Regionalization of Landscape Pattern Indices Using Multivariate Cluster Analysis

Regionalization, or the grouping of objects in space, is a useful tool for organizing, visualizing, and synthesizing the information contained in multivariate spatial data. Landscape pattern indices can be used to quantify the spatial pattern (composition and configuration) of land cover features. Observable patterns can be linked to underlying processes affecting the generation of landscape patterns (e.g., forest harvesting). The objective of this research is to develop an approach for investigating the spatial distribution of forest pattern across a study area where forest harvesting, other anthropogenic activities, and topography, are all influencing forest pattern. We generate spatial pattern regions (SPR) that describe forest pattern with a regionalization approach. Analysis is performed using a 2006 land cover dataset covering the Prince George and Quesnel Forest Districts, 5.5 million ha of primarily forested land base situated within the interior plateau of British Columbia, Canada. Multivariate cluster analysis (with the CLARA algorithm) is used to group landscape objects containing forest pattern information into SPR. Of the six generated SPR, the second cluster (SPR2) is the most prevalent covering 22% of the study area. On average, landscapes in SPR2 are comprised of 55.5% forest cover, and contain the highest number of patches, and forest/non-forest joins, indicating highly fragmented landscapes. Regionalization of landscape pattern metrics provides a useful approach for examining the spatial distribution of forest pattern. Where forest patterns are associated with positive or negative environmental conditions, SPR can be used to identify similar regions for conservation or management activities.     

Impacts of changes in land use and fragmentation patterns on Atlantic coastal forests in northern Spain

Changes in forested landscapes may have important consequences for ecosystem services and biodiversity conservation. In northern Spain, major changes in land use occurred during the second half of the 20th century, but their impacts on forests have not been quantified. We evaluated the dynamics of landscape and forest distribution patterns between 1957 and 2003 in Fragas do Eume Natural Park (northwestern Spain). We used orthoimages and a set of standard landscape metrics to determine transitions between land cover classes and to examine forest distribution patterns. Eucalypt plantations showed the greatest increase in area (197%) over time. Furthermore, transitions to eucalypt plantations were found in all major land cover classes. Forest showed a net decline of 20% in total area and represented 30% of the landscape area in 2003. Forest losses were mainly due to eucalypt plantations and the building of a water reservoir, while forest gains were due to increases in shrubland, meadows and cultivated fields which had been recolonised. Forest patch size and core area decreased, and edge length increased over time. In turn, increases were obtained in mean distance between forest patches, and in adjacency to eucalypt plantations and to a water reservoir. These results suggest an increase in forest fragmentation from 1957 to 2003, as well as a change in the nature of the habitat surrounding forest patches. This study shows that land use changes, mostly from eucalypt plantation intensification, negatively affected forested habitats, although some regeneration was ongoing through ecological succession from land abandonment.     

The dimensions of land use change in rural landscapes: lessons learnt from the GB Countryside Surveys

Rural landscapes are highly dynamic and their change impacts on a number of ecological processes such as the dynamics of biodiversity. Although a substantial amount of research has focused on quantifying these changes and their impact on biodiversity, most studies have focused on single dimensions of land use change. This lack of integration in land use change studies can be explained by the fact that data on the spatial, temporal, and ecological dimensions of land use are seldom available for the same geographical location. In this paper, the benefits of taking into account these three dimensions are illustrated with results derived from the Great Britain Countryside Surveys (CS), a large-scale monitoring programme designed to assess change in the extent and ecological condition of British habitats. The overview of CS results presented in this paper shows that (1) changes in land use composition will translate into a variety of spatial patterns; (2) the temporal stability of land use is often lower than can be expected; and (3) there can be large-scale shifts in the ecological condition of the land use types that form our rural landscapes. The benefits of integrated rural landscape studies are discussed in the context of other national monitoring programmes.     

