EN中文

Journal of Resources and Ecology >

2022 , Vol. 13 >Issue 1: 129 - 141

DOI: https://doi.org/10.5814/j.issn.1674-764x.2022.01.015

Industry Ecology and Regional Development

Alpine Grassland Aboveground Biomass and Theoretical Livestock Carrying Capacity on the Tibetan Plateau

  • ZHANG Xianzhou , 1, 2 ,
  • LI Meng 3 ,
  • WU Jianshuang 4 ,
  • HE Yongtao 1, 2 ,
  • NIU Ben , 1, *
Expand
  • 1. Lhasa Plateau Ecosystem Research Station, Key Laboratory of Ecosystem Network Observation and Modeling, Institute of Geographic Sciences and Natural Resources Research, Chinese Academy of Sciences, Beijing 100101, China
  • 2. College of Resources and Environment, University of Chinese Academy of Sciences, Beijing 100190, China
  • 3. School of Geographic Sciences, Nantong University, Nantong, Jiangsu 226007, China
  • 4. Institute of Environment and Sustainable Development in Agriculture, Chinese Academy of Agricultural Sciences, Beijing 100081, China
* NIU Ben, E-mail:

ZHANG Xianzhou, E-mail:

Received date: 2021-09-13

  Accepted date: 2021-10-13

  Online published: 2022-01-08

Supported by

The Second Tibetan Plateau Scientific Expedition and Research Program (STEP)(2019QZKK1002)

The National Natural Sciences Foundation of China(41807331)

The West Light Foundation of the Chinese Academy of Sciences(2018)

Fold

Abstract

The accurate simulation and prediction of grassland aboveground biomass (AGB) and theoretical livestock carrying capacity are key steps for maintaining ecosystem balance and sustainable grassland management. The AGB in fenced grassland is not affected by grazing and its variability is only driven by climate change, which can be regarded as the grassland potential AGB (AGBp). In this study, we compiled the data for 345 AGB field observations in fenced grasslands and their corresponding climate data, soil data, and topographical data on the Qinghai-Tibetan Plateau (TP). We further simulated and predicted grassland AGBp and theoretical livestock carrying capacity under the climate conditions of the past (2000-2018) and future two decades (2021-2040) based on a random forest (RF) algorithm. The results showed that simulated AGBp matched well with observed values in the field (R2 = 0.76, P < 0.001) in the past two decades. The average grassland AGBp on the Tibetan Plateau was 102.4 g m-2, and the inter-annual changes in AGBp during this period showed a non-significant increasing trend. AGBp fluctuation was positively correlated with growing season precipitation (R2 = 0.57, P < 0.001), and negatively correlated with the growing season diurnal temperature range (R2 = 0.51, P < 0.001). The average theoretical livestock carrying capacity was 0.94 standardized sheep units (SSU) ha-1 on the TP, in which about 54.1% of the areas showed an increasing trend during the past two decades. Compared with the past two decades, the theoretical livestock carrying capacity showed a decreasing trend in the future, which was mainly distributed in the central and northern TP. This study suggested that targeted planning and management should be carried out to alleviate the forage-livestock contradiction in grazing systems on the Tibetan Plateau.

Key words： alpine grassland; aboveground biomass; carrying capacity; climate change; random forest; Tibetan Plateau

Cite this article

ZHANG Xianzhou , LI Meng , WU Jianshuang , HE Yongtao , NIU Ben . Alpine Grassland Aboveground Biomass and Theoretical Livestock Carrying Capacity on the Tibetan Plateau[J]. Journal of Resources and Ecology, 2022 , 13(1) : 129 -141 . DOI: 10.5814/j.issn.1674-764x.2022.01.015

1 Introduction

Grassland accounts for approximately 40% of the terrestrial ecosystems in the world and plays important roles in biogeochemical cycles, energy flow and climate change (Scurlock and Hall, 1998; Piao et al., 2004). Aboveground biomass (AGB) is an important component of the ecosystem carbon cycle and is often used to monitor and evaluate changes in ecosystem structure and function (Tan et al., 2010; Yang et al., 2010). Grassland aboveground biomass is the main food source for grazing livestock (such as sheep and cattle), which is the material basis for the development of grassland animal husbandry and an important index for estimating grassland carrying capacity. Accurate simulation and prediction of the spatial and temporal patterns of grassland aboveground biomass can help to protect grassland ecosystems, rationally plan livestock production, and improve the sustainable management of grassland ecosystems (Li et al., 2016; Cao et al., 2019).
The Tibetan Plateau (TP) is mostly composed of grassland ecosystems, with 1.29×106 km2 of alpine grassland accounting for 65% of the total area of the TP (Long et al., 1999). The structure and function of alpine grassland not only determines the development of local animal husbandry and ecological environmental protection (Yao et al., 2012; Yang et al., 2017), but also affects the sustainable development of the social economy in the middle and lower reaches of the Yangtze River. In recent decades, the climate changes are dramatically becoming warmer and more humid (Karhu et al., 2010; Chen et al., 2013), dominating the greening of vegetation on the TP (Zhang et al., 2015b). However, human activities associated with the economic growth and social development, especially grazing activities, are increasingly disturbing the ecological environment of alpine grasslands on the TP. It was predicted that there were approximately 4.1×107 sheep and 1.3×107 yaks on the TP at the end of the 20th century (Long et al., 1999; Long et al., 2009). At the same time, the alpine grassland also supports large numbers of wild animals (over 200 species), thus the grassland is likely overloaded. Under the dual influence of climate change and overgrazing, grassland degradation on the TP is becoming increasingly prominent (Long et al., 2009; Dong et al., 2010; Chen et al., 2014; Wu et al., 2017). By the 1990s, the degraded alpine grassland area had reached 0.4×108 ha, accounting for approximately 33% of the naturally available grassland area on the TP (Long et al., 2009). Grassland degradation will cause many ecological, economic, environmental and social problems (Harris, 2010; Shang et al., 2014; Zhang et al., 2015c). Therefore, accurate estimation of the aboveground biomass and theoretical carrying capacity of the alpine grassland is of great significance for the formulation of restoration policies to control grassland degradation on the TP.
The establishment of an aboveground biomass assessment model for alpine grassland is the premise and basis of assessing grassland carrying capacity. Many studies have tried to estimate the aboveground biomass of the alpine grassland, but their results and conclusions are quite different for the TP due to the differences and limitations of data collection standards and estimation methods (Table 1). The traditional method is to obtain aboveground biomass at a regional scale based on statistical data and area weighting (Luo et al., 2002). However, this method can only estimate average biomass at the regional scale. Alternatively, ecosystem process models can explain the mechanism of carbon formation, and have also been applied in estimating the aboveground biomass of grassland. However, the process models have complex structures and many parameters, and most of the parameters require field verification (Niu et al., 2020; Xu et al., 2021). The uncertainty of these process model parameters (such as the aboveground and underground distribution ratio of productivity) directly affects the accuracy of the aboveground biomass simulation (Niu et al., 2019). With the development and application of remote sensing technology, many studies have established regression models between vegetation indices (such as NDVI and EVI) and use observed biomass values to estimate grassland aboveground biomass, thus the simulation accuracy often depends on a single vegetation index (Niu et al., 2016; Niu et al., 2017; Wu and Fu, 2018). Random Forest (RF) is a machine learning algorithm that can integrate multi-factor data and effectively deal with complex nonlinear relationships among the data (Breiman, 2001). RF models are also commonly used to simulate vegetation biomass, and show higher simulation accuracy than traditional estimation methods (Xia et al., 2018). Based on vegetation index, climate, soil and topographic data, the RF algorithm can be used to simulate the aboveground biomass of alpine grassland on the TP (Table 1). However, these simulation results mostly represent the remaining aboveground biomass of grassland after grazing by cattle and sheep, and they cannot accurately evaluate the theoretical carrying capacity of the alpine grassland.
Table 1 Estimated mean aboveground biomass (AGB) of alpine grassland on the Tibetan Plateau in various published studies
Time period Study area (104 km2) Methods Variables considered AGB (g m-2) References
1960-2002 147.74 Century Climate and soil data 70.00 Zhang et al., 2007
2002-2004 139.00 Orchidee Climate, soil and LAI data 119.78 Tan et al., 2010
1980-1990 113.60 Area-weighted average - 58.11 Ni, 2004
- 101.10 Area-weighted average - 61.15 Luo et al., 1998
2001-2004 - Filed observations - 59.30 Yang et al., 2010
2001-2004 112.80 Linear regression EVI 68.80 Yang et al., 2009
1982-2006 129.50 Exponential regression NDVI 74.11 Ma et al., 2010
2005 122.80 Exponential regression NDVI 43.33 Xu et al., 2017
- 124.00 Exponential regression NDVI 78.02 Piao et al., 2004
1982-2013 154.48 Exponential regression NDVI 104.40 Jiao et al., 2017
2000-2014 151.11 Random forest Climate, terrain and NDVI 77.12 Zeng et al., 2019
- 132.00 Random forest Climate and NDVI 78.40 Xia et al., 2018
2000-2017 - Random forest Climate, terrain and NDVI 59.63 Gao et al., 2020

