- Email: [email protected]
Temporospatial patterns of human appropriation of net primary production in Central Asia grasslands
Ecological Indicators 91 (2018) 555–561
Contents lists available at ScienceDirect
Ecological Indicators journal homepage: www.elsevier.com/locate/ecolind
Original Articles
Temporospatial patterns of human appropriation of net primary production in Central Asia grasslands
T
⁎
Xiaotao Huanga,b, Geping Luoa,b, , Qifei Hanc a
State Key Laboratory of Desert and Oasis Ecology, Xinjiang Institute of Ecology and Geography, Chinese Academy of Sciences, Urumqi 830011, Xinjiang, China University of Chinese Academy of Sciences, Beijing 100049, China c Nanjing University of Information Science & Technology, Nanjing 210044, Jiangsu, China b
A R T I C LE I N FO
A B S T R A C T
Keywords: HANPP Temporospatial patterns Dr Model Grazing
Quantifying and mapping grassland human appropriation of net primary production (HANPP) is vital for the sustainable use of grasslands. However, grazing process was not effectively considered in previous studies, leading to biased results. Although grasslands are widespread in Central Asia, temporospatial patterns of HANPP in Central Asia grasslands are still unclear. In this study, to effectively consider grazing process, we used the Biome-Biogeochemical Cycles grazing model to estimate HANPP and explore its temporospatial patterns in Central Asia grasslands from 1979 to 2012. In our study, net primary production includes aboveground and belowground components. Our estimates showed that HANPP was 47 g C/m2/yr, which represented 34% of Central Asia grassland potential net primary production (NPPpot) and HANPP efficiency was 70% in this region. Interannual variations in HANPP and HANPP as a percentage of NPPpot (HANPP%NPPpot) were significantly positively related to grazing intensity (P < 0.01). Interannual variation in HANPP efficiency was negatively related to grazing intensity (P < 0.1). HANPP showed strong regional variation. High HANPP values were mainly observed in temperate grassland and some forest meadow. Low HANPP values were mainly observed in desert grassland and some forest meadow. The spatial pattern of HANPP%NPPpot was similar to that of HANPP in this region. Interannual variations in HANPP were mainly determined by population change and economic development. Spatial patterns of HANPP were primarily determined by grazing intensity and grazing system. This study contributes to a better understanding of the temporospatial patterns of HANPP in Central Asia grasslands and provides data to support the rational use of grassland resources.
1. Introduction Human appropriation of net primary production (HANPP) is the difference between potential net primary production (NPPpot) and actual net primary production (NPPact) remaining in the ecosystem after harvest (Haberl et al., 2007; Ma et al., 2012). It derives from consumption of terrestrial photosynthesis products, as well as the loss of biomass caused by changes in human land use (Haberl et al., 2007; Ma et al., 2012). It represents the degree to which human society influences natural ecological systems and is based on the knowledge that the material and energy required for human survival and development is dependent on the net primary production (NPP) (Rojstaczer et al., 2002; Vitousek et al., 1986). NPP is the net amount of carbon assimilated in a given period by vegetation and determines the amount of
energy available for transfer from plants to other trophic levels (Artacho and Bonomelli, 2017; Chen et al., 2017; Deng et al., 2017). HANPP not only reduces the food sources available to other species, but it also distinctly alters the material and energy flow in biogeochemical cycles (BGC) and within food webs (Haberl et al., 2005, 2004a). Quantifying HANPP is important for ecological assessment of regional sustainable development. Furthermore, HANPP estimation has received considerable attention in the international community and has become a major method for ecological assessment (Krausmann et al., 2013, 2012). Grasslands are widespread in Central Asia and serve important economic and ecological functions, such as material production, climate regulation, soil and water conservation, sand stabilization, soil improvement, and biodiversity preservation (Eichelmann et al., 2016;
Abbreviations: BGC, biogeochemical cycles; HANPP, human appropriation of net primary production; HANPP%NPPpot, human appropriation of net primary production as a percentage of potential net primary production; KAZ, Kazakhstan; KYR, Kyrgyzstan; NPPact, actual net primary production; NPPpot, potential net primary production; TJK, Tajikistan; TKM, Turkmenistan; UZB, Uzbekistan ⁎ Corresponding author at: State Key Laboratory of Desert and Oasis Ecology, Xinjiang Institute of Ecology and Geography, Chinese Academy of Sciences, Urumqi 830011, Xinjiang, China. E-mail address: [email protected] (G. Luo). https://doi.org/10.1016/j.ecolind.2018.04.045 Received 2 December 2017; Received in revised form 15 April 2018; Accepted 17 April 2018 1470-160X/ © 2018 Elsevier Ltd. All rights reserved.
Ecological Indicators 91 (2018) 555–561
X. Huang et al.
Huang et al., 2017; Zhang et al., 2016). Due to lack of water resources, grassland ecosystems are considered highly sensitive and vulnerable to human disturbance and climatic change in this region (Han et al., 2016; Zhang et al., 2012). These grassland resources are commonly used for grazing (Han et al., 2016; Zhang et al., 2012). The vast rangelands of Central Asia form the world’s largest contiguous area of grazed land (Freitag and Wucherer, 2005; Han et al., 2016). Quantifying HANPP can help us to further assess grazing effects on grassland ecosystems in terms of carbon cycling, ecosystem services, and sustainability (Haberl et al., 2012, 2004b; Rojstaczer et al., 2002). However, the current studies in this region lack sufficient detail to provide adequate knowledge in this field. Thus, we do not have a clear understanding of the temporospatial patterns of HANPP in this region, which is detrimental to promoting rational use of Central Asia’s grassland resources. Previous studies mainly combined models with statistical data to estimate HANPP in grazed land. However, the grazing process was not effectively considered, leading to biased results (Haberl et al., 2007; Krausmann et al., 2013, 2012). Luo et al. (2012) developed the BiomeBGC grazing model by integrating a defoliation formulation (Seligman et al., 1992) into the Biome-BGC model, producing a grazing model for describing the effects of grazing on the carbon cycle of grassland ecosystems. The Biome-BGC grazing model is a process-based model that effectively estimates NPPpot, NPPact, and the carbon consumed by animals (Han et al., 2016; Luo et al., 2012). Therefore, HANPP can be estimated in Central Asia grasslands using this model. In this study, based on the Biome-BGC grazing model, HANPP was estimated in Central Asia grasslands from 1979 to 2012. The primary objectives of this study were to investigate interannual variation in grassland HANPP from 1979 to 2012 and to explore spatial patterns of grassland HANPP in Central Asia. In our study, NPP includes aboveground and belowground components.
2. Materials and methods
Fig. 1. Distribution of grassland types (a) and grazing intensity (b) in Central Asia from 1979 to 2012.
2.1. Study area
2.2. Methods
Central Asia, with widespread grasslands, occupies nearly 5.7 × 106 km2 and includes five republic countries [Kazakhstan (KAZ), Kyrgyzstan (KYR), Tajikistan (TJK), Turkmenistan (TKM), and Uzbekistan (UZB)] and Xinjiang in China (Fig. 1a) (Cowan, 2007). Because Central Asia is in the Eurasian hinterland and far from ocean, it has the typical continental arid climate characteristics. Average annual precipitation is less than 300 mm, however, significant spatial differences in precipitation exist. For example, annual precipitation in the desert is less than 100 mm, or even less than 50 mm, while that in the mountains is more than 500 mm, and even up to 1000 mm (Buslov et al., 2007). Three grassland types occur depending on terrain and climate: forest meadow, temperature grassland, and desert grassland. Grassland above 1650 m a.s.l. is forest meadow. Grassland below 1650 m a.s.l. is either temperate or desert grassland, depending on the climate. Temperate grassland occurs where plant growth is more restricted by temperature than by precipitation, while desert grassland occurs where plant growth is more restricted by precipitation than by temperature (Fig. 1a). Due to its unique geography and natural conditions, the Central Asia grassland ecosystem is very fragile. Furthermore, the vast rangelands of Central Asia form the world’s largest contiguous area of grazed land. Grazing is the main human disturbance on grasslands in this region. Forest meadow is dominated by seasonal pasture. Desert grassland is dominated by seasonal pasture and rotational grazing, and temperate grassland is dominated by annual pasture (Han et al., 2016; Zhang et al., 2012). Overgrazing (i.e., pronounced decrease in NPP caused by grazing) is serious in some parts of Central Asia, and is not sustainable utilization of grasslands (Fig. 1b).
