Determining the contributions of climate change and human activities to vegetation dynamics in agro-pastural transitional zone of northern China from 2000 to 2015

Highlights

• NPP increased significantly with the proportion of 48.77%.

• The vegetation greening promoted by ecological conservation projects (ERP) is 19.8%, and Climate-promoted 26.93%.

• The greening areas promoted by ERP is mainly in the Loess Plateau, and climate-promoted in northeast and south-center.

Abstract

The vegetation in the agro-pastoral transitional zone of northern China (APTZNC) was significantly restored, and both climate change and ecological restoration projects contributed to vegetation activities with varied proportion. Since few decades ago, APTZNC has undergone significant land degradation and climate change, threatening regional sustainable development, and in response to such ecological crises, multiple ecological restoration projects were implemented, which have caused a profound impact on the terrestrial ecosystem. Taking agro-pastural transitional zone of northern China (APTZNC) as the study area, this study used 16-year (2000–2015) net primary productivity (NPP) as an important indicator of the arid and semi-arid ecosystem's productivity, combing meteorological data in same period to (1) monitor the vegetation dynamics affected by both climate and ecological restoration projects; (2) detect climate changing trend, including annual precipitation, air temperature, and sunlight hours; (3) explicitly distinguish driving forces of climate change and ecological restoration projects on vegetation dynamics based on correlation analysis. The results demonstrated that (1) the annual NPP indicated overall greening (48.77% significant restoration) and partial degradation (0.39% significant degradation) in APTZNC; (2) the annual precipitation was the main factor that widely influences vegetation growth, and the area with significant influence accounted for 55.53%; however, the area with significant temperature influence only accounted for 1%, and the area affected significantly by sunshine hours accounted for 14.33%; (3) In the area of significant greening with proportion of 48.77%, of 26.93% was related to climate change, of 19.80% was related to ecological conservation programs, and of 2.05% was related to multiple factors. In the significantly degraded area with proportion of 0.39%, of 0.1% is related to climate change and of 0.29% is abnormally degraded. Our study is expected to accelerate the understanding of vegetation dynamics and its driving mechanisms, and provide support for scientifically formulating and adjusting ecological restoration projects in APTZNC.

Introduction

Landscapes include ecological processes and socio-economic processes, and are affected by a variety of factors, not only climate change but also human activities (Wu and Hobbs, 2002, Wu and Hobbs, 2007, Ker Rault et al., 2019, Jorda-Capdevila et al., 2019). Separating the relative roles of the two drivers is challenging, but it is critical to understanding and managing the landscape (Burgi et al., 2010, Li et al., 2012, Ye et al., 2013). This is especially true for agro-pastural transitional zone of northern China (APTZNC), which is characterized by extremely fragile ecosystem due to its relatively sparse vegetation and infertile soil. But it is an important ecological barrier in the central and eastern regions of China, providing a wealth of ecosystem services such as livestock products, tourism recreation, and climate regulation (Xu et al., 2009, Xu et al., 2014, Qiao et al., 2018, Liu et al., 2019).

APTZNC is sensitive to the changing of climatic factors, such as precipitation, temperature and sunlight hours, and the resilience to maintain the stability of ecosystem is weak. These climatic factors could directly influence the vegetation growth since the changes in temperature and precipitation can determine the hydrothermal conditions of vegetation growth, especially for arid and semi-arid ecosystems (Chen et al. 2019a). Global temperatures in the past century have increased by approximately 0.74 ℃, and the intensity of precipitation events is expected to increase (IPCC, 2007). Most regions of APTZNC have shown a trend of warming and drying since 1980s (Shi et al., 2014, Zhou et al., 2017). According to forecasts, the annual average temperature in APTZNC continues rising, and the temperature increase will reach to 0.3 ~ 1.5℃ in the future ten decades (Yan et al., 2008). The vegetation activities in this area also suffered from special natural conditions such as less precipitation, strong winds, and severe soil erosion, highly prone to evolution as the climatic factors change (Wang et al., 1999, Xu et al., 2006).

In addition, since the 1950s, political and socio-economic reforms, exponential population growth, over grazing, urbanization have accelerated the long-term decline in ecosystem functions (e.g. net primary production, NPP), culminating in large-scale sustainable development emergencies in APTZNC (Bryan et al., 2018). Therefore, some positive human interventions, especially multiple national ecological restoration projects, were spurred to halt devastating land degradation and to improve human well-being from both socioeconomic and ecological perspectives (Cao et al., 2011, Lv et al., 2012, Mao et al., 2018). Examples of these projects include the the Grain for Green Program, Three Norths Shelter Forest System Project, the Natural Forest Conservation Program, the Sand Control Programs for areas near Beijing and Tianjin, the Wildlife Protection and Nature Reserve Development Program, and the Fast-Growing and High-Yielding Timber Base Construction Program (Zhang et al., 2000, Liu and Diamond, 2005, Guo et al., 2019). APTZNC is the key implementation region for the large-scale ecological restoration projects.