Note: The conversion coefficient between plant biomass (AGB, unit: g m-2) and carbon (unit: g C) is 0.45 as reported by Piao et al. (2004).

Fig. 1 Spatial distribution of eco-geographical regions and the sample sites on the Tibetan Plateau

Note: (a) The eco-geographical regions of TP (Zheng, 1996) are the same as those listed in Table 2. (b) Sample sites in this study are from field observations and a previous study (Fu et al., 2017), which basically included all alpine grassland ecosystems on the TP.

In this study, the aboveground biomass data for 345 field observations (230 observations from 19 sites per year in each county of the north Tibet from 2009 to 2017 and 115 literature data, see Methods) from inside the fences on the TP were collected (Table 1 and Fig. 1). The grassland biomass in the fenced areas is not affected by cattle and sheep grazing and is only driven by climate and soil factors, so it can be regarded as the potential aboveground biomass of the grassland (AGBp). Based on climate, terrain and soil data, a grassland AGBp estimation model was constructed using the RF algorithm to simulate the grassland AGBp under the current climate conditions (2000-2018), and the alpine grassland AGBp in the next 20 years (2021-2040) was predicted by combining the four latest future climate scenarios. Our aims are: 1) To accurately simulate and predict the aboveground biomass, 2) To discuss the temporal and spatial dynamics of the theoretical livestock carrying capacity of alpine grassland on the TP.

2 Materials and methods

2.1 Study area

The TP covers an area of 2.57×106 km2, accounting for 26.8% of China’s total land area from the southern foothills of the Himalayas to the northern side of the Kunlun and Qilian mountains, and from the Pamir Plateau in the west to the Hengduan mountains in the east (Fig. 1a) (Zhang et al., 2002). The latitude and longitude ranges are 26°00'12"- 39°46'50"N and 73°18'52"-104°46'59"E, respectively. With an average elevation of more than 4000 m, the TP is known as “the roof of the world” and the “third pole”. In addition, TP is also known as the “Water tower of Asia” because it is the birthplace of many important Chinese rivers, such as the Yellow River, Yangtze River, Lancang River, Nu River and Yarlung Zangbo River. As the ecological security barrier in China, the TP plays an important role in maintaining biodiversity, climate and water cycles (Yao et al., 2012). Grassland is one of the most important ecosystem types on the Tibetan Plateau. Under the comprehensive effect of temperature and rainfall, alpine meadow, alpine steppe and alpine desert steppe are widely developed on the TP from east to west (Fig. 1b). The available pastures on the TP cover approximately 1.29×106 km2, accounting for 32.5% of the total grassland area in China, and making it one of the most important pastures in China (Long et al., 1999).

