The Pink and White Terraces: still lost?

Reports of the 'rediscovery' of the Pink and White Terraces may be premature, writes Bill Keir.

On 2 February 2011 a post on the blog of GNS Science's outreach educator Julian Thomson announced

"Pink Terraces found! Yes - the unbelievable news is that in spite of being located at the centre of New Zealand's most violent eruption in historic times, shaken by volcanic earthquakes, covered by many metres of mud and ash, and then flooded underneath a large lake, a large area of New Zealand's iconic Pink Terraces of Rotomahana has been rediscovered!"1

Newspapers and television gave extensive coverage, and blogs posted in March 2012 gave updates with higher resolution data. The project leader, Dr Cornel de Ronde, promoted the finds with public lectures, television interviews, and YouTube video presentations. In one video he stated categorically, "There is no doubt about it … [the Pink Terraces] were not modified by the eruption."2

These conclusions came from imagery obtained during the Lake Rotomahana Survey Project using sophisticated underwater sonar instruments, GPS systems and cameras capable of yielding very accurate location data and images of physical features under water. The project's main aim was to improve knowledge of the local geothermal fields. Any evidence of the Pink and White Terraces would be coincidental spinoff, but of great public interest.

The published images appear to show scallop-shaped stepped features of solid material similar to the famous silica terraces, which were either buried or destroyed during the eruption of the Tarawera-Rotomahana rift on 10 June 1886. However, these terrace-like objects were discovered at depths of 50-60 m in the lake. A simple calculation of the relative lake levels demonstrates that the objects found by the GNS team are almost certainly not in-situ remains of the terraces - they are too deep in the lake.

I based my calculation on the well-recorded fact that before the eruption the outlet of Lake Rotomahana, the Kaiwaka Stream, flowed into Lake Tarawera. Ferdinand von Hochstetter's 1859 map of the area labelled this stream with the words, "Disembogues into Lake Tarawera."3 Therefore, before the eruption, Lake Rotomahana must have been at least one metre higher than Lake Tarawera to give sufficient fall for the water to flow the 1.5 km between the two lakes.

We also know that, before the eruption, the Pink and White Terraces descended to the shore level of Lake Rotomahana - many pre-eruption photos verify this. The precise vertical height of the terraces above the lake was never recorded prior to the eruption. The most commonly quoted approximate figure for the White Terraces is 30 m. The White Terraces were probably slightly higher than the Pink Terraces.

We also know that outlet blockage and other factors caused Lake Tarawera to rise after the eruption. Ronald F Keam, who published a book on the eruption in 1988, estimated this rise as no more than three metres.4 A visual comparison of two well-known photos of Lake Tarawera's Wairoa inlet and surrounding peninsulas - one taken before the eruption by George D Valentine, the other taken after the eruption by Charles Spencer - confirms that Keam's estimate is about right.5 Lake Tarawera has maintained about this level until today, with one dramatic fall of two or three metres in 1904 when the outlet blockage burst, causing a flash flood that affected communities downstream.

There are no accurate data for the true elevations of the lakes before the eruption, but today's mean elevations are found on the latest topographical maps. My calculation starts with these known values.

Today, Lake Tarawera is 299 m above mean sea level (asl), and Lake Rotomahana is 337 m asl. Therefore, today, Lake Rotomahana is about 38 metres above Lake Tarawera.

Taking today's level of Lake Tarawera to be three metres above its pre-eruption level gives a nominal pre-eruption height of 296 m asl for that lake. This demands at least 297 m above mean sea level for the pre-eruption height of Lake Rotomahana because its water flowed into Lake Tarawera.

It follows from these facts, and one confident estimate, that, (a) Lake Rotomahana could not have risen more than 40 metres from its pre-eruption level to today's level, (b) The base of the Terraces, if still in situ, could not be more than 40 m below the surface of today's Lake Rotomahana, and the top of the terraces may be as little as 10 m below the surface if they are still in situ. A shallower depth than this would be yielded if the original height of Lake Rotomahana were higher above Lake Tarawera than the one metre I have assumed. A shallower depth would also result if the pre-eruption level of Lake Tarawera were higher than I have assumed. A variation of one or two metres in my calculation is possible from the known small variations in mean sea level and the seasonal levels of the lakes.

In-situ Pink and White Terraces at 60 m depth in today's Lake Rotomahana would require the pre-eruption level of Lake Tarawera to be about 20 m lower than it is today to enable Rotomahana's water to flow into it. Given the photographic evidence, this is not credible.

