Grape harvest dates and the evidence for global warming

I mentioned in my previous post (Statistical variance and global warming) that in Europe there are long-term records of the starting dates for grape harvests, and that this can be used to study changing weather patterns. This is because grape harvest dates are highly correlated with temperature — the warmer the season then the earlier will be the grape harvest. This fact has been much in the news this year, with very hot summer temperatures followed by very early grape harvests in many northern-hemisphere regions.

Written records of harvest dates exist in western Europe because the harvest dates are usually officially decreed, based on the ripeness of the grapes. The grapes are used for wine-making, and this activity has traditionally been under some sort of official control. Thus, we have historical records for many locations over many years.

I have previously shown two long-term datasets for wine-growing regions, one for Two centuries of Bordeaux vintages and one for Three centuries of Rheingau vintages. These both show very large changes in the timing of the start of the grape harvests, especially in recent decades. In this post I will look at some more data.

Much of this data has been compiled into a publicly accessible database archived at the World Data Center for Paleoclimatology (see V. Daux, I. Garcia de Cortazar-Atauri, P. Yiou, I. Chuine, E. Garnier, E. Le Roy Ladurie, O. Mestre, J. Tardaguila. 2012. An open-database of grape harvest dates for climate research: data description and quality assessment. Climate of the Past 8:1403-1418). This database comprises time series for 380 locations (see the map above), mainly from France (93% of the data) as well as from Germany, Switzerland, Italy, Spain and Luxembourg. The series have variable lengths up to 479 years, with the oldest harvest date on record being for 1354 CE in Burgundy.

Burgundy

I have taken the harvest-start data for Burgundy and supplemented it with data from another study (I. Chuine, P. Yiou, N. Viovy, B. Seguin, V. Daux, E. Le Roy Ladurie. 2004. Grape ripening as a past climate indicator. Nature 432:289-290). I have graphed the data below, which shows a complete record of the official start of the grape harvest for every year from 1370 to 2006 CE, inclusive.

The harvest dates are shown relative to the beginning of September (day 0); and the pink line shows the 9-year running average.

This graph shows some very interesting patterns. First, in spite of the ups and downs in the graph, there is no long-term change — harvest starts have pretty much remained within a 4-week period after September 10.

However, the long-term pattern does show two long cycles, with harvest dates getting progressively later through time and then moving earlier again — the first cycle was from 1370 to 1700, and the second from 1700 to now. Super-imposed on these two long cycles, there were smaller 20-year cycles before 1700, and 30-year cycles after that time. For mathematicians, this might be an interesting dataset on which to perform some Fourier time-series analysis.

For our purposes here, there has been a dramatic change in harvest date in recent years, with the earlier and earlier harvests since 1984 being attributed to global warming. However, there was just as rapid a change in harvest dates from 1420 to 1450, although at that time the harvests became rapidly later, due to cooling of the weather (not warming).

This graph thus illustrates what the climate-change skeptics are on about. There have been recordings of previous large changes in the weather, which have affected European agriculture. In that sense, the current change in the weather is not necessarily unusual. The skeptics suggest that we should continue to "suck it and see", to find out whether the weather turns around and becomes cooler again. However, there have been no longer trends of change, and, unlike for the previous longest occasion, there is no current indication that our recent warming trend will reverse itself.

Australia

By way of contrast, we could also look at some shorter harvest trends from elsewhere in the world. The data I have chosen come from Australia (L.B. Webb, P.H. Whetton, E.W.R. Barlow. 2011. Observed trends in winegrape maturity in Australia. Global Change Biology 17:2707-2719).

The longest grape-maturity record presented by these authors is 115 years, for the McLaren Vale region, in South Australia. Unfortunately, the data for 1992-1997, inclusive, are missing, which reduces its utility for studying global warming.

So, instead I will show the graphs for two shorter time-series from central Victoria, one for Shiraz grapes (red) and one for Marsanne (white). Note that the harvest in Australia is in March, not September. These graphs cover 70 years; and the pink lines show the 5-year running average.

Shiraz

Marsanne

The graphs both show relatively short-term cycles in grape maturity, superimposed on longer-term cycles, the same as I noted above for Burgundy. For example, for the Marsanne grapes the shorter cycles seems to be c. 20 years long. Maybe this is a common pattern for wine grapes?

In any case, the move towards earlier harvests is just as obvious in these data as it is in the data for the Burgundy region (and also for Bordeaux and the Rheingau, as shown in my earlier posts). The recent change in agriculture patterns truly is global.

The skeptics

Sadly, Australia is one of the official political homes of climate-change denial. For instance, take this comment by renowned Australian government viticulturist John Gladstones, which is from Wine, Terroir and Climate Change (Wakefield Press, 2011):

How much warming, then, can justly be attributed to anthropogenic greenhouse gases? Taking all evidence into account, the proven amount is: none ... from a viticultural viewpoint we can conclude that any anthropogenic changes to mean temperatures will be small and, for some decades to come, unlikely to have major effects beyond those of natural climate variability.

And yet, here we are, several years later, and we have already reached the limit of climate variability that we have recorded for the past 6+ centuries. How much longer are we expected to suck it and see?

The key word in the climate debate is "prove". We cannot, in the strict sense, ever prove a causal connection between anthropogenic activities and climate change. But, by the same token, we can never prove that the sun will rise tomorrow morning, either. Both cases involve forecasts about the future, and we will only be able to evaluate them in hindsight. By then, of course, it is usually too late, if something has gone amiss.