Alexander Sokolov & Fredrik Wulff
Department of Systems Ecology, Stockholm University
DAS (Data Assimilation System) is a computer software that has been developed as a tool to query and analyze hydrographic and chemical data from the Baltic Sea. It has primarily been interfaced to a database, called BED (Baltic Environmental Database) maintained at the Department and accessible both locally and via the Internet, but it can also be used against local databases, generated by the user.
The primary objectives for DAS is to:
Extract data (salt, temperature, nutrients, oxygen & chlorophyll) for specific time period and regions of the Baltic. These data can then be used to
Plot distribution of stations and no of observations (in order to evaluate if the sampling frequency is sufficient for further analyses)
Interpolate. The observations are used to calculate data for each grid cell of a 3D grid with a user-specified resolution. The interpolated data can then be used for:
Plot horizontal surfaces and vertical transects of concentrations for the specified space and time regions. Plots are done as contour plots in color and/or as isopleths and the user can specify scales. These plots are used for descriptive purposes but also for evaluating the quality of data; erroneous data points are usually easily seen in these color plots. The user can go back to the extracted data (stored locally), delete erroneous data and repeat the interpolation-visualization. It is then possible to:
Calculate total amounts (an average concentrations) of the interpolated variables within a specified region, either bounded by space coordinates (latitude, longitude, depth) or by a concentration interval of another variable. The concentrations are multiplied with the corresponding volumes of the hypsographic curve.
Perform various other arithmetic calculations for a single variable or between two interpolated variables. The results can be stored and used in further chained operation, i.e. in various stochiometric calculations (N/P ratios, nitrate anomalies, etc)
The interpolated gridded data can also be extracted and used for other purposes, outside DAS, i.e. for initiation data fields for numerical models.
The original version of DAS is described in detail by Sokolov et al. (1997). Numerous modifications, suggested by users have been added to later versions. We are now at DAS version 4, which is significantly rewritten and available on a CD together with public data for 1970-1990 as well as from the Internet. The modifications are described in the accompanying readme.html file.
There is a multitude of uses for this program but we will here give a simple example, to get you started! When you have installed and started the program (from the CD or from the Internet (see http://data.ecology.su.se/models) you will see the following screen:
The DAS window is initially empty, except the menu bar at the top at this stage. Click on the Database item on the menu and the window will look like this:
The DAS window now contains 4 windows, one displaying a map of the Baltic to the left and three windows to the right, above each other.
The map in the right window has been generated from a hypsographic data with a horizontal resolution of 5 nautical miles. This is the default resolution but by clicking on the File menu, you can choose a file with a 1 nm resolution (by selecting bs1852.rgn instead of bs9260.rgn). You can also change the size of the area you will be working with by specifying the coordinates in the ‘Request area’, which by default is the entire Baltic Sea. You can also, while the cursor in the map-window, click on the right button and select a smaller area (by moving the mouse holding down the right button) as in this example
The arrows in the ‘request area window’ are now highlighted (in red): by clicking on the arrow pointing to the right, you will select coordinates corresponding to the borders of the map-window (and vice-versa by clicking on the other arrow). By selecting the ‘unzoom’ alternative in the map-window, you can return to using the entire Baltic Sea.
If you had selected to use the 1nm resolution, the map window would looks like this:
Use of the higher-resolution hypsographic data will of course generate a more detailed map and is primarily useful for smaller areas that the entire Baltic Sea. All calculations using the fine resolution matrix will take longer time and will not necessarily produce better results, since the interpolations are using the same number of observations that are used in coarser resolved hypsographic data. The interpolation schemes are discussed further below.
Now, it is time to start using DAS and request data. We will return to use the entire Baltic Sea. Click with the mouse on the Account button and a new sub-panel will appear:
Here, you have the option to select which database you will access and query and also to give your username and password. By default you are user ‘guest’ with password ‘guest’ with Internet access to the BED database located at Department of Systems Ecology (we assume that your computer is connected to the Internet). You will now have access to all data except for those from the last five years (additional data can be accessed if permission is granted and new user/passwords are given). You also have the alternative to work with a local database.
For now, use the default settings which also means that you will request all data for August 1985, as shown in the ‘Time period’ Panel. Press the ‘Request’ button and the arrow on the screen representing the mouse point will change into a time-glass with the letters SQL below. This means that you instruction is translated into a database command and sent to the remote computer/database. After a few seconds, you screen will look like this:
The left map-window will show a number of dots in positions where measurements found in the database were observed. The observations, with their ID, position (latitude and longitude), data and time, are listed is the right panel together with the name of this research vessel used. You can use the mouse to double click on a line in this window and the corresponding point in the map will be highlighted in red and the actual measurements will be shown in the lower right panel. In this example a station visited by ‘Hydromet’ on Aug 8 outside the Polish coast is shown. Only salinity, temperature and oxygen were measured at three depths (0, 5 and 12 meters). You can also click on a point on the map and obtain the corresponding information.
