Climatic Rather than Edaphic Variables Determine Leaf C, N, P Stoichiometry of Deciduous Quercus Species

Purpose Leaf elemental stoichiometry is indicative of plant nutrient limitation, community composition, ecosystem function. Understanding the variations of leaf carbon (C), nitrogen (N), and phosphorus (P) stoichiometry at genus-level across large geographic regions and identifying their driving factors are important to predict species’ distribution range shifts affected by climate change. Methods Here, we determined the patterns of leaf concentrations ([ ]) and ratios ( / ) of C, N, P of ve deciduous oaks species (Quercus) across China covering ~ 20 latitude (~21–41˚ N) and longitude (~99–119˚ E) degrees, and detected their relationships with climatic, edaphic variables. Results


Introduction
Under the scenarios of climate change, the temperature and the precipitation regimes are predicted to increase and shift, respectively. Affected by climate change, some mid-latitude and semi-arid regions are projected to have ~ 1.5 to 2 times of the global warming rate in the next several decades; furthermore, most regions are projected to have more frequent and intensi ed precipitation extremes (IPCC 2021). The warming as well as the precipitation regime shifts have considerable effects on the cycles of carbon (C), water, and energy regionally and globally (IPCC 2021; Zeng et al. 2017), in consequence, integratively affect the vegetation activities (Gao et al. 2019), and ecosystems functions (Yuan and Chen 2015). For instance, the warming and the drought altered litter decomposition and organic matter mineralization via changing soil physical-chemical properties and microbial activities, leading to variations of plant nutrient availability and leaf stoichiometry (Reich and Oleksyn 2004;Yuan and Chen 2015).
Leaf stoichiometry drives basic physiological and ecological processes and indicates plant nutrient status, community composition, ecosystem functions (Lü et al. 2017). Carbon, nitrogen (N), and phosphorus (P) are important elements affecting the sustainability and the balance of biogeochemical cycles in terrestrial ecosystems (Gruber and Galloway 2008). Plant N and P regulate productivity and C sequestration in terrestrial ecosystems. Both the concentrations ([ ]) and the ratios ( / ) of leaf C, N and P are closely related to the utilization, the allocation of photosynthetic products and nutrients, and to a certain extent, to the ecological strategies of plant species (Ye and Wang 2021).
The spatial variability of plant stoichiometry can be affected by multiple environmental factors either at regional or global scales. At the regional scale, plant [N] and [P] decreased, and [C] increased, with increasing mean annual temperature (MAT) across China's terrestrial biomes (Tang et al. 2018); while leaf [N] and [P] of 753 terrestrial plant species increased with decreasing MAT but leaf N/P did not show signi cant changes in China (Han et al. 2005). However, when much more plant species (1900) across China were taken into consideration, plant functional type exhibited the greatest effects on most leaf nutrients, and precipitation explained more variations than temperature for leaf [N] but not for leaf [P] (Han et al. 2011). At the global scale, leaf [N] and [P] increased from the tropics to the mid-latitudes and remained stable or decreased at higher latitudes, and leaf N/P increased with temperature (Reich and Oleksyn 2004).
Up to date, the main large-scale (regional and global) patterns of plant stoichiometry have mainly been Large-scale ecological sampling on plant species at genus-level helps to better understand variations on leaf stoichiometry of C, N, P and provides opportunities to investigate the impacts of climate change on species' range shift (e.g., expansion, contraction, or displacement) ( [P], and C/N, C/P, N/P, and detected their relationships with climatic and edaphic variables. We aimed to explore the patterns of leaf stoichiometry of C, N, P at Quercus-level and to examine its potential linkage with climatic and edaphic variables. We hypothesized that: 1) the geographical patterns of leaf C, N, P stoichiometry of Quercus genus are similar to those at communitylevel at regional and global scales; 2) climatic variables determine the leaf patterns of C, N, P stoichiometry of the deciduous Quercus species, since climate governs the functional diversity and controls the geographic shifts of plants ( (Fick and Hijmans 2017). According to the distribution of the deciduous Quercus species, we sampled the oaks and ensured that the sampling sites were evenly distributed across latitude and longitude. At each sampling location, the mature and healthy leaves were collected from at least three individual trees for each species growing in natural forests.
Uniformity, the individual trees with diameter at breast height larger than 5 cm and tree height above 2 m were selected. Soils at 0 -20 cm depth below each tree were sampled at the same time. The leaf and soil samples were put in sealed plastic bags with ices and transported back laboratory within 8 hours.
In the laboratory, a sub-sample of the leaves was oven-dried to constant weight, and ground for [C], [N], and [P] measurements. Leaf [C] and [N] were determined using an elemental analyzer (Costech Analytical Technologies, Valencia, USA). Leaf [P] was determined by the molybdenum-antimony anti-colorimetric method (Dong 1997). Leaf C/N, C/P and N/P were calculated accordingly to leaf [C], [N] and [P]. Soil samples were air-dried, ground, and sieved for the measurement of pH, the concentrations of soil organic matter (SOM), soil total N (SN), soil total P (SP), and soil water content (SWC), according to the standard procedures (Liu et al. 1996). Speci cally, SWC was determined by oven drying about 20 g of fresh soil at 105°C for 48 h. The soil pH was measured in a 1: 2.5 soil: water (w/v) mixture using a glass-electrode meter (FiveEasyPlusTM FE28, Mettler Toledo, Switzerland). Soil organic matter (SOM) and total nitrogen (SN) concentrations were determined by using K 2 Cr 2 O 7 titration method and micro-Kjeldahl method, respectively. The concentration of soil total phosphorus (SP) was determined by the molybdenum antimony colorimetric method after digestion with H 2 SO 4 -HClO 4 .

