The ecological protection and sustainable development of Urumqi have become an important part of the high-quality growth of the urban agglomeration on the northern slope of Tianshan Mountain. Under the impacts of multi-source factors, the ecological landscape pattern of Urumqi has changed due to it being in a fragile eco-environment, so an ecological network is desperately needed to enhance ecological security patterns. Taking Urumqi city as the study area, the ecological risk evaluation model and the minimum cumulative resistance model were integrated to analyze the spatial and temporal features of landscape ecological risk from 2000 to 2020, and the future land use simulation model was used to predict the ecological risk pattern of Urumqi in 2030, construct a landscape ecological network, and propose ecological security protection strategies. Since 2000, land use in Urumqi has undergone drastic changes: the built-up land area has increased significantly, the landscape has diversified, and landscape fragmentation has shown a decreasing trend from the main urban area as the core to the urban fringe. The high-risk landscape ecology shows a decreasing trend from east to west, mainly in the bare land areas with sparse vegetation, whereas the risk is relatively low in woodland, arable land, and built-up areas. The change of risk in the study area is mainly influenced by the typical defective factors of oasis cities such as urban expansion, land desertification, and sparse vegetation. The landscape ecological network is mainly located in the southwest, central, and east of the study area, whereas there is no corridor distribution in the north and southeast, which is mainly caused by the special geographical location and climatic conditions. The ecological network mainly consists of 10 ecological sources and 10 ecological corridors and proposes conservation strategies for the optimization of the landscape pattern and for the construction of the ecological security pattern in Urumqi, providing a guide for the improvement of ecological security.

It has become undeniable that global land surface temperature (LST) has continued to rise in recent years. The threat of extreme heat to humans has become self-evident, especially in arid regions. Many studies have clarified the temperature rise/fall mechanism of LST from the perspective of influencing factors. However, there are few studies on mitigating LST from the standpoint of regional networks. This paper first combines Landsat 8 with Sentinel-2 imagery for LST downscaling based on the Google Earth engine as a way to match local climate zone (LCZ) with 17 classification types. Then, the thermal environment resistance surface is constructed according to LCZ, and the essential cold sources are identified using morphological spatial pattern analysis (MSPA) and circuit theory to form the thermal environment green corridor and obtain the pinch point and barrier point areas. The results show that (1) The downscaling of LST based on random forest (RF) for the Urumqi–Changji–Wujiaqu metropolitan area has an R² of 0.860 and an RMSE of 3.23, with high downscaling accuracy. (2) High temperature (HT), medium temperature (MT), and low temperature (LT) have the largest proportions in the study area; HT dominates in Urumqi, LT in Changji, and MT in Wujiaqu. (3) The natural types (LCZ-D, LCZ-C, and LCZ-F) in the LCZ classification occupy a large area, and the building types are mainly concentrated in Urumqi; LCZ-D, LCZ-G, and LCZ-A contribute the most to the cooling of LST, and LCZ-F, LCZ-C, and LCZ-10 contribute the most to the warming of LST. (4) After identifying critical cold source patches according to MSPA to arrive at 253 green corridors, subsensitive corridors and sensitive corridors need to take certain measures to prevent corridor blockage; pinch point areas, as well as barrier point areas, need to be protected and repaired according to their respective characteristics. In summary, corresponding cooling measures to specific areas can improve the connectivity between cooling sources and slow down the temperature increase of the whole area. This study and experimental approach can provide new insights for urban planners and climate researchers.

Assessing how land-use changes will affect water-producing ecosystem services is particularly important for water resource management and ecosystem conservation. In this study, the InVEST model and geographical detector were used to assess the water ecosystem service functions of the Ebinur Lake Basin and analyze their relationship with land-use changes. The results show that in the past 25 years, the water yield of the study area showed a trend of a strong yield at first and then a weaker one; there was a relatively large water yield in the west and southeast regions of the basin. The order of water yield for different land-use types is as follows: forest land > grassland > water area > unused land > crop land > construction land. After 2010, the output load of nitrogen and phosphorus increased; thus, the water purification ability weakened. The main land-use types in areas that demonstrate a large change rate in water purification capacity in the basin are cultivated land and construction land. Changes in the two water ecosystem services were associated with land-use changes. Geodetector analysis results further validated this conclusion. This study proposes a viable, replicable framework for land-use decisions in ecologically fragile watersheds. This study not only helps to gain insight into urban growth patterns in the study area but also helps to inform different land-use stakeholders.