Stage 21: Rambouillet – Paris

The grand finale takes us, as usual, to Paris. This year, we started at the Château de Rambouillet and headed past Versailles and into Paris. The Palace of Versailles is well known for its fountains and the Grand Canal, but all that water has to come from somewhere.

Le Bassin d’Apollon by Adam Perelle (source)

The expansion of Versailles and its gardens by Louis XIV came with a need for far more water than was accessible from the palace. The engineer Vauban was asked to solve this problem and came up with a plan to divert water from the Rambouillet plateau to the southwest of the palace. Part of this plan included the construction of six ponds near the town of Les Bréviaires. These ponds, named the Étangs de Hollande after an old local château, were connected by a long aqueduct to the Étang des Noës and the Étang de Saint-Quentin and then to Versailles.

A map of the ponds and canals that brought water to Versailles (source)

The Étang de Hollande (source: Albanet, CC-BY-SA 3.0)

After the Étangs de Hollande were constructed in 1685, Vauban designed a massive canal to take water some 80 km from the Eure River to Versailles. The canal proved incredibly expensive. The soldiers employed in constructing it were needed to fight in the War of the League of Augsburg in 1688, and they never returned to the canal, leaving it unfinished.

The aqueduc de Maintenon, part of the planned Canal de l’Eure (source: Selbymay, CC-BY-SA 3.0)

And that is a wrap! I hope you enjoyed the racing and the waterscapes of the 2019 Tour de France as much as I did. Check back in September for a special water-themed project during the Vuelta a España and in July of 2020, where the Tour will start in the seaside city of Nice!

Stage 20: Albertville – Val Thorens

Had the stage followed its original route today, the peloton would have climbed the Cormet de Roselend and made their way past the Lac de Roselend, an Alpine reservoir.

The Barrage and Lac de Roselend (source: Opsylac, CC-BY-BA 4.0)

The Barrage de Roselend is part of a large system of hydroelectric projects in the valley of Beaufortain. Roselend was constructed first, between 1955 and 1967, followed by the dams of Saint-Guérin and Gittaz which lie on either side of the Lac de Roselend. The Barrage de la Girotte nearby in Hauteluce was constructed in 1949 at the site of a lake that had been used for hydropower since 1900. The three dams of Beaufortain are connected to generating stations by a network of underground tunnels that run for kilometers.

The Barrage de la Gittaz (source: Florian Pépellin, CC-BY-SA 3.0)

Stage 19: Saint-Jean-de-Maurienne – Tignes (or the Col de l’Iseran)

Today’s stage was cut short by a storm in the Tarentaise Valley. The sudden thunderstorm brought hail and snow and a whole lot of water onto the roads and hillsides of the final descent towards the Lac du Chevril. The sudden addition of all that water created some landslides which completely blocked the road.


A landslide blocking the road today (source)

Landslides have also occurred on the route of the stage tomorrow, so the organizers have shortened that stage to just the climb of the Val Thorens.

When a lot of water falls on the slope in a short period of time (such as during a thunderstorm), the water starts accumulating in the pore spaces of the soil, exerting a force on the particles of soil and pushing them apart. Once that force exceeds the strength of the soil, it breaks the soil apart. Steep slopes mean gravity pulls hard on this loose soil and it runs down the mountainsides. Mass wasting like we saw today is thus fairly common in alpine regions, and there is growing concern that increased precipitation combined with warming soils may make them [even more common](, which does not bode well for future editions of the Tour de France.

Stage 18: Embrun to Valloire

The first proper Alpine mountain stage today took us from Embrun over the col d’Izoard and the col du Galibier to the ski resort of Valloire. Embrun lies at the head of the Lac de Serre-Ponçon, which we encountered two years ago during Stage 18, which finished on the col d’Izoard. The post below is what I wrote about the lake on that occasion.

The Alpine stages are all squeezed in right next to each other, so we start today in Briançon, just downstream from Serre-Chevalier, then head down the valley of the Durance before turning around and heading back up to the col d’Izoard, over which you would head down to get back to Briançon. And tomorrow we’ll start from Embrun, which we encountered along the Durance today. We’ll talk more about the river tomorrow as we follow it into Salon-de-Provence, but today is all about the Lac de Serre-Ponçon.


The Durance, like many of the rivers we’ve encountered this year, has a tendency to flood violently. Significant floods in the mid-19th century drove the investigation of damming the river, but it was not until the 1950s that technology progressed to the point that a dam on the Durance was possible.

The Barrage de Serre-Ponçon was completed in 1959 and holds back the second largest artificial lake (after the lac du Der-Chantecoq). Two villages were flooded by the construction of the dam: Savines was moved to the banks of the new lake while Ubaye was evacuated. The construction of the Barrage is the background for the film L’Eau Vive by François Villiers which features a soundtrack written by Guy Béart, including the title track:

Stage 17: Pont du Gard – Gap

Today’s stage took us from the Pont du Gard outside Nîmes to Gap in the Alps. Along the way, the peloton spent some time in the valley of the Ouvèze where, in September of 1992, a flood devastated the town of Vaison-la-Romaine, the site of today’s intermediate sprint.

Flood waters in Sarrians, downstream from Vaison-la-Romaine (source: JPS68, CC-BY-SA 3.0)

A convective storm system over the 21st and 22nd of September dropped record-setting amounts of precipitation on the Ouvèze watershed: nearly a foot of rain was recorded in some locations within the 24 hour period. Forty-seven people died in the ensuing flood. Damages amounted to around 500 million francs, which was enormous, especially considering that the population of Vaison is only six thousand. Much of the subsequent analysis of the flood of September 1992 has pointed to urbanization within the watershed — with the accompanying increase in impervious surface area — as exacerbating the effect of the high rainfall.

