds being in this ecodormant state.During ecodormancy, the buds are "physiologically primed" to deacclimate and resume growth when exposed to adequate temperatures (Kalberer et al. 2006;Arora and Taulavuori 2016).Chilling requirements are often met early in the dormant season in this region, meaning that bud storage temperature and duration could have a substantial impact on measurements of apparent cold hardiness through much of the winter (Guak and Neilsen 2013).
The findings from the current study align with previous research on sweet cherry and peach that has observed increased LTE temperatures when buds are exposed to warmer pretest temperatures.Proebsting and Mills (1972) exposed peach and sweet cherry flower buds to 21 • C for 24 h and observed a rapid loss in cold hardiness; buds were collected in the early winter (approximately mid-December) to early spring (when peach calyx were exposed and redpigmented) for peach and early spring (a season divided arbitrarily by when peach anthers become yellow in the earliest buds [approximately mid-January] to when peach calyx are exposed and red-pigmented) for cherry.Proebsting and Mills (1972) observed that cherry was less sensitive to warm storage temperatures in early winter (approximately mid-December to mid-January) than peach.The authors hypothesized this may be due to differences in timing of chilling requirement fulfilment, where cherry may require more chilling than peach.Similarly, Kovaleski (2022) observed an increase in deacclimation rate with increased chilling accumulation in the Prunus species apricot and Canadian plum (Prunus nigra).
The findings from the current study highlight that flower bud pretest storage temperatures at 12.5 • C and room temperature can result in losses in measured cold hardiness after just a few hours (Experiment 1 & 2).These findings also suggest that exposure to warming temperatures (e.g., ∼23 • C) later in the dormant season can result in such rapid deacclimation that LTEs cease to be detected using DTA.As these experiments were conducted on days that had recently experienced fairly mild temperatures, with the average temperature from 7 days prior ranging from 4.5 to −3.1 • C, we hypothesize that the effects of warmer pretest conditions may be exaggerated if ambient outdoor conditions are colder in the days preceding the day of flower bud sampling.Furthermore, the effects of pretest conditions during sample preparation must be more carefully considered when conducting cold hardiness measurements on crops that have a   high deacclimation potential as they will be more susceptible to rapid deacclimation when exposed to warming temperatures (Kovaleski 2022).

Pretest moisture conditions
Bud moisture was observed to significantly increase mLTE temperatures on two of the three sampling days at the 0 h time period (Experiment 1).These results align with observations in a variety of species where surface moisture was found to raise ice nucleating temperature (Modlibowska 1962;Cary and Mayland 1970;Ashworth et al. 1985), including in sweet cherry (Andrews et al. 1983).This may be a result of differences in the efficiencies of ice nucleating substances when surface moisture is present, allowing initiation of the formation of ice to occur more readily (Ashworth 1992).However, storing buds in a sealed bag with a moist paper towel versus in a sealed bag without any additional moisture at 12.5 • C for 2 and 5 h did not result in statistically significant differences in mLTE.This could be a result of a reduction in the overall surface moisture content over time caused by water absorption by the bud scales or even some evaporation within the sealed bags.Overall, the results of both studies observing the effects of pretest temperatures and moisture conditions over various durations indicate that samples should always be stored in similar ambient conditions and for similar durations prior to processing for DTA.

