Thermal Dynamics of Xylem and Soil–Root Temperatures in Olive and Almond Trees and Their Relationship with Air Temperature

Abstract

Air temperature is commonly used to represent plant thermal conditions, although temperatures within woody tissues and the soil–root zone can differ substantially under field conditions. This study characterized the thermal dynamics of xylem tissue and the soil–root interface in almond and olive orchards under Mediterranean field conditions in Southern Spain. Using long-term in-field measurements, temperatures were monitored in branch and trunk xylem tissues and at the soil–root interface, and regression models were developed to provide empirical correction relationships between air and internal temperatures across seasons and sensor position. Branch xylem temperatures closely matched air temperature for both minima and maxima. In contrast, trunk xylem and the soil–root interface showed pronounced thermal buffering. Trunk xylem maximum temperature was significantly (3.4 to 5.4 °C) lower than air temperature during summer. Shaded soil–root interface temperatures were 5.2 to 9.0 °C lower than air temperature in spring and summer but 5.9 to 11.7 °C higher than air temperature in autumn and winter. These patterns indicate a strong capacity of woody tissues and the soil–root system to moderate external thermal conditions. By quantifying air-to-tissue and air-to-soil relationships under field conditions, this study provides microclimatic data that can improve agronomic models and temperature-driven disease risk frameworks for vascular pathogens infecting woody crops.

1. Introduction

Olive (Olea europaea L.) and almond (Prunus dulcis L.) are two of the most representative tree crops in the Mediterranean Basin, covering 9.91 million and 482,212 hectares, respectively [1]. Spain leads in their acreage within this region. Beyond their economic, cultural, and social relevance, olive and almond orchards play crucial roles in maintaining the integrity of Mediterranean ecosystems and rural landscapes [2]. Both tree species share similar structural traits (perennial woody crops with deep roots and high xylem conductance).

Crops are exposed to multiple abiotic and biotic stresses that can reduce and limit their productivity [3]. Among the primary abiotic drivers, temperature is particularly relevant because it modulates physiological processes, including metabolic processes [4,5], water and photosynthate transport [6,7], and the transition from vegetative to reproductive stages [4]. In Mediterranean orchards, temperature strongly influences critical phenological events such as flowering and pollen release. However, despite its central role, little is known about the actual temperatures experienced within olive and almond xylem tissues throughout the year under field conditions. Soil temperature also plays a pivotal role in regulating plant developmental and metabolic processes, including root growth and distribution, nutrient uptake, root respiration, and microbial activity, among others. Quantifying the soil–root thermal environment is also necessary to interpret whole-tree thermal dynamics in Mediterranean orchards. In managed systems, the magnitude of soil thermal buffering can vary substantially with canopy cover and surface exposure, potentially creating contrasting microclimates within the orchard. Beyond its importance for plant functioning, temperature also influences microbial activity in soil and plant tissues, and shapes the thermal environments associated with root, foliar, and vascular pests and pathogens.

Temperature measurements are widely used in phenology and crop modeling [8], and they also underpin temperature-driven disease risk frameworks for plant pathogens [9,10]. These approaches rely on air temperature recorded by conventional weather stations or shielded sensors positioned aboveground and often outside the canopy. However, air temperature may not accurately represent the thermal regimes within internal plant tissues or at the soil–root interface, where strong seasonal buffering can occur [11,12]. Consequently, field-based measurements of internal plant and root-zone temperatures are needed to refine air-based temperature inputs and improve the biological realism of models applied to woody orchards [4,13]. These differences are mainly attributed to the continuous energy flux among the canopy, woody plant organs, and the surrounding soil [4,14,15]. Moreover, standard meteorological sensors are positioned at 1.5–2 m above ground and outside the canopy, which further limits their representativeness of internal microclimatic conditions within tree structures.

Accurate measurement and characterization of internal temperature are crucial for understanding the physiological and ecological processes occurring within tree canopies and internal plant tissues, such as the xylem [6,8]. The use of thermocouples offers precise and reliable point field temperature measurements [16]. These sensors can be manufactured in very small sizes, enabling the capture of real-time temperature data at exact points with minimal disturbance to plant tissues, thus providing valuable insights into tree thermal dynamics [17]. However, the use of thermocouples embedded in specific plant tissues under field conditions remains relatively uncommon. Previous studies have primarily focused on the physiological responses of conifers and angiosperms during winter periods and freeze–thaw events [18], insect thermal behavior [19], and the heat insulation capacity of bark in forest species [20]. The use of thermocouples in internal plant tissues has been reported in only one woody crop species, grapevine, where Peña Quiñones et al. [12,21] measured daytime temperatures within xylem tissues at various trunk heights. Internal xylem and root-zone temperatures may diverge from air temperature and could therefore shape the thermal conditions experienced by beneficial and pathogenic microorganisms inhabiting the xylem [22,23,24,25,26]. Moreover, soil and root temperatures may be critical for infection or survival of some vascular plant pathogens, as they can migrate and overwinter in the soil–root system [27,28], ensuring their survival for subsequent seasons. This is the case for two major vascular pathogens threatening Mediterranean olive and almond crops (i.e., the bacterium Xylella fastidiosa and the soilborne fungus Verticillium dahliae).

The primary objective of this study was to investigate daily and seasonal temperature dynamics in the xylem tissues and the soil–root interface of olive and almond trees under Mediterranean field conditions. The secondary objective was to determine how these temperature dynamics vary with the type of plant tissue (trunk versus branches), xylem depth within the trunk, and tree orientation and position within the orchard. To address these objectives, we conducted a long-term in-orchard monitoring of branch and trunk xylem temperatures and soil–root interface temperatures and evaluated their relationships with local air temperature across seasons and positions. We hypothesized that internal trunk and soil–root interface temperatures could exhibit thermal buffering relative to air temperature, and that the magnitude of this buffering would vary seasonally and with tissue depth. By quantifying these internal microclimates and their relationships with air temperature, this study aims to provide field-based evidence that can improve agronomic models and temperature-driven risk frameworks for vascular plant pathogens of woody crops. To our knowledge, this is the first multi-year study to continuously monitor xylem and soil–root temperatures in field-grown Mediterranean woody crops.

