Search tips
Search criteria 


Logo of scirepAboutEditorial BoardFor AuthorsScientific Reports
Sci Rep. 2017; 7: 10780.
Published online 2017 September 7. doi:  10.1038/s41598-017-11133-z
PMCID: PMC5589743

Changes in specific leaf area of dominant plants in temperate grasslands along a 2500-km transect in northern China


Specific leaf area (SLA) is a key trait with great ecological importance as it correlates with whole plant growth. We aimed to investigate how SLA varies with environmental factors at a geographical scale in temperate grasslands. We measured SLA and mass-based leaf nitrogen content (N mass) of four dominant plant genera along a 2500 km climatic gradient in northern China grassland, and correlated SLA with mean annual precipitation (MAP), mean annual temperature (MAT), soil nitrogen concentration (soil N), soil C:N and N mass. Climate accounts much more for SLA variation than soil variables for Stipa, Cleistogens and Carex. SLA of Stipa is negatively associated with MAP and soil N, while positively with MAT, but Cleistogenes and Carex show the opposite. For Leymus, soil N promotes SLA and accounts for largest fraction of SLA variation. Overall, SLA was positively correlated with N mass in semi-arid regions, but not significant in arid regions. The genus-dependent responses of SLA may have consequences on ecosystem functioning, thus may help to predict the community composition and ecosystem functions under future climate scenario. The finding of SLA-N mass trade-off and its susceptibility to precipitation will advance our understanding on plant resource use strategies.


Plant leaves play an essential role in biogeochemical cycles in ecosystems and thus changes of leaf traits strongly affect plant growth and production13.The investigation of plant leaf traits and their responses to the environmental change has gained increasingly more attention among ecologists in recent decades48. Specific leaf area (SLA), defined as the ratio of total leaf area to total leaf dry mass9, 10, has been shown one of the leaf traits best reflecting whole plant growth11, 12. SLA plays an important role in linking plant carbon (C) and water cycles because it describes the distribution of leaf biomass relative to leaf area, and thus refers to carbon gain relative to water loss, within a plant canopy10, 13.

Ecologists have shown that SLA varies with environmental factors: positively with precipitation at regional and global scales2, 14, 15, as well as with soil fertility and nutrient availability2, 1619, whereas negatively with soil C:N ratio7. Unlike the consistent positive correlation between SLA and precipitation, the relationship between SLA and air temperature seems quite controversial20, 21. In general, the responses of SLA to air temperature were species-specific22, suggesting that different species have different mechanisms to cope with temperature changes which are largely related to the plant’s structure as a whole23.

Besides being correlated with environmental variables, SLA was shown to vary closely with other leaf traits: positively with mass-based leaf nitrogen content (N mass) and negatively with leaf lifespan1, 2, 19, 24, 25. It has been well established that the positive correlation between SLA and N mass reflects the trade-off between two opposing resource strategies, i.e., conservation vs. rapid acquisition of soil water and nutrients2, 26. However, it is important to point out that these results were all obtained in areas with relatively high annual rainfall values, or the datasets used in these studies included very few cases from dry biomes comparable to those from wet or moist biomes. The relationship between SLA and N mass might be different along an extensive resource gradient (in water and/or nutrient availability). It is well established that plants growing in nutrient-rich environments generally grow leaves with high N contents and a relatively short lifespan producing large amount of nutrient-rich litter, whereas plants growing on dry soils conserve water and nutrients in long-lived and recalcitrant tissues2, 4, 7, 24, 26. Thus, along a long gradient in climate and corresponding soil characters, there might be a complex relationship between SLA and N mass indicating a trade-off between growth rate and nutrient conservation for plants.

The effects of soil and climate on SLA and other leaf traits have been evaluated at a global scale7, 8, however, the strength of these syntheses comes from their use of extraordinarily rich datasets, thus suffer from the difficulty in assuring standard data collection3. Moreover, the global studies encompassing all kinds of biomes might have overlooked more subtle patterns, which might be discovered from single-biome-focused and thus more refined studies. Therefore, large-scale empirical investigations conducted following a same protocol and focusing on a single type of biome are badly needed. After all, it is important to investigate how a “general” relationship found at global scale applies to more specific systems, to improve our understanding of the possibility to downscale such general findings. Indeed, it has been reported that relationships in the grassland biome were significantly different from the global average27.