2.2 Data

In this study, the aboveground biomass observed in the fenced grassland areas mainly came from two datasets that basically included all of the alpine grassland ecosystems on the TP (Fig. 1). One dataset is the field survey data for the grassland transection of the Northern Tibet Plateau from 2009 to 2017, covering three zonal alpine grassland vegetation communities (alpine meadow, alpine steppe and alpine desert steppe) with a total of 230 sampling data points (Fig. 1b). The aboveground biomass in this dataset was harvested when most local plants were at the flowering or fruiting stages (late July to early August). The data for each sample is the mean AGB of all species from five 0.5 m×0.5 m quadrats arranged in 20 m along a 100 m transect line at each sampling site. The other dataset of 115 AGB observations is from the previous literature (Fu et al., 2017), in which the sampling sites were concentrated in the eastern part of the TP (Fig. 1b). Therefore, the AGB of 345 fenced grassland sites were obtained in this study, covering the three main grassland types on the Tibetan Plateau.
Climate data are from WorldClim: Global Climate Data (www.worldclim.org), including monthly minimum temperature, maximum temperature, and precipitation for the current period (2000-2018) and the next 20 years (2021- 2040). Future climate data are multi-model datasets from the Coupled Model Intercomparison Project Phase 6 (CMIP6). In this study, the output data of four global climate models (CanESM5, CNRM-CM6-1, IPSL-CM6A-LR, AND MRI-ESM2-0) were selected. Each model contains four sets of data for the latest possible future scenarios (SSP1-2.6, SSP2-4.5, SSP3-7.0 and SSP5-8.5) with a spatial resolution of 2.5'×2.5'. The SSPX-Y scenario is based on different shared socio-economic paths and the latest anthropogenic emissions data, combined with a rectangle of radiative forcing magnitude by 2100 (Si et al., 2020). Specifically, SSP1-2.6 SSP2-4.5, SSP3-7.0, and SSP5-8.5 are the scenarios of low radiative forcing (2.6 W m-2), moderate radiative forcing (4.5 W m-2), medium to high radiative forcing (7.0 W m-2), and high radiative forcing (8.5 W m-2) in 2100 under the social development conditions. Climate data from these four global climate models are averaged in this study because the multi-model averaging method is better than any single model in simulating regional climate.
Topography data include elevation (m), slope (°) and slope direction (°). Shuttle Radar Topography Mission (SRTM) Digital Elevation Model (DEM) data are the radar digital elevation data with a total area of more than 1.19×108 km2 which were obtained by the SRTM system carried by the US Space shuttle Endeavour, and covers more than 80% of the world’s land surface. The spatial resolution of the original DEM data is 30 m×30 m (https://lta.cr.usgs.gov/GTOPO30), which was resampled to 2.5'×2.5' in ArcGIS10.0 (Environmental Systems Research Institute, Inc., ESRI) to match the spatial resolution of the other data sources. Correspondingly, slope and aspect were also calculated at 2.5'×2.5'.
Soil data include soil texture (sand, silt, clay content), soil pH and soil organic matter (SOM, g kg-1). The spatial distribution data for soil texture on the TP were obtained from the Data Center for Resources and Environmental Sciences, Institute of Geographic Sciences and Natural Resources Research, CAS (http://www.resdc.cn/data.aspx?DATAID=260). The data were compiled based on the 1:1000000 soil type map and soil profile data obtained from the second national soil survey. The soil texture can be divided into Sand, Silt and Clay according to the contents (%) of sand, silt and clay. Soil pH and SOM are derived from Land-Atmosphere Interaction Research Group at Sun Yat-sen University (http://globalchange.bnu.edu.cn/research/soil2). The vegetation type data comes from the 1:1000000 Chinese Vegetation Atlas published in 2001, which is a basic map of the Chinese natural resources and natural conditions (Hou, 2001).

2.3 Methods

2.3.1 Estimation of potential aboveground biomass

Random forest is a high-precision combinational machine learning algorithm suitable for processing high-dimensional and multicollinear datasets (Tin, 1998; Breiman, 2001; Xia et al., 2018). The RF algorithm uses the Bootstrap method to repeatedly and randomly extract samples from the original data and construct sets of decision trees to output the classification or mean prediction of individual trees during training (Lopatin et al., 2016). Meanwhile, RF can calculate the nonlinearity and reflect the interactions among variables (Li, 2013). In this study, the RF algorithm was used to calculate AGBp on the TP from 2000 to 2018. The prediction variables included 14 variables of climate, soil and topography data (Table 2). Specifically, the six climate variables included growing season (May to September in each year), diurnal temperature range (GSDR, ℃), growing season precipitation (GSP, mm), growing season temperature (GST, ℃), non-growing season diurnal temperature range (NGSDR, ℃), non-growing season precipitation (NGSP, mm), and non-growing season temperature (NGST, ℃). The three topography included elevation (m), slope (°) and slope direction (°). The five soil variables included soil texture (sand, silt, clay content, %), soil pH and soil organic matter (SOM, g kg-1).
Table 2 List of abbreviations used in this study
Classification Abbreviation Meaning
Eco-geographical regions TP Tibetan Plateau
IB1 Golog-Nagqu high-cold shrub-meadow zone
IIAB1 Western Sichuan-eastern Tibet montane coniferous forest zone
IC1 Southern Qinghai high-cold meadow steppe zone
IC2 Qiangtang high-cold steppe zone
ID1 Kunlun high-cold desert zone
IIC1 Southern Tibet montane shrub-steppe zone
IIC2 Eastern Qinghai-Qilian montane steppe zone
IID1 Nagri montane desert-steppe and desert zone
IID2 Qaidam montane desert zone
IID3 Northern slopes of Kunlun montane desert zone
OA1 Southern slopes of Himalaya montane evergreen broad-leaved forest zone
Grass
and livestock		 AGBp (Potential, only climate-derived) Aboveground biomass of the grassland
LCCT Theoretical livestock carrying capacity
SSU The standardized sheep unit (daily feed of 1.33 kg hay in this study)
Climate and soil GSDR Growing season (May to September in each year) diurnal temperature range
GSP Growing season precipitation
GST Growing season temperature
NGSDR Non-growing season diurnal temperature range
NGSP Non-growing season precipitation
NGST Non-growing season temperature
SOM Soil organic matter
We predicted the dynamics of AGBp under four scenarios in the next 20 years based on the 2021-2040 WorldClim scenario data. In this study, the RF model used to predict AGBp is run in R randomForest packages (https://cran.rproject.org/web/packages/randomForest). The RF algorithm requires two key parameters: 1) Mtry, which specifies the number of independent variables used for binary data by each node of the classification tree, usually one-third of the number of variables; and 2) Ntree, the number of decision trees, generally 500 by default (Li, 2013). In this study, mtry is set to 4 and ntree is set to 500. We calculated the coefficient of determination (R2), normalized root mean square error (equation (1)), and the Bias (1 minus the slope of the regression between the predicted and observed values) to validate the AGBp simulated by the RF model compared to the observed grassland AGBp. A higher R2 value and lower nRMSE and Bias indicate that the model fit is better.
$nRMSE=\frac{RMSE}{[\max (AG{{B}_{p}})-\min (AG{{B}_{p}})]}\times 100、%$

2.3.2 Estimation of theoretical livestock carrying capacity

The theoretical livestock carrying capacity (LCCT) was established by the following equation (2):
$LC{{C}_{T}}=\frac{AG{{B}_{p}}\times U\times C\times H}{S\times Gt}$
where AGBp is the potential aboveground biomass of grassland per unit area (g m‒2), U is pasture utilization rate (0.70), C is the availability of grassland (0.84) (Fan et al., 2010; Cao et al., 2020), H is the edible forage ratio of the grassland, with values for alpine meadow, alpine grassland and alpine desert grassland of 0.76, 0.69 and 0.76 according to the field observation data on the TP, respectively. S is the daily feed of 1.33 kg hay for the standardized sheep unit (SSU) (Fan et al., 2010), and Gt is grazing time in each year (365 days) (Cao et al., 2019).

