Managing/Repatriating Local Biological Records of Multiple Taxa in the Ocean Biogeographic
Information System (OBIS)
Abstract: In this
exercise you'll go back to the OBIS with a view toward obtaining data
for a single area, possibly for many species. This opens up all
sorts of possibilities, especially if you are seeking to develop a local
collection of biological data. The methods in the companion
Exercise 3.6 also
apply here, so they are not repeated.
Preliminary Reading (in
OceanTeacher, unless otherwise indicated):
|1. Open the OBIS website, and
take some time to get acquainted with the materials.
|2. Click on the MAPS tab to
see graphics that give a general overview of the collections.
number of records is over 40 million.
|3. This graphic shows the
number of species, per 5-degree block.
|4. And this graphic shows the
results of a specific "diversity" calculation, named ES50. Because it
is a derived quantity, it only partially parallels the above figures.
example, the northeast Atlantic, which is heavily emphasized in the records
and species counts, appears here to have a distinctly low diversity
|5. In this
exercise, we'll investigate how to manage the data for many species (in this
case all species) in a given small area of interest, i.e. Liberia.
You can scale this exercise down to a selected number of species, and scale
it up to a larger area, but just be realistic about the size of the dataset
and don't overload either your computer or the OBIS server with an
6. Click on SEARCH DATA to see this interface.
[You may see an temporary HELP SCREEN over the map; ignore it and it will go
|7. This is the
author's summary of the the 5 selection tools on the left margin:
- TAXA - Huge hierarchical index to the species records.
You can burrow down to any one (or multiple) species here. If you
do not select any specific taxon, it is a good idea to specify BIOTA to
make sure that everything is searched. It is not certain that the
default setting is to search everything.
- DATASETS - Master list of the datasets that cover the map
area. There is a control to SELECT ALL, and there is an OTHERS
category, but it is not clear that they work as you would expect.
The author suggests that unless you have a good reason to select
specific collections to search, don't choose anything here, not even
- REGION - Several ways to pick the area of interest.
Includes selecting one of a few specially named areas (would benefit by
website), a drawing tool, and a coordinate entry form (Top, Bottom, Left
- DATE & SEASON - Not explored yet; would range in utility from
hugely helpful (for well-represented species), to tantalizing (for
poorly sampled areas or species).
- OCEANOGRAPHY - Selection criteria based on the few
"oceanographic" variables that have been gleaned from the species
accounts, such as temperature, salinity, depths, nutrients, etc.
And above these selectors are:
- UPDATE MAP - Click this to see the map, or to make sure that
it has been synchronized with your REGION settings.
- SHOW RESULTS - Click this to see all the results of a search,
by tabs. Includes the important DOWNLOAD tab for the actual data
records, and the data map.
Learn to navigate the interface
mainly using these 2 controls to move from selection criteria to results.
Be a good data citizen and do not specify huge searches (many species/large
areas), unless you have a real analysis plan in place. The server is
fast, but you do not want to overload it and cause problems for other users.
In general, the author finds that TAXA and REGION are the most
important tools for ordinary use, but you will probably have a need for all
of them after you become an avid OBIS user.
|8. OPTION: You can
open the TAXA selection function and select any number of individual
species. You could, for example, select ANNELIDA to see data for those
worms, etc. etc.
NOTE: The author, a life-long biologist groupie, is
shocked now to see that a beloved phylum of his youth, COELENTERATA, has now
evolved into other forms. Where does the time go!
|9. But if you want to see
everything, then make sure that you select BIOTA only.
|10. Now we can selection
REGION from the menu on the left. You have 3 choices:
- USER-DEFINED/USE DRAWING TOOLS - Draw a rectangle by hand;
not precise enough
- USER-DEFINED/ENTER COORDINATES - Typical box coordinate
entry; preferred method
- PRE-FIXED REGION - Only 5 choices available now; needs to be
much larger collection
Select CLICK HERE TO ENTER COORDINATES.
|11. Now you can enter the
side coordinates for Liberia, enlarged by 1 degree in all directions.
[This is the same thing we did with WOD data in
automatically updates the data selection and the map, so there is nothing
else to do here.
|12. Select UPDATE MAP to see
the map. The author doesn't know what the inner dark rectangle means,
but the outer one seems to be the correct Liberia area of interest, as
|13. Use the magnify/zoom
functions along the upper left margin of the map to move into the area of
interest. It seems to be a gridded product, possibly showing the
number of records per cell area.
|14. Find and select the
LAYERS tool in the upper right corner of the map. Now you can see the
color legend for the map.