You can also use the panel with arrows, just above the data on stations, to move to different records. The right-hand button with a minus sign give you the option to delete a record. You will not delete data in the BED, of course! Extracted from the BED data are stored in the computer memory. Clicking on the panel ‘Local tables’ will give you the option to store and later retrieve and continue to work with these data. The row with titles for the records can be used to sort these: just click on the column-title that you want to use!
The initial map shows positions of all stations that contain data. You may want to see how many stations that contain a specific variable in a specified depth interval. This option, you can select by right clicking on the map and select option from the sub-panels that appear:
If you then select ‘distribution of the stations’’ only those stations are plotted that confirms to those criteria. For instance, there are fewer stations that have nutrient data, compared to those that have only salinity and temperature.
By selection the zoom option, you can select a sub-area in the Baltic and by clicking to the time panes, you can select other time periods, as well. You move the mouse on the year, month and day section in the panels and selectively change time intervals. Check it out!
You have now a large number of options for further use of DAS. We will here only work through one example. This example will hopefully give you ideas on other possible uses. We will here calculate to ‘cod spawning volume’ i.e. the amount of water in the Baltic proper with salinity above 11 PSU and an oxygen concentration higher that 3 ml/l. These are the environmental conditions necessary for a successful survival of fertilized cod eggs.
For this exercise, we have to extract and interpolate both salt and oxygen. We will use our default time period, i.e. August 1994. After the data have been extracted and stored locally, we start interpolation by choosing this command when the mouse in the map field by ‘right-click’:
When you click on salinity, you will get an additional sub-menu:
You can now change the size of the cell, within which all observations are averaged, before interpolation takes place. By default the size is 5 nautical miles = 9260 meters, which is the horizontal grid step used (if you had selected the high resolution hypsographic data, the size would have been 1852 or one nautical mile). You also have the option to exclude data marked as ‘questionable’ from further calculations.
When interpolation is ready (it takes usually a few seconds, with a ‘time glass’ displayed) you can leave the database window by selecting 1 in the top panel under Window:
You then have the option to select a variable to plot and for which depth interval. If you don’t select one the depth distribution is selected.
We have now selected salinity at surface (averaged for 0 – 2.5 meters). The depth intervals are those for which usually measurements are available in Baltic monitoring cruises. Select draw (default) by right-click on the mouse and a map of the interpolated surface concentration of salinity will be drawn:
By default the map is drawn with a color scale corresponding to the minimum and maximum values in the entire interpolated matrix (from north to south, from surface to bottom). You can select another scale and also to plot contour lines by selection in the ‘options’ header menu:
You can also select the size of an output as a postscript page (useful for sending plots to a file and/or a postscript printer). You can also change the projection when you select a 3D drawing (try it out, we will not use these features here).
Notice that in the lower left corner of the window is shown the position of the pointer in latitude and longitudes and the corresponding salinity.
You have many more plotting options here: we will just demonstrate the option of drawing a vertical transect. You first select the option ‘cross section’ in the window that is shown under the mouse, just before plotting:
You first select ‘set new position’ i.e. select the points in the transect (alternative, you have already done so and stored these positions. Then you select the option ‘read’). You then click on the map in the position you want (press Alt before the last point):
The next step is to plot the positions:
And then finally, ‘draw distribution’ in terms of a contour plot. You can select different scales (horizontal and vertical scales and the quality of the plot as well. You can also store the plot in a file. You start the plot by clicking on ‘start’.
By resizing the window and using different X/Y ratios and color scales (again through the ‘options’ menu, you can create different plot of the same data:
Plotting, is not only useful for descriptive purposes, it is also a powerful way to find and delete erroneous data. Color plots will show erroneous data points that can then be deleted from the local data (as described earlier).
To continue, it is time to extract and interpolate oxygen data, in exactly the same manner as those for salinity. As an illustration of the result, here is plotted the bottom oxygen concentrations:
The color scale has been reversed and isolines used to contour concentrations below 0. The only remaining item in our exercise is to calculate our ‘cod spawning volume’. This we can do now since we have the two interpolated matrices of salinity and oxygen. We will also do this only for the southern Baltic proper which is the relevant area of interest. Thus, we change the area by selecting the zoom command and redraw the area:
In the same menu, select ‘set area boundary’ and plot and then store as a file the coordinates you have used. Switch to the salinity data, though creating a new plot from the ‘window’ menu.
Then, select the ‘calculate option’ in the drawing menu:
Select ‘isosurfaces’ between 11-999 PSU for salinity and between 3 and 999 for oxygen and click ‘run’ and the result of 104.6 km3, above a bottom area of 7117.1 km2 is shown as well as the average salinity and oxygen concentrations within this water package.
You can also plot where this water occurs and the thickness of the package by creating a new window and by selecting thickness as the plot variable
That’s it, ladies and gentlemen, we hope that you will enjoy using DAS and soon be comfortable with it.