Monday, August 31, 2009

Stratigraphic concepts and agetagging

Stratigraphic 'homonyms' are quite common, a result of ambiguous placenames which have been used to name stratigraphic units. The stratigraphic terms used in Jens' post on Table Mountain stratigraphy nicely illustrate this:
He wrote about the South African 'Peninsula Formation' which is a synonym for the 'Peninsula Sandstone'. Unfortunately, this 'Peninsula Formation' is homonym to the 'Peninsula Formation' from British Columbia. The latter is of Cretaceous age, whereas the South African was formed during the Ordovician!

Geoscientists using stratigraphic terms surely have a clear concept of what they mean. However, whereas for human readers it may be easy to determine such a concept by lithological descriptions or the broader geological or tectonic context, this is much harder for computer algorithms and thus, agetagging.
Fortunately, the regional context can easily be discovered by geotagging technologies, therefore the geographic concept of a stratigraphic term is most important. Stratigraphic homonyms may exist at much lower regional level, e.g. from different provinces or states and most likely there are also significant conceptual differences between scientists. etc etc. However, such regional concepts are sufficient for the approximation for our agetagging purposes. The concept of a South African 'Peninsula Formation' is surely completey different from the Canadian.

.. a lot of complications with 'strato-semantic' text analyis.. stratigraphic analysis of geoscientific texts requires a combined approach, good agetagging is not possible without geotagging.

Friday, August 28, 2009

Petrified wood from the moon

Did Apollo 11 bring petrified wood from the moon? A treasured piece at the Dutch national museum - a supposed moon rock from the first manned lunar landing - is nothing more than petrified wood, curators say.

It was given to former Prime Minister Willem Drees during a goodwill tour by the three Apollo-11 astronauts shortly after their moon mission in 1969. The US agency gave moon rocks to more than 100 countries following lunar missions in the 1970s.

The Rijksmuseum, which is perhaps better known for paintings by artists such as Rembrandt, says it will keep the piece as a curiosity. "It's a good story, with some questions that are still unanswered," Xandra van Gelder, who oversaw the investigation that proved the piece was a fake, was quoted as saying by the Associated Press news agency.

A jagged fist-size stone with reddish tints, it was mounted and placed above a plaque that said: "With the compliments of the Ambassador of the United States of America . . . to commemorate the visit to The Netherlands of the Apollo-11 astronauts." The plaque does not specify that the rock came from the moon's surface.

It was on show in 2006, and a space expert told the museum it was unlikely NASA would have given away any moon rocks three months after Apollo's return.

But frankly, does it look like lunar basalt? Researchers from Amsterdam's Free University said they could see at a glance that the rock was probably not from the moon. They followed the initial appraisal up with extensive testing.

"It's a nondescript, pretty-much-worthless stone," geologist Frank Beunk concluded in an article published by the museum.

US officials said they had no explanation for the Dutch discovery.

Table Mountain stratigraphy update

During my recent holiday in the Cape I visited my alma mater, the University of Cape Town and its Department of Geosciences. At his visit John Rogers of the geology department told me about his recent work on Table Mountain and later showed me some outcrops along the northern and western face of the mountain.

The stratigraphy of Tale Mountain (Western Cape Province, South Africa) always seemed to be a perfect example of "layered cake" stratigraphy: the base is Malmesbury Shale and Cape Granite with the famous Sea Point Contact between the two. Both are truncated at the top by an unconformity. The succession continues upwards with some basal conglomerate in places, the tidal or fluvial Graafwaater Formation, followed by the fluvial quartzites of the Peninsula Formation and topped by some glacial till, remnants of the Pakhuis Formation. This package of Graafwater Fm. and Peninsula Fm. can be found in the Cape Mountains where it is highly deformed.

The contact between the Graafwater Fm. and the Peninsula Fm. had always been interpreted as a conformable contact. By coincidence John Rogers noticed an angle between the Graafwater Fm. and the overlying Peninsula Fm., prompting the question, whether the contact between these two formations is actually conformable. Dating on zircons by Rogers et al. points to a significant hiatus between the two formations.

The picture above shows Cape Point and Rooikrans at the southern tip of the Cape Peninsula. It is one of the prime tourism and field-trip stops. I had been here a few times before in the past twenty years, but neither me, nor seemingly anybody else, had noticed the prominent syncline/anticline in the Graafwater Fm. beneath an undeformed Peninsula Fm.

My guess is that the folding is due to compression along a north-south axis because all observed folds dip along this axis. The quartzites of the Peninsula Fm. seem to be more competent. Instead of ductile deformation they show brittle deformation, e.g. a prominent reverse fault under Constantiaberg, the fault plain dipping to the south.

This puzzles me and I will have to do more reading on the regional geology of the Western Cape. Why is the principal stress from the south if the nearby Cape Mountains were folded by a main stress from the west? How far did the Graafwater Fm. extend? Is there an angular unconformity at Lekkranskloof (near Heerenlogement and Graafwater)? Interesting new detail about the assemby of Gondwana.

Wednesday, August 19, 2009

Charting taxonomic knowledge through ontologies and ranking algorithms - Post-print at GFZ

For those who are interested in reading our paper on Taxonrank , feel free to download a post-print copy at:

TaxonRank is a ranking algorithm based on bibliometric analysis and Internet page ranking algorithms. TaxonRank uses published synonymy list data stored in TaxonConcept, a taxonomic information system. The basic ranking algorithm has been modified to include a measure of confidence on species identification based on the Open Nomenclature notation used in synonymy list, as well as other synonymy specific criteria...

Wednesday, August 12, 2009

How much money is in the scientific data management 'business'?

The answer to this question probably nobody knows... But, based on some numbers I know, I will give a rough estimate on how much money potentially could be made with scientific data management - in Europe.

The 7th European Framework Programme (fp7) provides more than 50 billion € for research projects between 2007-2013. The structure of fp7 is quite complicated and it is almost impossible to find out how much of this money exactly is spent for bio- and geosciences.
The themes 'Agriculture and Fisheries, Biotechnology' and 'Environment (including Climate Change)' alone are funded with > 3.8 billion €. So let's estimate the European Commission spends around 4 billion € for bio- and geosciences.

SourceMio €
fp7 (total)50000
fp7 (bio/geo themes)~4000
fp7 (potential for DM)~45-50

To my knowledge, those European research projects which provide some money for data management, reserve 0.5-3% of the total project funding for this purpose. The mean percentage is about 1.5%. From the 4 billion I mentioned above, some projects may not need data management at all. However, if at least 75-80% of all research projects produce data, 1.5% of the remaining 3 billion € would include around 45-50 million € for data management. This seems to be much money, but the fp7 started in 2007 and will last until 2013, so the yearly amount of money for data management (DM) potentially(!) spent is around 6.5-7 Mio €. The potential market for scientific data management still seems to be considerable.

Of course this money is not really spent. The percentage of projects which include proper data management is certainly below 75%. I would estimate less than 30% of all projects reserve money for data handling. Further, we need to consider that proponents from 27 member countries compete for this money (as little appropriate data centers exists the competition for the remaining millions is not as hard in reality...).

I used the fp7 example, as the European Commission has shown considerable interest in improving open access to scientific data. As far as I know, this issue is also considered in project proposal evaluations. However, the money the European Commission spends is only one possibility for funding, ideally national funding agency will also support access to data. If you are living in a lucky country, national funding for DM could equal the amount the Commission spends.

In summary, I think the scientific data management niche is still interesting, while there is not as many money you might have expected. However, as the importance of e-science infrastructures and open access to scientific information has only recently been recognized, this sector may still grow in the future.