Multiple Dimensions of Functional Traits in Subtropical Montane Mosses

Abstract

The study of functional traits and their relationship to trade-offs has provided valuable insights into how plants adapt to environmental changes. Nonetheless, further research is necessary to fully comprehend the subtropical montane trade-off patterns in moss functional traits and the impact of environmental gradients on the correlation of these traits. To address this gap, we conducted a study of 11 moss species (7 families, 9 genera) in 54 patches from two subtropical mountain ranges, examining 40 functional traits related to photosynthesis, nutrients, water retention, and architecture. Through principal component analysis (PCA) and bi-variate correlation analysis, our findings reveal a strong correlation between light capture and nutrient assimilation strategies, as evidenced by the coordination between the traits of light capture and nutrient per area along a main principal component. Interestingly, we observed a trend towards smaller leaves and leaf cells in species with higher capacity for photosynthesis and metabolism, consistent with global trait spectra in vascular plants. However, we found that the trade-off between light capture and nutrient assimilation strategies was independent of water-holding capacity at shoot levels. Instead, we found that water-holding capacity was closely associated with nutrient utilization, energy metabolism, chlorophyll synthesis, and the primary process of photosynthesis. Our results highlight the multiple dimensions of functional traits in subtropical montane mosses and demonstrate that variation in these traits is driven by water availability, slope, and canopy density. Overall, our study provides valuable insights into the co-variation of moss traits and how environmental changes may impact mosses and ecosystem function.

1. Introduction

Understanding the impact of environmental changes on biological processes and the potential for mitigating or exacerbating these changes is a significant challenge in plant ecology [1]. Functional traits are key indicators that can clarify plant adaptation strategies to environmental changes [2,3]. Response traits reflect the response of plant species to environmental gradients, while effect traits mediate the impact of ecosystem processes on the environment [3]. Despite bryophytes (hornworts, liverworts, and mosses) being the second most diverse group of land plants, with about 20,000 extant species, and playing vital roles in forest ecosystems, the number of studies focused on their comparative physiological performance is much smaller than those on tracheophytes [4,5]. The ecological strategies of bryophytes cause functional trait variation and species distribution differences in microhabitats [6]. Studying the differences in functional traits of moss species across various environments can reveal how they adapt to their surroundings. This information can help us to better understand how plants in general deal with environmental changes [7,8].

Compared to tracheophytes, mosses are primitive plants with lower leaf mass per area and gas exchange. However, they exhibit similar trait coordination between nutrient investment, dry mass, and photosynthesis during ecological differentiation [9,10]. In contrast to tracheophytes, the economic spectrum in mosses has been found in the shoot- or canopy-levels of moss rather than the leaf [8,9,11], because the shoot and canopy area of mosses is an analog of the leaf area of vascular plants. Additionally, mosses are typical poikilohydric plants that can rapidly absorb and lose water depending on the water availability in their immediate environment. Different species of mosses have varying strategies for adapting their photosynthesis to water surpluses and shortages, probably through taking more time to ‘prepare’ or adjust photosynthesis quickly [12,13]. Although water-holding capacity and economic traits were both reported to have a tight connection to photosynthesis progress [5,10,13,14], the decoupling of water-holding capacity to economic traits has been reported in boreal mosses [14]. Therefore, it is necessary to conduct a more comprehensive study encompassing a broader range of species and geographic locations to understand the pattern of the economic spectrum in bryophytes.

Different factors can determine the variability of traits, such as the impact of environmental gradients, trait trade-offs (which refer to the covariation of traits among species), and interspecific trait variability [15]. Variations in the environment can influence resource allocation between different types of traits, such as architectural traits (e.g., structural support) and physiological functional traits (e.g., photosynthesis) [5,16]. These traits impose physical and energetic constraints on the use and allocation of resources in plants. Trait trade-offs in the plant trait spectrum derive from covariation between architectural traits related to plant maintenance (high allocation to large structures) and physiological traits related to their metabolism (high resource capture rate), reflecting the costs and benefits of leaf investment, not only for spermatophytes, but also found in the shoot level (measured by the area projected on the ground or transversal to the shoot, based on individuals) and the canopy level (measured by the ground area of the moss canopy in its natural station) of bryophytes [6,9,14]. Increasing the number of species in a community may broaden the trait distribution as new species are introduced. Environmental variability, trait trade-offs, and interspecies trait variability in bryophytes are intermingled in generating the large diversity observed in the economic spectrum [6,14,15].