PROCESS DOMAINS AND THE RIVER CONTINUUM1

ABSTRACT: The concept of process domains is proposed as an alternative to the River Continuum Concept for the influence of geomorphic processes on aquatic ecosystems. Broadly defined, the Process Domain Concept is a multi-scale hypothesis that spatial variability in geomorphic processes governs temporal patterns of disturbances that influence ecosystem structure and dynamics. At a coarse scale, regional climate, geolog vegetation, and topography control the suite of geomorphic processes that are distributed over a landscape. Within the broad context so defined, stream channel classification can guide identification of functionally similar portions of a channel network, but the response of otherwise similar reaches can depend upon their geologic and geomorphic context. Within geomorphic provinces defined by differences in topography, climate history, and tectonic setting, areas with generally similar geology and topography define lithotopo units, which are useful for stratifying different suites of dominant geomorphic processes. Process domains are spatially identifiable areas characterized by distinct suites of geomorphic processes, and the Process Domain Concept implies that channel networks can be divided into discrete regions in which community structure and dynamics respond to distinctly different disturbance regimes. The concepts of process domains and lithotopo units provide both a framework for the application of patch dynamics concepts to complex landscapes and a context for addressing the effects of watershed processes on the ecology of mountain drainage basins.     

Spatial and temporal dynamics of hedgerows in three agricultural landscapes of southern Quebec,Canada

Noncrop areas such as hedgerows in agricultural landscapes can perform several ecological and agronomic functions (e.g., habitat, movement corridors, windbreak, etc.), but their dynamics and drivers of changes are often poorly known. We conducted a study in three agricultural landscapes of southern Quebec, Canada, to assess and compare the spatial and temporal (1958–1997) dynamics of three hedgerow networks in relation to geomorphic conditions (marine, glacial, and mixed deposit) and land-use changes. Hedgerow networks were mapped and described in terms of their structure (density, degree of connectivity, and presence of trees or shrubs) and their relationship to other components of the landscape (connection to woodland). Relationships were assessed in time and space using nonparametric correlation, Mantel test, and principal components analysis (PCA). Results show significant differences between hedgerow structure for the three landscapes and distinct temporal and spatial dynamics that can be related to changes in management practices and agricultural policies. On marine deposits, increases in hedgerow density did not always correspond to an increase in their degree of connectivity, suggesting a possible reduction in network quality. On glacial deposits, hedgerow density declined following abandonment of agricultural land, but rather than disappearing, these linear structures were integrated into adjacent brush or forested areas. Our analysis reveals the complex spatial and temporal dynamics of the hedgerow networks and highlights the need to take into account spatial attributes such as connectivity and connection to woodland to evaluate more accurately overall network quality.     

A landscape ecological approach to address scaling problems in conservation management and monitoring

Management of many African game reserves is today often still an art based on experience and intuition, rather than a science. Decision-making is based on an informal integration of accumulated individual knowledge and keen field observations. Data are generally poorly captured and curated. Until fairly recently, denominators of biological parameters (such as the unit of land or unit of plant production used as measurement) have generally been treated as being homogenous. The patchiness of landscapes and the issue of ecological scaling were ignored, often because of a lack of appropriate technical tools. The ecological data available on the 49,000-ha Songimvelo Game Reserve (SGR) result from a number of discrete survey and monitoring projects undertaken by different researchers, with different objectives, at different spatial and temporal scales. A landscape ecological approach towards research and monitoring is appropriate for an area of the size and diversity of the SGR. A combination of a database approach and spatial representation was used to consolidate and integrate data across temporal and spatial scales. Herbivore spatial and temporal distribution patterns were explored across three spatial scales. An understanding was achieved of the importance of landscape patchiness in controlling resource availability for herbivores. This insight is important in guiding management and monitoring of the SGR by placing perceived patch overutilization in its proper landscape context. The landscape ecological approach bridges the traditional scale-independent view to a more contemporary scale-related understanding of ecosystem diversity and functioning.     

