GGooseAtlasThe archive

Piarco · 1941 to 2026 · 416,149 hourly observations

Trinidad runs on air.

1,663 of the 2,074 MW on this grid are gas turbines, and a gas turbine does not make electricity from fuel alone: it makes it from the mass of air it can pull through a compressor. Hot air weighs less. So the fleet is smaller on a hot afternoon than it is on the nameplate, and the island's own weather station has been writing down exactly how much, every hour, since 1941.

Breathing ambient air
1,663 MW
80.2% of the installed fleet
Given back every afternoon
75 MW
05:00 against 13:00, on an average day
Off the plate at the record peak
215 MW
15 May 2024, 33.4°C at 14:00
Hours above 32°C a year
58 → 496
1942-1975 against 2016-2025
Piarco, 23 Oct, 02:0025°C89% RHwet bulb 23.6°C101.29 kPaturbine fleet worth about 1,596 MW right now

Which machines are exposed, and which are not

Steam turbines answer to cooling water, not to intake air, and reciprocating engines barely notice either. Gas turbines are the ones that track the weather hour by hour, and in Trinidad and Tobago they are almost the whole fleet. The station list is the Ministry of Energy's own.

StationInstalledAir exposed
Point Lisas
4 x 85 MW Westinghouse, 4 x 62.5 MW ABB, 2 x 104 MW Siemens Westinghouse, 1 x 20 MW GE for black start
PowerGen
818818
Penal
2 x 65 MW gas turbines and a 66 MW steam turbine on two HRSGs, plus 2 x 20 MW GE gas turbines
PowerGen
236170
Union Estate, La Brea
6 x 75 MW GE gas turbines and 2 x 135 MW steam turbines
Trinidad Generation Unlimited
720450
Trinity Power, Point Lisas
3 x 75 MW GE gas turbines, simple cycle
Trinity Power Ltd
225225
Cove, Tobago
4 x 16 MW Wartsila dual-fuel, commissioned 3 October 2009
T&TEC
64—
Scarborough, Tobago
2 x 5.5 MW Wartsila, standby
T&TEC
11—
Total2,0741,663

Ministry of Energy and Energy Industries, Electric Power. Tobago's Wärtsilä sets are counted as installed capacity and left out of the air-exposed column.

Five records in a fortnight, and the air on each of those days

In May 2024 T&TEC passed its all-time system peak five times in two weeks and said plainly why: hot spells. Those peak figures are the utility's own. Everything in the three right-hand columns is Piarco's hourly record for the same day, and the capability figure is what the published derate makes of that temperature.

DaySystem peakPiarco maxAtFeels likeTurbine fleet worthHottest day rank
13 September 2023
the record before May 2024
1,410 MW33.6°C14:0043.2°C1,447 MWof 1,663723 / 16,055
2 May 2024
first of five records in a fortnight
1,422 MW34°C13:0040.6°C1,442 MWof 1,663665 / 16,055
6 May 2024
1,447.1 MW32.8°C14:0040.1°C1,455 MWof 1,6632,522 / 16,055
7 May 2024
1,448 MW32°C13:0035.9°C1,465 MWof 1,6635,501 / 16,055
13 May 2024
1,455.8 MW33.6°C14:0039.2°C1,447 MWof 1,663724 / 16,055
15 May 2024
highest in the reported run
1,456.2 MW33.4°C14:0038.4°C1,448 MWof 1,663812 / 16,055
Reserve on paper
617.8 MW

2,074 MW installed less the 1456.2 MW peak of 15 May 2024. T&TEC put its own installed capacity at 2,037 MW that month.

Reserve in that afternoon's air
402.8 MW

The same sum with the turbine fleet taken at 33.4°C instead of at its rating point.

Every one of the six
13:00–14:00

The hour Piarco reached its maximum on all six record days. Not one of them peaked in the evening.

