A Country That Declared Two Emergencies for Opposite Reasons
In February 2026, Madagascar’s government filed for a state of emergency — for cyclones. Two of them, in quick succession, along a coastline already bracing for a season of unusually intense storms. In the very same stretch of months, the same government was still operating under a separate national response plan for a drought that has not meaningfully broken since 2023.
One island. One capital. One emergency operations center, filing paperwork for too much water and too little water within weeks of each other.
The geography explains the split — cyclones favor the east coast, drought has settled over the south — but the coincidence is worth sitting with anyway. Ambovombe, Amboasary, and Tsihombe, three districts in the Grand Sud, have gone through a rainfall deficit exceeding 300 millimeters over ninety days this year, on top of a run of failed harvests going back nearly a decade. The World Food Programme’s response plan for the country is already short $18.3 million for the next six months alone, according to the UN’s own tally.1
There is a detail in that shortfall that separates this crisis from most others this space has covered. No militia runs a checkpoint on the road into Ambovombe. No government policy diverted the harvest to someone else’s table. Nobody is profiting from the empty granaries. The rain simply did not arrive on schedule, for the fourth year running, and a region built across centuries to survive on that rhythm cannot absorb its disappearance.
WFP field reports out of the Grand Sud, filed across several successive drought years, keep returning to the same short list of coping measures: families surviving on cactus fruit and on wild tubers usually left for the cattle, and women and children walking several hours each way to reach a water source that no longer runs on any dependable schedule.
That absence — no toll-taker, no siege, no ledger to audit — is exactly what makes the question underneath this crisis harder to answer than the ones this series usually asks.
A Different Dust, A Familiar Verdict
Two days ago, this space traced the mechanics of Haiti’s checkpoint economy, where gangs had turned the ordinary act of moving food into a taxable transaction — a modern echo, it turned out, of medieval river tolls charged on grain barges along the Rhine (see The Toll Road). And two months before that, a piece examined a military siege in North Darfur, where a closed road produced a malnutrition rate near 53 percent within months (see The Siege Ration). Both crises had an author. Somebody built the checkpoint. Somebody closed the road.
This series has never really been about forecasting the next disaster. It exists to ask what a disaster reveals about the people watching it happen — and Southern Madagascar offers no equivalent villain to reveal, which makes it the harder case. History has already forced Americans to sit with a version of that same absence once, on their own soil.
In the 1930s, roughly 100 million acres of the southern Great Plains lost their topsoil to wind, in storms dense enough to blacken the sky over cities hundreds of miles away and dark enough that streetlights came on at noon. Contemporary journalists, and later the historian Timothy Egan in The Worst Hard Time, documented families who woke to find dirt sifted through window frames onto their beds, and children who died of what doctors at the time called “dust pneumonia” — soil packed so thick in young lungs that it showed on autopsy like something closer to drowning.
The Dust Bowl had a natural trigger: a multi-year drought over a region that, unlike Madagascar’s arid south, had never evolved to expect one. But it also had a human layer. Decades of deep plowing had stripped away the prairie grass that once held the topsoil down through dry years, and farmers, chasing wartime wheat prices with new mechanized plows, had converted grassland to cropland faster than the soil could survive once the rain stopped. The drought was weather. The soil loss was a choice, made a thousand small times over a generation.
Preachers in the Plains states argued from the pulpit over which of the two it really was — a natural cycle, a verdict on a nation loosening its grip on old land ethics, or simply bad luck compounded by bad farming. None of the three positions fully won. What survived instead was a policy response: the Soil Conservation Service, founded in 1935, built to repair what the plowing had undone.
Madagascar’s south has no comparable plow-driven backstory. Its farmers practice subsistence agriculture largely unchanged for generations, on land that was already marginal before the current drought sequence began. If the Dust Bowl was a human failure that weather made possible, the Grand Sud may be closer to weather with no human failure standing behind it at all — which raises, more starkly than Haiti or Darfur ever did, the question this series exists to sit inside rather than resolve.
When the Text Refuses to Simplify Its Own Claim
Scripture speaks about withheld rain more directly, and more often, than it speaks about almost any other natural catastrophe. It does not, however, speak about it with one voice.
“And also I have withholden the rain from you, when there were yet three months to the harvest: and I caused it to rain upon one city, and caused it not to rain upon another city: one piece was rained upon, and the piece whereupon it rained not withered. So two or three cities wandered unto one city, to drink water; but they were not satisfied: yet have ye not returned unto me, saith the LORD.” — Amos 4:7-8 (KJV)
Scripture does not require that every catastrophe be interpreted as divine judgment; it does, however, insist that societies eventually reveal the moral conditions under which they have chosen to live.
Amos 4 is a strange passage to lean on for a simple verdict, because the text itself won’t allow one. God, speaking through the prophet, lists five separate withheld mercies — famine, drought, blight, plague, and finally military defeat — and after each one repeats the same refrain: yet have ye not returned unto me. The passage reads less like a punishment log and more like a parent recounting every attempt at getting a child’s attention before finally raising their voice. The drought in Amos 4 is not the point of the chapter. The non-response to the drought is.
“The field is wasted, the land mourneth; for the corn is wasted: the new wine is dried up, the oil languisheth… Be ye ashamed, O ye husbandmen; howl, O ye vinedressers, for the wheat and for the barley; because the harvest of the field is perished.” — Joel 1:10-12 (KJV)
Joel offers something closer to grief than accusation. The prophet does not diagnose the drought’s cause before mourning it. He simply names the loss, crop by crop, in language a farmer in the Grand Sud could recite without translation — the cassava that didn’t set, the maize that never tasseled, the cattle sold off one by one to buy what the field no longer produced.