We used the estimating method of HANPP in grazed lands per Haberl et al. (2007). Grazing is the most important human disturbance on grasslands in Central Asia. Compared with grazing, disturbance from other human activities is very limited in this region (Han et al., 2016; Zhang et al., 2012). In addition, the effects of other human activities are very difficult to estimate because much of the necessary data are unavailable. Therefore, grazing can be considered the only human disturbance when HANPP is estimated in Central Asia grasslands. The Biome-BGC grazing model is a process-based ecosystem model that can effectively simulate the fluxes and storage of carbon, both in grazed and ungrazed grasslands over large areas (Han et al., 2016, 2014). In the model, NPP is estimated as (Luo et al., 2012)
NPP = C′veg + Clitter + Dr
(1)
where C′veg is vegetative carbon, Clitter is litter carbon, and Dr (g C/ (ha·d)) is the carbon consumed by animals (NPP harvest) (Seligman et al., 1992), and
Dr = Ge Sr (Cleaf −(Cleaf ) U)(0 < Dr < Sr Dx )
(2)
where Ge is the grazing efficiency of the livestock (ha/d per sheep unit), Sr is the grazing intensity (sheep/ha), Cleaf is the C in the leaf biomass (g C/m2), (Cleaf)U is the residual aboveground Cleaf that is unavailable to livestock (g C/m2), and Dx is the consumption rate of the livestock based on satiation (g C/(d·sheep)). In accordance with Luo et al., (2012), the grazing efficiency of the livestock is 0.011 ha/d per sheep unit, (Cleaf)U is 6.75 g C/m2, and Dx is 2.4 g C/(d·sheep). In this study, HANPP is calculated as (Haberl, 1997; Haberl et al., 2005, 2004a) 556
Ecological Indicators 91 (2018) 555–561
X. Huang et al.
Table 1 Data sources for Biome-BGC grazing model inputs. Data
Feature
Range
Source
Meteorological data Grazing data
PREC, SRAD, VPD, Tmax, Tmin, Tday, and DAYL Grazing time, grazing intensity
1979–2012 1979–2012
DEM CO2 Soil
Elevation, slope, and aspect Atmospheric CO2 concentrations Sand/silt/clay percentage and effective soil depth
1979–2012 1979–2012 1979–2012
CFSR FAO and corrected in accordance with livestock statistics from the respective governments WorldClim – Global Climate Data (Hijmans et al., 2005) Mauna Loa (ftp://ftp.cmdl.noaa.gov/ccg/- co2/trends/co2annmeanmlo.txt) HWSD
Note: PREC–precipitation, SRAD–solar radiation, VPD–daytime average vapor pressure deficit, Tmax–maximum temperature, Tmin–minimum temperature, Tday–daytime average temperature, DAYL–day length.
HANPP = NPPpot−NPPact + Dr
(3)
where NPPact is the NPP under grazed conditions (g C/m2) and NPPpot is the NPP under ungrazed conditions (g C/m2) (Dr = 0). In this study, we followed the “HANPP efficiency” definition from Fetzel et al. (2014) to analyze the efficiency of HANPP. HANPP efficiency is calculated as
HANPP efficiency = Dr /HANPP
(4)
A high HANPP efficiency is related to a high share of consumption by livestock to total HANPP and indicates the bulk of the appropriated biomass entering the socio-ecological system (Fetzel et al., 2016). 2.3. Data In this study, the required data included model input data and validation data. The model input data included meteorological data, grazing data, and ancillary data (Table 1). All the regional data were smoothed to 40 km × 40 km resolution. Fig. 2. Interannual variations in grazing intensity for the different administrative regions in Central Asia from 1979 to 2012.
2.3.1. Meteorological data Precipitation, solar radiation, daytime average vapor pressure deficit, maximum temperature, minimum temperature, daytime average temperature, and day length were the required daily meteorological data that drove the Biome-BGC grazing model. The regional meteorological data from 1979 to 2012 over Central Asia grasslands were derived from the Climate Forecast System Reanalysis (CFSR). The derived data were compared with observations from meteorological stations to support the accuracy of the dataset (Hu et al., 2014; Lioubimtseva et al., 2005).
Data (Hijmans et al., 2005) (http://www.worldclim.org). Atmospheric CO2 concentrations for 1979–2012 were obtained from observations taken at Mauna Loa (ftp://ftp.cmdl.noaa.gov/ccg/co2/trends/co2 annmean mlo.txt). 2.3.4. NPP validation data The NPP data that were used for model validation were collected from previous publications (Han et al., 2016) and from field observations in Central Asia grasslands. We used 58 plots to sample annual NPP data, among which, 42 plots were outside the enclosure and grazed. 16 plots were inside the enclosure and ungrazed.
2.3.2. Grazing data The grazing data, including grazing intensity and grazing calendar, were extracted from “Gridded Livestock of the World” (GLW) (http:// www.fao.org/AG/againfo/resources/en/glw/home.html) provided by the Food and Agriculture Organization of the United Nations (FAO). The spatial grazing data provided by the FAO were created through the spatial disaggregation of sub-national statistical data, based on empirical relationships with environmental variables in similar agro-ecological zones. We corrected these grazing data in accordance with livestock statistics for the different regions from the respective governments to ensure high precision. In this study, all the grazers were converted into sheep units according to the standard provided by the Ministry of Agriculture of the People’s Republic of China (http://www. chinaforage.com/standard/zaixuliang.htm): one cow equals six sheep, one goat equals 0.9 sheep, one horse equals to six sheep, one camel equals to eight sheep, and one yak equals 4.5 sheep. Fig. 2 shows interannual variations in grazing intensity for the different administrative areas in Central Asia, from 1979 to 2012.
3. Results 3.1. Model validation The Biome-BGC grazing model has been successfully used in studies of carbon emissions in Central Asia grasslands (Han et al., 2016). In this study, NPP estimates were further validated using field data collected from previous publications and through actual observations in Central Asia grasslands. The Biome-BGC grazing model well-simulated NPP for grazed (R2 = 0.88, RMSE = 26.92) (Fig. 3a) and ungrazed grassland (R2 = 0.87, RMSE = 38.01) (Fig. 3b). 3.2. Temporal variation in HANPP in Central Asia grasslands Fig. 4a shows the interannual variations in subcomponents of NPP (HANPP, NPPpot, NPPact, and Dr) in Central Asia grasslands from 1979 to 2012. The correlation coefficient between interannual variations in HANPP and grazing intensity is 0.7849, which suggests a significant positive correlation (P < 0.01). From 1979 to 1994, grazing intensity increased. During this period, HANPP also increased from 34 g C/m2/yr
2.3.3. Ancillary data Soil data, including sand/silt/clay percentages and effective soil depth, were derived from the Harmonized World Soil Database (HWSD). Elevation data were derived from WorldClim – Global Climate 557
Ecological Indicators 91 (2018) 555–561
X. Huang et al.
Fig. 3. Comparisons of annual NPP based on simulated and observed field data under grazed (a) and ungrazed (b) conditions (NPP – net primary production; RMSE – root-mean-square error).
increase of 2.3 g C/m2/yr. Annually, Dr was the main contributor to HANPP in Central Asia grasslands from 1979 to 2012. In general, the difference between NPPpot and NPPact was relatively high during the years with high grazing intensity, and relatively low during the years with low grazing intensity. Fig. 4b shows the interannual variations in HANPP efficiency and HANPP%NPPpot in Central Asia grasslands from 1979 to 2012. HANPP efficiency fluctuated frequently from 1979 to 2012. The correlation coefficient between interannual variations in HANPP efficiency and grazing intensity is −0.2843, which suggests a negative correlation (P < 0.1). This indicates that high HANPP efficiency is more likely to occur during years with low grazing intensity, while low HANPP efficiency is more likely to occur during years with high grazing intensity. The correlation coefficient between interannual variations in HANPP% NPPpot and grazing intensity is 0.7527, which suggests a significant positive correlation (P < 0.01). From 1979 to 1994, grazing intensity increased. During this period, HANPP%NPPpot also increased from 25% to 40%, with an average annual increase of 0.9%. From 1994 to 2000, grazing intensity decreased due to the dissolution of the Soviet Union (Li, 2008; Walker and Blackburn, 2015). HANPP%NPPpot decreased in this period, with an average annual decrease of 2.4%. From 2000 to 2012, grazing intensity increased due to the gradual recovery of livestock production. HANPP%NPPpot increased in this period, with an average annual increase of 0.5%. 3.3. Spatial patterns of HANPP in Central Asia grasslands Fig. 5a displays the spatial distribution of HANPP in Central Asia grasslands. HANPP evaluates the extent to which humans perturb ecosystems (Chen et al., 2015; Haberl et al., 2007). High HANPP values were mainly observed in temperate grassland and some forest meadow. Low HANPP values were mainly observed in desert grassland and some forest meadow. In rare cases, HANPP values were negative, indicating significant overcompensation due to moderate grazing. Fig. 5b displays HANPP as a percentage of NPPpot (HANPP%NPPpot) in each grid cell. The spatial distribution of HANPP%NPPpot is a useful indicator of land use intensity that can quantify and localize changes in ecosystem processes due to human activity (Haberl et al., 2007). In general, the spatial distribution pattern of HANPP%NPPpot was similar to the spatial distribution pattern of HANPP (Fig. 5b). Our estimates showed that HANPP was 47 g C/m2/yr, which represented 34% of Central Asia grassland NPPpot. Among the different administrative regions, HANPP values were ranked from high to low: KAZ, KYR, Xinjiang, TJK, UZB, and TKM. HANPP [% of NPPpot] were ranked from high to low: UZB, TKM, KAZ, TJK, Xinjiang, and KYR. Overall, consumption by livestock grazing (Dr) was 33 g C/m2/yr in
Fig. 4. Interannual variations in subcomponents of NPP (a), and HANPP efficiency and percentage that HANPP accounted for NPPpot (HANPP%NPPpot) (b) in Central Asia grasslands from 1979 to 2012 (NPP – net primary production; NPPpot – potential net primary production or NPP under ungrazed conditions; NPPact – actual net primary production or NPP under grazed conditions; HANPP – human appropriation of net primary production).