We conducted a study to investigate the functional traits of mosses in subtropical montane regions. The study was carried out in the Chebaling and Guanshan National Nature Reserves in South China. To conduct the investigation, we analyzed the co-variation of architectural, physiological, and chemical traits in 11 dominant moss species in subtropical montane regions and explored their responses to environmental gradients. Our study aimed to achieve the following objectives: (1) to investigate the subtropical montane pattern of trade-offs in moss functional traits, and (2) to determine how environmental gradients shape the co-variation of functional traits in subtropical montane regions. Our study provides insights into the co-variation of moss traits and how environmental changes may affect mosses’ functions and ecosystem function.

2. Materials and Methods

2.1. Study Site, Focal Species, Sampling, and Pretreatment

Sampling was conducted at the Chebaling National Nature Reserve (CBL, 16,110.7 hm², 114°07′ E–114°16′ E, 24°40′–24°46′ N) and Guanshan National Nature Reserve (GS, 11,500.5 hm², 114°29′–114°45′ E, 28°30′–28°40′ N) of China (Figure 1), between August and September 2021. The mean annual temperature is 15.99 °C in Guanshan and 19.31 °C in Chebaling, and the mean annual precipitation is 1790.5 mm in Guanshan and 1449.9 mm in Chebaling. We selected 11 prominent species as objective species in the two regions, though only 7 of these species were present in both regions (refer to

Table 1. The family, life forms, substrate, and sampled region of the 11 species of mosses in the study. CBL: Chebaling, GS: Guanshan.

Family	Species	Life-Form	Substrate	Region
Fissidentaceae	Fissidens nobilis	Fan	Soil/Rock	CBL/GS
Hypnaceae	Calohypnum plumiforme	Weft	Soil/Rock	CBL/GS
	Vesicularia reticulata	Weft	Soil/Rock/Rotten wood	GS
Leucobryaceae	Leucobryum bowringii	Cushion	Soil	CBL/GS
	Campylopus umbellatus	Turf	Soil/Rock	GS
Mniaceae	Plagiomnium acutum	Rough mat	Soil	CBL/GS
	Plagiomnium succulentum	Rough mat	Rock/Rotten wood	CBL/GS
	Plagiomnium maximoviczii	Turf	Soil	GS
Polytrichaceae	Pogonatum contortum	Turf	Soil	CBL/GS
Thuidiaceae	Thuidium cymbifolium	Weft	Soil/Rock	CBL/GS
Entodonaceae	Entodon macropodus	Weft	Rock/Rotten Wood	CBL

). Moss patches that form monospecific layers of our objective species were selected to collect samples for trait analysis. A total of 54 distinct moss patches in the two regions (24 patches in CBL and 30 patches in GS, Table S2) were selected (with 3 replicates for each studied species), ensuring a minimum distance of 20 m between each. To provide comprehensive documentation, we carefully recorded the environmental conditions of each patch, including its latitude, longitude, elevation, substrate, slope, and canopy density. To determine canopy density, we captured an image of the vascular plant cover above the moss surface using a smartphone camera and a wide-angle lens. The canopy density was then calculated from the captured image using ImageJ (1.54d) image analysis software. Furthermore, we developed a scoring system to assess the water availability in microhabitats based on the distance between the patch’s shade and its water source. The scoring system assigns a score of 1 if there is no significant water source within 2 m of the patch and a canopy shade of less than or equal to 0.3. A score of 2 indicates no significant water source within 2 m but with a canopy shade greater than 0.3. A score of 3 indicates a significant water source within 2 m of the patch, directly influenced by water. The higher the score, the stronger the water availability.

