Dynamic change of habitat quality and its key driving factors in Ningxia Hui Autonomous Region, China

Ding Wang; Haiguang Hao; Hao Liu; Lihui Sun; Yuyang Li

doi:10.3897/natureconservation.53.102810

Abstract

Habitat quality reflects the level of biodiversity, and habitat maintenance functions are related to human well-being and ecosystem stability. Ningxia Hui Autonomous Region is a typical ecologically fragile region in Western China with complex human-nature relationships. Maintaining good habitat is not only a fundamental requirement for biodiversity conservation but also a necessary path for sustainable regional development. In this study, we assessed and analysed the spatial and temporal patterns and changes in habitat quality in Ningxia from 2000 to 2020, and explored the driving factors of habitat quality using a geographically weighted regression (GWR) model. The results indicated: (1) The overall habitat quality level in Ningxia was low to intermediate, with an upwards and then downwards trend during the past 20 years, showing a small change in overall magnitude. (2) The high- and higher-level habitat quality patches in Ningxia were mainly distributed in areas with high vegetation cover, such as the Helan Mountain and Liupan Mountain. The patches of moderate-level habitat quality mainly included cultivated land, while the low- and lower-level patches were mainly distributed in areas subjected to more frequent human activities, such as cultivated land and construction land. (3) The spatial and temporal distribution patterns and changes in habitat quality in Ningxia from 2000 to 2020 were mainly influenced by fractional vegetation cover (FVC), soil moisture content (SMC), proportion of construction land area (PCL), and proportion of cultivated land area (CLP). Among them, FVC and SMC were positive driving factors, and PCL and CLP were negative driving factors. The results support that increasing vegetation cover and reducing anthropogenic disturbance to natural habitats are important measures to maintain fragile habitats and that key ecological function areas such as nature reserves are crucial for habitat quality protection in ecologically fragile areas.

Key words

Driving factors, geographically weighted regression (GWR), habitat quality, Ningxia Hui Autonomous Region, spatiotemporal pattern

Introduction

Habitat quality refers to the ability of an ecological environment to provide suitable conditions for the sustainable survival and development of individuals, populations or communities, reflecting the richness of biodiversity in a region, and it is related to human well-being (Yohannes et al. 2021). The concept of habitat was first introduced by Grinnel in 1917 (Riedler and Lang 2018), and the study of habitat quality can be traced back to the 1960s (Goertz 1964; Rosenzweig and Winakur 1969; Janzen 1970) which fully developed by field surveys (Van Horne 1983; Congdon 1974), habitat indices (Berger and Hodge 1998), model simulations (Dunning et al. 1992; Roth et al. 1996) and other stages. Currently, since the introduction of the United Nations Sustainable Development Goals (SDGs) and the convening of the Conference of the Parties to the United Nations Convention on Biological Diversity (UN CBD-COP), governments and researchers have agreed to stabilize and enhance the global biodiversity level (Hale and Swearer 2016).

A favourable habitat condition means that various ecological factors in the ecosystem meet the needs of population survival and reproduction, and the orderly differentiation of biological ecological niches will achieve a balanced and stable ecosystem function. Additionally, habitat maintenance is an ecosystem service that is of great concern to humans (Celina et al. 2022), and habitat quality is a comprehensive representation of habitat maintenance capacity, which fully connects ecological processes with human needs. Habitat quality is deeply related to regional biodiversity levels, ecosystem service trade-offs and synergies, and ecological security patterns (Wu et al. 2013; Wang et al. 2022); habitat quality determines the balance of ecosystems, laying the foundation for sustainable development prospects of social-economic-natural complex systems (Wu et al. 2017). Managing and maintaining habitat functions and improving habitat quality can effectively maintain biodiversity and provide a good base of ecosystem services that ultimately meet the needs of the human economy and society (Peggy et al. 2021). Ecologically fragile areas are located in the cross-transition zone of different types of ecosystems, with weak system resistance to disturbance, sensitivity to global climate change, and generally low habitat quality as well as biodiversity levels (Prasad and Ramesh 2019). For ecologically fragile areas with poor ecological backgrounds, measures such as curbing urban expansion and maintaining landscape integrity can enhance regional habitat quality, thereby improving and enhancing ecosystem structure and function, which can promote a continuous supply of ecosystem services to human society (Ramachandra et al. 2019).

Habitat quality assessments include both ecological and geographic perspectives. Early studies focused on the substantial impacts of human activities on plant and animal habitats, and the research methods and contents were more biased towards natural and ecological properties (Dallimer et al. 2012). Such studies usually used relevant parameters obtained by field survey methods to construct indicator systems; additionally, hierarchical analysis, grey correlation models, and entropy weighting methods were used to synthesize habitat conditions, and the natural or anthropogenic driving factors of habitats of single or homogeneous species were analysed. These studies were mostly conducted on a small scale, such as small cities, rivers, and nature reserves, and usually used the sample strip method or sample method of investigation to obtain various parameters related to the quality of plant and animal habitats; moreover, these studies commonly used indicators including species richness, vegetation types, topographic indicators, and water quality (Harper and Everard 1998). The advantage of this method is that it can reflect the habits and habitat conditions of the evaluated objects more comprehensively, and the indicators are more sensitive and detailed. However, due to the high time and labour costs and the difficulty of obtaining data over long-time spans for dynamic analysis, such methods cannot be applied to habitat quality studies at larger spatial and temporal scales (Wang et al. 2017).

As land use/land cover (LULC) change has become the focus of global change research (van Vliet et al. 2015), habitat quality assessment models developed by remote sensing and GIS technologies have been widely used in practice, forming a paradigm for habitat quality research based on a geographic perspective (Romero-Calcerrada and Luque 2006). In recent years, many scholars have performed many studies on habitat quality at large-scale scales, such as across regions, from the perspectives of ecological service function, ecological risk and early warning and the relationship between urbanization development and ecosystem conservation (Nagendra et al. 2013; Chen et al. 2016; Gomes et al. 2021; Duan and Yu 2022). Commonly used mature habitat quality assessment models include Integrated Valuation of Ecosystem Services and Tradeoffs (InVEST) (Bao et al. 2015), Habitat Suitability Index Model (HSI) (Wang et al. 2009), Maximum Entropy Model (MaxEnt) (Radosavljevic and Anderson 2014), Artificial Intelligence for Ecosystem Services (ARIES) (Villa et al. 2009), and Multiscale Integrated Models of Ecosystem Services (MIMES) (Roelof et al. 2015), among others. Among them, the Habitat Quality module in the InVEST model is widely used (Kareiva et al. 2011) because of its low application cost and high assessment accuracy, and the evaluation results support the spatial visual representation of regional habitat distribution as well as habitat degradation.

Currently, humans are facing a serious biodiversity crisis, and habitat destruction is one of the most serious challenges threatening biodiversity conservation (Crooks et al. 2017). Population growth, economic development, and continuous changes in land use and its structure profoundly affect the material and energy flow circulation processes between habitat patches, which in turn change the distribution patterns and functions of regional habitats (Haddad N et al. 2015). The level of habitat quality is highly dependent on natural conditions and on its proximity to human land uses and the intensity of land use (Sobhani et al. 2022). Therefore, exploring the relationships among ecological factors, ecological processes and habitat quality is important for maintaining regional biodiversity and even ecosystem functions and provides a basis for improving regional ecological security patterns and the sustainable development of land resources (Kalacska et al. 2017; Hao et al. 2019).

The current methods used to investigate the factors affecting habitat quality mainly include spatial exploratory analysis, spatial econometric analysis, multiple regression analysis, grey correlation analysis, and Moran’s I spatial autocorrelation index (Zhu and Alimujiang 2020; Moëzzi et al. 2022; Raimundo Lopes et al. 2022); moreover, these methods focus on the spatial and temporal patterns, multiscale change characteristics, and evolutionary mechanisms of habitat quality (Wu et al. 2021; Xiao et al. 2022). However, most of the studies have been conducted at the scale of administrative units, usually using methods such as indicator methods and spatial autocorrelation analysis, and they have failed to fully consider the heterogeneity of different geographic spaces and the spatial scales at which different influencing factors act (Chisholm et al. 2011; Smith et al. 2011). Based on Tobler’s (1970) first law of geography, Brunsdon proposed the geographically weighted regression (GWR) model (Brunsdon et al. 1996), which allows for spatial heterogeneity in the coefficients of independent variables and can effectively detect the spatial nonstationarity characteristics of regression variables; this method is widely used in geography, economics, ecology and the environment (Zhu et al. 2020; Hu et al. 2022). The GWR model performs distance-weighted regression with the help of observations from neighbouring sample points, which can reveal the quantitative relationships between factors and impact factors more accurately, thus improving the goodness of fit of the model and solving the deficiency of using exploratory spatial data analysis (ESDA) tools that can solve only time-sectional data (Qin 2007).