These calculations rule out the possibility that GNS scientists have discovered in-situ remains of the Terraces at a depth of 50-60 m in Lake Rotomahana.

So, what have the scientists found at these depths? There are several possibilities. They could be lower sections of the Terraces that were below lake level prior to the eruption and never seen by humans. They could be remnants of other silica terraces that formed before earlier eruptions in prehistoric times and were buried in the lake long before humans arrived. They could be fragments of the Terraces displaced into the crater by the explosion. They could be other step-shaped objects exposed in the crater by the explosion. Or they could be something else entirely. The evidence presented by the scientist is not sufficient to settle this question. The published images are far from convincing.6 They are open to a variety of interpretations.

Efforts to determine the fate of the Pink and White Terraces began immediately after the eruption. At that time the site of the original Lake Rotomahana was a huge explosion crater more than 100 m deep and more than a kilometre wide, with a small hot lake at the bottom, and surrounded by walls of ejected mud. It took about 10 years for this hole to fill with water to the level of the lake we know today. Investigators had plenty of time to find the Terraces if they were in the slightest discernible in the crater, and there were plenty of investigations. Scientists, surveyors, photographers and newspaper reporters were on site assessing and recording within a few days of the eruption. The landscape was dramatically changed, but they knew roughly where to look for the Terraces. Over ensuing months some daring investigators ventured into the crater to look for evidence of them. In July 1886 guide Alfred Warbrick, accompanied by Auckland Evening Star reporter Jim Philp, climbed down a rope. Warbrick thought he identified a portion of the Pink Terraces under a series of mud banks, but there was nothing conclusive.

There was no doubt that the Terraces had vanished. The debate was always about whether they had been completely blown to pieces, partly damaged, or just buried in mud, and it was never resolved. The recent high-tech findings have not added much to this debate, telling us little more than we learnt from the first investigators, who had the advantage of not having to search under water. Certainly the arithmetic of the lake levels is strong evidence that the GNS scientists are mistaken about what they have found.

On 20 September 2012 The Daily Post in Rotorua published an article by me presenting this case,7 followed by an invited response from Dr Cornel de Ronde the following day. Dr de Ronde was scornful of my arithmetical argument, insisting that his high-tech data is superior and conclusive that the Pink Terraces are unmodified at 50-60 m deep in the lake. The Daily Post published my letter of defence on 2 October 2012.

The GNS blogs and YouTube links are still on the websites in their original form, showing how easily popular fallacies can become entrenched in the public record - in this case, because scientists rushed to the mass media with hyperbole and haste unbecoming of their profession, and failed to do some basic preliminary research and simple arithmetic.

References and notes

1. Julian's blog:juliansrockandiceblog.blogspot.co.nz/2011/02/final.html
2. YouTube video by Cornel de Ronde: How We Found the Pink Terraces. Julian's blog February 2011.
3. www.aucklandcitylibraries.com/ DigitalLibrary/ resourcepages/ heritageimagesonline.aspx(search: maps only advanced search: NZ Map 5694d, 1859). Caution: Hochstetter's map is more of a sketch map than an accurate survey map. Compare Keam's sketch map (derived from photos) posted on Julian's blog 10 June 2011.
4. Keam, RF 1988: Tarawera: The volcanic eruption of 10 June 1886, page 400. Published by the author.
5. Keam 1988: ibid. pp. 31, 192. 6. For example: img.scoop.co.nz/stories/images/1106/26f24baacdbdb8cb98ab.jpeg Note some unexplained anomalies in this GNS image and caption: The convex edges of the scalloped objects in the image face south, but the original White Terraces faced west or northwest. There was probably a small tongue of the White Terraces spreading somewhat more southerly, but certainly not anywhere near 180 m wide. (Refer to Hochstetter's 1859 map and several pre-eruption photos of the White Terraces available online that clearly indicate the Terraces' orientation in relation to surrounding landscape features such as Mount Tarawera and the distant southern skyline.)
6. www.nzherald.co.nz/rotorua-daily-post/opinion/news/article.cfm?c_id=1503435&objectid=11075825

Bill Keir is a freelance journalist and member of the Auckland Astronomical Society.

Cornel de Ronde's response in the Rotorua Daily Post ('Fault found in Terraces scepticism') can be read at www.nzherald.co.nz/rotorua-daily-post/opinion/news/article.cfm?c_id=1503435&objectid=11075824