The squeeze happens in one hour, and it is the same hour

Two measures, two panels, one x axis: what the turbine fleet is worth through the day, and what the day feels like to the people switching on air conditioners. The heat index peaks at 12:00 and the fleet bottoms out at 13:00, an hour apart, which for planning purposes is the same moment. Both curves are 416,149 hours of Piarco averaged onto the clock.

What the 1,663 MW of gas turbines is actually worth, by hourMW, local time
1,5001,55000:00 · 24.47°C · 1553 MW of 1663 (93.4% of rating)01:00 · 24.27°C · 1555 MW of 1663 (93.5% of rating)02:00 · 24.02°C · 1558 MW of 1663 (93.7% of rating)03:00 · 23.92°C · 1559 MW of 1663 (93.8% of rating)04:00 · 23.77°C · 1561 MW of 1663 (93.9% of rating)05:00 · 23.62°C · 1563 MW of 1663 (94.0% of rating)06:00 · 23.61°C · 1563 MW of 1663 (94.0% of rating)07:00 · 24.42°C · 1553 MW of 1663 (93.4% of rating)08:00 · 26.39°C · 1530 MW of 1663 (92.0% of rating)09:00 · 28.08°C · 1511 MW of 1663 (90.8% of rating)10:00 · 29.01°C · 1500 MW of 1663 (90.2% of rating)11:00 · 29.53°C · 1494 MW of 1663 (89.8% of rating)12:00 · 29.86°C · 1490 MW of 1663 (89.6% of rating)13:00 · 29.94°C · 1489 MW of 1663 (89.5% of rating)14:00 · 29.78°C · 1491 MW of 1663 (89.7% of rating)15:00 · 29.51°C · 1494 MW of 1663 (89.8% of rating)16:00 · 29.01°C · 1500 MW of 1663 (90.2% of rating)17:00 · 28.21°C · 1509 MW of 1663 (90.8% of rating)18:00 · 27.11°C · 1522 MW of 1663 (91.5% of rating)19:00 · 26.18°C · 1533 MW of 1663 (92.2% of rating)20:00 · 25.52°C · 1541 MW of 1663 (92.6% of rating)21:00 · 25.23°C · 1544 MW of 1663 (92.8% of rating)22:00 · 24.95°C · 1547 MW of 1663 (93.0% of rating)23:00 · 24.66°C · 1551 MW of 1663 (93.2% of rating)1,563 MW1,489 MW036912151821
Heat index, the cooling load's own clock°C, local time
25303500:00 · 24.47°C and 90.6% RH feels like 24.6°C01:00 · 24.27°C and 91.3% RH feels like 24.36°C02:00 · 24.02°C and 92% RH feels like 24.06°C03:00 · 23.92°C and 92.4% RH feels like 23.95°C04:00 · 23.77°C and 92.8% RH feels like 23.79°C05:00 · 23.62°C and 93.1% RH feels like 23.63°C06:00 · 23.61°C and 93.1% RH feels like 23.62°C07:00 · 24.42°C and 90.7% RH feels like 24.71°C08:00 · 26.39°C and 83.3% RH feels like 28.26°C09:00 · 28.08°C and 76% RH feels like 31.37°C10:00 · 29.01°C and 70.9% RH feels like 32.57°C11:00 · 29.53°C and 68% RH feels like 33.09°C12:00 · 29.86°C and 66.5% RH feels like 33.46°C13:00 · 29.94°C and 66% RH feels like 33.46°C14:00 · 29.78°C and 66.1% RH feels like 33.12°C15:00 · 29.51°C and 67.2% RH feels like 32.81°C16:00 · 29.01°C and 69.2% RH feels like 32.13°C17:00 · 28.21°C and 72.6% RH feels like 31°C18:00 · 27.11°C and 78.1% RH feels like 29.3°C19:00 · 26.18°C and 83% RH feels like 27.55°C20:00 · 25.52°C and 86.1% RH feels like 26.13°C21:00 · 25.23°C and 87.4% RH feels like 25.72°C22:00 · 24.95°C and 88.6% RH feels like 25.27°C23:00 · 24.66°C and 89.6% RH feels like 24.81°C33.5°C23.6°C036912151821

The gap between 05:00 and 13:00 is 4.5% of rated output, or 75 MW, on an ordinary day. It is free capacity that exists overnight and is gone by lunchtime, which is an argument for moving load and for charging storage on the night side of that curve rather than for building more plant.