In summary, this study selected the Ningxia Hui Autonomous Region (hereinafter referred to as Ningxia), a typical ecologically fragile region in Western China, as the study area, collected raw data on land use, NDVI, and temperature from 2000 to 2020, and analysed the spatial and temporal patterns of habitat quality and their changes in the past 20 years based on remote sensing and GIS analysis. Based on the objective fact that the distribution of habitat quality in Ningxia is spatially heterogeneous, the GWR model with optimal fitting parameters was finally used to investigate the key factors driving the distribution and changes in habitat quality in Ningxia.

Materials and methods

Study area

Ningxia (35°14'–39°23'N, 104°17'–107°39'E) (Fig. 1) is located in Western China in the middle and upper reaches of the Yellow. The climate is temperate continental, with an average annual temperature of 6–10 °C, an average annual precipitation of approximately 220 mm, and more than 3,200 h of sunshine. The topography is high in the south and low in the north, with an altitude of 1,100–1,200 m. The ecological geography is divided into 3 parts. The Yellow River flows through the northern irrigation area, which has a gentle topography and superior soil and water conditions. Тhe central area is the arid wind-sand belt, which is subject to perennial drought and poor soil and water conditions; and the southern mountainous area is full of ravines and gullies, with complex topography and a cold and wet climate. Ningxia has a well-developed agriculture and animal husbandry industry, but water resources are concentrated with a small and uneven spatial distribution. Ningxia is located in the interlocking agricultural and pastoral areas of northern China, with a fragile ecological environment and simple species composition and ecosystem diversity. Due to its location in the transition zone of the arid and semiarid climate zones, Ningxia has become an important ecological security barrier in Western China. Тhere was still 15,534.84 km² of soil erosion in the region, accounting for 23.40% of the total area of the region, and the contradiction between ecological and environmental problems and economic and social development was still relatively prominent.

Figure 1.

Topography, natural reserves and administrative division of Ningxia.

Research methods

Habitat quality assessment methods

In this study, the habitat quality module of the InVEST model was used to assess habitat quality in Ningxia, and the habitat quality index was calculated as follows:

$Q_{x j} = H_{j} [1 - (\frac{D_{x j}^{Z}}{D_{x j}^{Z} + k^{z}})]$ (2.1.1)

where Q_xj is the habitat quality of raster x in land use type j; k is the half-saturation parameter, whose value is half of the resolution of the raster data in the study area and is generally 1/2 of the maximum value of habitat degradation; H_j is the habitat suitability of land use type j, whose value is usually 0~1; z is the normalization constant, which is usually set to 2.5; and D_xj is the level of stress to which raster x of land use type j is subjected, i.e., the degree of habitat degradation. The degree of habitat degradation is the intensity of habitat disturbance by threat sources and is calculated as follows:

$D_{x j} = \sum_{r = 1}^{R} \sum_{y = 1}^{y_{r}} (\frac{ω_{r}}{\sum_{r = 1}^{R} ω_{r}}) r_{y} i_{r x y} β_{x} S_{j r}$ (2.1.2)

$i_{r x y} = 1 - (\frac{d_{x y}}{d_{r m a x}})$ (Linear decay) (2.1.3)

$i_{r x y} = \exp (\frac{- 2.99 d_{x y}}{d_{r m a x}})$ (Exponential decay) (2.1.4)

where D_xj is the degree of habitat degradation; R is the number of stressors; y is the number of grids in the raster layer of stressor r; y_r is the number of grids occupied by stressors; w_r is the stressor weight; r_y is the stressor value of raster y; β_x is the accessibility level of raster x, which is not considered in this study; s_jr is the sensitivity of habitat type j to stressor r; i_rxy is the stress factor value r_y of raster y on the stress level of habitat raster x; d_xy is the linear distance between raster x and raster y; and d_rmax is the maximum stress distance of threat source r. The higher the calculated score is, the greater the threat level caused by the threat factor to the habitat and the higher the degree of habitat degradation.

Based on the InVEST model manual and with reference to previous research results on habitat quality in Ningxia and the arid and semiarid regions of Northwest China (Wu et al. 2020; Bao 2022; Ren et al. 2022), this study used paddy fields, drylands, urban land, rural settlements and other construction land as threat factors and determined the habitat suitability of habitat types and the sensitivity of different habitat types to stress factors (Tables 1, 2).

Table 1.

Download as

CSV

XLSX

Ecological threat factors and their maximum impact distances and weights.

Threat Factor	Impact Distance/km	Weight	Spatial Decline Type
Paddy Field	4	0.15	Linear Decline
Dryland	3	0.2	Linear Decline
Urban Land	5	0.3	Exponential Decline
Rural Settlements	4	0.3	Exponential Decline
Other Construction Land	8	0.2	Linear Decline

Table 2.

Download as

CSV

XLSX

Habitat suitability and relative sensitivity to threat factors.

Type	Habitat suitability	Paddy field	Dryland	Rural settlement	Urban land	Other construction land
Paddy Field	0.6	0.3	0.2	0.35	0.5	0.45
Dryland	0.4	0.3	0.2	0.35	0.5	0.4
Forested Land	1	0.8	0.7	0.85	1	0.6
Shrubland	1	0.4	0.3	0.45	0.6	0.4
Sparse Woodland	1	0.85	0.75	0.9	1	0.65
Other Forest Land	1	0.9	0.8	0.95	1	0.7
High Coverage Grassland	0.85	0.4	0.3	0.45	0.6	0.6
Medium Coverage Grassland	0.8	0.45	0.35	0.5	0.65	0.7
Low Coverage Grassland	0.75	0.5	0.4	0.55	0.7	0.8
Canal	1	0.7	0.6	0.75	0.9	0.5
Lake	1	0.7	0.6	0.75	0.9	0.5
Reservoir Pit	1	0.7	0.6	0.75	0.9	0.5
Beach Land	0.6	0.75	0.65	0.75	0.95	0.55
Urban Land	0	0	0	0.8	0	0
Rural Settlements	0	0	0	0	0	0
Other Construction Land	0	0	0	0	0	0
Unused Land	0	0	0	0	0	0

The rate of change in habitat quality was calculated using the terminal habitat quality minus the initial habitat quality with the following equation:

$K_{T} = \frac{H Q_{i} - H Q_{0}}{H Q_{i}} \times 100 %$ (2.2)

where K_T is the rate of change in habitat quality over time T. This study had a 5-year cycle; HQ_o is the size of habitat quality at the beginning of the study, HQ_i is the size of habitat quality at the end of the study, and the raster resolution is 30 m.

Spatial autocorrelation analysis

In this study, the global Moran’s I index was used to describe whether habitat quality in the study area had a clustering effect on a global scale, and the local Moran’s I index was used to reflect the spatial autocorrelation of habitat quality in the subregion. The spatial autocorrelation analysis was performed in ArcGIS 10.7 software.

$G l o b a l M o r a n' s I = \frac{n \sum_{i = 1}^{n} \sum_{j = 1}^{n} ω_{i j} (x_{i} - \bar{x}) (x_{j} - \bar{x})}{\sum_{i = 1}^{n} {(x_{i} - \bar{x})}^{2} (\sum_{i} \sum_{j} ω_{i j})} (i \neq j)$ (2.3.1)

$l o c a l M o r a n' s I = \frac{n (x_{i} - \bar{x})}{\sum_{i = 1}^{n} {(x_{i} - \bar{x})}^{2}} \sum_{j = 1}^{n} ω_{i j} (x_{i} - \bar{x}) (i \neq j)$ (2.3.2)

where x_i and x_j are the values of variable x taken on neighbouring cells, x is the attribute value of the n location variables, x̄ is the mean of the attribute values of the spatial variables, ω_ij is the spatial weight matrix of raster i and raster j, and n is the total number of rasters.

Geographically weighted regression

Pearson correlation regression, the least squares model (OLS), and geographically weighted regression models were used to explore the characteristics of driving factors acting on habitat quality in Ningxia. The GWR model is a local regression model that embeds the geographic location of the data into the regression parameters, allowing for local parameter estimation. In this study, the geographically weighted regression weight function was chosen as a Gaussian function (Adaptive Gaussian), and its calibration was performed using an adaptive approach (Adaptive).

$y_{i} = β_{0} (μ_{i}, v_{i}) + \sum_{i = 1}^{k} β_{k} (μ_{i}, v_{i}) x_{i k} + ε_{i}$ (2.4)

where y_i is the dependent variable at sample point i, x_ik is the observed value of the kth variable at the ith point, (µ_i, v_i)is the location coordinate of the ith point, β_o(µ_i, v_i) is the intercept, β_k(µ_i, v_i) is the regression coefficient of the ith, and ε_i is the error term.

Based on the results of previous studies on the driving factors of habitat quality in the Loess Plateau and Western China and the actual characteristics of the ecological environment in Ningxia (Yang et al. 2021; Bai et al. 2020), 16 indicators of physical geographic and socioeconomic factors were selected to analyse the key driving factors and characteristics of their effects on the status and dynamic change of habitat quality in Ningxia, as shown in Table 3.