2. Materials and Methods

2.1. Study Site

The study was carried out in two experimental olive and almond orchards located at the Alameda del Obispo experimental farm in Córdoba, Spain (37°51′37.4″N, 4°47′47.0″ W at an elevation of 96 m), close to the Guadalquivir River (Figure 1a), whose soils are classified as typical Xerofluvent with a sandy loam texture. The region experiences a typical Mediterranean climate, with an average annual rainfall of approximately 600 mm, primarily during the autumn and spring seasons. The average annual evapotranspiration (ET₀) is 1390 mm [29].

For the study, 14-year-old olive trees of cv. Picual and 10-year-old almond trees of cv. Marinada were selected. The trees ranged in height from 3 to 3.5 m, with trunk diameters of 8 to 10 cm, and were spaced at 4 × 2 m intervals for olive (Figure 1b) and at 5 × 4 m for almond (Figure 1c). The mean canopy cover was 58.4 ± 2.47% in the olive orchard and 70.7 ± 3.86% in the almond orchard. Both orchards were drip-irrigated during the warm season following standard commercial practice, with an annual water supply of 308 mm and 500 mm, for olive and almond trees, respectively. Fertilization followed standard orchard management practices.

2.2. Temperature Measurements

Temperature within the xylem tissue was measured using thermocouples. Eighteen sensors were installed in olive trees (six per tree), and nine in almond trees (three per tree). Sensors were placed in three trees per orchard, located at the southern border, northern border, and center of the orchard, surrounded by other trees (Figure 1b,c). The temperature probes were manufactured at IAS-CSIC, using measurement-grade type-K wire (Omega Engineering, Inc., Norwalk, CT, USA) encased in 2 mm diameter stainless steel tubes. The sensing junction was positioned at the tube tip, and thermal conductive paste was used to improve contact with plant tissue. To minimize radiation-related artifacts, the insertion point and cable near the probe were insulated using a foam cap covered with aluminum foil.

In each olive tree, sensors were installed in south- and north-oriented main branches with diameters of 4 to 5 cm, at a height of 1.5 m and a depth of 1 cm, and in the trunk at a height of 80 cm and at depths of 1 cm and 4 cm (Figure 1d,e). In almond trees, sensors were installed in a south-oriented branch at a depth of 1 cm and in the trunk at depths of 1 cm and 4 cm, oriented to the north and south (Figure 1d,e). Depths of 1 cm and 4 cm were selected to capture temperature within the active xylem region of adult trees [30,31,32], and north/south orientations were chosen to represent contrasting probabilities of direct solar thermal input.

In each orchard, one thermocouple was installed in a standard radiation shield, placed between the rows, at a height of 1.5 m to measure air temperature (T_air), and two thermocouples were buried at a depth of 20 cm to measure soil temperature, one in a sunny area and the other in a shaded area (Figure 1d). The area where the soil sensors were installed was cleared of adventitious plants, leaving the soil bare during the experiment. To validate the assumption that soil–root interface temperature at 20 cm may represent root temperature, we conducted a preliminary test in two trees per species. In each tree, two identical thermocouples were installed at a depth of 20 cm: one positioned in close proximity to the main roots and a second inserted approximately 1 cm inside the tissue of the main root at the same depth. After installation, excavated soil was carefully replaced to restore field soil conditions. Daily temperatures recorded by the paired sensors did not differ significantly (p ≥ 0.05), supporting the use of soil–root interface measurements as a proxy for root-zone temperature (Supplementary Figure S1). All sensors were connected via a multiplexer (model AM16/32, Campbell Scientific Inc., Logan, UT, USA) to a data logger (model CR1000, Campbell Scientific Inc., Logan, UT, USA), with temperature measurements recorded at 10 min intervals. Temperature data collection spanned two and a half years (from December 2017 to June 2020) for the olive orchard and just over one year (from February 2018 to March 2019) for the almond orchard. At the end of the experiment, just over 133,000 measurements were obtained from each sensor for the olive trees, and nearly 55,000 for the almond trees.

2.3. Data Analyses

Monitoring duration differed between orchards (olive: 2.5 years; almond: 1 year) due to the availability of the almond plot and sensor deployment during the study period. Because unequal time series can limit direct interspecies comparability, all statistical analyses were conducted for each crop independently. Accordingly, the results are interpreted as crop-specific characterizations of internal thermal dynamics under Mediterranean orchard conditions rather than as formal quantitative comparisons between olive and almond.

Ten-minute temperature time series were collected for each sensor, and the dataset was partitioned according to meteorological seasons [33]; namely, winter (from December to February), spring (from March to May), summer (from June to August), and autumn (from September to November). Each tree served as an experimental unit. Maximum and minimum temperature values were extracted for each sensor on an hourly, daily, and monthly basis and were used for statistical analyses.

Temporal dynamics of minimum and maximum monthly temperatures: The variation in temperature during the experimental period for each sensor was characterized based on the monthly average of minimum and maximum temperatures. For each month and crop, data were analyzed using a standard one-way analysis of variance (ANOVA) with sensor position and orientation in the tree or soil as the main factor. Pairwise comparisons of means among the main factor levels were performed using the post hoc Tukey’s honestly significant difference (HSD) test, at p < 0.05.

Temporal dynamics of minimum and maximum daily temperatures: The variation of temperature over the 24 h profile for each sensor was assessed based on the monthly average of daily minimum and maximum temperatures. We conducted one-way ANOVA for each individual one-hour period with sensor position and orientation in the tree or soil as the main factor. Differences between sensor positions and orientations were evaluated using pre hoc one-degree-of-freedom contrasts. To account for multiple comparisons, a Sidak correction was applied to adjust the p-value based on the number of contrasts performed at p < 0.001.

