1. Introduction

The geographic distributions of plant productivity and species diversity provide a comprehensive picture of long-term adaptation to ecological processes shaped by environmental and geographical factors (Whittaker et al., 2001; Willis and Whittaker, 2002; Ricklefs, 2004). Accordingly, biogeographical patterns have historically been the focus of botanists and ecologists (Lennon et al., 2004; Freestone and Inouye, 2006; Kraft et al., 2011). One benefit of clearly characterizing patterns of plant productivity and species diversity is that it helps researchers understand regional biodiversity and provides a theoretical basis for biodiversity conservation. Thus, exploring the spatial patterns of plant productivity and species diversity along environmental gradients is an important prerequisite for the conservation of species diversity and the exploration of resource acquisition among different species (Whittaker et al., 2001; Reilly et al., 2006).

Previous research has shown that spatial patterns of plant productivity and species diversity vary greatly at different scales and in response to environmental gradients. Environmental factors, such as temperature, precipitation, light and soil, are influenced by changes in latitude, longitude and their combinations in space. Thus, environmental and geographic factors jointly affect the spatial pattern of plant productivity and species diversity (Austrheim, 2002). However, some researches suggest latitude and longitude do not strongly affect plant productivity or species diversity (e.g. Zhang et al., 2022). Instead, altitude is thought to be a major factor that influences plant productivity and species diversity. Specifically, patterns of species diversity are similar along altitudinal and latitudinal gradients, i.e., species diversity decreases along altitudinal gradients (Tang et al., 2004). Researches have also indicated that species diversity and productivity vary along altitudinal gradients in a "hump-shaped" patterns (e.g., "middle elevation uplift"; "hump-shaped" patterns are subdivided into the partial peak pattern and the mid-peak pattern) (Austrheim, 2002; Fu et al., 2004; Manish et al., 2017), "U-shaped" pattern (Peet, 1978), "negative monotonic" pattern (Vazquez and Givnish 1998; Bruns and Kennedy, 2009; Salas-Morales and Meave, 2012), "positive monotonic" pattern (Silva et al., 2014), "bimodal patterns" (Tian et al., 2012), and even patterns with "no significant regularity or irrelevance" (Wilson and Sydes, 1988). The most common relationship between altitude and plant productivity and species diversity include "negative monotonic" and "hump-shaped" patterns (Liu et al., 2015).

Patterns of plant productivity and species diversity along altitudinal gradients also differ across regions. For example, in the tropics diversity and productivity decrease monotonically along altitudinal gradients, in subtropical regions they are hump-shaped, and in temperate regions they show no significant regularity (Ohsawa, 1991, 1995; Fang, 2004; Liu et al., 2007). These results show that patterns of plant productivity and species diversity are different across regions or mountains, spatial scales, habitat types, and geographical locations. However, few studies have examined how plant productivity and species diversity of distinct community types vary along an altitudinal gradient.

Plant productivity and species diversity are also affected by environmental factors, including temperature, humidity, soil nutrients, and solar radiation (Yang et al., 2004). At low altitudes, key environmental factors include human disturbance, soil nutrients, and precipitation (Körner et al., 2013; Feng et al., 2016; Liu et al., 2018; Oliveira et al., 2021), whereas at high altitudes temperature (Whittaker et al., 1975; Campbell et al., 1995) and solar radiation (Dorji et al., 2014) are important. Several environmental factors, including the combined effects of hydrothermal conditions (Beck and Chey, 2008), the scale of the study area (Liu et al., 2015), topographical changes (Lomolino, 2001; Willis and Whittaker, 2002; Rahbek, 2005) and evolutionary history (Wu et al., 2014), have been shown to alter plant community structure and species diversity at different altitudes.

