Carbon capture, photosynthesis, and leaf gas exchange of shade tree species and Arabica coffee varieties in coffee agroforestry systems in Veracruz state, Mexico

Selection of shade tree species and Arabica coffee varieties in agroforestry systems

Seven native tree species were selected from an initial screening of 50 conducted by Flores-Ortiz et al. (2025) based on conservation status, growth rate, and agroecological utility for coffee production. Species used for firewood or fuel were excluded, along with cultivated Persea americana Mill. (Hass avocado) due to its intensive domestication. For the selected shade trees species: Inga inicuil Schltdl. and Cham. Ex G. Don (Ii); Inga vera Willd. (Iv); Inga punctata Willd. (Ip); Erythrina americana Mill. (Ea); Psidium guajava L. (Pg); Persea schiedeana Turcz. (Ps), and Heliocarpus appendiculatus Nees (Ha), dendrometric, chlorophyll fluorescence, and gas exchange characteristics were measured in trees aged ≤30 years.

Due to the heterogeneity of C. arabica varieties found in the coffee farms in the region, chlorophyll fluorescence and gas exchange parameters measured were focused on Oro Azteca (Oa), Garnica (G), Costa Rica 95 (Cr), Tipica (T), and Catuai amarillo (Ca) plants aged 4–6 years. To enable comparison, identical physiological measurements were conducted for Oro Azteca variety plants under unshaded conditions.

This selection aimed to represent a functional diversity of native species with potential for enhancing ecosystem services in coffee farms.

Study area

The shade tree species and coffee varieties studied make up agroforestry systems in a traditional polyculture configuration and in the intermediate secondary succession stage. Coffee plants in unshaded conditions were part of an unshaded monoculture system. The tree density, considering trees between 10–15 m in height and >5 cm in diameter at breast height (DBH, at 1.30 m) for this type of coffee agroforestry system and coffee region, has been reported at ∼1,000 trees ha⁻¹ (López-Gómez, Williams-Linera & Manson, 2008; Williams-Linera & Lorea, 2009).

Study sites were located in central Veracruz on shaded and unshaded coffee farms in the municipalities of Teocelo (19°23′36″N, 96°59′9.4″W, at an elevation of 1,117 m a. s. l., average air temperature of 23.43 ± 0.37 °C and 70.4 ± 5.95% of relative humidity, or RH) and Xico (19°25′23.5″N, 96°55′42.6″W, at an elevation of 1,053 m a. s. l., average air temperature of 27.53 ± 0.86 °C and 60.23 ± 4.16% of RH for shaded and unshaded conditions, respectively (Fig. 1; Map created using the Free and Open Source QGIS).

To ensure environmental homogeneity, measurements were taken during two consecutive cool-season months (October–November 2022). This period exhibited typical climatic conditions for 2016–2022 trends, with precipitation, temperature, and cloud cover within expected ranges (Weather Spark, 2024). The cool season brought average maxima of 24 °C (Teocelo) and 26 °C (Xico), alongside ≥1 mm/h precipitation and 80% cloud cover. Historical records show Teocelo’s temperature extremes (21–35 °C, 1945–2020), while Xico ranged from 9–12 °C (minima) to 21–22 °C (maxima, 1966–2023) (Servicio Meteorológico Nacional, https://smn.conagua.gob.mx/es/climatologia/informacion-climatologica/normales-climatologicas-por-estado?estado=ver (accessed 17 November 2025)). Both sites share comparable annual precipitation (1,847 and 2,091.8 mm for Xico and Teocelo, respectively; period 1991–2020) and stable conditions due to their proximity.