The mosses were collected with the underlying substrate, sealed in plastic bags, and brought to the laboratory as soon as possible. The samples were kept moist and shaded. Before the experiments, the litter, bark, and mixed-in mosses of other species were removed carefully. All the samples were washed with distilled water to clean off the dust and mud. Dead tissue was removed, and the green moss sections were collected as the final samples.

2.2. Photosynthetic Traits

The content of chlorophyll pigments per mass (Chl_mass) was measured using the spectrophotometer method, which is detailed in method S1. A portable open-flow infrared gas analyzer (GFS-3000, Walz, Effeltrich, Germany) with a moss cuvette (cuvette for Lichens/Mosses 3010-V32) was used to measure photosynthetic light responses (see Methods S2 of Supplementary Materials). We determined light response parameters from plots of CO₂ assimilation rate (A) vs. PAR. The assimilation rate per mass (A_mass), dark respiration rate per mass (R_mass), the light compensation point (I_c), and the saturation irradiance (I_s) were estimated by fitting the light response curves to a nonrectangular hyperbola photosynthetic model [17]. The fitting coefficients were all above 0.99. Assimilation rates according to nitrogen (A_N), phosphorus (A_P), and chlorophyll (A_Chl) were calculated by dividing Amass by N, P, and chlorophyll concentration per mass (N_mass, P_mass, and Chl_mass), respectively.

Light-saturated assimilation rate, dark respiration rate, nitrogen, phosphorus, and chlorophyll per shoot area (A_area, R_area, N_area, P_area, Chl_area) by multiplying A_mass, R_mass, N_mass, P_mass, and Chl_mass by the shoot mass per area (SMA) will be described in Section 2.4.

A PAM fluorometer (PAM-2500, Walz, Effeltrich, Germany) was used to measure chlorophyll fluorescence. The measurement and calculation of the optimal photochemical efficiency of PSII (Fv/Fm), the actual photochemical efficiency of PSII (φPSII), photochemical quenching (qP), and photochemical quenching (NPQ) were described in Methods S3 of Supplementary Materials.

2.3. Water Retention Capacity

Water-holding capacity and the rate of water loss were determined through water loss experiments. Samples underwent the same wetting pretreatments and spatial arrangements used for measuring the photosynthetic water-response curve. Wet samples’ initial weight was recorded and they were kept under moderately cool and moist conditions (around 25 °C and 80% humidity) until dry. Sample weight was measured every 10 min for the first 60 min and at incremental intervals of 12, 15, 20, 30, 60, 120, 180, and 270 min, and then every 3 h until constant weights were obtained. After measurements, samples were dried in an oven at 70 °C for 48 h to determine the dry weight.

Water content (WC) was calculated as:

Water content = (fresh weight − dry weight)/dry weight,

and WC_max of each sample was calculated using the maximum (initial) fresh weight.

The water loss curves were fitted using exponential decay regressions [18]:

water content = y + a × e^(−DC*t)

where t (min) is the evaporative time and a, DC (decaying constant), and y are coefficients. The R² values were all above 0.98.

2.4. Architectural Traits

Architectural traits potentially linked with cell- and leaf-level structure and function were measured for five shoots of each species that were mounted and photographed at 20–400× magnification. We measured the cross-sectional area of the central strand (CAA), cross-sectional area of the shoot (CSA), leaf area (LA), leaf length (LL), leaf width at the widest part (LW), length/width ratio (L/W), perimeter²/area ratio (P²/A), costa length (CL) and costa width at the widest part (CW) using ImageJ software (National Institutes of Health, Bethesda, MD, USA), and 4–6 cells from the central area of the lamina were measured to determine cell lumen area (CLA), cell lumen length (CLL), cell lumen width (CLW), and cell wall thickness (CWT) of the leaf; cells per area (CPA) of the leaf were determined from 15 to 20 cells by circling a center area (diameter = 0.05 mm) on leaves. We determined SMA as leaf mass divided by shoot projected area [9,14].