成都市热力景观空间格局分析

城市热力景观空间格局是城市附属物和城市活动性在热场方面的综合表现。使用2008年11月23日的Landsat5/TM数据,将第6波段反演为亮温,并按温度值的高低把热力景观分为7个类型;借鉴景观生态学的原理和方法,以破碎度指数、多样性指数、均匀度指数和优势度指数为评价指标,对成都市三环路内约200km2热力景观空间格局的研究结果表明：5个行政区的热力景观格局指数存在显著差别,如锦江区的破碎度指数最大,为0.924,成华区的最小,为0.559,指数大小与城市热力景观类型的分布特点密切相关。这些结果对于分析和研究城市整体格局、城市规划和环境保护等具有应用意义。     

Analysis of Changes in Farm Pond Network Connectivity in the Peri-Urban Landscape of the Taoyuan Area,Taiwan

The farm pond system for irrigation is the most prominent feature in the Taoyuan area, Taiwan, giving the region a unique landscape and hydrological character. Although this area had more than 3,290 ponds in the 1970s, fewer than 1,800 now remain. This study analyzes changes in irrigation farm ponds and the canal network landscape in the Taoyuan area. The spatial and temporal changes to ponds and the canal network on the Taoyuan plain were examined graphically for each spatial unit (2,765 m × 2,525 m) using aerial photographs for 1979 and 2005. Landscape metrics were calculated to analyze landscape change associated with increased urbanization. Landscape indices of connectivity and circuitry were utilized to describe changes in the configuration of ponds and canal networks. The total length of canals and total number of ponds in the study area decreased significantly during 1979-2005. The average values of connectivity indices (γ- and α-index) also decreased during 1979-2005, reflecting degradation of canal networks due to urban sprawl. A multivariate technique was applied to portion the study area into three zones according to changes to land cover, ponds, and canal networks. The effects of urban sprawl on the spatial pattern of ponds and canal networks are discussed.     

Examining Landscape Dynamics at a Watershed Scale Using Landsat TM Imagery for Detection of Wintering Hooded Crane Decline in Yashiro,Japan

It is usually inappropriate to define rectangular land areas or administrative units as the extent for quantifying landscapes that possess hierarchical structure. As a functional unit established by geophysical relationships, the watershed is one of many natural scales in the hierarchical landscape. We examined the dynamics of the Yashiro watershed of Japan at the landscape level using pattern metrics based on Landsat thematic mapper (TM) imagery from 1985 to 1998. This watershed provides important habitats for the hooded crane (Grus monachus), a vulnerable species. While its world population has remained stable, the number wintering at Yashiro has declined in recent years. Changes in landscape metrics reveal that the spatial pattern within the watershed underwent homogenization due to depopulation of local people and shifts in local energy requirements and forest management policy at Yashiro. Specific changes include: a decrease in bare land area from 6.2% to 1.0% of the landscape, increased forest cover from 69.2% to 76.1%, reduction in patch number from 1194 to 616 and enlarged mean patch size, and a decrease in total edge from 223,740 m to 158,040 m. The rate of change in landscape metrics indicates a rapid change towards homogeneity in the landscape since 1990. The temporal changes in hooded crane populations corresponded to the changes in landscape. An alternative explanation has been proposed that decline of the species is influenced by landscape dynamics affecting both habitat selection and food resources. Conservation at the watershed scale is suggested to be complementary to the current conservation measures of the species.     

Maintaining the Conservation Value of Shifting Cultivation Landscapes Requires Spatially Explicit Interventions

Fallow vegetation within landscapes dominated by shifting cultivation represents a woody species pool of critical importance with considerable potential for biodiversity conservation. Here, through the analysis of factors that influence the early stages of fallow vegetation regrowth in two contrasting forest margin landscapes in Southern Cameroon, we assessed the impact of current trends of land use intensification and expansion of the cultivated areas, upon the conservation potential of shifting cultivation landscapes. We combined the analysis of plot and landscape scale factors and identified a complex set of variables that influence fallow regrowth processes in particular the characteristics of the agricultural matrix and the distance from forest. Overall we observed a decline in the fallow species pool, with composition becoming increasingly dominated by species adapted to recurrent disturbance. It is clear that without intervention and if present intensification trends continue, the potential of fallow vegetation to contribute to biodiversity conservation declines because of a reduced capacity, (1) to recover forest vegetation with anything like its original species composition, (2) to connect less disturbed forest patches for forest dependent organisms. Strategies to combat biodiversity loss, including promotion of agroforestry practices and the increase of old secondary forest cover, will need not only to operate at a landscape scale but also to be spatially explicit, reflecting the spatial pattern of species reservoirs and dispersal strategies and human usage across landscapes.     

