Science Data Challenge 3

Overview


Purpose

As with our previous two data challenges (SDC1 and SDC2), our goal is to prepare the radio-astronomical community for the novel nature of the data expected from the Square Kilometre Array. Given the order-of-magnitude improvement in sensitivity, new analysis methods are required for both the challenging nature of resulting data, but also for the previously untouched science. Thus, realistic, synthetic datasets emulating the telescope's capabilities will be disseminated to the community to test the suitability of existing methods and foster the development of new ones on these next-generation, scientific datasets. Ultimately, results of each of the competing teams' approaches will be compared via a standard figure-of-merit, instigating a competitive nature to our challenges.

Epoch of Reionisation

Given our current understanding, cosmological history prior to the current state of the universe can be divided in to several distinct epochs:

  1. Cosmic Dark Ages (1100 < z < 30)

  2. Cosmic Dawn (30 < z < 15)

  3. Epoch of Reionisation (15 < z < 6)

After the Big Bang, our universe expanded and cooled to the point at which ionised hydrogen recombined to its atomic state, with 21cm emission subsequently being observed in absorption against the background. With the appearance of a significant population of galaxies, the first stars formed and began to heat their surroundings. This heating shifted the absorption profile of the 21 cm Hydrogen line into emission. However, as this heating continued, more of the surrounding gas became ionised to the point at which this emission ceased. Having only roughly constrained the time periods at which these important events occurred, a primary science goal of the SKA is to constrain that timeline, which is where Science Data Challenge 3 comes in.

Credit: Robertson et al. (2010)

Participants of this data challenge will be tasked with elucidating exactly when (given a realistic, artificial dataset) the Epoch of Reionisation occurred. However, given that the observation itself is extremely challenging, our challenge will be broken down in to two parts:

  1. Removal of foreground emission from Galactic and Extragalactic sources ('Foregrounds', or SDC3a)

  2. Inference of important parameters of the Epoch of Reionisation ('Inference', or SDC3b)

Timeline

SDC3 will run for a period of approximately one year divided into two, independent challenges, 'Foregrounds' and 'Inference'. Both of these challenges will run for 6 months each with a short pause in-between. It is currently anticipated that the first part of SDC3, 'Foregrounds', will run from January 2023 - June 2023.