The city records those trees together, as part of one urban forest. On paper, that seems reasonable. The inventory holds 82,064 tree records and 680 scientific-name categories. It sounds like a picture of abundance and variety.
Yet the two numbers conceal something important. They say nothing about which species are doing most of the work of filling that inventory, or how the mix changes between a park and the streets around it.
I wanted to understand that difference, because an urban forest is not only a count on a council website. It is the shade people seek, the trees outside homes and workplaces, and the green infrastructure a city must maintain through pests, disease and a changing climate. The question matters to people who use those spaces, as well as those deciding what to plant and protect.
Melbourne has recognised the importance of variety for years. In 2011, the City of Melbourne proposed a 2040 goal under which no species would exceed 5% of the urban forest, no genus 10%, and no family 20% (City of Melbourne, 2011). Those were long term, citywide concentration guidelines—not deadlines that had already passed.
I went back to the city’s September 2025 inventory (City of Melbourne, 2025). What I found was not one simple story about a green city, but a succession of contrasts. The first appears before we even leave the citywide totals.
The citywide figures reveal a tension the overall count does not show. River red gum (Eucalyptus camaldulensis) accounts for 10.2% of these tree records more than double the 5% species benchmark. At broader levels, Eucalyptus represents 21.6% of the inventory, and Myrtaceae 36.0%.
These are not three independent groups. River red gum belongs to Eucalyptus, and Eucalyptus sits within Myrtaceae. That overlap is precisely why the family level result matters: diversity can look substantial at one level while the inventory remains heavily weighted towards related trees.
It would be wrong to call this a failure. The benchmarks are intended for 2040, and this inventory is only one snapshot of recorded public trees. But the figures do challenge a comfortable assumption: a long list of species is not the same thing as a balanced forest.
And that is before location enters the story. The citywide average gives no hint of what happens when the same inventory is separated into park trees and street trees.
River red gum dominates a much larger share of the park inventory: 15.8%, against just 2.1% of street trees. London plane (Platanus × acerifolia) turns the picture around. Its share is 13.6% on streets and only 1.0% in parks.
Those two species almost exchange roles. A name that looms large in one environment becomes relatively uncommon in the other. Count the trees together and that contrast fades into a single citywide average.
This matters for more than the appearance of a graph. A resident’s everyday experience of urban trees depends on where they spend time. For someone managing the forest, a species mix concentrated in streets can raise different questions from a mix concentrated in parks. The inventory describes those differences; it does not prove what caused them.
Still, it would be tempting to settle on a tidy explanation: parks are the botanically varied places, while streets are more uniform. The neighbourhood figures show why that assumption would be premature.
In the precinct labelled Melbourne, park trees have an effective diversity of approximately 101.0, compared with 21.7 for street trees. It is easy to look at that contrast and assume that parks generally contain a more balanced mix of trees.
Across the city, however, the result is not consistent. 4 of 11 precincts in the comparison actually have a higher street tree diversity score: West Melbourne, Parkville, Southbank, Docklands. This is not a Parkville-only exception. The direction of the park street difference genuinely changes by precinct.
Parkville makes the reversal particularly revealing. Its streets have an effective diversity of 38.9, against 24.0 in its parks. Looking back at the original records provides a clue: 25.7% of Parkville’s park trees are river red gums. That’s a substantial share for a single species, although it does not, by itself, establish why Parkville’s diversity index is lower.
Different precinct sizes, park configurations and inventory coverage may also shape these comparisons. The next step is to see where the two prominent species from the park street chart appear in Melbourne’s geography.
Parkville now connects two of the charts. Its recorded park trees have a strong river red gum presence, and some of the highest count river red gum grid cells also appear around Parkville. London planes show a different spatial pattern, with selected high-count cells in and around central Melbourne.
These are related observations, not proof of a cause-and-effect relationship. The map doesn’t measure each precinct’s entire distribution, and the high-count cells do not prove that river red gums caused Parkville’s lower park diversity. They do show why a citywide species average can conceal highly local concentrations.
Only the ten highest count grid cells for each species are displayed, among cells containing at least 50 trees overall. Areas without markers may still contain the species. The symbol sizes show selected counts, not canopy cover or trees per hectare.
London planes are prominent in the existing street inventory. But are they equally prominent in the records assigned more recent planting years?
The contrast is hard to miss. London planes make up 17.9% of street trees assigned a planting year of 2008, against 3.3% of those assigned 2024. The line isn’t a smooth decline; there are bumps and reversals. But several newer recorded cohorts contain London planes at much lower shares than some earlier ones.