Table 3.

Download as

CSV

XLSX

Selection of regressors for habitat quality driving factors in Ningxia.

Indicators	Abbreviation	Unit
Net Primary Productivity	NPP	gC/(m²*a)
Fractional Vegetation Cover	FVC	%
Mean Annual Precipitation	MAP	mm
Drainage Density	DRA	km/km²
Elevation	ELE	m
Slope	SLP	°
Degree of Relief	DRF	m
Soil Moisture Content	SMC	m³/m³
Average Annual Temperature	AAT	°C
Proportion of Cultivated Land	PCL	%
The Proportion of Construction Land	CLP	%
Population Density	POP	person/km²
Road Network Density	RND	km/km²
Nighttime Light Index	NLI	DN
Regional GDP	GDP	10⁴ Yuan (¥) /km²
Closest Distance to Road Network	DRN	m

Data sources and processing

The data in this study included habitat quality assessment data and driving regression data, and the InVEST model habitat quality assessment mainly used five periods of land use dataset from 2000 to 2020. The dataset were obtained from the Resource and Environmental Science and Data Center of the Chinese Academy of Sciences (https://www.resdc.cn/) at a resolution of 30 m. NDVI data were obtained from the 30 m annual maximum NDVI dataset of China at the National Ecological Science Data Center of China (http://www.nesdc.org.cn/). Net primary productivity data were obtained from the MOD17A3HGF Version 6.0 product (https://www.earthdata.nasa.gov/). Geospatial information data included the 2021 version of 1:1 million public geographic basic information dataset (https://www.webmap.cn/commres.do?method=result100W) and ALOS 12.5 m DEM data (https://www.gscloud.cn/). GDP and Population data were obtained from the 1 km-grid GDP dataset of China (https://www.resdc.cn/) and the 1 km-grid population dataset of China (https://www.resdc.cn/). Soil water content data were obtained from the Soil Moisture in China dataset (2002–2018) (http://data.tpdc.ac.cn/zh-hans/). A 1-km monthly mean temperature dataset for China (1901–2021), a 1-km monthly precipitation dataset for China (1901–2022), and the Prolonged Artificial Nighttime-light Dataset of China (1984–2020) were obtained from the National Tibetan Plateau Scientific Data Center (http://www.tpdc.ac.cn/).

Тhe original spatial raster data of river network density, road density, distance to the nearest road, elevation, slope, topographic relief, GDP, population density, nighttime lighting index, proportion of construction land area, and proportion of cultivated land area for the whole Ningxia region in 2000, 2005, 2010, 2015 and 2020 were obtained by processing the above datasets. Based on the zoning of Ningxia and the accuracy of the data, a suitable 5 km × 5 km fishing grid was built, excluding the grid with null values, to obtain the final 2290 grids. The raw data were partitioned in tabular form to obtain the final data results for each impact factor as well as the raw data results for habitat quality for the period 2000–2020, with 17 categories and 85 datasets in five periods.

Results and analysis

Spatial and temporal patterns of habitat quality in Ningxia

The habitat quality of Ningxia was classified into five levels: low level (0–0.25), lower level (0.25–0.4), moderate level (0.4–0.6), higher level (0.6–0.75), and high level (0.75–1.0) (Fig. 2). The results showed that the proportion of high-level habitat quality patches in Ningxia in 2020 was the highest (29.27%), followed by lower-level (25.08%) and higher-level (22.62%) patches, while the proportions of low-level and moderate-level habitat quality patches were lower, at 13.41% and 9.61%, respectively. The proportion distribution of habitat quality patches in Ningxia in 2000, 2005, 2010 and 2015 was similar to that in 2020, which showed the distribution of high level > lower level > higher level > low level > moderate. From 2000 to 2020, the average habitat quality of the whole region of Ningxia was approximately 0.58, which was at a moderate-good level. Combined with the proportional distribution of high-level and higher-level habitat quality patches, the habitat maintenance function in Ningxia has been relatively healthy over the past 20 years.

Figure 2.

Proportion of habitat quality levels in Ningxia from 2000–2020.

In terms of spatial distribution (Fig. 3a–f), the high-level habitat quality patches in Ningxia from 2000 to 2020 were mainly distributed in nature reserves with high forest cover and excellent ecological conditions, such as in the south, Helan Mountain in the north, Shixia Gorge in the east and Luoshan in the centre. The higher-level habitat quality patches were scattered around the higher-level habitat quality patches. The moderate-level habitat quality patches mainly included most of the cultivated land types. The lower-level habitat quality patches contained some cultivated land and were more randomly distributed. The low-level habitat quality patches mainly included most of the construction land and unused land with harsh natural conditions, such as the central wind-sand region area, the sandpots in the west, and the foothills of the Helan Mountain and the edge of the Ningxia Plain, which are strongly disturbed by humans.

Figure 3.

Spatial distribution of habitat quality and nature reserves in Ningxia from 2000 to 2020.

The results of spatial autocorrelation analysis showed (Fig. 4) that the distribution of habitat quality in Ningxia from 2000 to 2020 was significantly autocorrelated geographically and spatially, with Moran indices of 0.619, 0.604, 0.588, 0.581, and 0.144 (P < 0.01), all with a 90%+ credibility level. The results of local spatial autocorrelation coefficients showed (Fig. 5a–e) that the habitat quality aggregation in Ningxia mainly included two types of high-high value aggregation and low-low value aggregation, while the high-low value aggregation and low-high value aggregation types had poor significance levels and more random distributions.

Figure 4.

Spatial autocorrelation global Moran’s I values of habitat quality in Ningxia from 2000–2020.

Figure 5.

The results of spatial autocorrelation analysis of habitat quality in Ningxia from 2000 to 2020.

Combining the distribution of patches with the different habitat quality levels in Ningxia, we found that habitat quality in Ningxia was closely related to patch type and was influenced by both natural conditions and human activities. Our study found that habitat quality levels were highest in primary forest reserves that were not disturbed by human activities, where precipitation, temperature, topography and elevation were suitable for the survival and reproduction of organisms. In contrast, habitat quality was significantly lowest in construction sites strongly disturbed by human activities, where the climate is arid, vegetation is sparse, and land use changes are frequent, i.e., they lacked the basic conditions needed to meet biological survival. In addition, although cultivated land is used as an artificial landscape, it possesses a moderate level of habitat maintenance function, and the habitat quality was generally categorised at the moderate level. Nature reserves concentrate the most fully functional ecosystems, which are crucial for protecting habitat quality and improving biodiversity levels. By identifying key areas and delineating priority protection areas, it will further contribute to the stability and improvement of regional biodiversity levels.

Dynamic changes of habitat quality in Ningxia

In terms of different levels of habitat quality (Fig. 6a), the proportion of high-level habitat quality patches in Ningxia increased by 1.26% from 2000 to 2020, while the low-level habitat quality patches also increased by 1.53%. The proportions of lower-level, moderate-level and higher-level habitat quality patch areas decreased by 1.29%, 0.62% and 0.87%, respectively. From different periods (Fig. 6a, b), the area of low-level habitat quality patches in Ningxia from 2000 to 2020 increased, the area of moderate-level habitat quality patches decreased, and the area of lower-level habitat quality patches was more volatile. While the area of higher-level habitat quality patches decreased and then increased, the high-level patches showed a trend of increasing and then decreasing. From 2000 to 2020, the average habitat quality in Ningxia was maintained at a moderate level with small changes (<0.005) (Fig. 6b). On the one hand, ecological protection measures such as returning farmland to forests and grasses and ecological restoration have increased the habitat quality of some patches, but factors such as population growth and construction land expansion have caused the habitat quality of some patches to decrease, resulting in the contradiction between human activities and natural habitats remaining very prominent.

Figure 6.

Dynamic changes in the area proportion of patches with different levels of habitat quality in different periods in Ningxia from 2000 to 2020.

The spatial variation in habitat quality in Ningxia was divided into five classes: significantly decreasing (-1– -0.5), slightly decreasing (-0.5–-0.25), remaining stable (-0.25–0.25), slightly increasing (0.25–0.5), and significantly increasing (0.5–1) (Fig. 7a–f). In terms of spatial variation, habitat quality decreased more (-1–0.25) in areas such as Ningxia with a high density of construction land and some scattered cultivated land from 2000 to 2020. In contrast, significant increases in habitat quality were more concentrated in the central area where the cultivated land was returned to forest and grass (0.25–1). Habitat quality was stable in most areas of Ningxia over the 20-year period, with small changes (-0.25–0.25). In different periods (Figs 6a, b, 7a–f), the patches with decreasing habitat quality in Ningxia basically decreased continuously during the 20 years from 2000 to 2020. Most of the patches with habitat quality changes from 2000–2005 showed increases. From 2005 to 2010, habitat quality decreases were dominant. From 2010–2015, the decreasing trend remained severe. The decreasing trend improved only during 2015–2020.

Figure 7.