Here, we sampled the leaves in 38 sites following the same protocol and investigated relationships between SLA and both climate factors and soil properties along a 2500 km long climatic gradient through the temperate grassland biome in northern China (Fig. 1; Table 1). Along the transect, the most dominant genera are Stipa, Cleistogenes, Leymus and Carex, with one to three species present for each genus (Table 2). We quantified the relationship of SLA of the four dominant genera with environmental variables along the transect. We addressed the following questions: (i) To what extent do mean annual precipitation (MAP), mean annual temperature (MAT), aridity index (AI), soil nitrogen content (soil N) and soil C:N account for the variation in SLA along the transect? (ii) How does SLA of the dominant genera change with MAP, MAT, AI, soil N and soil C:N? (iii) What is the relationship between SLA and N mass along the transect, and is such relationship subject to precipitation level?

Figure 1
The transect and locations of sampling sites in the temperate grasslands of Northern China. The map was created using ArcGIS 9.3 (Esri, CA,
Table 1
Basic information of the study sites along the transect.
Table 2
Main species and the distribution of all four genera mentioned in this study.


Relative importance of climate and soil factors accounting for SLA variation

Multivariate analyses declared that the four genera have very different models to predict SLA variation. For Stipa, the best model included MAP, MAT, soil N, soil C:N and the interactions of soil N with MAP and with MAT, which together explained 71.65% of SLA variation. For Cleistogenes, the best model included MAT, soil N and their interaction, together accounting for 56.64% of SLA variation. For Leymus, the best model included MAT and soil N, accounting for 7.14% and 38.11% of SLA variation respectively. For Carex, the best model included MAP, MAT, soil N, soil C:N and the interaction between soil N and MAT, which explained 85.65% of SLA variation (Table 3).

Table 3
Multiple mixed regression relationships between SLA of the four genera and environmental variables.

Bivariate relationships of SLA versus climate and soil factors

On the entire transect, the SLA vs. MAP and SLA vs. AI relationships were positive in Carex, not significant in Leymus, while negative in Stipa (Fig. 2). For Cleistogenes, the SLA vs. MAP relationship was positive while the SLA first increased and then decreased with AI. With increasing MAT, the SLA of Cleistogenes first increased and then decreased while the SLA vs. MAT relationship was negative in Carex but positive in Stipa. For Leymus, however, SLA was not significantly correlated with MAT (Fig. 2). Stipa decreased and Leymus increased in SLA with the increase of soil N, while SLA of Cleistogenes and Carex showed increasing trends with soil N (Fig. 2). With increasing soil C:N, Stipa decreased in SLA while the other three genera were not significantly correlated (Fig. 2).

Figure 2
The statistics from regressions of SLA against MAP, MAT, AI, soil N and soil C:N in the whole transect. Panels (a–e) are for Stipa; panels (f–j) are for Cleistogenes; panels (k–o) are for Leymus; panels (p–t) are for Carex ...

Relationship between SLA and aboveground biomass

The SLA and aboveground biomass were negatively correlated in Stipa (Fig. 3a), but positively correlated in Leymus (Fig. 3b). There is a trend that SLA of Cleistogenes increases with the aboveground biomass of the genera in the community, but statistically, such trend is only close to marginal significant (Fig. 3c). For Carex, probably due to the low number of the observations for the genera, the SLA is not significantly correlated at all with the aboveground biomass of the genera in the community (Fig. 3d).

Figure 3
Relationships between SLA and aboveground biomass in each genus. Panel a is for Stipa; panel b is for Leymus; panel c is for Cleistogenes; panel d is for Carex.

Correlations between SLA and Nmass

On the entire transect, SLA was positively correlated (P < 0.05) with N mass in Stipa, but such correlation was not significant in the other three genera (Fig. 4). In the arid region where MAP was below 200 mm, SLA was negatively correlated with N mass in Carex (Fig. 4d), although such correlation was not true in Stipa and Cleistogenes (Fig. 4a and b). In the semi-arid region where MAP was above 200 mm, however, SLA vs. N mass relationship was significantly positive in Stipa, Cleistogenes and Carex, but not significant in Leymus (Fig. 4c).

Figure 4
The relationship between SLA and N mass for the regions with MAP < 200 mm (open circle) and MAP > 200 mm (open triangle). Panel a is for Stipa; panel b is for Cleistogenes; panel c is for ...


