The Unfinished Map: Observing the Indian Ocean through Model-Corrected, Fleet-Based Sensing

Derin George 

We make decisions about the ocean-the largest and most consequential system on Earth-from a sample so sparse it would be judged inadmissible in almost any other science. Less than one-thousandth of one percent of the deep seafloor has ever been seen by a human eye or camera; most deep-ocean observation has been carried out by a handful of wealthy nations; and the Indian Ocean-which has absorbed a disproportionate share of the planet's added heat and governs a monsoon on which more than a billion people depend-remains among the least-observed basins of all. This observation deficit, I argue, is not fundamentally a problem of reach but of method and economics-and both have quietly changed.
I set out an approach to observing the Indian Ocean built on two premises. The first is that the ocean should be known the way a brain knows the world: through a continuously running model corrected by measurement only where it is uncertain-a case for data assimilation and prediction over the impossible ambition of measuring everything. The second is that the in-situ observing platform the field has waited for is already at sea, in the form of the artisanal fishing fleet-a route to dense, low-cost, participatory ocean sensing across a data-poor basin. I describe the scientific basis for this approach, the inference techniques that let a single modest sensor payload report on far more than it directly measures, the specific case for beginning in India and the Indian Ocean, and-at length, and without flattery-the problems that make it hard. This is not a report of results. It is a research and engineering programme, and an invitation to improve it or to refute it.