to 54 g C/m2/yr, with an average annual increase of 1.2 g C/m2/yr. From 1994 to 2000, grazing intensity decreased remarkably, due to the dissolution of the Soviet Union (Li, 2008; Walker and Blackburn, 2015). HANPP noticeably decreased in this period, with an average annual decrease of 3.7 g C/m2/yr. From 2000 to 2012, grazing intensity increased due to the gradual recovery of livestock production. HANPP showed increased volatility in this period, with an average annual 558
Ecological Indicators 91 (2018) 555–561
X. Huang et al.
noticeably increase in this period (Niedertscheider et al., 2012; Prasad and Badarinth, 2004). This is because husbandry production suffered serious damage after the dissolution of the Soviet Union (Li, 2008; Walker and Blackburn, 2015). During the period 1979–1994, NPPpot values were distinctly higher than NPPact, indicating there was overgrazing in Central Asia grasslands in this period. HANPP values were generally higher from 1979 to 1994. Higher HANPP would leave less resources for other species, and excessive HANPP inevitably led to ecological destruction (Niedertscheider et al., 2012; Saikku et al., 2015). Although the dissolution of the Soviet Union negatively affected livestock production in Central Asia (Li, 2008; Walker and Blackburn, 2015), it was beneficial to the ecological health of grassland ecosystems. As husbandry production gradually recovered from 2000 to 2012, it led to an increase in HANPP in Central Asia grasslands. Interannual variation in HANPP was mainly determined by two factors in this region: population change and economic development (Ma et al., 2012). In general, population growth was relatively fast in Central Asia from 1979 to 1994 and from 2000 to 2012. Population growth led to increasingly intense land use and an increase in HANPP in Central Asia grasslands. Meanwhile, Central Asia experienced rapid economic development during the same periods. Economic development also affected HANPP, as a rising standard of living entailed higher per capita food consumption. However, in the 1990s, slow population growth and slow economic development in Central Asia occurred due to the dissolution of the Soviet Union, which led to a decrease in HANPP in Central Asia grasslands (Li, 2008; Walker and Blackburn, 2015). It should be noted that economic development and grassland protection do not conflict with each other. Economic development does not necessarily lead to high HANPP, because human consumption is the Dr, which is part of HANPP. Increasing HANPP efficiency is a way to solve this problem. For example, technological developments and regionally appropriate grazing management could bring about the growth of grassland NPPact, resulting in increasing harvests without increasing HANPP (Fetzel et al., 2014; Niedertscheider et al., 2012). Around the year 2000, HANPP efficiency was relatively high in Central Asia grasslands and NPPact was almost the same as NPPpot, which helped to achieve grass-livestock balance (Han et al., 2016, 2014). However, in these years, NPP harvest was low due to low grazing intensity, which cannot satisfy the needs for modern economic development. To obtain enough resources for economic development and achieve sustainable utilization of grassland resources, we should rely on technological progress to obtain high NPPact. Moreover, practicable grazing management should be adopted to obtain the optimal HANPP not only to satisfy human needs but also to leave enough resources for other species, under the current technology. HANPP showed a strong regional variation in Central Asia grasslands, as determined by the different grazing densities and grazing systems. In some forest meadow, HANPP values were high because of high grazing intensity. In temperate grassland, the grazing density was low, but HANPP values were high because grazing time was long in this region. Most areas of forest meadow were seasonal pastures, while those of temperate grassland were annual. Low HANPP values were mainly distributed in desert grassland because of low grazing intensity and short grazing time. Most desert grassland areas were seasonal pastures or rotational grazing (Han et al., 2016; Zhang et al., 2012). In rare case, negative HANPP values occurred due to significant overcompensation for moderate grazing, which were supported by previous studies (Han et al., 2016; Leriche et al., 2001; Luo et al., 2012; Zhang et al., 2015). There were distinct differences in government management, market demand, climate, and terrain, which caused differences in grazing densities and grazing systems (Han et al., 2016; Zhang et al., 2012). Therefore, we should adopt management measures in accordance with regional situations.
Fig. 5. Maps of the average annual human appropriation of net primary production (HANPP) (a) and HANPP as a percentage of NPPpot (HANPP%NPPpot) (b) in Central Asia grasslands from 1979 to 2012 (NPPpot – potential net primary production). Table 2 Values of subcomponents of NPP among the different administrative regions in Central Asia grasslands from 1979 to 2012.
2
HANPP (g C/m /yr) HANPP [% of NPPpot] Dr (g C/m2/yr) HANPP efficiency (%)
KAZ
KYR
TJK
TKM
UZB
XJ
CA
53 33 49 92
29 18 26 90
22 25 15 68
5 38 3 60
21 39 9 43
27 23 19 70
47 34 33 70
NPPpot – potential net primary production under ungrazed conditions; HANPP – human appropriation of net primary production; HANPP [% of NPPpot] –percentage that HANPP accounted for NPPpot; Dr– – carbon consumed by livestock; KAZ – Kazakhstan, KYR – Kyrgyzstan, TJK – Tajikistan, TKM – Turkmenistan, UZB – Uzbekistan, XJ – Xinjiang, and CA – Central Asia).
Central Asia grasslands, contributing 70% to total HANPP. Among the different administrative regions, Dr values were ranked from high to low: KAZ, KYR, Xinjiang, TJK, UZB, and TKM. HANPP efficiency were ranked from high to low: KAZ, KYR, Xinjiang, TJK, TKM, and UZB (Table 2). 4. Discussion 4.1. Temporospatial patterns of HANPP in Central Asia grasslands In general, HANPP values in Central Asia grasslands, as observed in this study, did not notably increase from 1979 to 2012, indicating that pressures on grassland ecosystems, caused by livestock grazing, did not 559
Ecological Indicators 91 (2018) 555–561
X. Huang et al.
considered in the model. Trampling can change the soil density and influence ecological processes in grassland ecosystems. However, previous studies showed that the grazing process in this model was more suitable in arid and semiarid grasslands than in humid grasslands (Han et al., 2016; Luo et al., 2012). In addition, further model validation was conducted with respect to NPP simulation results in Central Asia grasslands, which confirmed the reliability of the model. Nevertheless, we hope to improve the grazing process in the model in the future to effectively consider trampling effects. In addition to the above, uncertainty in the results can also arise from the model input data. Meteorological data are the major drivers influencing the model results (Luo et al., 2010; Raj et al., 2014; Sun et al., 2017). The meteorological data used in this study were extracted from CSFR and their accuracy was supported by several statistical measures (Hu et al., 2014). A previous examination justified the use of CSFR in our analysis of the climatic variation in Central Asia. Grazing data were the key inputs used to simulate the effects of grazing (Luo et al., 2012). The grazing data used in this study were derived from the GLIS (http://www.fao.org/docrep/010/a1259e/a1259e00.htm). We corrected the data in accordance with statistics from the respective governments, which ensured high data precision. Nevertheless, we hope that input data with higher precision will be produced in the future.
4.2. Compared with previous studies Certain studies regarding grassland HANPP have been conducted on a global or regional scale. However, previous studies were limited to estimates of total values in grassland HANPP, which did not provide details or adequate knowledge about temporospatial patterns of grassland HANPP. The estimates of grassland HANPP differed in different studies (Chen et al., 2015; Haberl et al., 2007; Kastner et al., 2015; Krausmann et al., 2013; Niedertscheider et al., 2012). Haberl et al. (2007) showed that HANPP in a global grazing area was 94 g C/m2/yr, accounting for 19.4% of NPPpot, while Saikku et al. (2015) showed that HANPP in a grazing area in Finland was 377 g C/m2/yr, accounting for 80% of NPPpot. Chen et al. (2015) showed that human appropriation of aboveground NPP is around 80% in China’s grasslands. In our study, HANPP was 72 g C/m2/yr, accounting for 35% of NPPpot in Central Asia grassland. HANPP efficiency was 70% in this region. In addition to different study areas, methodology and data may also be important factors causing the differences in grassland HANPP estimates. Haberl et al. (2007) estimated NPPpot and NPPact using the Lund-Potsdam-Jena (LPJ) model, and Dr was estimated using statistical data. Saikku et al. (2015) estimated NPPpot based on the growth and yield of natural Norway spruce stands, while NPPact was estimated as 80% of NPPpot, and Dr was estimated using statistical data. In the study by Chen et al. (2015), estimation did not refer to total NPP (including aboveground and belowground components) but to aboveground NPP, which was different from our study and the studies above. They estimated aboveground NPPpot using the Guangsheng Zhou Model, while the aboveground NPPact was estimated using MODIS data, and Dr was estimated using statistical data. In our study, NPPpot, NPPact, and Dr were estimated using the process-based Biome-BGC grazing model, which is functionally holistic in approach. Of course, different methods have their own advantages. In our study, grazing process was effectively considered in the Biome-BGC grazing model to estimate grazing effect on NPP. And the model in our study was calibrated and validated using field data, which was another advantage compared with previous studies. Thus, using this model, we can better estimate HANPP in grazed land. In addition, using outputs of the model, the temporospatial patterns of HANPP in Central Asia grasslands were analyzed in detail. Our study contributes to a better understanding of HANPP in Central Asia grasslands and provides more detailed and reliable data to support the rational use of grassland resources.
5. Conclusions In this study, based on the assumption that grazing is the only human disturbance in Central Asia grasslands and using the Biome-BGC grazing model, we estimated HANPP and explored its temporospatial patterns in Central Asia grasslands from 1979 to 2012. Our estimates showed that HANPP was 47 g C/m2/yr, which represented 34% of NPPpot in Central Asia grasslands and HANPP efficiency was 70% in this region. Interannual variations in HANPP and grazing intensity were significantly positively correlated (P < 0.01). Interannual variations in HANPP efficiency and grazing intensity were negatively correlated (P < 0.1). Interannual variations in HANPP%NPPpot and grazing intensity were significantly positively correlated (P < 0.01). In addition, HANPP showed a strong regional variation because of the differences in grazing intensities and grazing systems. High HANPP values were mainly observed in temperate grassland and some forest meadow. Low HANPP values were mainly observed in desert grassland and some forest meadow. In rare case, negative HANPP values occurred due to significant overcompensation for moderate grazing. The spatial pattern of HANPP%NPPpot was similar to that of HANPP in this region. This study contributes to a better understanding of the temporospatial patterns of HANPP in Central Asia grasslands and provides data to support the rational use of grassland ecosystems. In the future, we should use technological developments and regionally appropriate grazing management to obtain higher NPPact and optimal HANPP.
