1 Introduction
Land use changes caused by human activities have led to the habitat loss and fragmentation at local, regional, landscape and global scales, which can hinder the migration and dispersal of species at gene, individual, population levels, further alter the structure and configuration of the landscape, and have become an important factor threatening biodiversity (Kruess and Tscharntke, 1994 ;Ewers and Didham, 2006 ;Fischer and Lindenmayer, 2007 ). Compounding the effects of habitat loss and fragmentation are key drivers of global biodiversity loss (Fahrig, 2003 ). Loss of natural habitats reduce the proportion of patches large enough to sustain a population over time, and make species highly dependent on connections between habitat patches (Baguette et al., 2013 ; Sahraoui et al., 2017 ).
Landscape connectivity is the degree to which the landscape facilitates or impedes movement among resource patches (Taylor et al., 1993 ;Tischendorf and Fahring, 2000 ). Connectivity determines the effective and availability habitat area within a territory for a given species (Saura and Pascual-Hortal, 2007 ). High Connectivity can facilitate the capacity movement to satisfy ecological requirements and reduce population isolation (Minor and Urban, 2008 ). An animal’s ability to utilize a resource patch is determined not only by the distance between patches, but also by the biophysical nature of the routes between two patches and the biology and behavior of the organism (Henein and Merriam, 1990 ). Some routes facilitate or allow unimpeded movement among patches; others impede to varying degrees the amount, or success, of movement. Improving landscape connectivity can promote species migration, dispersal, foraging, reproduction and survival, and reduce the risk of extinction. Hence landscape connectivity will obviously vary according to both organism properties and landscape features. Maintaining and improving landscape connectivity is today considered a key part of efforts to protect biodiversity (Taylor et al., 1993 ).
Preserving and restoring connectivity has become a major conservation priority on biodiversity conservation. Conservation practitioners have long recognized ecological connectivity as a global priority for preserving biodiversity and ecosystem function. Understanding broad-scale ecological processes that depend on connectivity, and making effective conservation planning decisions, requires quantifying how connectivity is affected by landscape features. Connectivity among populations and habitats is important for a wide range of ecological processes. The assessment of landscape functional connectivity needs to comprehensively consider the spatial distribution of habitat patches, species movement characteristics, and landscape context. But field observations are extremely labor intensive and fail to provide a complete understanding of functional connections at a large scale, and it is often difficult to assess landscape functional connectivity because of data limitations (Sahraoui et al., 2017 ;Petsas et al., 2020 ).
A graph-theoretic approach employs fast algorithms and compact data structures that are easily adapted to landscape-level focal species analysis. Graphs is built using GIS coverages to define habitat patches and determine the functional distance between the patches with least-cost path models. In landscape graphs, a graph is a set of nodes and links such that each link connects two nodes. Nodes represent sites of suitable habitat patches of a given species while links symbolize the potential ability of a species to directly disperse between two nodes. This method has been shown to be a powerful way of performing complex analysis of potential functional connectivity at relatively large scales (Bunn et al., 2000 ;Urban and Keitt, 2001 ;Jordán et al., 2003 ;Pascual-Hortal and Saura, 2006 ,2008 ). The advantage of a graph-theoretic approach is that it has relatively modest data requirements and improved from the initial results. Graph-theoretic approaches provide a new research idea for powerful leverage on ecological processes concerned with connectivity as defined by dispersal. Graph theory-based models and least-cost path models have gained increasing attention and are widely applied in connectivity modeling, conservation planning, and monitoring of land use changes and their environmental impacts (Minor and Lookingbill, 2010 ; Theobald et al., 2011 ; Piquer-Rodriguez et al., 2012 ; Rubio et al., 2012 ;Devi et al., 2013 ;Wang et al., 2016 ;Sahraoui et al., 2017 ;Liu et al., 2020 ;Machado et al., 2020 ).