2.5. Chemical Analysis

The amount of organic carbon (C_mass) and nitrogen (N_mass) in each unit of mass was measured using a Vario Macro Cube Elemental Analyzer (Elementar Analysensysteme GmbH, Langenselbold, Germany). The phosphorus content per unit of mass (P_mass) was determined using the Mo–Sb antispectrophotography method. By multiplying the mass-based nutrient concentrations and SMA, the C, N, and P contents based on area (C_area, N_area and P_area) were derived. Finally, the mass-based concentrations were used to calculate the stoichiometric ratios of C:N, C:P, and N:P.

2.6. Data Analysis

We used R software version 4.3.1 and the package ggplot2 for all statistical and graphical analyses [19,20]. To assess the resource use strategy of the samples, we examined trait associations using principal component analysis (PCA) through the package FactoMineR [21]. We first used a total of 42 traits (

Table 2. Functional trait symbols and units.

Traits	Code	Unit	Mean ± SD	CV
Photosynthesis progress
Saturation irradiance	Is	µmol photons m−2 s−1	748.71 ± 247.57	0.33
Compensation irradiance	Ic	µmol photons m−2 s−1	21.72 ± 13.45	0.62
Light-saturated assimilation rate per mass	Amass	nmol CO2 g−1 s−1	21.01 ± 11.79	0.56
Light-saturated assimilation rate per area	Aarea	μmol CO2 m−2 s−1	75.24 ± 44.34	0.69
Light-saturated assimilation rate per chlorophyll	AChl	nmol CO2 g−1 s−1	2.86 ± 2.15	0.75
Photosynthetic N use efficiency	AN	nmol CO2 g−1 s−1	1.14 ± 0.62	0.54
Photosynthetic P use efficiency	AP	nmol CO2 g−1 s−1	16.9 ± 11.52	0.68
Dark respiration per mass	Rdmass	nmol CO2 g−1 s−1	20.51 ± 9.60	0.47
Dark respiration per area	Rdarea	μmol CO2 m−2 s−1	0.79 ± 0.47	0.60
Dark respiration rate/maximum photosynthetic rate	Rd/A	Dimensionless	1.55 ± 2.06	1.32
Optimal photochemical efficiency of PSII	Fv/Fm	Dimensionless	0.66 ± 0.03	0.05
Actual photochemical efficiency of PSII	φPSII	Dimensionless	0.39 ± 0.14	0.35
Photochemical quenching	qP	Dimensionless	0.44 ± 0.12	0.28
Nonphotochemical quenching	NPQ	Dimensionless	3.36 ± 1.56	0.46
Chlorophyll per mass	Chlmass	mg g−1	8.36 ± 2.91	0.35
Chlorophyll per area	Chlarea	g m−2	3.58 ± 1.63	0.47
Water retention capacity
Maximum water content	WCmax	% of d.w.	432.16 ± 191.13	0.44
Water loss decaying constant	DC	×10−3	16.01 ± 7.82	0.48
Nutrient content
Organic carbon per mass	Cmass	mg g−1	377.57 ± 31.26	0.08
Organic carbon per area	Carea	g m−2	14.87 ± 7.94	0.47
Total nitrogen per mass	Nmass	mg g−1	18.82 ± 4.60	0.24
Total nitrogen per area	Narea	g m−2	0.69 ± 0.28	0.40
Total phosphorus per mass	Pmass	mg g−1	1.44 ± 0.60	0.41
Total nitrogen per area	Parea	g m−2	0.05 ± 0.02	0.45
C:N ratio	C/N	Dimensionless	21.22 ± 5.23	0.25
C:P ratio	C/P	Dimensionless	313.67 ± 141.89	0.45
N:P ratio	N/P	Dimensionless	14.45 ± 4.13	0.28
Architectural traits
Costa length	CL	mm	3.75 ± 2.70	0.72
Costa width	CW	mm	24.52 ± 21.51	0.88
Leaf length	LL	mm	4.12 ± 2.32	0.56
Maximum leaf width	LW	mm	1.29 ± 0.84	0.65
Leaf area	LA	mm	4.09 ± 4.47	1.09
Shoot mass per area	SMA	g m−2	39.01 ± 17.88	0.46
Cross-axis area of central strand	CAA	mm2	0.003 ± 0.002	0.79
Cross-sectional area of shoot	CSA	mm2	0.08 ± 0.03	0.40
Perimeter2⁄area	P2/A	Dimensionless	52.29 ± 30.75	0.58
Leaf length-to-width ratio	L/W	Dimensionless	3.75 ± 2.38	0.63
Cells per area	CPA	cells mm−2	666.70 ± 536.87	0.81
Cell lumen length	CLL	µm	41.98 ± 30/32	0.72
Cell lumen width	CLW	µm	12.83 ± 12.4	0.97
Cell lumen area	CLA	µm2	582.44 ± 981.12	1.68
Cell wall thickness	CWT	µm	2.23 ± 0.80	0.36