Relationships between xylem tissues, soil–root system temperatures, and air temperatures: Linear models were fitted to determine the relationships between sensor-recorded temperatures and air temperatures with season, sensor type, and their interaction as covariates. Least-squares means (also known as marginal means) were estimated to interpret the main effects of these covariates on the relationship between sensor-recorded temperatures and air temperatures. Differences among sensor types and seasons were determined using the HSD post hoc test at p < 0.05 with Sidak correction.

All statistical analyses were performed using R software version 4.0.3 [34] using the “agricolae” v.1.3.5 [35], “multcomp” v.1.4.23 [36], and “emmeans” v.1.8.5 [37] packages.

3. Results

3.1. Temporal Dynamics of Sensor Monthly Average Temperature and Its Relationship with Air Temperature

Figure 2 provides a long-term overview of daily minimum and maximum temperatures recorded by air, trunks, branches, or soil sensors within each orchard. Across sensor types, deviations from air temperature were more pronounced for daily maxima, particularly during the summer. Soil–root interface temperatures exceeded air minima but were lower than air maxima (

Table 1. Average minimum and maximum temperatures (T) for air, soil, and xylem tissue measured with thermocouples at various positions in an olive and almond orchard.

T/ Crop/Month	T Air		T Soil Sunlit		T Soil Shade		T Branch N		T Branch S		T Trunk 1 cm N		T Trunk 1 cm S		T Trunk 4 cm N		T Trunk 4 cm S
Minimum temperatures
Olive orchard
January	3.24	b	8.72	a	8.36	a	2.97	b	3.05	b	3.72	b	3.73	b	3.57	b	3.58	b
February	4.26	b	10.88	a	9.69	a	3.94	b	4.04	b	4.84	b	4.80	b	4.65	b	4.67	b
March	7.01	cd	13.84	a	12.21	b	6.62	d	6.75	cd	7.72	c	7.58	cd	7.50	cd	7.51	cd
April	9.80	cd	16.17	a	14.74	b	9.41	d	9.53	d	10.64	c	10.59	c	10.42	cd	10.41	cd
May	13.20	d	22.43	a	19.06	b	12.88	d	13.04	d	14.51	c	14.45	c	14.21	c	14.23	c
June	15.84	d	26.72	a	22.96	b	15.56	d	15.62	d	17.57	c	17.66	c	17.16	c	17.21	c
July	18.37	d	30.05	a	26.08	b	18.09	d	18.04	d	20.17	c	20.24	c	19.71	c	19.79	c
August	19.58	d	31.16	a	28.15	b	19.26	d	19.31	d	21.42	c	21.43	c	20.92	c	20.96	c
September	18.05	de	26.81	a	25.38	b	17.74	e	17.79	e	19.16	c	18.76	cd	18.85	cd	18.83	cd
October	12.35	bc	20.89	a	20.06	a	12.08	c	12.14	c	13.45	b	13.38	b	13.16	bc	13.11	bc
November	8.24	b	14.27	a	14.32	a	7.95	b	8.04	b	8.98	b	8.87	b	8.77	b	8.79	b
December	5.11	b	10.47	a	10.36	a	4.82	b	4.90	b	5.62	b	5.60	b	5.48	b	5.47	b
Almond orchard
January	3.80	b	8.15	a	8.52	a			3.15	b	3.76	b					3.67	b
February	5.28	b	10.80	a	10.74	a			4.70	b	5.27	b					5.01	b
March	7.74	b	12.71	a	12.90	a			7.47	b	7.93	b					7.74	b
April	10.48	b	15.95	a	15.90	a			10.32	b	11.01	b					10.79	b
May	12.15	d	19.74	a	18.55	b			11.90	d	12.92	c					12.94	c
June	16.51	d	25.78	a	23.26	b			16.31	d	17.33	c					17.56	c
July	17.19	d	29.22	a	25.79	b			16.99	d	18.32	c					18.62	c
August	19.94	d	30.92	a	27.26	b			19.75	d	21.05	c					21.45	c
September	18.76	d	26.31	a	24.06	b			18.48	d	19.40	c					19.60	c
October	12.92	cd	18.97	a	18.15	b			12.63	d	13.28	cd					13.45	c
November	9.50	bc	13.86	a	14.35	a			9.16	c	9.78	bc					9.93	b
December	6.52	cd	10.84	b	11.68	a			5.92	d	6.41	c					6.29	cd
Maximum temperatures
Olive orchard
January	15.11	c	11.76	d	9.59	e	17.62	a	17.83	a	14.58	c	16.29	b	14.77	c	17.80	a
February	19.52	c	14.93	e	11.30	f	22.07	ab	22.60	a	16.84	d	19.40	c	17.53	d	20.98	b
March	20.99	b	17.82	d	14.05	e	23.21	a	23.96	a	18.16	cd	20.38	b	19.04	c	21.28	b
April	22.90	b	20.06	de	16.49	f	25.92	a	25.96	a	19.38	e	20.81	cd	20.69	cde	21.70	bc
May	29.99	c	27.84	d	21.01	g	34.18	a	32.33	b	24.51	f	25.60	ef	27.37	d	26.87	de
June	32.74	b	32.44	b	25.16	d	36.17	a	35.26	a	28.79	c	28.77	c	32.20	b	29.77	c
July	37.01	b	36.14	bc	28.61	f	40.23	a	39.76	a	32.22	e	32.17	e	35.93	c	33.35	d
August	39.15	b	36.92	c	30.48	f	42.29	a	42.16	a	33.71	e	34.08	e	35.60	d	35.55	d
September	33.56	b	31.84	c	27.07	e	35.32	a	36.51	a	29.77	d	30.10	d	30.68	cd	31.44	c
October	26.73	b	25.36	bc	21.58	d	28.54	a	29.40	a	24.00	c	24.28	c	24.53	c	25.34	bc
November	17.67	cd	16.82	d	15.29	e	19.69	ab	20.55	a	16.60	de	17.44	cd	16.86	d	18.56	bc
December	16.41	bc	13.14	d	11.50	e	18.35	a	18.72	a	15.72	c	17.30	b	15.78	c	18.96	a
Almond orchard
January	17.44	c	11.01	d	10.52	d			20.03	b	23.55	a					22.89	a
February	21.28	c	13.59	d	13.00	d			23.59	b	26.30	a					26.55	a
March	18.71	b	14.55	c	14.49	c			21.55	a	20.15	ab					20.71	a
April	23.27	b	17.85	d	17.79	d			25.39	a	21.92	c					20.95	c
May	26.14	b	24.56	c	21.25	e			27.29	a	24.72	c					23.75	d
June	30.73	a	30.84	a	26.03	c			31.00	a	28.03	b					26.15	c
July	33.86	a	34.52	a	28.62	c			34.24	a	30.63	b					28.75	c
August	37.05	a	35.17	b	29.58	d			37.26	a	33.48	c					29.70	d
September	31.91	b	29.89	c	26.54	e			34.57	a	32.00	b					28.64	d
October	23.78	b	22.05	c	20.28	d			26.05	a	26.64	a					23.36	bc
November	18.41	c	16.04	d	15.76	d			20.78	b	22.94	a					21.62	ab
December	19.16	c	13.40	d	13.43	d			21.71	b	25.90	a					24.90	a

