3.3.1. Analysis of Landscape Pattern Index Change on Patch Type Scale

As shown in Figure 5, the PD of the cultivated land is the largest among all land types, with a mean value of 0.24. This shows that the cultivated land has a profound influence on the landscape pattern of the Nanming River Basin. From 2000 to 2010, the PD of the forest, cultivated land, and constructed land increased. However, the PD of grassland showed a decreasing trend from 2010 to 2020. Except for the constructed land, the PD of the forest, constructed land, grassland, and water showed an increasing trend. The human ecological footprint has had a greater impact on the forest and cultivated landscapes over the past 20 years. Moreover, their ecological processes were more active, in which the expansion of the constructed landscapes was reasonably restrained during 2010–2020.

Among the landscapes in the watershed, the mean of the PLAND of the forest is the maximum of approximately 45.08%, which is about 28 times that of the smallest PLAND (waters), indicating that the forest offers more advantages in the landscape. This result agrees with the results of the spatial characterization of land use described above. From 2000 to 2010, the forest area and constructed land patches showed an increasing trend. From 2010 to 2020, the PLAND of the constructed land increased by 19.53%, which was 12% more than that of 2000. Nevertheless, the constructed land patches experienced continuously increasing trends during the study period. Since 2001, Nanming River Basin experienced comprehensive environmental improvement, which includes increasing projects on cultivated land, forest, and the construction of a comprehensive landscape on both sides of the river. With a rise in urbanization, a large construction continues to expand to the periphery of the city. However, patches of cultivated land and grassland continue to decline.

**Figure 5.** Change of landscape pattern index on patch type scale from 2000 to 2020.

From the LPI, the LPI of the forest land far exceeds that of other land types, indicating that forest is the major land substrate in the watershed. From 2000 to 2020, the LPI of constructed land continues to increase.

From the COHESION, the patch cohesiveness of the forest and constructed land is higher, indicating that the natural connectivity of these two types of land is efficient, and the distributions are patchy in spatial distribution. From 2000 to 2020, the COHESION of cultivated land, grassland, and waters generally declined, whereas the COHESION of constructed land continued to increase, reaching 35% from 2000 to 2010, which was three times that of 2010–2020. With large-scale land exploitation, the distribution of patches of cultivated land, grassland, and water fragments go through aggregation to fragmentation, with a decrease in the natural connectivity of the landscapes. However, the patches of constructed land gradually turn into aggregated blocks of spatial distribution.

3.3.2. Analysis of the Landscape Pattern Index Change on Landscape Scale

As shown in Figure 6**,** the PD and CONTAG have a more pronounced geographical variability. The CONTAG increased significantly in the southwestern part of the basin, and the PD was high in most of the northeastern part of the basin. Moreover, the spatial distribution of the CONTAG in these areas showed the opposite PD characteristics. We found that the high-value areas of CONTAG are concentrated in the urban center of Guiyang City, owing to its simple landscape structure, its constructed land as the matrix land, and a high degree of agglomeration, which greatly reduces the fragmentation of the

landscape patches. Moreover, the low-value areas comprise the interspersed distribution of cultivated land, constructed land, and grassland, consisting of an intricate landscape structure with a high degree of landscape fragmentation in the watershed. The lower stream of the basin, at Wudang and Longli, has a lower level of landscape diversity, indicating that human activities in this area are infrequent, and the development is low. The SHEI and SHDI in the basin have similar spatial variations. In the southern part of Nanming District and the eastern part of Guanshan Lake District and Yunyan District upstream, the SHEI and SHDI have decreased significantly with a rise in urbanization and an increase in the project construction, which have become the main advantages of the landscape.

**Figure 6.** Spatial distribution of landscape pattern index at landscape level in the Nanming River Basin from 2000 to 2020. Note: (**A1**) PD in 2000; (**A2**) PD in 2010; (**A3**) PD in 2020; (**B1**) CONTAG in 2000; (**B2**) CONTAG in 2010; (**B3**) CONTAG in 2020; (**C1**) SHEI in 2000; (**C2**) SHEI in 2010; (**C3**) SHEI in 2020; (**D1**) SHDI in 2000; (**D2**) SHDI in 2010; (**D3**) SHDI in 2020.