2.3.3 Statistical analysis

In this study, a linear trend, namely the slope derived from the least square fitting curve, was adopted to simulate the change trend and intensity of AGBp on the TP at the given time scale (equation (3)):
$S =\frac{n\times \sum\limits_{i=1}^{n}{(i\times {{x}_{i}})}-\sum\limits_{i=1}^{n}{i}\sum\limits_{i=1}^{n}{{{x}_{i}}}}{n\times \sum\limits_{i=1}^{n}{{{i}^{2}}}-{{\left( \sum\limits_{i=1}^{n}{i} \right)}^{2}}}$
where S is the slope of the regression trend, and a positive S indicates that the variables show an increasing trend during the study period while a zero or negative S indicates a stable or decreasing trend, n is the study time length, and xi represents the variable value of the grid in year i.

3 Results

3.1 Potential aboveground biomass estimation and validation

The results showed that GSP and NGSP had the highest correlations with AGBp among the 14 environmental factors in all sampling sites on the TP, with correlation coefficients of 0.59 (P < 0.01) and 0.51 (P < 0.01), respectively. AGBp decreased with the increase of altitude (R = -0.57, P < 0.01) (Fig. 2a). We found the AGBp estimations from the RF model had a significant linear correlation with the AGBp observations of alpine grasslands on the TP (R2=0.76, P < 0.01) with low nRMSE and Bias of 10.87 and 0.05, respectively (Fig. 2b).
Fig. 2 Correlation coefficients of grassland AGBp with 14 environmental factors (a) and the relationships between observed grassland AGBp and estimated AGBp on the TP based on the RF model (b)

3.2 Spatial and temporal patterns of potential aboveground biomass

The average annual AGBp of grassland on the TP was 102.4 g m-2 (Table 3). However, the spatial variation of AGBp on the TP was obvious, which showed a decreasing trend from southeast to northwest (Fig. 3a). For example, AGBp on the IB1 and IIC2 were generally over 150 g m-2, but IID1 and IC2 had lower AGBp values of approximately 50 g m-2 (Table 3). Overall, the AGBp of alpine grassland increased from 2000-2018, but not significantly (y = 0.14x-186.45, P > 0.05, Fig. 3). The fastest increase of the AGBp was in the northeastern Tibetan Plateau, especially in the Qilian Mountain steppe region (IIC2), where the AGBp increased 1.04 g m-2 yr-1 (Fig. 3); while in the southeastern and central Tibetan Plateau it showed a decreasing trend (Fig. 3b).
Table 3 The mean values and trends of grassland AGBp for each eco-region on the TP
Eco-geographical regions* AGBp mean (g m-2) AGBp trend (g m-2 yr-1)
Mean Standard deviation (SD) Mean Standard deviation (SD)
IB1 181.64 52.87 0.47 0.80
ICI 93.68 32.03 0.31 0.60
IC2 53.81 24.90 -0.11 0.30
ID1 55.90 7.70 0.02 0.08
IIAB1 196.47 46.29 -0.23 0.69
IIC2 167.11 44.29 1.04 0.59
IIC1 80.47 25.17 -0.16 0.29
IID2 109.63 24.70 0.09 0.22
OA1 228.15 35.00 -0.53 0.48
IID1 56.42 24.77 0.36 0.60
IID3 92.39 23.35 -0.02 0.11
TP 102.40 63.47 0.14 0.61

Note: * The eco-geographical regions of the TP are the same as indicated in Table 2.

Fig. 3 The spatial and temporal patterns of AGBp on the TP from 2000 to 2018

Note: The eco-geographical regions of the TP are the same as indicated in Table 2.

3.3 Climate effects on potential aboveground biomass

We assumed that no significant changes in soil and topographic conditions occurred over the decades, thus we only analyzed the effects of interannual dynamics of climate factors on AGBp from 2000 to 2018 on the TP. The results showed that the interannual variation of six climatic factors (GSDR, GSP, GST, NGSDR, NGSP and NGST) drove the changes of AGBp on the TP despite their insignificant trends (P > 0.05) (Fig. 4). In particular, the effects of GSDR and GSP on AGBp were the most obvious (Fig. 5). Specifically, the AGBp decreased by 15.13 g m-2 with each 1 ℃ increase of GSDR (Fig. 5a), while it increased by 0.19 g m-2 for every 1 mm increase of GSP (Fig. 5b). The other four climatic factors (GST, NGSDR, NGSP and NGST) had no significant effects on AGBp from 2000 to 2018 on the TP (P > 0.05, Figs. 5c-5f).
Fig. 4 The dynamics of growing season and non-growing season diurnal temperature, temperature, and precipitation on the Tibetan Plateau from 2000 to 2018.

Note: The six climate variables from (a) to (f) are the same as indicated in Table 2.

Fig. 5 Correlations between AGBp and climatic variables from 2000 to 2018 on the TP

Note: The six climate variables are the same as indicated in Table 2.

3.4 Spatial and temporal patterns of theoretical livestock carrying capacity

The average theoretical carrying capacity (LCCT) of grassland was 0.936 SSU ha‒1 for the past two decades on the TP. The spatial patterns of LCCT of grassland on the TP showed that it gradually decreases from southeast to northwest (Fig. 6a). Specifically, the eastern areas of the IB1 and IIC2 had higher LCCT with average LCCT over 2 SSU ha-1, while the western areas of IIDI, IC2, and ID1 had average LCCT that was generally lower than 0.5 SSU ha-1 (Fig. 6a). In addition, the temporal trends of grassland LCCT in different eco- geographical regions of the TP were also not consistent (Fig. 6b). Specifically, the LCCT showed an increasing trend for 54.1% of the area mainly concentrated on the eastern and western ends of the TP, while the decreasing LCCT was distributed in the central and southern parts of the TP (Fig. 6b).
Fig. 6 The temporal-spatial patterns of the theoretical livestock carrying capacity from 2000 to 2018 on the TP

Note: The eco-geographical regions of the TP are the same as indicated in Table 2.

3.5 Predictions of potential aboveground biomass and theoretical livestock carrying capacity

Compared with the past two decades, the AGBp will generally decrease on the TP in 2021-2040 under the four scenarios (SSP1-2.6, SSP2-4.5, SSP3-7.0 and SSP5-8.5) (Fig. 7 and Table 4). The largest decrease is by 7.96% in the AGBp of SSP5-8.5, while the smallest decrease is by 3.05% under SSP2-4.5 (Table 4). In the future, the spatial patterns of grassland AGBp changes are similar under the four scenarios, with grassland decreasing in the central and northern parts and increasing in the southwestern and southeastern areas of the TP (Fig. 7). Specifically, under the scenarios of SSP1-2.6, SSP2-4.0, and SSP3-7.0, the AGBp will decrease mainly in IC1, ID1, and IIC2 (Fig. 7a-c). Among them, it will decrease the most in IC1 by 16.33%, 14.05% and 14.37% of AGBp under the three scenarios, respectively (Table 4). However, the areas of AGBp increases are mainly in IID1, IC2 and IIAB1 on the TP (Figs. 7a-c). Among them, the AGBp increases the most in IID1 by 4.73%, 4.21% and 1.39% in the three scenarios, respectively (Table 4). Compared with the other three scenarios, the SSP5-8.5 scenario has a stronger CO2 emission intensity and higher radiative forcing. Therefore, the AGBp of most regions of the TP will show a declining trend for this scenario, except for slightly increasing trends in IIAB1 and OA1, in which the magnitudes of the decreases in IC1, IC2, and ID1 were over 10% (Fig. 7d).
Fig. 7 The grassland potential aboveground biomass (AGBp) changes in different future climate change scenarios for each eco-region on the TP compared to the past two decades

