Dynamic evaluation and prediction of the ecological environment quality of the urban agglomeration on the northern slope of Tianshan Mountains

In order to timely determine the dynamic changes of the ecological environment quality and future development laws of the urban agglomeration on the northern slope of the Tianshan Mountains, combined with the actual situation of the urban agglomeration, 11 indicators were selected from the three aspects of natural ecology, social ecology, and economic ecology. To reduce the dimensions of the indicators, principal component analysis, coefficient of variation, and analytic hierarchy process were used based on RS and GIS technology methods, and the ecological environmental quality (EQI) from 2000 to 2018 was dynamically evaluated. Further, the CA–Markov model was introduced to simulate the development status in 2026 for predictive purposes. The main results are as follows: the overall ecological environment of the area exhibited a gradually improving distribution change from southwest to northeast; the proportion of ecological environment classification exhibited a gradually decreasing change pattern; the spatial differentiation of ecological environment quality exhibited a significant spatial positive correlation; from the influencing factors, an observation can be made that natural ecological factors were highly significant; the prediction accuracy verification revealed that the CA–Markov model was suitable for the prediction of the ecological environment quality in the region and had high accuracy; and the comprehensive regional ecological environment quality indexes were 5.7392, 6.1856, and 6.4366, respectively, while the forecasted value for 2026 was predicted to be 6.6285, indicating that the overall ecological environment quality of the region will improve and develop well. The present research results reveal the law of dynamic changes and future development of the ecological environment quality in the region, which can be used as a theoretical reference for the formulation of ecological environmental protection measures.


Introduction
Based on ecological theory, the quality of the ecological environment refers to the pros and cons of the ecological environment. Through the quality of the ecological Responsible Editor: Philippe Garrigues * Zibibula Simayi zibibulla3283@sina.cn Yan Yibo yanyibo11060020@163.com 1 environment, the suitability of the ecological environment for human survival is reflected, as well as the sustainable development of social economy within a specific time and space range from the level of the ecosystem. The quality of the ecological environment is based on the specific requirements of human beings to evaluate the nature of the ecological environment and the results of its changing state , being directly affected by local natural resources and human life (Ma et al. 2021). Global changes and the intensification of human activities have had a significant impact on the ecosystems that humans rely on for survival and development, which resulted in the increased prominence of ecological problems .
The general evaluation of the quality of the ecological environment helps to determine the current status of sustainable development in the region (Tokatli 2022). This assessment relies on structured, semi-structured, and unstructured data analysis, as well as large data sets in many aspects of nature, society, and economy (Zhao et al. 2022). Using big data technology to manage big data sets, parallel computing provides a scientific and efficient method for regional ecology.
A good ecological environment is vital for people's livelihood and is also the basis for the sustainable development of people and society. Scientific cognition, evaluation, and rational regulation of the ecological environment are hot issues in the field of resources and environment research and are also urgent requirements for the construction of ecological civilization (Zhao et al 2019). However, most studies focus on the regional level, resulting in a lack of in-depth analysis of arid regions. Li (2007) first proposed an ecological environment quality evaluation index system that is closely related to the ecological environment. For the dynamic analysis of regional eco-environmental assessment, Ma and Shi (2016) used the comprehensive index of ecoenvironmental indicators to investigate the economic zone on the west coast of the Taiwan Strait, and developed an index system to evaluate the eco-environmental quality of the economic zone by using the objective weighting method. The overall ecological environment quality of the economic zone is described and analyzed. The results show that there are significant differences in the quality of the ecological environment within the region. Zheng (2020) took Jurassic Yushenfu Mining area in Shenfu-Dongsheng coalfield as an example; a comprehensive evaluation model of ecological geological environment quality is established by combining geographic information system (GIS) spatial analysis with multi-criteria decision analysis (MCDA) method, so as to (aiming to) describe the ecological geological environment quality status of ecologically fragile mining areas under different mining intensities.
The actual condition of the regional ecological environment cannot be measured by a single indicator alone. The development of remote sensing technology in recent years has provided a large amount of data for ecosystem monitoring and a reliable guide for ecological status evaluation at different scales (Klobucar et al. 2021). Wang et al. (2007) conducted a comprehensive evaluation of environmental quality in the middle and upper reaches of the Yellow River using GIS technology, Arc/Info, and ArcView based on remote sensing data and surveys and analyzed the environmental status and overall condition of each ecosystem region. Environmental quality determines the distribution law of environmental quality in the three regions. Hou et al. (2016) used remote sensing and GIS applications to analyze the dynamic changes in land use in Yan'an since the implementation of the policy of returning farmland to forests, and evaluated the ecological vulnerability of the Yan'an area using a spatial principal component analysis (SPCA) model, based on TM images of Yan 'an in 1997'an in , 2004'an in , and 2011'an in . Silva et al. (2018 used satellite imagery, GIS and GPS techniques, topographic maps, climate data and soil maps, and field surveys to analyze the ecological environment changes in Monteiro County, a semi-arid area in northeastern Brazil, from 1987 to 2010, and the ability to output the assessment results to accurately predict the most sensitive and least sensitive areas. Based on Landsat TM/ETM + /OLI/TIRS images for comprehensive evaluation of urban ecological quality, Li et al. (2021) combined entropy weight method and principal component analysis method and constructed an improved remote sensing integrated ecological index (IRSEI) evaluation model to evaluate the ecological environment of Wuhan City from 1995 to 2020, and performed spatial autocorrelation analysis on its geographic clustering.
In terms of the relevant research content, the ecological environment in a single year in a particular study area has been mainly analyzed from the quantitative characteristics, spatial distribution, spatial differentiation, zoning of different spatio-temporal scales, and others. However, there is a scarcity of analysis on the ecological environment changes over different periods of time and the driving forces thereof. In addition, there are few reports that include spatial clustering analysis results of the eco-environmental quality. The resolution of such issues is of considerable significance not only for the scientific evaluation of the effectiveness of current environmental regulation policies but also for encouraging policy makers to promote the optimization of relevant environmental regulation tools.
The focus of overall regional ecological environment quality evaluation has shifted from natural factors to integrated natural factors and socioeconomic conditions. With the development of other technologies such as remote sensing, geographic information systems, and GEE (Jia et al. 2021;Yuan et al. 2021), evaluation methods have also shifted from qualitative evaluation to quantitative evaluation by selecting appropriate indicators and developing evaluation models Zhong et al. 2022). These methods provide quantitative research for ecological or environmental analysis. So far, there is no unified and definite method for evaluating regional ecological environment quality, but the use of objective evaluation methods has become the mainstream of today's ecological environment quality evaluation (Ren et al. 2022).
In the present study, taking the urban agglomeration on the northern slope of Tianshan Mountains as an example, the ecological environment quality of the urban agglomeration on the northern slope of Tianshan Mountains was analyzed from the perspectives of natural ecology, social ecology, and economic ecology by using spatio-temporal big data, the quantitative analysis mathematical method, and GIS and RS remote sensing technology. The objectives of the present study were as follows: (1) use GIS technology and quantitative analysis of mathematical methods to establish a spatiotemporal big data based ecological environmental quality assessment model; (2) analyze the spatial distribution and main influencing factors of eco-environmental quality; (3) determine the spatial autocorrelation of the ecological environment quality; (4) simulate and predict the development of the ecological environment, providing reference for the development and protection of urban agglomeration on the northern slope of Tianshan Mountains and the construction of ecological civilization (Fig. 1).