Basically, our results about the relationship between SLA and plant aboveground biomass corroborates the notion that SLA is a proxy of the growth and abundance of the plants in the communities, which has been claimed in numerous published papers68, 26, 28, 29. This validates all the studies, including ours, that explicitly or implicitly used SLA as the proxy of plant growth. Our results declared that environmental variables have different impacts on SLA among different genera. This indicated that primary environmental factors determining SLA were genus-specific and just as Ordoñez et al. claimed, not all environmental variables were equally good predictors of SLA7. Between-genera discrepancy in the responses of SLA to environmental factors might be attributed to the fact that they belong to different functional groups and/or that they are different in anatomy and chemical composition6. Overall, climate variables play more important roles than soil fertility in SLA variation along this transect, probably because our study focused on a single biome, i.e., grassland, where the plant growth is primarily controlled by hydrothermal factors. In contrast, the relationship between SLA and N mass are much more consistent, which reflects a trade-off between leaf strategies of fast growth vs. leaf longevity, corresponding to rapid acquisition vs. conservation of soil water and nutrients respectively6, 7, 26.

Effects of climatic factors

Both multivariate and bivariate analysis indicated that climatic factors are more important than the investigated soil properties in determining SLA, which contrasted with Ordoñez et al.7. For instance, multivariate analysis showed that MAP and MAT account much more for SLA variation than soil nutrients in Stipa and Carex. The positive relationship between SLA and MAP in Cleistogenes, Leymus and Carex along the transect was consistent with Wright et al.2. Several other studies also found that plants from drier sites had lower SLA23, 30, which was interpreted as a mechanism whereby plants adapted to arid environment by keeping smaller but thicker leaves to decrease water loss. While plants growing in water- and/or nutrient-rich environments are able to obtain sufficient water and/or nutrients for fast growth, leaves become thinner and larger to promote their light capture in order to enhance their competitive capacities30.

SLA was negatively correlated with MAP and with AI in Stipa, which was really beyond our expectation, but some other inconsistent results had also been reported in Stipa plants before. Hu et al. found that SLA of Stipa purpurea varied little with mean growing season precipitation in alpine arid and semi-arid grasslands in Tibetan Plateau31. Tian et al. reported that SLA of Stipa krylovii was lowest under normal rainfall, and became greater whether the rainfall increased or decreased in typical steppe in Inner Mongolia32. In our study, both the leaf area and leaf dry mass of Stipa increased with increasing MAP, whereas the SLA decreased along the transect. We suspect that the unconventional responses of SLA of Stipa to precipitation might be attributed to the unique morphology or anatomy of Stipa leaves.

As perennial xerophytic herbaceous plant, Stipa adapts to drought environment by rolling up its leaves towards needle shaped, and this contributes to its dominance in the steppe communities. Moreover, the cross section of Stipa leaves completely curled up into a “V” or oval shapes to lessen the leaf area directly exposed to the air. Meanwhile, the outer surface of Stipa leaves is covered by a thick cuticle to increase the moisture diffusion resistance, leading to higher water use efficiency of leaf mesophyll cells31, 33. These mechanisms could contribute to Stipa’s dominance in the steppe communities by strong drought resistance (Table 2).

Correlations between SLA and MAT were quite inconsistent among the investigated plant genera in our study. Previous studies have drawn contrasting conclusions on the relationship between SLA and air temperature2, 6, 7. For instance, Wright et al. found that SLA was negatively correlated with MAT at global scale while Poorter et al. indicated that SLA increases with temperature2, 6. However, Ordoñez et al. found that there was no significant correlation between SLA and MAT7. Recently, De Frenne et al. did a meta-analysis and found that positive, negative and non-significant patterns in SLA vs. temperature relationship were all identified in previous studies34. Atkin et al. also claimed that the impact of air temperature on SLA was species-specific22. This suggests that different species have different mechanisms to cope with temperature changes, because distinct resource requirements across plant species can lead to different responses of species to the same environmental change35. It is possible that different responses of SLA to temperature are caused by plants’ fundamental growth characteristics and specific ecological niche36. Probably, these genus-specific responses are the pre-requisition and also consequences of an evolutionary process where different species fulfill complementary roles during community assembly (the niche complementarity hypothesis)6, 37.

