Janela · offshore energy · operational windows

The operational window, decided.

For production areas, platforms and oil and LNG terminals: the sea forecast with its error measured against satellites, decided against the published limit — the Harbour Master's (Capitania dos Portos), DNV's or yours. Each hour comes out with one of four verdicts and the reason.

CLEARED

Every limit holds at the unfavourable edge of the forecast band, at every step of the window.

BLOCKED

A limit fails even at the favourable edge. It comes with the witness: the hour and the variable.

UNDECIDED

The band straddles the limit. It comes with the threshold: how far the edge must move to flip.

NEEDS DATA

The forecast clears, but the rule limits something nobody forecasts here, such as current and visibility, and it says what.

sites
13
7 production areas, 1 platform, 4 terminals and 1 coastal point; the model node of each one, named
published limits
8 operations
from 3 acts of the Maritime Authority (NPCPs, the Harbour Masters' local rules, and ordinances), with page and literal quote
32 years counted
93,504
3-hourly sea states per area, 1993–2024: windows counted exactly, nothing fitted
scoreboard
42,403
forecasts committed before the target hour since 2026-10-06; 5522 scored so far

What is decided is the forecast band at the forecast's own steps, at the model node: not the sea between steps, not the berth, not a probability. It is not an approval of an operation; it is evidence a surveyor can redo.

A portrait of the ECMWF run of 2026-10-31 00 UTC. The numbers on this page — the week, the decisions, the scoreboard — are that day's, to document the method. Today's windows and today's scoreboard are in the app: /janela/.

who it is for

Six decisions, and the door to each one in the app.

Each card opens the app on the question already: the right operation, mode and criterion. The example limits say they are examples; replace them with your procedure's, and they stay in your browser only.

1 · the week

Seven days in 6-hour steps: each operation decided against its limit.

ECMWF forecast of 2026-10-31 00 UTC, read at 2026-10-31 09:08 UTC; Brasília time (UTC−3). In this run, 232 published decisions: 26 CLEARED, 0 BLOCKED, 44 UNDECIDED, 162 NEEDS DATA. Tap a cell: the reason, the witness or the threshold that would flip the verdict.

Production areas
Platform
Terminals
Florianópolis
your limit · decided in this tab

An example; replace it with your procedure's limits. The same measured band, the same modules as the record, decided on your machine.

  • CLEARED · every limit holds at the band's unfavourable edge
  • BLOCKED · a limit fails even at the favourable edge
  • UNDECIDED · the band straddles a limit
  • NEEDS DATA · the rule limits something with no measured band here
Terminals

Ilha Grande Bay (TEBIG)

Forecast read at node 23.00°S 44.00°W of ECMWF's 0.25° grid, 24 km from 23.05°S 44.23°W: the open sea of the APPROACH, not the berth. The measured band comes from Santos Basin — pre-salt: an altimeter cannot see inside a bay.

operation · Brasília time30Sat Oct 31Sun Nov 1Mon Nov 2Tue Nov 3Wed Nov 4Thu Nov 5Fri Nov 6
2103091521030915210309152103091521030915210309152103091521
TEBIG — East Channel transit (Aframax, Suezmax, VLCC)Hs ≤ 2.0 m · wind ≤ 20 knsource
Ilha Grande Bay out of redHs < 2.5 m · wind < 30 kn · visibility > 0.5 NMsource
Your limitHs ≤ 2.5 m · wind ≤ 20 kndecided in this tab
CLEAREDTEBIG — East Channel transit (Aframax, Suezmax, VLCC) · Sat Oct 31, 09:00 (12:00 UTC) · lead +12 h

Every limit holds at the unfavourable edge of the measured band: Hs reaches at most 1.33 m, and the limit is Hs ≤ 2.0 m; wind reaches at most 11.1 kn, and the limit is wind ≤ 20 kn.

decided over the measured band, claim 9/10: Hs [0.29, 1.33] m (335 pairs) · wind [0.0, 11.1] kn (335 pairs)

forecast ECMWF deterministic: Hs 0.35 m · Tp 11.3 s · from 176° · wind 5.1 kn · gust 7.0 kn. Ensemble, the central 40 of 50: Hs 0.33 to 0.36 m.

the operator's draft ≤ 13.2 m · speed ≤ 12 kn: vessel attributes, checked by whoever operates, never decided here.

00.511.522.53Hs (m)Sat 31Sun 1Mon 2Tue 3Wed 4Thu 5Fri 6< 2.5 m · yours ≤ 2.5 m≤ 2.0 m
  • measured band, what is decided: the forecast × the observed/forecast ratio measured at the lead
  • halo: the band is Santos Basin — pre-salt's, carried here, and nothing measured the error at this node
  • ECMWF deterministic forecast, dashed
  • ECMWF ensemble, the central 40 of 50
  • limit
Records: corpus/janela/feed/20261031.json.gz · sha256 b67d72525083; corpus/janela/feed-noaa/20261031.json.gz · sha256 f39741fb4fbe; certs/janela-bands.json · sha256 1f14a51fac10; certs/janela-bands-sergipe.json · sha256 175a5ce26d4a; certs/janela-bands-noaa.json · sha256 beaf366df295; certs/janela-bands-noaa-sergipe.json · sha256 991211a94b70The build refuses this figure if a published decision, re-decided by this page's own function over the band drawn here, differs from the record.
05101520253035wind at 10 m (kn)Sat 31Sun 1Mon 2Tue 3Wed 4Thu 5Fri 6< 30 kn≤ 20 kn · yours ≤ 20 kn
  • measured band, what is decided: the forecast ± the error measured in m/s at the lead
  • halo: the wind band was not validated outside calibration, here or anywhere
  • ECMWF deterministic forecast, dashed
  • gust, deterministic, dashed
  • limit
decided
exact arithmetic backs it
computed
float, or a number someone printed; nothing decided it
Records: corpus/janela/feed/20261031.json.gz · sha256 b67d72525083; corpus/janela/feed-noaa/20261031.json.gz · sha256 f39741fb4fbe; certs/janela-bands.json · sha256 1f14a51fac10; certs/janela-bands-sergipe.json · sha256 175a5ce26d4a; certs/janela-bands-noaa.json · sha256 beaf366df295; certs/janela-bands-noaa-sergipe.json · sha256 991211a94b70The build refuses this figure if a published decision, re-decided by this page's own function over the band drawn here, differs from the record.

At build: 232 of 232 published decisions redone by this page's function, equal to the record.

why these numbers stand as they do
  • computed — a forecast or an estimate in floating point; nothing decided it
2 · the month

32 years counted: how many windows the sea opens each month, and how many the table's α closes.

For the 6 production areas, every 3-hourly operation start from 1993 to 2024 is counted: a window of TR hours with Hs under the limit (OPLIM) fits or does not. The second bar redoes the count with the forecast limit DNV says to use, OPWF = α × OPLIM (Table 4-1, TPOP = TR/2); where there are enough satellite pairs, the third uses the α estimated at the site itself (“The site's alpha”, below). Exact fractions of integers, nothing fitted.

area
Hs limit (OPLIM)
window (TR)
the sea allows
24.2%

Hs ≤ 2.0 m (OPLIM) for the whole 48 h window

Table 4-1 leaves
3.3%

α 0.73 → OPWF 1.46 m

the site's alpha would leave
10.8%

α 0.86 → OPWF 1.72 m; α chosen in the estimated interval, exact count

Santos Basin — pre-salt, wave limit 2.0 m, window of 48 h: the sea allows 24.2% of operation starts over the year (22,657 of 93,488). With the α of DNV's Table 4-1 (0.73 × 2.0 m → OPWF 1.46 m, TPOP 24 h), 3.3% remain. The difference, 20.9 points, is what the table's α costs if the forecast were perfect. The month that loses most: Jan (36.8% → 5.2%). Mean wait until the next window: 158.3 h → 1,313.4 h. If the surveyor adopted the site's α, 0.86 (the lower end of the 90% interval of the estimate 0.880), the OPWF would be 1.72 m and 10.8% would remain, with a mean wait of 401.1 h. The site's α is an estimate and its lower end a choice; the count over it is exact.

