The increase in fine spatial resolution urban vegetation data and advances in high-performance computing has allowed for more detailed approaches to monitoring change in urban green space and vegetation mixes. Moving beyond standard measures such as area or percent cover of vegetation type, we draw on landscape metrics and novel measures of spatial arrangement to describe urban vegetation composition and configuration. To this end, we calculate the ‘spatial signature’ or probability distribution of land-cover class adjacencies for a predefined mapping unit to describe the diversity and complexity of vegetation patterns across all of Perth for multiple years.
Here, we use a mapping unit's ‘spatial signature’ to detect change in urban vegetation patterns across time. A city-scale landscape can be screened for locations where vegetation patterns are changing and measure the magnitude of this change. We also use 'spatial signatures' in machine learning and statistical analysis tasks to gain a more nuanced understanding of the relationship between vegetation patterning and a range of outcomes. We examine the relationship between spatial patterns of vegetation and Landsat 8 derived surface temperatures. The results indicate that, in certain instances, vegetation arrangement may have just as much of an impact on reducing surface temperatures as increasing the overall area of tree canopy. These results illustrate how the ‘spatial signature’ metric can support and enhance a range of urban vegetation monitoring tasks and how measures of spatial arrangement can inform contemporary planning debates concerned with mitigating the impact of canopy loss from densification and urban consolidation.