DTA cooling rate
Significant variability in mLTE among the cooling rates of −1, −4, and −8 • C h −1 was observed in this study (Experiment 3).These findings highlight that differences in DTA cooling rates of just a few degrees celsius per hour may result in measures of cold hardiness that are no longer comparable.This is an important factor to consider when contrasting measures of cold hardiness presented in the literature for supercooling Prunus species, as reported methods for DTA have indicated the use of cooling rates ranging from −1 to −6 • C h −1 .
The findings from the current study align with similar findings in peach (Ashworth 1982;Kang et al. 1998) where the LTE temperatures were observed to be significantly lower with slower cooling rates.Ashworth (1982) observed that a fastcooling rate of −20 • C h −1 resulted in the primordium of intact flower peach bud no longer supercooling.By contrast, the authors observed supercooling (presence of a LTE) when flower buds were cooled at a slower rate of −2 • C h −1 ; they hypothesized that rates of cooling between 0 and −10 • C h −1 are critical for supercooling.Slower cooling rates may allow for water to be withdrawn from the bud axis into the bud scales or surrounding tissues in contrast to faster cooling rates, where water may remain in this critical region (Quamme 1978(Quamme , 1983;;Ashworth 1982).It has been proposed that ice formation in peach is first initiated extracellularly in the bud axis and scales and that the redistribution of water to these regions allows the water in the primordium to supercool.If the cooling rate is too fast, supercooling may not occur (Ashworth 1982).Kang et al. (1998) highlights that since supercooling in peach is cooling-rate dependent, peach should fall within the supercooling category of extraorgan freezing as described by Ishikawa and Sakai (1982).The cooling rate data presented in this study support categorizing sweet cherry as a species that experiences supercooling by extraorgan freezing as well.
It is also possible that a slower cooling rate may more closely reflect field conditions and may, therefore, help improve measurement accuracy.The average hourly cooling rate observed in the Okanagan Valley over the past 10 years (2012-2023) was typically <−1 • C h −1 (Table 3).As such, slower DTA cooling rates may be more appropriate for the Okanagan Valley.However, one must consider that using a cooling rate this slow will also result in DTA measurements that will span the duration of 2 days and, thus, may be impractical for some research and extension programs.

Flower bud excision
Overall, the findings from this study suggest that the amount of material left attached to the flower bud and method of excision cause minimal variability in measures of LTE (Experiment 4).Although some statistically significant differences between mLTEs were observed among excision methods on one of the two study days, differences were only 1.6 -1.7 • C.However, retaining too much material on the bud base may prevent contact of the bud with the TEMs, especially if the plant material angles away from the bud, causing it to lift off the surface of the TEM.Use of a thermal conductive paste, like silicone grease (Kaya et al. 2018), may help facili-tate contact between the bud and the TEMs, if additional material is left attached to the base of the bud.
Even when the buds were excised just above the tapered area at the bud base, there were still significant vascular traces left attached below the primordia (Fig. 2).These results suggest that by not disturbing the area immediately below the primordia, which is critical in the supercooling capacity of peach (Quamme 1978;Quamme et al. 1995), the supercooling capacity of sweet cherry may remain undisturbed.This may help explain the differences between the observations in this study and findings in peach reported by Quamme (1978) and Proebsting and Sakai (1979).These authors observed that flower buds excised from the bud closer to the flower primordia experienced cold damage at warmer temperatures than buds excised with vascular traces and intact pieces of shoot.Bud morphology between cherry and peach also differ where cherry flower buds typically contain several floral primordia (Fadón et al. 2015), whereas peach flower buds have a single primordium.
Although the findings from these studies do not allow us to identify conditions that may result in "true" cold hardiness measurements (measurement accuracy), they do allow us to highlight that consistency in the pretest conditions, cooling rate, and bud excision may help reduce error and increase the precision of LTE measurements.Measuring LTE using DTA allows us to make meaningful and relatively rapid estimates of bud cold hardiness; this is biologically relevant information even though the absolute, in-field value of bud cold hardiness may differ from the in-lab LTE measurements (Londo et al. 2023).Increasing the precision of these measurements may allow for more meaningful observations and comparisons to be made among measurement dates and studies (e.g., comparing treatments or varietal differences).LTE 50 measured using DTA versus LT 50 measured using controlled freezing tests Controlled freezing tests provide an option for researchers, and extension workers, to measure plant lethal temperatures when the costly specialized equipment required to conduct DTA may not be available for use.In the current study, we observed good agreement between LTE 50 measured using DTA and LT 50 measured using controlled freezing tests.Overall differences in measures between the two methods ranged from 0.1 • C to 2.5 • C; these differences are comparable to findings by Andrews and Proebsting (1987) who observed differences between LTE 50 and LT 50 in sweet cherry of 0.1 • C to 2.8 • C. Similar agreement between the LT 50 determined using controlled freezing tests and the mLTE determined using DTA has been observed in other Prunus species including peach, sweet cherry, apricot, and plum (Quamme 1974;Mathers 2004;Liu et al. 2019).

