How Do Hurricanes Affect Gas Prices? 12 Gulf Storms Compared

In this 12-storm set the national pump price rose by more than 20 cents within two weeks only when refinery outages pushed wholesale gasoline up by 50 cents or more in New York Harbor as well as on the Gulf Coast: Katrina and Harvey added 34.5 and 31.9 cents a gallon. Shutting in offshore oil does not do it. Isaias, in October 2026, shut in 71.5% of Gulf oil output and the national average was 7.0 cents lower two weeks later.

The offshore number is the one that makes the headlines, and across twelve storms it predicts almost nothing. BSEE’s peak shut-in share ran from 24.5% (Harvey) to 100% (Rita and Gustav). The storm at the bottom of that range pushed the pump price up 32.5 cents at its peak, and Delta, with 91.7% of Gulf oil shut in, left it 0.5 cents lower. What did line up with the pump was whether the wholesale jump reached the East Coast market: where the New York Harbor spot price rose 55.8 cents or more above its pre-storm level (Harvey, Katrina), the pump rose 32.5 and 45.9 cents at its peak, and where it rose 36.3 cents or less, the pump never rose more than 18.7.

Twelve Gulf storms, one row each

The set runs from the two 2005 storms to the one that came ashore on October 9, 2026. It is a selected set, not a census: it covers the big refinery events (Katrina, Rita, Gustav, Ike, Harvey, Laura, Ida), two weaker hurricanes (Isaac, Barry), and the newest storms with a complete BSEE, NHC and EIA record (Delta, Francine, Isaias). Storms that BSEE lists but that shut in little production, such as Ian in 2022, which peaked at 11%, are left out. Table 1 gives the storm, the offshore response and what that response meant in barrels.

Federal agencies now label the water the Gulf of America, and the current BSEE and EIA series carry that name. Gulf of Mexico is used here for the same series because the older releases do. The October 2026 Isaias releases are issued under the name Marine Minerals Administration on the BSEE site, which is how the agency’s own pages print it.

Offshore shut-ins: big percentages, shrinking barrels

A shut-in is a precaution, not damage. BSEE’s Laura report describes the procedure plainly: personnel evacuate, and the operator closes the subsurface safety valves, “effectively shutting-in production from wells in the Gulf.” BSEE says the percentages are estimates built from operators’ daily reports of the oil they expected to produce that day, so they measure the precaution and not the damage. Seven of the twelve storms reached 90% or more at some point. The agency’s releases add that once standard checks are complete, production from undamaged facilities comes back immediately, while damaged facilities take longer.

How long the shut-in lasts varies far more than how high it goes. Laura’s share fell from 84.3% on August 28, 2020 to 9.07% on September 5, nine days after landfall. Ida was slower: 66.36% on September 10, 2021 and 16.18% in BSEE’s final report on September 23. After Ike, BSEE’s combined Gustav and Ike series still showed 48.0% of Gulf oil shut in on September 29, 2008, 16 days after landfall, with the earlier Gustav shut-in layered underneath. The weaker storms cleared faster: Isaias was down to 9.92% five days after landfall, Francine to 5.62% six days after and Delta to 7.94% seven days after.

What a given percentage means has changed. In July 2005 the Federal offshore Gulf produced 1,448 thousand barrels a day, 27.6% of the 5,253 thousand barrels a day the whole country produced. In August 2026, the latest month EIA has published, the Gulf produced 2,075 thousand barrels a day, but that was only 14.9% of a national total of 13,965. Katrina’s 95.2% shut-in removed an estimated 1,378 thousand barrels a day, about 26% of US crude output. Isaias’s 71.5% removed an estimated 1,484 thousand, 10.6%. Harvey’s 24.5% removed an estimated 430 thousand, 4.6%. The Gulf now supplies about 15% of US crude, not the 28% it did when Katrina hit, so the same shut-in percentage takes a smaller bite out of the national total.

Crude prices did not react to these shut-ins in a consistent direction. The Cushing WTI spot price two Mondays after the base Monday, expressed in cents per gallon (the dollar change divided by 42), was -9.8 for Katrina, -0.2 for Harvey, +2.3 for Isaac, +8.8 for Ida and +33.0 for Isaias, each against its own pre-storm Monday. None of these is a storm effect on its own, because each includes everything else crude was doing in those weeks. They do show that the offshore shut-in left no consistent mark on the crude price, and gasoline is made at a refinery, not at a wellhead.