Sampling strategy

The sampling design was structured to capture intra- and inter-specific ecophysiological variability. For each of the seven shade tree species, three mature and reproductive individuals were randomly selected (individuals = three per species). On each individual, three fully expanded, healthy leaves located on the first lateral branch at the base of the trunk were marked and measured (leaves = three per individual). This yielded in a total of nine measurements per species (measurements = nine). The same protocol was applied to the associated coffee plants under the canopy of each tree species, sampling three individuals per variety and three leaves attached to plagiotropic branches per coffee bush. For statistical analysis, data from the three leaves from the same individual were averaged to obtain a single representative value per individual (n = 3 individuals per species/variety). These averages were subsequently used for comparative analyses between species and varieties.

Dendrometric parameters of shade trees and carbon stocks determination

Measurements were made of the total height and the DBH of three mature and reproductive individuals of each of the seven shade trees. Using these morphometric parameters, the available allometric equations were applied to calculate each tree’s above ground biomass or AGB (Table 1). The allometric equations used to calculate the AGBs for Ha and Pg were at species level, at genus level for Inga spp. (Ii, Iv and Ip) and for Erythrin a sp. (Ea), and at tropical forest level for Ps (Rojas-García et al., 2015; Ortiz-Ceballos et al., 2020) (Table 1). AGB was converted to biomass carbon stock (CS) by multiplying by 0.47, representing the standard carbon fraction in tree biomass (IPCC, 2006). CS reflects both tree’s ability to grow new cells and its carbon storage potential (IPCC, 2022).

Due to variation in tree age, 10-year normalised CS were obtained by a simple proportionality rule for each tree species. This normalisation enables standardised comparison of carbon storage potential while recognising that species-specific traits, such as shade tolerance and wood density, may influence long-term sequestration rates. Finally, we extrapolated hectare-scale carbon storage using the tree density found in the literature (references in Table 2).

Chlorophyll fluorescence, gas exchange analyses of shade trees and coffee plants

In-situ measurements were performed on leaves from the seven tree species and associated- coffee plants located under the canopy of the seven tree species, which correspond to five varieties of Arabica coffee (Oa, G, Cr, T, and Ca) using a portable infrared gas analyser (IRGA) LI-6400XT (Licor, Lincoln, NE, USA) equipped with a fluorometric cell. For shade trees, we assessed leaves attached to lateral branches closest to the understory layer, between 2–4 m from the ground via ladders and climbing equipment. These branches exhibited sympodial growth with a predominantly horizontal or obliquely orientated architecture, characteristic of species such as Inga spp. (Troll model), Ea (Champagnat model), and Pg (Roux model) (Vester, 2002; de Reffye et al., 2008). Coffee measurements were taken from leaves attached to plagiotropic branches at 1.30–2 m height.

Chlorophyll fluorescence analysis was performed in dark-adapted leaf tissues (30 min). Minimum (F0) and maximum fluorescence (Fm) in light-adapted tissues was measured by the saturation pulse method (λ =630 nm, Q >7,000 µmol m⁻² s⁻¹, 6s). Based on these signals, variable fluorescence in the dark (Fv=Fm-F0) and quantum efficiency (Fv/Fm) were calculated (Silva et al., 2010; Rakocevic et al., 2022). Fv/Fm is frequently used to estimate the photochemical efficiency of PSII (Niinemets & Kull, 2001; Lepeduš et al., 2005; Zavafer & Mancilla, 2021). After that, the gas exchange parameters CO₂ assimilation rate, stomatal conductance, transpiration, and intercellular CO₂ (C_i) were measured in the same leaves attached to the branch. All measurements were conducted between 9:00 h and 11:00 h under controlled conditions, with the IRGA operated as an open system with a photon flux density of 1,000 µmol m⁻² s⁻¹, a leaf temperature of 25 °C, and an environmental air CO₂ concentration of ∼420 ppm.

Subsequent gas exchange measurements, it proceeded to estimate incident solar radiation in the understory layer between shaded and unshaded systems. PAR was measured under all tree species’ canopies and sun-exposed Oro Azteca coffee plants at midday; this period sees peak solar radiation and thermal stress (Meili et al., 2021; Kohl, Niether & Abdulai, 2024).

