When we look at a satellite map of a changing region, our eyes naturally pick up on patterns: a forest eaten away by roads, a wetland dissolving into patches of agriculture, or a city spreading like ink across a valley. But to manage ecosystems, design energy-efficient cities, or evaluate ecological risks, we need to move past visual intuition. We need to turn these visual textures into precise numbers.
Landscape pattern analysis bridges this gap. It treats a geographic area not just as a collection of land-cover types, but as a geometric configuration of discrete "patches." By calculating mathematical metrics across these patches, researchers can quantify exactly how fragmented, connected, or complex a landscape is—and understand the underlying environmental and human processes driving these changes.
The Anatomy of a Patch
The fundamental unit of landscape analysis is the patch: a contiguous non-linear area of a single land-cover type. If you have a continuous forest and cut a highway through it, you haven't changed the total forest area significantly, but you have fundamentally altered its patch geometry.
To quantify this structure, researchers look at three core dimensions of a patch: dominance, edge complexity, and internal stability.
Dominance and Core Zones
The simplest way to understand a landscape's integrity is to ask whether it is ruled by a massive, unbroken habitat or broken into minor pieces. The Largest Patch Index (LPI) measures the percentage of the total landscape area occupied by the single largest patch.
As shown by Mahdieh Shirmohammadi et al., tracking the LPI serves as an excellent proxy for habitat fragmentation. In their study of mangrove ecosystems under climate stress, a shrinking LPI signaled that core ecological zones were shattering into smaller, more vulnerable islands, even before the total vegetation area showed a drastic decline.
Boundary Complexity and Edge Effects
A square patch and a highly convoluted, star-shaped patch might have the exact same area, but they interact with the world completely differently. The convoluted patch has a much longer perimeter, creating an extensive "edge zone" where the interior habitat is exposed to outside disturbances.
To measure this, researchers use metrics like the Landscape Shape Index (LSI) and the Fractal Dimension. The fractal dimension uses an inverse power law to evaluate how patch perimeters scale with their areas.
- In agricultural conservation, Dan Yu et al. utilized fractal dimensions to measure the stability and utilization efficiency of peri-urban cultivated land, showing how urban encroachment forces farms into irregular, unstable shapes.
- In urban morphology, Qianchuan Zhao demonstrated that fractal dimension analysis can provide a quantitative signature that differentiates between rigid, top-down planned neighborhoods and emergent, self-organized urban fabrics.
Simple Shape (Low LSI / Fractal Dim) Complex Shape (High LSI / Fractal Dim)
┌───────────┐ ┌───┐ ┌───┐
│ │ │ └───────┘ │
│ Inside │ │ Inside │
│ │ ┌─────┘ └─────┐
└───────────┘ └───────────────────────────┘
Minimal Edge Exposure Maximal Edge Exposure (Vulnerable)
From Patches to Landscape Texture
While individual patch metrics tell us about local structure, we often need to characterize the "texture" of an entire region. This requires looking at how different patches are arranged relative to one another.
Contagion and Aggregation
How "clumpy" is the landscape? Contagion (CONTAG) measures the probability that two randomly chosen adjacent pixels belong to the same land-cover class.
- High Contagion: The landscape consists of large, contiguous, well-aggregated blocks.
- Low Contagion: The landscape is highly interspersed, resembling a fine-grained checkerboard or a peppered mixture of different land uses.
Ge Shi 0008 et al. and Yunhui Liu et al. rely on contagion alongside Shannon’s Diversity Index (SHDI) to capture how rapid urbanization shatters continuous rural habitats. When a city expands haphazardly, contagion drops sharply because the uniform rural matrix is interrupted by isolated pockets of built-up land.
Cohesion and Division
To understand whether species can move through a landscape, researchers look at connectivity metrics like Patch Cohesion and Landscape Division (DIVISION). Cohesion measures the physical connectedness of a land-cover type, while division quantifies the probability that two randomly chosen points in the landscape are not in the same patch.
As Ying Zhang et al. note in their development of an Integrated Spatial Landscape Index, metrics like DIVISION and Proportion of Landscape (PLAND) are highly sensitive indicators of structural shifts, capturing whether a nature reserve is consolidating or drifting apart.
Why the Geometry Works: Linking Form to Process
The power of landscape pattern metrics lies in their ability to act as mathematical proxies for real-world physical and socioeconomic processes. Instead of running computationally expensive simulations of animal movement, microclimates, or human behavior, researchers can read the geometric signatures left behind on the map.
Distinguishing Urban Growth Modes
Cities do not all grow the same way. A city might expand by filling in its empty central spaces (infilling) or by throwing out isolated clusters of development far into the countryside (leapfrog development).
Liyuan Qing et al. showed that by tracking patch density and shape complexity over time, algorithms can automatically distinguish between these growth modes. Leapfrog development causes an immediate spike in patch density and boundary irregularity, whereas infilling reduces patch density and smooths out the urban footprint.
Infilling Growth Leapfrog Growth
┌───────────────┐ ┌───────────────┐
│ ┌───┐ ┌───┐ │ │ ┌───┐ │
│ │ │─│ │ │ <--- Patches merge │ │ │ ┌───┐ │ <--- New isolated
│ └───┘ └───┘ │ (Density decreases) │ └───┘ │ │ │ patch appears
└───────────────┘ └─────────└───┘─┘ (Density spikes)
Thermodynamics and Public Health
The physical shape of a city directly dictates its thermodynamic behavior and environmental quality. Zi-gui Chen et al. demonstrated that urban centrality (measured via the proportion of the largest patch) and patch aggregation are directly tied to energy efficiency; consolidated cities minimize transportation overhead and heat loss.
Similarly, Chenlong Wang et al. found that landscape metrics like edge density and fractal dimension can modulate how air pollution affects human mortality, as the spatial arrangement of green spaces determines how effectively pollutants are dispersed or trapped in the built environment.
Quantifying Ecological Risk
When human activities like mining or farming encroach on natural territory, they don't just reduce the total acreage of wilderness; they degrade its structural resilience. To capture this, researchers like Wenjia Xu et al. build composite "Landscape Disturbance" indices.
By mathematically combining fragmentation (patch density), separation (how far apart patches are), and dominance (LPI), these models calculate an ecosystem's vulnerability. A landscape with high fragmentation and low cohesion is highly vulnerable because its sub-populations are isolated, making it much harder for the ecosystem to recover from localized shocks.
Ultimately, landscape pattern analysis works because geography is predictive. By translating the chaotic, visual reality of maps into structured metrics of patches, edges, and clusters, we gain a rigorous framework to evaluate the balance between human development and ecological survival.
Key Foundational Work
Spatial-temporal changes of regional sustainability: an empirical study in Taiwan (2006) — contributed 2 concept elements to this direction.
Rural Land Use Change during 1986–2002 in Lijiang, China, Based on Remote Sensing and GIS Data (2008) — contributed 1 concept elements to this direction.