The two curves as numbers
HourTemp °CRH %Wet bulb °CFeels like °CFleet MW% of ratingWind km/h
00:0024.4790.623.2424.6155393.44.1
01:0024.2791.323.1324.36155593.53.8
02:0024.029222.9824.06155893.73.6
03:0023.9292.422.9223.95155993.83.3
04:0023.7792.822.8323.79156193.93.3
05:0023.6293.122.7323.63156394.03.3
06:0023.6193.122.7123.62156394.03.2
07:0024.4290.723.1924.71155393.43.9
08:0026.3983.324.0928.26153092.08.3
09:0028.087624.6431.37151190.814.6
10:0029.0170.924.7332.57150090.217.3
11:0029.536824.7433.09149489.818.5
12:0029.8666.524.7933.46149089.619
13:0029.946624.7633.46148989.519.2
14:0029.7866.124.6533.12149189.719
15:0029.5167.224.5932.81149489.818.6
16:0029.0169.224.4432.13150090.217.6
17:0028.2172.624.2431150990.815.9
18:0027.1178.124.0429.3152291.512.4
19:0026.188323.8727.55153392.28.8
20:0025.5286.123.6526.13154192.66.8
21:0025.2387.423.5525.72154492.85.8
22:0024.9588.623.4425.27154793.05.1
23:0024.6689.623.324.81155193.24.6

And it is worst in the months you would least like it to be

Trinidad's hottest afternoons are April and May, at the end of the dry season, which is when both of the last two all-time system peaks were set. April averages 30.1°C in the 11:00 to 17:00 window and leaves the turbine fleet worth about 1,487 MW.

Afternoon capability by monthMW available from 1,663 MW of turbines, 11:00–17:00 mean
1,4901,500January · 28.6°C and 65.6% RH in the afternoon · 1504 MWFebruary · 29.0°C and 61.9% RH in the afternoon · 1500 MWMarch · 29.6°C and 59.7% RH in the afternoon · 1494 MWApril · 30.1°C and 60.4% RH in the afternoon · 1487 MWMay · 30.1°C and 65.3% RH in the afternoon · 1487 MWJune · 29.1°C and 72.1% RH in the afternoon · 1498 MWJuly · 29.4°C and 71.5% RH in the afternoon · 1495 MWAugust · 29.6°C and 71.3% RH in the afternoon · 1493 MWSeptember · 30.0°C and 70.2% RH in the afternoon · 1489 MWOctober · 29.6°C and 71.9% RH in the afternoon · 1493 MWNovember · 29.1°C and 73% RH in the afternoon · 1499 MWDecember · 28.6°C and 70.5% RH in the afternoon · 1504 MW1,504 MW1,487 MWJanFebMarAprMayJunJulAugSepOctNovDec

The whole distribution, not just the average

Every hour in the record, sorted. The median hour leaves the fleet at 1,535 MW. The hottest one hour in a hundred leaves it 81 MW below that.

1th percentile
1,453 MW
5th percentile
1,469 MW
25th percentile
1,500 MW
median hour
1,535 MW
75th percentile
1,557 MW
95th percentile
1,577 MW
99th percentile
1,593 MW

In the afternoon window alone the median is 1,491 MW and the hot tail is 1,455 MW.

Read the absolute numbers with this in mind

If a machine's published MW was set at site conditions rather than ISO, the absolute figures here overstate the loss. The hour-to-hour and season-to-season swings do not depend on the rating basis, and they are the numbers to act on.

