Cities worldwide are investing heavily in urban greening, heat adaptation, and climate resilience. But many climate decisions are still based on one-off maps that provide only a snapshot of conditions at a single moment in time.
The more important challenge is no longer simply identifying where cities are hottest today. It is understanding how heat exposure, canopy cover, and vulnerability change over time, whether interventions are improving conditions, and where future investment should be prioritised.
Building on the SmartSat CRC Heatwaves: Kanyini Waru research program, the Adelaide pilot was delivered by Geoneon in collaboration with SmartSat CRC, Green Adelaide, and Flinders University. The work brought together satellite-derived heat monitoring, AI-based canopy mapping, and social vulnerability analysis to explore how cities can move from one-off heat and canopy maps toward repeatable climate-risk monitoring.
Heatwaves are becoming more frequent, longer lasting, and more intense across many urban regions in Australia. At the same time, urban densification continues to increase the concentration of heat-retaining surfaces such as asphalt, rooftops, and concrete.
But urban heat is not distributed evenly.
Some neighbourhoods consistently experience higher thermal stress because of differences in canopy cover, built form, surface materials, and access to cooling infrastructure. These same areas often overlap with populations that are more vulnerable to heat impacts, including older residents, young children, and socioeconomically disadvantaged communities.
Tree canopy remains one of the most effective urban cooling interventions available to cities. Yet understanding where canopy is lacking, where heat concentrates, and which communities remain most exposed requires a repeatable monitoring system.
Most cities already possess baseline environmental datasets. The more difficult challenge is operationalising those datasets into monitoring workflows that:
remain consistent through time,
support repeat reporting,
allow comparison across years,
measure intervention outcomes, and
help prioritise investment spatially.
This was the basis for the Adelaide pilot, which tested how heat, canopy, and vulnerability data could be combined into a repeatable climate-risk monitoring framework.
A stand-alone heat map can identify where temperatures are currently highest.
A canopy map can show where vegetation cover is limited. Neither necessarily answers:
whether conditions are improving,
whether greening investments are working,
how exposure patterns are shifting, or
where future mitigation should be prioritised.
This becomes particularly important once climate adaptation is treated as an ongoing operational problem rather than a single planning exercise.
The Adelaide pilot project explored how heat, canopy, and vulnerability layers could be generated through repeatable workflows rather than one-off analytical outputs. The aim was not simply to create maps, but to establish a monitoring framework capable of supporting annual comparison, long-term reporting, and intervention tracking.
Urban climate risk emerges where environmental exposure intersects with vulnerable populations and limited local cooling capacity.
The Adelaide workflow combined three linked monitoring layers:
heat,
canopy, and
vulnerability.
Together, these layers create a more operational view of urban climate risk than any individual dataset alone.
The first layer captures where thermal stress concentrates spatially across the urban environment.
Using satellite-derived land surface temperature from Landsat 8 and 9 imagery collected between 2020 and 2026, Geoneon developed a Heat Susceptibility Index representing relative summer surface temperatures across metropolitan Adelaide.
The index classified temperatures into decile-based classes from 1 to 10:
1 represents the coolest 10% of locations,
10 represents the hottest 10% within the study area.
Areas with extensive hard surfaces, limited vegetation, and dense urban form were associated with elevated heat exposure.
Importantly, the heat layer was designed as a repeatable preparedness and mitigation product rather than an event-response forecast. The aim is not to predict daily heat behaviour, but to monitor where thermal stress consistently concentrates over time.
Fig. 2: Satellite-derived Heat Susceptibility Index showing relative summer surface temperatures across metropolitan Adelaide. Higher index classes indicate areas where urban heat accumulates most intensely over time.
The second layer captures local cooling capacity through urban tree canopy.
The canopy analysis used high-resolution aerial imagery supplied by the South Australian Department for Environment and Water. Geoneon then applied AI-based canopy classification models to estimate canopy extent across seven metropolitan Adelaide local government areas (LGAs).
The workflow generated:
Unlike broad vegetation indices, the canopy layer focused specifically on urban tree structure and distribution at decision-relevant scales including LGAs, suburbs, and ABS Mesh Blocks. ABS Mesh Blocks are Australia's smallest official statistical geography units, providing a neighbourhood-scale framework for analysing heat exposure, canopy cover and community vulnerability.
The pilot identified substantial variation in canopy cover across Adelaide:
At finer scales, nearly half of all Mesh Blocks had less than 15% canopy cover.
This finer-scale variability is important because local canopy deficits are often obscured within broader municipal averages.
Fig. 3: Fine-scale canopy mapping at Mesh Block scale revealed substantial local variation in urban tree cover, including widespread low-canopy areas not visible at broader municipal scales.
The third layer captures who is most affected by heat exposure.
The Social Vulnerability Index combined demographic age profiles with socioeconomic disadvantage metrics derived from ABS census datasets.
This layer identified areas where populations may be less resilient to extreme heat because of:
older age,
younger age,
social disadvantage, or
reduced adaptive capacity.
Importantly, vulnerability changes more slowly than thermal conditions or canopy distribution. This makes it well suited for integration into longer-term monitoring frameworks and multi-year comparison.
The key insight is that climate-risk prioritisation depends on more than heat alone. In the Adelaide pilot, the Residential Heat Risk Index combined heat exposure and social vulnerability to identify where residents face the greatest risk during extreme heat conditions. Canopy was analysed alongside these risk layers as a measure of local cooling capacity and potential intervention need.
This combined view is important because urban heat alone does not determine where action is most urgent. Vulnerability helps identify where thermal stress is most likely to produce harmful outcomes, while canopy data helps show where greening and shade interventions may be most needed.
Fig 4: Residential Heat Risk Index combining heat exposure and social vulnerability to identify areas where thermal stress and vulnerable populations overlap.