Refineries, where the weekly data shows the damage

The Gulf’s importance moved from the wellhead to the refinery gate while its oil share was falling. EIA’s annual capacity series puts the Gulf Coast (PADD 3) at 8.09 million barrels a day of operable atmospheric distillation capacity on January 1, 2005, 47.2% of the 17.12 million national total. On January 1, 2026 the figures were 9.88 million and 18.16 million, 54.4%. Offshore wells supply a smaller share of US crude than they did in 2005, while the coast’s refineries hold a larger share of US refining capacity. Our guide to the PADD districts shows how the Gulf Coast plants feed the East Coast and the Midwest by pipeline.

EIA reports refinery utilization every Wednesday for the week ending the previous Friday. Table 2 takes the last weekly reading before each storm’s base Monday as the starting point, the lowest weekly reading within 14 days after landfall as the low, and the first later week in which utilization had regained half the points lost as the recovery date. A full return to the starting level is a poor test, because late September is when refiners begin their autumn maintenance; our refinery turnaround guide explains why utilization tends to drift down after Labor Day. Half the loss regained is a test that a late-season storm can still pass.

Two pairs of storms belong together. Katrina took weekly utilization from 97.1% to 86.8%, and Rita, less than four weeks later, took it from 90.8% to 69.8%. Measured from the pre-Katrina week to the post-Rita low, the loss was 27.3 points. Gustav cut it from 88.7% to 77.4% and Ike, twelve days later, to 66.7%: 22.0 points from the pre-Gustav week. For these pairs the combined figure is the one to use, and the second storm’s “loss” in Table 2 is partly the first storm’s unrecovered damage.

The single-storm refinery cases are Harvey, Laura and Ida. Harvey cost 17.7 points nationally, and 34.6 points on the Gulf Coast, from 95.3% to 60.7%; EIA’s Today in Energy for September 11, 2017 reported that gross inputs to Gulf Coast refineries fell by 3.2 million barrels a day, or 34%, in the week ending September 1, the largest drop since Gustav and Ike in 2008. Laura took the Gulf Coast from 83.7% to 63.9%. Ida took it from 91.7% to 72.9%, and EIA’s This Week in Petroleum for September 9, 2021 counted at least nine refineries, about 2.3 million barrels a day of capacity, shut or running at reduced rates, with roughly 1.6 million barrels a day still offline on September 7.

The no-refinery cases are just as informative. Isaias, with 71.5% of Gulf oil shut in, left national utilization at 92.7% the week before, 93.5% in its landfall week and 92.8% two weeks later. The Gulf Coast series went from 95.0% to 94.9% at its lowest. Barry cost refiners 1.7 points, Francine 1.9. The refinery loss does not rank the pump response, either. Katrina’s 10.3 points ended in a bigger price rise (+45.9 cents) than Harvey’s 17.7 (+32.5), and Laura’s and Ida’s 10 points or so ended with almost no rise. The weekly refinery numbers say whether the supply side of the gasoline market was hit.

Cents at the pump, and what crude did in the same weeks

The pump series is EIA’s weekly national average for regular gasoline, all formulations, which is dated to Mondays. For each storm the base is the last Monday on or before the US landfall date, which is the last price reading before landfall (Katrina’s base Monday is the landfall day itself, so the first hours of its run-up are already in the base). Changes are then taken to the Monday one, two and six weeks later, in cents per gallon. In Table 3 every pump change is paired with the change in Cushing WTI over the same dates, converted to cents per gallon, because a pump price moves with crude whether or not there is a storm.

Two storms moved the pump. Katrina’s national price rose 45.9 cents in the first week, from $2.610 on August 29, 2005 to $3.069 on September 5, was still 34.5 cents above its base at two weeks and 23.8 cents above at six, with crude lower at both points. Harvey’s first Monday after landfall, August 28, 2017, showed only +3.9 cents, two days after a Friday-night landfall. The price then went up 28.0 cents in a single week, from $2.399 to $2.679 on September 4, in line with EIA’s own account of a 28-cent weekly rise, and peaked at +32.5 cents on September 11. Crude was flat for Harvey (-0.2 cents at two weeks) and 7.6 cents higher at six.

A middle group moved the pump a little. Isaac’s price was 7.1 cents above its base at two weeks and 7.4 at six, with crude 14.5 cents lower at six weeks, and its peak was 10.2 cents in week three. Rita added 14.2 cents in two weeks, from a base that still carried part of Katrina’s premium, and then fell 30.6 cents by week six, of which 17.6 cents is crude. Ike’s week-one gain was 18.7 cents.