Study area
The urban agglomeration on the northern slope of the Tianshan Mountains is located in the hinterland of Eurasia and the western edge of the second topological ladder of China, adjacent to Gurbantünggüt Desert in the north and Tianshan Mountain in the south. With superior resource endowment and favorable geographical conditions, the region is an emerging urban agglomeration in the inland arid area of northwest China (Lifang et al. 2021). At the same time, the region is also the core urban agglomeration for the development and construction of the "One Belt and One Road" Economic Belt, being a highly significant factor in the border consolidation of Xinjiang (Shen et al. 2021). The total land area of each county and city is about 8.7 × 104 km 2 . The spatial scope includes 11 counties (Urumqi City, Changji City, Shihezi City, and Karamay City, Fukang City, Kuitun City, Ili Prefecture, Wusu City, Hutubi County, Manas County, and Shawan County in the Tacheng area City) (Heni et al. 2019).
As shown in Fig. 2, the geographical location of urban agglomeration on the northern slope of Tianshan Mountains is between 81° 33′-93° 32′ E and 42° 78′-45° 59′ N. The annual average temperature and precipitation are about 7.5 ℃ and 185.34 mm, respectively, which belong to the mountainous, oasis-desert zone. The desert oasis region has obvious characteristics of "developing by soil and water, expanding with Wells and canals, spreading around basins, and being entrenched along the mountain front" (Qian Li et al. 2021). Alpine snow and melted ice water provide a water source for regional development , and grassland is the main green space cover type . Among the background of the construction of the silk road economic belt, the urban agglomeration is affected by the rapid development of economy and industry and the advancement of urbanization. Notably, such factors seriously threaten human health and the ecological environment.

Data source and preprocessing
In the present study, for urban agglomeration analysis, an index system of the ecological environment of urban agglomerations on the northern slope of Tianshan Mountains was constructed from three aspects: ecology, economic ecology, and social ecology. The data used in the study include the following: normalized difference vegetation index (NDVI) data (Yue et al. 2022)-Landsat remote sensing images from NASA (https:// gpm. nasa. gov/) were selected as the main data in the three time periods of 2000, 2010, and 2018, the selected data seasons were the same, and the element status was relatively consistent, thereby ensuring that the research was comparable and could meet the research accuracy requirements; digital elevation model (DEM) data (Raczkowska and Cebulski 2022)-the spatial resolution was 90 m × 90 m, and the data came from the geospatial data cloud platform (http:// www. gsclo ud. cn), being spliced after downloading; temperature and precipitation data (Mao et al. 2022)

Data preprocessing
(1) DEM data: The slope and topographic undulation data were obtained through the surface and neighborhood tools on the ArcGIS software platform.
(2) Comprehensive index data of land use degree: The comprehensive index of land use degree comprehensively reflects the degree of land use in a certain area. The calculation formula is shown in Formula (1).