Click on the POINTS control.
|15. Now you can see the
individual data points in OBIS. Several interesting things appear:
- A surprising number of fresh water/terrestrial data points
- A nearly continuous thick band of points along the equator.
These may reflect fascination about that line, but a data manager's
first question would be is there a possibility that the formatting of
the data or the conversions from original formats could lead to
erroneous latitudes that default to a value of zero.
- There is no similar feature in the longitude at the prime meridian,
so only the latitude values are puzzling.
|16. Now select SHOW RESULTS
to see much more about the data.
The RECORDS tab shows us that there are
247,657 records in the Liberia area. This would seem to be a great
beginning for a regional/national biological database. Below, we'll
find out how to download these data for local assembly and use.
|17. The TAXA tab shows us the
complete list of all taxa identified for Liberia, and you can click on the
5th column (#records) to sort them. Whatever Acanthocybium
solandri is, nearly 7000 of them have been recorded. Etc.
And the total number of taxa found so far in Liberia is 4242.
|18. Here we've selected the
DATASETS tab to examine where these data have come from. You can see
that 62 of the datasets had data for Liberia, and the FISHBASE collection
has the most records, 66,697.
|19. Now, select the GRAPH
tabs to see the history of the biological identifications in Liberia.
OBIS automatically breaks the historical record into sections, as you see
here. You can see periods of little or no activity, and periods of
Look closely at the Y maximum on the graphs and
you'll see that they are 35, 240, 870, 9825 and 15450, so the number of
observations increases exponentially with time.
|20. To get the actual data,
click on DOWNLOAD to see this settings window.
You can explore the WFS,
WMS and KML options on your own. These are graphics-oriented products.
For our own purposes, we'll get the CSV file, because it contains the
actual data. So select POINTS and click on CSV.
Save the file in the folder DATA > OCEAN > OBIS with the filename biota_liberia_obis_20140217_csv.zip
|22. You can unzip the file
and examine it in Excel to see the field structure.
|23. Here is a summary of the
fields in the database records. There is a record for each
individual specimen encountered, and no aggregation is indicated.
BASIS OF RECORD
|24. If you plan on using
these data in any other software, such as Saga, then you'll need to remove
extraneous commas that the OBIS editors have left in some of the original
Use Excel's FIND/REPLACE ALL function to replace all the commas with nulls.(i.e. nothing), as you see
here You need to do this in these fields:
- SNAME (recent discovery)
If you find any problems with the CSV file, you can always come back here
to clean them up.
Now save the cleaned data back into the same file location and filename.
|25. Convert the table to a
point shape with the module SHAPES-POINTS > CONVERT TABLE TO POINTS, as you
have done in
a Data Table to a Point Shape in Saga.
Here it is displayed as a
map in Saga. Notice that there are some extraneous points to the far
|26. You can open the data
table in Saga, and sort on longitude to find the strange points. They
all come from the same source, Seamounts Online. Apparently OBIS
delivers these points even if they fall outside the REGION box. You
can delete them or ignore them.
speaking you will probably never try to work with all the bio data for a
specific area, no matter how small (unless you are trying out diversity
algorithms). So the above methods are only useful for learning the
basic methods available in OBIS. You will probably want to return to
the TAXA control and limit the desired taxa to a group level or perhaps even
a single taxon. Tha author wanted to show you approximately what you
might expect if you do this. In the panels below he has again gone
after the albacore, Thunnus alalunga, singled out for study in the
There are about a dozen hierarchical levels in the OBIS taxonomy.
Starting at ANIMALIA, the author has followed the taxa levels down to
Thunnus alalunga to see how the numbers of records drop off. Not
all levels are shown here, for simplicity.
|28. Animalia > Chordata:
|29. Animalia > Chordata >
Gnathostomata > Pisces > Actinopterygii > Perciformes: 117054 records
|30. Animalia > Chordata >
Gnathostomata > Pisces > Actinopterygii > Perciformes > Scombridae:
|31. Animalia > Chordata >
Gnathostomata > Pisces > Actinopterygii > Perciformes > Scombridae >
Thunnus alalunga: 340 records
|32. As the author
has stated in other exercises with the World Ocean Database and the World
Ocean Optics Database, these archives are excellent places for you to begin
building (or restoring) regional and national data collections. In
many cases these data were taken in your area without any formal
notification or without any donation of the data to you. Now you have
the opportunity and the tools to repatriate these data, as you see fit.
But in all cases, you should give credit to OBIS and the the data
collection, when known, for access to the records.