Recommendations
Based on our observations, we can make recommendations for sample collection, storage, and preparation for DTA to estimate sweet cherry flower bud cold hardiness.Firstly, we suggest preparing buds for DTA immediately after collection, to avoid potential deacclimation of buds resulting from warmer pretest temperatures and, when possible, not storing buds for prolonged periods prior to analysis.If buds need to be transported or temporarily stored prior to conducting DTA, the findings from this study suggest that buds should be stored at temperatures closest to ambient outdoor air temperatures, to reduce potential deacclimation.Furthermore, bud samples should be transported at similar moisture conditions.For example, if bud samples are collected on days with high levels of precipitation, excess moisture should be avoided by patting the buds dry with an absorbent material before transportation and preparation.When excising the flower bud from the bud spurs, it is important that enough material is left on the bud base so that the region subtending the primordia remains undisturbed.The findings of this study suggest that excising the flower buds from the bud spur anywhere from the bud base to directly below the leaf scar for "Sweetheart" sweet cherry will result in similar LTE measures.However, we recommend ensuring a consistent excision point for all samples.
There remain several areas of further study to strengthen our understanding of sweet cherry cold hardiness and to improve consistency in DTA methodologies.In this research, we used the sweet cherry cultivar "Sweetheart."Varietal differences in morphology that may influence the location of the region that supports supercooling below the primordia should be investigated to determine where buds can be excised from the spurs to avoid disturbing this critical region.Furthermore, in the current study all sampled buds were prepared for DTA at room temperature (∼45 min) and placed in a programable freezer with an initial temperature of 3 • C. Future research investigating how sample preparation temperatures and differences between outdoor ambient air temperature and initial freezer temperature influence cold hardiness estimates would be beneficial (Kaya and Kose 2020).
Much of this research focused on aspects of DTA methodology that could help improve estimate precision.However, it may be beneficial to growers, researchers, and extension workers to improve our understanding of the accuracy of cold hardiness estimates obtained using DTA.Investigating how measured LTEs compare to in-field cold hardiness, by comparing measures of LTE to in-field bud damage prior to and after predicted cold snaps, could help strengthen our understanding of DTA accuracy.With similar experiments, we also recommend investigating whether slower DTA cooling rates may allow for cold hardiness estimates that more closely reflect the cold hardiness of buds in the field.In the future, DTA cooling rates may need to be tailored to the local climate to best match ambient rates of cooling.

Conclusion
The current research evaluated the effects of a variety of pretest storage temperatures, duration, and moisture conditions, as well as DTA cooling rates, and flower bud excision methods on sweet cherry flower bud cold hardiness measurements made using DTA.It also compared sweet cherry cold hardiness measurements made using DTA (LTE 50 ) and controlled freezing tests (LT 50 ).This study provides several key findings that should be considered by researchers interested in conducting cold hardiness measurements in sweet cherry: (i) pretest temperature, duration, and moisture conditions can influence LTE temperatures, making consistency in these conditions prior to DTA, as well as reporting on these conditions, important; (ii) the cooling rate used for DTA significantly influences the LTE temperature and findings from this study suggest that LTEs measured using slower cooling rates (e.g., −1 • C h −1 ) are not comparable to LTEs measured using faster cooling rates (e.g., −4 • C h −1 and faster); (iii) in general, the amount of plant material left attached to the bud spur (up to and including the leaf scar) or excised from the base of the flower bud does not significantly alter the LTE temperature as long as the region directly subtending the primordia is not disturbed; (iv) DTA and controlled freezing tests provide comparable measures of flower bud cold hardiness.Future work to improve our understanding of the relationship between LTE 50 and LT 50 at a higher resolution throughout the season may be beneficial to account for differences in plant cold hardiness measurements when comparing LTE 50 and LT 50 .This may be particularly beneficial if collaboration between labs with different access to equipment to measure lethal temperature (DTA vs. controlled freezing tests) arises.

