Iris 2.0 Features Tour
IRIS 2.0 has added several science capabilities built upon the powerful SED building and modeling already implemented in the previous versions of the tool, and that will make IRIS even more useful for astronomers.
Last Update: 05 Aug 2013 - Moved site location from /science/ to /v2.0_features/
- Interpolation and Smoothing
- Co-plotting different SEDs
- Interoperability with other Virtual Observatory tools
- Red- and Blue-shifting
- Integration of SEDs
This is a guided tour through the new IRIS science capabilities, complemented by examples and screenshots obtained with actual SED data. In the following, I will use the SED of one source, BL Lacertae, which is an AGN located at z=0.07. We can create a new SED in the SED Builder window of IRIS. Then, we can create a first segment by adding the photometric data available through the NED SED service. More segments can be added to the SED as discussed in the SED Builder documentation. The more useful visualization of the SED of this source can be obtained in the Viewer window switching from flux density to flux in the "Units" field (using Hz*Jy vs Hz is a nice choice). In this visualization (Fig.1), the grand pattern followed by the SED becomes clear, with a very well designed first component ranging from radio to near-ultraviolet wavelengths, and a second component that appears to rise from soft X-ray energy to very high energy (this component is much less clear as very few measurements are available for energies higher than X-ray). These two components are thought to be produced from the same population of electrons inhabiting the jet that extends from the AGN. BL Lac is the prototype of an entire class of AGNs, the blazars, that are characterized by a very small angle between the jet axis and the line of sight (in other words, the jets are pointing towards us), rapid variability, radio loudness, and almost featureless optical spectra.
Many data points of the BL Lac SED that we have constructed have been taken at different times, so their scatter may reflect an intrinsic shape of the SED or simply be due to the variability of BL Lac emission. While the most obvious approach would be to extract a subset of data points that have been observed in a small time window, we will try to follow a different approach, which is based on the new interpolation capabilities that have been implemented in IRIS 2.0.
In the main IRIS window, clicking on the "Shift, Interpolate, Integrate" icon will open the Science window (Fig2). The left column shows the list of SEDs available, while on the right the first tab, named "Redshift and Interpolation", allows users to move the SED of a source in redshift (upper panel), or produce new SEDs by interpolating the SED selected in the list on the left, using different techniques (lower panel). The second tab, called "Calculate Flux", allows users to evaluate integrated photometric quantities (see later). After checking that our BL Lac SED is selected in the left, we can choose the method for the interpolation, the interval limits, and the units of the final interpolated SEDs. In this case, the default values of the parameters ("Linear Spline" as the interpolation method, "Angstrom" as Units, and 1000 as the number of bins in the final interpolated SED) are good. In order to produce a smoothed version of the interpolated SED that would minimize the effect of the small scale scatter, we can also tick the "Smooth" box and choose the size of the box (expressed as a number of data points). Then, clicking on "Create New SED" will create the smoothed and interpolated version of the BL Lac SED. The new interpolated SED will appear in the left column in the "Science" window (and will also appear in the "Open SED" frame in the SED Builder window).
In coplot mode, we can still access the metadata of each photometric point by right-clicking on it, and the Metadata Browser (that can be access by clicking on the "Metadata" button in the main Viewer window) can be used to explore the distribution of data and metadata for all photometric points, extract subsets of points based on their metadata. In this case, since we are visualizing two different SEDs, the metadata and data of both SEDs are listed in the same window, but data belonging to different sources can be selected easily using a boolean filter based on one of the metadata, which is added to the original metadata of each photometric point when the Metadata Browser is opened in coplot mode. The second column of the metadata in the "Point metadata" tab of the metadata browser, named "b:SED", contains the name of the SED for each point, and can be used to extract subsets of points belonging to one SED using a simple boolean filter. For example, if the original SED is simply named "BL Lac" and the interpolated SED is named "BL Lac_interpol" (SED names can always be modified in the SED Builder window by editing the SED ID field and clicking on "Change"), the photometric points belonging to the original BL Lac can be selected by entering the boolean filter: "b: == 'BL Lac'" in the "Type boolean expression" field of the Metadata tab of the Metadata Browser, and clicking on "Select points". The selected points (belonging to the BL Lac SED) will be highlighted in the window (Fig.4).
boolean filters are powerful tools that can be used to extract subsets of data points of the SEDs based on their metadata values. The Metadata Browser window also allows users to create a new SED from the subset selected (with the "Extract" button), or simply highlight the points selected in the current Viewer window (with the "View selected" button). In IRIS 2.0 a new capability has been implemented that allows the subsets of points to be sent to other tools outside IRIS for further analysis. To showcase this capability, we can exit the coplot mode of the Viewer simply by selecting the original BL Lac SED in the "Open SEDs" panel on the left of the "SED builder" window. At this point, the Viewer will show only the BL Lac SED. After opening the Metadata Browser window, let's say that we want to select all photometric points that have error estimates, and, for some reason, have flux values larger that a given threshold (measured in ergs/cm2/s). We can achieve this using the boolean filter: "d: == 'uncertainty' and m: > 1E-10" where d: and m: are the IDs of the "DataSignificance" and "DataFluxPublishedValue" metadata columns, which contain the information whether an error has been published with the measured value of the flux and the actual observed flux for each data point. At this point, after visually checking if the filter has selected the right subset of photometric points (for example, by clicking on "View Selected" and looking at our SED in the Viewer window), we might want to investigate further the distribution of other metadata and data for this subset of points (Fig. 5).