Fig. 2 Study area
In the formula, L a is the comprehensive index of land use degree; A i is the grading index of land use degree of the ith level; and C i is the area ratio of the grading land use degree of the ith level.
(3) Meteorological data: The average annual temperature and annual precipitation are based on the data of 7 meteorological stations around the urban agglomeration on the northern slope of Tianshan Mountains. For precipitation, Anusplin tool was used to complete the spatial interpolation of meteorological data. (4) Socio-economic data: Population density is the total population/total area of each city (county), while the proportion of the secondary industry is the total output value of the secondary industry of each city (county)/ the total output value of the secondary industry of the city. For the socio-economic data, ArcGIS software was used to complete the inverse distance weight interpolation model.
The spatial visualization of index data could be realized by obtaining various index factors that characterize the quality of the ecological environment. Because of the differing projection methods, coordinate systems, and scales of various data, the spatial resolution of all data was unified into a 1 km × 1 km grid form, and the same Krasovsky ellipsoid coordinates and Albers projection were used. Data with missing spatial attributes in data processing were obtained indirectly based on data collection technology in reference to existing research results. Subsequently, the natural ecological index (NE1), social ecological index (SEI), and economic ecological index (EEI) were calculated on the ArcGIS platform, and grid calculation was conducted. The natural fracture method was used to obtain the classification map of ecological environment quality evaluation. Finally, the dynamic changes and influencing factors of the ecological environment quality of the urban agglomeration on the northern slope of Tianshan Mountains were analyzed, and the future development of ecological environment quality was predicted.

Rationality of indicators
The local conditions of the regional ecological environment are formed by the long-term interaction of various natural, social, and economic elements in the regional ecological (2) L a ∈ 100, 400 environment (Eckenwiler 2018). In the present study, the basic principles of comprehensiveness, scientificity, systemicity, easy accessibility, independence, and simplicity were followed , and the ecological environment and social economy of the urban agglomeration on the northern slope of Tianshan Mountains were combined. To construct the northern slope of Tianshan Mountains, 11 indicators were selected for the evaluation index system of the ecological environment of urban agglomerations. To be specific, a single indicator cannot reflect the relationship between urban agglomerations and the ecological environment, and factors affecting the quality of the regional ecological environment need to be considered (Komarova et al. 2021).
To ensure the accuracy of the evaluation results, the 11 indicators were verified in terms of whether there was overlap. Thus, the method of multicollinearity diagnosis was adopted to make judgments (Mihreteab et al. 2020). Commonly used diagnostic indicators of multivariate collinearity mainly include variance inflation factor (VIF) and tolerance (TOL) (Sahani and Ghosh 2021), which have a reciprocal relationship. When VIF < 10 (that is, TOL > 0.1, there is no obvious multiple collinearity in the selected indicator (Arabameri et al. 2019). For the specific method, in ArcGIS, a 5 km × 5 km fishing net was used to penetrate the entire boundary layer of the urban agglomeration on the northern slope of Tianshan Mountains, and a total of 5275 points were uniformly generated. With said points, 11 indicators and EQI values were read, and a collinearity diagnosis was calculated in SPSS. The statistics of the two indicators are shown in Table 1. All variables were subjected to a correlation test, and the correlation coefficients between the variables can be seen in Fig. 3. The results show that there was no obvious collinearity among the 11 indicators, and there was no information overlap. Therefore, the 11 indicators were reasonable for the study.

Ecological environment quality evaluation model
Evaluation indicators often have different dimensions, positive and negative terms, and there is significant variance in the data (Gottero and Cassatella 2017). The original data of each evaluation index needs to be standardized before determining the weight of indicators, so as to achieve comparability, testability, and ease of comparison among indicators, it is necessary to standardize (Pozsgai et al. 2021).

Natural ecological index
Under the premise of ensuring the minimum loss of data and information, the spatial principal component analysis method transforms multiple related indicators into several uncorrelated comprehensive indicators by rotating the Table 1 Results of multicolliearity diagnostics p1 NDVI, p2 annual average rainfall, p3 annual average temperature, p4 surface undulation, p5 distance to the river, p6 soil type, p7 soil organic carbon, p8 land use type, p9 distance from road, p10 population density, p11 the proportion of tertiary industry original spatial coordinate axis. As such, the information of the original indicators can be maximized (Arabameri et al. 2019). At the same time, the whole process of spatial principal components does not require weights to be artificially set, and the evaluation results are objective. Based on the aforementioned theories, the ArcGIS platform was used in the present study for evaluation of the index system. To calculate the natural ecological index (NEI), surface undulation, annual average precipitation, annual average temperature, river network density, soil type value, soil organic carbon content, and other indicators related to natural ecology in the index system were analyzed by spatial principal component analysis. The calculation formula is as follows: In the formula, R i is the contribution rate corresponding to the ith principal component, and X i is the ith principal component.
When the cumulative variance contribution rate is greater than or equal to 80%, most of the relevant information of the original data can be replaced (Zemni et al. 2022). In order to truly and objectively obtain the natural ecological information of the urban agglomeration on the northern slope of Tianshan Mountains, the cumulative contribution rate of the first four principal component factors needed to reach over 80% (Table 2). Therefore, the first four principal component factors were selected for fitting calculation.

Social ecological index
The importance of social ecological index indicators is closely related to the amount of information. The coefficient of variation method starts from the attributes of the data and uses the standard deviation of each indicator as the amount of information. The weighted average is used to determine the weight of each indicator . Considering that the impact on the ecological environment of the basin is difficult to determine using the three indicators of social ecology-related population density, land use type, and road network density, the characteristics of the selected indicators were first determined. Subsequently, the weight was determined by the importance of the indicators, and the coefficient of variation method was used to calculate the social ecological index (SEI). The calculation formula is as follows: where V i is the coefficient of variation of the ith index; i and X i are the standard deviation and average of the ith index, respectively; W i is the weight of the ith index; and Y i is the ith index after indexing, which is used to calculate the coefficient of variation and weight of the three indicators of SPI, as shown in Table 3.