The remaining storms barely registered, even where the physical event was large. Laura, with 84.3% of Gulf oil shut in and the Gulf Coast refineries at 63.9%, left the national price 2.9 cents higher at two weeks and 1.0 cent lower at six. Ida, at 95.65% shut in and 72.9% on the Gulf Coast, was +3.1 and +4.5 cents. Barry, Delta, Francine and Isaias were between +0.7 and -7.0 cents at two weeks. Table 3 gives the crude column next to each row for a reason: Isaias’s -7.0 cents came while crude, in cents per gallon, rose 33.0 cents, a combination in which the storm played no visible part.

Eight Mondays after four storms

Table 3 reads three dates. The weekly path shows the shape. Table 4 follows the two storms that moved the pump, Katrina and Harvey, beside Ida and Laura, the two refinery-hit storms that did not, with the crude equivalent for each Monday in brackets.

Katrina’s premium decayed fast and then Rita re-opened it: +45.9 cents in week one, +17.6 in week three, then +31.8 in week five as Rita’s shut-ins and refinery losses arrived, and -0.7 by week eight. Harvey’s was slower and smoother: +31.9 in week two, +32.5 in week three, +20.5 in week six and +12.9 in week eight, by which time crude had risen the equivalent of 10.6 cents, leaving almost none of the gain attributable to anything else. Laura’s path is flat: +4.0 in week one and slightly negative from week four.

Ida’s last two rows are a caution about reading late numbers as storm effects. The pump was +12.2 cents in week seven and +17.7 in week eight, October 11 and 18, 2021, six and seven weeks after landfall, and the crude equivalent over the same dates was +35.7 and +40.4 cents. That is a crude rally reaching the pump on a lag, and no sequel to the hurricane.

Wholesale prices lead, and only some spikes reach the East Coast

The pump price follows the wholesale gasoline market with a delay. EIA publishes daily spot prices for conventional regular gasoline at two hubs that matter here: the Gulf Coast, where a refinery outage shows up first, and New York Harbor, the East Coast market that the Gulf Coast supplies by pipeline and ship. Table 5 gives the highest daily price at each hub in the three weeks after the base Monday, less the base Monday price, beside the pump price peak that followed. Our rockets and feathers guide covers why retail follows wholesale up fast and down slowly.

Only two storms lifted New York Harbor by more than 40 cents. Katrina’s spot price there peaked 117.8 cents above its base on August 31, 2005, the same day the Gulf Coast price stood 82 cents up (its own peak of 89.8 cents came on September 1), and Harvey’s peaked 55.8 cents above on August 31, 2017. No other storm in the set took the New York Harbor price more than 36.3 cents above its base. A count of trading days at 20 cents or more above the base at the Gulf Coast hub (EIA’s daily Gulf Coast conventional regular gasoline spot price) in the three weeks does not sort the storms the same way: Katrina had 5 (the window stops on September 19, before Rita), Harvey 8, Rita 6, Ida 5 and Ike 3. At the New York Harbor hub the same count gives Katrina 3, Harvey 9, Rita 0, Ida 3 and Ike 2. How high the East Coast market went mattered more than how long the Gulf Coast premium lasted.

The 2008 pair shows a Gulf Coast spike can be enormous and still not carry. On September 12, the day before Ike made landfall, the Gulf Coast spot price stood 182.3 cents above the September 1 base used for Gustav and 162.4 cents above the September 8 base used for Ike, the highest peaks in the set by a wide margin. New York Harbor peaked only 36.3 and 35.5 cents above the same bases, and the pump price rose 15.5 cents by September 15 against the Gustav base and 18.7 cents against the Ike base. A Gulf Coast premium that does not reach the East Coast market stayed, in this set, a regional event with a small national pump effect.

Where the jump reached the East Coast, the pump followed there too. In Katrina’s first week the EIA regional average rose 68.1 cents in the Central Atlantic and 54.4 in the Lower Atlantic, against 38.5 on the Gulf Coast. In Harvey, the Lower Atlantic rose 47.4 cents by week three against 38.8 on the Gulf Coast by week two. The data show the spread and do not isolate its cause, but EIA’s Harvey account noted that the Colonial Pipeline, which connects 29 refineries and carries up to 2.5 million barrels a day, briefly curtailed operations and resumed at reduced rates of flow on September 6, 2017. The pipeline link between the Gulf Coast and the East Coast is why the PADD 1 and PADD 3 relationship matters to a Gulf Coast storm at all.