Climatic variables
Climatic variables including MAT, the minimum temperature of the coldest month (Tmin), the maximum temperature of the warmest month (Tmax), mean temperature of the driest quarter (Tdry), mean temperature of the wettest quarter (Twet), temperature seasonality (TS), MAP, precipitation of the driest month (Pmin), precipitation of the wettest month (Pmax), precipitation of the coldest quarter (Pcold), precipitation of the warmest quarter (Pwarm), and precipitation seasonality (PS) during 1970-2000 at the sampling locations were obtained from the WorldClim Bioclimatic variables for WorldClim version 2 (Fick and Hijmans 2017). Temperature seasonality (TS) was the difference between the annual maximum and minimum temperature, i.e. the standard deviation of annual temperature multiplied by 100. Precipitation seasonality (PS) was the coe cient of variation calculated based on monthly rainfall data. The aridity index (AI, annual) is the ratio of MAP to the potential evapo-transpiration (PET), which was obtained from the global aridity index and potential evapo-transpiration climate database v2 (https://cgiarcsi.community/category/data/) (Trabucco and Zomer 2019).

Statistical analysis
Prior to analyses, the data of leaf C, N, P stoichiometry was tested for approximate normality (Shapiro-Wilk test) and log 10 -transformed to achieve normality when necessary. The climatic and edaphic variables were standardized via equation (A) to a mean of 0 and standard deviation of 1 to reduce the magnitude and multicollinearity (Du et al 2020).
To eliminate the interference of species on the relationships between leaf traits and latitude and longitude, general linear-mixed effect models (GLMEMs) were employed with species as the random effects (Crawley, 2012). We used step and lmer function in lmerTest R package to establish best-t models based on AICc of each model to nd out which and how environmental factors control the geographical patterns of leaf C, N, and P stoichiometry. We selected the rst ve environmental factors that signi cantly affected leaf C, N, P stoichiometry to calculate their relative importance through relaimpo R package (Chevan and Sutherland 1991). In addition, the relationships between the most important factors and leaf C, N, P stoichiometry were analyzed by general linear models (GLMs). The correlations among leaf stoichiometry were estimated by idaFast function in R package pcalg (Kalisch et al. 2012). Finally, the structural equation model (SEM) was utilized to explore the pathway of environmental factors that in uence leaf stoichiometry (Gerlach et al. 1979). Based on the goodness-oft criteria, including the probability level (P), comparative t index (CFI), the ratio of χ 2 to degrees of freedom (χ 2 / df), and root mean squared error of approximation (RMSEA), we selected the model when P > 0.05, χ 2 / df ≤ 2, CFI ≥ 0.98, and RMSEA having the lowest value (Schermelleh-Engel et al. 2003). The signi cant pathway coe cients were determined using 95% bootstrap con dence intervals. The following three potential pathways were considered in a hypothesis-oriented model: 1) climatic or edaphic variables will primarily in uence leaf [C], [N] or [P]; 2) leaf C/N, C/P and N/P that are controlled by leaf [C], [N] or [P] might relate directly to climatic or edaphic variables; 3) there are potential pathways among these leaf stoichiometric traits. All statistical analyses were performed using the R v.3.6.3 software (R Core Team 2019), except for SEMs, by IBM SPSS Amos 24 (SPSS Inc., Chicago, IL, USA). Signi cance was set at P < 0.05.