Economic ecological index
Economic development will affect the changes of the ecological environment in surrounding areas to a certain extent. An indicator related to economic ecology in the index system is the proportion of the tertiary industry representing the economic development status of the basin , which can reflect the intensity of economic development on the environmental protection and capital investment in the surrounding areas to a certain extent. Thus, the economic ecological index (EEI) in the present study is equivalent to the proportion of the tertiary industry.

Eco-environmental quality evaluation index
The three aspects of nature, society, and economy were taken as the foundation in the present study, and NEI, SEI, and EEI were respectively calculated to characterize the ecological environment quality status of the urban agglomeration on the northern slope of Tianshan Mountains. Combining the unique characteristics of the urban agglomeration on the northern slope of Tianshan Mountains and the importance of the factors affecting the regional ecological environment, the factors were ranked from high to low as follows: natural factors > social factors > economic factors (Ren et al. 2022). Hierarchical analysis (Zhong et al. 2022) was used to determine the weight of NEI, SEI, and EEI. The specific method involved constructing a pairwise discriminant matrix according to the interrelationship between the indicators, and then performing a consistency test. Finally, the weight value of each indicator was calculated. Table 4 shows the discriminant matrix constructed based on said calculation. The maximum characteristic root Xmax = 3.014, the consistency index CI = 0.007, the random consistency index RI = 0.520, and the random consistency ratio CR. = 0.013 < 0.1 passed the consistency test (Jahanger 2021). After calculation, the weights of NEI, SEI, and EEI were respectively 0.570, 0.333, and 0.097. The calculation formula of the eco-environmental quality evaluation index (EQI) is as follows: In order to compare and analyze the differences in ecological environment quality in local areas, the EQI needs to be classified. In the present study, the natural break point method (Jenks) was used for classification (Guo and Yuan 2021). The classification criteria for each period should be unified; otherwise, no comparative analysis can be performed (Bjelle et al. 2021). Therefore, the 2018 grading standard was adopted for both 2000 and 2018. The grading standards are shown in Table 5.

Comprehensive index of ecological environment quality
The comprehensive index of ecological environment quality is an objective indicator to measure the overall condition of the regional ecological environment. The model is: In the formula, EEQI is the comprehensive index of ecological environment quality; P i is the ecological environment level; A i is the number of grids of the ith level; n is the total number of levels; and S is the total number of grids. The smaller the value of E in the study, the worse the overall ecological environment quality of the region.

Spatial clustering model
Spatial autocorrelation analysis involves examining a certain geographical phenomenon or the overall dispersion state of a certain variable, and then determining whether there are agglomeration characteristics in space (Haak et al. 2022). The global spatial autocorrelation index is used to verify the spatial correlation index of a certain element in the entire research area. In the present study, the global Moran's I index was selected, and with the support of the GeoDa software platform, the agglomeration characteristics of the ecological environment quality index in 2000, 2010, and 2018 were respectively analyzed. The calculation formula is as follows (5):  In the formula, I represents Moran's I index; X i and X j represent the mean value of the ecological environment quality of the ith and jth grids; W ij refers to the spatial weight matrix; and S represents the sum of the elements of the spatial weight matrix.
Based on the calculation of the global Moran's I index (Bai et al. 2021), the Moran scatter plot was obtained, and the ecological environment quality index was further divided into 5 different types, namely, high-high aggregation area (H-H), high-low aggregation area (H-L), low-high aggregation area (L-H), low-low aggregation area (L-L), and not significant (not significant). The specific definitions are given in Table 6.

Geodetector
As proposed by Wang Jinfeng, Geodetector is a new statistical method for detecting spatial differentiation and revealing the driving factors ). In the present study, a factor detector was used to analyze the causes of the ecological environment quality of the urban agglomeration on the northern slope of the Tianshan Mountains. The factor detector could detect whether a factor was the cause of the spatial and temporal distribution pattern of the ecological environment quality and to what extent said factor explains the space of the ecological environment quality. The specific method involved using EQI as the dependent variable, taking the selected 11 indicators as independent variables, using the natural breakpoint method for stratification, and converting the numerical value to the type value. In ArcGIS, the Create Fishnet tool was used to construct the fishing nets of 5 km × 5 km to cover the entire study area, and a total of 5275 fishing nets were uniformly generated. Subsequently, the dependent variable values and the independent variable values were matched through the fishing nets to detect the influence of each factor. For the available Q value measurement, the expression is as follows: where h = 1,…, L is the stratification (Strata) of variable Y or factor X, that is, classification or division; N h and N are the number of units in layer h and the whole area respectively; and 2 are layers, respectively, of the variance of h and the Y value of the whole area. The value range of p is [0,1], with a larger value showing more obvious spatial differentiation of Y. Meanwhile, if the stratification is generated by the independent variable X, the larger the value of q, the greater the influence of the independent variable X on the attribute Y and vice versa.