Instantaneous leaf water-use efficiency (WUE = A/E) was calculated as the CO₂ assimilation-transpiration ratio (Hatfield & Dold, 2019).

Gravimetric and analytical chemical analysis: moisture and nitrogen content of coffee leaves samples

Following immediately after chlorophyll fluorescence and leaf gas exchange measurements, the same leaves from Oro Azteca coffee plants grown under both shaded and unshaded conditions were excised at the base of the petiole. Collected leaves were stored in sealed plastic containers to prevent moisture loss until further analysis. Subsequently, the samples were processed within the next two days after field sampling. The leaves from each condition were macerated and pooled, then 0.5 g was taken for moisture and nitrogen analysis. Gravimetric methods comparing weight before and after drying determined moisture content based on the following equation: (1)${\displaystyle H\left(\text{\%}\right)=\frac{FW-DW}{FW}\times 100.}$

This parameter refers to the proportion of water present in fresh leaf tissue relative to its total weight, where H (%) is the moisture content expressed as a percentage, FW is the fresh weight, and DW is the dry weight.

We employed the semimicro–Kjeldahl method (Nelson & Sommers, 1980; DOF, 2019) for digested 50 mg of dried leaf material using a digestion/distillation apparatus (Labconco^®) to quantify organic nitrogen content.

Statistical analysis

Statistical analyses were performed on the averaged data from nine measurements (described above). The assumption of normality (Shapiro–Wilk, p > 0.01) was verified for all datasets. All data met the assumption; thus, no transformations were required.

One-way analysis of variance (ANOVA) was used to identify significant differences (p < 0.001) among shade tree species for each parameter. For comparisons involving coffee varieties under different shade conditions, a one-way ANOVA was also employed (p < 0.001). Where ANOVA indicated significant differences, the post-hoc Tukey’s test was applied for pairwise comparisons. Comparisons were conducted for each parameter between all shade tree species. Coffee variety comparisons were performed between the Oro Azteca variety in shaded conditions vs unshaded conditions. Additionally, comparisons were conducted considering only coffee varieties in shaded conditions. WUE statistical differences between shade trees and between coffee varieties were identified following the same methodology (p < 0.001). Comparisons of Oro Azteca moisture and nitrogen content between shaded and unshaded conditions were performed using a two-tailed Student’s t-test (p < 0.01) for independent samples.

To analyse the possible clustering of the seven shade tree species and coffee varieties based on their physiological traits, a principal component analysis (PCA) was conducted on both datasets comprising all measured parameters (Fv/Fm, CO₂ assimilation rate, stomatal conductance, transpiration rate, and intercellular CO₂ concentration). PCA analysis proceeded by extracting principal components from the correlation matrix of the variables, with the selection criterion being the retention of components that collectively explained at least 75% of the total variance.

GraphPad Prism^® version 9.5.1 for macOS (GraphPad Software, San Diego, CA, USA; http://www.graphpad.com) was used for all statistical analysis (accessed in January 2023).

Dendrometric parameters of shade trees and carbon stocks determination

Ea and Ps showed the highest mean AGB and CS values per tree, while Ha and Ii displayed intermediate levels, and Iv, Pg, and Ip had the lowest (Table 1). The 10-year normalized CS values showed similar patterns across species, where Ps and Ea had the highest carbon storage capacities, followed by lower values for the remaining species (Table 1). At the hectare scale, Ii achieved the highest annual CS (20.99 Mg C ha⁻¹), followed by Ea (19.83 Mg C ha⁻¹), Ps (11.01 Mg C ha⁻¹), and Ha (6.23 Mg C ha⁻¹), with Iv, Ip and Pg showing the lowest values (0.72 and 0.53 Mg C ha⁻¹, respectively) (Table 2).