Note: The eco-geographical regions of the TP are the same as indicated in Table 2. The different climate change scenarios of SSP1-2.6 SSP2-4.5, SSP3-7.0, and SSP5-8.5 (a-d) are the scenarios of low radiative forcing (2.6 W m-2), moderate radiative forcing (4.5 W m-2), medium to high radiative forcing (7.0 W m-2), and high radiative forcing (8.5 W m-2) in 2100 under the different social development conditions.

Table 4 Grassland potential aboveground biomass (AGBp) changes in different future climate change scenarios for each eco-region on the TP compared to the past two decades (Unit: g m-2)
Eco-geographical regions* SSP1-2.6 SSP2-4.5 SSP3-7.0 SSP5-8.5
IB1 -2.70±9.54 -2.24±9.31 -2.16±9.60 -2.40±10.10
IC1 -16.33±10.27 -14.05±10.32 -14.37±10.69 -16.29±11.05
IC2 -0.67±18.05 0.81±18.09 0.63±18.76 -11.51±17.43
ID1 -7.79±8.29 -7.74±8.37 -8.59±8.73 -17.47±7.44
IIAB1 2.75±5.73 2.48±5.64 2.84±6.07 2.28±6.07
IIC2 -7.11±8.8 -8.14±8.54 -8.43±8.28 -6.01±9.10
IIC1 -0.98±9.29 -0.72±9.31 -0.22±9.45 -1.24±8.96
IID2 -4.66±7.11 -4.71±6.85 -4.76±6.81 -4.33±7.11
OA1 1.94±2.69 1.30±1.93 1.61±2.28 0.93±2.01
IID1 4.73±17.9 4.21±18.01 1.39±17.27 -1.07±21.35
IID3 -5.03±8.21 -4.18±7.24 -4.11±7.46 -5.27±8.00
Tibetan Plateau -3.75±13.82 -3.05±13.7 -3.25±14.05 -7.96±14.14

Note: *The eco-geographical regions of the TP are the same as indicated in Table 2. The different climate change scenarios are the same as indicated in Fig. 7.

The LCCT on the TP will also show significant decreasing trends under the four future scenarios compared with the past two decades, with decreasing rates of 3.74%, 3.42%, 3.53% and 5.34%, respectively (Fig. 8). In contrast, the magnitude of the decrease in LCCT under the SSP5-8.5 scenario is significantly higher than those of the other three climate scenarios (Fig. 8).
Fig. 8 The theoretical livestock carrying capacity of alpine grasslands on the TP in the past two decades (2000-2018) and in the future two decades (2021-2040) under four climate change scenarios

Note: The different climate change scenarios are the same as indicated in Fig. 7.

4 Discussion

4.1 Comparison of alpine grassland aboveground biomass

The results showed that the average annual AGBp of grassland on the TP was 102.4 g m-2, which was higher than reported in previous studies, such as Gao et al. (2020) (59.63 g m-2), Xia et al. (2018) (78.40 g m-2) and Zeng et al. (2019) (77.12 g m-2). The higher AGBp estimations based on the RF model may be caused by the different data sources.
RF is a data-driven algorithm, which is an estimation model based on the relationship between the aboveground biomass of grassland and remote sensing and/or environmental variables (Zeng et al., 2019). Therefore, differences in sampling sites, sampling times and corresponding prediction variables of biomass will affect the accuracy of the model performance. In contrast, previous simulations of grassland AGB on the TP did not distinguish whether the sampling data were inside or outside of fenced areas. The biomass outside the fenced area is affected not only by climate change, but also by land use changes, mowing, feeding, and trampling by livestock and wild animals. Therefore, a model constructed using the AGB of grassland outside of the fenced areas simulates the actual AGB of residual grassland driven by both climate and human activities, thus it is not suitable for the assessment of grassland carrying capacity. However, the AGB of only the fenced grassland excludes the influences of large herbivorous animals, and is thus only driven by climate change. Therefore, in this study, we used the AGB of fenced grassland and corresponding environmental variables to construct an RF model for more accurate estimations on AGBp and the livestock carrying capacity of natural grassland on the TP.

4.2 Climate effects on potential aboveground biomass

The AGBp showed no significant increase overall, but there was a large inter-annual fluctuation on the TP in the past two decades (Fig. 3a). Alpine grassland on the TP is sensitive and vulnerable to climate change (Shen et al., 2015; Yao et al., 2018; Piao et al., 2019; Niu et al., 2021). In contrast, precipitation during the growing season (GSP) was one of the main factors controlling the inter-annual fluctuation of alpine grassland AGBp on the TP. Previous studies also suggested that precipitation was the limiting factor of grassland vegetation growth (Knapp and Smith, 2001; Bai et al., 2004; Li et al., 2019). Many studies on the TP also found that alpine grassland vegetation was more sensitive to precipitation changes. For example, Wu et al. (2016) revealed that GSP explained the AGB variation of grassland in the northern Tibet Plateau much more than other climatic factors. Fu et al. (2018) also found that increasing precipitation had a stronger effect on grassland productivity than increasing temperature. Generally, increasing temperature can improve vegetation productivity by enhancing soil nitrogen mineralization and lengthening the growing season (Wang et al., 2017; Liu et al., 2019). However, in this study, the effects of average temperature during growing season (GST) on AGBp were not significant (Fig. 5c). The diurnal temperature range of the growing season (GSDR) had a significant effect on AGBp (Fig. 5a). This result suggested that the diurnal temperature variation may be more important than the mean temperature for grassland vegetation on the TP. The diurnal range is the difference between daily maximum and minimum temperatures and is an important indicator of global climate change. Compared with average temperature, the diurnal temperature range can reflect the global and regional characteristics of the magnitude of temperature variations (Chen and Chen, 2007). In addition, the diurnal temperature range is more sensitive to cloud cover, soil moisture, precipitation and relative humidity (Dai et al., 1999). Since photosynthesis in the alpine grassland plants takes place during the day time and respiration lasts throughout the whole day and night, the change in the diurnal temperature range specifically affects vegetation productivity by changing the scale of photosynthesis and respiration (Phillips et al., 2011). Actullay, the atmospheric CO2 concentration may also accelerate the land carbon sequestration, thus stimulate the AGBp (Liu et al., 2019), which was not considered in this study. However, recent studies also suggested that CO2 fertilization effects on vegetation photosynthesis was declined (Wang et al., 2020; Winkler et al., 2021), even has no effects in some semi-arid grassland (Song et al., 2019). Besides, the decline trends and magnitude of CO2 fertilization effects on vegetation photosynthesis were fraught with uncertainty (Frankenberg et al., 2021; Sang et al., 2021; Zhu et al., 2021). Therefore, more studies on CO2 fertilization effects on vegetation photosynthesis are still needed for a better predict in potential vegetation productivity.