CA-Markov model
Cellular automata (CA) is a mathematical model that can simulate the spatiotemporal evolution process of a highly complex system ). The model is: In the formula, S is the state of the ijth cell; t and t + 1 are moments; f is the transfer function; and q is the neighborhood.
The basic principle of Markov is to realize the simulation prediction of the future development status by using the experience transfer probability of the existing discrete state of the system. If there is Markov property in the change process of a system, Se is the state at the initial moment, then the state after e cycles can be defined as: Table 6 The connotations of different Moran clustering models

Clustering types Connotation
High-high clustering (H-H) The spatial agglomeration characteristics of the region's own ecological environment and the surrounding level are high High-low clustering (H-L) The region's own ecological environment is of high quality, but the surrounding area has low spatial agglomeration characteristics Low-high clustering (L-H) The fragility of the ecological environment of the region itself is low, but the surrounding area has high spatial agglomeration characteristics Low-low clustering (L-L) The spatial agglomeration characteristics of the region's own ecological environment and the surrounding level are low No significant There is no significant spatial agglomeration feature where S e is the state after e cycles; e is the number of cycles; and P e is the system experience transfer probability matrix. The present research was based on IDRISI17. 0 software, which is used as a tool to calculate the probability matrix based on the data of 2000 and 2010. In the software, 2010 was the base period, the number of iterations was set to 8, the filter was 5 × 5, and the proportional coefficient was set to 0.15. The spatial pattern of the ecological environment quality in 2018 was simulated and compared with the actual ecological environment quality status in 2018. The CROSSTAB module was used to complete the calculation and accuracy verification of kappa coefficients of real and simulated results. In the same way, the simulation forecast was completed for the region in 2026.

Spatial distribution
The results of the present study combined with the analysis of the spatial distribution characteristics (Fig. 4) show that the overall ecological environment vulnerability of the urban agglomeration on the northern slope of Tianshan Mountains gradually increased from southwest to northeast, and the quality of the ecological environment in the southwest was relatively lower than that in the northeast.
At the same time, the spatial distribution of the ecological environment quality of each level of the urban agglomeration also exhibited obvious regional differences. In general, the quality of the ecological environment has gradually improved. In 2000, the areas below the medium-level ecological environment had the widest spatial distribution, accounting for about 73.2% of the entire area. Combined with the type of land use, such areas can be regarded as the core residential area and the main concentrated area of construction land for the social and economic development of urban agglomerations. In 2010, the ecological environment (13) S e = S 0 P e was relatively improved compared to the ecological environment above the middle level in 2000. Areas above the middle level accounted for about 35% of the entire area. Most of said areas were concentrated in the middle mountainous areas on both sides of the river. The population density was relatively low and human activity was not obvious. Combined with the analysis of topographic data, an important buffer zone and key monitoring and prevention area for regional ecological environmental protection was identified. The main ecosystem is relatively complex and diverse, and the land landscape is dominated by mediumcoverage grassland. Compared with 2010, areas with an excellent ecological environment improved significantly by 2018. Such improvement could be mainly attributed to the implementation of China's "returning farmland to forest" and "returning farmland to grass" and other policies and ecological projects. At the same time, people's awareness of environmental protection and ecological protection has been strengthened.

Structural features
Statistical analysis of the grid ratios of different levels of ecological environment quality can help to further reveal the differences and change characteristics of the regional ecological environment quality. Through analyzing Fig. 5, there were obvious differences in the structure distribution of the grid ratios of each level. The ratio in 2000 was Better > Moderate > Good > Poor > Bad > Extremely bad > Excellent, the scale relationship in 2010 was Moderate > Better > Good > Poor > Bad > Extremely bad > Excellent, and the scale relationship in 2018 was Better > Good > Moderate > Poor > Excellent > Bad > Extremely bad. Meanwhile, from 2000 to 2018, the proportion of grids above the intermediate level accounted for more than 65% of the entire area. Objectively, the overall ecological environment quality of the area continued to be at a medium level during the study period. The proportion of high ecological environment quality index exhibited a gradual increase, reflecting that the

Spatial aggregation mode of ecological environment quality
Using GeoDa and GIS software, the Queen spatial weighting method was used to conduct spatial autocorrelation analysis of the ecological environment quality. To quantitatively describe the spatial dependence of elements, Moran's I was used to determine whether a variable was spatially correlated and the degree of correlation.
An observation can be made from Fig. 6 that the regional ecological environment quality was concentrated, with fewer discrete points, showing a continuous distribution trend. The scattered points were mainly distributed in the first and third quadrants. The Moran's I index from 2000 to 2018 was higher than 0.85, indicating that the ecological environment quality of the study area had a significant positive spatial correlation. In other words, the ecological environment quality presented a significant clustering state of high-value areas close to each other, and low-value areas close to each other.
The analysis of the local spatial autocorrelation spatial distribution map (Fig. 7) shows that the ecological environment quality of the urban agglomeration on the northern slope of the Tianshan Mountains roughly presented a "banded" distribution pattern, with obvious spatial clustering characteristics and uniform distribution. On the whole, the ecological environment of the urban agglomeration on the northern slope of Tianshan Mountains. There were two types of spatial correlation patterns in environmental quality, namely high-high aggregation type (H-H) and low-low spatial aggregation type (L-L). For the characteristics of the spatial distribution of the area, the southeast area was mainly high-high agglomeration area, the low-low agglomeration type was mainly in the central and northwestern areas of the urban agglomeration on the northern slope of Tianshan Mountains, and most of the areas were located in the lower elevation area on the northern slope of Tianshan Mountains, desert area and low mountain and hilly area. The quality of the ecological environment in such areas had positive spatial autocorrelation. The eco-environmental quality of the areas was significantly affected by the eco-environment of the surrounding areas. In the remaining areas, the agglomeration was not obvious, the spatial autocorrelation was not obvious, and the ecological environment quality was randomly distributed. That is, the ecological environment quality had minimal impact on the surrounding areas. When conducting ecological management, priority can be given to areas with high spatial autocorrelation, so as to improve work efficiency and quickly improve the quality of the ecological environment.