) in the PCA analysis and found that the traits mainly projected their coordination relationship in the first four axes (Figure S1A). Based on our analysis, we found that the variance in CSA and CAA projected in each of the first four axes was less than 20%. Additionally, the total variance of CSA and CAA projected in the first four axes was less than 40%. This information suggests that CSA and CCA are not closely associated with the first four axes. As a result, we decided to exclude them from our analysis. The selected variables were log-transformed as appropriate to improve linearity and scaled to unit variance before analysis. Considering the possible non-linear relationship between traits, Spearman and Pearson correlation coefficients were both computed to estimate the bi-variate relationship between functional traits (Figures S2 and S3).

To assess how trait spectra relate to environmental gradients, we used the function envfit in the vegan package to fit environment gradients onto the ordination and to test their fit through permutation tests [22]. To investigate the relationship between the environment gradients with significant influence and the four primary PCA axes, we performed additional tests using fixed-effects and mixed-effects models. In these models, we used the first four PCA axis sample scores as the response variables, with slope, canopy density, and wetness as fixed effects and species as random effects in the mixed-effects models. We achieved this by fitting fixed-effects models using the lm function and linear mixed-effects models using the lmer function in the lme4 package [23] and the expansion packages, including lmerTest [24], and sjstats [25]. To test whether regions had an effect for those species occurring in more than one region, we performed likelihood-ratio tests comparing the full model (i.e., species/habitat combination as a random effect) with a nested subset (i.e., species as a random effect). To further analyze the influence of environmental gradients on the functional traits, we evaluated the bi-variate relationship between functional traits and environmental gradients using Spearman and Pearson correlation.

3. Results

Principal Component Analysis (PCA) was performed to examine the relationship between physiological and morphological traits based on the functional characteristics screened. The analysis resulted in the identification of the first four principal component axes (PCs), which accounted for 67.91% of the total variance and effectively captured the trade-off between physiology and morphology traits, as depicted in Figure 2. Fitting environmental variables onto the ordination provided strong evidence that moss trait variation, as described by the PCA, was overall associated with slope (R² = 0.35, p < 0.001), with canopy density (R² = 0.21, p = 0.026), and with water availability (R² = 0.38, p < 0.001). Upon further analysis of fixed effects models (

Table 3. The effects of environmental gradients on the first four principal components estimated using fixed effect and mixed effect models. Mixed-effect models with species or species nested within regions as random effects were, respectively, established.