Plant growth is also impacted by various environmental factors, including changes in the barometric pressure (BP), and the quality (e.g., increases or decreases in air pollutants) and quantity (such as changes in concentrations of CO₂ and O₂) of atmospheric components (Liu et al., 2000; Guo et al., 2008). Plant productivity increases in low-oxygen conditions, as molecular diffusive rates increase and photosynthetic and transpiration rates increase (Andre and Massimino, 1992). Several factors have also been reported affect plant growth at low pressure, including the partial pressure of oxygen (pO₂) being lower than the total pressure. Low pO₂ breaks the reaction balance of the photosynthesis equation and reduces the dark respiration rate, which increases photosynthetic rate and ultimately promotes plant growth. In addition, other factors, such as the partial pressure of carbon dioxide (pCO₂), which affects the photosynthetic and transpiration rates of plants (Daunicht and Brinkjans, 1992; Corey et al., 1997; He et al., 2003). Overall, most previous studies on plant productivity and species diversity have only considered common environmental factors, such as climate, water, temperature and disturbance, but have ignored the effects of BP, pO₂ and pCO₂ on plants.

The Qinghai-Tibet Plateau (QTP) is the largest plateau in the world (Zhang et al., 2002; Chen et al., 2022). Grasslands on the QTP account for about 12% of China's land area (Zhang et al., 2016; Sun et al., 2019), playing a pivotal role in regulating climate, preventing soil erosion, purifying the air environment, protecting biodiversity and protecting ecological security (Sun et al., 2012). In addition, as the highest geographical unit in the world, with an average elevation higher than 4000 m above sea level (Zhang et al., 2002; Chen et al., 2022), the QTP has unique climatic conditions, including strong solar radiation, long durations of sunlight exposure, low temperatures, low air pressures, low CO₂ and O₂ contents, large temperature differences between day and night, and uneven humidity and precipitation (Tibetan Plateau Scientific Expedition Team, Chinese Academy of Sciences, 1984; Qing et al., 1987; Hou et al., 1993; Liu et al., 2000; Wu and Kayser, 2006; Li et al., 2012). Therefore, the QTP is an ideal region in which to study patterns and processes of alpine grassland (Sun and Liu, 2021).

In this study, we asked how plant productivity and species diversity vary with altitude, longitude and latitude in alpine grasslands of the QTP. We then identified the environmental factors that drive these observed patterns of plant productivity and species diversity. We hypothesized that (1) species diversity and productivity vary greatly with grassland types, (2) species diversity and productivity are restricted by BP, pO₂ and pCO₂, and (3) alpine grassland plant above-ground biomass (AGB) is significantly correlated with species diversity indexes. The findings of this study offer a foundation for conservation of plant diversity and sustainable protection of the alpine grasslands of the plateau. This study also provides new ideas for the further exploration of the influence of anoxic environments on vegetation growth.

2. Materials and methods 2.1. Study area description

Our study area is located in the QTP region of southwestern China. As the largest and highest geographical plateau in the world, it is characterized by low temperatures, with its annual mean temperature (MAT) ranging from −15 ℃ to 10 ℃ (You et al., 2013). These sustained low temperatures make the QTP the largest expanse of high-elevation permafrost in the world. The mean annual precipitation (MAP) ranges from > 1000 mm to < 50 mm and conforms to a southeast–northwest gradient, with rainfall generally concentrated between June and September. Solar radiation ranges between 6000 and 9000 MJ m⁻² yr⁻¹. Additional climatic and ecological features of the QTP include low pO₂ and pCO₂ levels (Liu et al., 2000).

The harsh climatic features of the QTP have led to the development of a range of unique alpine grassland ecosystems (Zhong et al., 2010), which provide ideal locations for investigating alpine grassland responses to a changing climate over a large region (Fig. 1). In this study, we examined three different grassland types (alpine meadow, alpine steppe and alpine desert steppe) at four ecological functional zones, including the Qilian Mountain water conservation ecological functional zone (Qilian Mountain Zone), the Three-River Headwaters grassland meadow ecological functional zone (Three-River Headwaters), the Zoige grassland wetland ecological functional zone (Zoige Grasslands), the Changtang Plateau Desert ecological functional zone (Changtang Plateau Desert). The dominant plants in each ecological functional area are detailed in Table 1.

