3.1. Changes in the Ecosystem Pattern in the National Barrier Zone
The largest proportion of the ecological system area in the NBZ is the grassland ecosystem and the forest ecosystem, accounting for approximately 39% and 27%, respectively. The grassland ecosystem is mainly distributed in the ecological barrier of the Qinghai–Tibet plateau, covering an area of 647,616 km2 and accounting for 52.79% of the total grassland ecosystem. The forest ecosystem is mainly distributed in the northeast forest shelter and the southern hilly and mountainous shelter, with areas of 388,032 and 191,616 km2, respectively. The river ecosystem is the most widely distributed in the ecological barrier of the Qinghai–Tibet plateau, accounting for 54% of the total river ecosystem. The farmland ecosystems are mainly distributed in the northern sand-prevention shelter and the ecological barrier of the Sichuan–Yunnan–Loess plateau, accounting for 31.2% and 25.4% of the total farmland ecosystem, respectively.
Between 2000 and 2015, the ecosystem of the NBZ changed significantly. The increase of grassland ecosystem is the most obvious, changing from 1,212,928 km2 in 2000 to 1,226,880 km2 in 2015, with an increase of 13,952 km2, accounting for 0.45%. The urban ecosystem increased by 6720 km2, increasing the ratio by 0.21%, so the urbanization process was accelerated. The farmland ecosystem and river ecosystem increased slightly, and the respective proportions of the increased area were 0.068% and 0.076%. The desert ecosystem decreased significantly; the area decreased by 4544 km2, and the area proportion decreased by 0.15%, mainly in the northeast forest shelter. The forest ecosystem remained basically unchanged, with a change range of less than 0.01%.
The NBZ ecosystem distribution from 2000 to 2015 is shown in Figure 2
. There are significant differences in the changes in the ecosystem in different areas of the NBZ, as shown in Figure 3
. The forest ecosystem increased within the ecological barrier of the Sichuan–Yunnan–Loess plateau, with an area of 768 km2
, accounting for 0.19%, while the area within the northeast forest shelter decreased by 576 km2
, accounting for 0.08%. The grassland ecosystem increased by 1280 km2
in the southern hilly and mountainous shelter, with an area increase of 0.45%, while the area in the northern sand-prevention shelter decreased by 5248 km2
, with an area decrease of 0.6%. The ecological barrier of the Qinghai–Tibet plateau decreased by 1792 km2
, with an area decrease of 0.21%. The farmland ecosystem in the ecological barrier of the Sichuan–Yunnan–Loess plateau decreased by 3648 km2
, accounting for 0.1%; the area of the southern hilly and mountainous shelter decreased by 1280 km2
, accounting for 1.66%. The area of the farmland ecosystem in the northern sand-prevention shelter increased by 6016 km2
, accounting for 0.7%. The urban ecosystem in the ecological barrier of the Sichuan–Yunnan–Loess plateau increased by 1536 km2
, accounting for 0.38%, mainly in the middle and north of the region. In the northern sand-prevention shelter, the area increased by 3712 km2
, accounting for 0.43%. The desert ecosystem decreased by 3904 km2
and 0.45% in the northern sand-prevention shelter, and decreased by 768 km2
and 0.12% in the northeast forest shelter. The change in the river ecosystem was the most obvious in the ecological barrier of the Qinghai–Tibet plateau, where the area increased by 1600 km2
, accounting for 0.17%, and in the ecological barrier of the Sichuan–Yunnan–Loess plateau, where the area increased by 576 km2
, accounting for 0.14%.
The transition matrix analysis method was used to explore the transition relationship between ecosystems, as shown in Figure 4
. The ecological environment in the NBZ was improved, with 1.54% of farmland ecosystems and 0.65% of desert ecosystems transferred to forest and grassland ecosystems, respectively. Thanks to the implementation of the policy of Grain for Green Project policy in this region, 0.66% and 1.32% of the farmland ecosystem in the ecological barrier of the Sichuan–Yunnan–Loess plateau were converted into forest ecosystem and grassland ecosystem, respectively. The encroach phenomenon in the northern sand-prevention shelter is obvious; that is, there is mutual flow between the desert ecosystem and other ecosystems. From 2000 to 2015, 0.81% of desert ecosystems was transferred to grassland ecosystems, while 0.58% and 0.29% of grassland ecosystems and farmland ecosystems were transferred to desert ecosystems, respectively. In addition, 1.17% of the farmland ecosystems was transferred to grassland ecosystems in the northern sand-prevention shelter to prevent and control desertification, which has had a significant effect. However, the desertification problem still cannot be ignored because of arid and semiarid climate conditions and long-term unsustainable land use (e.g., overgrazing, extensive rainfed herbaceous crops, woody crops, irrigated crops linked to groundwater exploitation) [38
]. The distribution of river ecosystems in the barrier area is scattered, and what deserves high attention is that 1.32% of river ecosystems was transferred into desert ecosystems, especially in the northern sand-prevention shelter, accounting for 2.25%; therefore, water resources urgently need to be improved. It should also be noted that 1.12% and 1.45% of the farmland ecosystems in the northern sand-prevention shelter and the ecological barrier of the Sichuan–Yunnan–Loess plateau have been separately transferred to urban ecosystems, increasing urbanization and obvious expansion of the building land.