Models	Statistics		PC1	PC2	PC3	PC4
Fixed	Estimate	Slope	−0.01	−0.04 **	−0.03 *	−0.02
		Canopy density	−1.62 *	0.63	0.41	−1.01 *
		Water availability	−2.16 **	0.40	−0.71	−1.52 **
	R2	Slope	0.01	0.12 **	0.12 *	0.05
		Canopy density	0.10 *	0.02	0.01	0.09
		Water availability	0.16 **	0.01	0.03	0.18 **
Mixed (Species)	Estimate	Slope	0.00	−0.03 *	0.00	0.00
		Canopy density	−0.10	0.44	−0.13	−0.00
		Water availability	−0.14	−0.26	0.11	−1.23 *
	Marginal R2	Slope	0.00	0.09 *	0.00	0.00
		Canopy density	0.00	0.01	0.00	0.00
		Water availability	0.00	0.00	0.00	0.12 *
	Conditional R2	Slope	0.938 ***	0.30 *	0.81 ***	0.71 ***
		Canopy density	0.935 ***	0.33 **	0.81 ***	0.72 ***
		Water availability	0.935 ***	0.36 **	0.81 ***	0.72 ***
Mixed (Species nested within region)	Estimate	Slope	−0.00	−0.03 *	0.01	0.00
		Canopy density	−0.14	0.79	−0.14	0.04
		Water availability	−0.19	−0.54	0.02	−1.29 *
	Marginal R2	Slope	0.00	0.10 *	0.00	0.00
		Canopy density	0.00	0.04	0.00	0.00
		Water availability	0.00	0.02	0.00	0.13 *
	Conditional R2	Slope	0.95 ***	0.56 ***	0.85 ***	0.73 ***
		Canopy density	0.95 ***	0.59 ***	0.83 ***	0.74 ***
		Water availability	0.95 ***	0.60 ***	0.83 ***	0.83 ***

Significant results are bolded. * p ≤ 0.05. ** p ≤ 0.01. *** p ≤ 0.001.

), PC1 and PC4 exhibited moderate associations with canopy density and water availability, whereas PC2 and PC3 were moderately affected by slope. However, the mixed-effect models only support the significant relationship of PC2 to slope and PC4 to water availability. Coupled with the slight difference in the conditional R² between the mixed-effect and random-effect models presented in Table S3, we concluded that the fixed effects (slope, canopy density, and water availability) contributed much less to the variation in PC scores intra-species than inter-species. Moreover, likelihood-ratio tests showed only PC2 with a significant difference between region/species combination and species alone across the four axes (Table S3), indicating that only a small portion of the variation in traits within species was attributed to regions. Our results provide strong evidence that the variation of functional traits was driven by the environment gradients, although mainly through interspecific differentiation other than intraspecific variation.

The present study has observed a trade-off between economic and light-capturing traits in PC1. Species situated on the negative side of PC1 exhibited larger leaves and leaf cells (CLW: −0.91, LW: −0.84, CLA: −0.77, LA: −0.71, CLL: −0.43), as well as a higher nutrient concentration based on mass (N_mass: −0.69, P_mass: −0.53). This suggests that they have adapted to shaded environments and are primarily focused on capturing light. Conversely, species located on the positive side of PC1 displayed a higher capacity for nutrient and carbon assimilation (SMA:0.91, C_area: 0.93, C/N: 0.79, CPA: 0.77, N_area: 0.68, Rd_area: 0.60, C_mass: 0.56, A_area: 0.44, P_area: 0.43). Further bivariate correlation analysis demonstrated that several variables, including A_area, N_area, P_area, C_area, and SMA, displayed significant correlations with CLW, LW, LA, CLA, and CPA, as illustrated in Figures S2 and S3. It is noteworthy that the traits associated with water retention (WC_max, DC) projected their negligible variance on PC1, as illustrated in Figure S1. Further bivariate correlation analysis did not support a significant relationship between water retention traits and economic traits, except for Nmass and C/N (Figures S2 and S3). Our result revealed the independence of water retention ability from the economic spectrum observed on PC1.

A trade-off between energy and nutrient utilization was observed on PCA axis 2. The nutrient utilization efficiency in photosynthesis progress (A_P: 0.84, A_N: 0.71) decreased along the negative direction of the axis, while dark respiration (Rd/A: −0.73, I_c: −0.57, Rd_area: −0.44) increased. A significant relationship was observed only between A_P, A_N, and Rd/A, and these variables were found to have an explicit relationship with water-holding capacities (WC_max), as illustrated in Figures S2 and S3.