4.3 Theoretical carrying capacity and adaptive strategies

The theoretical livestock carrying capacity (LCCT) of the alpine grassland on the TP will significantly decrease in the future two decades under any of the four scenarios (Fig. 8). The most significant decreases of the AGBp and LCCT were in the central IC1, northeastern IIC2 and northwestern ID1 areas, while the most significant increases were in the western IIDI and IC2 areas on the TP (Fig. 7). Taking necessary measures to mitigate and adapt to climate change in grassland ecosystems is of great significance to the sustainable development of the grassland (Zhang et al., 2015a; Zhang et al., 2016). The grazing policy on the TP should be adjusted in a timely manner according to the temporal and spatial changes in the amount of livestock carried on the grassland (Cao et al., 2019). Reasonable and effective management schemes should be adopted in accordance with local conditions, rather than remaining static and unchanged (Zhang et al., 2015a; He et al., 2016). Specifically, we suggested optimizing the management of pasture production and livestock production in areas where the grassland carrying capacity decreases greatly, such as in IC1, IIC2, and ID1 on the TP. On the one hand, the negative impact of climate change on grassland can be alleviated by legitimately developing artificial grassland in areas with reasonable hydrothermal conditions (Zhang et al., 2016), strengthening the combination of agriculture and animal husbandry (He et al., 2016), and improving grain and feed production (Shi and Zhang, 2020). On the other hand, the grazing pressure of grassland can be alleviated by improving the structure of herds, increasing the strength of corral feeding and so on. However, for the areas with a large increase in carrying capacity, such as IIDI and IC2 on the TP, we suggest further clarifying the balance of grassland and livestock, and identifying potential grazing areas and grazing pressure areas (Cao et al., 2019). Future grazing activities should be transferred to the potential areas as much as possible, and the favorable climate change should be utilized to promote vegetation protection and restoration in grazing pressure areas, in order to realize the complementary regional advantages.

5 Conclusions

In this study, the aboveground biomass observations of the fenced grasslands and their corresponding climatic, soil and topographic data were used to simulate and predict the AGBp and livestock carrying capacity (LCCT) of the alpine grassland on the TP by using the random forest algorithm. This analysis showed that the average AGBp and LCCT of alpine grassland on the TP were 102.4 g m-2 and 0.94 SSU ha-1, respectively, which showed increasing but insignificant trends in the past two decades. The inter-annual fluctuation of AGBp was mainly dependent on precipitation and diurnal temperature range in the growing season. If we do not consider the divergent effects on CO2 fertilization effects on vegetation photosynthesis among two periods, future climate change will likely cause the AGBp in the central IC1, northeastern IIC2 and northwestern ID1 areas on the TP to have significant decreases, while in the western IIDI and IC2 areas it will have significant increases. Therefore, our study suggests that taking necessary measures to mitigate and adapt to climate change in the grassland ecosystem is of great significance to the sustainable development of the TP alpine grassland.

References

Publishing order | Descend order by publishing year | Descend order by cited within
[1]
Bai Y F, Han X G, Wu J G, et al. 2004. Ecosystem stability and compensatory effects in the Inner Mongolia grassland. Nature, 431: 181-184.

DOI

[2]
Breiman L. 2001. Random forests. Machine Learning, 45(1): 5-32.

DOI

[3]
Cao Y N, Wu J S, Zhang X Z, et al. 2019. Dynamic forage-livestock balance analysis in alpine grasslands on the Northern Tibetan Plateau. Journal of Environmental Management, 238: 352-359.

DOI

[4]
Cao Y N, Wu J S, Zhang X Z, et al. 2020. Comparison of methods for evaluating the forage-livestock balance of alpine grasslands on the Northern Tibetan Plateau. Journal of Resources and Ecology, 11(3): 272-282.

DOI

[5]
Chen B X, Zhang X Z, Tao J, et al. 2014. The impact of climate change and anthropogenic activities on alpine grassland over the Qinghai-Tibet Plateau. Agricultural and Forest Meteorology, 189-190: 11-18.

DOI

[6]
Chen H, Zhu Q A, Peng C H, et al. 2013. The impacts of climate change and human activities on biogeochemical cycles on the Qinghai-Tibetan Plateau. Global Change Biology, 19(10): 2940-2955.

DOI

[7]
Chen T X, Chen X. 2007. Variation of diurnal temperature range in China in the past 50 years. Plateau Meteorology, 26(1): 150-157. (in Chinese)

[8]
Dai A G, Trenberth K E, Karl T R. 1999. Effects of clouds, soil moisture, precipitation, and water vapor on diurnal temperature range. Journal of Climate, 12(8): 2451-2473.

DOI

[9]
Dong S K, Li J P, Li X Y, et al. 2010. Application of design theory for restoring the “black beach” degraded rangeland at the headwater areas of the Qinghai-Tibetan Plateau. African Journal of Agricultural Research, 5(25): 3542-3552.

[10]
Fan J W, Shao Q Q, Liu J Y, et al. 2010. Assessment of effects of climate change and grazing activity on grassland yield in the Three Rivers Headwaters Region of Qinghai-Tibet Plateau, China. Environmental Monitoring and Assessment, 170(1-4): 571-584.

DOI

[11]
Frankenberg C, Yin Y, Byrne B, et al. 2021. Comment on “Recent global decline of CO2 fertilization effects on vegetation photosynthesis”. Science, 373(6562): eabg2947. DOI: 10.1126/science.abg2947.

DOI

[12]
Fu G, Shen Z X, Zhang X Z. 2018. Increased precipitation has stronger effects on plant production of an alpine meadow than does experimental warming in the Northern Tibetan Plateau. Agricultural and Forest Meteorology, 249: 11-21.

DOI

[13]
Fu G, Sun W, Li S W, et al. 2017. Modeling aboveground biomass using MODIS images and climatic data in grasslands on the Tibetan Plateau. Journal of Resources and Ecology, 8(1): 42-49.

DOI

[14]
Gao X X, Dong S K, Li S, et al. 2020. Using the random forest model and validated MODIS with the field spectrometer measurement promote the accuracy of estimating aboveground biomass and coverage of alpine grasslands on the Qinghai-Tibetan Plateau. Ecological Indicators, 112: 106114. DOI: 10.1016/j.ecolind.2020.106114.