2.2. Sampling sites and methods

From July and August 2021, we conducted field surveys and sampling in four ecological functional zones: Qilian Mountain Zone, Three-River Headwaters, Zoige Grasslands, and the Changtang Plateau Desert. Field work corresponded to a period of vigorous plant growth and followed technical specifications for inventorying plant communities (Fang et al., 2009). We established two 1-km transects along different altitudinal gradients in different ecological functional areas. We then recorded environmental factors along each altitudinal gradient, including geographic location, terrain factors and disturbance levels. Three replicate subplots (1 m × 1 m each) were randomly selected at each altitudinal gradient, for which we recorded species names, species numbers, natural height, density and coverage of the plants in the subplots.

pO₂ was measured using a CY-12C digital oxygen meter. Air temperature (AT), Air humidity (AH) and BP were measured using a Casio DPH-103 barometer (Chen et al., 2022). pCO₂ was measured using a SW-723 gas detector. AGB of the plants in the subplots was collected and put into envelopes, and then classified by species. After collecting AGB, the below-ground biomass (BGB) was randomly sampled using a standard stainless-steel root auger (STEPS, DEU; sampler size: 10 cm in width, 100 cm in length) in the subplots, and the samples were then placed in nylon mesh bags at depths of 0–5 cm, 5–10 cm, 10–20 cm, 20–40 cm, 40–60 cm, 60–80 cm and 80–100 cm.

All plant samples were transported to the laboratory and dried in an oven at 65 ℃ for 48 h before recording the constant weight. Soil in nylon mesh bags was washed with running water, and the remaining roots were collected through a 1 mm screen to remove dead roots and other matter. Similarly, three soil samples from seven depths (0–5; 5–10; 10–20; 20–40; 40–60; 60–80 and 80–100 cm) were collected using a standard stainless soil collection steel tube (STEPS, DEU; sampler size: 5 cm in width, 100 cm in length) to analyze the soil organic carbon (SOC) and total nitrogen (STN) contents, pH and soil electrical conductivity (EC). All plant and soil samples were analyzed at the Analysis and Test Center, State Key Laboratory of Earth Surface Processes and Resources Ecology of Beijing Normal University (Beijing, China), according to the standard methodology (LY/T 1271–1999 Forest Soil Analysis Method).

2.3. Soil nutrient assay and species diversity calculation

The SOC and STN contents (g·kg⁻¹ dry soil) were analyzed using an elemental analyzer (CN802, VELP, Italy). The pH and EC of the soil were measured by a pH (pHs-3C) meter and a DDSJ-308 conductivity meter, respectively. The soil-carbon-to-nitrogen ratio (C: N) is the ratio of SOC and STN. Species diversity was estimated using the four standard multi-dimensional biodiversity indexes (i.e., the Margalef index (R), the Pielou index (J), the Shannon–Weiner index (H') and the Simpson index (C)), and it was calculated based on the equations below as previously reported (Fang et al., 2004; Fayiah et al., 2019).

The Shannon–Wiener index (H') is mainly used to measure the level of community species diversity, and the Pielou index is used to measure the evenness of species distribution in the community. The formulae are as follows:

Shannon - Wiener\; index: {H^\prime } = - \sum\limits_{i = 1}^S {{{\rm{p}}_{\rm{i}}}} \ln {p_i}

(1)

Pielou \;index: J = \frac{{ - \sum\limits_{{\rm{i}} = 1}^S {{p_i}} \ln {p_i}}}{{\ln S}}

(2)

In formulae 1 and 2, S is the number of species in the community, and p_i is the ratio of the number of individuals in the species i to the total number of individuals in all species in the community. The Shannon index varies from 0 for communities with only a single taxon to high values lnS, for communities with many taxa each with few individuals. The Pielou index, which generally takes a value between 0 and 1, can be used to indicate the uniformity of species distribution. When the uniformity is low, that is, the value is small, it indicates that the distribution of species in the system is not uneven, and when the value is large, it indicates that the distribution of species is uniform.

The Simpson index, also known as the dominance index, is a measure of concentration, the opposite of diversity. The formula of the Simpson index is as follows:

Simpson\; index : C=1-\sum\limits_{i=1}^S P_i^2

In formula 3, $ {\sum {{p_i}} ^2}$ reflects the concentration of the species, and $1 - {\sum {{p_i}} ^2} $ reflects the dispersion of the species. The Simpson index varies from 0-Cmax, with 0 indicating all individuals belong to one species, while Cmax (1–1/S) indicates each individual belongs to different species. The closer the measured value is to 1, the greater the dispersion of the species.