Effects of soil N and soil C:N

Taken together the results of multivariate and bivariate regressions, soil N, as the most commonly used proxy of N supply, accounts much more than soil C:N in explaining SLA variation along the transect. Although Ordoñez et al. deemed soil N as a rough estimate of nutrient supply7, we showed that high plant SLA were found at high soil N region, indicating that soil N was a decent predictor of SLA variation in temperate grassland, especially for nitrophilous Stipa and Leymus 38. Since SLA reflects plants’ abilities to access resources, our results implied that plants allow fast use of nutrients and rapid growth in fertile habitats, while small SLA occurred at low nutrient supply areas where conservation of nutrients is more important. With the increase of soil N, plants allow to obtain more nutrients to increase leaf area and further increase the photosynthetic capacity. This was in line with previous studies4, 7. Although Soil C:N can provide more information about the potential to mineralize or immobilize nutrients, it turns out to be a weaker predictor for SLA variation than soil N in temperate grassland, in contrast with the global synthesis incorporating various biomes7. Probably, this should be attributed to the nice linear relationship between soil C and N content in the temperate grassland, giving rise to a relatively very narrow range of C:N values.

The SLA vs. soil factor relationship was most distinctive in Stipa in our study. The variation of this association coincides with the SLA vs. MAP relationship in Stipa. With opportunistic growth patterns39, Stipa own distinctive morphological features and traits to deal with extreme environmental stress and allow its dominance in dry areas40. As soil N was positively correlated with humidity, interspecific competition become more intense with increasing MAP and soil N, and Stipa may keep longer, heavier leaves of smaller SLA to retain its competitive advantages39.

Trade-offs between SLA and Nmass

The relationship between SLA and N mass was found to be dependent on precipitation levels. When MAP > 200mm, SLA was positively correlated with N mass, in line with previous studies from regional to global scale with MAP ranging from 133 to 5,300 mm1, 2, 19, 24, 25; while in the region of MAP < 200 mm, SLA was unrelated or negatively correlated with N mass. It has been claimed that the correlation between SLA and N mass reflects the trade-off between two opposing resource strategies, i.e., conservation vs. rapid acquisition of soil water and nutrients2, 26, 41. In those regions that receive plentiful rainfall, plants obtain sufficient nutrients for fast growth in order to enhance their competitive capacities, implying that high SLA and N mass reflect the plant strategy of rapid resource acquisition. However, in the regions where MAP < 200 mm, both nutrients and water are in shortage, plants have smaller SLA to reduce the transpiration6, while N mass were increased to improve cellular osmotic pressure and strengthen water protection in plants, in order to resist drought stress42, 43. Therefore, low SLA and high N mass reflect the resisting drought stress strategy of plants. Tian et al. investigated the responses of leaf traits of 14 plant species in the typical steppe to simulated rainfall, but failed to find simple linear relationship between SLA and rainfall, instead they found that N mass was affected by simulated rainfall32. This result might explain why SLA vs. N mass relationship in our case was contrasting among genera and was subject to MAP level.


Summarily, climatic factors account much more for SLA variation than soil variables for Stipa, Cleistogens and Carex. SLA of Stipa is negatively associated with MAP, AI and soil N, while positively with MAT, and Cleistogens and Carex show the opposite. For Leymus, SLA variation can be explained primarily by soil N, with a positive correlation. Such genus-dependent responses of SLA may have implications in terms of plant coexistence under the future climate scenario, as different species fulfill complementary roles during community assembly. In addition, since SLA is a key determinant of plant growth and playing an important role in plant community assembly2, 6, it is bound to have consequences on ecosystem functioning. For example, high SLA plants are preferentially attacked by herbivores, and also decompose much faster, facilitating carbon and nutrient cycling, which in turn enhances the primary production6. Therefore, our results may help to predict the changes of community productivity, species distribution range, and other ecosystem functions under the context of climate change. Based on our results, we predict that with the aridifying and warming climate, the productivity of the four dominant genera will decrease and their distribution range will shift northeastward in China. Moreover, as SLA is a key functional trait reflecting the trade-off between resource capture and conservation2, 6, our findings on the SLA-N mass relationship and its susceptibility to precipitation are bound to advance our understanding on plant adaptive strategies.