0%25%50%75%100%37516Jan39821Feb401022Mar33417Apr2128May24311Jun1515Jul1014Aug601Sep1304Oct2017Nov33414Dec
  • the sea allows: Hs ≤ 2.0 m (OPLIM) for the whole window
  • with the table's α: Hs ≤ 1.46 m (OPWF)
  • if the site's α were adopted: Hs ≤ 1.72 m (chosen: the lower end of the estimated α; the count over it is exact)
decided
exact arithmetic backs it
chosen
one member of a set the data admits
Hindcast Ifremer WAVEWATCH III GLOBMULTI at each area's node, 1993–2024 (384 monthly files, each checked by sha256 against corpus/ww3-points/meta.json). Each bar: starts with a window ÷ starts determined in the month; the count is instruments/window/workability.js's, in integers. Mean wait: hours from a start to the next start with a window. Record: certs/janela-workability.json · sha256 7e9a70222fe3The build refuses this figure if the battery (apps/janela/battery.js) re-derives a workability cell from the pinned hindcast unequal to the record, or a module the record was written with has changed.
the exact counts, month by month
monthOPLIMstarts with a window / determinedthe table's OPWFstarts with a window / determinedcost of the alpha (points)the site alpha's OPWFstarts with a window / determined
Jan36.8%2,920 / 7,9365.2%411 / 7,93631.616.1%1,279 / 7,936
Feb39.2%2,833 / 7,2328.1%587 / 7,23231.121.3%1,542 / 7,232
Mar39.7%3,152 / 7,9369.6%762 / 7,93630.121.7%1,724 / 7,936
Apr33.3%2,557 / 7,6804.3%329 / 7,68029.016.8%1,289 / 7,680
May21.4%1,698 / 7,9362.1%166 / 7,93619.38.4%665 / 7,936
Jun23.9%1,837 / 7,6803.3%256 / 7,68020.610.6%815 / 7,680
Jul15.2%1,207 / 7,9360.9%74 / 7,93614.35.0%395 / 7,936
Aug10.3%815 / 7,9360.9%73 / 7,9369.33.6%284 / 7,936
Sep6.0%464 / 7,6800.1%9 / 7,6805.91.2%90 / 7,680
Oct13.0%1,032 / 7,9360.3%25 / 7,93612.73.8%304 / 7,936
Nov20.0%1,535 / 7,6801.3%100 / 7,68018.77.1%548 / 7,680
Dec32.9%2,607 / 7,9203.8%297 / 7,92029.214.3%1,132 / 7,920
year24.2%22,657 / 93,4883.3%3,089 / 93,48820.910.8%10,067 / 93,488
why these numbers stand as they do
  • chosen — counted exactly over the site's α at the lower end of its 90% interval, one member of the interval, chosen for caution
  • computed — a forecast or an estimate in floating point; nothing decided it
3 · the site's alpha

The α the satellites measure here, beside the α of the North Sea table.

DNV's α shrinks the forecast limit (OPWF = α × OPLIM) to cover the forecast's error. The table was calibrated in the North Sea; here the same method is applied to each site's ECMWF × altimeter pairs. The method reproduces DNV's Table 4-1 in 16 of 20 cells (±0.02) before looking at Brazil.

what this is, and what it is not

The site's α is a statistical estimate (floating point, resampled by day, 90% interval), not a decided interval; the counts beside it are exact. The 1-in-10,000 tail behind the α is a model extrapolation: three years of satellite data do not observe it. It is evidence for the surveyor to redo, never an approval.

In each cell: the site's α, the 90% interval (rounded outward) and the α of Table 4-1. “above”: the whole interval lies above the table, and the North Sea table is conservative here; “below”: the table is not conservative here; “contains”: it cannot be told apart from the table. Under 30 pairs, the cell says only how many pairs there are.

site · design Hs 2 mTPOP 12 hTPOP 24 hTPOP 36 hTPOP 48 hTPOP 72 h
Campos0.910.88–0.93table 0.76 · above0.870.85–0.90table 0.73 · above0.890.87–0.92table 0.71 · above0.870.85–0.90table 0.68 · above0.860.84–0.88table 0.63 · above
Platform P-250.880.86–0.90table 0.76 · above0.860.84–0.88table 0.73 · above0.870.85–0.90table 0.71 · above0.870.85–0.89table 0.68 · above0.850.83–0.87table 0.63 · above
Espírito Santo0.930.91–0.95table 0.76 · above0.910.89–0.93table 0.73 · above0.900.88–0.92table 0.71 · above0.900.88–0.92table 0.68 · above0.880.86–0.90table 0.63 · above
Florianópolis0.760.73–0.80table 0.76 · contains0.750.73–0.78table 0.73 · above0.730.70–0.76table 0.71 · contains0.720.70–0.75table 0.68 · above0.720.70–0.74table 0.63 · above
Amazon Mouth0.960.94–0.99table 0.76 · above0.950.92–0.97table 0.73 · above0.930.90–0.95table 0.71 · above0.920.90–0.94table 0.68 · above0.900.88–0.93table 0.63 · above
Pelotas0.880.86–0.90table 0.76 · above0.890.87–0.91table 0.73 · above0.840.82–0.86table 0.71 · above0.840.82–0.86table 0.68 · above0.790.77–0.82table 0.63 · above
Potiguar0.960.95–0.98table 0.76 · above0.950.92–0.97table 0.73 · above0.940.93–0.96table 0.71 · above0.930.92–0.95table 0.68 · above0.920.90–0.93table 0.63 · above
Santos0.890.87–0.92table 0.76 · above0.880.86–0.91table 0.73 · above0.860.83–0.89table 0.71 · above0.880.85–0.90table 0.68 · above0.840.81–0.86table 0.63 · above
Sergipe-Alagoas0.690.66–0.72table 0.76 · below0.720.70–0.74table 0.73 · contains0.690.67–0.72table 0.71 · contains0.700.68–0.72table 0.68 · above0.680.67–0.71table 0.63 · above

What this changes in the month. The same exact count as “The month”, redone with the site's α in place of the table's: the lower end of the 90% interval, rounded down to 0.01, so as not to gain windows from the optimistic part of the estimate. It is what would remain if the surveyor adopted the site's α.

areaOPLIM · TRthe sea allowsTable 4-1 leavesthe site's alpha would leave
Santos2.0 m · 24 h35.0%9.1% (α 0.76)19.6% (α 0.87)
Santos2.0 m · 48 h24.2%3.3% (α 0.73)10.8% (α 0.86)
Santos2.0 m · 72 h17.1%1.4% (α 0.71)4.8% (α 0.83)
Campos2.0 m · 24 h40.7%9.6% (α 0.76)23.2% (α 0.88)
Campos2.0 m · 48 h30.1%3.5% (α 0.73)12.1% (α 0.85)
Campos2.0 m · 72 h23.1%1.3% (α 0.71)9.5% (α 0.87)
Espírito Santo2.0 m · 24 h58.8%19.3% (α 0.76)43.6% (α 0.91)
Espírito Santo2.0 m · 48 h49.1%9.6% (α 0.73)30.9% (α 0.89)
Espírito Santo2.0 m · 72 h41.5%5.2% (α 0.71)22.9% (α 0.88)
Pelotas2.0 m · 24 h26.2%6.3% (α 0.76)13.2% (α 0.86)
Pelotas2.0 m · 48 h16.2%1.9% (α 0.73)7.1% (α 0.87)
Pelotas2.0 m · 72 h10.2%0.6% (α 0.71)2.4% (α 0.82)
Potiguar2.0 m · 24 h78.4%24.2% (α 0.76)68.8% (α 0.95)
Potiguar2.0 m · 48 h71.4%12.0% (α 0.73)53.9% (α 0.92)
Potiguar2.0 m · 72 h65.8%6.3% (α 0.71)50.1% (α 0.93)
Amazon Mouth2.0 m · 24 h67.4%36.6% (α 0.76)60.9% (α 0.94)
Amazon Mouth2.0 m · 48 h63.1%26.3% (α 0.73)54.3% (α 0.92)
Amazon Mouth2.0 m · 72 h59.9%18.9% (α 0.71)48.8% (α 0.90)
Campos · 5,963 pairs over 594 days
design HsTPOP 12 hTPOP 24 hTPOP 36 hTPOP 48 hTPOP 72 h
1 m0.850.78–0.93table 0.65 · above0.870.84–0.91table 0.63 · above0.840.78–0.89table 0.62 · above0.860.83–0.89table 0.60 · above0.800.76–0.84table 0.55 · above
2 m0.910.88–0.93table 0.76 · above0.870.85–0.90table 0.73 · above0.890.87–0.92table 0.71 · above0.870.85–0.90table 0.68 · above0.860.84–0.88table 0.63 · above

Design Hs 4 and 6 m: not enough pairs at any lead (at most 20; the minimum is 30).

