Three floods, one monsoon, and the lifeboats we are still not filling.

Three floods in one monsoon

We are a small foundation working out of Siliguri, in the corridor of land where Nepal, Bhutan, Bangladesh and north-east India are all a few hours’ drive away. In an ordinary year that means we respond to one emergency and read about the rest. This year the work came in three.

They were not the same event. One was a river basin filling the way it has filled for centuries, only harder. One was a mountain failing above a river. One was a plain draining water that fell in another country. What they share is a season, and a set of odds that have been moving underneath all three for decades.

Line those three up and the geography stops being a coincidence. The rain that fills the Brahmaputra, the ice that sits above the Bhotekoshi and the Nepali catchments that drain south into the Ganga are all one system, and it is a system with a thermometer in it.

What warming actually does to rain

The mechanism is not complicated and it is not disputed. Warm air holds more water vapour — roughly 7 per cent more for every degree Celsius, which is a nineteenth-century result, not a modelling assumption. Warming does not conjure storms out of nothing. It loads them. When the conditions for heavy rain do arrive, there is simply more water available to come down.[2]

The IPCC’s Sixth Assessment states the consequence plainly: the frequency and intensity of heavy precipitation have likely increased over a majority of land regions with good observational records, and human influence is likely the main driver of that intensification.[1] The report also puts it in the language that engineers and district officers actually use. A rainfall day that used to arrive once in ten years is now about 1.3 times as likely, and about 7 per cent wetter, at one degree of warming.[1] 2024 was measured at roughly 1.55 degrees above the 1850–1900 average, the first calendar year over 1.5.[3]

In this catchment specifically

The regional picture is sharper than the global one. Work led by M. K. Roxy at the Indian Institute of Tropical Meteorology found a threefold rise in widespread extreme rain events over central India between 1950 and 2015, driven by increasingly variable monsoon westerlies pushing surges of moisture in off the Arabian Sea.[4] The total rainfall of a monsoon has not changed much. Its delivery schedule has. The same water arrives in fewer, heavier events, which is precisely the pattern that drainage designed for a gentler distribution cannot take.

Above Nepal the change is in the ice. The ICIMOD assessment of the Hindu Kush Himalaya found glaciers there disappearing 65 per cent faster in 2011–2020 than in the previous decade, with up to 80 per cent of current volume projected to go by 2100 on present emissions.[5] Ice that retreats leaves behind slopes that were being held together by it, and lakes that were not there before.

The honest limit of all that

None of this says that warming caused the avalanche above the Bhotekoshi on 26 August, or the Ganga’s level at Sultanganj this month. Attributing a single event is a separate piece of science with its own methods, and for these three floods no such study has been published. Anyone telling you otherwise this soon is guessing.

What the evidence describes is the deck, not the hand. The odds of a heavy-rain day have moved, measurably, in the direction of heavier. Whether any particular flood was one of the extra ones is usually unanswerable and, for a family in a relief camp, beside the point.

The most expensive near-miss in the history of being sure

Everyone knows the story, and the version everyone knows is the least useful part of it. Arrogant people built a ship, called it unsinkable, hit an iceberg and got what was coming. It is a satisfying story because it lets the listener off. We would not have been that stupid.

The record is duller and considerably worse.

The word had a qualifier, and the qualifier went missing

Nobody official ever said “unsinkable” flat out. The phrase came from a trade magazine: in 1911 The Shipbuilder ran a special issue on the Olympic-class liners and described the watertight-compartment arrangement as making the vessel “practically unsinkable”. The White Star Line’s own promotional copy was hedged too — “as far as it is possible to do so, these two wonderful vessels are designed to be unsinkable.” Harland and Wolff, who built her, have always insisted they claimed no such thing.[10]

Between the trade press and the dinner table, the qualifier fell off. By the time the word reached the passengers it had become a flat statement of fact, and several survivors later testified that this was exactly what they had understood. A carefully hedged engineering claim had been sanded down into a promise nobody had made.

That trip is not a historical curiosity. It is what happens to climate language every day, in both directions. “A one-in-a-hundred-year flood” is a statement about a one per cent chance each year, on a distribution that is currently moving — and it gets heard as “not in my lifetime.” The hedge is doing the opposite work, and it still gets stripped.