Chlorophyll fluorescence, gas exchange and PCA of shade trees and coffee plants

The Fv/Fm values of shade tree species (range 0.74–0.81) and coffee plants (0.73–0.80) showed consistent trends, with no significant differences observed (Table S1). Similarly, no significant differences were found between shaded and unshaded Oa or among other coffee varieties under shaded conditions (Table S2).

CO₂ assimilation rates differed significantly among species (F _(6,56) = 118.3, p < 0.001), with highest values in Ea (8.37 ± 0.90 µmol CO₂m²/s), Pg (9.18 ± 1.00), and Ha (7.99 ± 1.16), and lowest in Ps (2.45 ± 0.40), Iv (2.21 ± 1.19), and Ii (1.67 ± 0.38) (Fig. 2B, Table S1). For Oa coffee, shaded plants showed higher rates than unshaded (F _(3,32) = 54.42, p < 0.001), with Ip-Oa (5.68 ± 0.65) and Ha-Oa (5.76 ± 0.91) exhibiting the highest values among shaded varieties (F_(6,56) = 16.82, p < 0.001) (Figs. 3A, 4A, Table S2).

Transpiration rates varied significantly (F_(6,56) = 65.43, p < 0.001), with highest values in Pg (2.59 ± 0.22 mmol H₂O m²/s) and Ha (2.34 ± 0.33), and lowest in Ip (1.15 ± 0.29), Ea (0.71 ± 0.19), and Ps (0.48 ± 0.10) (Fig. 2B, Table S1). Among Oa coffee variety, unshaded Oa (0.53 ± 0.04) and Ip-Oa (0.53 ± 0.30) showed highest transpiration, while Ha-Oa (0.30 ± 0.18) and Ps-Oa (0.23 ± 0.02) showed lowest (F_(2,32) = 6.488, p < 0.001) (Fig. 3B, Table S2). Among shaded coffee varieties, Ii-G showed the highest transpiration rate (1.02 ± 0.25) while Pg-Ca the lowest (0.27 ± 0.06) (Fig. 4B, Table S2).

Stomatal conductance followed similar patterns, with highest values in Pg (266.88 ± 29.01 mmol H₂O m²/s), Iv (245.44 ± 85.25), and Ha (224.88 ± 35.51), and lowest in Ps (26.57 ±12.98) with significant differences between them (F_(6,56) = 19.33, p < 0.001) (Fig. 2C, Table S1). For the Oa cultivar, stomatal conductance varied significantly between unshaded conditions and shaded Ha-Oa and Ps-Oa (F_(3,32) = 8.697, p < 0.001) (Fig. 3C, Table S2). For shaded coffee, Iv-Cr (145.70 ± 67.86) and Ii-G (100.17 ± 34.05) showed highest conductance, while Pg-Ca (37.76 ± 24.04), Ip-Oa (31.65 ± 11.74), and Ha-Oa (15.60 ± 4.06) showed lowest (F_(6,56) = 24.02, p < 0.001) (Fig. 4C, Table S2).

C_i values among shade trees exhibited notable variations, with the highest values observed in Iv (444 ± 17.34 µmol CO₂/mol), Ha (427.44 ± 12.97), and Ii (419.78 ± 41.88) and lower values in Pg (359.33 ± 10.54) and Ea (325.78 ± 21.76), with significant differences between them (F_(6,56) = 48.22, p < 0.001) (Fig. 2D, Table S1). For the Oa variety, C_i values were influenced by shaded and unshaded conditions, with Ip-Oa showing the highest values (448.77 ± 82.59) and Ps-Oa the lowest (223.44 ± 46.47) (F_(6,56) = 48.22, p < 0.001) (Fig. 3D, Table S2). Among shaded coffee varieties, Ip-Oa (448.77 ± 82.59), Iv-Cr (443.33 ± 162.69), and Ii-G (410.55 ± 26.82) demonstrated higher C_i values, while Ea-T (284.88 ± 34.46), Pg-Ca (284.88 ± 34.46), Ha-Oa, (305.66 ± 82.79) and Ps-Oa (223.44 ± 46.47) showed lower values, showing significant differences between them (F_(6,56) = 11.89, p < 0.001) (Fig. 4D, Table S2).