Materials and Methods

Study area

The investigation was conducted along an essentially west-east transect of approximately 2500 km long in temperate grassland in Northern China, from Alxa Left Banner, Inner Mongolia in the west to Changling County, Jilin Province in the east, ranging from 40.7°N to 50.1°N in latitude and 104.5°E to 123.5°E in longitude (Fig. 1; Table 1). The range of elevation was between 142 m to 1613 m a.s.l. The range of soil water content was from 2% to 28% and nutrient availability gradually increased from west to east along the transect44. The vegetation types along the transect included desert steppe, typical steppe and meadow steppe from west to east according to Vegetation of China 45. Three climatic variables AI (calculated as the ratio of precipitation to potential evapotranspiration), MAT and MAP were examined in this study, to analyze the effects of climate on the changes in SLA. MAP ranges from 93 to 436 mm and MAT ranges from −1.85 to 6.93 °C along the climatic gradient. The data of MAT and MAP were extracted from the WorldClimdatabase46 from 1950 to 2000.


The four most widely distributed dominant genera in the temperate grasslands of northern China and adjacent Mongolia, Stipa, Cleistogenes, Leymus and Carex were selected as the plants of interest (Table 2). Sampling was carried out during the first half of August in 2012, a period during which the phenological changes in these plants were relatively negligible and plants were fully mature throughout the transect. We selected 38 sample sites along the transect (Fig. 1; Table 1). All the sampling sites were selected inside the enclosures, which were very common in China grassland region due to grazing ban policy. So our study sites had been subject to minimal grazing and other anthropogenic disturbances. Each site was geo-referenced with GPS (eTrex Venture, Garmin, ±3 m accuracy) for its latitude, longitude and elevation. At each site, five quadrats (1 m × 1 m each) were selected for both vegetation survey and soil sampling. In each quadrat, after removing floor litter, 20 random soil samples (0–10 cm) were collected using a soil core (2.5 cm diameter), and were mixed together into one sample. Soil samples were air-dried right after sampling and then stored in a plastic bag for measurements. Plant samples were collected randomly next to the five quadrats at each site. Where present, five to ten adult and healthy individuals without obvious symptoms of pathology or herbivore attack were randomly sampled for each of the four genera11. We collected the whole aboveground parts and put them in sealed plastic bags, storing them in a refrigerator at 4 °C until further processing in the laboratory later the same day.


In the laboratory at least 20 young but fully expanded and hardened leaves were taken from the 5 to 10 randomly-selected individual plants for each of the four genera at each site. Each leaf was cut off from the culm and gently rubbed dry before measurement11. We used a scanner (HP Scanjet G3110) to scan each leaf image; then leaf area of each leaf was calculated using software Image J47. For very small or very narrow leaves like Stipa, transparent scotch tape was used to fix them onto blank paper before scanning. Each leaf sample was then dried at 80 °C for 48 h and weighed.

Soil samples were passed through a 2 mm mesh sieve to remove fine roots and plant debris in laboratory, then the samples were homogenized by hand mixing and oven-dried at 65 °C to constant weight. Both plant and soil samples were ground in a ball mill (NM200, Retsch, Haan, Germany) and then stored in a plastic bag until further analysis. Leaf and soil total N concentration (expressed here per unit of dry mass, mg·g−1) and C concentration (per unit of dry mass, mg · g−1) were determined for each sample with an elemental analyzer (Vario EL III, Germany). Since all the plant and soil samples were collected by a single team of researchers using uniform sampling method within a period of 15 days under constant weather conditions, our analysis avoids the difficulty of heterogeneous data as encountered in previous large-scale surveys48 or synthesis studies7, 8.

Statistical analysis

Levene’s test was used to test for normality of all data before statistical analysis. Bivariate analysis and multivariate analysis were performed to relate SLA and environmental variables (MAP, MAT, AI, soil C:N, soil N) separately for each genera. When analyzing the SLA vs. N mass tradeoff, we divided the total sites pool into two parts: semi-arid areas (sites in the area with >200 mm MAP) and arid areas (sites in the area with <200 mm MAP) according to Climate classification in China 49. Relationships between SLA and N mass in the two different MAP parts were analyzed using linear regression. All the above statistical analyses except multivariate analysis were performed with SPSS 18.0 (SPSS Inc., Chicago, IL, USA) for windows. Multivariate analysis of the effects of soils and climate on SLA was performed using the R software package (version 3.2.0) and the Akaike information criterion (AIC) was selected as a goodness of fit measure: the lower the AIC value, the better the model.