Spatial distribution of habitat quality and land use change in Ningxia in different periods from 2000 to 2020.

The analysis of habitat quality in Ningxia showed (Fig. 7a–f) that the overall habitat quality of construction land and unused land was decreasing over the past 20 years, and these areas were severely disturbed by humans or had poor natural conditions. The patches with improved habitat quality were more randomly distributed, mostly in areas returned to forest and grass. By comparing the results with the land use change, we found that the changes from other types of patches to forest, grassland and water improved habitat quality, while the change from other types of patches to construction land, cultivated land and unused land obviously forced the habitat quality to decrease. When the habitat quality remained stable, the land use type remained the same over 20 years or the habitat quality remained the same after the transformation. Human activities not only cause damage to biological habitats, but also take proactive measures to control urban expansion and arable land development. The application of reasonable ecological restoration technologies and projects can promote habitat improvement. In this process, natural reserves, especially those designated for biodiversity, provide a stable and complete habitable environment, which has received significant attention and protection from humans, Integrating ecosystem functions and human needs.

Analysis of the driving factors of habitat quality in Ningxia

Analysis of the applicability of the driving factor regression model

First, the analysis of 16 driving factors using Pearson’s method found that the R² values for 2000–2020 (Fig. 8) were 0.327, 0.312, 0.325, 0.325, and 0.121, and the adjusted R² values were 0.322, 0.307, 0.320, 0.320, and 0.114, respectively (P < 0.05). From the sum of relative coefficients over the 20-year period (Fig. 9), factors such as AAT (-1.562), PCL (-1.302), and CLP (-0.895) had a strong negative effect on habitat quality in Ningxia, and factors such as RFI (2.158), SLP (2.142), ELE (1.604), and NPP (0.915) had a strong positive effect. Since the correlation analysis could not determine the covariance between the factors, OLS linear regression and GWR models were further adopted to explore the role of the 16 factors on habitat quality.

Figure 8.

Pearson correlation significance results for each driving factor from 2000–2020.

Figure 9.

Relative sum of coefficients of Pearson correlation analysis of driving factors from 2000–2020.

A comparative analysis (Table 4) revealed that the 2000–2020 GWR model AIC and AICc values were -4,365.10, -4,436.45, -4,302.96, -4,425.03, and -2,263.94 and -3,060.87, -2,542.82, -2,402.95, -2,525.42, and -1,769.53, respectively, which were significantly smaller than the OLS model AIC and AICc values. The R² and adjusted R² of the GWR model for 2000–2020 were 0.735, 0.779, 0.762, 0.777, and 0.202 and 0.692, 0.721, 0.700, 0.719, and 0.158, respectively, with a higher fitting effect and higher accuracy than the OLS model (0.70). In addition, OLS regressions of K(BP)-Prob and JB-Prob were significant (P < 0.01) in terms of regression coefficients and significance levels, indicating that the OLS regressions were robust and redundant in terms of the effect of each factor on habitat quality. The relative sum of OLS regression coefficients over the 20-year period indicated that habitat quality in Ningxia was strongly driven negatively by PCL (-2.418), CLP (-0.485), and SLP (-0.083) and positively by SMC (1.215), NPP (1.109), and FVC (1.004). However, OLS regression is not applicable to the analysis of the driving effect of the factor random distribution due to the limitations of the small absolute values of the coefficients, the unmet significance level, and the significant aggregation of the residual results (Fig. 10a–e) (Global Moran’s I = 0.618, 0.589, 0.591, 0.589, 0.086). Based on the results of Pearson correlation and OLS regression analysis (Fig. 11), the five driving factors with the strongest effects of multicollinearity (VIF: AAT, ELE, NPP, RFI, SLP > 7.5) were removed, the GWR model considering spatial heterogeneity was used to explore the driving effects of 11 factors including DRA, DRN, FVC, GDP, PCL, MPA, NLI, CLP, POP, RND and SMC on habitat quality in Ningxia.

Table 4.

Download as

CSV

XLSX

Comparison of regression test indicators between OLS and GWR models.

Year	AIC		AICc		R²		Adj R²
Year	OLS	GWR	OLS	GWR	OLS	GWR	OLS	GWR
2000	-2,818.34	-4,365.10	-2,818.02	-3,060.87	0.327	0.735	0.322	0.692
2005	-2,754.45	-4,436.45	-2,754.14	-2,542.82	0.312	0.779	0.307	0.721
2010	-2,829.25	-4,302.96	-2,828.93	-2,402.95	0.325	0.762	0.320	0.700
2015	-2,832.13	-4,425.03	-2,831.81	-2,525.42	0.333	0.777	0.328	0.719
2020	-2,198.45	-2,263.94	-2,198.14	-1,769.53	0.121	0.202	0.114	0.158

Figure 10.

Results of spatial autocorrelation analysis of OLS regression residuals for habitat quality in Ningxia, 2000–2020.

Figure 11.

OLS regression coefficients, P values and VIF values of the driving factors from 2000 to 2020.

Key driving factors of habitat quality in Ningxia

The results of GWR analysis showed (Table 5) that GDP (-0.001), NLI (-0.184), CLP (-1.413), and PCL (-3.260) negatively affected the habitat quality level in Ningxia, while FVC (1.765), SMC (0.619), DRA (0.023), POP (0.016), DRN (0.014), MPA (0.009), and RND (0.000) were positively correlated with habitat quality in Ningxia. Combining the absolute values of coefficients and driving effects over 20 years, we found that FVC (0.431, 0.392, 0.419, 0.256, 0.267), SMC (0.324, 0.097, 0.116, 0.089, -0.008), PCL (-0.687, -0.733, -0.687, -0.674, – 0.479), and CLP (-0.269, -0.356, -0.322, -0.327, -0.139) (Fig. 12) were the four factors with the strongest combined explanatory power for the spatial and temporal patterns and changes in habitat quality in Ningxia. Among them, FVC and SMC were positively correlated with habitat quality, PCL and CLP were negatively correlated with habitat quality, and the explanatory power of the remaining driving factors was weaker. This result indicates that higher vegetation cover and soil water content can significantly promote ecosystem habitat maintenance function. In contrast, a larger area of cultivated land and construction land within the unit grid will decrease the habitat quality.

Table 5.

Download as

CSV

XLSX

GWR regression coefficient values of habitat quality in Ningxia from 2000 to 2020.

Variable	2000	2005	2010	2015	2020	Relative sum of coefficients
Intercept	0.041	0.025	0.121	0.163	0.498	0.849
DRA	0.007	0.002	0.007	0.005	0.002	0.023
DRN	0.003	0.002	0.002	0.002	0.004	0.014
FVC	0.431	0.392	0.419	0.256	0.267	1.765
GDP	0.002	-0.001	0.000	-0.002	0.000	-0.001
PCL	-0.687	-0.733	-0.687	-0.674	-0.479	-3.260
MPA	0.002	0.002	0.001	0.001	0.002	0.009
NLI	-0.061	-0.054	-0.059	-0.010	0.000	-0.184
CLP	-0.269	-0.356	-0.322	-0.327	-0.139	-1.413
POP	0.001	0.004	0.001	0.010	0.000	0.016
RND	-0.002	-0.002	-0.001	0.000	0.004	0.000
SMC	0.324	0.097	0.116	0.089	-0.008	0.619

Figure 12.

GWR regression coefficients of the main driving factors of habitat quality in Ningxia, 2000–2020.

From the local R² distribution map of Ningxia (Fig. 13a–e), it can be seen that from 2000 to 2005, the R² was higher in the Helan Mountain in northern Ningxia and in the Liupan Mountain in the south, while the R² was relatively smaller in the central and eastern regions, indicating that nature reserves such as the Helan Mountain and Liupan Mountain strongly affected habitat quality. Combined with the decreasing R² explanatory power of the key driving factors of FVC, SMC, PCL, CLP and habitat quality from 2000 to 2020, it was clear that the driving factors affecting habitat quality in Ningxia had complex nonlinear intersection characteristics during the 20 years, while the role of human disturbance became increasingly prominent.

Figure 13.

Local R² distribution of GWR regressions of the driving factors of habitat quality in Ningxia from 2000 to 2020.

Selecting the spatial distribution of the GWR regression coefficients of the main driving factors of habitat quality in 2020 as an example, spatially (Fig. 14a–d), the positive regression coefficients of FVC in Ningxia from 2000 to 2020 were concentrated in the central and northern Helan Mountain and the southern Liupan Mountain. These areas had a relatively high level of habitat quality, and the positive impact of FVC on habitat quality was very significant. The areas with positive SMC regression coefficients mainly included the central cultivated areas with good irrigation conditions and sufficient soil moisture, and a higher SMC had a stronger effect on habitat quality in these areas. The areas with negative PCL regression coefficients were mainly concentrated in the northern part of Ningxia with a high density of construction land and high developed urbanization level, and the habitat quality in these areas was generally low. The areas with negative CLP regression coefficients were mainly located in cultivated land in the yellow irrigation area, which retained part of the habitat maintenance function and had a moderate level of habitat quality. The spatial distribution of the regression coefficients of FVC, SMC, PCL, CLP and GWR of habitat quality had a small threshold of change during the 20 years, and the driving effects spread from the most accurate and close areas of their respective regression effects to the surrounding areas and gradually weakened, together with the interactions of other influencing factors, thus driving the distribution and evolution of habitat quality in Ningxia in the past 20 years. The driving effect of these key driving factors is particularly evident in the nature reserves of Ningxia, and the changes in key driving factors of habitat quality in nature reserves will largely drive changes in biodiversity, reflecting the importance of nature reserve delineation for biodiversity conservation under conditions of avoiding more human activities and sufficient ecological factors.