To a certain extent, the effectiveness of China's ecological conservation program is controversial. On the one hand, successfully prevented land desertification, increased biomass and restoration of ecosystem functions (Zhang et al., 2000, Wu et al., 2013, Tian et al., 2014). On the contrary, some scholars believe that the results of ecological conservation projects in dryland may be exaggerated, because a large number of afforestation has not been tailored to local environmental conditions, for instance, overemphasis on tree and shrub planting caused increase in evaporation, forming a dry layer in soil, affecting regional hydrological processes, thereby decreasing vegetation growth (Jiang, 2005, Wang et al., 2010, Cao et al., 2011, Ye et al., 2018). So, before assessing the effectiveness of the human positive interventions, it is necessary to distinguish between climate-driven and human-induced vegetation dynamics in APTZNC. Besides, increasing vegetation activities (e.g. NPP or fractional vegetation coverage) is one of the goals of the implementation of ecological conservation programs, and is also the direct criterion for judging the success of those projects (Xu, 2019). Thus, vegetation dynamics are also supposed to be monitored and detected in APTZNC.

Although numerous researches have been carried out to elucidate climate-oriented and human-oriented vegetation dynamics, studies on quantifying the relative contributions of the two factors were rare (Sun et al., 2015, Zheng et al., 2019). To date, several efforts have attempted to separately quantify the influence of climatic and anthropogenic factors on an ecosystem within a specific region (Zhang, 2006, Mahmoud and Gan, 2018, Wang et al., 2019). The challenges to separate out them hinge on (i) both vegetation and precipitation in APTZNC typically have large inter-annual variations (Buyantuyev and Wu, 2009, Li et al., 2012); (ii) the indicators adopted to separate the contributions are various but uneven; (iii) well-established procedures for analysis are still lacking (Li et al., 2012, Gu et al., 2019). There are three main methods to address the relative contribution: (i) the regression model-based method; (ii) the residual trend-based method; (iii) the biophysical model-based method. First, the regression model-based evaluation method uses mathematical regression statistical methods to establish the relationship between vegetation change process and climate or human factors (Mueller et al., 2014). The operation is relatively simple and the data is relatively easy to obtain. But it is hard to depict the complex nonlinear processes of interaction between vegetation and the atmosphere, between that and human activities (Turner and Carpenter 2017). Second, the residual trend method is to indirectly estimate the impact of human activities by simulating the difference between vegetation changes without human disturbance and actual changes with human disturbance (Evans and Geerken, 2004, Jiang et al., 2017). The drawback of method is that model calibration in the year assuming without human disturbances would introduce errors into the model itself. Third, the biophysical model-based method is a new way to study the driving mechanism of vegetation change (Zhang et al., 2018a, Zhang et al., 2018b). For example, Terrestrial Ecosystem Model (TEM) (Raich et al., 1991, Melillo et al., 1993), Integrated Biosphere Simulator Dynamic Global Vegetation Model (IBIS-DGVM) (Foley et al., 1996) and Carnegie-Ames-Stanford Approach (CASA) (Field et al., 1995, Potter et al., 1996) were adopted to estimated potential NPP and actual NPP, then to determine the relative contribution. Although process-based models can accurately reflect vegetation dynamics and ecological processes, a large number of vegetation physiological and ecological parameters need to be measured and increase the uncertainty of the model (Jiang, 2019). In our work, the threshold segmentation method was adopted to conduct the evaluation the relative contribution of climate change and ecological programs. It is a relatively new method based on the statistical correlation analysis, and has been applied to detect vegetation activities in dryland areas (Tian et al., 2015). Multiple climate factors can serve as the assessment index, and simpler significant correlation can serve as the criteria to separate the coupled effects.

Although NDVI can represent vegetation productivity in different ecosystems (Wessels et al., 2004, Piao et al., 2005), some studies demonstrated that NDVI was only a proxy for the productivity of single land type because different vegetation types have various relationships between NDVI and NPP (Boelman et al., 2003, Santin-Janin et al., 2009, Zheng et al., 2019). This may weaken the proxy ability of NDVI to vegetation status. Net primary production (NPP) indicates as the net amount of solar energy converted to chemical energy through photosynthesis (Rosenzweig, 1968, Liu et al., 2019). It is the basis for ecosystem vitality, functions and processes, and a fundamental parameter for earth system science and global climate change research. Moreover, NPP is more sensitive to climate variation and especially vulnerable to human activities, which can accurately reflect the responses of vegetation growth to climate and human disturbances (Potter et al., 1999, Liu et al., 2019). So NPP served as an ideal candidate for characterizing vegetation activities in our work.

Our placed-based study aims to quantitatively understand and distinguish driving forces of vegetation significant restoration and degradation from 2000 to 2015 at 1 km spatial resolution by using APTZNC as a case research area. Thus, our study was designed to explicitly focus on the following questions: (1) did vegetation productivity in APTZNC improve or deteriorate in the recent 16 years? (2) what were the main drivers for vegetation dynamics with significant change—climate variations or human activities? (3) was the threshold segmentation method effective in addressing the above two questions? Hopefully, knowledge of the variation characteristics and patterns of the vegetation will promote regional sustainability and maintain ecological construction efforts under both the impacts of climate change and human activities.