The circuit theory modeling is an alternative approach to model gene flow and the dispersal or movement routes of organisms,and widely applied for assessing ecological functional connectivity across heterogeneous landscapes. This theory does not require new ways of representing landscape data; rather, it takes advantage of graph-theoretic data structures, and can be applied in graph-theoretic or raster GIS frameworks (McRae et al., 2008 ). Circuits are defined as networks of nodes connected by resistors (electrical components that conduct current) and are used to represent and analyze graphs. In circuit theory-based models, landscape is transformed into a cost map (called resistance surface), effective resistance, indicating landscape’s resistance to current flow between two nodes separated by a network of resistors, and is calculated through multiple pathways. Comparing to least-cost path models that measured along a single optimal pathway or corridor with the lowest cost, circuit theory integrates multiple possible pathways in a landscape into distance calculations and offers a measure of isolation assuming a random walk. The multipath results obtained from the circuit theory-based connectivity models make it more realistic and objective in practical application. By predicting net movement probabilities through nodes, a new landscape raster is created (called current map), where every cell is assigned a value that reflects the probability of an animal traveling from one node to another. Current density can be used to identify landscape corridors or “pinch points,” high current through a node indicates that removing it will have a high impact on connectivity (McRae and Beier, 2007 ; McRae et al., 2008 ). Circuit theory has already been shown to be useful for predicting movement patterns and probabilities of successful dispersal moving across complex landscapes, generating measures of connectivity or isolation of habitat patches, or protected areas, and to identifying important connective elements for conservation planning and designing robust reserve network (Epps et al., 2011 ; Howey, 2011 ;Walpole et al., 2012 ;Koen et al., 2014 ;Pelletier et al., 2014 ;McClure et al., 2016 ;Dickson et al., 2019 ;An et al., 2021 ;Hyseni et al., 2021 ;Leskova et al., 2022 ).
The North China leopard (Panthera pardus japonensis , commonly referred to as Panthera pardus fontanierii in the Chinese literature), is the most widespread subspecies of leopard and one of the rare and endangered species in China. Recent camera-trap surveys and other evidence revealed the presence of the North China leopard in Shanxi, Shaanxi, northern Hebei, Ningxia, and northern Henan. Most populations in these provinces are small, and occur in several isolated protected areas (Consolee et al., 2020 ). More and more attention has been paid to the protection of North China leopard. In Tieqiaoshan Provincial Nature Reserve, Shanxi Province, studies indicated that the North China leopard is in serious conflict with the locals, causing them personal economic losses due to lack of prey (Consolee et al., 2020 ; Vitekere et al., 2020 ). The North China leopard density was 4.23 individuals/100 km2, the population density increased with the distribution of wild boars (Vulpes vulpe s), and decreased with the distribution of roe deer (Capreolus capreolus ) (Zhu et al., 2021 ). Habitat environmental factors and anthropogenic interference also significantly affected the population density and spatial distribution of the North China leopard. Previous studies were mostly based on literature, expert and villager interviews, and camera-trap surveys to determine the current status and distribution of North China leopard, and most studies were conducted within protected areas. But there are few studies on the impact of land use changes on the ecological network connectivity of North China leopard at a large scale.
The Great Taihang Region is the dominant distribution area of North China leopard and contains the most densely distributed population of North China leopard in China (Laguardia et al., 2017 ;Cao et al., 2020 ). Due to the Human-Leopard conflict, North China leopard is mainly distributed in several isolated protected areas. In the past 30 years, with the implementation of a series of forestry projects, such as National Natural Forest Protection Program, Three-North Forest Shelterbelt Program and Grain for Green Project, great changes have taken place in the study area. Yet there is little insight concerning on the impacts of land use changes on the habitat ecological network connectivity of North China leopard. The present study, therefore, was an attempt to clarify the influences of land use changes on landscape connectivity for North China leopard. The main objectives of this study were: (1) to detect the land use changes in the Great Taihang Region from 1990 to 2020, (2) to clarify the influences of land use changes on landscape connectivity for North China leopard, and to answer the following research question, (3) how are the changes of the key barrier areas that affect the habitat ecological network connectivity of North China leopard? The purpose of this study was to explore the key barrier areas affecting the migration of North China leopard, and to provide the basis for this unique and rare wildlife species conservation planning in China.