These are surviving tree records, not a planting ledger. Some trees planted in earlier years may since have been removed, and records may not capture their original planting dates consistently. The series starts in 2006 because there were no eligible street records in 2003, only one in 2004 and 92 in 2005. It ends in 2024 because the September 2025 snapshot cannot represent a complete 2025 planting year. The chart therefore cannot establish changes in planting policy.
What the chart can say is narrower, and still worth knowing: the London plane’s place in the existing street population is not mirrored equally across newer recorded cohorts. The forest we count today and the trees assigned more recent planting years are not necessarily composed in the same way.
Melbourne’s inventory looks reassuring when reduced to one total. Its botanical makeup is much less uniform. One river red gum can count towards a species, a genus and a family already prominent in the inventory. Parks and streets have contrasting species mixes. And in four of the eleven precincts suitable for comparison, streets not parks show higher effective diversity.
Parkville makes the point concrete: river red gums represent more than a quarter of the recorded park trees there. The selected map cells show that some river red gum concentrations also fall nearby. Together, the charts make a case for looking below the citywide average, without pretending the map explains every ecological difference.
The practical takeaway is specific: as Melbourne works towards its long-term botanical diversity ambitions, species, genus and family shares should be tracked separately across parks, streets and precincts. Repeated inventories could show whether those concentrations change. A single snapshot cannot establish why they arose, whether individual trees are healthy or how resilient each neighbourhood will be to future threats.
The useful question is no longer only how many trees Melbourne has. It is which trees are common, where they are common, and whether that mix becomes more balanced over time.
The investigation draws on the City of Melbourne’s public Trees, with species and dimensions (Urban Forest) dataset, using the CSV snapshot updated on 22 September 2025 (City of Melbourne, 2025). The source data are read from a local CSV file rather than entered manually. The dataset describes inventoried tree assets, not every tree in the wider metropolitan area.
Melbourne’s botanical concentration benchmarks come from its 2011 Urban Forest Diversity Guidelines (City of Melbourne, 2011). These were proposed for the urban forest as a whole with a 2040 horizon, not as requirements for each neighbourhood or a measure of current ecosystem health.
The dataset labels each record as either Park or
Street in the located_in field, and
provides the named administrative area in its precinct
field. I used these existing source classifications rather than
inventing park boundaries or matching neighbourhoods by map coordinates.
Chart 2 compares the ten most frequently recorded scientific names
citywide, selected before splitting them into park and street shares;
its purpose is to show how the most common species differ between
environments.
The precinct comparison uses the effective number of species, derived from the exponential Shannon diversity index (Jost, 2006). Only precincts with at least 200 park-tree and 200 street tree records are included; this leaves 11 comparable precincts. The threshold avoids very small samples but does not equalise geographic area, tree density, sampling conditions or precinct size.
The spatial map displays the ten highest-count approximately 0.003 degree coordinate cells for river red gum and London plane among cells with at least 50 inventoried trees. It shows selected species concentrations rather than the complete distribution, or density per unit area.
Planting year analysis treats pre 2003 dates as unverified historical values and omits 2003–2005 because the surviving street inventory contains too few records in those years for a meaningful annual comparison (zero, one and 92, respectively). It excludes 2025 because the inventory was updated in September, before the year finished. The analysis is based on records remaining in that snapshot, not all historical plantings.
The diameter_breast_height field is missing for about
54.7% of records, and
useful_life_expectency is missing for about
21.2%. Because these fields are incomplete and a small
number of diameter values appear anomalously large, I did not use them
to infer the health or remaining life of trees. The original scientific
name spelling is retained for matching, including the source’s
x notation for hybrids; for readability, the narrative uses
the botanical multiplication sign (×). No independent taxonomic
reconciliation was performed.
City of Melbourne. (2011). Urban forest diversity guidelines: 2011 tree species selection strategy for the City of Melbourne. https://mvga-prod-files.s3.ap-southeast-4.amazonaws.com/public/2024-07/urban-forest-diversity-guidelines.pdf
City of Melbourne. (2025). Trees, with species and dimensions (Urban Forest) [Data set]. data.gov.au. https://data.gov.au/data/en/dataset/trees-with-species-and-dimensions-urban-forest
Jost, L. (2006). Entropy and diversity. Oikos, 113(2), 363–375. https://doi.org/10.1111/j.2006.0030-1299.14714.x
OpenAI. (2026). ChatGPT [Large language model]. https://chatgpt.com/
Generative AI tools, including ChatGPT, were used to assist with identifying and correcting R code errors, checking syntax, improving grammar and sentence structure, and suggesting relevant references. All data analysis, visualisations, interpretations, and final content were independently developed, reviewed, and verified.