#### *3.4. Landscape Ecological Safety Evaluation*

#### 3.4.1. Landscape Ecological Safety Changes of Nanming River Watershed

The ecological safety index of the Nanming River Basin was divided into five classes (refer to related studies regarding each grade [55]). As shown in Figure 7 and Table 1, the overall landscape ecological safety shows a continuous positive trend in the last 20 years. The medium ecological safety zone is the largest area. The low-security area is concentrated in the southwest part of the watershed, which is characterized by a farming economy, underdeveloped economy, extensively cultivated land, and interspersed distribution of grassland, forest, and waters, which destroys the stability of the landscape. Driven by urbanization, a large patch of arable land is used for infrastructure construction to improve the standard of living. The construction of residential housing is increasing on the large arable land, leading to a rise in the landscape ecological risk. However, the low-security areas decline by about 25% in 2020. The shrinkage rate in the lower security zone was 11.13%. Spatially, the constructed land is connected to patches. The patches also converge and shift to the middle-security zone, revealing a gradual increase in ecological security. The higher security zone is increasing yearly, which is mainly concentrated in the northern part of the watershed, showing that the forest offers advantages to the landscape. Despite the recent government's policy of encouraging people to return to cultivated land, the forest has not yet been encroached upon by other landscape types, providing policy support to maintain a higher security-level state. The trend of a high-security zone is increasing, which is distributed in the center of Guiyang City and at a lower elevation at the watershed. With the obvious landscape advantages and contiguous urban housing, this area has become a stable landscape structure and has low landscape fragmentation.

**Figure 7.** (**a**) Spatial distribution of landscape ecological security in 2000; (**b**) Spatial distribution of landscape ecological security in 2010; (**c**) Spatial distribution of landscape ecological security in 2020.



3.4.2. Spatial Autocorrelation Analysis of Landscape Ecological Security Index

As shown in Figure 8, the Moran's I values of the landscape ecological safety index of the Nanming River Basin were 0.394, 0.464, and 0.488, greater than 0 for the study periods 2000, 2010, and 2020 at a significance level of *p* < 0.05. This indicates that the landscape ecological safety index in the study zone is correlated, and the spatial convergence is gradually increasing.

**Figure 8.** Scatter map of ecological security index of landscape pattern in the study area from 2000 to 2020.

As shown in Figure 9, the spatial clustering pattern of landscape ecological safety values in the Nanming River basin is characterized by high–high clustering and low–low clustering. The percentage of high–high concentration sample areas gradually increased over the three periods, ranging from 7.66% to 12.66%. From the local autocorrelation of the study area, the cluster structure of "high–high" values of the landscape ecosystem security index continues to extend outward from 2000. However, the range of "low–low" values continues to shrink. In terms of spatial distribution, the ecological safety high-value catchment area of the watershed is concentrated in the center of Guiyang City, with little distribution in the eastern part of the downstream Wuzhong. In this area, the terrain is relatively flat, and the topographic conditions are simple, with a single land-use landscape type as the main feature. Low-value ecological security catchment areas are focused on the central Huaxi and southern Pingba in the upper part of the watershed. At the same time, we found that the internal structure of the landscape in the adjacent areas of the region is finely fragmented. Additionally, each land use landscape type is disturbed by human activities and interspersed with each other; so the ecological safety is in a low-value state, and its stability may be difficult to maintain.

**Figure 9.** (**a**) LISA Map of Landscape Ecological Security in 2000; (**b**) LISA Map of Landscape Ecological Security in 2010; (**c**) LISA Map of Landscape Ecological Security in 2020.

#### **4. Discussion**

#### *4.1. Landscape Ecological Safety Evolution Rules*

(1) Change in the Landscape Ecological Safety Index

From 2000 to 2020, the landscape ecological security index showed an increasing trend, indicating that the ecological security of the Nanming River Basin gradually increased. Driven by the market economy, crop cultivation has increased in the upper watershed, making a certain amount of forest and grassland reclaimed as cropland. Moreover, a huge number of cultivated land has been converted to forest and constructed land, which was driven by the policy of returning farmland to forest and the construction of land expansion. This policy has increased the fragmentation of forest and cropland landscapes, decreased connectivity, and decreased the ecological security index. With urbanization and rapid economic development, the non-farm population has increased, and the disturbance to the watershed landscape from human activities has risen. This development enhances the distribution of the construction of landscapes favoring human ecological and living needs. Moreover, the contiguous distribution of constructed land decreases the fragmentation of landscape patches and increases the degree of stability because of the low vulnerability in the flat topography. Thus, the ecological safety index of the constructed land has maintained its maximum value in the last 20 years. Additionally, the implementation of artificial landscaping projects in the watershed has rationalized the layout of the grassland landscape and increased the landscape ecological safety index.

