Public benchmarks

Anonymised aggregate benchmark stats

Overall cross-port benchmark statistics from submitted baselines. This public view applies small-cohort suppression and never shows individual ports, port-specific markers, or current-user comparisons.

Anonymised — indicative

All submitted ports

Cross-port aggregates from submitted baselines. Individual ports, port names, port IDs, and current-user comparisons are not included on this public page.

n = 9

Benchmark distribution

Published cohort ranges

Band = P25-P75, marker = mean

Total CO2

t CO2/yr

Lower is better

P25

672.9

Mean

1,051.6

P75

1,041.3

Emission intensity

kg CO2/t cargo

Lower is better

P25

2.61

Mean

256.39

P75

186.15

Energy intensity

kWh/t cargo

Lower is better

P25

20

Mean

1,046

P75

874

Renewable share

%

Higher is better

P25

9.8

Mean

10.5

P75

11.3

Radar overview

Cohort performance shape

Each axis normalises the published cohort mean against its P25-P75 band. Better aggregate performance extends farther from the centre.

Radar overview of anonymised cohort benchmark metricsRadar chart normalising CO2 load, emission intensity, energy intensity, and renewable share from public aggregate benchmark statistics.CO2 loadCO2 / tonneEnergy / tonneRenewables

Cohort maturity

Aggregate Port Score profile

Port Score averages are published only as cohort statistics. No individual port self-assessment values are visible.

Average Port Score

56 / 100

P25-P75: 55-59

Strongest aggregate dimension

Management (2.9 / 5)

Weakest aggregate dimension

Infrastructure (2.4 / 5)

Strategy

2.8 / 5 mean

Management

2.9 / 5 mean

Infrastructure

2.4 / 5 mean

Superstructure

2.6 / 5 mean

External participants

2.6 / 5 mean

Total CO2

t CO2/yr

Mean
1,051.6
Median
962.5
P25-P75 range
672.9-1,041.3

Emission intensity

kg CO2/t cargo

Mean
256.39
Median
2.82
P25-P75 range
2.61-186.15

Energy intensity

kWh/t cargo

Mean
1,046
Median
21
P25-P75 range
20-874

Renewable share

%

Mean
10.5
Median
10.5
P25-P75 range
9.8-11.3

Outliers outside mean +/- sigma are excluded before publication. Small-cohort suppression is applied per KPI, including the aggregate maturity panel.