Figure 14.

Distribution of GWR regression coefficients for the main driving factors of habitat quality in Ningxia from 2000 to 2020.

Discussion

Habitat quality and biodiversity distribution in Ningxia

We conclude that the distribution and evolution of habitat quality in Ningxia were mainly driven by fractional vegetation cover, soil moisture content, cultivated land expansion, and construction land expansion, where high vegetation cover and soil moisture content were suitable for biological habitats; in contrast, cultivated land and construction land expansion reduced habitat suitability. The habitat quality of forestland, grassland, water area and some cultivated land in Ningxia was high, and these patches were in good condition as ecological source land and were far from human activity areas, so they were less disturbed by resource development and utilization. The habitat quality of areas such as urban land, which had a high intensity of human activities, was obviously extremely low. The relationships between soil moisture content, some cultivated land, pasture land and habitat quality were more specific. On the one hand, higher soil water content in the natural state means lush vegetation, and the expansion of cultivated land is not conducive to habitat maintenance. However, on the other hand, due to the construction of artificial cultivated land and artificial wetland, the investment of human and material resources, green funds and ecological technology can maintain the fragile habitat to a certain extent. Therefore, there is uncertainty regarding the role of factors such as cultivated land and soil water content on habitat quality. In response to the weak natural foundation, from 2000 to 2020, Ningxia continuously improved its vegetation cover through the project of returning farmland to forest and grass, and the level of habitat quality in nature reserves such as Liupanshan and Luoshan increased significantly. Due to frequent industrial and agricultural activities such as food production and mineral extraction, the habitat quality fluctuated in some areas of Ningxia over the course of 20 years, and the results showed a decrease in habitat function. In conclusion, urban and cultivated land expansion are the most critical factors reducing habitat suitability in Ningxia, and protecting and utilizing grassland, vegetation, wetland and other ecosystems can effectively improve habitat suitability in Ningxia.

In addition, although other factors, such as topographic relief, had insufficient explanatory power for habitat quality in Ningxia, related studies have shown that rainfall, slope, elevation, and temperature are important conditions that influence the distribution and ecological niche of organisms within small-scale habitats (Rahbek et al. 2019; Chase et al. 2020). Combining field surveys and indoor ecological experiments to obtain detailed habitat data for all or key species in a region is essential in driving factors studies of habitat quality (Forister et al. 2015). Therefore, in specific habitats and typical ecosystems such as grasslands, deserts, and forests, we need more refined work to obtain the total number of various ecological factors and their thresholds required for birth, death, migration, dispersal, and reproduction during the life cycles of key species in the region to determine the causes of habitat quality heterogeneity within small and medium scales (Fahrig 2017). Additionally, these elements are the focus of ecological niche, biodiversity, and ecosystem conservation studies (Xiao et al. 2004).

GWR modelling can fully reflect the spatial non-smoothness of the region and select the optimal spatial weights in combination with spatial heterogeneity; the results have higher accuracy and credibility than Pearson correlation analysis and OLS regression for exploring the role of driving factors at large spatial and temporal scales. Although, the results are consistent with the trend of habitat quality changes in Ningxia over the past 20 years. However, some studies have shown that GWR is essentially a one-dimensional linear regression with parameters that fail to consider multivariate and correlational settings (Fotheringham and Oshan 2016; Fotheringham et al. 2017), while the occurrence of some human irresistible factors, causing the Moran Index and the R² results after regression in 2020 in this study were less satisfactory. To improve the exploration method of driving factors, future research can utilize multivariate nonlinear regression, image-by-image correlation analysis, and artificial neural network models (Fotheringham et al. 2015; Li and Fotheringham 2020). Meanwhile, the sampling accuracy of the unit grid was improved, which eventually made the results match the actual influence of driving factors (McGarigal et al. 2016).

Habitat maintenance and ecosystem protection in ecologically fragile areas

Ningxia is a typical ecologically fragile area with relatively poor natural foundations, and the habitat quality is at an intermediate level with a wide scope for improvement. This study explored the driving effects of habitat quality in Ningxia based on 16 physical geographic, economic and social factors and found that habitat quality tended to be higher in areas with higher vegetation cover and lower in patches with a high proportion of construction land and cultivated land. The results demonstrate that for ecologically fragile areas with similar characteristics to Ningxia’s ecological environment, enhancing vegetation cover can effectively improve habitat structure and function. For ecologically fragile areas, poor conditions such as low precipitation, loose soils, and low biodiversity, coupled with the frequent use of natural resources due to economic development, have led to further anthropogenic damage to the already fragile habitats. Ecologically fragile areas often lack biologically beneficial ecological factors, and due to the harsh natural conditions and prominent human-land conflicts, humans must sacrifice habitat quality in exchange for improved well-being, thus leading to a vicious circle of economic and social development and ecosystem decline in ecologically fragile areas (Tong and Long 2003). Therefore, it is of great practical significance and scientific value to find a way to maintain habitat quality, ecological environmental protection and sustainable and healthy economic and social development in both directions in ecologically fragile areas (Len and Liu 1999; Aksoy and Bayram Arli 2020). Of course, achieving a synergistic promotion of natural systems and economic society depends on rich biodiversity and on the frequency and cost of investment of time, money, technology, and methods by government, society, the public, and researchers and conservation agencies.

In this study, we considered 16 factors that are critical for influencing biodiversity levels in ecologically fragile areas, and these factors fully reflected the natural climatic conditions and human activity disturbances in ecologically fragile areas. The results can be used as a reference for the conservation of ecosystems in other ecologically fragile areas internationally. In the future, studies on biodiversity conservation and habitat quality in ecologically fragile areas should combine model simulations and biodiversity field surveys to summarize the distribution and change characteristics of habitat quality over a long time series. It is also important to understand the habitats of specific species from the key areas of habitat quality maintenance, such as nature reserves and ecological functional areas, which will give more full play to the natural stability (resistance and resilience) and human maintenance of the habitat systems in ecologically fragile areas (Geldmann et al. 2013; Gray et al. 2016). For example, Ningxia has designated a priority area for biodiversity conservation (24,409.7 km²), which accounts for 47.1% of the national territory, effectively covering the typical ecosystems, biodiversity hotspots, and important ecological function areas within the territory and will play a leading role in long-term biodiversity conservation. How to build ecological barriers, ecological corridors and ecological nodes using nature reserves, nature protection areas and other ecosystems with strong habitat suitability to protect the landscape diversity and functional integrity of ecologically fragile areas for biological survival and reproduction will be important for future policy-makers and academic research. Our research supports that focusing on important natural conditions such as vegetation and precipitation in ecologically fragile areas and reducing excessive resource claims will benefit the natural vitality and well-being of human in ecologically sensitive and fragile areas.

Global changes such as climate change and cultivated land expansion have increased the instability of ecosystems in ecologically fragile areas (Garcia et al. 2014; Hautier et al. 2015; Venter et al. 2016; Newbold 2018), leading to a deterioration in biodiversity levels. Studies on global biodiversity suggest (Waldron et al. 2013; Gray et al. 2016; Lepczyk et al. 2017; Wintle et al. 2019; Maxwell et al. 2020; Xu et al. 2021) that the main measures that can be applied to biodiversity conservation in ecologically fragile areas are (1) the use of small patches to protect the minimum suitable habitat for very small populations and in this way consolidate the basis of biodiversity; (2) strengthening the intensification of cultivated land, the construction of nature reserves and the investment of green funds at the regional scale and focusing on the creation of good biodiversity landscapes in urban spaces; and (3) relying on international cooperation such as the Convention on Biological Diversity, in which we will link biodiversity and ecosystem services and improve human well-being to enhance genetic, species, and ecosystem diversity. Under the above opportunities, the identification of habitat suitability and ecological factors that cause habitat changes, the introduction of nature-based solutions to biodiversity dilemmas, and the realization of the systematic assessment, monitoring and management of biodiversity and ecosystems in multiple spatial and temporal sequences will further contribute to the achievement of the UN CBD goals.