4.3. Uncertainties in the calculations Our study is the first analysis of temporospatial patterns of HANPP in Central Asia grasslands. However, because carbon cycling is complex, the results in this study are naturally subject to limitations. In this section, we first point out the limitations and then illustrate that the results are unlikely to be strongly impacted. The limitations resulted from the assumption that grazing is the only human activity, model structure, and model input data. In this study, the method for estimating HANPP assumed that grazing was the only human disturbance. However, grazing was the most important human disturbance in Central Asia grasslands and many necessary data regarding the effects of other human activities were unavailable in this region, like many other natural grasslands (Han et al., 2016; Zhang et al., 2012). Thus, other human activities can reasonably be omitted, since they contribute much less than grazing, when estimating grassland HANPP. The results are unlikely to be strongly impacted by this limitation. The model structure itself is vital to the output. The grazing process is complex; thus, it is impossible to fully consider all the components in the Biome-BGC grazing model (Luo et al., 2012; Sun et al., 2017; Ueyama et al., 2010). This led to uncertainty in the simulation results, like many other models. For example, because the mechanism regarding trampling effects remains unclear, it is not effectively
Acknowledgments This work was supported financially by the Strategic Priority Research Program of Chinese Academy of Sciences (Grant No. XDA2006030201) and the National Natural Science Foundation of China (Grant No. 41361140361). The authors declare no conflict of interest. References Artacho, P., Bonomelli, C., 2017. Net primary productivity and allocation to fine-root production in field-grown sweet cherry trees under different soil nitrogen regimes. Sci. Hortic. 219, 207–215. Buslov, M.M., De Grave, J., Bataleva, E.A.V., Batalev, V.Y., 2007. Cenozoic tectonic and geodynamic evolution of the Kyrgyz Tien Shan Mountains: a review of geological, thermochronological and geophysical data. J. Asian Earth Sci. 29 (2–3), 205–214. Chen, A., Li, R., Wang, H., He, B., 2015. Quantitative assessment of human appropriation
560
Ecological Indicators 91 (2018) 555–561
X. Huang et al.
doubled in the 20th century. PNAS 110 (25), 10324–10329. Krausmann, F., et al., 2012. Long-term trajectories of the human appropriation of net primary production: lessons from six national case studies. Ecol. Econ. 77, 129–138. Leriche, H., et al., 2001. Which functional processes control the short-term effect of grazing on net primary production in grasslands? Oecologia 129 (1), 114–124. Li, C., 2008. Analysis of potential between Xinjiang and Central Asia Countries cooperation on fruit industry and animal husbandry. J. Anhui Agric. Sci. 36 (13), 5652–5653. Lioubimtseva, E., Cole, R., Adams, J.M., Kapustin, G., 2005. Impacts of climate and landcover changes in arid lands of Central Asia. J. Arid Environ. 62 (2), 285–308. Luo, G., et al., 2012. Moderate grazing can promote aboveground primary production of grassland under water stress. Ecol. Complexity 11, 126–136. Luo, Z., Sun, O.J., Wang, E., Ren, H., Xu, H., 2010. Modeling productivity in Mangrove forests as impacted by effective soil water availability and its sensitivity to climate change using Biome-BGC. Ecosystems 13 (7), 949–965. Ma, T., Zhou, C., Pei, T., 2012. Simulating and estimating tempo-spatial patterns in global human appropriation of net primary production (HANPP): a consumption-based approach. Ecol. Ind. 23, 660–667. Niedertscheider, M., Gingrich, S., Erb, K.-H., 2012. Changes in land use in South Africa between 1961 and 2006: an integrated socio-ecological analysis based on the human appropriation of net primary production framework. Reg. Environ. Change 12 (4), 715–727. Prasad, V.K., Badarinth, K.V.S., 2004. Land use changes and trends in human appropriation of above ground net primary production (HANPP) in India (1961–98). Geogr. J. 170, 51–63. Raj, R., Hamm, N.A.S., van der Tol, C., Stein, A., 2014. Variance-based sensitivity analysis of BIOME-BGC for gross and net primary production. Ecol. Model. 292, 26–36. Rojstaczer, S., Sterling, S.M., Moore, N., 2002. Human appropriation of net primary production – response. Science 296 (5575) 1969–1969. Saikku, L., Mattila, T., Akujarvi, A., Liski, J., 2015. Human appropriation of net primary production in Finland during 1990–2010. Biomass Bioenergy 83, 559–567. Seligman, N.G., Cavagnaro, J.B., Horno, M.E., 1992. Simulation of defoliation effects on primary production of a warm-season, semiarid perennial-species grassland. Ecol. Model. 60 (1), 45–61. Sun, Q., et al., 2017. An improved Biome-BGC model for estimating net primary productivity of alpine meadow on the Qinghai-Tibet Plateau. Ecol. Model. 350, 55–68. Ueyama, M., et al., 2010. Simulating carbon and water cycles of larch forests in East Asia by the BIOME-BGC model with AsiaFlux data. Biogeosciences 7 (3), 959–977. Vitousek, P.M., Ehrlich, P.R., Ehrlich, A.H., Matson, P.A., 1986. Human appropriation of the products of photosynthesis. Bioscience 36 (6), 368–373. Walker, R., Blackburn, J., 2015. Biothreat reduction and economic development: the case of animal husbandry in Central Asia. Front. Public Health 3 270–270. Zhang, T., Peng, J., Liang, W., Yang, Y., Liu, Y., 2016. Spatial-temporal patterns of water use efficiency and climate controls in China's Loess Plateau during 2000–2010. Sci. Total Environ. 565, 105–122. Zhang, T., Sun, Y., Shi, Z., Feng, G., 2012. Arbuscular Mycorrhizal fungi can accelerate the restoration of degraded spring grassland in Central Asia. Rangeland Ecol. Manage. 65 (4), 426–432. Zhang, Y., et al., 2015. 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. Rangeland J. 37 (1), 57–65.
of aboveground net primary production in China. Ecol. Model. 312, 54–60. Chen, H.Y.H., Brant, A.N., Seedre, M., Brassard, B.W., Taylor, A.R., 2017. The contribution of litterfall to net primary production during secondary succession in the boreal forest. Ecosystems 20 (4), 830–844. Cowan, P.J., 2007. Geographic usage of the terms middle Asia and Central Asia. J. Arid Environ. 69 (2), 359–363. Deng, Q., et al., 2017. Effects of precipitation changes on aboveground net primary production and soil respiration in a switchgrass field. Agric. Ecosyst. Environ. 248, 29–37. Eichelmann, E., Wagner-Riddle, C., Warland, J., Deen, B., Voroney, P., 2016. Evapotranspiration, water use efficiency, and energy partitioning of a mature switchgrass stand. Agric. For. Meteorol. 217, 108–119. Fetzel, T., Gradwohl, M., Erb, K.-H., 2014. Conversion, intensification, and abandonment: A human appropriation of net primary production approach to analyze historic landuse dynamics in New Zealand 1860–2005. Ecol. Econ. 97, 201–208. Fetzel, T., Niedertscheider, M., Haberl, H., Krausmann, F., Erb, K.-H., 2016. Patterns and changes of land use and land-use efficiency in Africa 1980–2005: an analysis based on the human appropriation of net primary production framework. Reg. Environ. Change 16 (5), 1507–1520. Freitag, G., Wucherer, W., 2005. Gintzburger G., Toderich K. N., Mardonov B. K., Mahmudov M. M. Rangelands of the arid and semi-arid zones in Uzbekistan. Montpellier, 2003. 432 p. Bot. J. 90 (3), 471–473. Haberl, H., 1997. Human appropriation of net primary production as an environmental indicator: Implications for sustainable development. Ambio 26 (3), 143–146. Haberl, H., et al., 2007. Quantifying and mapping the human appropriation of net primary production in earth's terrestrial ecosystems. PNAS 104 (31), 12942–12945. Haberl, H., et al., 2005. Human appropriation of net primary production as determinant of avifauna diversity in Austria. Agric. Ecosyst. Environ. 110 (3–4), 119–131. Haberl, H., et al., 2004a. Human appropriation of net primary production and species diversity in agricultural landscapes. Agric. Ecosyst. Environ. 102 (2), 213–218. Haberl, H., et al., 2012. Natural and socioeconomic determinants of the embodied human appropriation of net primary production and its relation to other resource use indicators. Ecol. Ind. 23, 222–231. Haberl, H., Wackernagel, M., Krausmann, F., Erb, K.H., Monfreda, C., 2004b. Ecological footprints and human appropriation of net primary production: a comparison. Land Use Policy 21 (3), 279–288. Han, Q., et al., 2016. Simulated grazing effects on carbon emission in Central Asia. Agric. For. Meteorol. 216, 203–214. Han, Q., Luo, G., Li, C., Xu, W., 2014. Modeling the grazing effect on dry grassland carbon cycling with Biome-BGC model. Ecol. Complexity 17, 149–157. Hijmans, R.J., Cameron, S.E., Parra, J.L., Jones, P.G., Jarvis, A., 2005. Very high resolution interpolated climate surfaces for global land areas. Int. J. Climatol. 25 (15), 1965–1978. Hu, Z., Zhang, C., Hu, Q., Tian, H., 2014. Temperature changes in Central Asia from 1979 to 2011 based on multiple datasets. J. Clim. 27 (3), 1143–1167. Huang, X., Luo, G., Lv, N., 2017. Spatio-temporal patterns of grassland evapotranspiration and water use efficiency in arid areas. Ecol. Res. 32 (4), 523–535. Kastner, T., Erb, K.-H., Haberl, H., 2015. Global human appropriation of net primary production for biomass consumption in the European Union, 1986–2007. J. Ind. Ecol. 19 (5), 825–836. Krausmann, F., et al., 2013. Global human appropriation of net primary production
561
Contents lists available at ScienceDirect
Ecological Indicators journal homepage: www.elsevier.com/locate/ecolind
Original Articles
Temporospatial patterns of human appropriation of net primary production in Central Asia grasslands
T
⁎
Xiaotao Huanga,b, Geping Luoa,b, , Qifei Hanc a
State Key Laboratory of Desert and Oasis Ecology, Xinjiang Institute of Ecology and Geography, Chinese Academy of Sciences, Urumqi 830011, Xinjiang, China University of Chinese Academy of Sciences, Beijing 100049, China c Nanjing University of Information Science & Technology, Nanjing 210044, Jiangsu, China b
A R T I C LE I N FO
A B S T R A C T
Keywords: HANPP Temporospatial patterns Dr Model Grazing
Quantifying and mapping grassland human appropriation of net primary production (HANPP) is vital for the sustainable use of grasslands. However, grazing process was not effectively considered in previous studies, leading to biased results. Although grasslands are widespread in Central Asia, temporospatial patterns of HANPP in Central Asia grasslands are still unclear. In this study, to effectively consider grazing process, we used the Biome-Biogeochemical Cycles grazing model to estimate HANPP and explore its temporospatial patterns in Central Asia grasslands from 1979 to 2012. In our study, net primary production includes aboveground and belowground components. Our estimates showed that HANPP was 47 g C/m2/yr, which represented 34% of Central Asia grassland potential net primary production (NPPpot) and HANPP efficiency was 70% in this region. Interannual variations in HANPP and HANPP as a percentage of NPPpot (HANPP%NPPpot) were significantly positively related to grazing intensity (P < 0.01). Interannual variation in HANPP efficiency was negatively related to grazing intensity (P < 0.1). HANPP showed strong regional variation. High HANPP values were mainly observed in temperate grassland and some forest meadow. Low HANPP values were mainly observed in desert grassland and some forest meadow. The spatial pattern of HANPP%NPPpot was similar to that of HANPP in this region. Interannual variations in HANPP were mainly determined by population change and economic development. Spatial patterns of HANPP were primarily determined by grazing intensity and grazing system. This study contributes to a better understanding of the temporospatial patterns of HANPP in Central Asia grasslands and provides data to support the rational use of grassland resources.