3.2. Driving Mechanism of Changes in the Ecosystem in the National Barrier Zone
The complex driving mechanism of changes in the ecosystem is composed of natural and socioeconomic factors. Among them, natural conditions such as climate and water resources are the constraints and conditions caused by changes in the ecosystem, and human activities are the main reason affecting changes in the ecosystem [39
In this paper, temperature and precipitation, which are the representative factors of the natural environment, are selected, and the drought index and water yield, calculated by the combining temperature and precipitation models are taken as supplements. The influence of natural factors on the changes in the ecosystem in the NBZ is quantitatively analyzed based on three factors: vegetation coverage, net primary productivity, and the leaf area index. The social and economic factors that affect ecosystem change can be divided into direct factors and indirect factors. The former includes population change, technological development, economic growth, and political and economic policies. The latter includes urbanization degree, land use intensification degree, land ownership, and land input [40
]. Due to the special geographical location of the NBZ, the data are difficult to obtain, and the political and economic policy, land ownership, and other indicators are difficult to quantify; therefore, we ultimately selected the population and GDP to quantitatively study the impact of social and economic factors on changes in the ecosystem. Based on the above nine factors, this study explores the driving mechanism of changes in the ecosystem in the NBZ.
3.2.1. Analysis of Driving Factors of Changes in the Ecosystem in the National Barrier Zone
Analysis of natural factors (Figure 5
): There are significant differences between the regions. Between 2000 and 2015, precipitation in the NBZ increased slightly, from 540.13 mm in 2000 to 688.92 mm in 2015, with an annual average precipitation of 570.86 mm in 16 years. The average annual temperature of the NBZ varies little, and the annual average temperature of 16 years is 3.68 °C. The average temperature of 18.1 °C in the southern hilly and mountainous shelter is the highest, and the average temperature of the ecological barrier of the Qinghai-Tibet Plateau is −2.84 °C. In 16 years, the normalized vegetation index (NDVI), vegetation net primary productivity (NPP), and leaf area index (LAI) all showed a fluctuating growth trend: the NDVI increased from 0.343 to 0.352, the NPP increased from 360.74 to 394.03 gmm−2
, and the LAI increased from 1.22 to 1.27. The southern hilly and mountainous shelter has the highest mean value, while the northern sand-prevention shelter has the lowest mean value. The change of drought index is complicated, and the occurrence of drought has some potential relation with the regional temperature and precipitation. The drought index (PDSI) showed an increasing trend in the northern sand-prevention shelter, reaching 4.44 in 2015. The water yield (WY) in the NBZ showed a trend of fluctuating growth, from 64.24 mm in 2000 to 187.5 mm in 2015. The interannual fluctuation is obvious in some regions: the minimum value of the annual water yield in the southern hilly and mountainous shelter was −111.69 mm in 2011, and the maximum value was 722.52 mm in 2015.
Socioeconomic factors (Figure 5
): Between 2000 and 2015, the GDP of the NBZ increased significantly, from 243,200 yuan/km2
in 2000 to 2,072,300 yuan/km2
in 2015, with a growth rate of 7.52/a. The population in the NBZ shows an increasing trend, with a growth rate of 0.005/a, and the population distribution in the area is extremely uneven, which affects not only the uneven distribution of the farmland and urban ecosystems but also some driving forces, such as the GDP. The southeastern part of the ecological barrier of the Sichuan–Yunnan–Loess plateau is the most concentrated area with the highest value at 24,926.1 people/km2
and with a value of 0 in the no-man zone.
3.2.2. Analysis of the Driving Factor Contribution Rates and Correlations
The contribution rates of the driving factors for changes in the ecosystem in the NBZ based on redundancy analysis are shown in Table 1
. Combined with the results of the sequence of driving factors in Figure 6
, the comprehensive redundancy analysis shows that from 2000 to 2015, the changes in the ecosystem in the NBZ were mainly driven by social and economic factors, with the contribution rates of the GDP and population (POP) being 0.75 and 0.65, respectively.
With the increase of GDP in the NBZ, the investment in ecological protection projects was further increased. There is a close correlation between the urban ecosystem, river ecosystem, and cultivated land ecosystem and various driving factors of the NBZ. Comparing the ecological barrier of the Qinghai–Tibet plateau, the northeast forest shelter, and the southern hilly and mountainous shelter, the influence intensity of the temperature is higher than the precipitation in north China, while the influence intensity of precipitation is higher than the temperature in south China.