Influencing factors of ecological environment quality
The ecological environment quality of the urban agglomeration on the northern slope of the Tianshan Mountains is inseparable from the natural ecology, social ecology, and economic ecology. Fig. 8 shows the relationship between NEI, SEI, EEI and EQI. Points with better ecological conditions were mainly distributed at the top of the three-dimensional scatter plot, which were clustered in areas with large SEI values and high NEI. Points with weaker ecological conditions were located at the bottom where the pooled EEI value was higher, and the NEI value was lower.
An observation can be made from Fig. 9 that the high value of NEI was mainly concentrated in the middle andlower reaches of the artificial oasis area, and the low value was mainly concentrated in the lower reaches of thedesert, with the junction of the desert and the vast area to the north being significantly scarce. The spatial distribution of the SEI value was relatively fragmented, and the high value was mainly concentrated in the woodland, grassland and water area of the artificial oasis. The low value was mainly located in the urban construction land, desert area and low mountain and hilly area, while the high value of EEI was mainly concentrated in the economically developed southeast area. Regarding the entire urban agglomeration on the northern slope of the Tianshan Mountains, the industrial structure has always been in the order of "two-three-one." In the tertiary industry, Urumqi, Shihezi, Kuytun, and Changji have a better foundation, among which Shihezi has stronger competitiveness. The tertiary industry in Fukang is a fast-developing industrial sector with a poor foundation. Karamay and Usu have a poor tertiary industry foundation and weak competitiveness. Energy and chemical machinery manufacturing and electric power are the main industries of Karamay, Urumqi, Changji, Fukang, and other cities in the urban agglomeration on the northern slope of the Tianshan Mountains, but an orderly gradient level has not been formed for the industrial division of labor.
In order to further analyze the internal reasons for the changes in the ecological environment quality of the urban agglomeration on the northern slope of the Tianshan Mountains, a factor detector was introduced into the geographic detector as an analysis tool to reveal the internal driving factors of the ecological environment quality. The q value indicates the degree of influence of the factor on the ecological environment, while the P value indicates whether the factor passes the significance test level. As shown in Table 7, the results of the geographic detector show that from the overall ranking of 11 indicators according to the q value, the natural ecology occupied a dominant position. Among the indicators, the vegetation coverage and annual precipitation were both greater than 0.5. The natural ecological spatial distribution of the urban agglomeration on the northern slope of the Tianshan Mountains was quite different. The vegetation coverage, altitude gradient distribution, soil type, precipitation status, and water resource reserves of different regions were relatively different. The impact of social ecology was the second largest, among which, the q value of land use types were all above 0.54, and the impact was relatively high, mainly because the urban construction land and large-scale desert area have brought economic development to the urban agglomeration on the northern slope of Tianshan Mountains. The urban agglomeration on the northern slope of the Tianshan Mountains is deep inland, with sparse precipitation, strong evaporation, and Fig. 7 Spatial agglomeration of ecological environment quality sparse vegetation. The vegetation coverage is mainly low and medium-low. Affected by topographical factors, river flow, flow direction, oasis distribution and human activities, the different levels of ecological environmental quality indexes were roughly distributed in a northwest-southeast direction. The regional population is concentrated in the oasis, and there was minimal human activity in the mountains and deserts of the region. The quality of the ecological environment was relatively stable. The stable pattern of the change map mostly overlapped with the regions with low ecological environment quality index, which also verifies the rationality of the EQI construction from the side. The EQI was established in the order of the importance of natural factors>social factors>economic factors in the initial stage of construction, which is consistent with the actual situation, and the results are true and reliable (Fig. 9).
During the changes from 2000 to 2018, the influence of vegetation coverage has gradually increased, and the increase  Table 7 Geographical detector results on 11 impact factors of eco-environmental quality in vegetation coverage has been a significant factor in the restoration and management of the ecological environment. The local government should take the initiative to improve policies and measures related to ecological environment protection, implement the responsibilities of all parties, and guide and encourage local residents to increase their awareness of environmental protection. For urban expansion and economic development zones, the ecological environment and socioeconomic development should be coordinated. For the oasis and the desert Gobi area where the ecological environment has been damaged to a certain extent, the cause needs to timely determined and remediated as soon as possible. The ecological environment is the foundation and necessary condition for people's survival and development, and requires due attention.

