Field notes: the week the port pile outgrew the mine's story

It started as a rounding error. Mid-month, a coal terminal on the east coast was running a stockpile about four percent above where the loading schedule and the mine's reported dispatch figures said it should sit. Four percent on a pile that size isn't nothing, but it's the kind of gap a desk usually waves off as timing noise between a bill of lading and a weighbridge ticket.

By week three it was closer to eleven percent, and the mine hadn't said a word.

That's the part worth sitting with. A stockpile that grows faster than the production and shipment data on file can account for isn't a mystery in the dramatic sense. It's almost always one of a handful of boring, plumbing-level things: a rail cycle that's slipped and left wagons sitting at the yard instead of turning, a vessel that missed its laycan and pushed cargo back onto the pad, a blend change at the washplant that's holding more low-grade material on site while higher-grade tons get prioritized for the ships that are loading now. None of these show up in a quarterly release. Most of them don't show up in a monthly dispatch number either, because dispatch is reported as a total, not as a pile-by-pile reconciliation.

What tips you off, if you're watching the stacking pattern instead of waiting for the filing, is the shape of the pile itself. A normal build from steady production looks like a pile growing evenly across its length, the stacker making consistent passes. A build from a stalled vessel queue looks different: material keeps arriving at the same windrows because the reclaimer isn't drawing them down, so the pile gets taller and narrower instead of longer. Watch the same site for three or four weeks running and that pattern reads clearly, long before anyone puts out a statement about vessel delays or a weather event at the anchorage.

This particular case turned out to be a laycan slip. Two Panamax berths had been held up by a weather system that never made the shipping press because it was a non-event everywhere except that one anchorage. The mine's production hadn't changed. The port's inbound rail hadn't changed. What changed was the ship side of the ledger, and for nine days the pile just kept climbing because nothing was coming off it.

Stockpile growth doesn't always mean a problem. Sometimes it means a desk is about to get a cheap entry on a spread that's priced in disruption that was never on the production side to begin with. The point is narrower than that: shipment data and quarterly disclosure tell you what happened over a window long enough that the short, tradeable gap has already closed by the time you see it. The pile itself doesn't wait for a filing deadline. It just sits there, visibly bigger or smaller than the story being told about it, for anyone checking often enough to notice.

That's the whole case for watching the physical pile on a cadence tighter than the reporting cycle around it. It's also the reason we built the index that does exactly that off weekly stereo satellite passes rather than another layer of derivative data. The volumetrics don't care what the press release says the mine is doing. They care what's actually stacked at the site this week versus last week, at the mine, the smelter, and the port, independently of each other.

Nine days of unexplained growth sat there as an open question on the pile before anyone called it a disruption. A desk checking the volumetrics every week saw that gap open on day nine. A desk waiting on the next quarterly filing would have seen it, if at all, weeks after the laycan slip had already resolved and the spread had already moved.

If your watchlist has a site worth checking weekly instead of quarterly, that's what Bulk Stockpile Index is built for.