Values are the mean of the minimum and maximum temperature values for each day throughout the measured period by sensor and month. For each sensor and tree species within a month, values with the same letter are not significantly different according to the post hoc Tukey’s honestly significant difference (HSD) test, at a significance level of p < 0.05.

; Figure 2).

3.1.1. Air Temperatures

Air-temperature ranges are reported separately for each orchard to provide the baseline needed to interpret xylem and soil–root interface thermal buffering and the air-to-sensor relationships presented below (Table 1). Across the monitoring period, the monthly average minimum T_air in each orchard spanned the seasonal range shown in Table 1. The coldest period was characterized by average minimum T_air below 11 °C and spanned from November to April (Table 1). Minimum T_air typically occurred between 00:00 and 8:00 solar time from October to March, and between 00:00 and 07:00 from April to September, often reaching its lowest values around 07:00 and 06:00, respectively (Figure 3). In general, the absolute minimum T_air stayed above 0 °C, with a few exceptions noted during the winter (only 14 independent days in the olive orchard and 1 day in the almond orchard) (Figure 2a,c). Freezing T_air was mainly recorded between 00:00 and 08:00 solar time, with T_air rising above 0 °C as the day progressed.

Monthly average maximum T_air values for each orchard are summarized in Table 1. The warmest period, with average maximum T_air above 23.5 °C, lasted from May to October within each orchard (Table 1). The highest monthly average maximum T_air was typically recorded between 10:00 and 17:00 from October to March, and between 09:00 and 19:00 from May to September (Figure 4).

3.1.2. Xylem Temperatures

Average monthly minimum temperatures in xylem tissue for branches and trunks are summarized in Table 1. Across the monitoring period, branch minima ranged from 3 to 20 °C and trunk minima from 4 to 22 °C within each orchard (Table 1). During the coldest period, branch minimum temperatures fell to approximately 3 to 10 °C and trunk minimum temperatures to approximately 4 to 11 °C within each orchard (Table 1).

Within each orchard, the average monthly minimum temperatures recorded in branches did not differ significantly from T_air throughout the year (p ≥ 0.05), regardless of sensor orientation (Table 1). For trunk sensors, no significant differences (p ≥ 0.05) among orientations and depths were observed during the coldest period (Table 1). In contrast, during the warmest months (from May to August), trunk minimum temperatures were consistently higher (p < 0.05) than T_air, averaging 1.0 to 1.9 °C higher in the olive orchard and 0.8 to 1.5 °C in the almond orchard, with no detectable effect of sensor depth or orientation (Table 1). These results were also supported by hourly data, except from 00:00 to 09:00 when absolute minimum temperatures typically occurred (Figure 3).

Average monthly maximum xylem temperatures for branches and trunks are summarized in Table 1. In the olive orchard, branch maximum temperatures ranged from 17.6 ± 0.32 to 42.3 ± 0.35 °C, while trunk maximum temperatures ranged from 14.6 ± 0.29 to 35.9 ± 0.16 °C (Table 1). In the almond orchard, branch maximum temperatures ranged from 20.0 ± 0.39 to 37.3 ± 0.29 °C and trunk maximum temperatures from 20.2 ± 0.61 °C to 33.5 ± 0.17 °C (Table 1). Within the olive orchard, monthly average branch maximum temperatures were consistently higher (p < 0.05) than T_air throughout the year, with mean increases of 2.6 and 2.8 °C for north- and south-oriented branches, respectively (Table 1). Within the almond orchard, branch temperatures were higher (p < 0.05) compared to T_air in all months except summer (June to August), with increases of 1.2 to 2.8 °C from September to May (Table 1).

Conversely, trunk maximum temperatures in the olive orchard were consistently lower (p < 0.05) than T_air, with an average difference of 0.8 to 3.1 °C, depending on sensor orientation and depth (Table 1). In the almond orchard, this pattern was mainly observed during the spring and summer, with a decrease in temperature of 2.2 °C for north-oriented trunks with sensors at a depth of 1 cm, and 4.5 °C lower for south-oriented 4 cm depth sensors (p < 0.05). However, during winter, trunk temperatures were 4.6 to 5.1 °C higher (p < 0.05) than T_air (Table 1).

Regarding depth effects, trunk sensors at a depth of 4 cm recorded significantly higher (p < 0.05) temperatures during winter and summer for south-oriented sensors, and mainly during summer for north-oriented sensors (Table 1). In the almond orchard, 1 cm depth trunk sensors oriented north recorded significantly higher (p < 0.05) temperatures than south-oriented sensors during summer (Table 1).

Overall, hourly data showed that branch temperatures were higher and trunk temperatures were lower than T_air from 09:00 to 19:00 in spring and summer, when air temperatures peaked (Figure 4). During these periods, branch temperatures were consistently higher (p < 0.05) than trunk temperatures, regardless of trunk depth or orientation.