The Margalef index is a simple species diversity index emphasizing species richness. It attempts to correct for the increasing number of species collected with the greater number of organisms sampled by dividing the species count by the natural log of the number of sampled organisms. The formula for the Margalef index is as follows:

M \text { arg}\; alef \;index : R=\frac{(S-1)}{\ln \mathrm{N}}

(4)

In formula 4, S is the number of species in the community, and N is the total number of individuals in all species in the community. This can also be a simple measure of mean population size. Using the form (S–1) gives a zero value for just one species present in a sample. The maximum appears for S ═N.

2.4. Data processing and analysis

A one-way ANOVA in SPSS analysis (SPSS software v.22.0, Chicago, Illinois, USA) was employed to compare the mean values and test the significance of differences (α = 0.05 and 0.01). Pearson's simple correlation test function in SPSS was performed to assess the plant productivity and species diversity relationships with longitude, latitude and altitude. Redundancy analysis (RDA) was based on a covariance matrix, and was used to examine the effects of environmental factors on plant productivity and diversity indices using the package CANOCO v.5.0 (Microcomputer Power, Inc., Ithaca, NY). Statistical drawing was performed using Origin pro 2023 (Originlab Inc., Northampton, MA., USA) and CanoDraw (for Windows). Productivity was calculated as the dry weight of above-ground biomass and below-ground biomass, that is, the total dry weight of the above-ground and below-ground parts of the plants in each quadrat, and it is expressed in the units g·m⁻² as previously reported (Qiu and Du, 2004).

3. Results 3.1. Geographical distribution patterns of plant productivity 3.1.1. Plant productivity

The variation ranges of the AGB and BGB in the alpine grassland were 25.45 ± 10.29–197.19 ± 34.53 g m⁻² and 86.84 ± 29.00–289.04 ± 92.48 g m⁻², respectively (Table 2). In addition, AGB increased linearly with the increase in longitude and latitude (longitude: y = 6.545x-530.056, R² = 0.035, p < 0.01; latitude: y = 7.2002x-147.488, R² = 0.035, p < 0.05) (Fig. 2a and e). However, there was no significant trend in BGB with latitude or longitude (p > 0.05) (Fig. 2b and f). Among the grassland types, AGB decreased from alpine meadow to alpine steppe and then to alpine desert steppe, but there was no significant difference between alpine steppe and alpine desert steppe (p > 0.05). However, BGB did not change with respect to grassland type (p > 0.05) (Fig. 3a and b). In addition, AGB showed a significant linear decrease with altitude (p < 0.01), but there was no significant change in BGB (p > 0.05) (Fig. 4a).

3.1.2. Vegetation coverage and height

The variation ranges of the VC and H in the alpine grassland were 48.73 ± 10.35%–9.91 ± 6.53% and 2.55 ± 0.40–13.53 ± 8.90 cm, respectively (Table 2). H increased linearly with longitude and latitude (p < 0.01) (Fig. 2c and g). VC also increased significantly with longitude but not with latitude (p > 0.05) (Fig. 2d and h). Among the grassland types, the VC and H of alpine meadow were significantly higher than those of alpine steppe and alpine desert steppe (p < 0.05) (Fig. 3c and d). In addition, H and VC decreased linearly with the altitude gradient (p < 0.01) (Fig. 4b).

3.2. Geographical distribution pattern of species diversity

As shown as in Table 3, the variation ranges of the Shannon–Wiener, Margalef, Simpson and Pielou indexes were 0.70 ± 0.22–1.88 ± 0.08, 4.51 ± 0.48–25.50 ± 7.86, 0.32 ± 0.27–0.78 ± 0.05 and 0.27 ± 0.11–1.26 ± 0.23, respectively. The optimal linear mixed model showed that the Shannon–Wiener, Simpson and Pielou indexes increased linearly with longitude and latitude (p < 0.01) (Fig. 5(a, b, d, e, f and h)), but the Margalef index showed the opposite trend with longitude and latitude (p < 0.01) (Fig. 5(c and g)). Similarly, the Shannon–Wiener, Simpson and Pielou indexes showed an extremely significant negative correlation with altitude (Fig. 7a and b), while the Margalef index showed the opposite trend with altitude (Fig. 7b). Among the different grassland types, the Shannon–Wiener, Simpson and Pielou indexes all decreased significantly from alpine meadow to alpine steppe and then to alpine desert steppe, but the Margalef index showed the opposite trend from the alpine meadow to alpine steppe and then to the alpine desert steppe (Fig. 6).

