This Data Portal contains data from Mermaid floating seismometers that have been used for testing the first generation Mermaids. Only one of those (Mermaid 31) is still operational. The data from an experiment with 9 Mermaids near Galapagos (2014-2016) are still under emargo but will be added in the course of 2017.
With the experience obtained we have redesigned the robots to have a larger capacity, and in principle able to carry up to eight different instruments. See the Mermaid page for more detail and a tutorial for analysing Mermaid data.
The logistics of a Mermaid are fundamentally different from those of fixed land stations, and the current SEED format is not fully compatible with it, primarily because the station coordinates change with every recording. Although we make the raw data available in mini-SEED, the format that is probably most easily accessible is the SAC format. We store our best estimate of the station coordinates at the time of recording in the header: latitude STLA, longitude STLO and station depth below sealevel STDP (in m).
For strong events, the Mermaid will immediately come to the surface and take its position with GPS. This may take several minutes, during which it drifts at the surface. Depending on the wind, the surface drift is usually much stronger than the drift at depth. Sukhovich et al. (Nature Geosc., 2015) estimate a maximum drift of 500 m between recording and GPS fix, but usually much less. We determine the best estimate of station location by extrapolating from the first GPS fix taking the surface drift (as measured by subsequent GPS fixes) into account. The location error may be of the order of 100 meter. For a P wave with a slowness of 0.1 s/km this translates into a delay time error of the order of 0.01 s. See also Joubert et al. (2016) .
Access by ftp is here: Mermaid data , or, if your site has problems with ftp access click here
The records are up to four minutes long and generally contain 100 seconds of noise before the trigger (generally a P or PKP arrival, depending on distance). Since the data volumes are still small, for now we bundle all data in one zip file for each Mermaid. This file is automatically kept up to date. Whenever the NEIC PDE listing enables us to link a record with an earthquake, the record is moved from directory 'unidentified' into directory 'identified' and the event is added to *events.txt. This is also done automically but may not catch all events, so be sure to also consult the directory 'unidentified', especially for weaker earthquakes. Some small oceanic earthquakes may never enter in the NEIC catalogue but are recorded by the Mermaids (see for example the swarm of 235 quakes after Nov 25, 2013, recorded by Mermaid 14). Seismogram images in png format can be found in subdirectory images.
Important: Users of these experimental Mermaid data should be aware that this was a development project, in which prototypes were tested under various conditions, errors in design were discovered during tests, and improvements were made continuously. That implies that the data are not (yet) always error-free. Below we describe the difficulties encountered and the way we have handled them.
The largest headache during development was to correct the on board clock for drift using GPS. This required procedures that are robust even in cases when the GPS fix after surfacing is delayed (e.g. by high waves).
File *clk_drift.txt lists the clock drift estimates, obtained by comparison with GPS at each surfacing. The four earliest floats (m17, m18, m12 and m13) still contain bugs discovered during their testing phase. We are pessimistic that on can reconstruct correct timing for these four floats (m17 and m18 have been recovered off the coast of Spain, but m12 and m13 were abandoned in the Indian Ocean). Mermaid 31 (Mediterranean) has shown erratic behaviour, probably due to malfunction of its electronics, and its time must be considered as very unreliable. However, for differential measurements such as PKPdf-bc its data are perfectly useful.
A bug in the Garmin software caused GPS to forget occasionally the value of the leap second (the second that is sometimes added at the end of the year to UTC). Though Garmin has since published a fix, the first generation Mermaids still have the original software and suffer from this. By inspecting files like m14_clk_drift.txt one is usually able to find out when this happened (it causes a large but momentary change in the calculated daily drift).
In the data published on the web we have corrected for clock drift (with the exceptions mentioned above). Please refer to the Mermaid Data User Manual and to the Mermaid Data Format Manual for more detailed information on how to access and use the meta data.For a full description of all issues encountered during development see Float Issues during Development. However, you would normally not need this information to use the data.
File *phases.txt lists the times of start/end of descent/ascent. Finally, a *kml file allows you to view the Mermaid's trajectory with Google Maps.
See also Alexey Sukhovich' page