Results
Geographic patterns of leaf C, N, P stoichiometry  Table 1). The patterns of all leaf stoichiometric traits of the studied deciduous oaks varied signi cantly as indicated by the coe cient of variation (Table 1). Speci cally, leaf [C], [N], C/P, and N/P signi cantly decreased but leaf [P] and C/N increased with latitude (P < 0.01), however, only leaf [C] decreased signi cantly with increasing longitude (P < 0.01) (Figure 2). Determinants of leaf C, N, and P stoichiometry According to the stepwise regression models and relative importance analysis, we found that climatic rather than edaphic variables signi cantly affected the leaf stoichiometries of the deciduous oaks ( Figure  3, Table S1). Speci cally, both the MAP and AI controlled the variation of leaf [N], C/N and N/P (P < 0.05), while PET and TS determined the patterns of leaf [C] (P < 0.05). The MAT, Tmax, and TS were the driving factors on leaf [P] and C/P (P < 0.001).
For the temperature-related factors, leaf [C], [N], C/P, and N/P (Figure 4a, b, e, f) were positively while leaf [P] and C/N (Figure 4c and d) were negatively related to MAT; leaf C, N and P stoichiometry except for leaf [P] and C/N were negatively related to TS (Figure 4a, b, d, e); only leaf [P], C/P and N/P were signi cantly affected by Tmax (Figure 4c, The north-tosouth and west-to-east transects across China both re ect shifts from cold, dry to warm, humid conditions, although the temperature gradient is more obvious in the former and the moisture gradient is more pronounced in the latter ( Figure S1, S2). In this study, the signi cant relationships between MAT, Tmax, TS, MAP, PET, AI and leaf C, N, P stoichiometry implied that the nutrient status of the studied deciduous Quercus species was more likely affected by the co-regulation of temperature-and moisturerelated factors, due to the slowdown of litter decomposition and soil N mineralization at low precipitation and temperature (Finger et (Figure 4). The results revealed how climatic variables in uence leaf C, N, P stoichiometry of the deciduous Quercus species, which provided evidence for predicting nutrient strategies, which have considerable impacts on plant growth and survival, and further leading to potential distribution shifts of the eurytopic species, under the ongoing climate change.

Conclusions
We analyzed the spatial patterns of leaf C, N, P stoichiometry of ve deciduous Quercus species and their linkages to the environmental variables across a broad geographic range in China. We found that the deciduous Quercus species signi cantly decreased leaf [C], [N], C/P, and N/P, but increased leaf [P] and C/N, with the increasing latitude. Leaf stoichiometry except for leaf [C] had no signi cant trends along the longitudinal gradient. The climatic variables, i.e. mean annual temperature and precipitation, the max temperature of the warmest month, temperature seasonality, aridity index, and potential evapotranspiration, rather than the edaphic variables, were the determinants of the geographic patterns of leaf C, N, P stoichiometry of the studied deciduous Quercus species. Affected by climate change, Quercus species will alter leaf [P] to regulate leaf C, N, and P stoichiometry. The patterns of leaf C, N, and P stoichiometry at genus level (Quercus) and their association with climatic variables suggest that the eurytopic species will adjust nutrient strategies and potentially shift the distribution range affected by climate change.  Latitudinal (yellow) and longitudinal (green) patterns of leaf C, N, and P stoichiometry of the deciduous Quercus species.

Figure 3
Relative importance of environmental factors on leaf stoichiometric traits of the deciduous Quercus species presented by radar maps. MAT, mean annual temperature (℃); Tmin, the minimum temperature of the coldest month; Tmax, the maximum temperature of the warmest month; Twet, mean temperature of the wettest quarter; TS, temperature seasonality; MAP, mean annual precipitation (mm); Pmin, precipitation of driest month; Pcold, precipitation of coldest quarter; PET, annual potential evapotranspiration; AI, aridity index; SWC, soil water content (%). The abbreviations of leaf stoichiometric traits can be found in Figure 2. "*" "**" "***" showed P < 0.05, 0.01, < 0.001, respectively. The relationships between the temperature-related (a -f), the moisture-related (g -l) factors and leaf C, N, P stoichiometry of the deciduous Quercus species. The red, green, and black colors in a -f indicated the relationships of leaf C, N, and P stoichiometry with mean annual temperature (MAT), the maximum temperature of the warmest month (Tmax), and temperature seasonality (TS), respectively; while those in g -l indicated the relationships of leaf C, N, and P stoichiometry with mean annual precipitation (MAP), potential evapo-transpiration (PET), and aridity index (AI), respectively. The abbreviation of leaf traits can be found in Figure 2.

Supplementary Files
This is a list of supplementary les associated with this preprint. Click to download. SupplementaryInformation.docx