Every MW figure on this page uses the derate manufacturers publish: 0.7 % of rated output per degree of intake air above the ISO rating point of 15 C, the middle of the 0.5 to 0.9 % range for industrial gas turbines.

Beside it is moist-air density from each hour's temperature, station pressure and relative humidity, as a ratio to density at ISO intake conditions. A gas turbine breathes by volume, so density sets the mass of air it can draw. Density explains about half the derate: the rest is compressor work, which rises with intake temperature. The density figure is measured physics from the station record and moves about 0.33 % per degree.

T&TEC publishes no demand series. The peak figures on this page are its own published records, quoted as such. Nothing here models demand.

Why fogging the inlet will not fix it here

The standard answer to a hot intake is evaporative cooling, and it is cheap. It also cannot cool air below its wet-bulb temperature, and Trinidad's afternoons are humid: the gap between dry bulb and wet bulb averages 5.2°C at 13:00, from 4.0°C in November to 6.1°C in March. That gap is the whole budget an evaporative cooler has to spend.

Wet-bulb depression, afternoons°C of cooling an evaporative stage could ever reach
0246January · dry bulb 28.65°C, wet bulb 23.6°C, depression 5.08°CFebruary · dry bulb 28.98°C, wet bulb 23.3°C, depression 5.7°CMarch · dry bulb 29.55°C, wet bulb 23.4°C, depression 6.11°C6.1°CApril · dry bulb 30.14°C, wet bulb 24.1°C, depression 6.06°CMay · dry bulb 30.09°C, wet bulb 24.8°C, depression 5.28°CJune · dry bulb 29.15°C, wet bulb 25.0°C, depression 4.14°CJuly · dry bulb 29.4°C, wet bulb 25.1°C, depression 4.25°CAugust · dry bulb 29.62°C, wet bulb 25.3°C, depression 4.3°CSeptember · dry bulb 29.98°C, wet bulb 25.5°C, depression 4.5°COctober · dry bulb 29.62°C, wet bulb 25.4°C, depression 4.21°CNovember · dry bulb 29.09°C, wet bulb 25.1°C, depression 3.99°CDecember · dry bulb 28.63°C, wet bulb 24.3°C, depression 4.33°CJFMAMJJASOND

What each option recovers, across the whole turbine fleet

Evaporative or fogging, 85% effective48 MW

2.86% of rating on an average afternoon, 78 MW on the best of them. Capped by the humidity, not by the equipment.

Chilled coil to 10°C, gross226 MW

13.59% of rating. Not capped by the wet bulb, because it is moving heat rather than evaporating water.

Chilled coil, net of auxiliaries169 MW

Gross figures. A quarter is taken off for chiller auxiliaries and condensate handling to give the net, which is an assumption, not a measurement: the real number comes from a vendor study at one site.

The ranking is the point, not the decimals: in this climate the cheap option buys roughly a fifth of what the expensive one does, and the expensive one competes with new plant at a fraction of the capital. A vendor study at one station would turn these into real numbers.

The same megawatt-hour costs more gas when the air is hot

Output is only half of it. Heat rate moves the other way, by roughly 0.15% per degree of intake air, so an afternoon at 29.4°C burns about 2.16% more fuel for the same electricity than the rating point does. In a country whose gas has somewhere else to go, that is the number with a price attached.

Scaled up: 1,000 MW held for 7 afternoon hours every day of the year carries about 562.9 mmscf of gas a year in ambient penalty alone.

Simple-cycle heat rate 10,500 Btu/kWh (HHV) and 1030 Btu/scf pipeline gas. Heat rate rises about 0.15 % per degree of intake air above the rating point. Substitute your own heat rates and the arithmetic is the same.