The contributions of driving factors in different regions are also significantly different. The contributions of NDVI, POP, GDP, and temperature (TEMP) in the northern sand-prevention shelter are significant, and the total contribution rate is 99.9%; the contribution rate of the GDP (0.79) is the largest, followed by POP (0.66). The correlation between the GDP and POP and the farmland ecosystem and urban ecosystem is very high, indicating that the driving mechanism of ecosystem changes in this region is relatively simple, climate change is relatively regular, and the influence of human activities on the change of ecosystem pattern in this region is dominant and stabilized. The largest impact on the changes in the ecosystem in the northeast forest belt is GDP (0.48), the smallest is NPP (0.02), and the contribution rate of all driving factors is greater than 0.01, forming a driving mechanism together. The correlation between the farmland ecosystem and urban ecosystem and all driving factors is high in this region, while the correlation between the desert ecosystem and all driving factors is low; the precipitation (PREC) (0.33) has an impact on all driving factors, indicating that precipitation in this region varies greatly and that drought has a serious impact on vegetation growth. The GDP (0.78) still has the greatest impact on the changes in the ecosystem in the southern hilly and mountainous shelter. Additionally, the contribution rates of the NPP, NDVI, and LAI are very high, which is directly related to the large distribution of the forest ecosystem in this region. The grassland ecosystem and urban ecosystem are closely related to all driving factors, and the social and economic factors (POP, GDP) have a high influence on the river ecosystem. It is worth noting that TEMP has an impact on all driving factors in this region, but that PREC has a low impact, which indicates that the precipitation variation range in this region is small and that water can guarantee the growth of vegetation especially under drought conditions. POP (0.63) and NDVI (0.32) contributed the most to changes in the ecosystem of the ecological barrier of the Sichuan–Yunnan–Loess plateau. Due to the implementation of policies, such as the Grain for Green Project, the river ecosystem, grassland ecosystem, and forest ecosystem in the region are highly correlated with the POP, NDVI, and NPP, while the relationship between the farmland ecosystem and each factor is low. The ecological barrier of the Qinghai–Tibet plateau is the region where human activities have the greatest impact on changes in the ecosystem. The respective contribution rates of the GDP and POP are 0.83 and 0.72, and the contribution rates of the PPEC, NPP, and other factors to the driving mechanism in this region are less than 0.05. The influence of socioeconomic factors on the forest ecosystem and river ecosystem is especially large, but it has little effect on the grassland ecosystem.
3.2.3. ANOVA of the Driving Factors Contribution Rate
According to the results in Table 1
, it can be seen that the sum of the contribution rates of each driving factor obtained after redundancy analysis is much higher than 1. The reason is that the factor contribution rates of redundancy analysis are not independent and consist of two parts: the independent contribution rates of factors and the interaction between the factors and other factors. Therefore, it is necessary to further use ANOVA to decompose the results of the redundancy analysis, isolate the most prominent factors to make them independent, and further quantitatively analyze the driving mechanism of the changes in the ecosystem in the NBZ.
The significance of the correlation between each driving factor and the distribution of the ecosystem is calculated by using the Vegan pack in the R language, and the driving factor with the highest significance is the typical factor for variance analysis, as shown in Table 2
. Only the GDP and POP are typical factors in the ecological barrier of the Sichuan–Yunnan–Loess plateau. The typical factors of the northern sand-prevention shelter are the GDP, POP, and TEMP. The typical factors of the northeast forest shelter are the GDP, POP, PREC, and PDSI. There are many typical factors in the southern hilly and mountainous shelter, including the GDP, WY, LAI, NDVI, NPP, and PDSI, respectively. In the analysis of variance, we chose the four factors with the highest significance, including the GDP, WY, LAI, and NDVI. The typical factors of the ecological barrier of the Qinghai–Tibet plateau are GDP, POP, TEMP, and PDSI. The typical factor of the overall NBZ is only GDP, so we choose the GDP, POP, TEMP, and PREC as the typical factors in this paper.
shows the results of the variance analysis: The POP of the ecological barrier of the Sichuan–Yunnan–Loess plateau is the most typical driving factor, with an independent contribution rate of 0.49. The proportion of GDP and POP in the northern sand-prevention shelter is 0.44, while the proportion of TEMP is only 0.02. The proportion of social and economic factors in the northeast forest shelter is 0.73, and the interaction of PREC and GDP is 0.22. The ratio of GDP in the southern hilly and mountainous shelter is 0.28, and the interaction of GDP and NDVI is 0.25. The impact of socioeconomic factors on the ecological barrier of the Qinghai–Tibet plateau is very obvious, with the combined effect of GDP and POP reaching 0.77. The overall NBZ is relatively balanced, where the influence of social and economic factors is greater than natural factors. At the same time, it can be seen that the contribution rate of each factor as the driving factor is relatively low, such as the PDSI, which did not even exceed 0.01. However, the joint action of each factor after the interaction is important, which obviously changes the ecosystem.