3.1.3. Soil Temperature

Minimum soil temperatures, irrespective of the position of the sensor on the shaded or sunlit side, were consistently higher (p < 0.05) than T_air (Table 1). For both crops, monthly sunlit minimum soil temperatures varied from 8.2 ± 0.36 to 31.2 ± 0.12 °C and from 8.4 ± 0.14 to 28.2 ± 0.07 °C in shaded soil. (Table 1). During the coldest period (November to April), the average monthly minimum temperature ranged from 8.2 ± 0.36 to 16.2 ± 0.20 °C in sunlit soil and from 8.4 ± 0.14 to 15.9 ± 0.34 °C in shaded soil, for both olive and almond orchards. (Table 1). During these months, the monthly minimum temperature in soil increased by 5.4 to 6.8 °C in sunlit soil and 4.9 to 6.1 °C in shaded soil relative to air temperature. From May to October, the soil monthly minimum temperatures increased by 9.2 to 11.7 °C in sunlit soil and 5.9 to 8.6 °C in shaded soil in the olive orchard, and from 7.6 to 12.0 °C in sunlit soil and from 6.4 to 8.6 °C in shaded soil in the almond orchard (Table 1). Both soil sensors recorded similar average minimum temperatures during the coldest period. Within each orchard, minimum sunlit soil temperatures peaked seasonally (p < 0.05) from spring to early autumn (e.g., olive: March to September; almond: May to October) (Table 1).

For both crops, the monthly maximum soil temperatures varied from 11.0 ± 0.29 to 36.9 ± 0.14 °C in sunlit soil and from 9.6 ± 0.13 to 30.5 ± 0.11 °C in shaded soil (Table 1). During the warmest months (May to October), maximum soil temperatures ranged from 22.1 ± 0.71 to 36.9 ± 0.14 °C in sunlit soil and from 20.3 ± 0.55 to 30.5 ± 0.11 °C in shaded soil. (Table 1). During the coldest months, the monthly maximum soil temperatures decreased by 0.9 to 4.6 °C in sunlit soil and by 2.4 to 8.2 °C in shaded soil in the olive orchard, and by 2.4 to 7.7 °C in sunlit soil and by 2.6 to 8.3 °C in shaded soil in the almond orchard. During the warmest period, from May to October, the monthly minimum temperature in soil decreased by 0.3 to 2.2 °C in sunlit soil and by 5.2 to 9.0 °C in shaded soil in the olive orchard, and by −0.7 to 2.0 °C in sunlit soil and by 3.5 to 7.5 °C in shaded soil in the almond orchard (Table 1). Differences in monthly maximum temperatures between the soil sensor positions varied by crop species. The sunlit sensor consistently recorded the highest temperatures (p < 0.05) in olive trees throughout the year, while in almond trees, this trend was noted only from May to September (Table 1). At hourly intervals, both minimum and maximum temperatures recorded by soil sensors were nearly constant during the day in winter, as illustrated in Figure 3 and Figure 4.

3.2. Seasonal Relationships Between Air and Sensor Temperatures

Air temperature showed a strong positive linear relationship with temperatures recorded in branch xylem, trunk xylem, and the soil–root interface within each orchard (Figure 5, Figure 6 and Figure 7). In these models, the slope parameter describes the change in sensor-recorded temperature per 1 °C change in T_air. Slopes close to 1 indicate that the internal compartment closely tracks air temperature, whereas slopes < 1 reflect thermal buffering of the measured plant compartment. Slopes > 1, observed occasionally for branch sensors, suggest a steeper exposure-related response than air temperature.

3.2.1. Minimum Temperatures

The regression models analyzing the relationship between T_air and sensor-recorded minimum temperatures demonstrated a strong linear trend across all seasons for branch and trunk sensors (R² > 0.96) (Figure 5) and soil sensors (0.60 < R² < 0.89) (Figure 7a–d). The slope parameter estimated from the tree sensor models ranged from 0.95 to 1.05 for the olive orchard and from 1.00 to 1.10 for the almond orchard, with most cases significantly different from 1 (p < 0.05). Slopes for soil sensor models showed a wider range, from 0.46 to 1.03 in the olive orchard (Figure 7a,c) and from 0.55 to 1.16 in the almond orchard (Figure 7b,d).

In each orchard, the relationship between T_air and sensor-recorded minimum temperatures was modulated by season (F > 194.78, p < 0.0001), sensor position within the tree (F > 3170.12, p < 0.0001), and their interaction (F > 70.61, p < 0.0001). However, most of the variation in the mean square error (MSE) explained by the linear model was attributed to the position of the sensor, accounting for 92.3% of the MSE in both orchards. Season and interaction accounted for 6.2 and 1.4% of the MSE in the olive orchard, and 5.7 and 2.1% of the MSE in the almond orchard, respectively (Supplementary Tables S1 and S2). These results indicate that, for minimum temperatures, the main source of variation relative to air temperature is the compartment sampled (branch, trunk, or soil–root interface), whereas seasonal effects and their interaction with position play a comparatively smaller role.

The estimated marginal means of sensor-recorded minimum temperatures for the different seasons are shown in

Table 2. Least-square adjusted means from the linear model relating minimum and maximum plant and soil-system temperatures with air minimum and maximum temperatures for each sensor type and season in olive and almond orchards.