Exact counts, summed over the 5 leads (each pair falls in one only): when the forecast was at or below Table 4-1's OPWF, how many times the satellite saw Hs above OPLIM.

OPLIMforecasts ≤ OPWFsatellite > OPLIMsatellite > 1.5 × OPLIM
1.5 m700
2.0 m33720
2.5 m1,03720
3.0 m1,67300
3.5 m2,15300
4.0 m2,43500
Platform P-25 · 6,038 pairs over 593 days
design HsTPOP 12 hTPOP 24 hTPOP 36 hTPOP 48 hTPOP 72 h
1 m0.760.71–0.84table 0.65 · above0.870.84–0.91table 0.63 · above0.750.70–0.82table 0.62 · above0.840.82–0.88table 0.60 · above0.770.73–0.80table 0.55 · above
2 m0.880.86–0.90table 0.76 · above0.860.84–0.88table 0.73 · above0.870.85–0.90table 0.71 · above0.870.85–0.89table 0.68 · above0.850.83–0.87table 0.63 · above

Design Hs 4 and 6 m: not enough pairs at any lead (at most 13; the minimum is 30).

Exact counts, summed over the 5 leads (each pair falls in one only): when the forecast was at or below Table 4-1's OPWF, how many times the satellite saw Hs above OPLIM.

OPLIMforecasts ≤ OPWFsatellite > OPLIMsatellite > 1.5 × OPLIM
1.5 m600
2.0 m36200
2.5 m1,08440
3.0 m1,75300
3.5 m2,23400
4.0 m2,48600
Espírito Santo · 5,997 pairs over 605 days
design HsTPOP 12 hTPOP 24 hTPOP 36 hTPOP 48 hTPOP 72 h
1 m0.840.82–0.88table 0.65 · above0.850.82–0.89table 0.63 · above0.810.77–0.85table 0.62 · above0.830.80–0.86table 0.60 · above0.820.80–0.84table 0.55 · above
2 m0.930.91–0.95table 0.76 · above0.910.89–0.93table 0.73 · above0.900.88–0.92table 0.71 · above0.900.88–0.92table 0.68 · above0.880.86–0.90table 0.63 · above

Design Hs 4 and 6 m: not enough pairs at any lead (at most 6; the minimum is 30).

Exact counts, summed over the 5 leads (each pair falls in one only): when the forecast was at or below Table 4-1's OPWF, how many times the satellite saw Hs above OPLIM.

OPLIMforecasts ≤ OPWFsatellite > OPLIMsatellite > 1.5 × OPLIM
1.5 m2700
2.0 m52700
2.5 m1,40040
3.0 m2,00300
3.5 m2,33300
4.0 m2,51600
Florianópolis · 5,429 pairs over 604 days
design HsTPOP 12 hTPOP 24 hTPOP 36 hTPOP 48 hTPOP 72 h
1 m0.690.64–0.76table 0.65 · contains0.680.63–0.75table 0.63 · above0.660.60–0.73table 0.62 · contains0.640.60–0.69table 0.60 · above0.610.56–0.67table 0.55 · above
2 m0.760.73–0.80table 0.76 · contains0.750.73–0.78table 0.73 · above0.730.70–0.76table 0.71 · contains0.720.70–0.75table 0.68 · above0.720.70–0.74table 0.63 · above

Design Hs 4 and 6 m: not enough pairs at any lead (at most 10; the minimum is 30).

Exact counts, summed over the 5 leads (each pair falls in one only): when the forecast was at or below Table 4-1's OPWF, how many times the satellite saw Hs above OPLIM.

OPLIMforecasts ≤ OPWFsatellite > OPLIMsatellite > 1.5 × OPLIM
1.5 m15240
2.0 m876110
2.5 m1,550110
3.0 m1,965120
3.5 m2,23160
4.0 m2,37460
Amazon Mouth · 4,768 pairs over 517 days
design HsTPOP 12 hTPOP 24 hTPOP 36 hTPOP 48 hTPOP 72 h
1 m0.910.88–0.94table 0.65 · above0.890.87–0.92table 0.63 · above0.880.86–0.91table 0.62 · above0.880.86–0.91table 0.60 · above0.860.84–0.89table 0.55 · above
2 m0.960.94–0.99table 0.76 · above0.950.92–0.97table 0.73 · above0.930.90–0.95table 0.71 · above0.920.90–0.94table 0.68 · above0.900.88–0.93table 0.63 · above

Design Hs 4 and 6 m: not enough pairs at any lead (at most 0; the minimum is 30).

Exact counts, summed over the 5 leads (each pair falls in one only): when the forecast was at or below Table 4-1's OPWF, how many times the satellite saw Hs above OPLIM.

OPLIMforecasts ≤ OPWFsatellite > OPLIMsatellite > 1.5 × OPLIM
1.5 m4000
2.0 m46000
2.5 m1,16100
3.0 m1,68500
3.5 m1,96100
4.0 m2,12100
Pelotas · 7,847 pairs over 699 days
design HsTPOP 12 hTPOP 24 hTPOP 36 hTPOP 48 hTPOP 72 h
1 m0.830.79–0.88table 0.65 · above0.840.80–0.89table 0.63 · above0.800.75–0.85table 0.62 · above0.810.77–0.84table 0.60 · above0.780.75–0.81table 0.55 · above
2 m0.880.86–0.90table 0.76 · above0.890.87–0.91table 0.73 · above0.840.82–0.86table 0.71 · above0.840.82–0.86table 0.68 · above0.790.77–0.82table 0.63 · above
4 m0.850.81–0.89table 0.79 · above0.850.80–0.91table 0.76 · above0.800.75–0.86table 0.73 · above0.820.78–0.87table 0.71 · above0.810.77–0.84table 0.68 · above

Design Hs 6 m: not enough pairs at any lead (at most 6; the minimum is 30).

Exact counts, summed over the 5 leads (each pair falls in one only): when the forecast was at or below Table 4-1's OPWF, how many times the satellite saw Hs above OPLIM.

OPLIMforecasts ≤ OPWFsatellite > OPLIMsatellite > 1.5 × OPLIM
1.5 m1100
2.0 m32830
2.5 m90600
3.0 m1,68340
3.5 m2,34520
4.0 m2,81930
Potiguar · 4,752 pairs over 487 days
design HsTPOP 12 hTPOP 24 hTPOP 36 hTPOP 48 hTPOP 72 h
1 m0.920.89–0.96table 0.65 · above0.860.83–0.90table 0.63 · above0.900.87–0.93table 0.62 · above0.870.84–0.90table 0.60 · above0.850.82–0.87table 0.55 · above
2 m0.960.95–0.98table 0.76 · above0.950.92–0.97table 0.73 · above0.940.93–0.96table 0.71 · above0.930.92–0.95table 0.68 · above0.920.90–0.93table 0.63 · above

Design Hs 4 and 6 m: not enough pairs at any lead (at most 0; the minimum is 30).

Exact counts, summed over the 5 leads (each pair falls in one only): when the forecast was at or below Table 4-1's OPWF, how many times the satellite saw Hs above OPLIM.