PAR levels differed significantly among shade tree species (F_(6,56) = 64.15, p < 0.001), ranging from 23.3 ± 4.06 µmol m⁻² s⁻¹ (Ps) to 92.2 ± 6.66 (Ea) (Table S1). For Oa coffee variety, unshaded coffee received 1,427 ± 124.0 µmol m⁻² s⁻¹ versus 8.44 (Ps-Oa)–25.11 (Ip-Oa) under shade (F_(6,56) = 47.68, p < 0.001) (Table S2). Among shaded coffee varieties, Iv-Cr (47.1 ± 9.18) and Pg-Ca (32.2 ± 5.38) displayed the highest PAR values, while Ps-Oa (8.44 ± 1.08) and Ha-Oa (14.8 ± 6.64) showed the lowest measurements, with significant differences between conditions (F_(6,56) = 47.68, p < 0.001) (Table S2).

WUE values showed significant variation among the seven tree species, ranging from 0.83 ± 0.15 µmol CO₂/mmol H₂O (Ii) to 16.92 ± 11.05 µmol CO₂/mmol H₂O (Ea) (F_(6,56) = 5.810; p < 0.001) (Fig. 5A, Table S1). For the Oa coffee variety, WUE values ranged from 4.61 ± 10.06 µmol CO₂/mmol H₂O in unshaded conditions to 21.62 ± 11.27 µmol CO₂/mmol H₂O (Ha-Oa) (F_(3,32) = 8.538; p < 0.001) (Fig. 5B, Table S2). Among shaded Arabica cultivars, values varied from 7.12 ± 3.78 µmol CO₂/mmol H₂O (Iv-Cr) to 21.62 ± 11.27 µmol CO₂/mmol H₂O (Ha-Oa) (Fig. 5C, Table S2); no statistically significant differences were observed.

For the PCA for the seven shade tree species, PC1 accounted for 43.57% of the variance (eigenvalue = 2.178), while PC2 contributed 33.43% (eigenvalue = 1.671), cumulatively explaining 77% of the dataset variability (Table S3). In contrast, the PCA for the coffee varieties (including Oro Azteca) under shaded and unshaded conditions revealed a stronger influence of PC1, which explained 49.04% of the variance (eigenvalue = 2.452), with PC2 adding 29.98% (eigenvalue = 1.499), resulting in a slightly higher cumulative variance (79.02%) (Table S4).

For the seven shade tree species, PC1 was strongly associated with stomatal conductance (loading = 0.882) and transpiration (0.866), with Pg and Ps contributed most to PC1 (42.5% and 37.1%, respectively) (Fig. 6A, Table S3). PC2 showed positive loading for CO₂ assimilation rate (0.594) and negative loading for intercellular CO₂ concentration (−0.937), with Ea contributing 73.9% to this axis (Fig. 6A, Table S3).

For coffee varieties, PC1 had negative loadings for stomatal conductance (−0.906), transpiration (−0.897), and intercellular CO₂ (−0.884) (Fig. 6B; Table S4), with Ps-Oa (22.0%) and Ha-Oa (5.9%) showing highest contributions (Table S4). PC2 showed positive loading for photochemical efficiency (Fv/Fm = 0.766) and negative loading for CO₂ assimilation rate (−0.884) (Fig. 6B, Table S4), with unshaded Oro Azteca contributing 39.3% (Table S4).

Correlation analyses for shade trees, stomatal conductance, and transpiration were strongly positively linked (r = 0.825), while CO₂ assimilation and intercellular CO₂ were negatively correlated (r = −0.689) (Table S3). For coffee varieties, stomatal conductance, transpiration, and intercellular CO₂ showed strong positive correlations (r = 0.698–0.731), whereas CO₂ assimilation and Fv/Fm were moderately negatively associated (r = −0.420) (Table S4).