Temperature Crop/Sensor	Season
	Winter		Spring		Summer		Autumn
Minimum temperatures
Olive orchard
Branch N	9.97	C d	10.17	AB d	10.60	A e	10.38	BC d
Branch S	10.06	C d	10.30	AB d	10.63	A e	10.45	BC d
Trunk 1 cm N	10.78	C c	11.49	B c	12.69	A c	11.65	B c
Trunk 1 cm S	10.77	C c	11.41	B c	12.74	A c	11.46	B c
Trunk 4 cm N	10.63	C c	11.24	B c	12.23	A d	11.38	B c
Trunk 4 cm S	10.63	C c	11.25	B c	12.29	A d	11.37	B c
Shade soil	15.53	D b	15.87	C b	18.66	A b	17.72	B b
Sunlit soil	16.06	D a	18.02	C a	22.25	A a	18.46	B a
Almond orchard
Branch N	10.71	B c	11.24	A c	11.54	A d	11.29	A d
Trunk 1 cm N	11.27	C b	11.94	B b	12.76	A c	12.02	B c
Trunk 4 cm S	11.11	C bc	11.81	B b	13.07	A c	12.20	B c
Shade soil	16.46	C a	17.10	B a	19.30	A b	16.72	BC b
Sunlit soil	16.10	C a	17.41	B a	22.51	A a	17.59	B a
Maximum temperatures
Olive orchard
Branch N	25.51	D a	28.16	B a	31.02	A a	27.20	C b
Branch S	25.88	C a	27.80	B a	30.49	A a	28.17	B a
Trunk 1 cm N	21.92	B c	21.07	C d	23.07	A d	22.81	A d
Trunk 1 cm S	23.85	A b	22.65	C c	23.16	B d	23.29	B d
Trunk 4 cm N	22.22	D c	22.76	C c	26.10	A b	23.37	B d
Trunk 4 cm S	25.43	A a	23.67	C b	24.37	B c	24.46	B c
Shade soil	17.01	D e	17.56	C e	19.57	B e	20.66	A e
Sunlit soil	19.47	D d	22.30	C c	26.67	A b	24.03	B c
Almond orchard
Branch N	25.94	B b	26.36	C a	27.61	A a	27.31	A a
Trunk 1 cm N	29.37	B a	23.97	B b	24.15	A b	27.37	A a
Trunk 4 cm S	28.94	A a	23.60	C b	21.64	C c	24.71	B b
Shade soil	16.41	C c	19.45	B d	21.51	A c	21.04	A d
Sunlit soil	16.78	D c	20.52	C c	26.95	A a	22.84	B c

For each sensor type (rows), values sharing the same uppercase letter and for each season (columns), values sharing the same lowercase letter are not significantly different according to the post hoc Tukey’s honestly significant difference (HSD) test, at a significance level of p < 0.05.

. Within each orchard, a consistent seasonal pattern was observed across sensor positions. The average sensor minimum temperature predicted by the linear model at the average T_air was significantly higher (p < 0.05) in summer, decreased in spring and autumn, and was lowest (p < 0.05) in winter (Table 2; Figure 5). Within each season, soil sensors recorded the highest minimum temperatures compared to tree sensors, indicating stronger buffering of minimum temperatures at the soil–root interface. Specifically, sunlit soil sensors registered the highest average minimum temperature, followed by shaded soil sensors. Among tree sensors, trunk sensors at depths of 1 and 4 cm, and branch sensors recorded progressively lower temperatures in that order. No significant differences (p ≥ 0.05) were found between the north- and south-oriented sensors (Table 2; Figure 5 and Figure 7).

3.2.2. Maximum Temperatures

There was also a strong linear relationship between T_air and sensor-recorded maximum temperatures across all sensor positions in both olive and almond trees and soil, across sensor positions within each orchard and season, with 43 out of 52 models showing 0.70 < R² ≤ 0.98 (Figure 6 and Figure 7). The slopes of the fitted models for temperatures recorded by sensors in tree branches ranged from 0.90 to 1.06 within each orchard, similar to the values estimated for minimum temperature. Interestingly, the slopes of the models fitted to trunk sensors were lower than 0.83 for olive orchard sensors and 0.71 for almond orchard sensors, except during the winter months when almond trees are leafless. For spring and summer, the slopes ranged from 0.51 to 0.76 for olive orchard sensors and from 0.42 to 0.71 for almond orchard sensors. These lower slopes for trunk sensors indicate stronger thermal buffering of maximum temperature in the trunk relative to T_air and to branch xylem (Figure 6). A wider range was found for the slopes in models fitted to soil sensors, varying from 0.34 to 0.84 in olive orchard sensors and from 0.35 to 0.855 in almond orchard sensors (Figure 7e–h).

Maximum temperatures showed a similar trend to that described for minimum temperatures. The relationship between T_air and sensor-recorded maximum temperatures within each orchard was influenced by the season (F > 72.25, p < 0.0001), sensor position within the tree (F > 932.67, p < 0.0001), and their interaction (F > 115.93, p < 0.0001) (Supplementary Tables S3 and S4). However, most of the variation in MSE explained by the linear model was attributed to sensor position, accounting for 84.0% of MSE in the olive orchard and 75.9% of MSE in the almond orchard. Season and the interaction contributed 12.0 and 4.0% of MSE, respectively, in the olive orchard (Supplementary Table S3), and 5.9 and 18.3% of MSE, respectively, in the almond orchard (Supplementary Table S4). These results indicate that, for maximum temperatures, the primary source of variation relative to air temperature is the sampled compartment (branch, trunk, or soil–root interface), whereas seasonal effects and their interaction with position play a comparatively smaller role.

The estimated marginal means of sensor-recorded maximum temperatures across different seasons are presented in Table 2. Within each orchard, a consistent seasonal pattern was observed across sensor positions. The average maximum temperature recorded by the sensors, as predicted by the linear model at the average T_air was significantly higher in summer (p < 0.05), decreased in autumn and spring, and was lowest in winter (p < 0.05) (Table 2; Figure 6 and Figure 7). An exception was noted for trunk sensors at a depth of 1 or 4 cm with a southern orientation, which recorded higher average maximum temperatures in winter and lower ones in spring.

For olive trees within each season, the average maximum temperatures relative to average T_air were highest in branches, followed by trunk sensors at a depth of 4 cm, trunk sensors at a depth of 1 cm, and sunlit soil, with the lowest temperature recorded in shaded soil. For almond trees, during spring, summer, and autumn, branch sensors recorded significantly higher average maximum temperatures (p < 0.05), followed by trunk sensors at a depth of 1 cm, sunlit soil, with the lowest temperatures being recorded in trunks at a depth of 4 cm and in shaded soil. In contrast, in winter, the lowest average maximum temperatures were recorded by both soil sensors, increasing in branches and trunk at a depth of 1 cm, and being highest for the trunk at a depth of 4 cm (Table 2, Figure 5, Figure 6 and Figure 7).