DOI

[15]
Harris R B. 2010. Rangeland degradation on the Qinghai-Tibetan Plateau: A review of the evidence of its magnitude and causes. Journal of Arid Environments, 74(1): 1-12.

DOI

[16]
He Y T, Zhang X Z, Yu C. 2016. Coupling crop farming and pastoral system for regional development and their ecological effects on the Tibetan Plateau. Bulletin of Chinese Academy of Sciences, 31(1): 112-117. (in Chinese)

[17]
Hou X Y. 2001. Vegetation atlas of China. Beijing, China: Science Press. (in Chinese)

[18]
Jiao C C, Yu G R, Ge J P, et al. 2017. Analysis of spatial and temporal patterns of aboveground net primary productivity in the Eurasian steppe region from 1982 to 2013. Ecology and Evolution, 7(14): 5149-5162.

DOI

[19]
Karhu K, Fritze H, Tuomi M, et al. 2010. Temperature sensitivity of organic matter decomposition in two boreal forest soil profiles. Soil Biology and Biochemistry, 42(1): 72-82.

DOI

[20]
Knapp A K, Smith M D. 2001. Variation among biomes in temporal dynamics of aboveground primary production. Science, 291(5503): 481-484.

PMID

[21]
Li M, He Y T, Fu G, et al. 2016. Livestock-forage balance in the Three River Headwater Region based on the terrestrial ecosystem model. Ecology and Environmental Sciences, 25(12): 1915-1921. (in Chinese)

[22]
Li M, Wu J S, Song C Q, et al. 2019. Temporal variability of precipitation and biomass of alpine grasslands on the Northern Tibetan Plateau. Remote Sensing, 11(3): 360. DOI: 10.3390/rs11030360.

DOI

[23]
Li X H. 2013. Using “random forest” for classification and regression. Chinese Journal of Applied Entomology, 50(4): 1190-1197. (in Chinese)

[24]
Liu Y W, Piao S L, Gasser T, et al. 2019. Field-experiment constraints on the enhancement of the terrestrial carbon sink by CO2 fertilization. Nature Geoscience, 12(10): 809-814.

DOI

[25]
Liu Y X, Zhao W W, Hua T, et al. 2019. Slower vegetation greening faced faster social development on the landscape of the Belt and Road Region. Science of the Total Environment, 697: 134103. DOI: 10.1016/j.scitotenv.2019.134103.

DOI

[26]
Long R J, Apori S O, Castro F B, et al. 1999. Feed value of native forages of the Tibetan Plateau of China. Animal Feed Science and Technology, 80(2): 101-113.

DOI

[27]
Long R J, Shang Z H, Guo X S, et al. 2009. Case study 7: Qinghai-Tibetan Plateau rangelands. Rangeland Degradation & Recovery in China’s Pastoral Lands, 25(4): 29-36.

[28]
Lopatin J, Dolos K, Hernández H J, et al. 2016. Comparing generalized linear models and random forest to model vascular plant species richness using LiDAR data in a natural forest in central Chile. Remote Sensing of Environment, 173(315): 200-210.

DOI

[29]
Luo T X, Li W H, Leng Y F, et al. 1998. Estimation of total biomass and potential distribution of net primary productivity in the Tibetan Plateau. Geographical Research, 17(4): 2-9. (in Chinese)

[30]
Luo T X, Li W H, Zhu H Z. 2002. Estimated biomass and productivity of natural vegetation on the Tibetan Plateau. Ecological Applications, 12(4): 980-997.

DOI

[31]
Ma W H, Fang J Y, Yang Y H, et al. 2010. Biomass carbon stocks and their changes in Northern China’s grasslands during 1982-2006. Science China (Life Sciences), 53(7): 841-850.

[32]
Ni J. 2004. Forage yield-based carbon storage in grasslands of China. Climatic Change, 67(2-3): 237-246.

DOI

[33]
Niu B, He Y T, Zhang X Z, et al. 2016. Tower-based validation and improvement of MODIS gross primary production in an alpine swamp meadow on the Tibetan Plateau. Remote Sensing, 8(7): 592. DOI: 10.3390/rs8070592.

DOI

[34]
Niu B, He Y T, Zhang X Z, et al. 2020. Satellite-based estimates of canopy photosynthetic parameters for an alpine meadow in Northern Tibet. Journal of Resources and Ecology, 11(3): 253-262.

DOI

[35]
Niu B, Zeng C X, Zhang X Z, et al. 2019. High below-ground productivity allocation of alpine grasslands on the Northern Tibet. Plants, 8(12): 535. DOI: 10.3390/plants8120535

DOI

[36]
Niu B, Zhang X Z, He Y T, et al. 2017. Satellite-based estimation of gross primary production in an alpine swamp meadow on the Tibetan Plateau: A multi-model comparison. Journal of Resources and Ecology, 8(1): 57-66.

DOI

[37]
Niu B, Zhang X Z, Piao S L, et al. 2021. Warming homogenizes apparent temperature sensitivity of ecosystem respiration. Science Advances, 7(15): eabc7358. DOI: 10.1126/sciadv.abc7358.

DOI

[38]
Phillips C L, Gregg J W, Wilson J K. 2011. Reduced diurnal temperature range does not change warming impacts on ecosystem carbon balance of Mediterranean grassland mesocosms. Global Change Biology, 17(11): 3263-3273.

DOI

[39]
Piao S L, Fang J Y, He J S, et al. 2004. Spatial distribution of grassland biomass in China. Acta Phytoecologica Sinica, 4: 491-498. (in Chinese)

[40]
Piao S L, Zhang X Z, Wang T, et al. 2019. Responses and feedback of the Tibetan Plateau’s alpine ecosystem to climate change. Chinese Science Bulletin, 64(27): 2842-2855. (in Chinese)

[41]
Sang Y X, Huang L, Wang X H, et al. 2021. Comment on “Recent global decline of CO2 fertilization effects on vegetation photosynthesis”. Science, 373(6562): eabg4420. DOI: 10.1126/science.abg4420.

DOI

[42]
Scurlock J M O, Hall D O. 1998. The global carbon sink: A grassland perspective. Global Change Biology, 4(2): 229-233.

DOI

[43]
Shang Z H, Gibb M J, Leiber F, et al. 2014. The sustainable development of grassland-livestock systems on the Tibetan Plateau: Problems, strategies and prospects. The Rangeland Journal, 36(3): 267-296.

[44]
Shen M G, Piao S L, Cong N, et al. 2015. Precipitation impacts on vegetation spring phenology on the Tibetan Plateau. Global Change Biology, 21(10): 3647-3656.