1. Introduction Human appropriation of net primary production (HANPP) is the difference between potential net primary production (NPPpot) and actual net primary production (NPPact) remaining in the ecosystem after harvest (Haberl et al., 2007; Ma et al., 2012). It derives from consumption of terrestrial photosynthesis products, as well as the loss of biomass caused by changes in human land use (Haberl et al., 2007; Ma et al., 2012). It represents the degree to which human society influences natural ecological systems and is based on the knowledge that the material and energy required for human survival and development is dependent on the net primary production (NPP) (Rojstaczer et al., 2002; Vitousek et al., 1986). NPP is the net amount of carbon assimilated in a given period by vegetation and determines the amount of
energy available for transfer from plants to other trophic levels (Artacho and Bonomelli, 2017; Chen et al., 2017; Deng et al., 2017). HANPP not only reduces the food sources available to other species, but it also distinctly alters the material and energy flow in biogeochemical cycles (BGC) and within food webs (Haberl et al., 2005, 2004a). Quantifying HANPP is important for ecological assessment of regional sustainable development. Furthermore, HANPP estimation has received considerable attention in the international community and has become a major method for ecological assessment (Krausmann et al., 2013, 2012). Grasslands are widespread in Central Asia and serve important economic and ecological functions, such as material production, climate regulation, soil and water conservation, sand stabilization, soil improvement, and biodiversity preservation (Eichelmann et al., 2016;
Abbreviations: BGC, biogeochemical cycles; HANPP, human appropriation of net primary production; HANPP%NPPpot, human appropriation of net primary production as a percentage of potential net primary production; KAZ, Kazakhstan; KYR, Kyrgyzstan; NPPact, actual net primary production; NPPpot, potential net primary production; TJK, Tajikistan; TKM, Turkmenistan; UZB, Uzbekistan ⁎ Corresponding author at: State Key Laboratory of Desert and Oasis Ecology, Xinjiang Institute of Ecology and Geography, Chinese Academy of Sciences, Urumqi 830011, Xinjiang, China. E-mail address: [email protected] (G. Luo). https://doi.org/10.1016/j.ecolind.2018.04.045 Received 2 December 2017; Received in revised form 15 April 2018; Accepted 17 April 2018 1470-160X/ © 2018 Elsevier Ltd. All rights reserved.
Ecological Indicators 91 (2018) 555–561
X. Huang et al.
Huang et al., 2017; Zhang et al., 2016). Due to lack of water resources, grassland ecosystems are considered highly sensitive and vulnerable to human disturbance and climatic change in this region (Han et al., 2016; Zhang et al., 2012). These grassland resources are commonly used for grazing (Han et al., 2016; Zhang et al., 2012). The vast rangelands of Central Asia form the world’s largest contiguous area of grazed land (Freitag and Wucherer, 2005; Han et al., 2016). Quantifying HANPP can help us to further assess grazing effects on grassland ecosystems in terms of carbon cycling, ecosystem services, and sustainability (Haberl et al., 2012, 2004b; Rojstaczer et al., 2002). However, the current studies in this region lack sufficient detail to provide adequate knowledge in this field. Thus, we do not have a clear understanding of the temporospatial patterns of HANPP in this region, which is detrimental to promoting rational use of Central Asia’s grassland resources. Previous studies mainly combined models with statistical data to estimate HANPP in grazed land. However, the grazing process was not effectively considered, leading to biased results (Haberl et al., 2007; Krausmann et al., 2013, 2012). Luo et al. (2012) developed the BiomeBGC grazing model by integrating a defoliation formulation (Seligman et al., 1992) into the Biome-BGC model, producing a grazing model for describing the effects of grazing on the carbon cycle of grassland ecosystems. The Biome-BGC grazing model is a process-based model that effectively estimates NPPpot, NPPact, and the carbon consumed by animals (Han et al., 2016; Luo et al., 2012). Therefore, HANPP can be estimated in Central Asia grasslands using this model. In this study, based on the Biome-BGC grazing model, HANPP was estimated in Central Asia grasslands from 1979 to 2012. The primary objectives of this study were to investigate interannual variation in grassland HANPP from 1979 to 2012 and to explore spatial patterns of grassland HANPP in Central Asia. In our study, NPP includes aboveground and belowground components.
2. Materials and methods
Fig. 1. Distribution of grassland types (a) and grazing intensity (b) in Central Asia from 1979 to 2012.
2.1. Study area
2.2. Methods
Central Asia, with widespread grasslands, occupies nearly 5.7 × 106 km2 and includes five republic countries [Kazakhstan (KAZ), Kyrgyzstan (KYR), Tajikistan (TJK), Turkmenistan (TKM), and Uzbekistan (UZB)] and Xinjiang in China (Fig. 1a) (Cowan, 2007). Because Central Asia is in the Eurasian hinterland and far from ocean, it has the typical continental arid climate characteristics. Average annual precipitation is less than 300 mm, however, significant spatial differences in precipitation exist. For example, annual precipitation in the desert is less than 100 mm, or even less than 50 mm, while that in the mountains is more than 500 mm, and even up to 1000 mm (Buslov et al., 2007). Three grassland types occur depending on terrain and climate: forest meadow, temperature grassland, and desert grassland. Grassland above 1650 m a.s.l. is forest meadow. Grassland below 1650 m a.s.l. is either temperate or desert grassland, depending on the climate. Temperate grassland occurs where plant growth is more restricted by temperature than by precipitation, while desert grassland occurs where plant growth is more restricted by precipitation than by temperature (Fig. 1a). Due to its unique geography and natural conditions, the Central Asia grassland ecosystem is very fragile. Furthermore, the vast rangelands of Central Asia form the world’s largest contiguous area of grazed land. Grazing is the main human disturbance on grasslands in this region. Forest meadow is dominated by seasonal pasture. Desert grassland is dominated by seasonal pasture and rotational grazing, and temperate grassland is dominated by annual pasture (Han et al., 2016; Zhang et al., 2012). Overgrazing (i.e., pronounced decrease in NPP caused by grazing) is serious in some parts of Central Asia, and is not sustainable utilization of grasslands (Fig. 1b).
We used the estimating method of HANPP in grazed lands per Haberl et al. (2007). Grazing is the most important human disturbance on grasslands in Central Asia. Compared with grazing, disturbance from other human activities is very limited in this region (Han et al., 2016; Zhang et al., 2012). In addition, the effects of other human activities are very difficult to estimate because much of the necessary data are unavailable. Therefore, grazing can be considered the only human disturbance when HANPP is estimated in Central Asia grasslands. The Biome-BGC grazing model is a process-based ecosystem model that can effectively simulate the fluxes and storage of carbon, both in grazed and ungrazed grasslands over large areas (Han et al., 2016, 2014). In the model, NPP is estimated as (Luo et al., 2012)
NPP = C′veg + Clitter + Dr
(1)
where C′veg is vegetative carbon, Clitter is litter carbon, and Dr (g C/ (ha·d)) is the carbon consumed by animals (NPP harvest) (Seligman et al., 1992), and
Dr = Ge Sr (Cleaf −(Cleaf ) U)(0 < Dr < Sr Dx )
(2)
where Ge is the grazing efficiency of the livestock (ha/d per sheep unit), Sr is the grazing intensity (sheep/ha), Cleaf is the C in the leaf biomass (g C/m2), (Cleaf)U is the residual aboveground Cleaf that is unavailable to livestock (g C/m2), and Dx is the consumption rate of the livestock based on satiation (g C/(d·sheep)). In accordance with Luo et al., (2012), the grazing efficiency of the livestock is 0.011 ha/d per sheep unit, (Cleaf)U is 6.75 g C/m2, and Dx is 2.4 g C/(d·sheep). In this study, HANPP is calculated as (Haberl, 1997; Haberl et al., 2005, 2004a) 556
Ecological Indicators 91 (2018) 555–561
X. Huang et al.
Table 1 Data sources for Biome-BGC grazing model inputs. Data
Feature
Range
Source
Meteorological data Grazing data
PREC, SRAD, VPD, Tmax, Tmin, Tday, and DAYL Grazing time, grazing intensity
1979–2012 1979–2012
DEM CO2 Soil
Elevation, slope, and aspect Atmospheric CO2 concentrations Sand/silt/clay percentage and effective soil depth
1979–2012 1979–2012 1979–2012
CFSR FAO and corrected in accordance with livestock statistics from the respective governments WorldClim – Global Climate Data (Hijmans et al., 2005) Mauna Loa (ftp://ftp.cmdl.noaa.gov/ccg/- co2/trends/co2annmeanmlo.txt) HWSD
Note: PREC–precipitation, SRAD–solar radiation, VPD–daytime average vapor pressure deficit, Tmax–maximum temperature, Tmin–minimum temperature, Tday–daytime average temperature, DAYL–day length.