Stage 16: Nîmes – Nîmes

The peloton crossed over the Gardon River several times today on their loop out and back to Nîmes. On the first of these crossings, they made their way across the Pont du Gard, the tallest Roman aqueduct bridge.

The Pont du Gard (source: Benh LIEU SONG, CC-BY-SA 3.0)

The Roman colony of Nemausus did not grow up around a major river, so as it became a regional capital, it needed more water. It was decided that the best nearby source was a spring in Uzès, some 20 kilometers away. Fortunately the Romans were pros at building aqueducts. There was just one problem: the Gardon River. The river sits in a gorge that would need to be spanned to get the spring water to the city, and the needed bridge would have to be massive.

A cross section of the Pont du Gard drawn in 1875 by Alfred Léger. The “pont moderne” is the modern road bridge that was constructed alongside the Pont du Gard in the 18th century. (source)


The Pont du Gard consists of three levels of arches and reaches a height of 49 meters above the low water mark on the river. It weighs around 50,000 tons, solid limestone quarried from a site downstream. The water conduit is carried by the top level and transported around 8 million gallons of water every day to the city.

Stage 15: Limoux – Foix

Today’s mountain stage took us into the town of Foix before heading up the Prat d’Albis. Foix, the hometown of French canal engineer Charles de Freycinet, sits upon the Ariège River, a tributary of the Garonne. But just northwest of the town lies an underground river, the Labouïche.

The underground river (source: site officiel de la rivière souterraine de Labouiche)

The Labouïche sits within a large cave system and is navigable for around 1500 meters, making it the longest navigable underground river in Europe. It was discovered in 1908 by Jules Dunac who, along with his two sons and a couple of young army officers, followed a stream known as the Fajal into the Aïgo Perdent, the Lost Water. Their exploration was taken up by pioneering speleologists Édouard-Alfred Martel in the 1910s and Norbert Casteret in the 1930s, who revealed the currently known extent of the river.

The entrance to the underground river (source: site officiel de la rivière souterraine de Labouiche)

The river terminates in a siphon, a permanently submerged chamber that must be dived to pass. No one has so far managed to cross that siphon. In 2015, Franck Bréhier tried and managed to pass about 100 meters into the siphon before he was forced to turn back.

Stage 14: Tarbes – Tourmalet

The ride up to the Col du Tourmalet takes us up the valley of the Bastan, which was the site of a massive flood in the summer of 2013. The Bastan arises on the western slopes of the Col du Tourmalet and flows through Barèges until it joins the Gave de Gavarnie in Luz-Saint-Saveur. These waters flow down into the Gave de Pau through Lourdes and then through Pau.

Over the winter of 2012 to 2013, precipitation was significantly higher than normal, with a very large accumulation of snow in the mountains surrounding the valley. An avalanche on May 31, 2013, blocked up part of the river, setting the stage for the flood in June. Over the weekend of June 15 and 16, a heat wave developed, melting that tremendous snow pack, and on the 17th and 18th, a storm passed through, dropping an additional seven inches of rain in 48 hours. The Bastan left its banks, overwhelmed the dam at Cabadur and flooded the valley down to Lourdes. Barèges was completely devastated by the flood, and Lourdes suffered significant damage as well.

Stage 12: Toulouse – Bagnères-de-Bigorre

The first of the Pyrenean stages takes us through two bath towns, Bagnères-de-Luchon, just before the climbing started today, and Bagnères-de-Bigorre at the finish line.

The baths at Bagnères-de-Luchon (source)

In 1835, the Scottish physicist James David Forbes took a trip to the Pyrenees and visited hot springs across the mountain range, including those at Luchon and Bigorre. In his travel narrative/geological survey/treatise on thermometers, Forbes had this to say about the springs of Bagnères-de-Luchon:

It is a curious fact that all the chief springs of Bagnères de Luchon issue from the granite within a few feet of one another, although their properties are believed to differ considerably, and their temperatures certainly do. They are kept separate by partitions connected with a vertical wall, into which slabs of stone (which may be removed) are cemented. The springs are called La Reine, La Grotte Supérieure, La Source aux Yeux, La Blanche, and La Froide; but all of these, excepting the two first, are apt to mix with one another; and I even learned that such a mixture was practised in order to give a greater apparent supply to some of the more esteemed of the springs. It is quite certain too that rain water mixes with some of these, which with other facts immediately to be noticed, render observations of temperature here of little avail. My observations were made on the springs as they flowed from beneath the wall just mentioned. These springs are highly sulphureous, and the two whose temperature I measured were copious.

Forbes’s main interest in the springs was their temperature. He would later devote a good portion of his career to the study of heat conduction, so the apparent warming of the springs at Bagnères-de-Luchon was absolutely fascinating:

The springs of Luchon have undergone most surprising changes. Camperdon, a writer of credit, and himself physician for thirty years at this place, assures us that the Source de la Reine was cold until 1755, when (on occasion of the great earthquake of Lisbon) it assumed a temperature of 41° Reaumur. The hot springs of many parts of Europe were affected by the same event. What shows that these springs are much connected with the sources of ordinary springs, is a curious fact mentioned to me by M. Barrau, one of the Inspecting Physicians of Luchon. In 1835, after great rains (in the month of May if my memory serves me rightly), the same spring, La Reine, delivered four times as much water as usual, and its temperature fell 16° Reaumur. This continued for twenty-five days, when it resumed its former state.

Forbes also cared a lot about the accuracy of thermometers. Thus, he (and I) appreciated his correspondent’s method of checking his thermometer before taking the temperature of the springs.

Boisgiraud’s observations were made with great care; I had an opportunity of verifying the accuracy of the copy I obtained of them by the originals. M. Boisgiraud informed me that he had verified the zero of his thermometer by frequently plunging it in melting snow.