The lifeboats were legal

Here is the part that should be uncomfortable. The Titanic was not under-equipped by the standards of her day. She was over-equipped. The British Wreck Commissioner’s report states it in so many words: the ship “carried boat accommodation for 1,178 persons, a number far in excess of the requirements” of the Board of Trade table then in force.[6]

Those requirements came from rules made under the Merchant Shipping Act 1894. The table set a floor of sixteen boats under davits for every vessel of 10,000 gross tons and above — and then it stopped. There was no further gradation, because when the table was drawn nothing much larger than 10,000 tons existed. Titanic measured 46,328 gross tons. She carried twenty boats against a requirement of sixteen, and her owners could accurately describe themselves as exceeding the safety regulations.[7]

Eighteen years passed between the rule and the ship. In those eighteen years the ships quadrupled and the table did not move. Nobody was corrupt and nobody was stupid. A number that had once been generous simply stayed where it was while the world it described walked away from it.

And then the boats went out half empty

The second failure is the one that is genuinely hard to read about, because it happened after everyone knew.

The first boat lowered, number seven, went down the side with 28 people in a boat built for 65. Boat number one left with twelve people in a boat for forty. Across the night, roughly 711 people came off in boats certified for 1,178. About 467 seats — seats that existed, that were paid for, that were legally required to be there, that were successfully launched — went into the North Atlantic carrying nobody.[8]

The reasons were not cowardice or malice. They were ordinary. The ship was lit and warm and very nearly level, and the boat swinging over eighteen metres of black water was obviously the less safe place to be. The first boats went at one in the morning into a situation that did not look like an emergency. Some officers were unsure the davits would take the weight of a full boat. Passengers below decks could not get up to the boat deck in time. Every one of those is a reasonable, explicable, locally sensible decision, and they add up to four hundred and sixty-seven empty seats.

Two years later the first international Safety of Life at Sea convention required lifeboat places for every person aboard.[11] The rule changed afterwards. That is the only part of this story that happens reliably.

We are one of the half-empty boats

It would be easy to end this on governments and design codes. The more honest ending is closer to home.

Our Nepal response worked. The convoy loaded in Siliguri on 31 August, crossed after several days of delays, and the goods were handed over at holding sites in Trishuli and Rasuwa on 10 September. Nearly all of the money that paid for it arrived within about three days of a single video going up, and very little arrived after that. The need did not change in those three days. The attention did.

Then there is Bihar. The appeal page has been open since 13 September. It carries the official figures with their sources, the dates, an explicit statement that we have not been there and have no partner there yet, a ring-fenced fund, and a written undertaking to refund every rupee if we cannot get there within sixty days. It is, by our own standards, the most carefully built thing on this website.

As of this morning it has received no donations at all. Not a small number. Zero, from a page that has everything a donor could reasonably ask for.

Nobody decided that 46.88 lakh people did not matter. There was simply no photograph, no reel, no moment that made anyone look — and a launched, seaworthy, fully compliant boat sat in the water with nobody in it. We built the boat and did not fill it. That failure is ours, and it is the same failure as 1912’s, at a scale we can actually do something about.

Which is the small, unglamorous point this whole piece exists to make. The argument about whether the climate is changing was settled by people with instruments a long time ago. The argument that is still live is the one that ran on a lit, level deck at one in the morning: whether anybody gets into the boat while there is still time, or waits until the angle of the floor makes the decision for them.

The boat that is currently empty

Assam is closed and Nepal’s first phase is done. The appeal that is open, and unanswered, is Bihar. Gifts through these buttons are tagged to the Bihar Flood Relief 2026 fund, which was opened on 16 September 2026, so they are reported separately from our general donations rather than disappearing into them.

Or give another amount. Payments are handled by Seva Stack. Donations to Humrahi Foundation carry 50 per cent tax exemption under Section 80G and the receipt is issued automatically. If we cannot get to Bihar within sixty days of the appeal opening, every rupee is refunded — that undertaking, and the full reasoning behind opening an appeal before we have been anywhere, are set out on the appeal page. Our registrations, accounts and reports are on the transparency page.

Not in a position to give money? Two things help about as much. Goods work for Bihar in a way they do not for Nepal, because there is no customs border between Siliguri and Araria — our year-round collection point takes clothes and sealed dry food. And if you found this piece worth reading, send it to one person who has not seen it. That is not a throwaway line. On the evidence of the last three weeks, it is the single variable that decides whether an appeal is answered.