What the weather record is actually good for here

  • Dispatch order. Two machines with the same quoted heat rate do not have the same heat rate at 31°C. The one with the newer compressor and the cleaner filter house wins by more in the afternoon than it does at night.
  • Maintenance timing. Compressor washes and filter changes are worth most just before the April to May window, and worth least in December.
  • Capacity planning. A reserve margin computed on nameplate is the wrong margin. The right one is computed at the temperature of the hour you are planning for, and that hour has a known distribution.
  • Contract renewal. With a third of the fleet's contracts reaching their end this decade, an ambient-corrected capability curve per station is a negotiating position rather than an opinion.

52 years of Piarco, one stripe a year

Each stripe is a year's mean temperature from this station's own hourly observations, cool to warm about the 1941-to-2026 mean. The dark stripes are the years the record does not hold: 1950 to 1971 and 1998 to 2003. Nothing here is modelled or interpolated.

1941: no record1942: 25.83°C mean, 365 days of observations1943: 25.39°C mean, 365 days of observations1944: 25.56°C mean, 366 days of observations1945: 25.47°C mean, 365 days of observations1946: 25.72°C mean, 365 days of observations1947: no record1948: 25.5°C mean, 366 days of observations1949: 25.39°C mean, 354 days of observations1950: no record1951: no record1952: no record1953: no record1954: no record1955: no record1956: no record1957: no record1958: no record1959: no record1960: no record1961: no record1962: no record1963: no record1964: no record1965: no record1966: no record1967: no record1968: no record1969: no record1970: no record1971: no record1972: no record1973: 26°C mean, 360 days of observations1974: 25.61°C mean, 357 days of observations1975: 25.57°C mean, 361 days of observations1976: 25.5°C mean, 365 days of observations1977: 26.17°C mean, 365 days of observations1978: 26.16°C mean, 365 days of observations1979: 26.16°C mean, 365 days of observations1980: 26.04°C mean, 365 days of observations1981: 26.28°C mean, 365 days of observations1982: 26.01°C mean, 364 days of observations1983: 26.49°C mean, 365 days of observations1984: 25.82°C mean, 366 days of observations1985: 25.73°C mean, 365 days of observations1986: 25.84°C mean, 365 days of observations1987: 26.82°C mean, 365 days of observations1988: 26.34°C mean, 366 days of observations1989: 26.13°C mean, 360 days of observations1990: 26.25°C mean, 365 days of observations1991: 26.16°C mean, 365 days of observations1992: 26.23°C mean, 366 days of observations1993: 26.56°C mean, 365 days of observations1994: 26.42°C mean, 365 days of observations1995: 26.84°C mean, 365 days of observations1996: 26.54°C mean, 366 days of observations1997: 26.58°C mean, 365 days of observations1998: no record1999: no record2000: no record2001: no record2002: no record2003: no record2004: no record2005: 27.15°C mean, 361 days of observations2006: 26.69°C mean, 365 days of observations2007: 26.64°C mean, 365 days of observations2008: 26.86°C mean, 366 days of observations2009: 27.19°C mean, 365 days of observations2010: 27.55°C mean, 365 days of observations2011: 27.09°C mean, 365 days of observations2012: 27.17°C mean, 366 days of observations2013: 27.25°C mean, 365 days of observations2014: no record2015: 27.25°C mean, 365 days of observations2016: 27.38°C mean, 366 days of observations2017: 26.98°C mean, 365 days of observations2018: 26.59°C mean, 365 days of observations2019: 27.03°C mean, 365 days of observations2020: 26.98°C mean, 365 days of observations2021: 26.68°C mean, 365 days of observations2022: 26.53°C mean, 365 days of observations2023: 27.14°C mean, 365 days of observations2024: 27.67°C mean, 365 days of observations2025: 26.79°C mean, 365 days of observations2026: no record
Mean temperature
25.6°C→27.0°C
annual mean of every hour
Cooling degree hours above 24°C
17,637→26,914
53% more cooling to do
Hours above 32°C
58→496
8.6 times as many, a year
Nights that never drop below 25°C
1→77
the cooling load no longer switches off

First figure is the mean of the first ten complete years in the record (1942-1975), second is the mean of the last ten (2016-2025).