OPLIMforecasts ≤ OPWFsatellite > OPLIMsatellite > 1.5 × OPLIM
1.5 m200
2.0 m17000
2.5 m88000
3.0 m1,64800
3.5 m1,97800
4.0 m2,08100
Santos · 5,193 pairs over 565 days
design HsTPOP 12 hTPOP 24 hTPOP 36 hTPOP 48 hTPOP 72 h
1 m0.870.82–0.93table 0.65 · above0.870.83–0.91table 0.63 · above0.850.80–0.90table 0.62 · above0.840.81–0.87table 0.60 · above0.810.77–0.85table 0.55 · above
2 m0.890.87–0.92table 0.76 · above0.880.86–0.91table 0.73 · above0.860.83–0.89table 0.71 · above0.880.85–0.90table 0.68 · above0.840.81–0.86table 0.63 · above

Design Hs 4 and 6 m: not enough pairs at any lead (at most 22; the minimum is 30).

Exact counts, summed over the 5 leads (each pair falls in one only): when the forecast was at or below Table 4-1's OPWF, how many times the satellite saw Hs above OPLIM.

OPLIMforecasts ≤ OPWFsatellite > OPLIMsatellite > 1.5 × OPLIM
1.5 m2100
2.0 m32120
2.5 m81710
3.0 m1,32810
3.5 m1,74700
4.0 m2,04700
Sergipe-Alagoas · 3,096 pairs over 366 days

Caution: The platforms stand in 13-38 m of water 5-20 km off the coast, where an altimeter footprint is land-contaminated: the passes that count (547 in three years, >= 5 points within 100 km) see the sea 40-100 km out (median mean distance 70 km), deeper and rougher than at the platforms, while the forecast is read at the model's sea node 5 km off the coast. The measured ratio band therefore carries the shelf's gradient as well as the model's error (Hs at 24 h: 1.13-1.62 x the forecast, against 0.86-1.20 at Santos): it leans HIGH for the platforms — conservative by construction, and stated wherever it decides.

design HsTPOP 12 hTPOP 24 hTPOP 36 hTPOP 48 hTPOP 72 h
1 m0.640.61–0.67table 0.65 · contains0.620.59–0.65table 0.63 · contains0.640.61–0.67table 0.62 · contains0.630.60–0.65table 0.60 · above0.630.61–0.65table 0.55 · above
2 m0.690.66–0.72table 0.76 · below0.720.70–0.74table 0.73 · contains0.690.67–0.72table 0.71 · contains0.700.68–0.72table 0.68 · above0.680.67–0.71table 0.63 · above

Design Hs 4 and 6 m: not enough pairs at any lead (at most 0; the minimum is 30).

Exact counts, summed over the 5 leads (each pair falls in one only): when the forecast was at or below Table 4-1's OPWF, how many times the satellite saw Hs above OPLIM.

OPLIMforecasts ≤ OPWFsatellite > OPLIMsatellite > 1.5 × OPLIM
1.5 m31350
2.0 m1,026240
2.5 m1,253170
3.0 m1,34300
3.5 m1,35300
4.0 m1,35300
why these numbers stand as they do
  • computed — a forecast or an estimate in floating point; nothing decided it
  • chosen — counted exactly over the site's α at the lower end of its 90% interval, one member of the interval, chosen for caution
4 · the scoreboard

Every band is committed before the sea happens, and scored afterwards against the satellite.

NOAA ensemble (WAVEWATCH III), the central 25 of 31
6,282 committed

822 scored, 350 covered so far. Claim: the band contains the satellite's measurement in 3/4 of cases. Targets from 2026-10-07 06:00 to 2026-11-07 00:00 UTC, at 9 open-sea sites; first on 2026-10-07.

UNDER EVALUATION 22 of 30 trial days: nothing tested yet; the first reading is on day 30

ECMWF ensemble, the central 40 of 50
6,274 committed

820 scored, 333 covered so far. Claim: the band contains the satellite's measurement in 4/5 of cases. Targets from 2026-10-06 18:00 to 2026-11-07 00:00 UTC, at 9 open-sea sites; first on 2026-10-06.

UNDER EVALUATION 23 of 30 trial days: nothing tested yet; the first reading is on day 30

ECMWF AIFS ensemble (data-driven), the central 40 of 50
6,057 committed

780 scored, 410 covered so far. Claim: the band contains the satellite's measurement in 4/5 of cases. Targets from 2026-10-08 00:00 to 2026-11-07 00:00 UTC, at 9 open-sea sites; first on 2026-10-07.

UNDER EVALUATION 21 of 30 trial days: nothing tested yet; the first reading is on day 30

Janela's measured band, v1
5,600 committed

760 scored, 661 covered so far. Claim: the band contains the satellite's measurement in 9/10 of cases. Targets from 2026-10-06 18:00 to 2026-11-07 00:00 UTC, at 8 open-sea sites; first on 2026-10-06.

UNDER EVALUATION 23 of 30 trial days: nothing tested yet; the first reading is on day 30

NOAA's measured band (WAVEWATCH III), v1
5,392 committed

720 scored, 623 covered so far. Claim: the band contains the satellite's measurement in 9/10 of cases. Targets from 2026-10-07 18:00 to 2026-11-07 00:00 UTC, at 8 open-sea sites; first on 2026-10-07.

UNDER EVALUATION 21 of 30 trial days: nothing tested yet; the first reading is on day 30

The deciding band: ECMWF ∪ NOAA, v1
5,392 committed

720 scored, 673 covered so far. Claim: the band contains the satellite's measurement in 9/10 of cases. Targets from 2026-10-07 18:00 to 2026-11-07 00:00 UTC, at 8 open-sea sites; first on 2026-10-07.

UNDER EVALUATION 21 of 30 trial days: nothing tested yet; the first reading is on day 30

ECMWF AIFS's measured band, v1
5,384 committed

720 scored, 687 covered so far. Claim: the band contains the satellite's measurement in 9/10 of cases. Targets from 2026-10-08 00:00 to 2026-11-07 00:00 UTC, at 8 open-sea sites; first on 2026-10-07.

UNDER EVALUATION 21 of 30 trial days: nothing tested yet; the first reading is on day 30

Janela's measured band, Sergipe-Alagoas, v1
674 committed

60 scored, 47 covered so far. Claim: the band contains the satellite's measurement in 9/10 of cases. Targets from 2026-10-07 18:00 to 2026-11-07 00:00 UTC, at 1 open-sea sites; first on 2026-10-07.

UNDER EVALUATION 8 of 30 trial days: nothing tested yet; the first reading is on day 30

NOAA's measured band, Sergipe-Alagoas, v1
674 committed

60 scored, 55 covered so far. Claim: the band contains the satellite's measurement in 9/10 of cases. Targets from 2026-10-07 18:00 to 2026-11-07 00:00 UTC, at 1 open-sea sites; first on 2026-10-07.

UNDER EVALUATION 8 of 30 trial days: nothing tested yet; the first reading is on day 30

The deciding band in Sergipe-Alagoas: ECMWF ∪ NOAA, v1
674 committed

60 scored, 58 covered so far. Claim: the band contains the satellite's measurement in 9/10 of cases. Targets from 2026-10-07 18:00 to 2026-11-07 00:00 UTC, at 1 open-sea sites; first on 2026-10-07.

UNDER EVALUATION 8 of 30 trial days: nothing tested yet; the first reading is on day 30

What will be scored. The Hs a satellite altimeter measures: the exact mean of NOAA RADS's 1 Hz values within 100 km of the site, on the pass nearest the target hour and within 3 h of it, with at least 5 points. A target with no pass is never scored; the passes do not depend on the forecast, so the scored set is not chosen by the outcome.

The admission rule. One trial per target day: a proposer's bands scored for the same day share the same satellite passes and are not independent, so one alone decides the day, drawn by sha256(proposer|day) — the draw does not look at the result. The exact binomial tail P[X ≤ k], X ~ Binomial(m, claim), is read only at m = 30, 60, 120, 240… days, against 1/40, 1/80, 1/160…: summed, 1/20 over the proposer's whole life, not 1/20 per day. The first reading under the bar prunes this version for good; only a recalibrated version, with a new record, comes back. Days are treated as independent: a declared convention, not a theorem. Covering more than it says does not prune; the Winkler score, exact, charges a band that is too wide. Rule fixed on 2026-10-06, before the first scoring. Admission is lost by record, never by opinion.