4. Discussion

This study provides the first comprehensive field-based analysis of seasonal and daily temperature dynamics within the xylem tissue and the soil–root interface of almond and olive trees under field conditions in a Mediterranean-type environment. The results showed how ambient air temperature relates to microclimates within specific plant tissues, highlighting the significant effects of sensor position within the tree and seasonal variation on temperature dynamics. In general, branch temperatures were closely coupled to air temperatures across all seasons. However, during spring and summer, trunk temperatures at depths of 1 and 4 cm, as well as within the soil–root system interface at a depth of 20 cm, remained notably lower than maximum air temperatures. In contrast, during autumn and winter, the soil–root system interface at a depth of 20 cm consistently exhibited higher temperatures than the air. These findings underscore the buffering capacity of plant tissues and soil, which mitigates temperature extremes experienced by the xylem tissues and soil–root interface. This buffering is critical for understanding plant physiological responses and pathogen dynamics, as temperature fluctuations in these microenvironments can influence both pathogen development and plant stress resilience.

4.1. Internal Temperature Dynamics and Tree Physiology

Thermocouples can provide accurate and reliable temperature measurements under field conditions, enabling precise measurement of xylem temperature in trees [17]. However, several considerations must be taken into account when using thermocouples to minimize potential sources of error. Physically, type K thermocouples may be subjected to slow thermal drift, but this is insignificant at temperatures lower than 200 °C. In our study, all the thermocouples were purposely obtained from the same reel to minimize the likelihood of alloy-composition variation during manufacturing, and additional aging-related effects during the measurement period are unlikely. In addition, each recorded measurement was the average of 10 individual measurements; therefore, random errors are reduced. The air temperature was measured under a shielded enclosure complying with World Meteorological Organization specifications; therefore, no relevant measurement errors are expected for air temperature. The effect of probe contact with the xylem was also considered, and in our study, the probes were positioned in the xylem after precision-drilled holes sized for a press-fit insertion; therefore, the contact with the xylem was as homogeneous and consistent as possible. The use of thermocouples has largely been confined to forest ecosystems [18,19,20], and their deployment within the xylem of field-grown crops is still relatively rare. In this study, we have shown that thermocouples can be effectively used in woody crops such as olive and almond trees.

Thermocouple measurements have shown variable outcomes depending on factors such as species, season, solar exposure, and position (height and depth) of the thermocouple within the tree. Gansert et al. [38] reported that in silver birch (Betula pendula), xylem sap temperatures measured at a depth of 8 mm in trunks were similar to air temperatures during March and April. However, in August, xylem sap temperature began to diverge from air temperature. Peña Quiñones et al. [21] also found that minimum xylem temperatures recorded at depths of 2 and 3 mm at various heights in grapevine trunks did not differ significantly from air temperature, while branches or thinner trunks were more prone to temperature variations induced by solar radiation due to greater exposure to direct sunlight, smaller diameters, and lower bark lignification [6,14,29]. These attributes are known to enhance convective heat transfer with the atmosphere compared to thicker trunks, which provide a buffering effect that moderates temperature fluctuations.

The buffering effect observed in the trunks at depths of 1 and 4 cm in olive and almond trees can be primarily attributed to the low thermal conductivity of mature wood and the insulating properties of well-lignified bark [14,39]. Edwards and Hanson [13] recorded trunk temperatures that were 6 to 8 °C lower than air temperature at a depth of 2 cm in oak and red maple trees in spring; this difference was attributed to the shading effect of leaves, which reduced direct solar exposure. Furthermore, Quick [40] noted that stems near the base of shrubs and roots experience less temperature variation than upper branches, which are more exposed to solar radiation, and experience reduced convective heat exchange, while the proximity to larger wood masses near the base acts as a thermal buffer.

Our study showed that xylem temperature varied with trunk depth. In general, sensors placed deeper within the trunk recorded lower temperatures. Lindroth et al. [41] reported that trunk temperatures were 5 to 7 °C lower than air temperatures at depths of 3 and 7 cm in Norway spruce and Scots pine. Similarly, Stockfors [5] reported lower trunk temperatures in Picea abies trees when measured with deeper trunk sensors. Our summer measurements in almond trees support these findings, as sensors at a depth of 4 cm recorded lower temperatures than those at a depth of 1 cm. However, in olive trees, during the same period, higher temperatures were registered at a depth of 4 cm. This variation may result from the anisotropic nature of wood, which leads to heterogeneous thermal properties [42]. Sap flow is known to influence xylem temperature, with higher sap flow associated with lower xylem temperature, likely due to a cooling effect from transpirational water uptake [43]. Variations in sap flow may explain the observed temperature patterns. For example, López-Bernal et al. [30] observed higher sap flow on the south side of olive trees at trunk depths of 1.3 and 2.2 cm, which may account for the lower temperature recorded at the depth of 1 cm in the olive trees in our study. If so, the cooler outer rings could be due to a larger flux of sap originating from cooler soil, assuming convective sap flux decreases with trunk depth. Reduced sap flow with increased depth has been observed in some species [31,32].

The xylem temperature in woody plants is affected by both conductive and convective processes. The rising sap is initially at the soil temperature, and it exchanges thermal energy with the conductive xylem, which transports it upward, and the parenchymal tissue around the vessels, until it approaches the leaf temperature by the time it reaches the leaves. This convective heat exchange often generates temperature gradients of significant magnitude along the pathway through the trunk and branches [44]. These gradients follow pseudo-sinusoidal patterns over daily and seasonal cycles and may influence the sap flow measurements made with thermal dissipation or heat balance methods [45], often requiring correction techniques [46]. The spatially dynamic modeling of trunk and branch temperature is extremely difficult, as it depends on the instantaneous temperatures of soil and atmosphere (the former varying with individual root depth), the sap flow and velocity, the wood thermal properties and local dimensions, the pathway length, and finally radiation and wind. The attempts made in literature [6] are very partial models and remain far from the descriptive approach of this study.