“He turneth rivers into a wilderness, and the watersprings into dry ground; a fruitful land into barrenness, for the wickedness of them that dwell therein. He turneth the wilderness into a standing water, and dry ground into watersprings.” — Psalm 107:33-35 (KJV)
The psalm is the outlier of the three, and it is worth reading in full rather than in half. Verse 33 sounds like the sentence a certain kind of preacher would stop on. Verse 35 refuses to let him. The same voice that dries a fruitful land also floods a wilderness a few lines later — the passage moves in both directions inside a single breath, and its point, read whole, is less about assigning blame for the drought than about insisting the drought is not the final word on the land.
Later rabbinic commentary on this passage cautions against exactly the shortcut a modern reader might reach for — treating unequal rainfall between two towns as a scorecard on those two towns, rather than a pattern meant for a whole people watching both at once. Taken together, scripture’s drought texts refuse the two easiest exits: they will not let the reader conclude that every dry season is a verdict, and they will not let the reader conclude that none of them are.
The Ledger With No Name on It
The humanitarian numbers out of southern Madagascar this year describe a crisis that is, in the strict operational sense, entirely legible. The IPC’s most recent projection put roughly 1.64 million people in the Grand Sud at high levels of acute food insecurity — a figure that has held roughly steady across three consecutive lean seasons rather than spiking and receding the way a single bad year normally would.2
What makes this ledger different from the Haiti checkpoint economy or the Darfur siege is that every line item traces back to the same root cause, and that cause has no office, no chain of command, and no interview to give. A drought is not a decision. It is a decision’s absence — and it exposes, more clearly than a checkpoint ever could, what a community does when there is no one left to blame.
That is where the human layer reappears, even in a crisis with no human trigger. The $18.3 million WFP funding gap is not weather. It is a choice made in donor capitals thousands of miles from Ambovombe, about which emergencies get funded on schedule and which get funded late, if at all. Southern Madagascar competes for the same finite pool of humanitarian dollars as Sudan, Gaza, Yemen, and a dozen other hotspots with louder headlines and more familiar names. The drought decided who would go hungry first. The funding gap is still deciding, in real time, who gets fed second.
This is the same structural question Nehemiah’s Jerusalem faced during its own famine, though the mechanism there was interest-bearing loans rather than a donor pipeline — a natural shortage, compounded by a human response to that shortage, becoming the layer that could still be changed even after the layer that couldn’t be changed had already done its damage. The drought in the Grand Sud cannot be un-happened. The $18.3 million can still be raised.
Psalm 107 does not end at verse 35. A few lines later, the psalmist describes a hungry people “he setteth… on high from affliction, and maketh him families like a flock” — the same God who dries the fruitful land is credited, within the same psalm, with resettling the hungry and multiplying what little remained. It is not a promise that the rain returns on request. It is closer to an insistence that famine, in scripture’s own accounting, is rarely presented as a closed case.
Two Emergencies, One Question
Madagascar’s government filed two emergency declarations this year for opposite reasons, in the same country, within the same season — one for a sky that gave too much, one for a sky that has given too little for going on four years. Both filings landed on the same desk. Only one of them came with an obvious villain, and it wasn’t the drought.
The Dust Bowl taught one generation of Americans that a catastrophe can be partly weather and partly a human hand no one measured until it was too late to matter. Southern Madagascar may be teaching the harder version of that same lesson: a crisis that really is, this time, mostly weather — met by a response that is not weather at all.
Scripture’s drought texts were never trying to settle which kind of catastrophe a given famine belongs to. Amos, Joel, and the 107th psalm sit in obvious tension with each other, and none of the three authors seem bothered by that. What they share instead is a refusal to let a dry season pass as simply unremarkable. God responds to withheld rain by naming it. Nehemiah’s Jerusalem had to respond to its famine by undoing what its own nobles had done to it. Southern Madagascar is waiting on a third kind of response, the kind that shows up as a wire transfer rather than a verdict — and every drought, in the end, becomes less a question about rain than about who still answers when it stops.
Sources: WFP, “Hunger Hotspots: FAO-WFP Early Warnings on Acute Food Insecurity,” June–November 2026 Outlook (2026) · IPC, “Madagascar: Acute Food Insecurity Situation, May 2025–April 2026” · FEWS NET, “Madagascar Food Security Outlook,” October 2025 · Action Against Hunger, “Super El Niño 2026: Food and Nutrition Crises on the Horizon” (2026) · UN OCHA, Madagascar Country Page (2026) · Timothy Egan, The Worst Hard Time (2006)
¹ WFP’s 2026 country operational plan for Madagascar requested $287 million to reach 1.2 million people; the $18.3 million cited in the text is the shortfall specifically against the next six months of that plan, per UN reporting.
² The IPC’s Grand Sud figure of 1.64 million includes roughly 110,000 people already in Emergency-level (Phase 4) need; FEWS NET separately projects continued humanitarian dependence for up to two million people through the first quarter of the year. Both draw on a joint government–humanitarian analysis under the Cadre Harmonisé framework, using district-level agricultural and market monitoring data collected through early 2026.
New essays every week. Delivered quietly.