Ida is the row to be careful with. EIA’s This Week in Petroleum of September 9, 2021 said estimated gasoline margins had passed 70 cents a gallon in late August and that margins would stay elevated while Gulf Coast operations were disrupted, and it forecast a September retail average of $3.14. The Gulf Coast spot price did spike 33.7 cents above its base on August 30, but it was only 14.1 cents above on the Monday after, and within six weeks the national price never rose more than 4.5 cents. Gasoline stocks do not account for it: they were 2.0% below the five-year mean on August 27, 2021, and EIA reported draws of 3.6 million barrels on the East Coast and 3.2 million on the Gulf Coast in the week that followed. The data here cannot say which factor separates Ida from Katrina, so no guess is offered.

Isaias, October 2026: 71.5% of Gulf oil shut in and a falling pump price

The newest storm in the set is the cleanest test of the offshore-only case. The Marine Minerals Administration releases on the BSEE site show Gulf oil shut in at 25.08% on October 7, 62.89% on October 8, 71.51% on October 9, 68.76% on October 10, 57.54% on October 11, 29.63% on October 12, 16.13% on October 13 and 9.92% on October 14. NHC’s public advisory 13, issued at 4 pm CDT on October 9, put Isaias’s sustained winds at 115 mph and expected landfall that evening on the coast of the western Florida Panhandle; by advisory 14, at 10 pm, the center was moving over the Panhandle with winds down to 90 mph. The storm passed east of the Louisiana and Texas refining centers.

EIA’s weekly refinery reports show nothing. National utilization was 92.7% in the week ending October 2, 93.5% in the landfall week, 93.4% a week later and 92.8% on October 23. The pump price was $4.354 on October 5, $4.331 on October 12, $4.284 on October 19 and $4.269 on October 26, which is 7.0 cents lower at two weeks. The Gulf Coast spot price drifted from $3.521 to $3.110 across the same four Mondays, a peak of only 4.3 cents above its base. Meanwhile WTI went from $96.13 to $110.01 on October 19 and $105.25 on October 26. The six-week Monday for Isaias is November 16, and the row in the tables will be completed when EIA publishes it.

The surrounding year adds a condition. On October 22, 2026 EIA’s Today in Energy reported that US refinery utilization averaged more than 95% between May and September, the highest for those months since 1998, with tight global product supplies boosting margins. Katrina and Harvey, the two storms that moved the pump, began at 97.1% and 95.4%. Had a storm taken Gulf Coast refineries offline in a summer month of 2026, it would have started from a similarly tight fleet. That describes the starting point and says nothing about whether any storm will do it.

Reading the six-week column without fooling yourself

The +6 column has three problems that the +2 column mostly avoids. First, crude: Gustav’s base Monday was September 1, 2008, six weeks before October 13, and WTI fell from $115.55 (the August 29 close, the last trading day on or before the Monday) to $81.17 in that time, the equivalent of 81.9 cents a gallon. The pump fell 52.9 cents. A reader who looks only at the pump column would conclude that Gustav and Ike cut gasoline prices, when the 2008 crude collapse did that and the pump fell less than crude did. Table 3 is built so that the crude column answers that question every time.

Second, the net of crude is a rough gauge. For Katrina the pump was 23.8 cents above its base at six weeks while the crude equivalent was 13.3 cents below, so the pump finished 37.1 cents above where crude alone would have put it. For Harvey the gap was 12.9 cents (+20.5 against +7.6) and for Rita -13.0 cents (-30.6 against -17.6), the last being Katrina’s premium draining away. Those differences include the normal lag of retail behind wholesale, which can run for weeks, and they include any seasonal changes in the gasoline blend, so they are an indication of the storm’s residue and not a measurement of it. Our summer versus winter gasoline guide covers the seasonal blend change that arrives in the autumn.

Third, storms overlap. Katrina’s weeks four to six include Rita and Gustav’s weeks two to six include Ike, which is why Table 3 labels them. The +1 week figures are the cleanest: Gustav’s first-week change was -3.2 cents, which says that the first of the 2008 storms did nothing to the pump on its own and that everything attributed to it at two weeks is Ike.

