There is more to Flores. The island is spectacular in any case. The Portuguese explorers called it Cabo de Flores (Cape of Flowers) because of the red-flowered flame trees, dotted between the palm trees of the north coast. The landscape is varied, from low-land savannah to volcanic rain forest. It is not as wet here as further west in Indonesia. There are other differences. Flores is east of the Wallace line, the invisible division that runs between Bali and Lombok where there is a sudden change in the flora and fauna. The division goes back a long way. West of the Wallace line was once part of Asia. East of the line it never was: there was always deep water separating the two. Flores is at the heart of a region called Wallacea, where very few large terrestrial mammals migrated from either west or east. Stegodons arrived, but otherwise this was a lost Gondwana. It left ecological room for the komodo dragons to become the top predator, with the stegodons acting as their food source.
There are some standard items on the tourist trails of Flores. Around the western tip of this 700-km long island, the komodo dragons can still be found. The towering thatched houses in the village of Waerobo have become famous, although not easy to reach. The reefs, dolphins and turtles of the coast are a must-see, and views of the rice fields cascading against the hill sides make a lasting memory. But don’t go there if you can’t live without social media. Broadband is still unknown in many places. It is in many ways like Bali but a century behind. And you should not expect much comfort when traveling around the island.
15 years ago the fame of science came to the island. Remains were discovered of a diminutive humanoid, no more than 1 meter tall, who had existed here as recent as 18,000 years ago. They had lived in and around the limestone cave of Liang Bua. They had picked a good home. The cave is huge, 50-meters across, and was once used as a school. The precise status of Homo floresiensis is still a matter of dispute, but a direct link to the African species Homo habilis is considered most likely. The remains have been re-dated, and now this branch of the human tree is thought to have gone extinct 54,000 years ago. They still managed to live through the Toba eruption, helped by the fact that eruption was several thousand kilometers away. But when modern humans expanded into the region, Homo floresiensis faded away, leaving only a few skeletons and their stone tools. After that, the modern humans found their own isolation. Even today, at least five different languages are spoken across the rugged interior.
The fertility of the soil of Flores points at the volcanic heartland. There are no major calderas here, but plenty of volcanic peaks. The twin peaks of Lewotobi erupt most frequently, with the latest event in 2003. The tallest volcanoes are Ebulobo at over 2100 meters, and Inieri at 2200 meters. But the most famous is the volcano of Kelimutu, just over 1600 meter tall. It has limited activity: in historical times there have been a few small phreatic explosions, but no major events. The celebrity status does not come from the eruptions, but from the small crater lakes left by past explosions. These lakes are a natural wonder. May the mountain never erupt again! Otherwise it could destroy a scene that, had it been in the US, would be visited by millions. As it is, many of the visitors are locals on annual pilgrimages to leave offerings. Local lore has it that these lakes are the soul’s final resting place. What a place it is.
Kelimutu is located in an isolated region on the eastern half and the southern side of Flores. The nearest town is Moni, and the closest city is Ende, 50 km away on the coast and of very limited touristic interest. Walking up the mountain from Moni takes a few hours (or you can drive up most of the way leaving just a short easy walk). The walk goes through dense forest; the vegetation on the slopes makes it look like Australia, as befits its location behind the Wallace line. She-oaks are common. The forest ends at the summit, leaving two of the craters clear of vegetation: the denuded rocks are a consequence of the gas emissions from the summit. A third crater is closely surrounded by trees, already suggesting it has less activity.
Most tours aim to be at the summit for sunrise. This may not the best time to be there: it can be a bit crowded, and the best views are in sunlight whilst at sunrise the sun is often obscured by cloud or fog. You may find yourself having to wait a few hours for the clouds to lift. In any case, the lakes are surrounded by high crater walls and remain in shade for some time after sunrise. For the same reason, it may be advisable to avoid the rainy season (December to March) for your visit. The best time is the dry season, June to September.