On 2026-10-31, the date of this portrait. 5522 bands scored up to that date. Today's scoreboard is in the app: /janela/ → SCOREBOARD.

5 · what each number rests on

The arithmetic is exact; the premises were checked to different rungs, and the decision is worth the weakest.

Each premise has a rung, from the verification and validation ladder of modelling (Roache 1998; Oberkampf and Roy 2010; ASME V&V 20): validated, checked against the sea itself, held out of calibration; compared, against a model that keeps more of the physics; verified, the code against a reference solution of its own equations; assumed, adopted from a source or chosen. The rung is where the checking stopped, never a probability.

By the measured band, at a measured site

  1. validated
    The Hs band, at the measured sites calibrated before 2025-07 and judged over 15 months it did not see: the union ECMWF ∪ NOAA covered 96.5% of 21,259 satellite passes (it claims 9 in 10); of the CLEARED windows with Hs ≤ 2 m, the sea broke 0.33% (21 of 6,368). At Sergipe-Alagoas, judged separately (1,603 passes): the union covered 94.8%; ECMWF's band alone, 87.2%, below what it claims.
  2. assumed
    The wind band measured on the same 45,987 pairs (the altimeter's wind); it claims 9 in 10 by theorem, if the next error resembles the archive's. That was not tested outside calibration: the scoreboard and the check of the decisions look at Hs only.

With an Hs-only limit, the weakest premise sits at the rung validated; with a wind limit, at the rung assumed. Going forward: since 2026-10-07, of this criterion's CLEARED windows that a satellite saw at the open-sea sites, 3,191 held and 0 broke (Hs only; a descriptive count, the one-day windows overlap).

At a production unit or a terminal

  1. validated
    The Hs band, at the measured site it comes from the same band validated above, at the measured sites.
  2. assumed
    The band carried to a unit or a terminal each unit uses the band and the α of the nearest measured site, up to 350 km away; a terminal, its open-sea neighbour's. The forecast error there is assumed to hold here; nothing measured it at the unit's node, and an altimeter cannot see inside a bay.

The weakest premise sits at the rung assumed: on the chart, the carried band has a halo.

By DNV's Table 4-1

  1. assumed
    The α of Table 4-1 (DNV-OS-H101) adopted from DNV, calibrated in the North Sea; the 1-in-10,000 tail behind it is extrapolated: three years of satellite data do not observe it. On the pairs here, the 90% interval of the estimated α lies above the table's in 85 of 95 estimated cells and below it in 1: the table is conservative for the most part, if the Gaussian tail model holds.
  2. assumed
    The wind α (Table 4-6) adopted from DNV; not redone here.

The weakest premise sits at the rung assumed. Going forward: since 2026-10-07, of this criterion's CLEARED windows that a satellite saw at the open-sea sites, 3,348 held and 45 broke (Hs only; a descriptive count, the one-day windows overlap).

By the site's α

  1. verified
    The site-α method reproduces DNV's Table 4-1 in 16 of 20 cells (±0.02) before looking at Brazil; it does not reproduce Hs 2 m with TPOP 72 h, Hs 4 m with TPOP 48 h, Hs 4 m with TPOP 72 h, Hs 6 m with TPOP 72 h.
  2. assumed
    The α estimated at each site from three years of forecast × satellite pairs, in floating point (resampled by day, 90% interval); the decision uses the lower end, rounded down, a choice made for caution. The same 1-in-10,000 tail, extrapolated.
  3. assumed
    The wind α (Table 4-6) adopted from DNV; not redone here.

The weakest premise sits at the rung assumed. Going forward: since 2026-10-07, of this criterion's CLEARED windows that a satellite saw at the open-sea sites, 872 held and 31 broke (Hs only; a descriptive count, the one-day windows overlap).

The month and the campaign

  1. assumed
    The hindcast (Ifremer WAVEWATCH III) at the areas' 6 nodes No buoy at any of the areas' nodes. Checked against 14 deployments of the Brazilian Navy's buoys (PNBOIA, the Navy's buoy network, 2009–2018) at 9 other nodes of the margin, 25,215 pairs: bias from −0.65 to +0.22 m, RMS from 0.34 to 2.18 m; Hs ≤ 2 m for 65.5% of the time at the buoys and 64.1% in the hindcast (certs/janela-pnboia-check.json).
  2. assumed
    The α of Table 4-1 (DNV-OS-H101) adopted from DNV, calibrated in the North Sea; the 1-in-10,000 tail behind it is extrapolated: three years of satellite data do not observe it. On the pairs here, the 90% interval of the estimated α lies above the table's in 85 of 95 estimated cells and below it in 1: the table is conservative for the most part, if the Gaussian tail model holds.

Each count is exact over the hindcast; the weakest premise sits at the rung assumed: the past sea is a model's, and no buoy checked it at the areas' nodes.

What is shown and never decided

  1. assumed
    The current (Copernicus Marine, 1/12°) shown, never decided. The truth it will be checked against (GlobCurrent, rule current-truth-v1, fixed on 2026-10-07) is being gathered; no comparison recorded yet.
why these numbers stand as they do
  • computed — a floating-point evaluation (certs/janela-providers-eval.json)
  • computed — a floating-point evaluation (certs/janela-providers-eval-sergipe.json)
  • computed — a floating-point evaluation (certs/janela-pnboia-check.json)
6 · why trust it

The verdict is exact arithmetic over the limit as printed; the band is a claim the scoreboard audits.

Exact arithmetic. Every comparison is made in exact rationals, integers with no rounding: the forecast comes from the GRIB's packed integers (value = (R + X·2E)/10D), the wind in knots by 1 knot = 463/900 m/s, the limits as printed, with the decimal comma read as a point. No floating point decides. The edge is exact: an edge of exactly 2.0 against “< 2.0” is UNDECIDED, never CLEARED.

The measured band. Hs: the deterministic forecast × the exact interval of the observed/forecast ratio in the satellite × forecast pairs of the site and the lead (blocks of 12 h). Wind: the forecast ± the error measured in m/s at the lead (observed − forecast), floored at zero. Each band claims coverage of 9/10, a counting theorem under the assumption that the next error resembles the archive's 45,987 pairs at the same site and lead; the scoreboard audits that assumption in public.

In the app, the band rounded outward. The app (/janela/) publishes the bands of each site and each production unit in short decimals, rounded OUTWARD: Hs to 0.001 m and wind to 0.01 kn, the lower edge down and the upper edge up; the deterministic forecast goes as the 1 mm interval that contains it. Over a wider band, a CLEARED or a BLOCKED still holds over the exact one; only UNDECIDED can grow. The build checks this against this page's exact decisions every day and refuses if it fails. The units borrow the band and the α of the nearest measured site, up to 350 km away; farther, they say NEEDS DATA.

The same arithmetic, in your tab. The grid above is decided again in your browser by the same modules that produced the record: instruments/window/q.js sha256 0cba8daebf19ad466f6737291d4ab96437257b39ca166e042d8572f86b4f5924 · instruments/window/decide.js sha256 42b1deeb865d9599c18d77b26e7a9bf7c457ec4d59608043f4cf4a3c2e8fac2a. With scripts off, the grid shows the published record.

The battery. 17 checks and 8 red controls at every build — inclusive and strict limits at the exact edge, the witness of a BLOCKED, the threshold of an UNDECIDED, and refusals for a swapped unit, a floating-point number, an unknown variable and an empty window.