HANPP = NPPpot−NPPact + Dr
(3)
where NPPact is the NPP under grazed conditions (g C/m2) and NPPpot is the NPP under ungrazed conditions (g C/m2) (Dr = 0). In this study, we followed the “HANPP efficiency” definition from Fetzel et al. (2014) to analyze the efficiency of HANPP. HANPP efficiency is calculated as
HANPP efficiency = Dr /HANPP
(4)
A high HANPP efficiency is related to a high share of consumption by livestock to total HANPP and indicates the bulk of the appropriated biomass entering the socio-ecological system (Fetzel et al., 2016). 2.3. Data In this study, the required data included model input data and validation data. The model input data included meteorological data, grazing data, and ancillary data (Table 1). All the regional data were smoothed to 40 km × 40 km resolution. Fig. 2. Interannual variations in grazing intensity for the different administrative regions in Central Asia from 1979 to 2012.
2.3.1. Meteorological data Precipitation, solar radiation, daytime average vapor pressure deficit, maximum temperature, minimum temperature, daytime average temperature, and day length were the required daily meteorological data that drove the Biome-BGC grazing model. The regional meteorological data from 1979 to 2012 over Central Asia grasslands were derived from the Climate Forecast System Reanalysis (CFSR). The derived data were compared with observations from meteorological stations to support the accuracy of the dataset (Hu et al., 2014; Lioubimtseva et al., 2005).
Data (Hijmans et al., 2005) (http://www.worldclim.org). Atmospheric CO2 concentrations for 1979–2012 were obtained from observations taken at Mauna Loa (ftp://ftp.cmdl.noaa.gov/ccg/co2/trends/co2 annmean mlo.txt). 2.3.4. NPP validation data The NPP data that were used for model validation were collected from previous publications (Han et al., 2016) and from field observations in Central Asia grasslands. We used 58 plots to sample annual NPP data, among which, 42 plots were outside the enclosure and grazed. 16 plots were inside the enclosure and ungrazed.
2.3.2. Grazing data The grazing data, including grazing intensity and grazing calendar, were extracted from “Gridded Livestock of the World” (GLW) (http:// www.fao.org/AG/againfo/resources/en/glw/home.html) provided by the Food and Agriculture Organization of the United Nations (FAO). The spatial grazing data provided by the FAO were created through the spatial disaggregation of sub-national statistical data, based on empirical relationships with environmental variables in similar agro-ecological zones. We corrected these grazing data in accordance with livestock statistics for the different regions from the respective governments to ensure high precision. In this study, all the grazers were converted into sheep units according to the standard provided by the Ministry of Agriculture of the People’s Republic of China (http://www. chinaforage.com/standard/zaixuliang.htm): one cow equals six sheep, one goat equals 0.9 sheep, one horse equals to six sheep, one camel equals to eight sheep, and one yak equals 4.5 sheep. Fig. 2 shows interannual variations in grazing intensity for the different administrative areas in Central Asia, from 1979 to 2012.
3. Results 3.1. Model validation The Biome-BGC grazing model has been successfully used in studies of carbon emissions in Central Asia grasslands (Han et al., 2016). In this study, NPP estimates were further validated using field data collected from previous publications and through actual observations in Central Asia grasslands. The Biome-BGC grazing model well-simulated NPP for grazed (R2 = 0.88, RMSE = 26.92) (Fig. 3a) and ungrazed grassland (R2 = 0.87, RMSE = 38.01) (Fig. 3b). 3.2. Temporal variation in HANPP in Central Asia grasslands Fig. 4a shows the interannual variations in subcomponents of NPP (HANPP, NPPpot, NPPact, and Dr) in Central Asia grasslands from 1979 to 2012. The correlation coefficient between interannual variations in HANPP and grazing intensity is 0.7849, which suggests a significant positive correlation (P < 0.01). From 1979 to 1994, grazing intensity increased. During this period, HANPP also increased from 34 g C/m2/yr
2.3.3. Ancillary data Soil data, including sand/silt/clay percentages and effective soil depth, were derived from the Harmonized World Soil Database (HWSD). Elevation data were derived from WorldClim – Global Climate 557
Ecological Indicators 91 (2018) 555–561
X. Huang et al.
Fig. 3. Comparisons of annual NPP based on simulated and observed field data under grazed (a) and ungrazed (b) conditions (NPP – net primary production; RMSE – root-mean-square error).
increase of 2.3 g C/m2/yr. Annually, Dr was the main contributor to HANPP in Central Asia grasslands from 1979 to 2012. In general, the difference between NPPpot and NPPact was relatively high during the years with high grazing intensity, and relatively low during the years with low grazing intensity. Fig. 4b shows the interannual variations in HANPP efficiency and HANPP%NPPpot in Central Asia grasslands from 1979 to 2012. HANPP efficiency fluctuated frequently from 1979 to 2012. The correlation coefficient between interannual variations in HANPP efficiency and grazing intensity is −0.2843, which suggests a negative correlation (P < 0.1). This indicates that high HANPP efficiency is more likely to occur during years with low grazing intensity, while low HANPP efficiency is more likely to occur during years with high grazing intensity. The correlation coefficient between interannual variations in HANPP% NPPpot and grazing intensity is 0.7527, which suggests a significant positive correlation (P < 0.01). From 1979 to 1994, grazing intensity increased. During this period, HANPP%NPPpot also increased from 25% to 40%, with an average annual increase of 0.9%. From 1994 to 2000, grazing intensity decreased due to the dissolution of the Soviet Union (Li, 2008; Walker and Blackburn, 2015). HANPP%NPPpot decreased in this period, with an average annual decrease of 2.4%. From 2000 to 2012, grazing intensity increased due to the gradual recovery of livestock production. HANPP%NPPpot increased in this period, with an average annual increase of 0.5%. 3.3. Spatial patterns of HANPP in Central Asia grasslands Fig. 5a displays the spatial distribution of HANPP in Central Asia grasslands. HANPP evaluates the extent to which humans perturb ecosystems (Chen et al., 2015; Haberl et al., 2007). High HANPP values were mainly observed in temperate grassland and some forest meadow. Low HANPP values were mainly observed in desert grassland and some forest meadow. In rare cases, HANPP values were negative, indicating significant overcompensation due to moderate grazing. Fig. 5b displays HANPP as a percentage of NPPpot (HANPP%NPPpot) in each grid cell. The spatial distribution of HANPP%NPPpot is a useful indicator of land use intensity that can quantify and localize changes in ecosystem processes due to human activity (Haberl et al., 2007). In general, the spatial distribution pattern of HANPP%NPPpot was similar to the spatial distribution pattern of HANPP (Fig. 5b). Our estimates showed that HANPP was 47 g C/m2/yr, which represented 34% of Central Asia grassland NPPpot. Among the different administrative regions, HANPP values were ranked from high to low: KAZ, KYR, Xinjiang, TJK, UZB, and TKM. HANPP [% of NPPpot] were ranked from high to low: UZB, TKM, KAZ, TJK, Xinjiang, and KYR. Overall, consumption by livestock grazing (Dr) was 33 g C/m2/yr in
Fig. 4. Interannual variations in subcomponents of NPP (a), and HANPP efficiency and percentage that HANPP accounted for NPPpot (HANPP%NPPpot) (b) in Central Asia grasslands from 1979 to 2012 (NPP – net primary production; NPPpot – potential net primary production or NPP under ungrazed conditions; NPPact – actual net primary production or NPP under grazed conditions; HANPP – human appropriation of net primary production).