IRIS is not the right tool to explore tabular data and we would use another application instead, for example Topcat. If Topcat is open (see TOPCAT documentation), you should be able to send the table containing all metadata of the selected subset of points to Topcat by simply clicking on the "Broadcast" button in the Single Point metadata browser tab (just make sure that IRIS is registered to the SAMP hub by checking whether IRIS icon appears in the lower-right margin of the main Topcat window). At this point, a table of the metadata of the selected points will appear in the Topcat file list window, where users will be able to produce scatterplots, histograms, and harness all of the powerful data manipulation and data discovery features available in Topcat (Fig.6). While we have described a simple case where the boolean filter has been applied to a set of metadata columns, the same procedure can also be replicated in the "Data" tab of the metadata browser. We can also send metadata and data for a whole set of points of a given SED if no boolean filter is defined and all rows are selected (for example, by clicking on the "Select from plot" button).
Let's go back to the description of the new science capabilities available in IRIS 2.0. Let's say that now we would like to understand how the SED of the BL Lac blazar would look if BL Lac was located at a much larger redshift than the actual redshift (z=0.07), for example z=0.3. The values of the spectral coordinates and fluxes for the observed data points of a SED at a redshift different from the observed one can be calculated in the "Science" window. The upper panel of the right side of the "Redshift and Interpolation" tab of the "Science" window shows two empty fields named "Initial redshift" and "Move to redshift", respectively. In our case, we will enter 0.07 in the "Initial redshift" field and 0.3 in the "Move to redshift" field (Fig.7).
A click on "Create new SED" will produce a new item in the list of SEDs on the left side of the "Science" window, containing a new SED obtained by redshifting all photometric points of the BL Lac SED to the new redshift, z=0.3. The resulting change in the shape of the SED can be inspected qualitatively by coplotting the original BL Lac SED and the redshifted BL Lac SED at z=0.3 using the technique described above. We might also be interested in how the SED of this source would look like if it was in the vicinity of the Milky Way, say at redshift z=0. We can reconstruct the "local" version of the BL Lac SED just setting to 0 the "Move to redshift" field and clicking again on "Create new SED" (bear in mind that setting either one of the initial or final values of the redshift to a negative number will trigger an "Invalid redshift values" error message, since we are only taking into account the effect of the cosmological redshift in this case). So, now that we have three different versions of the SED of BL Lac at redshifts z=0, z=0.07 (observed), and z=0.3, coplotting all of them can again be useful to compare the evolution of the shape as a function of different redshifts.
Another useful science capability that has been added to IRIS 2.0 is the ability to evaluate integrated fluxes from the SEDs in spectral intervals that can be either defined by the users (in multiple ways) or selected from a depository of existing photometric systems. We can showcase the highlights of this new capability using again the interpolated and smoothed SED of BL Lac that we have produced previously. We could have also used the original observed SED but, in that case, we should have made sure that at least a few data points are available in the the spectral interval where we want the integrated flux to be calculated for the calculation to succeed. We don't need to care about this point if we use the interpolated SED (provided that number of bins is large enough to densely sample the SED).
In the "Science" window, let's select the "Calculate Flux" tab. We might want to measure the integrated flux of the BL Lac emission in a simple spectral interval. In this case we want to calculate the observed flux emitted by BL Lac in the [0.2, 5] keV energy interval, which spans the soft and hard X-ray regions of the SED. It is worth stressing that we could have used frequency (in Hz) or wavelength (in Angstrom) to specify the same spectral interval. Since we are defining the spectral interval, let us select the "Passband" option and edit the two fields with the lower and upper limits of the spectral interval, 0.2 and 5 keV respectively. In the drop-down menu on the right, we have to select keV as unit of measure. Now, by clicking on "Add", the passband we have defined will be automatically added to the "Results" list (in this case, it will be the first), together with the calculated flux value and the effective wavelength measured in Angstrom. We can repeat the operation with any different spectral interval, for example [1E16, 2E17] Hz, and the resulting integrated flux will appear in the "Results" panel after clicking on "Add" (Fig.8).
At this point, let's say that we want to evaluate the integrated fluxes from the BL Lac source that would be collected in a real photometric system used at a given observatory/facility. For example, we could ask about the observed fluxes in the three MIPS filters (160 mu, 24 mu and 70 mu) on board of Spitzer. This can be done very easily selecting the "Photometry filter" option in the "Calculate flux" tab of the "Science" window and clicking on "Choose". A new window, named "Photometry Filter Selector" (Fig.9), will open showing on the left a long list of photometric systems provided by the Spanish Virtual Observatory arranged hierarchically (the same interface is also used by the SED builder to allow users to import photometry points observed in any of the filters belonging to one of these photometric systems). Now, we can browse the list or just highlight the Spitzer photometric filters by typing "Spitzer" in the "By String" search field on top of the right side of the tab, and clicking on "Search". On the left, after the search only the Spitzer folder is shown, and a click on the gray triangle on the left of the Spitzer folder reveals the available filters. We can select the three MIPS filters at once by clicking on the three items contained in the Spitzer folder and keeping the Command button pressed (on Mac systems). The basic information about the filters selected will be shown on the right. Clicking on "Done" will automatically close the "Photometry Filter Selector" window and calculate the integrated fluxes associated to the Spitzer filters selected. These filters will be added to the list of spectral intervals previously defined (the list is arranged in order of decreasing wavelength). All the flux values calculated can be saved to a ASCII file by clicking on "Save": a simple interface will allow users to pick the location, name of the file, and the units of measures used to express the effective wavelengths of the spectral intervals/photometric systems and flux measures respectively.
|02 Jun 2013||First issue|
|05 Aug 2013||Moved site location from /science/ to /v2.0_features/|