Summary

This study evaluated the habitat quality of Ningxia from 2000 to 2020 based on the InVEST model, analysed the spatial and temporal patterns and changes during the 20-year period, and explored the role of driving factors on habitat quality using correlation analysis, the OLS model, and the GWR model in combination with 16 physical-geo-socioeconomic factors. The main results of this study are as follows:

From 2000 to 2020, the average habitat quality in Ningxia was 0.576,428, 0.578,517, 0.576,102, 0.573,025, and 0.573,325, respectively, which increased and then decreased over 20 years, with a small overall decrease. The high-level habitat quality patches in Ningxia were mainly distributed in areas with high vegetation cover, while the low-level habitat quality patches were mainly distributed in areas subject to more frequent human activities, such as construction land and cultivated land. The habitat quality level in Ningxia had significant high-high value aggregation and low-low value aggregation characteristics, which basically overlapped with the distribution of high-level habitat quality and low-level habitat quality, respectively, while the area of patches with low-high value and high-low value aggregation characteristics was small.
From 2000 to 2020, the regression results R ² of 16 driving factors and the GWR model of habitat quality in Ningxia were 0.691,66, 0.721,169, 0.699,633, 0.718,556, and 0.158,344, respectively. By comparing R ², AIC, AICc and other test indicators, we found that the GWR regression model in this study was able to fit the driving effects of different factors on habitat quality in Ningxia at different spatial scales. The results showed that FVC, SMC, PCL and CLP were the most important driving factors affecting the spatial distribution and evolutionary characteristics of habitat quality in Ningxia. Among them, high vegetation cover and soil water content positively promoted habitat suitability, and construction land and cultivated land expansion negatively affected habitat suitability.
The integrity of habitats can ensure the ecological factors required for biological survival, while the multifunctionality of landscapes enriches the diversity of biological evolution. In natural ecosystems and socio-economic systems, the positive or negative role played by human activities is increasingly becoming an important driving factor for the quality of habitats. Whether policy formulation and public participation prioritize, coordinate, or lag the development of natural biodiversity, it determines the indispensable harmony between human and nature in the region. By integrating the distribution, evolution and driving factors of habitat quality in Ningxia, we found that for the conservation of biodiversity and habitat quality in ecologically fragile areas, it is first necessary to maintain the integrity of the original natural habitats as much as possible and increase the multifunctionality of the landscape so that organisms can survive, reproduce and spread smoothly. Second, it is necessary to minimize human interference with the natural landscape and to carry out human activities such as urban construction and cultivated land production in an appropriate and reasonable manner to avoid habitat fragmentation and improve the connectivity of habitat patches. Finally, it is necessary to protect already fragile natural habitats by delineating nature reserves and to develop suitable ecological environmental protection policies for targeted protection and restoration of habitat-sensitive and fragile areas.

Acknowledgments

We appreciate the constructive comments and suggestions from the reviewers that helped improve the quality of this manuscript.

Additional information

Conflict of interest

The authors declare no conflict of interest. The funders had no role in the design of the study; in the collection, analyses, or interpretation of data; in the writing of the manuscript; or in the decision to publish the results.

Ethical statement

No ethical statement was reported.

Funding

This research was funded by the National Natural Science Foundation of China (Grant Number 41871196).

Author contributions

H.H. constructed the concept and overall framework of the research and provided the necessary funding sources and other funding for the research; D.W. collected and sorted out the research data and designed and visualized the charting of the research results; L.S. and H.L. helped D.W., to skillfully use relevant software and charts; D.W., has written and reviewed the first draft, and Y.L. has given great creativity in polishing and improving the article; H.H., L.S., H.L., and Y.L. have all made great contributions to the writing, revision, editing of articles and the management, investigation and implementation of projects. All authors have read and agreed to the published version of the manuscript.

Author ORCIDs

Ding Wang https://orcid.org/0000-0003-1748-8383

Haiguang Hao https://orcid.org/0000-0003-3726-7254

Hao Liu https://orcid.org/0000-0002-1495-2120

Lihui Sun https://orcid.org/0000-0001-6380-8232

Yuyang Li https://orcid.org/0000-0001-6352-271X

Data availability

Not applicable.

References

Aksoy F, Bayram Arli N (2020) Evaluation of sustainable happiness with sustainable development goals: Structural equation model approach. Sustainable Development 28(1): 385–392. https://doi.org/10.1002/sd.1985

Bai LM, Feng XH, Sun RF, Gao H (2020) Spatial and temporal responses of habitat quality to urbanization: A case study of Changchun City, Jilin Province, China. Chinese Journal of Applied Ecology 31(4): 1267–1277. https://doi.org/10.13287/j.1001-9332.202004.012

Bao YB (2022) Habitat assessment and ecological corridor construction of priority areas for biodiversity conservation in Ningxia based on InVEST model. Journal of Ningxia University 43(03): 318–324. https://kns.cnki.net/kns8/defaultresult/index [Natural Science Edition]

Bao YB, Liu K, Li T, Hu S (2015) Effects of land use change on habitat based on invest model – Taking yellow river wetland nature reserve in Shaanxi province as an example. Ganhanqu Yanjiu 32(03): 622–629. https://doi.org/10.13866/j.azr.2015.03.29

Berger AR, Hodge RA (1998) Natural change in the environment: A challenge to the pressure-state-response concept. Social Indicators Research 44(2): 255–265. https://doi.org/10.1023/A:1006888532080

Brunsdon C, Fotheringham AS, Charlton ME (1996) Geographically weighted regression: A method for exploring spatial nonstationarity. Geographical Analysis 28(4): 281–298. https://doi.org/10.1111/j.1538-4632.1996.tb00936.x

Celina A, Jens-Christian S, German T, Francesc B, Unai P (2022) Wildness and habitat quality drive spatial patterns of urban biodiversity. Landscape and Urban Planning 228: e104570. https://doi.org/10.1016/j.landurbplan.2022.104570

Chase JM, Blowes SA, Knight TM, Gerstner K, May F (2020) Ecosystem decay exacerbates biodiversity loss with habitat loss. Nature 584(7820): 238–243. https://doi.org/10.1038/s41586-020-2531-2

Chen Y, Qiao F, Jiang L (2016) Effects of land use pattern change on regional scale habitat quality based on InVEST model – a case study in Beijing. Acta Scientiarum Naturalium Universitatis Pekinensis (3): 553–562. https://doi.org/10.13209/j.0479-8023.2016.057

Chisholm C, Lindo Z, Gonzalez A (2011) Metacommunity diversity depends on connectivity and patch arrangement in heterogeneous habitat networks. Ecography 34(3): 415–424. https://doi.org/10.1111/j.1600-0587.2010.06588.x

Congdon J (1974) Effect of habitat quality on distributions of three sympatric species of desert rodents. Journal of Mammalogy 55(3): 659–662. https://doi.org/10.2307/1379557

Crooks KR, Burdett CL, Theobald DM, King SRB, Di Marco M, Rondinini C, Boitani L (2017) Quantification of habitat fragmentation reveals extinction risk in terrestrial mammals. Proceedings of the National Academy of Sciences of the United States of America 114(29): 7635–7640. https://doi.org/10.1073/pnas.1705769114

Dallimer M, Irvine KN, Skinner AMJ, Davies ZG, Rouquette JR, Maltby LL, Warren PH, Armsworth PR, Gaston KJ (2012) Biodiversity and the feel-good factor: Understanding associations between self-reported human well-being and species richness. Bioscience 62(1): 47–55. https://doi.org/10.1525/bio.2012.62.1.9

Duan H, Yu X (2022) Land-use change, habitat connectivity, and conservation gaps: A case study of shorebird species in the Yellow River Delta of China using the InVEST model and network analysis. Remote Sensing 14(24): e6191. https://doi.org/10.3390/rs14246191

Dunning JB, Danielson BJ, Ronald-Pulliam H (1992) Ecological processes that affect populations in complex landscapes. Oikos 65(1): 169–175. https://doi.org/10.2307/3544901

Fahrig L (2017) Ecological responses to habitat fragmentation per se. Annual Review of Ecology, Evolution, and Systematics 48(1): 1–23. https://doi.org/10.1146/annurev-ecolsys-110316-022612

Forister ML, Novotny V, Panorska AK, Baje L, Basset Y, Butterill PT, Cizek L, Coley PD, Dem F, Diniz IR, Drozd P, Fox M, Glassmire AE, Hazen R, Hrcek J, Jahner JP, Kaman O, Kozubowski TJ, Kursar TA, Lewis OT, Lill J, Marquis RJ, Miller SE, Morais HC, Murakami M, Nickel H, Pardikes NA, Ricklefs RE, Singer MS, Smilanich AM, Stireman JO, Villamarín-Cortez S, Vodka S, Volf M, Wagner DL, Walla T, Weiblen GD, Dyer LA (2015) The global distribution of diet breadth in insect herbivores. Proceedings of the National Academy of Sciences of the United States of America 112(2): 442–447. https://doi.org/10.1073/pnas.1423042112

Fotheringham AS, Oshan TM (2016) Geographically weighted regression and multicollinearity: Dispelling the myth. Journal of Geographical Systems 18(4): 303–329. https://doi.org/10.1007/s10109-016-0239-5