3.3. Analysis of relative contribution rate of environmental factors to plant productivity and species diversity

Redundancy analysis (RDA) showed that, out of all environmental factors examined STN, pCO₂, AH and SOC were the main factors affecting the change in alpine grassland plant productivity, with contribution rates of 41.6%, 25.7%, 10.0% and 10.0%, respectively (p < 0.05) (Fig. 8a). Species diversity was affected most by pO₂, AH and BP, with contribution rates of 62.5%, 11.9%, and 9.6%, respectively (p < 0.05) (Fig. 8b).

Plant productivity and species diversity were highly correlated with SOC, STN, C: N, and pH (p < 0.01). In addition, the species diversity indexes were significantly correlated with pO₂ and BP (p < 0.05). In contrast, the relationships between the effect sizes of AH, AT and plant productivity and species diversity indexes were not significant (p > 0.05) (Table 4).

Statistical analysis showed that the effect sizes of AGB, VC and H and species diversity were positively correlated (p < 0.01), and the highest correlation was detected between AGB and BGB (p < 0.01). The effect sizes of Shannon–Wiener and Simpson and Pielou indexes were positively correlated (p < 0.01), but negatively correlated with the effect size of the Margalef index (p < 0.01). In contrast, the relationships between the effect sizes of VC, AGB, H, Shannon–Wiener index, Pielou index, Simpson index and BGB were not significant (p > 0.05) (Table 5).

3.4. Relationship between species diversity and above-ground biomass

Pearson analysis found that there was no significant linear relationship between the Shannon–Wiener index and the AGB, except for in Qilian Mountain Zone (p < 0.05). In addition, there was no significant correlation among the Simpson, Pielou and Margalef indexes and the AGB in the four ecological functional areas (p > 0.05; Fig. 9). However, after synthesizing all the data, we found that there was a significant correlation between AGB and species diversity indexes (p < 0.01), indicating that scale is important when studying the relationship between species diversity and AGB (Fig. 9).

4. Discussion 4.1. Species diversity and productivity vary greatly with grassland types

The spatial patterns of plant productivity and species diversity are a comprehensive reflection of the effects of various ecological factors. Our results show that the productivity and diversity of the different grassland types differ greatly. For example, AGB, H, and VC of alpine meadow were significantly larger than those of alpine steppe and alpine desert steppe, but there was no significant difference in BGB between the different grassland types (p > 0.05). The reason for this is that water has a great impact on AGB, H, and VC, but it has a relatively small impact on BGB (Yang et al., 2009). It is worth noting that plant species diversity (the Shannon–Wiener, Simpson, and Pielou indexes) also showed a similar trend, which is consistent with previous studies (Tang et al., 2015). Differences in grassland plant biomass may be due to the differential effects of climate change as well as biotic pressures such as grazing and tourism (Fayiah et al., 2019; Zhao et al., 2019). However, Wang et al. (2022) suggested these differences arise because resources in resource-rich habitats promote both plant productivity and diversity. In our study, species richness showed the opposite trend. This was likely because the large biomass of the alpine meadow community, which increased the competition among species in the community, caused some species in the community to disappear, resulting in a decrease in richness (Grime, 1979; Bonser and Reader, 1995). Likewise, we found that the changes in grassland productivity and diversity along environmental gradients also varied along the large-scale longitudinal gradient across the whole QTP. This finding is consistent with previous research that has indicated that longitude and its inherent differences in climatic properties are critical for grassland plant productivity and species diversity (Zhu et al., 2017). Similar to the variation in longitude, plant productivity and species diversity also showed large-scale latitudinal variation, indicating that the geographical distribution of plant productivity and species diversity of the alpine grasslands is controlled by both longitude and latitude. Interestingly, however, the longitudinal effects were strong. Based on this, it can be speculated that moisture is the dominant factor that alters large-scale distribution of plant productivity and diversity of the alpine grasslands in the QTP rather than temperature. RDA analysis also support this conclusion, which is similar to results of previous studies (Yang et al., 2004; Zhu et al., 2017).