Gravimetric and analytical chemical analysis: moisture and nitrogen content of coffee leaves samples

The shaded Oa samples showed a slightly higher moisture content (55.56%) compared to the unshaded samples (53.97%) (Table S5). Also, shaded Oa samples exhibited a significantly higher nitrogen content (2.83 ± 0.06%) compared to unshaded samples (2.54 ± 0.02%) (t(4) = 7.6; p < 0.01) (Table S5).

Dendrometric parameters of shade trees and carbon stocks determination

The seven shade tree species showed distinct photosynthetic traits, growth and biomass allocation patterns, leading to inherent variability in carbon accumulation, making direct interspecific comparisons of AGB and CS unreliable (Poorter et al., 2008; Chave et al., 2014). Works such as that of Garza-Lau et al. (2020) showed these types of variations in agroforestry systems in the state of Veracruz. Farmers typically manage tree density through propagation programmes, though optimal balances between competition and productivity need further study. Regional factors like altitude, slope, and shade management affect phenology and influence tree traits and ecosystem services (Bewley & Black, 1994; Cerda et al., 2017; Asanok et al., 2024). These constraints and cultivation practices collectively determine tree densities. Coffee plants carbon sequestration in the study area, although it was not determined in our work, is highly variable (0–12 Mg ha⁻¹; Valdés-Velarde et al., 2022), depending on variety and density, making extrapolation and comparison difficult, even within the same study area.

Three ecological patterns emerge when examining species contributions. First, the Inga genus (Fabaceae), particularly I. inicuil, achieves remarkable carbon capture through numerical dominance rather than individual tree performance. With 660 trees ha⁻¹, I. inicuil accounts for 33.77% of system carbon uptake, achieving 198 Mg C ha⁻¹, nearly double the sequestration reported for other Inga, Erythrina, and Musaceae species (91.64–115.5 Mg C ha⁻¹; Haber, 2001). Reported densities for I. inicuil range from 100 to 800 trees ha⁻¹, averaging 250–350 trees ha⁻¹ (Barradas & Fanjul, 1986; Soto-Pinto et al., 2001). After 10 years of growth, I. inicuil exhibited a CS of 20.9 Mg C ha⁻¹ at 200 trees ha⁻¹, exceeding values for Inga densiflora Benth. (24.3 Mg C ha⁻¹ at 400 trees ha⁻¹) with similar age and size parameters (Salazar-Figueroa, 1985; Kursten & Burschel, 1993). However, these values were three times lower than those reported for I. inicuil in Oaxaca, Mexico (64.3 Mg C ha⁻¹ at 164 trees ha⁻¹; Hernández-Vásquez et al., 2012; Alessandrini et al., 2011; Tellez et al., 2020), underscoring how regional factors like altitude and microclimate interact with species physiology. When comparing agroforestry systems across different states, it is important to consider that variations in management practices, soil characteristics, and climatic regimes may lead to substantial differences in carbon sequestration. Consequently, direct comparisons between regions should be interpreted with caution, as local variables shape ecosystem functioning. Although I. punctata had lower AGB and CS, its high density (representing 20–40% of total trees) contributed significantly to system-level carbon capture (Soto-Pinto et al., 2001). Incorporating both I. punctata and I. vera in agroforestry configurations may achieve carbon stocks of 91.64 Mg C ha⁻¹ (Haber, 2001).

Second, E. americana and P. schiedeana follow a quality-over-quantity approach. Their substantial trunk diameters and heights enable just 40 trees to capture carbon equivalent to 94.48% and 52.47%, respectively of the carbon captured by 200 I. inicuil trees. However, biological constraints, including seed dormancy in E. americana (Bewley & Black, 1994; Bonfil-Sanders, Cajero-Lázaro & Evans, 2008) and extensive crown-canopy development in P. schiedeana (Niembro, 1992; Vázquez-Torres, Campos-Jiménez & Juárez-Fragoso, 2017), naturally restrict their planting densities in managed agroforestry systems.