Sorting the next storm by how far the wholesale jump travels

Ranked by the New York Harbor spot price peak, the twelve storms fall into three groups, and the pump price peak sorts with them. This is an observed pattern in a small sample that overlaps with itself, so it is a way of reading evidence as it arrives and not a model with a threshold.

Three signals, in the order a market watcher meets them. The offshore shut-in percentage separates nothing. Laura, Delta, Isaias, Barry and Francine (84.3%, 91.72%, 71.51%, 72.82% and 41.85%) are all in the first group; Isaac, Ida, Rita, Gustav and Ike (94.99%, 95.65%, 100%, 100% and 99.9%) are in the second; and the third holds Katrina at 95.2% beside Harvey at 24.49%.

The refinery signal separates somewhat. Of the seven storms that cost refiners 10 points or more, the two that started at 95% or higher, Katrina and Harvey, are the two in the third group. Rita, Gustav and Ike began between 78% and 91%; Ida started at 92.4% and Laura at 82.0%.

The New York Harbor price draws the clearest line, and it is the same line. The gap between the second and third groups is 19.5 cents at the Harbor (36.3 against 55.8) and 13.8 cents in the pump peak (18.7 against 32.5). The Gulf Coast price alone does not sort: Gustav’s and Ike’s Gulf Coast peaks of 182.3 and 162.4 cents and Rita’s 55.5 all exceed Harvey’s 51.4. With twelve storms, two pairs of which share weather and damaged plants, that is a pattern to watch and not a threshold to trade on. It fits the sequence in EIA’s Harvey account: Gulf Coast refinery inputs fell by a third, the Colonial Pipeline curtailed, and retail gasoline rose 28 cents in a week.

What to watch the week of a landfall, then, is short. The New York Harbor spot price against its pre-storm level on the daily EIA table, beside the Gulf Coast price; the following Wednesday’s utilization figure, for the Gulf Coast as well as the nation, in the weekly status report (the EIA release schedule gives the dates); and whether the East Coast premium climbs past 50 cents. If it does, the market is in the 2005 and 2017 pattern, where the national price peaked 32 to 46 cents above its base. If it stays near 10 cents or less, the offshore percentage has not mattered to the pump in this set. For the conversion from wholesale to retail, see the crack spread guide and the refinery utilization explainer.

The next test is already on the calendar. When EIA posts the November 16 price, the sixth-week cell for Isaias will show whether 71.5% of Gulf oil shut in, with the refineries untouched, left any trace at the pump. The PetroEyes gas price page tracks the weekly figure, and the breakdown of what a gallon of gasoline costs shows where the refinery margin sits among taxes, crude and retail costs.

Data and sources, all checked November 3, 2026. U.S. Energy Information Administration weekly retail gasoline prices, regular all formulations: U.S., Gulf Coast, Lower Atlantic, Central Atlantic and Midwest; spot prices: Cushing WTI, Gulf Coast conventional gasoline and New York Harbor conventional gasoline; refinery utilization U.S. (WPULEUS3) and Gulf Coast (W_NA_YUP_R30_PER); total motor gasoline stocks (WGTSTUS1); crude production Federal Offshore Gulf of Mexico (MCRFP3FM2) and U.S. (MCRFPUS2); operable distillation capacity Gulf Coast and U.S., annual as of January 1. EIA Today in Energy, Hurricane Harvey caused U.S. Gulf Coast refinery runs to drop, gasoline prices to rise (September 11, 2017) and U.S. refineries have been running at very high rates in 2026 (October 22, 2026); EIA This Week in Petroleum, September 9, 2021. Offshore shut-in statistics: BSEE hurricane activity updates (daily releases for 2011 to 2026, including the Marine Minerals Administration releases for Isaias), Katrina and Rita (MMS evacuation and production shut-in statistics reports, 2005) and Gustav and Ike (MMS activity statistics, 2008). Three releases whose live pages did not load were read from Internet Archive copies of the official pages: Hurricane Isaac, August 29, 2012 (Wayback copy, saved September 8, 2015), Tropical Storm Harvey, August 30, 2017 (Wayback copy, saved August 12, 2022) and Hurricane Ida, August 29, 2021 (Wayback copy, saved August 30, 2021). Storm tracks and landfalls: NOAA National Hurricane Center Tropical Cyclone Reports for Katrina, Rita, Gustav, Ike, Isaac, Harvey, Barry, Laura, Delta, Ida and Francine; for Isaias, NHC advisory archive, public advisories 13 and 14. All changes, shares, counts and groupings are computed from the series listed here.