Periodic Reporting for period 1 - STORM-CHASER (Chasing plasma storms on exoplanets)
Periodo di rendicontazione: 2022-06-01 al 2024-11-30
Are there other worlds out there that can sustain life? A key factor in the ability of a planet to host life is its atmosphere. Stars can erode planetary atmospheres by bombarding the planet with plasma. How well planets can protect their atmospheres depends on the strength of their magnetic fields. We do not know the ultimate fate of exoplanet atmospheres because we have neither detected massive plasma ejections on stars nor measured the magnetic fields of exoplanets. Both of these shortcomings can be solved with sensitive radio observations at low frequencies. Previous radio observations did not simultaneously have the sensitivity and on-sky exposure to detect the faint, transient radio signals from stars and exoplanets. In this project, I will deploy a novel strategy to mine over 20,000 hours of existing data from the Low Frequency Array (LOFAR), the most sensitive telescope at low frequencies. I will achieve two orders of magnitude better sensitivity and/or time-on-sky than previous efforts to accomplish two goals: (1) the first detection of massive plasma ejection events on a star other than the Sun and (2) the first measurement of an exoplanet's magnetic fields. The ERC grant will allow my team to tackle petabytes of data and search for radio signals over trillions of image pixels to achieve the goals of this project. It will lead to a leap in our understanding of the plasma environment of exoplanets around stars of different types and ages and the underlying laws that determine the magnetic field strength of exoplanets.
The project is currently in its "methodology development" phase. This means that we are developing the software and algorithms necessary to process almost 10 paytebytes of data. To do this we have so far (a) produced a dedicated computer cluster (thanks to the infrastructure funds from the grant), (b) developed the software to calibrate and image the radio data and (c) developed the software to search for peculiar signals from nearby stellar systems.
The calibration and imaging software developed by a PhD student working on the project already goes significantly beyond the state-of-the-art because it can make images at frequencies as low as 15 MHz which was not possible before. Data processing is very challenging at such low frequencies because the Earth's own atmospheric layer called the ionosphere severely distorts cosmic signals at these low frequencies but it is at these low frequencies that we expect to find signals from exoplanets.
The software to search for peculiar signals was developed by another PhD student working on the project. It looks for so-called type-II bursts that are generated when plasma is ejected from a star. These plasma ejection events can be devastating for exoplanets but they are rare and occur unpredictably. So we have developed software that can sift through thousands of hours of telescope observational data to locate these smoking-gun signatures.
In the coming months we will publish these software and papers on the methodology behind them. Then begins the work of applying the software to a large fraction of data in the telescope archive to make the discoveries.
In parallel with the postdoc on the project, we are developing techniques to look for signals that are produced from the interactions between the stellar plasma ejection and the planet. We have broadened our search from our original plan to use radio observations to now using both radio and optical observations. This interaction signature will appear as a periodic signal in the data modulated at the orbital period of the exoplanet. We have now also developed a methodology to detect the presence of this periodicity in a statistically robust way.
The calibration and imaging software developed by a PhD student working on the project already goes significantly beyond the state-of-the-art because it can make images at frequencies as low as 15 MHz which was not possible before. Data processing is very challenging at such low frequencies because the Earth's own atmospheric layer called the ionosphere severely distorts cosmic signals at these low frequencies but it is at these low frequencies that we expect to find signals from exoplanets.
The software to search for peculiar signals was developed by another PhD student working on the project. It looks for so-called type-II bursts that are generated when plasma is ejected from a star. These plasma ejection events can be devastating for exoplanets but they are rare and occur unpredictably. So we have developed software that can sift through thousands of hours of telescope observational data to locate these smoking-gun signatures.
In the coming months we will publish these software and papers on the methodology behind them. Then begins the work of applying the software to a large fraction of data in the telescope archive to make the discoveries.
In parallel with the postdoc on the project, we are developing techniques to look for signals that are produced from the interactions between the stellar plasma ejection and the planet. We have broadened our search from our original plan to use radio observations to now using both radio and optical observations. This interaction signature will appear as a periodic signal in the data modulated at the orbital period of the exoplanet. We have now also developed a methodology to detect the presence of this periodicity in a statistically robust way.
Exoplanet detection efforts: The radio data calibration and imaging software developed in the project has broader applicability to any other science case that benefits from observations at very low radio frequencies (below about 40 MHz). As such, it can have a broad impact beyond the project. In particular, the project has made technical progress on two fronts. (a) Low frequency radio observations suffer from the cosmic radio signals being distorted by the Earth's ionosphere. Correcting these distortions is necessary to produce radio images. The project has pushed the state-of-the-art in this front by building a pipeline that can yield reliable sky images at frequencies as low as 15 MHz which was not possible before. (b) another problem that we deal with in calibration is the separation of the effects of the ionosphere and small but significant drifts in the clocks used in the telescope electronics. There existed a way to separate the two effects which was effective at higher frequencies where second order effects of the ionosphere are not important. The project has resulted in a new more robust algorithms that can also account for these higher order effects. This algorithm has been instrumental in obtaining images down to 15 MHz. These two technical developments have a broad impact as they can be used for any science case that needs radio images to be made at frequencies below around 40 MHz.
Further uptake by the broader radio astronomy community requires that the algorithm and code-base we have developed are integrated into standard workflows used by the community and the software made freely available. We are now doing so by integrating our modifications into the standard pipelines that are being widely used in the low-frequency radio astronomical community while also providing our codes freely via GitHub (see for e.g. https://github.com/cristina-low/lofar-low-pipelines(si apre in una nuova finestra))
Stellar plasma ejection detection efforts: Current state-of-the-art to detect stellar plasma ejection events used dedicated observations of one or a few stars covering just tens of hours. However these events are rare which needs thousands of hours of coverage. This means that manual searching for these bursts is no longer possible and we must develop new algorithms to automate the search. The project has done just that. We now have an initial implementation of such an algorithm that is showing promise. We have been able to use this algorithms to detect stellar bursts (not necessary related to coronal mass ejections). We are now deploying the algorithms on over 0.5 millions hours worth of data on nearby stars. Some aspects of the algorithms that reject artifacts in the data such as radio frequency interference and interference from unrelated nearby sources are more broadly applicable to any transient detection in radio interferometric data. This gives the algorithms a broader applicability. It is still early days to know what modifications may be necessary to ensure wider uptake (it depends on the nature of the artifacts we have top reject for our project and how general they are).
Further uptake by the broader radio astronomy community requires that the algorithm and code-base we have developed are integrated into standard workflows used by the community and the software made freely available. We are now doing so by integrating our modifications into the standard pipelines that are being widely used in the low-frequency radio astronomical community while also providing our codes freely via GitHub (see for e.g. https://github.com/cristina-low/lofar-low-pipelines(si apre in una nuova finestra))
Stellar plasma ejection detection efforts: Current state-of-the-art to detect stellar plasma ejection events used dedicated observations of one or a few stars covering just tens of hours. However these events are rare which needs thousands of hours of coverage. This means that manual searching for these bursts is no longer possible and we must develop new algorithms to automate the search. The project has done just that. We now have an initial implementation of such an algorithm that is showing promise. We have been able to use this algorithms to detect stellar bursts (not necessary related to coronal mass ejections). We are now deploying the algorithms on over 0.5 millions hours worth of data on nearby stars. Some aspects of the algorithms that reject artifacts in the data such as radio frequency interference and interference from unrelated nearby sources are more broadly applicable to any transient detection in radio interferometric data. This gives the algorithms a broader applicability. It is still early days to know what modifications may be necessary to ensure wider uptake (it depends on the nature of the artifacts we have top reject for our project and how general they are).