The visitor will find three lakes, each about 400 meters across, and with evocative names. The western lake, surrounded by trees, is called Tiwi Ata Mbupu, or lake of old people. The two eastern lakes, which share a crater wall, are called Tiwu Nua Muri Kooh Tai, or lake of young men and maidens, and Tiwu Ata Polo or enchanted lake (or lake of evil spirits, in a less tourist-friendly translation). At Kelimutu, souls are divided according to their age and disposition. These two lakes contain vigorous fumaroles below their surface, which may have something to do with the lack of vegetation. The lakes should be viewed from above, and not be approached: they are acidic, and not by a little. A measurement of Tiwu Nua returned a pH of 0.5! The gas emissions will keep you away from the lakes themselves, if the steep crater walls or fence are insufficient. As a warning, one tourist who got too close died here and his body was never recovered.
The lakes are famous for their ever-changing colours. You never know what you will find. If it is foggy, you have little choice but to wait until it lifts. If cloudy, the lakes are visible but the colours are not so clear and tourists can be disappointed, after their long and perhaps arduous journey. But when the clouds break and the sun lights up the lakes, the colours become vivid. Blue, green and black are common, but at other times it can be white and red. And each lake is different. Even the two adjacent ones are never alike. A cursory glance at the Landsat satellite maps of Flores will reveal how remarkable the lakes are. They featured on an old bank note, a telling sign in view of the famous sights of Indonesia it had to compete with. (Not of a particularly high value though: the 5000-rupiah note nowadays converts to about 25 British pence.)
What colours are you likely to see? Tiwu Nua can be light blue, light green, or white. Tiwu Ata Polo can be red, dark green, blue, or brown. Tiwu Ata Mbupu is mostly black or dark blue at present but has also been green or white in the past. The colours can change several times per year, in some cases related to the rainy season but in other cases without a clear cause. During 2016, six colour changes were observed. Tiwu Ata Mbupu has been seen to change rapidly, within days, starting from the crater walls.
The colour activity indicates a lively chemistry. The chemistry of volcanic lakes is a bit of a hot potato. Most research is done on the CO2 content of volcanic lakes, which are seen as a particular threat ever since the disaster at Lake Nyos; there is perhaps less research on other aspects. But the chemistry of the Kelimutu lakes is of interest to science: the chemical conditions are extreme due to amount of dissolved gasses and minerals. The ocean underneath the thick ice of Europa, the frozen moon of Jupiter may have such a chemistry. This ocean has been mentioned as a candidate for life but this may be optimistic, seeing that similar volcanic lakes provide a very hostile environment. What are the Kelimutu lakes like?
Tiwu Ata Mbupu, the westernmost crater, is the centre of a larger crater. It is not fully stable: during rain storms it suffers from landslides where the sides slump into the lake, including boulders. Part of the wall is stabilised by two peaks. Kemmerling in1929 reported fumarole activity both within and around the lake. There was still activity in the 1970’s, but none is present now: of the three lakes it has the smallest gas input. Gypsum crystals have grown in the cracks of the rocks around the lake. The lake is just over 60 meters deep.
Tiwu Nua, the northwestern of the pair of lakes, has vigorous hydrothermal activity, with a strong plume in the centre of the lake. There used to be fumaroles around the lake as well but they are not currently active. The plume convects the water and brings a yellow froth on the surface. Rain scatters and removes the froth. Tiwu Nua is the deepest of the three lakes of Kelimutu, with a reported depth of 127 meters. It is also the most volcanically active: the eruptions of 1938 and 1965 both took place here.
Tiwu Ata Polo, the southeastern one of the central pair of craters, has a thermal plume in the northwest part of the lake. Gas bubbles (probably CO2) show evidence for the hydrothermal activity but the activity is variable. Older fumaroles on the eastern wall are no longer there. The gas input into the water is intermediate between the other two. A white froth is present on the lake’s surface around the plume when the convective activity increases. Some of the waters finds its way into the river Watu Gana. The lake is just over 60 meters deep.
Measuring the water chemistry has been an adventure. The craters are deep and the water is difficult to approach, and both the equipment and the human operators need to be resistant to very aggressive corrosion. A small number of measurements have been made, both of the water and of the sediment at the bottom, but we lack long-term monitoring. It would be nice to know exactly the compositions year-round, for each lake and for each colour of the water, but we don’t. Here is a brief summary of what we do know.