The honest boundaries

  • The node, not the berth. The forecast is ECMWF's, read at the nearest open-sea node. At a terminal, it is the approach: a wave limit at the berth stays NEEDS DATA until the transfer into the bay exists.
  • The band is a claim. Today, ECMWF's ensemble (the central 40 of 50), NOAA's (WAVEWATCH III, the central 25 of 31) and Janela's measured bands over each provider; all scored in public against the same satellites, and the proposer that misses its own claim is pruned by the exact binomial rule. A raw ensemble is shown and scored, never decided. Where both measured bands exist, their union decides, checked as itself. Over 15 months no band saw, the union's CLEARED windows broke 0.33% of the time (Hs ≤ 2 m), against 1.10% for ECMWF's band alone and 0.72% for it widened until it covered 96.2% (certs/janela-providers-eval.json).
  • The forecast's steps. What is decided is the band at the forecast's own steps: not the sea between steps, not a probability.
  • The tail of the α. The 1-in-10,000 tail behind DNV's α is a model extrapolation; three years of satellite pairs do not observe it.
  • Not an approval. Marine warranty and the classification societies own that word. Janela delivers evidence a surveyor can redo.
the sources, the licences and the pins
sourcelicenceused forpin
ECMWF open forecast, 00 UTC: IFS 0.25°, waves and the wave ensemble (50 members)CC BY 4.0the day's forecastcorpus/janela/feed/20261031.json.gz sha256 2542857d7dc4e4dd… of the JSON; 87 GRIB files, each with its sha256
NOAA WAVEWATCH III, 00 UTC: GEFS-Wave (31 members) and GFS-Wave, since 2026-10-07public domainNOAA's band (in the union) and the ensemble (shown)corpus/janela/feed-noaa/20261031.json.gz sha256 58cf8c4fbb9e84db… of the JSON; 58 GRIB groups, each with its sha256
Copernicus Marine, surface current 1/12°, since 2026-10-07free, with creditthe current shown (not decided)Generated using E.U. Copernicus Marine Service Information, doi:10.48670/moi-00016
NOAA GFS-Wave, 00 UTC, 2023-07 to 2026-10 (historical)public domainNOAA's measured band: the same satellites and rulecorpus/janela/noaa 40 monthly files, sha256 in corpus/janela/noaa/MANIFEST.json; pairs in corpus/janela/matchups-noaa.json.gz
NOAA RADS, near-real-time altimetrypublic domainthe measured error (satellite × forecast pairs) and the scoreboardcorpus/janela/matchups.json.gz sha256 82d8f09af79fcc2c…, 45,987 pairs
Ifremer WAVEWATCH III GLOBMULTI, hindcast 1993–2024CC BY-SA 4.0the windows of “The month”corpus/ww3-points 384 monthly files, sha256 in corpus/ww3-points/meta.json
The Harbour Masters (Capitanias dos Portos): NPCPs and ordinancesofficial actthe published limits3 acts, each PDF with its sha256 (below)
DNV-OS-H101, Oct 2011, Section 4 B700, Tables 4-1 to 4-6DNV; values transcribed with the pagethe α and the OPWFsha256 4eca653e7cc4c528… of the PDF, not redistributed
the published limits, as printed, in Portuguese (8)

São Francisco do Sul bar open to pilotage

Normas e Procedimentos da Capitania dos Portos de Santa Catarina (NPCP-SC/2022, 1ª Revisão, consolidada até a Mod. 21). Portaria CPSC/Com5°DN/ComOpNav/MB nº 35, de 24 de junho de 2022 (aprova a 1ª Revisão; vigência 24/07/2022), modificada até a Mod. 21 (Portaria nº 62, de 29/07/25, vigência 31/07/25).

QUADRO DAS CONDIÇÕES AMBIENTAIS DE IMPRATICABILIDADE | Altura de Onda (Hs) // (m) | Intensidade do Vento Sustentado (V) // (nós) | Corrente Máxima (c) // (nós) | Visibilidade (Vis) (MN) || São Francisco do Sul | VERDE | Hs ≤ 1,5 e V ≤ 18 e C ≤ 1 e Vis ≥ 1,5 | AMARELO | 1,5 < Hs < 2,0 e/ou 18< V < 22 e/ou 1 < C < 3,0 e/ou 1,0 < Vis < 1,5 | VERMELHO | Hs ≥ 2,0 ou V ≥ 22,0 ou C ≥ 3,0 ou Vis ≤ 1,0 … Na ocorrência de parâmetros ambientais na faixa vermelha do Quadro de Parâmetros Ambientais de Impraticabilidade, o Capitão dos Portos ou Delegado declarará impraticabilidade total da barra. Os parâmetros descritos neste item deverão ser verificados anualmente, no mês de maio, sendo atualizados de acordo com a necessidade, caso exista.printed '– 4 – 22 de 37 –' (item 4.18 b); pdf p. 80 of sc_mod21.pdf; same table in Portaria nº 62/2025 p. 2 de 2

How Janela reads it: Hs < 2.0 m · wind < 22 kn · current < 3.0 kn · visibility > 1.0 NM.

in force, as transcribed: in force as of the act (Mod. 21, Portaria nº 62/2025, vigência 31/07/2025); text says it is to be reviewed every May

transcription note: Row is labelled only 'São Francisco do Sul'; the Mod.19 row was 'São Francisco do Sul e Itapoá' — Mod.21 does not say whether Itapoá is still covered by this row. Mod.21 also deleted the old 'Observação' that pointed to Cap. 5 seção I and Anexo 5-B. Hs column has no wave-height location qualifier (Mod.19 said 'Canal de Acesso Externo').

the act · sha256 40e478360f083b13805220223e693402c722d3c3d8642d67e3de8b4b0cd689dd · 2026-10-06

Terminal Gás Sul — LNG carrier manoeuvre (experimental parameters)

Normas e Procedimentos da Capitania dos Portos de Santa Catarina (NPCP-SC/2022, 1ª Revisão, consolidada até a Mod. 21). Portaria CPSC/Com5°DN/ComOpNav/MB nº 35, de 24 de junho de 2022 (aprova a 1ª Revisão; vigência 24/07/2022), modificada até a Mod. 21 (Portaria nº 62, de 29/07/25, vigência 31/07/25).

MANOBRA EXPERIMENTAL [...] II – MANOBRAS EXPERIMENTAIS Alterar os parâmetros operacionais, excepcionalmente, para a realização de 02 (duas) manobras experimentais de entrada e 02 (duas) manobras experimentais de saída, de navios-tipo FSRU e 06 (seis) manobras experimentais de entrada e 06 (seis) manobras experimentais de saída, de navios-tipo LNGC, no Terminal Gás Sul, em período diurno. [...] a) Manobras diurnas dos navios-tipo (FSRU) com as seguintes dimensões: A) LOA 297,57 metros; e B) Boca de 43,4 metros. b) Manobras diurnas dos navios-tipo (LNGC) com as seguintes dimensões: A) LOA 315 metros; e B) Boca de 50 metros. c) Manobra em horário próximo ao estofo de maré; d) 4 rebocadores com 70 BP; [...] f) 2 Práticos; g) FAQ de 2,5m para a FSRU e de 3,2m para o LNGC; h) Vento máximo de 16 nós, corrente de 1,2 nó e altura das ondas de 1,5m; i) Visibilidade mínima de 2 MN; [...] III – AVALIAÇÃO DAS MANOBRAS EXPERIMENTAIS Após o período experimental, a Autoridade Portuária deverá apresentar, juntamente com o Parecer Técnico da Praticagem São Francisco, um relatório [...] para subsidiar o lançamento definitivo dos parâmetros operacionais ou a redução do navio-tipo a ser empregado na área de navegação. IV – PERÍODO EXPERIMENTAL As manobras experimentais devem ocorrer até 30 de dezembro de 2024.printed '– 5 – C – 27 –' (Anexo 5-C, DelSFSul); pdf p. 243 (identical in Mod.19 pdf p. 243); same parameters in Portaria nº 55/2024 pp. 1–2 (item i/j)

How Janela reads it: wind ≤ 16 kn · current ≤ 1.2 kn · Hs ≤ 1.5 m · visibility ≥ 2 NM · length ≤ 315 m. Wave limited at the berth: not decided at the node.