Lakes,Wetlands, and Streams as Predictors of Land Use/Cover Distribution

The importance of the surrounding landscape to aquatic ecosystems has been well established. Most research linking aquatic ecosystems to landscapes has focused on the one-way effect of land on water. However, to understand fully the complex interactions between aquatic and terrestrial ecosystems, aquatic ecosystems must be seen not only as receptors of human modification of the landscape, but also as potential drivers of these modifications. We hypothesized that the presence of aquatic ecosystems influences the spatial distribution of human land use/cover of the nearby landscape (≤1 km) and that this influence has changed through time from the 1930s to the 1990s. To test this hypothesis, we compared the distribution of residential, agricultural, and forested land use/cover around aquatic ecosystems (lakes, wetlands, and streams) to the overall regional land use/cover proportion in an area in southeast Michigan, USA; we also compared the distribution of land use/cover around county roads/highway and towns (known determinants of many land use/cover patterns) to the regional proportion. We found that lakes, wetlands, and streams were strongly associated with the distribution of land use/cover, that each ecosystem type showed different patterns, and that the magnitude of the association was at least as strong as the association with human features. We also found that the area closest to aquatic ecosystems (<500 m) was more strongly associated with land use/cover distribution than areas further away. Finally, we found that the strength of the association between aquatic ecosystems and land use/cover increased from 1938 to 1995, although the overall patterns were similar through time. Our results show that a more complete understanding is needed of the role of aquatic ecosystems on the distribution of land use/cover.     

Assessing the spatial distribution of grassland age in a marginal European landscape

Grassland age is increasingly recognised to be an indicator for present-day biodiversity, e.g. plant species richness, and is also important for other landscape functions. We developed a methodological approach to systematically assess the spatial distribution of grassland age in marginal European landscapes. This approach - applied to the Lahn-Dill Highlands (1270 km(2)), a marginal landscape in Hesse, Germany - comprises three steps: (1) in a two-stage stratification process, we pre-stratified the study area according to recent land-cover patterns and their changes between 1955 and 1995 (stratification I) and classified grassland types by combining data on soil moisture, base-richness, and elevation (stratification II). From 50 grassland types, we randomly selected 1000 representative grassland patches. (2) We determined the age of these patches by means of aerial photograph interpretation of a chronosequence dating back to 1953 and classified each patch with respect to the age classes young (<18 years), mid-aged (18-47 years), and old (>47 years). (3) Based on this information, we calculated grassland type-specific probabilities for grassland patches to belong to the respective age classes. These probabilities were projected to districts by direct extrapolation. An exemplary validation of extrapolation results for two test areas was performed. The results revealed that 49% of the investigated patches were old grassland. The remaining patches were mid-aged (36%) or young grassland (15%). The extrapolation results indicated accordingly a predominance of old grassland at the district scale. Occurrences of mid-aged grassland were concentrated in districts with a pronounced land-cover change, whereas young grassland is apparently evenly distributed across the study area. Validation results suggest that our approach is suitable for a realistic estimation of grassland age in marginal European landscapes. The method may be applied in landscape models of various disciplines that rely on large-scale information on grassland age.     

From Forest Landscape to Agricultural Landscape in the Developing Tropical Country of Malaysia: Pattern,Process, and Their Significance on Policy

Agricultural expansion and deforestation are spatial processes of land transformation that impact on landscape pattern. In peninsular Malaysia, the conversion of forested areas into two major cash crops—rubber and oil palm plantations—has been identified as driving significant environmental change. To date, there has been insufficient literature studying the link between changes in landscape patterns and land-related development policies. Therefore, this paper examines: (i) the links between development policies and changes in land use/land cover and landscape pattern and (ii) the significance and implications of these links for future development policies. The objective is to generate insights on the changing process of land use/land cover and landscape pattern as a functional response to development policies and their consequences for environmental conditions. Over the last century, the development of cash crops has changed the country from one dominated by natural landscapes to one dominated by agricultural landscapes. But the last decade of the century saw urbanization beginning to impact significantly. This process aligned with the establishment of various development policies, from land development for agriculture between the mid 1950s and the 1970s to an emphasis on manufacturing from the 1980s onward. Based on a case study in Selangor, peninsular Malaysia, a model of landscape pattern change is presented. It contains three stages according to the relative importance of rubber (first stage: 1900–1950s), oil palm (second stage: 1960s–1970s), and urban (third stage: 1980s–1990s) development that influenced landscape fragmentation and heterogeneity. The environmental consequences of this change have been depicted through loss of biodiversity, geohazard incidences, and the spread of vector-borne diseases. The spatial ecological information can be useful to development policy formulation, allowing diagnosis of the country’s “health” and sustainability. The final section outlines the usefulness of landscape analysis in the policy-making process to prevent further fragmentation of the landscape and forest loss in Malaysia in the face of rapid economic development.

