Archived snapshot from the 18Z Thu 8 Oct 2026 model cycle. Forecasts have changed since. See the latest analysis.
NHC model cycle 18Z Thu 8 Oct 2026 · central Gulf of Mexico
Physics-based, statistical and AI models are all forecasting the same Gulf hurricane. This page shows where they agree, where they split, and whether the split is unusual. It uses NOAA's public model file and three seasons of verification.
Nearly every model agrees on the shape. Isaias strengthens for 12 to 24 hours, then weakens fast after +24 h as NHC's track takes the center inland over the north-central Gulf coast. NHC's track is inland by +36 h. The models disagree on two numbers: how strong the storm is at the coast, and how fast it falls apart over land.
One detail matters for reading Fig. 1. The early aids all start at 75 kt, the intensity NHC was using when they ran. Hurricane Hunter data then led NHC to raise the 18Z intensity to 85 kt in its 21Z advisory, so the official forecast starts 10 kt above every aid. NHC's discussion describes its intensity forecast as "at the high end of the model guidance."
Isaias sat nearly stationary in the Bay of Campeche for about two days, first as a weak disturbance and then as a depression. It then turned east-northeast and intensified rapidly, from 35 kt at 06Z Wednesday to 70 kt at 06Z Thursday. That is 35 kt in 24 hours. It reached 85 kt by 18Z Thursday in NHC's operational best track. Early on, the models disagreed about both the turn and the intensification. Some moved the system east quickly and strengthened it. Others kept it slow, weak and farther south. Those forecasts can now be checked against NHC's best track.
The 18Z Tuesday runs, made 48 h before Thursday's 18Z position, show who handled the turn. DeepMind's raw run placed the center 19 nmi from where it was, and NHC's official forecast 20 nmi. That makes two runs in a row where raw DeepMind was the closest at two days. The 12Z Tuesday run was 6 nmi off. The raw HAFS-A and HAFS-B runs from 18Z Tuesday missed by 183 and 163 nmi, after the 12Z Tuesday runs had been within 11 and 24 nmi. Single runs vary, which is why the averages below matter more. Over every forecast so far that all the aids below share, the interpolated DeepMind aid and NHC have the smallest average track errors, both 17 nmi.
The same raw 18Z Tuesday DeepMind run forecast 76 kt for 18Z Thursday, against 85 kt in the best track. Among the raw runs, only HMON, a previous-generation NOAA model, came closer, at 81 kt. The raw HAFS-A run had 50 kt and the Navy's COAMPS-TC 104 kt. Across the shared forecasts, the smallest intensity errors belong to the corrected consensus (HCCA), the neural-network consensus (NNIC), DeepMind, NHC and COAMPS-TC, all between 4 and 6 kt.
Nearly every aid was too slow to intensify Isaias. Every aid in the table except COAMPS-TC has a low bias. The global models ran the lowest, as expected at their resolution, and the statistical LGEM was 14 kt low. Two days out, the raw GFS and its ensemble mean put the center 106 and 154 nmi from where it was. COAMPS-TC, the best interpolated physics model on intensity, had the largest track error of the regional models.
| Aid | Type | Intensity error | Intensity bias | Track error |
|---|---|---|---|---|
| HCCA | Consensus | 4.4 kt | −3.4 | 30 nmi |
| NNIC | Machine learning | 5.2 kt | −4.4 | — |
| GDMI | AI (DeepMind) | 5.4 kt | −4.2 | 17 nmi |
| OFCL | NHC official | 5.6 kt | −5.0 | 17 nmi |
| CTCI | Regional physics | 5.8 kt | +4.6 | 86 nmi |
| HWFI | Regional physics | 7.1 kt | −7.1 | 42 nmi |
| HFAI | Regional physics | 8.9 kt | −8.4 | 45 nmi |
| HMNI | Regional physics | 9.0 kt | −8.9 | 57 nmi |
| SHIP / DSHP | Statistical | 9.2 kt | −8.4 | — |
| HFBI | Regional physics | 10.4 kt | −7.7 | 45 nmi |
| AVNI | Global physics | 13.4 kt | −13.4 | 48 nmi |
| LGEM | Statistical | 14.9 kt | −14.2 | — |
| AEMI | Global physics | 22.0 kt | −22.0 | 67 nmi |
Mean absolute errors at +12 to +48 h, over the 16 forecast cycle and lead time pairs that every listed aid shares (a homogeneous sample), scored against NHC's operational best track through 18Z Thursday. Statistical and ML aids borrow another model's track, so their track errors are not listed. This covers two to three days of one storm against a preliminary best track. It shows which models handled this storm's genesis and rapid intensification, not which models are better overall.