Hours a year above 32°Cthe derate's worst tail
05001,0001942: 52 hours at or above 32°C1943: 4 hours at or above 32°C1944: 8 hours at or above 32°C1945: 1 hours at or above 32°C1946: 8 hours at or above 32°C1948: 9 hours at or above 32°C1949: 70 hours at or above 32°C1973: 257 hours at or above 32°C1974: 67 hours at or above 32°C1975: 103 hours at or above 32°C1976: 140 hours at or above 32°C1977: 199 hours at or above 32°C1978: 319 hours at or above 32°C1979: 443 hours at or above 32°C1980: 292 hours at or above 32°C1981: 373 hours at or above 32°C1982: 191 hours at or above 32°C1983: 398 hours at or above 32°C1984: 91 hours at or above 32°C1985: 116 hours at or above 32°C1986: 110 hours at or above 32°C1987: 582 hours at or above 32°C1988: 484 hours at or above 32°C1989: 346 hours at or above 32°C1990: 323 hours at or above 32°C1991: 235 hours at or above 32°C1992: 301 hours at or above 32°C1993: 529 hours at or above 32°C1994: 309 hours at or above 32°C1995: 757 hours at or above 32°C1996: 377 hours at or above 32°C1997: 442 hours at or above 32°C2005: 917 hours at or above 32°C2006: 518 hours at or above 32°C2007: 516 hours at or above 32°C2008: 760 hours at or above 32°C2009: 777 hours at or above 32°C2010: 1172 hours at or above 32°C2011: 784 hours at or above 32°C2012: 781 hours at or above 32°C2013: 786 hours at or above 32°C2015: 890 hours at or above 32°C2016: 1091 hours at or above 32°C2017: 766 hours at or above 32°C2018: 249 hours at or above 32°C2019: 477 hours at or above 32°C2020: 502 hours at or above 32°C2021: 311 hours at or above 32°C2022: 295 hours at or above 32°C2023: 518 hours at or above 32°C2024: 571 hours at or above 32°C2025: 181 hours at or above 32°C'80'90'10'20
Nights that never fall below 25°Covernight cooling demand
0501001501942: 4 nights with a minimum of 25°C or more1943: 0 nights with a minimum of 25°C or more1944: 1 nights with a minimum of 25°C or more1945: 3 nights with a minimum of 25°C or more1946: 0 nights with a minimum of 25°C or more1948: 0 nights with a minimum of 25°C or more1949: 1 nights with a minimum of 25°C or more1973: 2 nights with a minimum of 25°C or more1974: 1 nights with a minimum of 25°C or more1975: 1 nights with a minimum of 25°C or more1976: 2 nights with a minimum of 25°C or more1977: 4 nights with a minimum of 25°C or more1978: 7 nights with a minimum of 25°C or more1979: 11 nights with a minimum of 25°C or more1980: 10 nights with a minimum of 25°C or more1981: 24 nights with a minimum of 25°C or more1982: 8 nights with a minimum of 25°C or more1983: 36 nights with a minimum of 25°C or more1984: 3 nights with a minimum of 25°C or more1985: 8 nights with a minimum of 25°C or more1986: 7 nights with a minimum of 25°C or more1987: 26 nights with a minimum of 25°C or more1988: 10 nights with a minimum of 25°C or more1989: 3 nights with a minimum of 25°C or more1990: 6 nights with a minimum of 25°C or more1991: 9 nights with a minimum of 25°C or more1992: 12 nights with