Decision Tree Algorithm for Detection of Spatial Processes in Landscape Transformation

The conversion of landscapes by human activities results in widespread changes in landscape spatial structure. Regardless of the type of land conversion, there appears to be a limited number of common spatial configurations that result from such land transformation processes. Some of these configurations are considered optimal or more desirable than others. Based on pattern geometry, we define ten processes responsible for pattern change: aggregation, attrition, creation, deformation, dissection, enlargement, fragmentation, perforation, shift, and shrinkage. A novelty in this contribution is the inclusion of transformation processes causing expansion of the land cover of interest. Consequently, we propose a decision tree algorithm that enables detection of these processes, based on three parameters that have to be determined before and after the transformation of the landscape: area, perimeter length, and number of patches of the focal landscape class. As an example, the decision tree algorithm is applied to determine the transformation processes of three divergent land cover change scenarios: deciduous woodland degradation in Cadiz Township (Wisconsin, USA) 1831–1950, canopy gap formation in a terra firme rain forest at the Tiputini Biodiversity Station (Amazonian Ecuador) 1997–1998, and forest regrowth in Petersham Township (Massachusetts, USA) 1830–1985. The examples signal the importance of the temporal resolution of the data, since long-term pattern conversions can be subdivided in stadia in which particular pattern components are altered by specific transformation processes.     

Indicators of visual scale as predictors of landscape preference; a comparison between groups

Landscape change alters the visual scale of agricultural landscapes, as production units get larger and strips of taller vegetation are removed between smaller units. Visual scale, defined as the perceptual units reflecting the experience of landscape rooms, visibility and openness, is considered a key factor shaping landscape preference. The visual expression of landscapes affects people in many ways; aesthetic appreciation, health and well being. In order to understand how landscape changes alter the visual scale of landscapes, and hence its effect on human beings, it is important to establish efficient and practical ways of measuring visual scale and its relation to landscape preferences. This paper presents an empirical test of the ability to predict landscape preference of two photo-based indicators of visual scale; percentage open land in the view and size of landscape rooms. A preference survey was conducted with two groups of respondents; a student group from the Norwegian University of Life Sciences (UMB) and a public group. The student group was chosen to represent future landscape professionals. Photos of landscapes varying in the expression of visual scale were shown to survey respondents, who were asked to give scores according to how much they liked the landscapes. Both indicators were found to be predictors of preference for the student group, but not for the public group. The results demonstrate that student preferences do not reflect the landscape preferences of the wider public. The different appreciation of visual scale in the landscape of future landscape professionals and the public implies that experts wanting to reflect preferences of the wider public should be cautious when evaluating the visual impact of landscape change.     

The application of local measures of spatial autocorrelation for describing pattern in north Australian landscapes

