Measuring the spatiotemporal variations of vegetation net primary productivity in Inner Mongolia using spatial autocorrelation

https://doi.org/10.1016/j.ecolind.2020.106108Get rights and content

Highlights

  • Spatial autocorrelation indices were used to analyse spatial variations of NPP.

  • NPP in Inner Mongolia demonstrated significant spatial clustering.

  • Distribution of high and low NPP was polarized in Inner Mongolia.

  • NPP internal agglomerations were observed in Inner Mongolia.

  • Inter-annual variations of NPP were holistic in internal agglomeration.

Abstract

Monitoring variations in vegetation net primary productivity (NPP) contributes greatly to the evaluation of terrestrial ecosystem structure and function. Current research has focused on the trend changes of vegetation productivity but has largely ignored spatial correlation. This study selected MOD17A3 NPP data in Inner Mongolia to explore the spatiotemporal variations of NPP using spatial autocorrelation indices (Global Moran’s I, Getis-Ord General G, Getis-Ord Gi*, and Anselin Local Moran’s I). This is helpful to examine whether NPP in Inner Mongolia has exhibited significant spatial cluster based on temporal variations and spatial patterns. NPP in Inner Mongolia has shown significant spatial cluster in high-value areas from 2000 to 2014 and demonstrated a polarized distribution. Although the east part of the study area generally showed a high NPP pattern, the northeast turned to low NPP pattern in 2001, 2004, 2007, and 2010 and the southeast turned to low NPP pattern in 2009 and 2013. The former variations were primarily affected by precipitation, while the latter variations were affected by both hydrothermal conditions and human activities. These results are useful to analyze spatial correlations and local abnormalities of NPP and provide references for integrated ecosystem management in Inner Mongolia.

Introduction

Vegetation net primary productivity (NPP) is a vital ecological index for estimating terrestrial ecosystem function and provides comprehensive high-quality information for renewable resource measurement (Wuliangha et al., 2016, Zhang et al., 2017, Ren et al., 2018, Xu and Ge, 2018). Studying NPP variations over a large area is potentially useful to evaluate the interactions between climate change and terrestrial ecosystems. A large body of research describes NPP variations using linear regression from regional to pixel scales. A linear regression approach quantifies the spatiotemporal variations of NPP using the unary linear regression equation and significance tests (Gang et al., 2015, Liu et al., 2015, Li et al., 2018, Mao et al., 2018). The approach is widely applied in studies that simulate and predict vegetation growth (Liao et al., 2018). Another approach, known as standard deviation analysis, is often calculated at pixel scale. It can be used to estimate NPP fluctuations over time. The method is widely used to study ecosystem stability (Li et al., 2012). Spatiotemporal variations of NPP based on these methods are known as “first-order effect” (Huang et al., 2013). The first-order effect does not include spatial correlations and local anomalies.

To alleviate this problem, the study used the geostatistics (Moran, 1950, Getis and Ord, 1992, Ord and Getis, 1995, Anselin, 1995) to compensate for the shortcomings of commonly used methods. The spatial autocorrelation in the geostatistics expresses the tendency of cluster (in high- or low-value) or outlier between the attributes of spatially adjacent positions. This method can be used to explore the spatial correlation of adjacent positions. This is called the “second-order effect” (Huang et al., 2013). Using geostatistics to reveal the second-order effect is essential for monitoring spatiotemporal variations of NPP and analyzing the influences of climate change and anthropogenic activities.

Inner Mongolia, an important ecological barrier, is located in north China. In recent decades, human activities such as cattle grazing and urban construction have caused serious damage to vegetation ecosystems in the study area (Ren and Zhou, 2018). The variations of hydrothermal conditions have also had great impacts on vegetation ecosystems in this region (Li et al., 2018, Qiao and Wang, 2019). Thus, the vegetation productivity in Inner Mongolia have shown obvious spatial dependence and local heterogeneity (Ren and Zhou, 2018). This study explored the spatiotemporal variations of NPP in Inner Mongolia using spatial autocorrelation, in particular global and local spatial autocorrelation. We hypothesized that the spatial autocorrelation would accurately reveal the spatiotemporal variations. We arrived at this hypothesis based on geostatistics theory and vegetation variations in Inner Mongolia. This study also investigated the main factors that may affect the spatiotemporal variations of NPP in anomalous areas in the region.

Section snippets

Study area

The Inner Mongolia Autonomous Region (referred to as Inner Mongolia) (97°12′E to 126°04′E longitude, 37°24′N to 53°23′N latitude) is situated in northern China. It extends obliquely from northeast China toward the southwest and is approximately 1.18 million km2 in area (Tong et al., 2019). The annual mean temperature increases from −1 °C in the northeast to 10 °C in the southwest, and the annual precipitation declines from 450 mm to 50 mm across the study area (Li et al., 2018). Due to the

Global correlation variations of NPP

As showed in Table 1, the Global Moran’s I, Getis-Ord General G, their expectations and z-scores were calculated with NPP data in the study area from 2000 to 2014. The values of Global Moran’s I varied from 0.59 to 0.87 and were significant at the 99% confidence level (Z (I) > 2.58). This indicated that NPP in the study area showed significant spatial correlation from 2000 to 2014. General G observations are all larger than expectations (P < 0.001). This suggested that NPP in the study area

Discussion

To analyze spatiotemporal variations of NPP correlations in Inner Mongolia, this study introduced eco-geographic regionalization of China. The eco-geographic regional system has been developed by the Chinese Academy of Sciences (Wu et al., 2003). The system was generated according to natural characteristics, major ecological elements (biotic and non-biotic), and geographic zonation (Wu et al., 2003). The Inner Mongolia was delineated into 12 natural regions that include 3 temperature and 4

Conclusions

This study explored the spatial cluster distributions and variations of NPP in Inner Mongolia from 2000 to 2014 using spatial autocorrelation. Results showed spatial correlation of NPP caused by high-value cluster. NPP was clustered with the maximum of Moran’s I (0.87) in 2009 and the minimum of Moran’s I (0.59) in 2007. Distribution of high and low NPP were polarized in Inner Mongolia. Two eco-geographic areas (IIB2 and IIB1 (South)-IIC1-IIIB3) with local anomalies were observed. In IIB2,

Declaration of Competing Interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Acknowledgments

This study is supported by the National Key R&D Program of China (Project No. 2016YFD0300201) and the National Science Foundation of China (Project No. 41801078).

Authors contributions sections

Hongrui Ren, Yingjie Shang, and Shuai Zhang conceived and designed the research. Hongrui Ren and Yingjie Shang analyzed the data and wrote the manuscript.

References (28)

  • Y. Huang et al.

    Graphic analysis of spatio-temporal effect for vegetation disturbance caused by coal mining: a case of Datong coal mine area

    Acta Ecol. Sinica

    (2013)
  • S.Y. Li et al.

    Spatial analysis on the fluctuation of regional NPP in the typical steppe, in China

    Adv. Mater. Res.

    (2012)
  • W. Li et al.

    Analysis of spatial-temporal variation in NPP based on hydrothermal conditions in the Lancang-Mekong River Basin from 2000 to 2014

    Environ. Monit. Assess.

    (2018)
  • C.G. Liao et al.

    Spatial and temporal variation pattern of vegetation NPP in Guangxi based on GIS

    Asian Agric. Res.

    (2018)
  • Cited by (60)

    View all citing articles on Scopus
    View full text