a minimum of 25°C or more1993: 47 nights with a minimum of 25°C or more1994: 17 nights with a minimum of 25°C or more1995: 41 nights with a minimum of 25°C or more1996: 11 nights with a minimum of 25°C or more1997: 33 nights with a minimum of 25°C or more2005: 57 nights with a minimum of 25°C or more2006: 33 nights with a minimum of 25°C or more2007: 33 nights with a minimum of 25°C or more2008: 59 nights with a minimum of 25°C or more2009: 89 nights with a minimum of 25°C or more2010: 119 nights with a minimum of 25°C or more2011: 74 nights with a minimum of 25°C or more2012: 60 nights with a minimum of 25°C or more2013: 95 nights with a minimum of 25°C or more2015: 98 nights with a minimum of 25°C or more2016: 84 nights with a minimum of 25°C or more2017: 64 nights with a minimum of 25°C or more2018: 30 nights with a minimum of 25°C or more2019: 89 nights with a minimum of 25°C or more2020: 61 nights with a minimum of 25°C or more2021: 56 nights with a minimum of 25°C or more2022: 41 nights with a minimum of 25°C or more2023: 110 nights with a minimum of 25°C or more2024: 156 nights with a minimum of 25°C or more2025: 77 nights with a minimum of 25°C or more'80'90'10'20
Year by year
YearDaysMean °CMean daily max °CCDH >24°CHours ≥32°CNights ≥25°CAfternoon °C
194115726.02————28.83
194236525.8330.291918852428.89
194336525.3929.3157254028.16
194436625.5629.37168968128.31
194536525.4729.32163411328.16
194636525.72—179108028.07
194732025.8529.76———28.5
194836625.529.6168529028.25
194935425.3929.921724370128.38
1972124—————
197336026—20354257229.14
197435725.61—1798167128.85
197536125.57—17876103128.83
197636525.5—17373140228.71
197736526.17—21173199429.55
197836526.16—21325319729.5
197936526.16—217754431129.48
198036526.04—204722921029.29
198136526.28—218163732429.44
198236426.01—19761191829.08
198336526.4930.94233563983629.6
198436625.8230.211861691328.86
198536525.73—18171116828.68
198636525.84—18888110728.94
198736526.82—261015822630
198836626.34—226734841029.65
198936026.13—21544346329.49
199036526.25—21876323629.47
199136526.16—20979235929.27
199236626.2330.83215753011229.43
199336526.5631.27240405294729.72
199436526.4231.05228393091729.48
199536526.8431.75264747574130.15
199636626.54—237253771129.57
199736526.58—240894423329.54
2004124.08—————
200536127.15—285569175729.98
200636526.6931.3248525183329.56
200736526.6431.33245675163329.52
200836626.8631.69265837605929.85
200936527.1931.78286717778930.13
201036527.5532.5131697117211930.55
201136527.0931.94277887847429.92
201236627.1731.93281527816030.05
201336527.2531.78289197869530.07
2014124.88—————
201536527.2532.05291888909830.39
201636627.3832.413020010918430.56
201736526.9831.82270947666429.88
201836526.5930.84239512493029.21
201936527.0331.34273944778929.8
202036526.9831.38270015026129.77
202136526.6830.84243613115629.29
202236526.5330.74233272954129.11
202336527.1431.292844251811029.8
202436527.6731.743249757115630.24
202536526.7930.54248721817729.22
202629627.2————29.51