#### (2) Spatial Distribution of Landscape Ecological Safety Index

Table 2 shows a comparative analysis of the relevant literature [56–68]. In karst areas, the regional characteristics above the higher ecological safety level are manifested as follows: <sup>1</sup> They are located at lower elevations and in urban centers with faster economic development. <sup>2</sup> These areas are also the most concentrated belt of forest or the constructed land and forest of the landscape. Low-security and lower-security areas are characterized by cultivated land, grassland, and water. In non-karst areas, areas above medium security level are dominated by natural ecosystems such as forest and grassland, which are concentrated in agricultural areas or natural landscape protection zones far from urban centers. However, low-security level areas are distributed on constructed land, and natural landscape structures are fragmented by human interference.

Thus, the higher safety-level areas of the watershed are distributed in the central urban areas, whereas the constructed land is concentrated and contiguous. Moreover, the low and lower ecological safety areas are characterized by a distribution of interspersed and scattered forest, grassland, and cultivated land.

These results contradict results presented by previous studies conducted in non-karst areas. Compared with related studies in karst areas, the Panlong River Basin located in Kunming [60] has a social and geomorphic environment comparable to the study area. Lin et al. found that the ecological safety index of the constructed land within the Panlong River Basin was higher than that of other landscape types. Moreover, the landscape types formed by human behavioral activities have the characteristics of being the most resistant and stable to external disturbances. Taking the Dianchi watershed as a case study, Wu et al. analyzed ecological security and found that the high-security areas were located in the urban areas because urban housing was distributed in a row, with almost no other land use landscape. Moreover, the patches are highly connected and less fragmented, making them capable of resisting external disturbances [61]. These findings are in line with the results of this study.

From the above comparison of the landscape ecological security research in non-karst and karst areas, it is recommended to give more attention to the fallibility and fragility of natural ecosystems during the process of karst ecosystem restoration and reconstruction, such as forests and grassland. In non-karst areas, we should pay more attention to the uncontrolled expansion of constructed land and the quality of the surrounding natural ecosystem. In addition, in the follow-up study, the factors influencing the spatial distribution difference of ecological security of landscape patterns in non-karst and karst areas should be discussed in depth.

**Table 2.** Relevant study on ecological security of landscape pattern.


#### *4.2. Limitations and Shortcomings*

In interpreting land use type data through RS images and analyzing land use changes and landscape patterns in the Nanming River Basin, this study evaluated the spatiotemporal variation characteristics of ecological security in the basin. However, some shortcomings are found that require improvement. The results inevitably have certain errors when interpreting images through RS technology because the results have been influenced by objective factors and human subjective factors, thereby affecting the accuracy verification. Thus, the ground-based field data surveys and historical record data must be rectified to

improve the accuracy of the interpretation results. The dynamic and landscape indicators of land use types only reflect the macro-structural changes influenced by topographic factors [68]. However, it is difficult to reveal the microstructural changes in the landscape. Thus, the microstructural changes in the landscape should be explored in future studies.

As a natural geographical unit, a watershed is the unification of multiple catchment areas within a natural environment [69]. However, watershed boundaries and administrative boundaries cannot completely overlap [70]. Thus, when exploring the influence of the natural environment and anthropogenic activities on watershed landscape patterns and ecological security, these influences cannot be fully quantified because of some constraints on the analysis of watershed change.