On PCA axis 3, a trade-off between the traits related to leaf size and structure (LL: 0.74, CW: 0.72, CL: 0.69, L/W: 0.67) and photosynthesis capacity (A_mass: −0.56, A_N: −0.48, A_area: −0.34) was found. However, only A_mass and L/W were significantly related. On PCA axis 4, a trade-off between the absorption and utilization of light energy was observed. Species with high photosynthetic efficiency per unit of chlorophyll (A_Chl: 0.69) were found in the positive direction, while species with higher chlorophyll content (Chl_area: −0.66, Chl_mass: −0.65) were found in the opposite direction. Chl_area and Chl_mass were significantly negatively related to CW, L/W, and CLL.

Additionally, we found that the influence of water availability on moss traits related to more aspects, including photosynthesis progress (A_area, Rd_area, I_s, Chl_mass), leaf and cell size (L/W, CPA, CLA, CLL, LW), and water retention capacity (DC), as shown in Figure 3. Slopes tended to relate to costa length (CL) and leaf length (LL), while canopy density showed a tight relationship to photosynthesis progress. Our results demonstrated the different mechanisms by which mosses adapt to environmental gradients.

4. Discussion

4.1. Morphological Adaptations to the Microhabitat Gradient

Morphological traits are physical characteristics that provide a clear indication of the functional strategies employed by different species [26]. A principal component analysis (PCA) of these traits, including CLW, LW, LA, CPA, CLA, and P²/A, revealed that most of their variability was projected onto PCA axis 1 (Figure S1), which was significantly related to gradients in water availability and canopy density. Our research found that as the gradient of canopy density and water availability decreased, leaves and leaf cells exhibited a trend towards miniaturization, while leaf margins (P^2/A) became more complex. These adaptations served to reduce heat loading and water loss in high light and xeric conditions. Similar morphological adaptations were reported in mosses distributed across three microhabitats along a vertical gradient, where branch-dwelling species had smaller and slimmer leaves and leaf cells than those on tree trunks [4]. Conversely, larger leaves facilitated greater levels of light interception [9], demonstrating the opposite end of the adaptation spectrum.

The costae found in mosses are known to provide structural support to leaves and water transportation. In this regard, leaves with longer and wider costae (Higher LL and LW) have been found to be associated with longer leaves (Higher LL) and higher leaf length/width ratios (L/W) [9]. A positive correlation between CL, CW, LL, and L/W was found to be significant in our study (Figures S2 and S3). These traits also played an essential role in the construction of PCA axis 3 (Figure S1), which was observed to be significantly related to the slope gradient. The slope gradient was negatively correlated with CL, CW, LL, and L/W (Figure 3). This result was consistent with previous studies, such as that by Turberville et al., which reported a smaller phyllid surface area in mosses growing on vertical substrates [27]. In general, there is a negative relationship between canopy and slope in montane regions, due to the high slope effect of strong intense light, scare water resource and frequent disturbance [28]. We speculate that steeper slopes led to a more rapid drain rate, which reduced the water retention capacity of mossy carpets, thereby limiting the growth of leaves.

Our result also showed the co-variance between morphological traits and water-retention capacity. WC_max was positively correlated with CLL, CLA, and CW, and DC was positively correlated with CPA. The significance of these relationships was demonstrated by bivariate correlation (Figures S2 and S3). Previous research studies have revealed that the covariance of morphological traits in moss is adjusts and adapts to the availability of light and water, leading to variations in the synergistic photosynthetic and water-holding capacity [4,7,29]. However, water-retention capacity was not related to the economical traits generally, such as SMA, N_area, and P_area.