Tiwu Ata Mbupu has the most diluted composition, as might be expected due the lower gas input. The pH is around 3 (orange juice), increasing to 4 (tomato juice) in spring at the end of the rainy season, SO4 is 1600 ppm, and calcium is 430 ppm. Chlorine is 90 ppm and sodium 500pm. The sediment at the bottom is higher in iron than found in the other lakes. Tiwu Nua is 10C warmer, has a pH of 0.5, SO4 is a staggering 50,000 ppm, chlorine 25,000 ppm, sodium 940 ppm, iron 2600 ppm, aluminium 8600 ppm, but calcium is similar to Tiwu Ata Mbupu. There are measurable amounts of lead (4 ppm) and strontium (11 ppm). It has very low oxygen levels. Tiwu Ata Polo is a bit less extreme, with a pH of 1.8 (coca cola), SO4 of 10,0000 ppm, chlorine 3,000 ppm, sodium 240 ppm, iron 1200 ppm and aluminium of 1600 ppm. The bottom sediment is enriched in sulfur and arsenicum The lake lacks the warm water of Tiwu Nua, in spite of being right next to it.
This make the water chemistry of Tiwu Ata Mbupu acid-sulfate, Tiwu Nua acid-brine, and Tiwu Ata Polo is an intermediate acid-saline. The differences are strongly related to the hydrothermal activity, which is almost absent in Tiwu Ata Mbupu, very strong and warm in Tiwu Nua, and weaker and colder in Tiwu Ata Polo. It changes over time. Tiwu Ata Mbupu used to be much more acidic but this lessened over the past century as the hydrothermal activity decreased.
There are two types of hydrothermal activity in lakes. In the usual one, ground water circulates through a layer heated by volcanic activity below, and comes back to the surface. The heating does not affect the composition of the water, and if it enters a lake, that lake becomes warm and perhaps enriched in H2S, but only mildly acidic and not particularly hostile. The other type is where the water interacts with volcanic gasses, and becomes enriched in sulfur, chloride, and fluoride. This can happen either underground, with the resulting water injected into the lake, or the lake can absorb gasses directly from fumaroles located below the surface. Such lakes become highly acidic, and salty. The Kelimutu lakes are clearly of the second type.
The three lakes are probably fed by volcanic gasses from the same source. The difference is in how the gas gets to the lakes. Tiwu Ata Mbupu has cool fumaroles with low output. Such fumaroles can lose HF and HCl due to interaction with rock, leaving an input into the lake which is rich mainly in H2S (oxidized to SO4) and CO2. Tiwu Nua, in contrast, is fed by hot fumaroles with a high volume, which inject HF and HCl, in addition to the H2S. The volume is so high that part of the sulfur precipitates out of the water, leaving the lake water a bit lower in S than would be expected from the abundances of the other constituents. Tiwu Ata Polo is intermediate. Not all elements remain in the water. The sediment below all the lakes have high levels of copper and vanadium, which are found in the water only as trace elements.
So what causes the colour changes? It is clearly related to the pH of the water. Tiwu Ata Mbupu changed from green in 1930’s, and white in the 1970’s, to black in recent years, as the pH increased. The other two lakes are usually green to turquoise; Tiwu Ata Polo can also be red but Tiwu Nua (the most acidic) never is. Some changes are seasonal, and probably related to the amount of oxygen in the water. The water temperature could play a role: when Tiwu Nua was heated to over 60C in the 1930’s, the water went white.
The colours are largely due to solid particles (precipitates) in the water which reflect certain colours, but absorption of some colours by molecules in the water also plays a role. In other places, colour is often due to life. Not so in Kelimutu where conditions are so extreme that even algae are unknown. Here the colour seems largely due to the chemistry of the water.