#### **5. Conclusions**

This study used 3S technology, the landscape pattern index method, and spatial autocorrelation theory to systematically analyze the landscape pattern evolution characteristics of the Nanming River watershed and evaluate the ecological security of the karst watershed in the Guizhou plateau. Several conclusions were obtained as follows. Forest was the leading landscape and mainland substrate in the watershed from 2000 to 2020. The cohesiveness of constructed land patches continues to increase. Moreover, the fragmentation and diversity of landscape pattern in the patchy distribution of forest and constructed land have declined. The growth rate of the Nanming River Basin Landscape Ecological Safety Index has increased by 5.80%, and the overall ecological safety has shown a continuous positive trend. The high-value ecological safety clusters are distributed in the central urban areas, where the constructed land is concentrated and contiguous. Moreover, the low-value clusters show the scattered distribution characteristics of forest, grassland, and cultivated land. Thus, the spatial clustering effect of the ecological security index is obvious, which is dominated by high–high clustering and low–low clustering types. The study reveals the landscape pattern evolution rules of the typical karst watershed in the Guizhou plateau since 2001 by systematically evaluating the spatiotemporal distribution characteristics of ecological security in the watershed. The findings provide scientific reference for maintaining the ecological balance in the watershed, optimizing land resource allocation and regulation, and improving the ecological environment of typical karst watershed geomorphic units.

**Author Contributions:** Y.L.: Methodology, Investigation, Formal Analysis, Visualization, Writing— Original Draft, Writing—Review and Editing. H.G.: Data Curation, Resources, Formal Analysis. All authors have read and agreed to the published version of the manuscript.

**Funding:** This study was jointly supported by the Youth Talent Growth Project of Guizhou Provincial Department of Education (Qian Jiao He KY [2022] No. 202), the Water Conservancy Science and Technology Funding Projects in Guizhou Province (KT202114), the Water Conservancy Science and Technology Funding Projects in Guizhou Province (KT202223), National Natural Science Foundation of China (42261044).

**Institutional Review Board Statement:** Not applicable.

**Informed Consent Statement:** Not applicable.

**Data Availability Statement:** Not applicable.

**Conflicts of Interest:** The authors declared that they have no conflict of interest to this work. We declare that we do not have any commercial or associative interest that represents a conflict of interest in connection with the work submitted.

#### **References**


**Yu Zhang 1, Chaoyong Shen 1,2,3,\*, Shaoqi Zhou 1,\*, Ruidong Yang 1, Xuling Luo <sup>3</sup> and Guanglai Zhu <sup>1</sup>**


**Abstract:** Remote sensing image with high spatial and temporal resolution is very important for rational planning and scientific management of land resources. However, due to the influence of satellite resolution, revisit period, and cloud pollution, it is difficult to obtain high spatial and temporal resolution images. In order to effectively solve the "space–time contradiction" problem in remote sensing application, based on GF-2PMS (GF-2) and PlanetSope (PS) data, this paper compares and analyzes the applicability of FSDAF (flexible spatiotemporal data fusion), STDFA (the spatial temporal data fusion approach), and Fit\_FC (regression model fitting, spatial filtering, and residual compensation) in different terrain conditions in karst area. The results show the following. (1) For the boundary area of water and land, the FSDAF model has the best fusion effect in land boundary recognition, and provides rich ground object information. The Fit\_FC model is less effective, and the image is blurry. (2) For areas such as mountains, with large changes in vegetation coverage, the spatial resolution of the images fused by the three models is significantly improved. Among them, the STDFA model has the clearest and richest spatial structure information. The fused image of the Fit\_FC model has the highest similarity with the verification image, which can better restore the coverage changes of crops and other vegetation, but the actual spatial resolution of the fused image is relatively poor, the image quality is fuzzy, and the land boundary area cannot be clearly identified. (3) For areas with dense buildings, such as cities, the fusion image of the FSDAF and STDFA models is clearer and the Fit\_FC model can better reflect the changes in land use. In summary, compared with the Fit\_FC model, the FSDAF model and the STDFA model have higher image prediction accuracy, especially in the recognition of building contours and other surface features, but they are not suitable for the dynamic monitoring of vegetation such as crops. At the same time, the image resolution of the Fit\_FC model after fusion is slightly lower than that of the other two models. In particular, in the water–land boundary area, the fusion accuracy is poor, but the model of Fit\_FC has unique advantages in vegetation dynamic monitoring. In this paper, three spatiotemporal fusion models are used to fuse GF-2 and PS images, which improves the recognition accuracy of surface objects and provides a new idea for fine classification of land use in karst areas.

**Keywords:** spatiotemporal fusion; land use; high resolution; FSDAF; STDFA; Fit\_FC