This study was financially supported by the National Key Basic Research Program (2015CB150802) and NSFC grant (41371077). We thank all the members of sampling campaign teams from Institute of Applied Ecology, Chinese Academy of Sciences for their assistance with field data collection. We gratefully acknowledge E. Bai for providing meteorological data and soil C and N data. We appreciate two anonymous reviewers for their helpful comments and suggestions on the earlier version of our manuscript.

Author Contributions

Author Contributions

X.H., Z.W., X.L. and X.W. conceived and designed the experiments. M.L. and S.L. performed the experiments. M.L analyzed the data, drew the figures. M.L. and Z.W. wrote the manuscript. X.L. and Z.W. revised the manuscript. All authors reviewed the manuscript.


Competing Interests

The authors declare that they have no competing interests.


Publisher's note: Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.


1. Reich PB, Walters MB, Ellsworth DS. Leaf life-span in relation to leaf, plant, and stand characteristics among diverse ecosystems. Ecol Monogr. 1992;62:365–392. doi: 10.2307/2937116. [Cross Ref]
2. Wright IJ, et al. The worldwide leaf economics spectrum. Nature. 2004;428:821–827. doi: 10.1038/nature02403. [PubMed] [Cross Ref]
3. He JS, et al. Stoichiometry and large-scale patterns of leaf carbon and nitrogen in the grassland biomes of China. Oecologia. 2006;149:115–122. doi: 10.1007/s00442-006-0425-0. [PubMed] [Cross Ref]
4. Westoby M, Falster DS, Moles AT, Vesk PA, Wright IJ. Plant ecological strategies: Some leading dimensions of variation between species. Annu Rev Ecol Syst. 2002;33:125–159. doi: 10.1146/annurev.ecolsys.33.010802.150452. [Cross Ref]
5. Wright IJ, Westoby M, Reich PB. Convergence towards higher leaf mass per area in dry and nutrient-poor habitats has different consequences for leaf life span. J Ecol. 2002;90:534–543. doi: 10.1046/j.1365-2745.2002.00689.x. [Cross Ref]
6. Poorter H, Niinemets Ü, Poorter L, Wright IJ, Villar R. Causes and consequences of variation in leaf mass per area (LMA): A meta-analysis. New Phytol. 2009;182:565–588. doi: 10.1111/j.1469-8137.2009.02830.x. [PubMed] [Cross Ref]
7. Ordoñez JC, et al. A global study of relationships between leaf traits, climate and soil measures of nutrient fertility. Global Ecol Biogeogr. 2009;18:137–149. doi: 10.1111/j.1466-8238.2008.00441.x. [Cross Ref]
8. Maire V, et al. Global effects of soil and climate on leaf photosynthetic traits and rates. Global Ecol Biogeogr. 2015;24:706–717. doi: 10.1111/geb.12296. [Cross Ref]
9. Evans, G. C. The quantitative analysis of plant growth (Blackwell Scientific, Oxford, 1972).
10. Gunn S, Farrar JF, Collis BE, Nason M. Specific leaf area in barley: individual leaves versus whole plants. New Phytol. 1999;143:45–51. doi: 10.1046/j.1469-8137.1999.00434.x. [Cross Ref]
11. Cornelissen JHC, et al. A handbook of protocols for standardised and easy measurement of plant functional traits worldwide. Aust J Bot. 2003;51:335–380. doi: 10.1071/BT02124. [Cross Ref]
12. Cheng J, Chu P, Chen D, Bai Y, Niu S. Functional correlations between specific leaf area and specific root length along a regional environmental gradient in Inner Mongolia grasslands. Funct Ecol. 2016;30:985–997. doi: 10.1111/1365-2435.12569. [Cross Ref]
13. Pierce LL, Running SW, Walker J. Regional-scale relationships of leaf-area index to specific leaf-area and leaf nitrogen-content. Ecol Appl. 1994;4:313–321. doi: 10.2307/1941936. [Cross Ref]