to 54 g C/m2/yr, with an average annual increase of 1.2 g C/m2/yr. From 1994 to 2000, grazing intensity decreased remarkably, due to the dissolution of the Soviet Union (Li, 2008; Walker and Blackburn, 2015). HANPP noticeably decreased in this period, with an average annual decrease of 3.7 g C/m2/yr. From 2000 to 2012, grazing intensity increased due to the gradual recovery of livestock production. HANPP showed increased volatility in this period, with an average annual 558
Ecological Indicators 91 (2018) 555–561
X. Huang et al.
noticeably increase in this period (Niedertscheider et al., 2012; Prasad and Badarinth, 2004). This is because husbandry production suffered serious damage after the dissolution of the Soviet Union (Li, 2008; Walker and Blackburn, 2015). During the period 1979–1994, NPPpot values were distinctly higher than NPPact, indicating there was overgrazing in Central Asia grasslands in this period. HANPP values were generally higher from 1979 to 1994. Higher HANPP would leave less resources for other species, and excessive HANPP inevitably led to ecological destruction (Niedertscheider et al., 2012; Saikku et al., 2015). Although the dissolution of the Soviet Union negatively affected livestock production in Central Asia (Li, 2008; Walker and Blackburn, 2015), it was beneficial to the ecological health of grassland ecosystems. As husbandry production gradually recovered from 2000 to 2012, it led to an increase in HANPP in Central Asia grasslands. Interannual variation in HANPP was mainly determined by two factors in this region: population change and economic development (Ma et al., 2012). In general, population growth was relatively fast in Central Asia from 1979 to 1994 and from 2000 to 2012. Population growth led to increasingly intense land use and an increase in HANPP in Central Asia grasslands. Meanwhile, Central Asia experienced rapid economic development during the same periods. Economic development also affected HANPP, as a rising standard of living entailed higher per capita food consumption. However, in the 1990s, slow population growth and slow economic development in Central Asia occurred due to the dissolution of the Soviet Union, which led to a decrease in HANPP in Central Asia grasslands (Li, 2008; Walker and Blackburn, 2015). It should be noted that economic development and grassland protection do not conflict with each other. Economic development does not necessarily lead to high HANPP, because human consumption is the Dr, which is part of HANPP. Increasing HANPP efficiency is a way to solve this problem. For example, technological developments and regionally appropriate grazing management could bring about the growth of grassland NPPact, resulting in increasing harvests without increasing HANPP (Fetzel et al., 2014; Niedertscheider et al., 2012). Around the year 2000, HANPP efficiency was relatively high in Central Asia grasslands and NPPact was almost the same as NPPpot, which helped to achieve grass-livestock balance (Han et al., 2016, 2014). However, in these years, NPP harvest was low due to low grazing intensity, which cannot satisfy the needs for modern economic development. To obtain enough resources for economic development and achieve sustainable utilization of grassland resources, we should rely on technological progress to obtain high NPPact. Moreover, practicable grazing management should be adopted to obtain the optimal HANPP not only to satisfy human needs but also to leave enough resources for other species, under the current technology. HANPP showed a strong regional variation in Central Asia grasslands, as determined by the different grazing densities and grazing systems. In some forest meadow, HANPP values were high because of high grazing intensity. In temperate grassland, the grazing density was low, but HANPP values were high because grazing time was long in this region. Most areas of forest meadow were seasonal pastures, while those of temperate grassland were annual. Low HANPP values were mainly distributed in desert grassland because of low grazing intensity and short grazing time. Most desert grassland areas were seasonal pastures or rotational grazing (Han et al., 2016; Zhang et al., 2012). In rare case, negative HANPP values occurred due to significant overcompensation for moderate grazing, which were supported by previous studies (Han et al., 2016; Leriche et al., 2001; Luo et al., 2012; Zhang et al., 2015). There were distinct differences in government management, market demand, climate, and terrain, which caused differences in grazing densities and grazing systems (Han et al., 2016; Zhang et al., 2012). Therefore, we should adopt management measures in accordance with regional situations.
Fig. 5. Maps of the average annual human appropriation of net primary production (HANPP) (a) and HANPP as a percentage of NPPpot (HANPP%NPPpot) (b) in Central Asia grasslands from 1979 to 2012 (NPPpot – potential net primary production). Table 2 Values of subcomponents of NPP among the different administrative regions in Central Asia grasslands from 1979 to 2012.
2
HANPP (g C/m /yr) HANPP [% of NPPpot] Dr (g C/m2/yr) HANPP efficiency (%)
KAZ
KYR
TJK
TKM
UZB
XJ
CA
53 33 49 92
29 18 26 90
22 25 15 68
5 38 3 60
21 39 9 43
27 23 19 70
47 34 33 70
NPPpot – potential net primary production under ungrazed conditions; HANPP – human appropriation of net primary production; HANPP [% of NPPpot] –percentage that HANPP accounted for NPPpot; Dr– – carbon consumed by livestock; KAZ – Kazakhstan, KYR – Kyrgyzstan, TJK – Tajikistan, TKM – Turkmenistan, UZB – Uzbekistan, XJ – Xinjiang, and CA – Central Asia).
Central Asia grasslands, contributing 70% to total HANPP. Among the different administrative regions, Dr values were ranked from high to low: KAZ, KYR, Xinjiang, TJK, UZB, and TKM. HANPP efficiency were ranked from high to low: KAZ, KYR, Xinjiang, TJK, TKM, and UZB (Table 2). 4. Discussion 4.1. Temporospatial patterns of HANPP in Central Asia grasslands In general, HANPP values in Central Asia grasslands, as observed in this study, did not notably increase from 1979 to 2012, indicating that pressures on grassland ecosystems, caused by livestock grazing, did not 559
Ecological Indicators 91 (2018) 555–561
X. Huang et al.
considered in the model. Trampling can change the soil density and influence ecological processes in grassland ecosystems. However, previous studies showed that the grazing process in this model was more suitable in arid and semiarid grasslands than in humid grasslands (Han et al., 2016; Luo et al., 2012). In addition, further model validation was conducted with respect to NPP simulation results in Central Asia grasslands, which confirmed the reliability of the model. Nevertheless, we hope to improve the grazing process in the model in the future to effectively consider trampling effects. In addition to the above, uncertainty in the results can also arise from the model input data. Meteorological data are the major drivers influencing the model results (Luo et al., 2010; Raj et al., 2014; Sun et al., 2017). The meteorological data used in this study were extracted from CSFR and their accuracy was supported by several statistical measures (Hu et al., 2014). A previous examination justified the use of CSFR in our analysis of the climatic variation in Central Asia. Grazing data were the key inputs used to simulate the effects of grazing (Luo et al., 2012). The grazing data used in this study were derived from the GLIS (http://www.fao.org/docrep/010/a1259e/a1259e00.htm). We corrected the data in accordance with statistics from the respective governments, which ensured high data precision. Nevertheless, we hope that input data with higher precision will be produced in the future.
4.2. Compared with previous studies Certain studies regarding grassland HANPP have been conducted on a global or regional scale. However, previous studies were limited to estimates of total values in grassland HANPP, which did not provide details or adequate knowledge about temporospatial patterns of grassland HANPP. The estimates of grassland HANPP differed in different studies (Chen et al., 2015; Haberl et al., 2007; Kastner et al., 2015; Krausmann et al., 2013; Niedertscheider et al., 2012). Haberl et al. (2007) showed that HANPP in a global grazing area was 94 g C/m2/yr, accounting for 19.4% of NPPpot, while Saikku et al. (2015) showed that HANPP in a grazing area in Finland was 377 g C/m2/yr, accounting for 80% of NPPpot. Chen et al. (2015) showed that human appropriation of aboveground NPP is around 80% in China’s grasslands. In our study, HANPP was 72 g C/m2/yr, accounting for 35% of NPPpot in Central Asia grassland. HANPP efficiency was 70% in this region. In addition to different study areas, methodology and data may also be important factors causing the differences in grassland HANPP estimates. Haberl et al. (2007) estimated NPPpot and NPPact using the Lund-Potsdam-Jena (LPJ) model, and Dr was estimated using statistical data. Saikku et al. (2015) estimated NPPpot based on the growth and yield of natural Norway spruce stands, while NPPact was estimated as 80% of NPPpot, and Dr was estimated using statistical data. In the study by Chen et al. (2015), estimation did not refer to total NPP (including aboveground and belowground components) but to aboveground NPP, which was different from our study and the studies above. They estimated aboveground NPPpot using the Guangsheng Zhou Model, while the aboveground NPPact was estimated using MODIS data, and Dr was estimated using statistical data. In our study, NPPpot, NPPact, and Dr were estimated using the process-based Biome-BGC grazing model, which is functionally holistic in approach. Of course, different methods have their own advantages. In our study, grazing process was effectively considered in the Biome-BGC grazing model to estimate grazing effect on NPP. And the model in our study was calibrated and validated using field data, which was another advantage compared with previous studies. Thus, using this model, we can better estimate HANPP in grazed land. In addition, using outputs of the model, the temporospatial patterns of HANPP in Central Asia grasslands were analyzed in detail. Our study contributes to a better understanding of HANPP in Central Asia grasslands and provides more detailed and reliable data to support the rational use of grassland resources.
5. Conclusions In this study, based on the assumption that grazing is the only human disturbance in Central Asia grasslands and using the Biome-BGC grazing model, we estimated HANPP and explored its temporospatial patterns in Central Asia grasslands from 1979 to 2012. Our estimates showed that HANPP was 47 g C/m2/yr, which represented 34% of NPPpot in Central Asia grasslands and HANPP efficiency was 70% in this region. Interannual variations in HANPP and grazing intensity were significantly positively correlated (P < 0.01). Interannual variations in HANPP efficiency and grazing intensity were negatively correlated (P < 0.1). Interannual variations in HANPP%NPPpot and grazing intensity were significantly positively correlated (P < 0.01). In addition, HANPP showed a strong regional variation because of the differences in grazing intensities and grazing systems. High HANPP values were mainly observed in temperate grassland and some forest meadow. Low HANPP values were mainly observed in desert grassland and some forest meadow. In rare case, negative HANPP values occurred due to significant overcompensation for moderate grazing. The spatial pattern of HANPP%NPPpot was similar to that of HANPP in this region. This study contributes to a better understanding of the temporospatial patterns of HANPP in Central Asia grasslands and provides data to support the rational use of grassland ecosystems. In the future, we should use technological developments and regionally appropriate grazing management to obtain higher NPPact and optimal HANPP.