Forecast results
The ecological environment quality classification raster data of 2000, 2010, and 2018 were imported into IDRISI, and the ecological environment quality transfer area matrix and transition probability matrix from 2010 to 2018 were obtained through the Markov module. The 2018 forecast was predicted through the CA-Markov model. The ecological environment quality spatial distribution map was plotted and compared with the actual 2018 ecological environment quality distribution map, and the calculated kappa value was 0.8972. The kappa coefficient value was between 0.75 and 1, indicating that the simulation effect was good. The model could be applied to the spatial simulation of ecological environment quality changes in the urban agglomeration on the northern slope of Tianshan Mountains. At the same time, in order to further study the future development of the fragility of the ecological environment in the area, a simulation forecast of the situation in 2026 was conducted (Fig. 10). Through the analysis of the prediction simulation results in 2018, an observation can be made that the spatial distribution characteristics were basically consistent with the real results. As shown in Fig. 11, the proportions of each grid in 2018 were 8. 24%, 10.63%, 17.35%, 16.02%, 19.16%, 17.31%, and 11.29%. Comparing the true proportions of each category, the accuracy of the predictions of each category presented a relatively consistent state of change. The accuracy test shows that the CA-Markov model not only had a higher overall prediction accuracy when realizing the prediction and analysis of the ecological environment quality in the region, the prediction accuracy of each category was also high. The prediction results in 2026 show that the urban agglomeration on the northern slope of Tianshan Mountains as a whole exhibited a gradual improvement in the quality of the ecological environment, and the grid ratios of each category were 11.07%, 7.01%, 12.05%, 15.86%, 22.07%, 16.77%, and 15.17%, respectively. Compared with the above-average level of ecological environment in 2018, the proportion of ecological environment grades exhibited an increasing trend, while the ecological environment quality below the intermediate level exhibited a decreasing change trend.

Trend changes
As shown in Fig. 12, Formula (8) was used to calculate the EEQI value of the area in 2000, 2010, and 2018, which were 5.7392, 6.1856, and 6.4366, respectively, showing a gradually increasing trend. As such, the ecological environment of the basin area in the past 20 years was objectively characterized. At the same time, the EEQI of the area is predicted to be 6.6285 in 2026, which also exhibited an increasing trend. Some areas with extremely poor ecological environment had an increasing trend, which may be due to the harsh natural environment in the northern slope of the Tianshan Mountains and the arid climate conditions in the area. Thus, the deterioration of the ecological environment was exacerbated; however, the number of areas where the ecological environment was improving gradually increased, which indicates that with the effective implementation of a series of environmental measures, the overall regional ecological environment will continue to develop in the direction of continuous improvement.

Discussion
The continuous changes in human activities and climate change have significantly affected and altered the ecological environment at different scales. Therefore, how to effectively assess and

Significance of the EQI
An ecological environment quality (EQI) evaluation index system was established to evaluate and detect the ecological quality and spatiotemporal changes of the urban agglomeration on the northern slope of the Tianshan Mountains from 2000 to 2018. It was found that the ecological quality of the urban agglomeration on the northern slope of the Tianshan Mountains is currently facing great challenges, suggesting the importance and urgency of ecological protection of oasis in arid areas. The results of the EQI-based ecological quality evaluation provide important and effective guidance for urban management and ecological protection of urban agglomeration on the northern slope of Tianshan Mountains. The results show that the EQI system based on spatiotemporal big data, quantitative analysis mathematical methods, and GIS and RS remote sensing technology for the urban agglomeration on the northern slope of the Tianshan Mountains is an effective method for ecological environment evaluation. Although the ecological environment of some regions in the urban agglomeration on the northern slope of the Tianshan Mountains fluctuated between 2001 and 2018, the types of changes showed a trend of improvement, and the area and percentage of poor ecological quality have been decreasing over the past two decades. This may be influenced by the policies of "returning farmland to forest" and "returning farmland to grassland" and the active implementation of ecological engineering, which have improved the EQI score to a certain extent. This is due in large part to the expansion of urbanization, which inevitably leads to an increase in areas with poor ecological quality. In addition, the highest proportion of moderate and good ecological quality levels is found in the urban agglomeration on the northern slope of the Tianshan Mountains. According to the characteristics of the ecological environment quality of the urban agglomeration on the northern slope of the Tianshan Mountains, we suggest that decision makers have a comprehensive and clear understanding of the spatial distribution of ecological quality, and strive to protect and maintain areas with good ecological quality, mainly including forests and grasslands. Areas with moderate and moderate levels of ecological quality need to be monitored for ecological changes to prevent deterioration, and areas with poor ecological quality are in urgent need of restoration and improvement, mainly in core areas with a high degree of urbanization.

