Harvest Dates, Century by Century

Burgundy has one of the longest continuous climate records anywhere, and nobody set out to create it. The ban des vendanges was the official declaration that harvest could begin, a legal and fiscal event recorded by town authorities from the fifteenth century onwards. The people writing it down were collecting dues, not building a dataset, which makes the series unusually free of the motivated reasoning that afflicts later records, and because the declaration was triggered by ripeness, it tracks growing-season temperature closely.

For five hundred years the mean start date hovered around the last week of September, with occasional very early or very late years at the edges of the distribution. Since the 1990s the whole distribution has moved forward by roughly two weeks. The 2003 harvest began in August, which had happened twice in six hundred years of records. Dates that were once remarkable are now ordinary, and the reference point everybody used has quietly stopped being a reference point.

The reason this matters more to Burgundy than to almost anywhere is that Burgundy's style was a consequence of struggling. It sat in a marginal climate: one where the grape only just ripens in an average year. That produced high acidity, modest alcohol, and aromatics that heat would have cooked away. Nobody chose those characteristics in a cellar; they were what the latitude delivered. A region that no longer struggles to ripen now has to produce that style deliberately, which is a different craft with entirely different ways of failing.

Growing degree days give the shift a number. It is the standard measure of viticultural climate, accumulating warmth above a threshold across the season, on the reasoning that what ripens fruit is accumulated heat rather than any single hot afternoon. The Côte's totals now resemble those recorded several hundred kilometres further south in the 1980s. Put that way the scale becomes legible: the region has effectively moved, without anybody digging anything up.

The most interesting consequence is one that rarely gets stated, because it sounds like good news and is more complicated than that. Warming has compressed the spread between vintages. A cool, wet year that would have produced thin, green, barely drinkable wine in 1977 now produces something perfectly respectable, because the baseline warmth carries it over the line. The great years have improved less, since they were already ripe. So the floor has risen much further than the ceiling.

That has quietly undermined the whole apparatus of vintage charts, which are still consulted as though it had not. A chart encodes the risk that fruit will not ripen, because for five hundred years that was the dominant risk. It is no longer. The dangers now are heat, hail and frost, and those are localised (a storm takes one commune and misses the next) which means they do not summarise into a single regional score at all. A buyer avoiding a vintage on an old chart's advice and finding the wines perfectly good is not experiencing a difference of palate. They are using an instrument built to measure one failure mode to predict another.

None of this says the region is finished, and the more excitable versions of the argument should be treated with suspicion. Burgundy in 2020 made extraordinary wine. What the record shows is narrower and harder to dismiss: the conditions that produced the style the region is famous for no longer obtain reliably, and everything the growers now do in the vineyard and the cellar is an attempt to reproduce deliberately what used to arrive by default.