Let’s visualize how it works. Light enters the water from above. Some light is absorbed in the water, and some is scattered by particles and molecules and goes off in different directions; a fraction of this makes it back to the surface and reaches our eyes. The lakes are deep enough that reflections from the substrate at the bottom can be ignored. The perceived colour depends on which colours survive the turn-around best. If a certain colour is efficiently scattered, it will be present in the light reflected back to us. If it is efficiently absorbed, it will be missing. If it is not easily scattered, it will travel deeper into the water and therefore suffer more absorption: such a colour will underrepresented in the reflected light.
For instance, sea water scatters blue light well, red light less so. This makes the sea look blue-ish. Put red algae in the water, and the blue and green light are absorbed by the algae, leaving only the red light to come back to you: the sea turns red – often dark-red as there isn’t much red light to begin with.
An example of scattering versus absorption comes from opal. Against a white background, it seems yellow, but against a black background it is blue. That is because blue light is scattered inside the opal, and comes out in other places where there is no background light. The light that pass straight through has lost some blue, and thus appears yellow. The iris of the eye provides another example. Your eye may look blue, but there is no blue pigment in it. Instead the person looking in your eyes sees the colour of the scattered light. Add some colour (melanin) to the iris, and it becomes darker.
Particle size also plays a role. Very small particles (less than a micron in size) scatter blue light much better than red – if they are present, water can go a vivid blue. Larger particles scatter all colours equally well. This is the case for water droplets in clouds, and is the reason why clouds are white.
The colour of the reflected light depends not only on what happens in the water. It also depends on what you put in, i.e. the illumination. A cloud in the shade of another cloud goes grey, which can be described as white but not bright. If the sea reflects a cloudy sky, it lacks the blue light to begin with, and so the sea turns darker. You can’t easily see blue water in the absence of direct sunlight. This affects not only the Cote d’Azur, but Kelimutu as well. Don’t expect bright blue colours if there is no sun. It is one reason why sunrise is not always the best time to see the lakes.
But in the presence of sunlight, with all the right conditions, which molecules and particulates cause the magic colours of the lakes? Why do the lakes show different colours even though the illumination is the same?
The yellow froth seen on Tiwu Nua is easiest to explain: it is pure sulfur, brought up by the intensive hydrothermal fumaroles. Its staggering concentration of SO4 already points at the oversupply of this element: the sulfur input is estimated at 85 tonnes per day. Tiwu Nua is thus very similar to the sulfur lake of Kawa Ijen, and other hyper-acidic lakes. The sulfur mats can form through a gas reaction involving SO2 and H2O, forming HSO4 and S. The sulfur precipitates out because of the sheer amount in the water of Tiwu Nua. The mats reflect strongest in the red and green, and the two colours combine to give their yellow colour.
The red colour in the water is due to precipitating hematite, or Fe2O3. It forms when oxygen levels are high, and disappears again when oxygen drops. The change between red and green water correlates with the rainy season: the rains oxygenate the water while it lasts. Put too much oxygen in, and the water turns brown or black. Hematite does not form at very low pH, below 0.5, and thus Tiwu Nua, with the most extreme pH, is never red.
The other colours are harder to pin down. Green has been attributed to iron (Fe2+ to be precise) which would explain that change from green to red and back as due to rust – iron reacting with oxygen to form hematite. During the change often the water turns yellow, which is in fact a combination of red and green. However, no water measurements have been taken during these changes so this is not confirmed. Sulfur is less likely, as it forms yellow mats but does not make the water itself yellow.
Blue is the hardest colour. It has been attributed to hydrated metaloxides, involving aluminium or iron. Many types of micro-particles can cause blue colour. In the famous example of the Rio Celeste, it is aluminosilicate. Hydrated copper-sulfate may be a possibility. All three lakes can show a blue/turquoise colour but the precise origin is not yet clear.
The colour white is often seen when levels of volcanic activity are high. It may be that an intrinsically bright sediment is stirred up by overactive fumaroles, such as salt or gypsum.
Colours of change
The figure above shows the colour changes over the years, derived from Landsat imaging. It is taken from the recent paper by Murphy et al. (see the bottom of the post). The bottom panel shows the temperature changes over the 30 years. Tiwa Nua shows temperature fluctuation with a spike around 1997 when there appears to have been a spike in its volcanic activity. The top panel shows the colour as perceived by our eyes. During the temperature spike, Tiwa Nua turned white.