Fig. 3 .
Fig. 3. Mean low temperature exotherm (mLTE ± SD) and linear regression lines of sweet cherry flower buds exposed to different ambient air temperatures (outdoor temperature, 4 • C, or room temperature (∼23 • C)) for 0, 2, 6, or 25 h prior to conducting differential thermal analysis.Measurements were conducted three times during the 2022-2023 dormant season.The 0 h measurements on 17 November 2022 are missing due to computer failure.

Fig. 3 )
, and mLTEs ranged from −12.6 to −12.2 • C (17 November 2022); from −15.8 to −15.3 • C (12 December 2022); and from −14.2 to −13.0 • C (11 January 2023) (Fig.3).Similar results were observed in the 4 • C treatment after up to 25 h after buds were sampled, with mLTEs ranging from −12.9 to −12.2 • C (17 November 2022); from −15.7 to −14.4 • C (12 December 2022) and from −14.5 to −13.0 • C (11 January 2023) (Fig.3).For the room temperature treatment, however, a progressive loss in cold hardiness (higher mLTE values) was observed as storage duration increased.The mLTE increased by 1.2 • C between the 2 and 25 h storage treatments on 17 November 2022, and by 4.2 and 5.0 • C between the 0 and 25 h storage treatments on 12 December 2022, and 11 January 2023, respectively.Furthermore, a reduction in the number of LTEs recorded per flower bud was observed with increasing storage time at room temperature.This pattern was most pronounced on the third sampling date (11 January 2023), when a major decrease in the number of LTE recorded per bud was observed after 6 and 25 h (Table

Table 3 .
Average cooling rate ± SD over 10 years(November  2012-March 2023)  in Summerland, BC (located in the Okanagan Valley).Cooling Rate ( • C h −1 Hourly weather data were obtained from the Government of Canada's online historical weather and climate database(Government of Canada 2023).

Fig. 4 .
Fig. 4. Linear regression of LTE 50 and LT 50 completed on three combined measurement dates in the 2021-2022 dormant season.The linear regression line is visually compared to a one-to-one regression (dashed line).LTE, low temperature exotherm.

0001 <0.0001 <0.0001 <0.0001 <0.0001 <0.0001 0.00127
Note: Differential thermal analysis were conducted three times during the 2021-2022 dormant season.a Within columns, means that share the same lowercase letters do not differ significantly (p ≤ 0.05) using Tukey's HSD.b Within rows, means that share the same uppercase letters do not differ significantly (p ≤ 0.05) using Tukey's HSD.c Value were squared to meet the assumptions of normality.
Note: Differential thermal analysis were conducted three times during the 2022-2023 dormant season.The 0 h measurements on 17 November 2022 are missing due to computer failure.

Table 4 .
Mean low temperature exotherm (mLTE ± SE) and percentage of sweet cherry flower buds that produced LTEs (LTE/Bud) after exposure to different differential thermal analysis (DTA) cooling rates.DTA were conducted two times during the 2022-2023 dormant season.mLTE values measured on the same date that share the same letter do not differ significantly (p ≤ 0.05). Note:

Table 5 .
Mean low temperature exotherm (mLTE ± SE) and percentage of sweet cherry flower buds that produced LTEs (LTE/Bud) after excision from the bud base with varying levels of plant material remaining.
Note: Differential thermal analysis were conducted two times during the 2022-2023 dormant season.mLTE values measured on the same date that share the same letter or have no letters do not differ significantly (p ≤ 0.05).