in force, as transcribed: experimental (period ended 2024-12-30 per NPCP text; Portaria nº 55/2024 says 2024-12-31); still printed unchanged in the current Mod.21 consolidation (Aug 2025); no later Mod (20, 21) or other text makes them final

transcription note: (1) 'máximo' grammatically qualifies only the wind; current 1,2 nó and wave 1,5 m read as maxima. Wind not stated as sustained/gust; wave not stated as Hs. (2) End date differs: NPCP Anexo 5-C-27 '30 de dezembro de 2024' vs Portaria nº 55/2024 'V - PERÍODO EXPERIMENTAL As manobras experimentais devem ocorrer até 31 de dezembro de 2024.' (3) The portaria's Art. 2 says '04 (quatro) manobras experimentais de entrada/saída ... FSRU' (= 2+2). (4) No final ('definitivo') parameters were found in the CPSC list of portarias (Mods 1–21) or in Mod.21; the TGS has no Anexo 5-B row. These are the only published TGS metocean numbers.

the act · sha256 40e478360f083b13805220223e693402c722d3c3d8642d67e3de8b4b0cd689dd · 2026-10-06

TEBAR — ship-to-ship manoeuvre at berth 1 (PP1)

Normas e Procedimentos da Capitania dos Portos de São Paulo (NPCP/SP, 3ª revisão) — pacote SUPP. Portaria CPSP/ComOpNav/MB nº 82, de 21 de janeiro de 2026 (vigência 21/01/2026; revoga a Portaria nº 78, de 15/01/2024).

5.7.3 Limites ambientais/operacionais estabelecidos para operação ship to ship (sts) atracado ao berço nº1 do tebar: BERÇO | LIMITES AMBIENTAIS PARA MANOBRAR | LIMITES AMBIENTAIS PARA MANTEREM-SE NOS BERÇOS || PP1 – SUEZMAX/SUEZMAX | CORRENTE < 2,0 nós VENTOS < 20 nós | CORRENTE < 2,1 nós VENTOS < 30 nós || PP1 – VLCC/SUEZMAX | CORRENTE < 2,0 nós VENTOS < 20 nós | CORRENTE < 2,1 nós VENTOS < 30 nósprinted '- 5-12 -' (item 5.7.3); zip member 'SEQ 030 - 27-01-2026 11:58:05 - OUTROS - K_CAPITULO 5.PDF', pdf p. 12

How Janela reads it: current < 2.0 kn · wind < 20 kn. Wave limited at the berth: not decided at the node.

in force, as transcribed: in force as of the act (Portaria CPSP nº 82/2026, vigência 21/01/2026)

transcription note: 'VENTOS' not stated as sustained or gust. 5.9 a) says 2-pilot STS parameters at PP1 'estão regulamentadas por Portaria específica da DelSSebastião' — that portaria is not on the DelSSebastião 'Portarias' page (checked 2026-10-06) and was not found.

the act · sha256 377b4318c84483ac8836a411b707ca636d5c67c576c1c47352a29c67b868a12e · 2026-10-06

São Sebastião port complex out of red

Normas e Procedimentos da Capitania dos Portos de São Paulo (NPCP/SP, 3ª revisão) — pacote SUPP. Portaria CPSP/ComOpNav/MB nº 82, de 21 de janeiro de 2026 (vigência 21/01/2026; revoga a Portaria nº 78, de 15/01/2024).

A condição de praticabilidade do porto seguirá a seguinte tabela: a) VERDE – PRATICABILIDADE IRRESTRITA [...] b) AMARELA – PRATICABILIDADE RESTRITA [...] c) VERMELHA – IMPRATICABILIDADE TOTAL [...] ÁREA GEOGRÁFICA | CONDIÇÕES DE PRATICABILIDADE | ALTURA DE ONDA (m) | INTENSIDADE DO VENTO (nós) | VISIBILIDADE (MIN) | MÉTODO DE PRATICAGEM || COMPLEXO PORTUÁRIO DE SÃO SEBASTIÃO | Verde | < 1,5 | < 17 | > 1 | Direta | Amarelo | ≥ 1,5 | > 17 ou Rajadas > 35 | < 1 | Indireta | Vermelho | ≥3 | > 30 ou Rajadas > 35 | < 0,5 | Impraticabilidadeprinted '- 4-23 -' (item 4.3.7 table; band definitions a)–c) on '- 4-22 -'); zip member 'SEQ 029 - 27-01-2026 11:58:03 - OUTROS - J_CAPITULO 4.PDF', pdf p. 23

How Janela reads it: Hs < 3 m · wind ≤ 30 kn · gust ≤ 35 kn · visibility ≥ 0.5 NM.

in force, as transcribed: in force as of the act (Portaria CPSP nº 82/2026, vigência 21/01/2026)

transcription note: (1) Visibility unit printed as '(MIN)' — read as MN (milhas náuticas); could be a typo for 'MN'. (2) Wave column says 'ALTURA DE ONDA', not Hs. (3) Boundary gaps: wind exactly 17 nós and visibility exactly 1 fall in no band. (4) 'Rajadas > 35' appears in both Amarelo and Vermelho — a gust >35 alone is ambiguous between the two. (5) Wind column does not say sustained; '> 17' read as sustained because gusts are named separately.

the act · sha256 377b4318c84483ac8836a411b707ca636d5c67c576c1c47352a29c67b868a12e · 2026-10-06

TEBIG — East Channel transit (Aframax, Suezmax, VLCC)

Normas e Procedimentos da Capitania dos Portos do Rio de Janeiro (NPCP-RJ, 3ª Revisão, Mod. 3). Portaria CPRJ/COMOPNAV/MB nº 168, de 04 de dezembro de 2025 (3ª Revisão; vigência 04/12/2025), com Mod. 1 (Portaria nº 11/2026), Mod. 2 (Portaria nº 110/2026) e Mod. 3 (Portaria nº 200/2026, de 15 de julho de 2026).

Estão autorizadas as manobras no trânsito pelo Canal Leste de Acesso ao TEBIG, para navios-tipo Aframax, Suezmax e VLCC, desde que sejam observadas as seguintes condicionantes: - calado máximo: 13,2 m (≤ 13,2 m); [...] - luminosidade: períodos diurno e noturno; - vento: não superior a 20 nós (≤ 20 nós); - altura significativa de onda: até 2,0 m (≤ 2,0 m); - velocidade limite de trânsito: 12 nós (≤ 12 nós); e - maré e corrente: sem restrições. As manobras de entrada e saída deverão ser realizadas apenas à luz do dia, podendo ser agendadas (POB) desde o nascer do sol, na meia hora, ou hora cheia posterior, até 2h antes do pôr do sol, na meia hora ou hora cheia anterior.printed '- 5-70 -'/'- 5-71 -' (5.1.2 b) I); pdf pp. 189–190; inserted by Portaria nº 200/2026 Art. 2º (rj-port200-2026, p. 1)

How Janela reads it: Hs ≤ 2.0 m · wind ≤ 20 kn · draft ≤ 13.2 m · speed ≤ 12 kn.

in force, as transcribed: in force as of the act (Mod. 3 — Portaria nº 200/2026, de 15 de julho de 2026, em vigor na data)

transcription note: Direct contradiction as printed: the Port. 200/2026 block allows 'períodos diurno e noturno', but the very next paragraph (pre-existing text, different indentation) says 'As manobras de entrada e saída deverão ser realizadas apenas à luz do dia ... até 2h antes do pôr do sol'; the general TEBIG paragraph above says 'não há restrições de horário'. Which governs night transits of the Canal Leste is not resolvable from the text.

the act · sha256 0d1706a26ea9ebc7489952c9872eee6ff99cb5bc306c5cddf7fe9c20ee425ecb · 2026-10-06

Ilha Grande Bay out of red

Normas e Procedimentos da Capitania dos Portos do Rio de Janeiro (NPCP-RJ, 3ª Revisão, Mod. 3). Portaria CPRJ/COMOPNAV/MB nº 168, de 04 de dezembro de 2025 (3ª Revisão; vigência 04/12/2025), com Mod. 1 (Portaria nº 11/2026), Mod. 2 (Portaria nº 110/2026) e Mod. 3 (Portaria nº 200/2026, de 15 de julho de 2026).