One of the most important findings of the Adelaide pilot was that different environmental indicators require different monitoring strategies.
No single dataset or update cycle is sufficient.
Heat monitoring was anchored to satellite-derived land surface temperature because satellite thermal imagery supports:
repeatable methodology,
broad spatial coverage,
annual updates, and
long-term trend comparison.
The Heat Susceptibility Index was calculated using median summer land surface temperature values derived from multiple Landsat scenes.
This creates a stable monitoring baseline that can be reproduced consistently over time. The value lies in the ability to maintain a consistent annual methodology that supports comparable monitoring across multiple years.
The resulting surface is therefore useful as a preparedness and mitigation layer that helps cities understand where thermal stress persistently concentrates. It does not serve as an emergency-response product that can be used to identify, for example, which public spaces are unsafe due to extreme heat at a given time of day or day of the year.
Canopy monitoring behaves differently. While thermal monitoring benefits from annual consistency, canopy monitoring depends more heavily on high-resolution imagery acquisition cycles.
The Adelaide workflow used aerial imagery originally captured at approximately 7.5 cm resolution and resampled to 50 cm and 1 m resolutions for AI-based canopy classification.
The workflow included:
Because canopy monitoring depends on high-resolution capture programs, updates need to be aligned with imagery acquisition cycles rather than annual release schedules.
Fig 5 and Fig 6: High-resolution aerial imagery (top) and AI-based canopy classification (bottom) workflow used to generate repeatable urban tree canopy layers across metropolitan Adelaide.
Population vulnerability evolves more slowly than environmental conditions and can therefore be updated using census and demographic cycles.
In the Adelaide pilot, vulnerability layers incorporated:
This provides the human context necessary for prioritisation.
Heat alone does not determine where intervention is most urgent. Vulnerability determines where heat exposure is most likely to translate into harm.
The Adelaide pilot integrated heat, canopy, and vulnerability into a combined climate-risk workflow covering seven metropolitan LGAs.
The pilot produced several linked environmental and risk-monitoring outputs:
a Heat Susceptibility Index showing where urban heat consistently concentrates spatially,
a Residential Heat Exposure Index identifying where residential populations are exposed to elevated heat conditions,
a Social Vulnerability Index highlighting populations with reduced resilience to heat impacts,
a combined Residential Heat Risk Index integrating exposure and vulnerability into a prioritisation layer,
detailed tree canopy distribution and percentage-cover layers,
canopy statistics and summaries at LGA, suburb, and Mesh Block scales, and
prioritisation outputs highlighting where elevated heat exposure, limited canopy, and vulnerability intersect spatially.
Several important spatial patterns emerged.
Areas with extensive built surfaces, fragmented green space, and limited canopy consistently aligned with elevated thermal exposure.
The analysis also revealed that canopy deficits and heat exposure frequently overlap at highly local scales that become difficult to identify through broader municipal averages alone.
Approximately 44.8% of residents fell within moderately high to extremely high heat-risk classes.
Importantly, the significance of the pilot was not simply the production of these maps.
The more important outcome was demonstrating that these layers can be generated through repeatable workflows that support future comparison, reporting, and intervention tracking over time.
The real value of environmental monitoring only becomes visible after multiple reporting cycles. A baseline establishes current conditions. Repeat monitoring makes it possible to evaluate change.
Over time, monitoring frameworks can answer questions such as:
Has canopy increased?
Has heat exposure decreased?
Are vulnerable communities benefiting from interventions?
Which greening programs are producing measurable outcomes?
Where should investment continue?
This transition from static mapping toward ongoing climate monitoring is central to climate adaptation planning.
The Adelaide pilot sets the foundation for a shift from producing environmental maps toward building operational climate intelligence systems that support decision-making over time.
Monitoring is ultimately about prioritisation.
Not every suburb can be treated simultaneously, and not every intervention produces equal benefit.
The Adelaide workflow demonstrates how integrated environmental monitoring can support:
urban forest planning,
street greening,
climate adaptation strategies,
infrastructure planning,
grant prioritisation,
public realm upgrades,
cooling corridor planning, and
resilience reporting.
At finer spatial scales such as Mesh Blocks, the monitoring layers become particularly actionable because they reveal highly localised canopy deficits and exposure concentrations that are often invisible at broader regional scales. This can support councils and community resilience stakeholders through consistent reporting and shared GIS delivery.
The Adelaide pilot reflects a broader transition occurring in urban climate analytics.
Cities are increasingly moving toward integrated climate intelligence systems built around:
Earth observation,
AI-assisted environmental mapping,
repeatable monitoring,
automated change detection,
multi-source environmental data, and
operational decision-support workflows.
The Adelaide pilot proposed a practical monitoring model built around two linked environmental streams:
This distinction is important because different environmental variables operate on different temporal and spatial cycles. Heat monitoring benefits from consistent annual observation, while canopy monitoring depends more heavily on the availability of high-resolution imagery and structural data.
The result is not simply a collection of maps. It is an evolving environmental monitoring framework designed to support repeat climate-risk assessment through time.
Urban climate resilience is not achieved by producing a heat map once. It is achieved by establishing repeatable monitoring systems that help cities understand where risk exists, prioritise intervention, and measure whether those interventions are improving conditions over time.
By combining Earth observation, AI-based environmental mapping, and repeatable climate-risk monitoring, the Adelaide pilot demonstrates how cities can move from isolated environmental baselines toward operational climate intelligence.
The Adelaide pilot was delivered by Geoneon in collaboration with Green Adelaide, SmartSat CRC and Flinders University, building on the SmartSat CRC Heatwaves: Kanyini Waru research program. The canopy analysis used aerial imagery supplied by the South Australian Department for Environment and Water.