All figures below are reproducible from the CSVs in this folder using
../scripts/summarize_campaign.py. Savings
are the weather-aware-vs-weather-blind fuel/CO₂ saving in percent, over the
status == ok instances. The full matrix is 12 routes × 4 seasons × 2 years =
96 instances per vessel.
| Vessel | metric | n | mean | median | p90 | max |
|---|---|---|---|---|---|---|
| S175 | route-only | 91 | 5.04 | 1.91 | 14.58 | 39.82 |
| S175 | joint route+speed | 91 | 11.01 | 8.82 | 19.75 | 45.56 |
| KVLCC2 | route-only | 67 | 6.14 | 3.91 | 16.07 | 26.25 |
| KVLCC2 | joint route+speed | 66 | 10.95 | 10.00 | 20.21 | 29.69 |
| JBC | route-only | 66 | 4.55 | 1.82 | 15.09 | 26.81 |
| JBC | joint route+speed | 65 | 9.57 | 7.82 | 17.30 | 30.33 |
The S175 slow-steaming (cube-law) component alone is mean 6.03 %, median 6.93 %.
| Basin | n | mean | median | max |
|---|---|---|---|---|
| N. Atlantic | 6 | 10.88 | 12.15 | 17.98 |
| NE Atlantic | 7 | 8.86 | 8.29 | 12.83 |
| N. Pacific | 6 | 10.10 | 6.69 | 32.48 |
| NW Atlantic | 16 | 6.32 | 2.62 | 39.82 |
| Indian | 16 | 5.08 | 3.64 | 14.58 |
| S. Atlantic | 16 | 3.87 | 0.00 | 25.07 |
| W. Pacific | 8 | 2.92 | 2.16 | 8.04 |
| Tasman | 8 | 0.42 | 0.20 | 1.47 |
| North Sea | 8 | 0.00 | 0.00 | 0.02 |
Savings concentrate where a strong weather gradient meets a resistance-sensitive operating point; calm or laterally-constrained routes save little, a heavy storm tail saves substantially.
On the ECA-affected lanes the two objectives diverge: cost-optimal routing trades a small fuel increase for a voyage-cost reduction by shortening time on the pricier ECA-compliant fuel.
| Vessel | diverged (of 96) | mean cost saving | max cost saving | mean fuel change |
|---|---|---|---|---|
| S175 | 31 | +2.36 % | +10.9 % | +1.48 % |
| KVLCC2 | 24 | +2.15 % | +10.0 % | +0.64 % |
| JBC | 20 | +1.34 % | +9.6 % | +0.97 % |
status == ok instances (route computed) of 96:
| Vessel | ok / 96 | Notes |
|---|---|---|
| S175 | 96 / 96 | all feasible |
| KVLCC2 | 75 / 96 | deep draft (20.8 m > Suezmax 20.1 m) makes some lanes infeasible irrespective of weather |
| JBC | 76 / 96 | deep draft (16.5 m > Panama 15.2 m) likewise |
Of the ok instances, a further subset have a weather-blind baseline that is
itself infeasible (the storm exceeds the engine MCR or seakeeping limit at the
service speed): these carry no defined saving percentage and are excluded from
the saving statistics above (hence n = 91 / 67 / 66 for the saving rows, below the
ok counts). On these "storm-enabled" voyages, weather routing does not merely
save fuel — it enables the crossing at all.
| Quantity | S175 min-fuel |
|---|---|
| Waypoints per instance | 17 – 1003 |
| Router solve time per route | median 183 s, mean 335 s, range 39 s – 24 min |
Provided so methods can be compared on computational cost as well as savings.