Fotheringham AS, Crespo R, Yao J (2015) Geographical and Temporal Weighted Regression (GTWR). Geographical Analysis 47(4): 431–452. https://doi.org/10.1111/gean.12071

Fotheringham AS, Yang W, Kang W (2017) Multiscale Geographically Weighted Regression (MGWR). Annals of the Association of American Geographers 107(6): 1247–1265. https://doi.org/10.1080/24694452.2017.1352480

Garcia RA, Cabeza M, Rahbek C, Araújo MB (2014) Multiple dimensions of climate change and their implications for biodiversity. Science 344(6183): e1247579. https://doi.org/10.1126/science.1247579

Geldmann J, Barnes M, Coad L, Craigie ID, Hockings M, Burgess ND (2013) Effectiveness of terrestrial protected areas in reducing habitat loss and population declines. Biological Conservation 161: 230–238. https://doi.org/10.1016/j.biocon.2013.02.018

Goertz JW (1964) The influence of habitat quality upon density of Cotton Rat populations. Ecological Monographs 34(4): 359–381. https://doi.org/10.2307/2937068

Gomes E, Inácio M, Bogdzevič K, Kalinauskas M, Karnauskaitė D, Pereira P (2021) Future scenarios impact on land use change and habitat quality in Lithuania. Environmental Research 197: e111101. https://doi.org/10.1016/j.envres.2021.111101

Gray CL, Hill SLL, Newbold T, Hudson LN, Börger L, Contu S, Hoskins AJ, Ferrier S, Purvis A, Scharlemann JPW (2016) Local biodiversity is higher inside than outside terrestrial protected areas worldwide. Nature Communications 7(1): e12306. https://doi.org/10.1038/ncomms12306

Haddad N, Brudvig L, Clobert J, Davies K, Gonzalez A, Holt R, Lovejoy T, Sexton J, Austin M, Collins CD, Cook W, Damschen E, Ewers R, Foster B, Jenkins C, King AJ, Laurance W, Levey D, Margules C, Melbourne B, Nicholls AO, Orrock J, Song D, Townshend J (2015) Habitat fragmentation and its lasting impact on earth’s ecosystems. Science Advances 1(2): e1500052. https://doi.org/10.1126/sciadv.1500052

Hale R, Swearer SE (2016) Ecological traps: Current evidence and future directions. Proceedings of The Royal Society B Biological Sciences 283(1824): e20152647. https://doi.org/10.1098/rspb.2015.2647

Hao Y, Zhang N, Du YJ, Wang YH, Zheng YD, Zhang CC (2019) Construction of ecological security pattern based on habitat quality in Tang County. Chinese Journal of Applied Ecology 30(03): 1015–1024. https://doi.org/10.13287/j.1001-9332.201903.007

Harper D, Everard M (1998) Why should the habitat-level approach underpin holistic river survey and management? Aquatic Conservation 8(4): 395–413. https://doi.org/10.1002/(SICI)1099-0755(199807/08)8:4<395::AID-AQC297>3.0.CO;2-X

Hautier Y, Tilman D, Isbell F, Seabloom EW, Borer ET, Reich PB (2015) Anthropogenic environmental changes affect ecosystem stability via biodiversity. Science 348(6232): 336–340. https://doi.org/10.1126/science.aaa1788

Van Horne B (1983) Density as a misleading indicator of habitat quality. The Journal of Wildlife Management 47(4): 893–901. https://doi.org/10.2307/3808148

Hu J, Zhang J, Li Y (2022) Exploring the spatial and temporal driving mechanisms of landscape patterns on habitat quality in a city undergoing rapid urbanization based on GTWR and MGWR: The case of Nanjing, China. Ecological Indicators 143: e109333. https://doi.org/10.1016/j.ecolind.2022.109333

Janzen DH (1970) Herbivores and the number of tree species in Tropical Forests. American Naturalist 104(940): 501–528. https://doi.org/10.1086/282687

Kalacska M, Arroyo-Mora JP, Lucanus O, Kishe-Machumu MA (2017) Land cover, land use, and climate change impacts on Endemic Cichlid Habitats in Northern Tanzania. Remote Sensing 9(6): e623. https://doi.org/10.3390/rs9060623

Kareiva P, Tallis H, Ricketts TH, Daily GC, Polasky S (2011) Natural Capital: Theory & Practice of Mapping Ecosystem Services. Ed, Oxford biology. Oxford University Press, Oxford [England], New York. https://doi.org/10.1093/acprof:oso/9780199588992.001.0001

Len SY, Liu YH (1999) The framework design of sustainable development index system for fragile ecological regions in China. China Population, Resources and Environment 02: 42–47.

Lepczyk CA, Aronson MF, Evans KL, Goddard MA, Lerman SB, MacIvor SJ (2017) Biodiversity in the City: Fundamental Questions for Understanding the Ecology of Urban Green Spaces for Biodiversity Conservation. Bioscience 67(9): 799–807. https://doi.org/10.1093/biosci/bix079

Li Z, Fotheringham AS (2020) Computational improvements to Multi-Scale Geographically Weighted Regression. International Journal of Geographical Information Science 34(7): 1378–1397. https://doi.org/10.1080/13658816.2020.1720692

Maxwell SL, Cazalis V, Dudley N, Hoffmann M, Rodrigues ASL, Stolton S, Visconti P, Woodley S, Kingston N, Lewis E, Maron M, Strassburg BBN, Wenger A, Jonas HD, Venter O, Watson JEM (2020) Area-based conservation in the twenty-first century. Nature 586(7828): 217–227. https://doi.org/10.1038/s41586-020-2773-z

McGarigal K, Wan HY, Zeller KA, Timm BC, Cushman SA (2016) Multi-scale habitat selection modeling: A review and outlook. Landscape Ecology 31(6): 1161–1175. https://doi.org/10.1007/s10980-016-0374-x

Moëzzi F, Poorbagher H, Eagderi S, Feghhi J, Dormann CF, Nergi SK, Amiri K (2022) Modelling habitat preference of Caspian Kutum, Rutilus kutum, using non-linear habitat suitability indices and generalized additive models. Regional Studies in Marine Science 56: e102715. https://doi.org/10.1016/j.rsma.2022.102715

Nagendra H, Lucas R, Honrado JP, Jongman RHG, Tarantino C, Adamo M, Mairot P (2013) Remote Sensing for conservation monitoring: Assessing protected areas, habitat extent, habitat condition, species diversity, and threats. Ecological Indicators 33: 45–59. https://doi.org/10.1016/j.ecolind.2012.09.014

Newbold T (2018) Future effects of climate and land-use change on terrestrial vertebrate community diversity under different scenarios. Proceedings of the Royal Society B, Biological Sciences 285(1881): e20180792. https://doi.org/10.1098/rspb.2018.0792

Peggy RM, Tridoyo K, Luky A, Achmad F (2021) Valuing habitat quality for managing mangrove ecosystem services in coastal Tangerang District, Indonesia. Marine Policy 133: e104747. https://doi.org/10.1016/j.marpol.2021.104747

Prasad G, Ramesh MV (2019) Spatio-temporal analysis of land use/land cover changes in an ecologically fragile area – Alappuzha District, Southern Kerala, India. Natural Resources Research 28(S1): 31–42. https://doi.org/10.1007/s11053-018-9419-y

Qin WZ (2007) The basic theoretics and application research on geographically weighted regression. PhD Thesis, Tongji University, Shanghai, China. https://kns.cnki.net/KCMS/detail/detail.aspx?dbname=CDFD0911&filename=2007222843.nh

Radosavljevic A, Anderson RP (2014) Making better maxent models of species distributions: Complexity, overfitting and evaluation. Journal of Biogeography 41(4): 629–643. https://doi.org/10.1111/jbi.12227

Rahbek C, Borregaard MK, Colwell RK, Dalsgaard B, Holt BG, Morueta-Holme N, Nogues-Bravo D, Whittaker RJ, Fjeldså J (2019) Humboldt’s enigma: What causes global patterns of mountain biodiversity? Science 365(6458): 1108–1113. https://doi.org/10.1126/science.aax0149

Raimundo Lopes ND, Li TX, Qian DY, Matomela N, Sá RM (2022) Factors influencing coastal land cover change and corresponding impact on habitat quality in the North-Western Coastline of Guinea-Bissau (NC-GB). Ocean & Coastal Management 224: e106181. https://doi.org/10.1016/j.ocecoaman.2022.106181

Ramachandra TV, Bharath S, Vinay SA (2019) Visualisation of impacts due to the proposed developmental projects in the ecologically fragile regions- Kodagu district, Karnataka. Progress in Disaster Science 3: e100038. https://doi.org/10.1016/j.pdisas.2019.100038

Ren J, Zhao XY, Xu XC, Ma PY, Du YX (2022) Spatial-temporal evolution, tradeoffs and synergies of ecosystem services in the middle Yellow River. Journal of Earth Environment 13(4): 477–490. https://doi.org/10.7515/JEE222019