The second panel, called hue stretch, shows what is the dominant colour in the reflected light, leaving everything else out. Grey with a slightly reddish tint would show here as bright red. It is the difference between the colour depicted on the paint tin, and the colour it becomes on the wall.
The third panel is the saturation. It shows how dominant the colour of the previous panel is. A low value means that a lot of the light is from other colours. A high value means the majority is the one shown in the hue stretch.
The fourth panel is the ‘value’, a term somewhat lacking in descriptive value. It gives the strength of the reflected light, where ‘1’ is white and ‘0’ is black.
Let’s look at Tiwu Ata Polo (TAP in the figure). From 1997 to 2009, the dominant colour was green, yellow or red. But the saturation was low, so these colours didn’t stand out so well, and the value was low so the lake appeared quite dark. The red would have appeared as brown: you would have needed very bright light to pick out the colour with your eye. (In the tropics, on a clear day the light can be fantastically bright. At times it felt you could get sunburn from moonlight. The person who called Africa ‘the dark continent’ clearly had never been there.) After 2009, both the saturation and value went up, and the colour changed to blue and green. Now the lake appeared colourful to the eye.
At Tiwa Nua, the 1997 event made the water blue/green, with low saturation but high value. It appeared white to the tourist’ eye. The reflectivity (value) of this lake is consistently higher than that of the other two lakes. Tiwu Ata Mbupu has the lowest reflectivity (value) and therefore may appear black.
The data shows that the saturation has two main states: it is either around 0.2, or around 0.6 ( Tiwu Ata Mbupu has been slowly moving from the low to the high value). The dfference coincides with a change from green/yellow to blue. The most efficient scatterer in the lakes is blue, either because it is intrinsically that colour, or because it consists of very small particles. But only Tiwu Ata Polo shows a good correlation between ‘value’ and saturation. For the other two lakes, the two parameters seem to be determined by different water components.
Rain makes a difference. Both reflectivity and saturation increase towards the end of the dry season. The combination gives the strongest colours.
The miracle of the three lakes will not last forever. Nothing does around an active volcano. A major eruption would destroy everything; minor explosions could do severe damage. Such explosions happened in 1938 and 1965 in Tiwa Nua. The wall between Tiwu Nua and Tiwu Ata Polo is currently 35 meters above the water level. It used to be much higher. Kemmerling, in 1929, stated it was 70 meters above the lakes, and he was told that 70 years earlier, the wall had been as high as the crater wall. It is crumbling under volcanic attack. At times, the two lakes may already be intermingling. Eventually, they will become one lake and the amazing contrast will no longer exist. It could happen quite soon, and is likely to happen within the next 50 years.
But new eruptions can also create new things. Who knows what new miracles may appear in the decades after the next eruption. The Earth is an amazing place. We should always expect the unexpected, and look down in wonder.
A brief introduction can be found on NASA’s Image of the day: Image of the day, based on the new paper by Murphy et al.
This paper by Sean Murphy et al. is good starting point for research into Kelimutu:
Color and temperature of the crater lakes at Kelimutu volcano through time, published in Bull Volcanol (2018) 80:2 (https://www.higp.hawaii.edu/~wright/bv80.pdf)
The classic work on the water chemistry is the paper by Pasternack and Varekamp, The geochemistry of the Keli Mutu crater lakes, Flores, Indonesia, published in Geochem J (1994) 28:243–262. (https://www.terrapub.co.jp/journals/GJ/pdf/2803/28030243.PDF)
For homo floresiensis, a good starting point is Baab, K.L. (2012) Homo floresiensis: Making Sense of the Small-Bodied Hominin Fossils from Flores. Nature Education Knowledge 3(9):4 https://www.nature.com/scitable/knowledge/library/homo-floresiensis-making-sense-of-the-small-91387735
On light colouring by scattering, try The colour of the sky by Dietrich Zawischa
Albert Zijlstra, July 2018