4.2. Species diversity and productivity are restricted by BP and pO₂

Studies have shown that the rate at which ecological factors change is much higher along an altitudinal gradient than along latitudinal and longitudinal gradients (Yoda, 1967; Körner, 2003). Consequently, changes in plant productivity and species diversity were also more dramatic across the altitudinal gradient. Currently, there are five patterns of change in species diversity along altitudinal gradients, namely, positive correlation (monotonically increasing linearly), negative correlation (monotonically decreasing linearly), uncorrelated, unimodal curves (i.e., a single peak that is high in the middle and low on both sides, or a variation pattern of partial peaks) and moderately low altitude (U-shaped or V-shaped) (He and Chen, 1997). Unimodal curves (Tian and Xing, 2008; Duan et al., 2011; Kong and Li, 2012; Sa et al., 2012) has been extensively reported, with positive correlation (Shimono et al., 2010; Dorji et al., 2014) and negative correlation less so (Wang et al., 2006). The results of this study show that, except for BGB, which had no significant correlation with altitude, vegetation productivity and diversity indicators were highly significantly correlated with altitude (p < 0.01). However, altitude changes the distribution of community productivity and species diversity by changing other local environmental factors, such as temperature, humidity and solar radiation (Gaston, 2000), and it has no effect on plants by itself.

Local disturbance factors on the altitudinal gradient mainly include topographic factors (shady and sunny slopes), soil factors (pH, carbon and nitrogen elements, and the carbon–nitrogen ratio) and climatic factors (AT, AH, BP, pO₂ and pCO₂). In fact, it has been demonstrated that environmental factors (e.g., climatic conditions and soil properties) play important roles in determining plant productivity in the alpine grasslands on the QTP (Li et al., 2011; Sun et al., 2013; Jiang et al., 2014). For instance, the results of some studies suggest that plant biomass is positively related to MAP (Li et al., 2011; Jiang et al., 2014) and that it is generally higher in the alpine meadow than in the alpine steppe and in the alpine desert steppe (Sun et al., 2016). Our study found that the changes in plant productivity were not related to AT or AH (p > 0.05, Table 2), indicating that climate conditions may not be key factors controlling plant productivity in the alpine grasslands on the QTP. Similarly, it has also been shown that climatic conditions are key factors in determining species diversity in the alpine grasslands on the QTP (Wang et al., 2006; Miehe et al., 2008). However, our study showed that the effect sizes of plant diversity indexes were weakly correlated with AT and AH (Table 4), suggesting that the differences in climatic conditions across grassland types might have little impact on changes in species diversity. This conclusion seems to be inconsistent with the results of previous studies (Zhang et al., 2023), but what has caused these inconsistencies? RDA analysis indicated that pO₂, AH and BP had a great impact on species diversity of alpine grasslands (p < 0.05), with contribution rates of 62.5%, 11.9% and 9.6%, respectively, implying that besides climate factors (AH), BP and pO₂ also have significant effects on species diversity. The more in-depth reason is probably that, when the oxygen diffusion in the environment cannot meet the requirements set by metabolic rates, the reaction balance of the photosynthetic process of the plant is disturbed, the dark respiration rate is reduced, and the steady-state level of the plant pchlide is affected (Abbas et al., 2022). In addition, the reduction in BP also affects the activity of plant enzymes (Andre and Massimino, 1992). Similar to species diversity, plant productivity was mainly affected by STN, pCO₂, AH and SOC (p < 0.05), with contribution rates of 41.6%, 25.7%, 10.0% and 10.0%, respectively, indicating that besides soil factors (STP and SOC) and climate factors (AH), pCO₂ also has a significant impact on plant productivity. Therefore, future research on the factors affecting alpine grassland plant productivity and species diversity on the Qinghai-Tibet Plateau should focus on changes in BP, pO₂ and pCO₂.