The sun arrives on the same schedule as the problem

No weather station in the Caribbean measures irradiance, so this comes from NASA's satellite-derived record at Piarco's own coordinates, every day since 2005. The resource is steady rather than spectacular: a 78% clearness index and 5.33 kWh/m² a day on average, with the dry-season months ahead of the wet ones.

Global horizontal irradiance at PiarcokWh/m² per day, monthly mean
024January · 5.02 kWh/m²/day, 82% of clear sky, 682 days in the meanFebruary · 5.51 kWh/m²/day, 82% of clear sky, 621 days in the meanMarch · 5.79 kWh/m²/day, 81% of clear sky, 682 days in the meanApril · 5.77 kWh/m²/day, 79% of clear sky, 660 days in the meanMay · 5.57 kWh/m²/day, 78% of clear sky, 682 days in the meanJune · 5.14 kWh/m²/day, 74% of clear sky, 660 days in the meanJuly · 5.43 kWh/m²/day, 77% of clear sky, 682 days in the meanAugust · 5.61 kWh/m²/day, 78% of clear sky, 682 days in the meanSeptember · 5.69 kWh/m²/day, 80% of clear sky, 659 days in the meanOctober · 5.16 kWh/m²/day, 77% of clear sky, 664 days in the meanNovember · 4.69 kWh/m²/day, 76% of clear sky, 630 days in the meanDecember · 4.54 kWh/m²/day, 77% of clear sky, 651 days in the meanJFMAMJJASOND

Brechin Castle, 92 MW, synchronised July 2025

Yield
1,555 kWh/kW
Capacity factor
17.8%
92 MW would make
≈143.1 GWh/yr
Lost to cell heat
10.9%

An estimate, stated as one: 0.8 performance ratio, NOCT 45°C, temperature coefficient -0.0037 per °C, a 54.4°C cell on a 29.4°C afternoon. Silicon dislikes the heat for the same reason the turbines do, and loses about a tenth of its output to it here.

The sea breezemean wind speed, km/h, by hour
102000:00 · 4.1 km/h01:00 · 3.8 km/h02:00 · 3.6 km/h03:00 · 3.3 km/h04:00 · 3.3 km/h05:00 · 3.3 km/h06:00 · 3.2 km/h07:00 · 3.9 km/h08:00 · 8.3 km/h09:00 · 14.6 km/h10:00 · 17.3 km/h11:00 · 18.5 km/h12:00 · 19 km/h13:00 · 19.2 km/h14:00 · 19 km/h15:00 · 18.6 km/h16:00 · 17.6 km/h17:00 · 15.9 km/h18:00 · 12.4 km/h19:00 · 8.8 km/h20:00 · 6.8 km/h21:00 · 5.8 km/h22:00 · 5.1 km/h23:00 · 4.6 km/h19 km/h3 km/h048121620

Six times as much wind at midday as at dawn, peaking in the same hours as the load. Modest for generation, and it is also the only thing cooling the inlets.

Where every number on this page comes from

The weather. Piarco International (WMO 78970), 416,149 hourly observations, 1941 to today, from NOAA's Integrated Surface Database for the history and Iowa State's METAR archive for the present. Tobago's A.N.R. Robinson International adds 237,147 hours from 1988. Irradiance is NASA POWER at the same coordinates, because no station in the region measures it.

The fleet. Ministry of Energy and Energy Industries, Electric Power. Which machines are air exposed is our classification: gas turbines yes, steam turbines and reciprocating engines no.

The demand figures. T&TEC publishes no demand series. The peak figures on this page are its own published records, quoted as such. Nothing here models demand.

Wet bulb and heat index. Stull (2011) approximation, valid across this record's whole range. NOAA Rothfusz regression.

Pressure. Measured pressure where the observation carries it (60.5 % of hours), otherwise that month's mean. Tropical pressure varies by a few tenths of a kPa, against a 3 to 4 % density swing from temperature.

Computed 23 October 2026 at 06:55 UTC from 416,149 observations.

Pull the same data yourself

Every series on this page is in the public API, which is the same archive this page reads. Piarco is station 9690, Tobago 9688.

# every hour Piarco has ever recorded
GET /api/v2/weather/observations
      ?station=9690&cadence=hourly
      &from=1941-07-15&fields=temp,rh,pressure_msl

# cooling degree days on any base you like
GET /api/v2/weather/degree-days
      ?station=9690&base=24&from=1941-01-01

# satellite irradiance at the same point
GET /api/v2/weather/solar?station=8825&from=2005-01-01
API documentation →The whole archive