4.3. Uncertainties in the calculations Our study is the first analysis of temporospatial patterns of HANPP in Central Asia grasslands. However, because carbon cycling is complex, the results in this study are naturally subject to limitations. In this section, we first point out the limitations and then illustrate that the results are unlikely to be strongly impacted. The limitations resulted from the assumption that grazing is the only human activity, model structure, and model input data. In this study, the method for estimating HANPP assumed that grazing was the only human disturbance. However, grazing was the most important human disturbance in Central Asia grasslands and many necessary data regarding the effects of other human activities were unavailable in this region, like many other natural grasslands (Han et al., 2016; Zhang et al., 2012). Thus, other human activities can reasonably be omitted, since they contribute much less than grazing, when estimating grassland HANPP. The results are unlikely to be strongly impacted by this limitation. The model structure itself is vital to the output. The grazing process is complex; thus, it is impossible to fully consider all the components in the Biome-BGC grazing model (Luo et al., 2012; Sun et al., 2017; Ueyama et al., 2010). This led to uncertainty in the simulation results, like many other models. For example, because the mechanism regarding trampling effects remains unclear, it is not effectively
Acknowledgments This work was supported financially by the Strategic Priority Research Program of Chinese Academy of Sciences (Grant No. XDA2006030201) and the National Natural Science Foundation of China (Grant No. 41361140361). The authors declare no conflict of interest. References Artacho, P., Bonomelli, C., 2017. Net primary productivity and allocation to fine-root production in field-grown sweet cherry trees under different soil nitrogen regimes. Sci. Hortic. 219, 207–215. Buslov, M.M., De Grave, J., Bataleva, E.A.V., Batalev, V.Y., 2007. Cenozoic tectonic and geodynamic evolution of the Kyrgyz Tien Shan Mountains: a review of geological, thermochronological and geophysical data. J. Asian Earth Sci. 29 (2–3), 205–214. Chen, A., Li, R., Wang, H., He, B., 2015. Quantitative assessment of human appropriation
560
Ecological Indicators 91 (2018) 555–561
X. Huang et al.
doubled in the 20th century. PNAS 110 (25), 10324–10329. Krausmann, F., et al., 2012. Long-term trajectories of the human appropriation of net primary production: lessons from six national case studies. Ecol. Econ. 77, 129–138. Leriche, H., et al., 2001. Which functional processes control the short-term effect of grazing on net primary production in grasslands? Oecologia 129 (1), 114–124. Li, C., 2008. Analysis of potential between Xinjiang and Central Asia Countries cooperation on fruit industry and animal husbandry. J. Anhui Agric. Sci. 36 (13), 5652–5653. Lioubimtseva, E., Cole, R., Adams, J.M., Kapustin, G., 2005. Impacts of climate and landcover changes in arid lands of Central Asia. J. Arid Environ. 62 (2), 285–308. Luo, G., et al., 2012. Moderate grazing can promote aboveground primary production of grassland under water stress. Ecol. Complexity 11, 126–136. Luo, Z., Sun, O.J., Wang, E., Ren, H., Xu, H., 2010. Modeling productivity in Mangrove forests as impacted by effective soil water availability and its sensitivity to climate change using Biome-BGC. Ecosystems 13 (7), 949–965. Ma, T., Zhou, C., Pei, T., 2012. Simulating and estimating tempo-spatial patterns in global human appropriation of net primary production (HANPP): a consumption-based approach. Ecol. Ind. 23, 660–667. Niedertscheider, M., Gingrich, S., Erb, K.-H., 2012. Changes in land use in South Africa between 1961 and 2006: an integrated socio-ecological analysis based on the human appropriation of net primary production framework. Reg. Environ. Change 12 (4), 715–727. Prasad, V.K., Badarinth, K.V.S., 2004. Land use changes and trends in human appropriation of above ground net primary production (HANPP) in India (1961–98). Geogr. J. 170, 51–63. Raj, R., Hamm, N.A.S., van der Tol, C., Stein, A., 2014. Variance-based sensitivity analysis of BIOME-BGC for gross and net primary production. Ecol. Model. 292, 26–36. Rojstaczer, S., Sterling, S.M., Moore, N., 2002. Human appropriation of net primary production – response. Science 296 (5575) 1969–1969. Saikku, L., Mattila, T., Akujarvi, A., Liski, J., 2015. Human appropriation of net primary production in Finland during 1990–2010. Biomass Bioenergy 83, 559–567. Seligman, N.G., Cavagnaro, J.B., Horno, M.E., 1992. Simulation of defoliation effects on primary production of a warm-season, semiarid perennial-species grassland. Ecol. Model. 60 (1), 45–61. Sun, Q., et al., 2017. An improved Biome-BGC model for estimating net primary productivity of alpine meadow on the Qinghai-Tibet Plateau. Ecol. Model. 350, 55–68. Ueyama, M., et al., 2010. Simulating carbon and water cycles of larch forests in East Asia by the BIOME-BGC model with AsiaFlux data. Biogeosciences 7 (3), 959–977. Vitousek, P.M., Ehrlich, P.R., Ehrlich, A.H., Matson, P.A., 1986. Human appropriation of the products of photosynthesis. Bioscience 36 (6), 368–373. Walker, R., Blackburn, J., 2015. Biothreat reduction and economic development: the case of animal husbandry in Central Asia. Front. Public Health 3 270–270. Zhang, T., Peng, J., Liang, W., Yang, Y., Liu, Y., 2016. Spatial-temporal patterns of water use efficiency and climate controls in China's Loess Plateau during 2000–2010. Sci. Total Environ. 565, 105–122. Zhang, T., Sun, Y., Shi, Z., Feng, G., 2012. Arbuscular Mycorrhizal fungi can accelerate the restoration of degraded spring grassland in Central Asia. Rangeland Ecol. Manage. 65 (4), 426–432. Zhang, Y., et al., 2015. 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. Rangeland J. 37 (1), 57–65.
of aboveground net primary production in China. Ecol. Model. 312, 54–60. Chen, H.Y.H., Brant, A.N., Seedre, M., Brassard, B.W., Taylor, A.R., 2017. The contribution of litterfall to net primary production during secondary succession in the boreal forest. Ecosystems 20 (4), 830–844. Cowan, P.J., 2007. Geographic usage of the terms middle Asia and Central Asia. J. Arid Environ. 69 (2), 359–363. Deng, Q., et al., 2017. Effects of precipitation changes on aboveground net primary production and soil respiration in a switchgrass field. Agric. Ecosyst. Environ. 248, 29–37. Eichelmann, E., Wagner-Riddle, C., Warland, J., Deen, B., Voroney, P., 2016. Evapotranspiration, water use efficiency, and energy partitioning of a mature switchgrass stand. Agric. For. Meteorol. 217, 108–119. Fetzel, T., Gradwohl, M., Erb, K.-H., 2014. Conversion, intensification, and abandonment: A human appropriation of net primary production approach to analyze historic landuse dynamics in New Zealand 1860–2005. Ecol. Econ. 97, 201–208. Fetzel, T., Niedertscheider, M., Haberl, H., Krausmann, F., Erb, K.-H., 2016. Patterns and changes of land use and land-use efficiency in Africa 1980–2005: an analysis based on the human appropriation of net primary production framework. Reg. Environ. Change 16 (5), 1507–1520. Freitag, G., Wucherer, W., 2005. Gintzburger G., Toderich K. N., Mardonov B. K., Mahmudov M. M. Rangelands of the arid and semi-arid zones in Uzbekistan. Montpellier, 2003. 432 p. Bot. J. 90 (3), 471–473. Haberl, H., 1997. Human appropriation of net primary production as an environmental indicator: Implications for sustainable development. Ambio 26 (3), 143–146. Haberl, H., et al., 2007. Quantifying and mapping the human appropriation of net primary production in earth's terrestrial ecosystems. PNAS 104 (31), 12942–12945. Haberl, H., et al., 2005. Human appropriation of net primary production as determinant of avifauna diversity in Austria. Agric. Ecosyst. Environ. 110 (3–4), 119–131. Haberl, H., et al., 2004a. Human appropriation of net primary production and species diversity in agricultural landscapes. Agric. Ecosyst. Environ. 102 (2), 213–218. Haberl, H., et al., 2012. Natural and socioeconomic determinants of the embodied human appropriation of net primary production and its relation to other resource use indicators. Ecol. Ind. 23, 222–231. Haberl, H., Wackernagel, M., Krausmann, F., Erb, K.H., Monfreda, C., 2004b. Ecological footprints and human appropriation of net primary production: a comparison. Land Use Policy 21 (3), 279–288. Han, Q., et al., 2016. Simulated grazing effects on carbon emission in Central Asia. Agric. For. Meteorol. 216, 203–214. Han, Q., Luo, G., Li, C., Xu, W., 2014. Modeling the grazing effect on dry grassland carbon cycling with Biome-BGC model. Ecol. Complexity 17, 149–157. Hijmans, R.J., Cameron, S.E., Parra, J.L., Jones, P.G., Jarvis, A., 2005. Very high resolution interpolated climate surfaces for global land areas. Int. J. Climatol. 25 (15), 1965–1978. Hu, Z., Zhang, C., Hu, Q., Tian, H., 2014. Temperature changes in Central Asia from 1979 to 2011 based on multiple datasets. J. Clim. 27 (3), 1143–1167. Huang, X., Luo, G., Lv, N., 2017. Spatio-temporal patterns of grassland evapotranspiration and water use efficiency in arid areas. Ecol. Res. 32 (4), 523–535. Kastner, T., Erb, K.-H., Haberl, H., 2015. Global human appropriation of net primary production for biomass consumption in the European Union, 1986–2007. J. Ind. Ecol. 19 (5), 825–836. Krausmann, F., et al., 2013. Global human appropriation of net primary production
561