DOI

[45]
Shi P L, Zhang X Z. 2020. Towards regional synergy: Reconciling rangeland ecological functioning with forage production of cultivated pasture. Journal of Resources and Ecology, 11(3): 247-252.

DOI

[46]
Si S, Bi X, Kong X, et al. 2020. Spatial-temporal characteristics of the emission intensities of several major greenhouse gases and aerosols under CMIP6 scenarios. Climatic and Environmental Research, 25(4): 366-384. (in Chinese)

[47]
Song J, Wan S Q, Piao S L, et al. Elevated CO2 does not stimulate carbon sink in a semi-arid grassland. Ecology Letters, 22(3): 458-468.

DOI

[48]
Tan K, Ciais P, Piao S L, et al. 2010. Application of the ORCHIDEE global vegetation model to evaluate biomass and soil carbon stocks of Qinghai-Tibetan grasslands. Global Biogeochemical Cycles, 24(1): GB1013. DOI: 10.1029/2009GB003530.

DOI

[49]
Tin K H. 1998. The random subspace method for constructing decision forests. IEEE Transactions on Pattern Analysis and Machine Intelligence, 20(8): 832-844.

DOI

[50]
Wang S H, Zhang Y G, Ju W M, et al. 2020. Recent global decline of CO2 fertilization effects on vegetation photosynthesis. Science, 370(6522): 1295-1300.

DOI

[51]
Wang S H, Zhang Y G, Ju W M, et al. 2021. Response to comments on “Recent global decline of CO2 fertilization effects on vegetation photosynthesis”. Science, 373(6562): eabg7484. DOI: 10.1126/science.abg7484.

DOI

[52]
Wang S Y, Zhang B, Yang Q C, et al. 2017. Responses of net primary productivity to phenological dynamics in the Tibetan Plateau, China. Agricultural and Forest Meteorology, 232: 235-246.

DOI

[53]
Winkler A J, Myneni R B, Hannart A, et al. 2021. Slowdown of the greening trend in natural vegetation with further rise in atmospheric CO2. Biogeosciences, 18(17): 4985-5010.

DOI

[54]
Wu J S, Feng Y F, Zhang X Z, et al. 2017. Grazing exclusion by fencing non-linearly restored the degraded alpine grasslands on the Tibetan Plateau. Scientific Reports, 7(1): 15202. DOI: 10.1038/s41598-017-15530-2.

DOI

[55]
Wu J S, Fu G. 2018. Modelling aboveground biomass using MODIS FPAR/LAI data in alpine grasslands of the Northern Tibetan Plateau. Remote Sensing Letters, 9(2): 150-159.

DOI

[56]
Wu J S, Wurst S, Zhang X Z. 2016. Plant functional trait diversity regulates the nonlinear response of productivity to regional climate change in a)Tibetan alpine grasslands. Scientific Reports, 6: 35649. DOI: 10.1038/s41598-017-15530-2.

DOI

[57]
Xia J Z, Ma M N, Liang T G, et al. 2018. Estimates of grassland biomass and turnover time on the Tibetan Plateau. Environmental Research Letters, 13(1): 014020. DOI: 10.1038/s41598-017-15530-2.

DOI

[58]
Xu L L, Niu B, Zhang X Z, et al. 2021. Dynamic threshold of carbon phenology in two cold temperate grasslands in China. Remote Sensing, 13(4): 574. DOI: 10.3390/rs13040574.

DOI

[59]
Xu W H, Xiao Y, Zhang J J, et al. 2017. Strengthening protected areas for biodiversity and ecosystem services in China. Proceedings of the National Academy of Sciences of the USA, 114(7): 1601-1606.

DOI

[60]
Yang G, Peng C H, Chen H, et al. 2017. Qinghai-Tibetan Plateau peatland sustainable utilization under anthropogenic disturbances and climate change. Ecosystem Health and Sustainability, 3(3): e01263. DOI: 10.1002/ehs2.1263.

DOI

[61]
Yang Y H, Fang J Y, Ma W H, et al. 2010. Large-scale pattern of biomass partitioning across China’s grasslands. Global Ecology and Biogeography, 19(2): 268-277.

DOI

[62]
Yang Y H, Fang J Y, Pan Y D, et al. 2009. Aboveground biomass in Tibetan grasslands. Journal of Arid Environments, 73(1): 91-95.

DOI

[63]
Yao T D, Thompson L G, Mosbrugger V, et al. 2012. Third pole environment (TPE). Environmental Development, 3: 52-64.

DOI

[64]
Yao Y T, Wang X H, Li Y, et al. 2018. Spatiotemporal pattern of gross primary productivity and its covariation with climate in China over the last thirty years. Global Change Biology, 24(1): 184-196.

DOI

[65]
Zeng N, Ren X L, He H L, et al. 2019. Estimating grassland aboveground biomass on the Tibetan Plateau using a random forest algorithm. Ecological Indicators, 102: 479-487.

DOI

[66]
Zhang X Z, He Y T, Shen Z X, et al. 2015a. Frontier of the ecological construction support the sustainable development in Tibet Autonomous Region. Bulletin of the Chinese Academy of Sciences, 30(3): 306-312. (in Chinese)

[67]
Zhang X Z, Wang X, Gao Q, et al. 2016. Research in ecological restoration and reconstruction technology for degraded alpine ecosystem, boosting the protection and construction of ecological security barrier in Tibet. Acta Ecologica Sinica, 36(22): 7083-7087. (in Chinese)

[68]
Zhang X Z, Yang Y, Piao S, et al. 2015b. Ecological change on the Tibetan Plateau. Chinese Science Bulletin, 60(32): 3048-3056. (in Chinese)

[69]
Zhang Y, Gao Q Z, Dong S K, et al. 2015c. Effects of grazing and climate warming on plant diversity, productivity and living state in the alpine rangelands and cultivated grasslands of the Qinghai-Tibetan Plateau. The Rangeland Journal, 37(1): 57-65.

DOI

[70]
Zhang Y, Li B, Zheng D. 2002. A discussion on the boundary and area of the Tibetan Plateau in China. Geographical Research, 21(1): 1-8. (in Chinese)

[71]
Zhang Y Q, Tang Y H, Jiang J, et al. 2007. Characterizing the dynamics of soil organic carbon in grasslands on the Qinghai-Tibetan Plateau. Science in China (Series D), 50(1): 113-120.

DOI

[72]
Zheng D. 1996. Natural region system research of Tibetan Plateau. Science in China (Series D), 26(4): 336-341. (in Chinese)

[73]
Zhu Z C, Zeng H, Myneni R B, et al. 2021. Comment on “Recent global decline of CO2 fertilization effects on vegetation photosynthesis”. Science, 373(6562): eabg5673. DOI: 10.1126/science.abg5673.

DOI

Options
Outlines

/

〈 〉