Third, the complementary roles of remaining species enhance system functionality. H. appendiculatus, representing 16–20% of tree strata in Chiapas coffee farms (Soto-Pinto et al., 2001; Castillo-Capitán et al., 2014), accounted for 10.03% of total carbon assimilation despite representing only 6% of trees in this study. P. guajava, which constituted 4–5% of tree density in coffee plantations (Soto-Pinto et al., 2001), showed the lowest CS values (6.53% of total CS at 10 years), consistent with prior findings (Nava et al., 2009) and attributable to its average height of 3–8 m (Heuzé et al., 2017). P. guajava can be incorporated into agroforestry systems to enhance carbon storage, particularly in leaves and roots, with a whole calculated CS ranging between 0.27 and 4.19 Mg ha–1 in 2- to 10-year-old orchards (Naik et al., 2021). Additionally, it provides valuable firewood and fruits (Somarriba, 1988; Pascarella et al., 2000; Miceli-Méndez, Ferguson & Ramírez-Marcial, 2008).

Our calculated carbon sequestration of shade tree species was highly specific. I. inicuil achieved system-level dominance through high density, whereas E. americana and P. schiedeana excelled through high individual tree efficiency. Optimising carbon capture therefore requires strategic species combinations that leverage these complementary ecological patterns and management practices.

Chlorophyll fluorescence, gas exchange and PCA of shade trees and coffee plants

The physiological performance among shade trees and coffee varieties revealed distinct functional strategies shaped by interspecific variation in quantum efficiency, stomatal behaviour, and carbon capture potential. Most species maintained Fv/Fm values above the 0.75 threshold for fully functional PSII (Genty, Briantais & Baker, 1989; Lepeduš et al., 2005), indicating robust photochemical activity. This divergence underscores how intrinsic physiological traits interact with environmental conditions to determine carbon capture efficiency.

PCA identified three clusters among shade tree species, revealing distinct adaptive strategies:

Gravimetric and analytical chemical analysis: moisture and nitrogen content of coffee leaves samples

The lower air temperature and light intensity in shaded areas contribute to higher moisture content in coffee leaves by increasing relative humidity, which lowers the VPD (Schwerbrock & Leuschner, 2017). In contrast, coffee plants in unshaded conditions are subjected to environmental variables more likely to trigger plant stress responses compared to shaded plants and previous evidence on the nitrogen content of leaves has demonstrated that, in stressful situations, increased leaf nitrogen availability promotes the activation and maintenance of photoprotective systems to avert photooxidation (Fahl et al., 1994; Ramalho et al., 2000), which allows shaded leaves to adapt more efficiently to different irradiation conditions than fully sun-exposed leaves (Araujo et al., 2008).

Gravimetric and chemical analysis suggests that shaded conditions could have promoted higher leaf moisture content due to increased humidity and lower VPD. Conversely, unshaded coffee plants exhibited traits suggesting environmental stress, with nitrogen likely allocated to support photoprotection mechanisms. These findings highlight a physiological divergence in coffee leaf characteristics mediated by shade management practices.

Some limitations of this study should be considered. The measurements were conducted during the cool season to ensure environmental homogeneity; therefore, physiological responses during warmer or drier periods remain to be investigated. Furthermore, while we infer the effects of shade on microclimate, direct measurements of air temperature, relative humidity, and VPD at the leaf level were not concurrently recorded. Future longitudinal studies across seasons, incorporating continuous microclimate monitoring alongside physiological measurements, would provide a more comprehensive understanding of plant responses to dynamic environmental conditions. Establishing permanent sample plots would also allow for tracking long-term carbon storage and physiological acclimation. Despite these limitations, our study offers a robust snapshot of the physiological mechanisms governing species performance and provides critical insights for the immediate selection and management of shade trees.