14. Aerts, R. & Chapin, F. S. III. The mineral nutrition of wild plants revisited: A reevaluation of processes and patterns. Adv Ecol Res30, 1–67 (2000).
15. Chown SL, Gaston KJ, Robinson D. Macrophysiology: large-scale patterns in physiological traits and their ecological implications. Funct Ecol. 2004;18:159–167. doi: 10.1111/j.0269-8463.2004.00825.x. [Cross Ref]
16. Chapin, F. S. III. The mineral nutrition of wild plants. Annu Rev Ecol Evol Syst11, 233–260 (1980).
17. Wright IJ, Reich PB, Westoby M. Strategy shifts in leaf physiology, structure and nutrient content between species of high- and low-rainfall and high-and low-nutrient habitats. Funct Ecol. 2001;15:423–434. doi: 10.1046/j.0269-8463.2001.00542.x. [Cross Ref]
18. Reich PB, Oleksyn J. Global patterns of plant leaf N and P in relation to temperature and latitude. Proc Natl Acad Sci USA. 2004;101:11001–11006. doi: 10.1073/pnas.0403588101. [PubMed] [Cross Ref]
19. Wright IJ, et al. Assessing the generality of global leaf trait relationships. New Phytol. 2005;166:485–496. doi: 10.1111/j.1469-8137.2005.01349.x. [PubMed] [Cross Ref]
20. Hovenden MJ. The influence of temperature and genotype on the growth and stomatal morphology of southern beech, Nothofagus cunninghamii (Nothofagaceae) Aust J Bot. 2001;49:427–434. doi: 10.1071/BT01001. [Cross Ref]
21. Loveys BR, Scheurwater I, Pons TL, Fitter AH, Atkin OK. Growth temperature influences the underlying components of relative growth rate: an investigation using inherently fast- and slow-growing plant species. Plant Cell Environ. 2002;25:975–987. doi: 10.1046/j.1365-3040.2002.00879.x. [Cross Ref]
22. Atkin OK, Loveys BR, Atkinson LJ, Pons TL. Phenotypic plasticity and growth temperature: understanding interspecific variability. J Exp Bot. 2006;57:267–281. doi: 10.1093/jxb/erj029. [PubMed] [Cross Ref]
23. Castro-Díez P, Villar-Salvador P, PérezRontomé C, Maestro-Martínez M, Montserrat-Martí G. Leaf morphology and leaf chemical composition in three Quercus (Fagaceae) species along a rainfall gradient in NE Spain. Trees–Struct Funct. 1997;11:127–134.
24. Reich PB, Walters MB, Ellsworth DS. From tropics to tundra: Global convergence in plant functioning. Proc Natl Acad Sci USA. 1997;94:13730–13734. doi: 10.1073/pnas.94.25.13730. [PubMed] [Cross Ref]
25. Reich PB. Generality of leaf trait relationships: A test across six biomes. Ecology. 1999;80:1955–1969. doi: 10.1890/0012-9658(1999)080[1955:GOLTRA]2.0.CO;2. [Cross Ref]
26. Ordoñez JC, et al. Leaf habit and woodiness regulate different leaf economy traits at a given nutrient supply. Ecology. 2010;91:3218–3228. doi: 10.1890/09-1509.1. [PubMed] [Cross Ref]
27. Feng XH, Dietze M. Scale dependence in the effects of leaf ecophysiological traits on photosynthesis: Bayesian parameterization of photosynthesis models. New Phytol. 2013;200:1132–1144. doi: 10.1111/nph.12454. [PubMed] [Cross Ref]
28. Garnier E, et al. Plant functional markers capture ecosystem properties during secondary succession. Ecology. 2004;85:2630–2637. doi: 10.1890/03-0799. [Cross Ref]
29. Villar R, et al. Variation in relative growth rate of 20 Aegilops species (Poaceae) in the field: The importance of net assimilation rate or specific leaf area depends on the time scale. Plant Soil. 2005;272:11–27. doi: 10.1007/s11104-004-3846-8. [Cross Ref]
30. Gong XY, et al. Tradeoffs between nitrogen- and water-use efficiency in dominant species of the semiarid steppe of Inner Mongolia. Plant Soil. 2011;340:227–238. doi: 10.1007/s11104-010-0525-9. [Cross Ref]