While several studies [6,13,14] have shown that tree orientation can influence trunk temperature, our model, implemented in the olive orchard with sensors placed on the north and south sides, did not reveal significant temperature differences. This may be due to the fact that most studies highlighting the impact of tree orientation, including those cited above, were conducted in natural forests, where tree orientation leads to differences in exposure to solar radiation. In contrast, orchards typically have more controlled and uniform conditions, with factors such as tree spacing and canopy management minimizing variations in solar exposure.

4.2. Soil Thermal Behavior

In our study, soil temperature measurements indicated a significant buffering effect in the soil environment for both minimum and maximum temperatures, which can extend to root tissues. Thermocouples installed within the root tissues of almond and olive trees recorded temperatures similar to those at the surrounding soil–root interface in the same orchards used in this study (Supplementary Figure S1). One key factor contributing to this buffering effect is the relatively low thermal conductivity of soil, which dampens temperature fluctuations with increasing depth [47,48]. Bertrand et al. [49] reported lower temperatures with increasing soil depth from 5 to 50 cm. Similarly, Pregitzer et al. [50] observed significant temperature fluctuations of over 10 °C between depths of 1 and 15 or 20 cm, particularly during May and June. Canopy coverage also plays a role in regulating soil temperature by moderating the amount of solar radiation reaching the soil surface [48], which can influence photosynthetic rates, evapotranspiration, and soil microbial activity. Our findings are consistent with these results, as sensors placed in sunny areas exhibited greater temperature fluctuations compared to those in shaded areas.

4.3. Potential Implications for Plant-Associated Pathogens

Temperature is a critical determinant of plant pathogen activity and the development of vascular diseases [23,24]. Among the most damaging vascular pathogens affecting olive and almond are X. fastidiosa and V. dahliae, both mesophilic microorganisms capable of persisting across a wide range of climatic conditions worldwide. Because these pathogens are xylem-restricted, the characterization of the thermal dynamics within xylem is essential for predicting their survival and potential spread. Our results showed a marked thermal buffering effect within tree trunks, with maximum trunk temperatures ranging from 26.1 to 35.6 °C during summer and minimum temperatures ranging from 3.7 to 7.7 °C during winter (Table 1). Importantly, these thermal extremes were typically sustained for no more than 12 h per day. When considered in the context of published survival thresholds for both pathogens [9,25,27,51], the buffered xylem temperatures recorded in our study are consistent with conditions that may favor pathogen persistence, particularly during periods of extreme environmental stress characteristic of Mediterranean-type climates.

Soil and root temperature are also important for pathogen survival. V. dahliae may persist in soil through microsclerotia, which remain viable for years [52]. Microsclerotia can tolerate high soil temperatures, with 30% survival at 35 °C [26]. For X. fastidiosa, root colonization has been documented in almond [53] and olive [54], among other crops. In an infected almond orchard, despite the air temperature dropping to −28 °C, soil temperature at a depth of 50 cm remained above 6 °C. This soil buffering effect allowed X. fastidiosa cells to overwinter in almond roots, although a reduction in bacterial load and remission of symptoms occurred the following year. In our study, maximum soil temperatures during the warmest month reached nearly 37 °C for sunlit sensors and slightly above 30 °C for shaded sensors. While these conditions approach the upper survival limit for V. dahliae, soil insulating properties may maintain microsclerotia viability. In contrast, minimum soil temperatures remained above 8 °C in winter, a range that would not impair X. fastidiosa survival according to several authors [25,51,53].

Overall, our findings highlight that the thermal buffering occurring within the xylem and the root–soil interface creates a microclimate that may allow vascular pathogens to withstand extreme temperatures. Understanding these microenvironmental conditions is crucial for refining disease-risk models in Mediterranean-type climates and explaining the climatic suitability of vascular pathogens threatening woody crops. Since temperature is a key predictor for pathogen risk models [9,10], incorporating air-temperature data corrected for trunk and soil buffering effects could improve the accuracy of temperature-driven disease-risk models.

5. Conclusions

The results of this study demonstrate that temperatures measured within olive and almond xylem tissues differ significantly from air temperature, particularly during spring and summer. Relying solely on air temperature data recorded from standard weather stations may therefore lead to inaccurate estimates of the thermal conditions experienced by woody tissues, with implications for understanding the physiological and ecological processes occurring within xylem tissues.

Xylem temperature sensors deployed under field conditions in woody crops, such as those used in the present study, highlight the potential of continuous plant-based monitoring systems. In addition to providing a simple and minimally invasive approach for monitoring tissue-level thermal dynamics, these systems could support a range of agronomic applications. Moreover, because pathogen risk models often rely on temperature as a key predictor of vascular pathogen establishment and spread, incorporating air-temperature data corrected for trunk and soil thermal buffering effects could contribute to developing more accurate and biologically realistic disease-risk models. These approaches could also support the early detection of water stress, improve irrigation management and water-use efficiency, and help identify canopy or soil-management practices that influence plant thermal dynamics. Such approaches also offer opportunities to contribute to the development of precision-agriculture tools, enabling better decision-making under variable environmental conditions.

Future modeling efforts would benefit from incorporating additional explanatory variables to better resolve the factors governing the relationship between air temperature and xylem thermal dynamics. In particular, the role of irrigation remains uncertain, given the relatively homogeneous field conditions of this study, and experimental comparisons across contrasting irrigation regimes would help clarify its influence on xylem thermal buffering. Similarly, direct measurements of sap flow from roots to canopy could clarify its contribution to xylem thermal regulation. Beyond plant physiological drivers, orchard structural and management factors—including ground-cover management, plant density, and exposure to solar radiation—are also likely to modulate microclimatic patterns within xylem tissues and across the soil–root temperature profile. From a methodological perspective, monitoring the soil–root interface at multiple depths would improve the representation of belowground thermal gradients, while targeted analyses during extreme events (e.g., heatwaves or cold spells) would help quantify the limits of thermal buffering under stress conditions. Together, these approaches would strengthen the interpretation of xylem thermal buffering and improve the transferability of empirical relationships across management systems and climatic contexts. Further research should therefore aim to quantify the relative contribution of physiological, structural, and environmental drivers shaping woody-tissue thermal buffering, particularly in the context of projected climate change scenarios.