Importance for ecological protection
Eco-environmental quality assessment is to evaluate the ecological environment quality assessment status of the study area by obtaining the spatial distribution information of each major factor, including basic geology, water resources, land resources, environment, ecology and environment, and other existing factors. With eco-environmental quality assessment, we can predict the development trend of ecological environment quality, evaluate the impact of human activities on the environment, and propose measures to rationally develop and utilize resources and protect the ecological environment. The ecological environment quality system of urban agglomeration on the northern slope of Tianshan Mountain is an effective method for ecological environment assessment based on spatio-temporal big data, quantitative analysis mathematical method, and GIS and RS remote sensing technology. Taking Chongming Island, a typical estuarine island, as an example, a comprehensive evaluation system of ecological vulnerability of estuarine island was established based on the pressure-state-response (PSR) conceptual model, and the spatial and temporal distribution characteristics of ecological vulnerability of estuarine island were discussed from 2005 to 2015 (Peng et al. 2021). This research takes chongqing wansheng economic development zone as an example, the urban planning from the land, the hydrology, resources, ecological and geological environment five aspects of selecting 20 evaluation indicators, using the analytic hierarchy process (AHP) to determine the weights of evaluation indexes, analyzes the geological environment quality present situation from 2007 to 2019, using fuzzy comprehensive evaluation method of urban geological environment quality evaluation (Lu et al. 2021).
Based on a series of Landsat images, the improved remote sensing ecological index was used to monitor and evaluate the ecological environment quality of Yangquan Coal mine in Shanxi Province from 1987 to 2020. The temporal and spatial evolution law of ecological environment quality was quantitatively evaluated quickly and accurately, which is of great significance to the ecological restoration and development planning of coal mine area (Nie et al. 2021). Primary productivity, land surface temperature, land exposure, and vegetation coverage were selected to reflect the ecological environment quality of Wuzhong District, Suzhou. The spatial principal component analysis method was used to construct the ecological environment comprehensive index to characterize the temporal and spatial variation characteristics of ecological environment quality in the red line area of Wuzhong District (Jia et al. 2017). Based on the modified remote sensing ecological index (MRSEI) retrieved from Google Earth Engine (GEE), we evaluated and analyzed the ecological and environmental quality of the Qaidam Basin from 1986 to 2019, combining meteorological and socioeconomic auxiliary data (Jia et al. 2021).
In contrast, our method has the ability to assess the ecological environment system of the urban agglomeration on the northern slope of the Tianshan Mountains from a more comprehensive perspective. Its advantages are: on the one hand, it takes into account the multiple pressures brought about by social changes (such as economic development, human influence, etc., combined with multiple status and response indicators). In addition, the method can also provide spatial ecological environment information. More importantly, since the urban agglomeration on the northern slope of the Tianshan Mountains is an arid area with harsh native environment, there are a large number of desert Gobi around the urban agglomeration. The resources mainly come from the melting water of ice and snow in summer in the Tianshan Mountains in the south, and the unreasonable development and utilization lead to the increasingly serious regional salinization phenomenon, which seriously affects the ecological environment of the region and restricts the management and restoration process of the ecological environment in the study area. At the same time, driven by my country's Belt and Road Initiative and the Western Development, the regional economy has developed rapidly, the level of urbanization has been continuously improved, the impact of urban built-up areas on the environment has increased sharply, and high-intensity human activities have exacerbated the changes in EQI. With the continuous improvement of the level of urbanization and the increasing emphasis on environmental protection in my country, economic construction with high intensity, low production capacity and serious pollution has been suspended. Before the increasing awareness of human environmental protection, human interference activities tend to improve the direction of the ecological environment. Development, to a certain extent, can improve the ecological environment of the region.
In contrast, our method has a more comprehensive perspective to evaluate the eco-environmental system of the urban agglomeration on the Northern Slope of the Tianshan Mountains. Its advantages are as follows: on the one hand, it considers multiple pressures brought by social changes (such as economic development and human influence, etc., and combines multiple state and response indicators). In addition, the method can also provide spatial ecological environment information. More importantly, by putting forward a comprehensive evaluation system, we can understand the spatial and temporal distribution of ecological environment quality from the perspectives of natural ecology, social ecology, and economic ecology, which is helpful for managers and decision makers to carry out protection planning and environmental management. At the same time, the ecological environment development of the study area in 2026 was simulated and predicted to provide reference for urban agglomeration development and protection and ecological civilization construction.

Research deficiencies and prospects
Some limitations and future works need to be discussed. The ecological environment is a complex, comprehensive and fuzzy aggregate, but the current evaluation methods and evaluation index system cannot be scientifically and comprehensively evaluated. This research still has some shortcomings in the selection of indicators. From the perspective of natural factors, the influence of factors such as groundwater, soil moisture, and windy days is not considered; from the perspective of human factors, this research only considers population, economy, and land. The impact of utilization methods on the ecological environment of the urban agglomeration on the northern slope of the Tianshan Mountains. Human disturbance activities include many aspects. Environmental issues such as mining intensity and industrial and agricultural development have certain impacts on the regional ecological environment. In future research, analysis should be conducted based on the specific conditions of the study area and comprehensive consideration of various ecological influencing factors. There is still a need for further exploration to establish a reasonable, scientific, and comprehensive evaluation index system for ecological environmental quality.

Conclusion
1. Based on RS and GIS, the ecological environment quality of the urban agglomeration on the northern slope of Tianshan Mountains was evaluated on a macro scale. Generally speaking, the ecological environment quality of the urban agglomeration on the northern slope of Tianshan Mountains exhibited a trend of improvement as a whole, and there was a strong positive space. Further, there was a certain internal connection, and the spatial distribution exhibited cluster characteristics, and not random distribution. Areas with poor ecological environment quality were mainly distributed in unused land areas such as bare land, deserts, and Gobi. Meanwhile, areas with good and excellent ecological environment quality were mainly concentrated in oasis areas and the middle and lower reaches of rivers. Natural ecological factors had a dominant influence on the area, where the q value of vegetation coverage, soil type, altitude, and annual precipitation was greater than 0.5, which had a greater impact; the second was social ecological factors, and the q value of land use type was 0.540. The degree of influence was relatively high, while the influence of economic and ecological factors was the least. Regression analysis was conducted on the selected sample points and a CA-Markov model suitable for evaluating and predicting the quality of the ecological environment was established. The prediction results reveal that the overall ecological environment quality of the region will continue to improve. 2. The research results show that the combination of RS and GIS for ecological environment vulnerability assessment had high accuracy, and the CA-Markov model was feasible for realize future development forecasting. At the same time, unlike previous studies, the exploration of long-term changes in the results of multiple periods was achieved in the present study, in addition to prediction and analysis of the future development of the regional ecological environment.