31. Hu MY, Zhang L, Luo TX, Shen W. Variations in leaf functional traits of Stipa purpurea along a rainfall gradient in Xizang, China. Chin J Plant Ecol. 2012;36:136–143. doi: 10.3724/SP.J.1258.2012.00136. [Cross Ref]
32. Tian Q, Wang G, Cao Z. The response of leaf traits of 14 plant species in typical steppe to the simulated rainfall. J Gansu Agric Univ. 2008;43:129–134.
33. Kang S, Niu J, Zhang Q, Chen L. Anatomical structure of Stipa breviflora leaves and its relationship with environmental factors. Acta Pratac Sin. 2013;22:77–86.
34. De Frenne P, et al. Latitudinal gradients as natural laboratories to infer species’ responses to temperature. J Ecol. 2013;101:784–795. doi: 10.1111/1365-2745.12074. [Cross Ref]
35. Kichenin E, Wardle DA, Peltzer DA, Morse CW, Freschet GT. Contrasting effects of plant inter- and intraspecific variation on community-level trait measures along an environmental gradient. Funct Ecol. 2013;27:1254–1261. doi: 10.1111/1365-2435.12116. [Cross Ref]
36. Donovan LA, Maherali H, Caruso CM, Huber H, de Kroon H. The evolution of the worldwide leaf economics spectrum. Trends Ecol Evol. 2011;26:88–95. doi: 10.1016/j.tree.2010.11.011. [PubMed] [Cross Ref]
37. Díaz S, et al. Incorporating plant functional diversity effects in ecosystem service assessments. Proc Natl Acad Sci USA. 2007;104:20684–20689. doi: 10.1073/pnas.0704716104. [PubMed] [Cross Ref]
38. Wang R, et al. Nitrogen acquisition strategies used by Leymus chinensis and Stipa grandis in temperate steppes. Biol Fertil Soils. 2016;52:951–961. doi: 10.1007/s00374-016-1128-2. [Cross Ref]
39. Pugnaire FI, Haase P. Comparative physiology and growth of two perennial tussock grass species in a semi-arid environment. Ann Bot. 1996;77:81–86. doi: 10.1006/anbo.1996.0010. [Cross Ref]
40. Balaguer L, et al. Ecophysiological significance of chlorophyll loss and reduced photochemical efficiency under extreme aridity in Stipa tenacissima L. Plant Soil. 2002;240:343–352. doi: 10.1023/A:1015745118689. [Cross Ref]
41. Freschet GT, Cornelissen JHC, van Logtestijn RSP, Aerts R. Evidence of the ‘plant economics spectrum’ in a subarctic flora. J Ecol. 2010;98:362–373. doi: 10.1111/j.1365-2745.2009.01615.x. [Cross Ref]
42. Osmond CB, et al. Stress physiology and the distribution of plants: The survival of plants in any ecosystem depends on their physiological reactions to various stresses of the environment. BioScience. 1987;1:39–48.
43. Wei H, Wu B, Yang W, Luo T. Low rainfall-induced shift in leaf trait relationship within species along a semi-arid sandy land transect in northern China. Plant Biol. 2011;13:85–92. doi: 10.1111/j.1438-8677.2010.00321.x. [PubMed] [Cross Ref]
44. Wang C, et al. Aridity threshold in controlling ecosystem nitrogen cycling in arid and semi-arid grasslands. Nat Commun. 2014;5:4799. doi: 10.1038/ncomms5799. [PubMed] [Cross Ref]
45. Wu, Z. Y. Vegetation of China (Science Press, Beijing, 1980).
46. Hijmans RJ, Cameron SE, Parra JL, Jones PG, Jarvis A. Very high resolution interpolated climate surfaces for global land areas. Int J Climatol. 2005;25:1965–1978. doi: 10.1002/joc.1276. [Cross Ref]
47. Abràmoff MD, Magalhães PJ, Ram SJ. Image processing with ImageJ. Biophotonics Int. 2004;11:36–42.
48. Luo WT, et al. Patterns of plant biomass allocation in temperate grasslands across a 2500-km transect in northern China. PLoS One. 2013;8:e71749. doi: 10.1371/journal.pone.0071749. [PMC free article] [PubMed] [Cross Ref]
49. Zhu, K. Z. Climate classification in China. In Anthology of Zhu Kezhen (ed. Zhu, K. Z.) 124–132 (Science Press, Beijing, 1979).

Articles from Scientific Reports are provided here courtesy of Nature Publishing Group