This paper tests the use of a spatial analysis technique, based on the calculation of local spatial autocorrelation, as a possible approach for modelling and quantifying structure in northern Australian savanna landscapes. Unlike many landscapes in the world, northern Australian savanna landscapes appear on the surface to be intact. They have not experienced the same large-scale land clearance and intensive land management as other landscapes across Australia. Despite this, natural resource managers are beginning to notice that processes are breaking down and declines in species are becoming more evident. With future declines of species looking more imminent it is particularly important that models are available that can help to assess landscape health, and quantify any structural change that takes place. GIS and landscape ecology provide a useful way of describing landscapes both spatially and temporally and have proved to be particularly useful for understanding vegetation structure or pattern in landscapes across the world. There are many measures that examine spatial structure in the landscape and most of these are now available in a GIS environment (e.g. FRAGSTATS* ARC, r.le, and Patch Analyst). All these methods depend on a landscape described in terms of patches, corridors and matrix. However, since landscapes in northern Australia appear to be relatively intact they tend to exist as surfaces of continuous variation rather than in clearly defined homogeneous units. As a result they cannot be easily described using entity-based models requiring patches and other essentially cartographic approaches. This means that more appropriate methods need to be developed and explored. The approach examined in this paper enables clustering and local pattern in the data to be identified and forms a generic method for conceptualising the landscape structure where patches are not obvious and where boundaries between landscape features are difficult to determine. Two sites are examined using this approach. They have been exposed to different degrees of disturbance by fire and grazing. The results show that savanna landscapes are very complex and that even where there is a high degree of disturbance the landscape is still relatively heterogeneous. This means that treating savanna landscapes as being made up of homogeneous units can limit analysis of pattern, as it can over simplify the structure present, and that methods such as the autocorrelation approach are useful tools for quantifying the variable nature of these landscapes.     

The Influence of Landscape Features on Road Development in a Loess Region,China

Many ecologists focus on the effects of roads on landscapes, yet few consider how landscapes affect road systems. In this study, therefore, we quantitatively evaluated how land cover, topography, and building density affected the length density, node density, spatial pattern, and location of roads in Dongzhi Yuan, a typical loess region in China. Landscape factors and roads were mapped using images from SPOT satellite (Système Probatoire d’Observation de la Terre), initiated by the French space agency and a digital elevation model (DEM). Detrended canonical correspondence analysis (DCCA), a useful ordination technique to explain species–environment relations in community ecology, was applied to evaluate the ways in which landscapes may influence roads. The results showed that both farmland area and building density were positively correlated with road variables, whereas gully density and the coefficient of variation (CV of DEM) showed negative correlations. The CV of DEM, farmland area, grassland area, and building density explained variation in node density, length density, and the spatial pattern of roads, whereas gully density and building density explained variation in variables representing road location. In addition, node density, rather than length density, was the primary road variable affected by landscape variables. The results showed that the DCCA was effective in explaining road–landscape relations. Understanding these relations can provide information for landscape managers and transportation planners.     

Patterns of biodiversity and habitat sensitivity in agricultural landscapes

Design of landscape is the process of the arrangement of spatial features with the objective of sustaining ecosystem services, and maintaining ecological functionality to meet societal needs. Along a gradient of cultivation intensity, the functional quality of agricultural landscape was explored and the relationships between landscape metrics and functional quality were analyzed, in order to make effective recommendations for landscape design aimed at sustainable land use schemes. The functional quality of landscape was calculated using the InVest model for 20 farm landscapes (North-Eastern Italy) where biodiversity (plant taxa) and sensitivity to disturbance (hemeroby) were used as model inputs. Results highlighted the importance of specific habitat types such as meadows and woodlands rather than other habitats for improving the biodiversity of agricultural landscapes. A high proportion of these habitats enhanced the functional quality of the landscape when the habitats were organized in large and not isolated patches in heterogeneous landscapes.     

A Hierarchical Approach to Forest Landscape Pattern Characterization

Landscape spatial patterns have increasingly been considered to be essential for environmental planning and resources management. In this study, we proposed a hierarchical approach for landscape classification and evaluation by characterizing landscape spatial patterns across different hierarchical levels. The case study site is the Red Hills region of northern Florida and southwestern Georgia, well known for its biodiversity, historic resources, and scenic beauty. We used one Landsat Enhanced Thematic Mapper image to extract land-use/-cover information. Then, we employed principal-component analysis to help identify key class-level landscape metrics for forests at different hierarchical levels, namely, open pine, upland pine, and forest as a whole. We found that the key class-level landscape metrics varied across different hierarchical levels. Compared with forest as a whole, open pine forest is much more fragmented. The landscape metric, such as CONTIG_MN, which measures whether pine patches are contiguous or not, is more important to characterize the spatial pattern of pine forest than to forest as a whole. This suggests that different metric sets should be used to characterize landscape patterns at different hierarchical levels. We further used these key metrics, along with the total class area, to classify and evaluate subwatersheds through cluster analysis. This study demonstrates a promising approach that can be used to integrate spatial patterns and processes for hierarchical forest landscape planning and management.