4.3.7. Impraticabilidade [...] conforme o quadro de condições ambientais abaixo (Quadro 1). [...] Área Geográfica | Faixa | Altura de Onda (Hs) (m) | Intensidade do Vento (V) (nós) | Visibilidade (Vis) (MN) || Baía da Ilha Grande (DelAReis) | Verde | Hs ≤ 1,5 e V ≤ 15 e Vis ≥ 1,0 | Amarela | 1,5 < Hs < 2,5 e/ou 15 < V < 30 e/ou 0,5 < Vis < 1,0 | Vermelha | Hs ≥ 2,5 ou V ≥ 30 ou Vis ≤ 0,5 — Quadro 1 – Parâmetros Ambientais de Impraticabilidade [...] Nas condições da faixa amarela, considerando a medição dos sensores específicos, a avaliação no local pela equipe de inspeção naval da CP/DL e a assessoria dos Práticos em escala do dia, o Capitão dos Portos ou o Delegado poderá declarar impraticabilidade total ou parcial [...] Na ocorrência de parâmetros ambientais na faixa vermelha do quadro de parâmetros ambientais de impraticabilidade, o Capitão dos Portos ou Delegado declarará impraticabilidade total da barra. || Baía da Ilha Grande (DelAReis) | Não há | Não há | Não há — Quadro 2 – Sensores de Parâmetros Ambientaisprinted '- 4-25 -' (Quadro 1) and '- 4-26 -' (faixa text, Quadro 2); pdf pp. 95–96

How Janela reads it: Hs < 2.5 m · wind < 30 kn · visibility > 0.5 NM.

in force, as transcribed: in force as of the act (NPCP-RJ 3ª Revisão, Portaria nº 168/2025, consolidada até a Mod. 3 de 15/07/2026)

transcription note: Quadro 2 lists NO reference sensor for any of the three parameters in the Baía da Ilha Grande ('Não há'), so the band cannot be read from a designated station there. Wind not stated as sustained/gust.

the act · sha256 0d1706a26ea9ebc7489952c9872eee6ff99cb5bc306c5cddf7fe9c20ee425ecb · 2026-10-06

Sepetiba — LNG carrier entry for ship-to-ship with the FSRU

Normas e Procedimentos da Capitania dos Portos do Rio de Janeiro (NPCP-RJ, 3ª Revisão, Mod. 3). Portaria CPRJ/COMOPNAV/MB nº 168, de 04 de dezembro de 2025 (3ª Revisão; vigência 04/12/2025), com Mod. 1 (Portaria nº 11/2026), Mod. 2 (Portaria nº 110/2026) e Mod. 3 (Portaria nº 200/2026, de 15 de julho de 2026).

- Para a realização de operações ship-to-ship, as manobras somente no período diurno, com ventos de até 16 nós para a entrada e até 12 nós para a saída, visibilidade não inferior a 2 milhas náuticas, e corrente de vazante, com velocidade ≤ 0,5 nó (na bifurcação de entrada do canal da Ternium); - Entrada: a partir do nascer até 5 horas antes do pôr do sol, na meia hora ou hora cheia anterior; e - Saída: a partir do nascer até 3 horas antes do pôr do sol, na meia hora ou hora cheia anterior.printed '- 5-103 -'/'- 5-104 -' (5.1.3 d) V – 'Manobra com LNGC – Ship-to-Ship'); pdf pp. 222–223

How Janela reads it: wind ≤ 16 kn · current ≤ 0.5 kn · visibility ≥ 2 NM. Wave limited at the berth: not decided at the node.

in force, as transcribed: in force as of the act (NPCP-RJ 3ª Revisão, Portaria nº 168/2025, consolidada até a Mod. 3 de 15/07/2026)

transcription note: (1) The current clause reads 'corrente de vazante, com velocidade ≤ 0,5 nó' — it is unclear whether flood (enchente) current is excluded altogether or simply not limited. (2) On p. 5-103 the abort rule refers to the current limit 'preconizado acima', but the limit is printed below (p. 5-104). (3) This block sits, oddly, under 5.1.3 d) 'Porto Sudeste do Brasil', item V). Wind not stated as sustained/gust.

the act · sha256 0d1706a26ea9ebc7489952c9872eee6ff99cb5bc306c5cddf7fe9c20ee425ecb · 2026-10-06

Sepetiba Bay out of red

Normas e Procedimentos da Capitania dos Portos do Rio de Janeiro (NPCP-RJ, 3ª Revisão, Mod. 3). Portaria CPRJ/COMOPNAV/MB nº 168, de 04 de dezembro de 2025 (3ª Revisão; vigência 04/12/2025), com Mod. 1 (Portaria nº 11/2026), Mod. 2 (Portaria nº 110/2026) e Mod. 3 (Portaria nº 200/2026, de 15 de julho de 2026).

4.3.7. Impraticabilidade [...] conforme o quadro de condições ambientais abaixo (Quadro 1). [...] Área Geográfica | Faixa | Altura de Onda (Hs) (m) | Intensidade do Vento (V) (nós) | Visibilidade (Vis) (MN) || Baía de Sepetiba (DelItacuruçá) | Verde | Hs ≤ 1,5 e V ≤ 15 e Vis ≥ 1,0 | Amarela | 1,5 < Hs < 2,5 e/ou 15 < V < 30 e/ou 0,5 < Vis < 1,0 | Vermelha | Hs ≥ 2,5 ou V ≥ 30 ou Vis ≤ 0,5 — Quadro 1 – Parâmetros Ambientais de Impraticabilidade [...] Nas condições da faixa amarela, considerando a medição dos sensores específicos, a avaliação no local pela equipe de inspeção naval da CP/DL e a assessoria dos Práticos em escala do dia, o Capitão dos Portos ou o Delegado poderá declarar impraticabilidade total ou parcial [...] Na ocorrência de parâmetros ambientais na faixa vermelha do quadro de parâmetros ambientais de impraticabilidade, o Capitão dos Portos ou Delegado declarará impraticabilidade total da barra. || Baía de Sepetiba (DelItacuruçá) | Estação meteoceanográfica TIG Boia 01 | Estação meteoceanográfica TIG Boia 01 | Não há — Quadro 2 – Sensores de Parâmetros Ambientaisprinted '- 4-25 -' (Quadro 1) and '- 4-26 -' (Quadro 2); pdf pp. 95–96

How Janela reads it: Hs < 2.5 m · wind < 30 kn · visibility > 0.5 NM.

in force, as transcribed: in force as of the act (NPCP-RJ 3ª Revisão, Portaria nº 168/2025, consolidada até a Mod. 3 de 15/07/2026)

transcription note: Reference sensor for Hs and wind is 'Estação meteoceanográfica TIG Boia 01'; for visibility 'Não há'. Wind not stated as sustained/gust.

the act · sha256 0d1706a26ea9ebc7489952c9872eee6ff99cb5bc306c5cddf7fe9c20ee425ecb · 2026-10-06

DNV's Table 4-1, transcribed
alpha (waves)design Hs 1 mdesign Hs 2 mdesign Hs 4 mdesign Hs ≥ 6 m
TPOP ≤ 12 h0.650.760.790.80
TPOP ≤ 24 h0.630.730.760.78
TPOP ≤ 36 h0.620.710.730.76
TPOP ≤ 48 h0.600.680.710.74
TPOP ≤ 72 h0.550.630.680.72

DNV-OS-H101 Marine Operations, General (Offshore Standard), October 2011. Linear interpolation between columns; the window TR is twice the TPOP when the contingency is not assessed in detail (B402).

Redo it. The code is open: apps/janela and instruments/window.

node instruments/window/battery.js
node apps/janela/battery.js
node apps/janela/build.js
node apps/janela/build-today.js
why these numbers stand as they do
  • computed — a floating-point evaluation (certs/janela-providers-eval.json)
  • computed — a floating-point evaluation