4.3. Spatial scale affected the relationship between plant productivity and species diversity

Many studies have confirmed that the relationship between species diversity and productivity takes many forms: a unimodal relationship (Kassen et al., 2000), a positive linear relationship (Tilman et al., 1997), a negative linear relationship (Thompson et al., 2005) or a relationship that is not obvious (Grace et al., 2007). Most common among these relationships between diversity and productivity are no obvious relationship (32%), unimodal curves (30%) and positive correlations (26%) (Waide et al., 1999). For alpine grasslands, the relationship between species diversity and productivity has been found to have a unimodal curve relationship (Kassen et al., 2000; Wang et al., 2004; Anivar et al., 2006; Zhu et al., 2008), a negative correlation (Li et al., 2007) and a positive correlation (Du et al., 2003; Liu et al., 2013). Our study found that, except for the Shannon–Wiener index of Qilian Mountain Zone, which had a significant linear relationship with the above-ground biomass (p < 0.05), grassland plant diversity in other regions had no significant correlation with productivity (p > 0.05). Most studies have found similar results; that is, no obvious relationship between species diversity and productivity has been found in 32% of cases (Waide et al., 1999). However, when we integrated all the quadrat data of the mountain transects in the four ecological functional areas, we found that the above-ground biomass and the diversity index were highly significantly correlated (p < 0.01), indicating that, within a single community type, the relationship between diversity and productivity was significantly different. The relationship is often not significant (Moore and Keddy, 1988), and when the study area includes several community types, the relationship between the two often shows a significant correlation (Liu et al., 2015). This demonstrates the importance of research scale in exploring the relationship between species diversity and community productivity (Chase and Leibold, 2002; Adler et al., 2011; Fraser et al., 2015). This conclusion can also be explained by the "habitat heterogeneity hypothesis" and the "geographic region hypothesis"; that is, less environmental heterogeneity in smaller geographic units leads to no significant relationship between species diversity and productivity. However, when the spatial scale of the study object increases, the environmental heterogeneity also increases, and the larger available area is conducive to the coexistence of more species in high-altitude areas dominated by alpine grasslands, with the relationship between species diversity and productivity often being positively correlated (Zhang et al., 2011).

5. Conclusions and implications

In summary, this work found that plant productivity and species diversity showed large-scale spatial variations with longitude and latitude and that the longitudinal effects were strong. These findings indicate that moisture has a greater impact than temperature on the plant productivity and species diversity of alpine grasslands in the QTP. We also found that besides soil factors and climate factors, pCO₂ also has significant effects on plant productivity, and BP and pO₂ also have significant effects on species diversity. Therefore, future research on the factors affecting alpine grassland productivity and species diversity on the Qinghai-Tibet Plateau should focus on changes in BP, pO₂ and pCO₂. Spatial scale affected the relationship between plant productivity and species diversity. Due to the limited selection of habitat types and the sampling time span, these conclusions do not represent the complete ecosystem of the QTP. However, our study is an important step in understanding the interaction between different habitats of the alpine meadow and plant communities in long-term hypoxic (low-oxygen) regions. Future studies on the mechanisms underlying these results are necessary so as to put forward more accurate suggestions for the protection of alpine grasslands.

Acknowledgments

We thank the following individuals for help with the experiments: Xiaofei Gao, Yingying Liu, Yingna Liu and Chunling Liu (teachers in the Analysis and Test Center, State Key Laboratory of Earth Surface Processes and Resources Ecology of Beijing Normal University, who helped with indoor sample analyses and tests), and Xinyan Chen and Youjun Bai (students in the Qinghai Normal University who helped with field sample collection). We sincerely thank the subject editor and the two anonymous reviewers for their careful review of this manuscript and for their valuable comments on the earliest version of the manuscript. This study was supported by the Second Tibetan Plateau Scientific Expedition and Research Program (STEP), Grant No. 2019QZKK0606.

Author contributions

Yunlong Pan: Conceptualization, Methodology, Formal analysis, Visualization, Software, Writing – original draft. Haiping Tang: Supervision, Resources, Writing – review & editing, Data curation, Investigation, Project administration, Funding acquisition. Dong Liu: Data curation, Investigation, Validation, Writing – review & editing. Yonggui Ma: Investigation. All authors have read and agreed to the published version of the manuscript.

Data and Materials Availability Statement

The data presented in this study are available on request from the corresponding author. The data are not publicly available due to privacy.

Declaration of competing interest

The authors confirm that there are no competing interests.