A dry winter is not a weather story

Every few years the winter rain fails and the reporting follows a familiar shape. A story about wheat. A story about a village whose spring has stopped. A story about tankers in Kathmandu. Each is filed under weather, each is true, and each disappears when the monsoon arrives.

Read together, they are not three stories. They are one system reporting the same failure at three points.

The chain#

Winter precipitation in the hills does two things. It supports the winter crop directly, and it recharges the shallow aquifers that feed hill springs. A spring is not a thing that exists independently of last season’s rain; it is a delayed signal from it.

So when winter rain fails:

• The wheat crop is hit first, within weeks. This is visible, and it gets reported.
• The springs decline months later, on a lag long enough that nobody connects the two events.
• Households that lose a spring shift to whatever else is available, which in the valley means groundwater, and in the hills often means walking further or buying water.
• The additional abstraction is unmetered and unrecorded, so the system’s response to the shock leaves no trace in any dataset.

By the time the consequence is severe enough to be a story, the cause is two seasons in the past and the connecting evidence was never collected.

The measurement problem is the policy problem#

We treat these as separate because our instruments treat them as separate. Precipitation is recorded by one agency. Crop yield by another. Spring discharge is recorded almost nowhere on a continuous basis. Groundwater abstraction in the Kathmandu Valley is recorded for licensed extraction and largely invisible for the rest.

You cannot manage a chain you can only observe at the first and last link.

This is not a call for a large new monitoring programme, which will not be funded. It is an argument for something narrower: a small number of instrumented springs, read continuously, in catchments where we also have rain gauges. A few dozen well-chosen sites would let us establish the lag and the response function. That is the piece of knowledge that turns a failed winter rain from a weather event into a forecast of what happens to domestic supply six months later.

Why the framing matters#

If a dry winter is a weather story, the response is relief: seed, cash transfers, tankers. Relief is necessary and it arrives after the damage.

If a dry winter is a water system story, it is forecastable. A poor winter rain in January is a signal about spring discharge in the following dry season. That is enough lead time to prepare supply, to stage tankering rationally rather than in a panic, and to tell farmers something useful before they plant.

The difference between the two responses is not money. It is whether anyone drew the chain.