4.2. Photosynthetic Adaptations to the Microhabitat Gradient

The physiological traits related to photosynthesis in moss varied depending on water availability and canopy density, as shown in Figure 3A,B. Moss photosynthetic capacity and dark respiration based on area were more sensitive to changes in gradients than those based on mass. This reflects the fact that moss photosynthetic capacity is dependent on photosynthetic area [10]. In our study, we observed that under light and xeric conditions, the species had higher A_area, I_s, φPSII, Rd_area, Rd_mass, and I_c, accompanied by smaller leaf size. A_area, φPSII, and I_s indicate an increased capacity to utilize greater available irradiance, while the higher Rd_area, Rd_mass, and I_c indicate a greater maintenance cost. These findings reflect the optimization of mosses to irradiance, like tracheophytes in their association of morphology with irradiance [9,30]. It is worth noting that the co-variance did not include the variation of chlorophyll per mass or area and fluorescence parameters, reflecting the decoupling of photosynthesis capacity to photosynthesis primary reaction in the study.

Our research has revealed that there is a trade-off between energy and nutrient use efficiency during photosynthesis, as indicated by the negative correlation between Rd/A and A_P and A_N. When the utilization efficiency of nitrogen (N) and phosphorus (P) is low, mosses increase their dark respiration rate to compensate for the loss of energy and nutrients. Conversely, when the utilization efficiency of N and P is high, mosses reduce their dark respiration rate and optimize the utilization of energy and nutrients produced during photosynthesis. This adjustment is closely linked to water-holding capacities (WC_max), suggesting that water retention capacity plays a critical role in photosynthesis [13].

4.3. Multiple Dimensions of Moss Functional Traits

The presence of leaf economics and hydraulic traits plays a crucial role in determining the rate of leaf photosynthesis in vascular plants. However, researchers have debated whether these two trait sets are entirely coordinated or if they allow for independent variation [2,31]. While the economic spectrum and trade-off between water relations and photosynthetic responses have been reported in mosses [9,13], there has been limited exploration of the relationship between the sets of traits related to economic and water retention capacity. Our result showed that the water retention ability of mosses was independent of the economic spectrum observed on the primary axes (Figure 2A,C). Grau-Andrés et al. [14] reported a similar result, finding that the economic spectrum of boreal mosses was driven by trade-offs in carbon, nutrient, and water conservation strategies. The architectural traits of the moss shoots and canopy were not significantly related to this spectrum, but instead formed a secondary axis.

Studies have shown that compared to vascular plants, mosses tend to exhibit higher susceptibility to environmental gradients, resulting in stronger effects of the environment on their traits [32,33]. As a result, unlike vascular plants, which tend to exhibit single main axes of trait variation, moss species tend to exhibit more independent dimensions [14]. However, most previous studies about moss functional traits have only focused on primary axes [9,10,34]. In our study, we found that the primary axis (PC1) was associated with architectural traits, nutrients, photosynthetic capacity, and dark respiration based on area, which are considered analogous to the functional traits of vascular plants that drive the “leaf economics spectrum” [9,35]. It is important to note that the main axis accounted for only 25.50% of the total trait variation, while the other three axes explained 42.41% of the variability. Additionally, although A_area showed a close relationship with economic traits (SMA, C_area, N_area) and water retention capacity, the correlation between economic traits and water retention capacity was weak. Our study showed functional trait combinations in more dimensions compared to previous moss studies by including as many traits as possible across a wider range of environmental gradients and species [5,11,12]. This allowed us to differentiate between architectural traits related to different functions, such as water-holding capacity and photosynthesis optimization, and assign them to different trait spectra. It is imperative to highlight that our findings require further scrutiny and evaluation across a broader range of moss species and regions in future research endeavors. This will enable us to validate our findings and establish a more comprehensive understanding of the underlying mechanisms that drive such outcomes.

5. Conclusions

Our research has revealed that subtropical montane mosses possess multifaceted functional traits. Despite sharing an economic spectrum with vascular plants, this spectrum appears to be unrelated to their water retention capacity at the shoot level and the primary process of photosynthesis. Because mosses are poikilohydric plants and can flourish in diverse global locations and microhabitats, the merging of their functions into a unified economic spectrum akin to vascular plants would be a momentous endeavor. To achieve this goal, additional research into the functional traits of mosses across a broader range of regions and species is imperative.