Exchange Report by Felipe Navarrete

Project affiliation: A1
Supervisor: Prof. Dr. Frank Bertoldi
Hosting institution: NRAO, USA, Chris Carilli
Duration of stay abroad: 22.11.2012 - 22.02.2013


My doctoral project focused on a submillimeter extragalactic Survey at 870 um of the COSMOS Field.

These observations cover the largest area for this Field at this wavelength, i.e. ~0.7 square degrees, to a variable depth, which is deepest at the center of the Field, i.e., 1.6 mJy/beam. A catalog of 39 sources has been produced from this Survey, with a subsample of the sources followed up with the Plateau de Bure interferometer (Smolcic et al. 2012).

The last part of the project consisted on studying the average properties of color selected galaxies, which are detected at other wavelengths, but not observed at 870 um. This was achieved with the stacking technique, which has been tested and optimized for the COSMOS Field. As a result of this project we have studied average properties of BzK galaxies and also galaxies selected according to its rest frame (NUV-r+) color.

This is the only way to study these galaxies nowadays, given that average detections of these galaxies require observations down to a 0.1 mJy level at 870 um, which makes difficult, even with ALMA, to study a large sample of these galaxies. Moreover, we study the average properties of these galaxies at 1.4 GHz, and plan to do also at 3 GHz with the outcome of the project which was initialized in this exchange program.


We observed the COSMOS Field with the Jansky Very Large Array (JVLA) at 3 GHz. The area of the Survey is ~2 square degrees aiming to a depth of 2 uJy, which makes this survey the deepest to date for such a large area.

Given that the JVLA is a powerful updated version of the VLA interferometer, which produces an incredible amount of data (~2TB only for the first part of this project) the reduction and imaging of the data presented new challenges to face. I was trained to reduce and image this radio interferometric data.

The observations at the 3 GHz band are new for the JVLA telescope, for this reason technical issues as the flagging of the data had to be thoroughly studied. For example,there are spectral windows in this wavelength band, which are highly contaminated by satellite signal. Also the large volume of data was not easy to handle, hence we had to parallelize the process using the powerful cluster in NRAO to produce a first version of the entire map, which is composed of 64 pointings.


The major benefit for me was learning the radio interferometric technique working with state-of-the-art data, which is of fundamental importance for my work as it heavily relies on radio interferometric data. Certainly, besides me, the data will benefit the A1 project in many areas, as two main science cases where this dataset will make major contributions to the astronomical community will be

i) thoroughly test the drivers of star formation as a function of cosmic time and environment.
ii) explore dust obscuration at high redshift (z > 2), derive the dust unbiased star formation history out to z ~ 4, and probe extreme starbursts out to z ~ 7.

Ultimately, the JVLA observations of the COSMOS Field at 3 GHz will contribute to pave the way, as a pathfinder, for the planned SKA radio interferometer.

I did not feel anything going wrong with this exchange. Clearly, there were some difficulties when reducing and analyzing the data, but that is inherent to any novel project like this. For this same reason, certainly, having the experts in the field just next door it was a great advantage for the project, which otherwise would have had a slower progress.

Current position

Since 2015 Felipe works as a data scientist at in Amsterdam.