Riedler B, Lang SA (2018) A spatially explicit patch model of habitat quality, integrating spatio-structural indicators. Ecological Indicators 94: 128–141. https://doi.org/10.1016/j.ecolind.2017.04.027

Roelof B, Joe R, Irit A, Les K (2015) The multiscale integrated model of ecosystem services (MIMES): Simulating the interactions of coupled human and natural systems. Ecosystem Services 12: 30–41. https://doi.org/10.1016/j.ecoser.2015.01.004

Romero-Calcerrada R, Luque S (2006) Habitat quality assessment using Weights-of-Evidence based GIS modelling: The case of Picoides Tridactylus as species indicator of the biodiversity value of the Finnish Forest. Ecological Modelling 196(1): 62–76. https://doi.org/10.1016/j.ecolmodel.2006.02.017

Rosenzweig ML, Winakur J (1969) Population ecology of desert rodent communities: Habitats and environmental complexity. Ecology 50(4): 558–572. https://doi.org/10.2307/1936246

Roth NE, Allan JD, Erickson DL (1996) Landscape influences on stream biotic integrity assessed at multiple spatial scales. Landscape Ecology 11(3): 141–156. https://doi.org/10.1007/BF02447513

Smith AC, Fahrig L, Francis CM (2011) Landscape size affects the relative importance of habitat amount, habitat fragmentation, and matrix quality on forest birds. Ecography 34(1): 103–113. https://doi.org/10.1111/j.1600-0587.2010.06201.x

Sobhani P, Esmaeilzadeh H, Barghjelveh S, Sadeghi SMM, Marcu MV (2022) Habitat integrity in protected areas threatened by LULC changes and fragmentation: A case study in Tehran Province, Iran. Land 11(1): 1–6. https://doi.org/10.3390/land11010006

Tong YQ, Long HL (2003) Study on the sustainable development of poor areas under the coupling of fragile ecological environment. China Population, Resources and Environment 02: 50–54.

van Vliet J, de Groot HLF, Rietveld P, Verburg PH (2015) Manifestations and underlying drivers of agricultural land use change in Europe. Landscape and Urban Planning 133: 24–36. https://doi.org/10.1016/j.landurbplan.2014.09.001

Venter O, Sanderson EW, Magrach A, Allan JR, Beher J, Jones KR, Possingham HP, Laurance WF, Wood P, Fekete BM, Levy MA, Watson JEM (2016) Sixteen years of change in the global terrestrial human footprint and implications for biodiversity conservation. Nature Communications 7(1): e12558. https://doi.org/10.1038/ncomms12558

Villa F, Bagstad K, Johnson G, Krivov S, Ceroni M (2009) ARIES (Artificial Intelligence for Ecosystem Services): A new tool for ecosystem services assessment, planning, and valuation. 11^th Annual Bioecon Conference on Economic Instruments to Enhance the Conservation and Sustainable Use of Biodiversity, Venice (Italy), September, 2009.

Waldron A, Mooers AO, Miller DC, Nibbelink N, Redding D, Kuhn TS, Roberts J, Gittleman JL (2013) Targeting global conservation funding to limit immediate biodiversity declines. Proceedings of the National Academy of Sciences of the United States of America 110(29): 12144–12148. https://doi.org/10.1073/pnas.1221370110

Wang ZQ, Chen ZC, Hao CY (2009) Breeding habitat suitability evaluation of red-crown crane in Zhalong national nature reserve by the method of habitat suitability index. Wetland Science 7(03): 197–201. https://doi.org/10.13248/j.cnki.wetlandsci.2009.03.002

Wang Q, Lu C, Li FY, Fan ZP (2017) River habitat quality assessment based on principal component analysis and entropy weight in Qinghe River as a case. Ecologic Science 36(4): 185–193. https://doi.org/10.14108/j.cnki.1008-8873.2017.04.025

Wang Z, Xiao L, Yan H, Qi Y, Jiang Q (2022) Optimization of the ecological network structure based on scenario simulation and trade-offs/synergies among ecosystem services in Nanping. Remote Sensing 14(20): e5245. https://doi.org/10.3390/rs14205245

Wintle BA, Kujala H, Whitehead A, Cameron A, Veloz S, Kukkala A, Moilanen A, Gordon A, Lentini PE, Cadenhead NCR, Bekessy SA (2019) Global synthesis of conservation studies reveals the importance of small habitat patches for biodiversity. Proceedings of the National Academy of Sciences of the United States of America 116(3): 909–914. https://doi.org/10.1073/pnas.1813051115

Wu JS, Zhang LQ, Peng J, Feng Z, Liu HM, He SB (2013) The integrated recognition of the source area of the urban ecological security pattern in Shenzhen. Acta Ecologica Sinica 33(13): 4125–4133. https://doi.org/10.5846/stxb201208081123

Wu JS, Mao JY, Qian L, Li JC (2017) Urban growth boundary based on the evaluation of habitat quality: Taking the Yangtze River delta as an example. Dili Kexue 37(1): 28–36. https://doi.org/10.13249/j.cnki.sgs.2017.01.004

Wu D, Li H, Ai N, Huang T, Gu JS (2020) Predicting spatiotemporal changes in land use and habitat quality based on CA-Markov: A case study in central Ningxia, China. Chinese Journal of Eco-Agriculture 28(12): 1969–1978. https://doi.org/10.13930/j.cnki.cjea.200221

Wu J, Li X, Luo Y, Zhang D (2021) Spatiotemporal effects of urban sprawl on habitat quality in the Pearl River Delta from 1990 to 2018. Scientific Reports 11(1): 1–15. https://doi.org/10.1038/s41598-021-92916-3

Xiao Y, Ouyang ZY, Zhu CQ, Zhao JZ, He GJ, Wang XK (2004) An assessment of giant panda habitat in Minshan, Sichuan, China. Acta Ecologica Sinica 07: 1373–1379.

Xiao PN, Zhou Y, Li MY, Xu J (2022) Spatiotemporal patterns of habitat quality and its topographic gradient effects of Hubei Province based on the InVEST model. Environment, Development and Sustainability 11: e13981. https://doi.org/10.1007/s10668-022-02310-w

Xu H, Cao Y, Yu D, Cao MC, He YX, Gill M, Pereira HM (2021) Ensuring effective implementation of the post-2020 global biodiversity targets. Nature Ecology & Evolution 5(4): 411–418. https://doi.org/10.1038/s41559-020-01375-y

Yang J, Xie BP, Zhang DG (2021) Spatial-temporal evolution of habitat quality and its influencing factors in the Yellow River Basin based on InVEST model and GeoDetector. Journal of Desert Research 41(4): 12–22. https://doi.org/10.7522/j.issn.1000-694X.2021.00026

Yohannes H, Soromessa T, Argaw M, Dewan A (2021) Spatio-temporal changes in habitat quality and linkage with landscape characteristics in the Beressa Watershed, Blue Nile Basin of Ethiopian Highlands. Journal of Environmental Management 281: e111885. https://doi.org/10.1016/j.jenvman.2020.111885

Zhu ZY, Alimujiang K (2020) Spatial-temporal evolution of habitat quality in Yili Valley based on geographical detector and its influencing factors. Chinese Journal of Ecology 39(10): 3408–3420. https://doi.org/10.13292/j.1000-4890.202010.009

Zhu C, Zhang X, Zhou M, He S, Gan M, Yang L, Wang K (2020) Impacts of urbanization and landscape pattern on habitat quality using OLS and GWR models in Hangzhou, China. Ecological Indicators 117: e106654. https://doi.org/10.1016/j.ecolind.2020.106654

﻿Abstract

Key words

﻿Introduction

﻿Materials and methods

﻿Study area

﻿Research methods

Habitat quality assessment methods

Spatial autocorrelation analysis

Geographically weighted regression

﻿Data sources and processing

﻿Results and analysis

﻿Spatial and temporal patterns of habitat quality in Ningxia

﻿Dynamic changes of habitat quality in Ningxia

﻿Analysis of the driving factors of habitat quality in Ningxia

Analysis of the applicability of the driving factor regression model

Key driving factors of habitat quality in Ningxia

﻿Discussion

﻿Habitat quality and biodiversity distribution in Ningxia

﻿Habitat maintenance and ecosystem protection in ecologically fragile areas

﻿Summary

﻿Acknowledgments

﻿Additional information

Conflict of interest

Ethical statement

Funding

Author contributions

Author ORCIDs

Data availability

References

Abstract

Introduction

Materials and methods

Study area

Research methods

Data sources and processing

Results and analysis

Spatial and temporal patterns of habitat quality in Ningxia

Dynamic changes of habitat quality in Ningxia

Analysis of the driving factors of habitat quality in Ningxia

Discussion

Habitat quality and biodiversity distribution in Ningxia

Habitat maintenance and ecosystem protection in ecologically fragile areas

Summary

Acknowledgments

Additional information