At +24 h the regional physics models span 83–96 kt, with HAFS-A highest and the Navy's COAMPS-TC close behind at 94 kt. The global physics, statistical and interpolated DeepMind aids sit between 76 and 82 kt. Regional hurricane models follow the storm with nested grids (about 2 km for HAFS) that can resolve the eyewall. Global models, with grid spacings of about 10 km or more, cannot, and they usually run weak at the peak. NHC's 85 kt at +24 h is mid-range. Its 95 kt peak at +12 h is above every aid at that time, the highest being COAMPS-TC at 93 kt, partly because NHC starts from 85 kt and the aids from 75 kt.
At +48 h the DeepMind aid (GDMI) is at 15 kt, the regional physics models are at 19–34 kt and the GFS is at 33 kt. The UK global model and the Canadian ensemble mean are still at 44 and 54 kt, and the deterministic Canadian run (CMCI, not plotted) is at 40 kt. SHIPS, a statistical model with no land in it by design, holds 52 kt. Its land-aware version, Decay-SHIPS, drops to 37 kt. The 15 kt between those two is the land effect as that statistical scheme models it, which makes the pair a useful control.
Some of the intensity spread is really track spread. A model that brings the center ashore earlier, or moves it farther inland, also weakens it earlier. At +36 h HAFS-A puts the center at 30.6°N 87.6°W and the DeepMind aid at 30.8°N 86.9°W, about 70 km apart. HAFS-A is at 64 kt there and the DeepMind aid at 44 kt (Fig. 4).
The DeepMind line on most spaghetti plots is GDMI, an interpolated aid. The raw model output (GDMN) arrives too late for the forecast cycle it would serve. NHC therefore takes the previous run and adjusts it to the storm's intensity at the start time.
For Isaias, the raw 12Z DeepMind run went from its 70 kt starting value to 106 kt at 18Z, six hours later. The best-track intensity at 18Z is 85 kt, and the aids were adjusted to 75 kt, the value NHC was using when they ran. After the adjustment, the 18Z aid runs 18–31 kt below the raw run at 0, +12 and +24 h, and 5–11 kt below it at +36 and +48 h. The raw run peaks at 109 kt at +6 h, above NHC's 95 kt. In the previous cycle the raw run also started high (82 kt against 70 kt observed) and peaked at NHC's 95 kt. In both cycles the interpolated aid sits well below the raw model.
"AI is less aggressive" is partly an artifact of this adjustment. Compare raw and interpolated output before drawing conclusions about any model class.
Intensity forecasts over land depend on when and where the center crosses the coast. The tracks below cluster tightly toward the Alabama and western Florida Panhandle coast. At +36 h every plotted aid has the center between 30.6°N and 31.5°N and between 86.8°W and 88.1°W, at or north of the coast. Their small east–west and timing differences explain part of the +36 to +48 h intensity spread.
Not among the core models. The split comes from the models outside that core.
We took eight aids that ran in every season from 2023 to 2025: HAFS-A, HAFS-B, HWRF, HMON, COAMPS-TC, GFS, Decay-SHIPS and LGEM. For every Atlantic forecast cycle that started at hurricane strength, we computed their spread, the standard deviation across models. By that measure, Isaias's disagreement is ordinary at +24 and +48 h (46th and 27th percentiles). At +72 h it is low, in the 8th percentile, when every core model has the storm well inland and weak. Spread in knots shrinks over land, so the low +72 h value partly reflects that (see the Fig. 5 caption).
Isaias stands out once the other models come back in. Adding the DeepMind aid, the UK global model, the Canadian ensemble mean and SHIPS raises the +48 h spread from 6.2 kt to 11.9 kt. That figure uses a different set of models, so it can't be compared to the histograms. It does show where the disagreement comes from: models that treat land very differently.
NOAA's global ensemble (GEFS) runs 31 versions of the GFS, each with slightly perturbed starting conditions and model physics. At this resolution its intensities are much too weak. At 12Z the members started between 49 and 70 kt against an observed 70 kt, so read the plume for timing, not strength. On the 18Z clock, 20 of the 31 members weaken fastest between +30 and +42 h, and all 31 lose at least 15 kt in that window. It spans 00Z to 12Z Saturday (7 PM Friday to 7 AM Saturday CDT), consistent with NHC's forecast landfall late Friday or early Saturday. That is independent evidence that the timing of the move inland is fairly well constrained.
The European (ECMWF) model and ensemble are not in NOAA's public model file, so they are not shown. Google's model is itself an ensemble, but the public file carries one forecast per cycle.
Each version of this page, one per NHC model cycle, is committed to a public repository before landfall, and the commit history timestamps it. After the storm, each aid will be scored against NHC's best track from +12 to +72 h, on the questions this storm raises:
One storm can't settle which model is better. NHC's annual verification reports are where models get judged. This page is a worked example of reading the guidance.
This page is updated as new NHC model cycles arrive. Each earlier version is preserved unchanged.