2008A&A...489..359Y

Massive stars in low-metallicity environments may produce exotic explosions such as long-duration gamma-ray bursts and pair-instability supernovae when they die as core-collapse supernovae (CCSNe). Such events are predicted to be relatively common in the early Universe during the first episodes of star-formation. To understand these distant explosions it is vital to study nearby CCSNe arising in low-metallicity environments to determine if the explosions have different characteristics to those studied locally in high-metallicity galaxies. Many of the nearby supernova searches concentrate their efforts on high star-formation rate galaxies, hence biasing the discoveries to metal rich regimes. Here we determine the feasibility of searching for these CCSNe in metal-poor dwarf galaxies using various survey strategies. We determine oxygen abundances and star-formation rates for all spectroscopically typed star-forming galaxies in the Sloan Digital Sky Survey, Data Release 5, within z=0.04. We then estimate the CCSN rates for sub-samples of galaxies with differing upper-metallicity limits. Using Monte-Carlo simulations we then predict the fraction of these CCSNe that we can expect to detect using different survey strategies. We test survey capabilities using a single 2m telescope, a network of 2m telescopes, and the upcoming all-sky surveys of the Pan-STARRS and LSST systems. Using a single 2m telescope (with a standard CCD camera) search we predict a detection rate of ∼1.3 CCSNe/yr in galaxies with metallicities below 12+log(O/H)<8.2 which are within a volume that will allow detailed follow-up with 4m and 8m telescopes (z=0.04). With a network of seven 2m telescopes we estimate ∼9.3 CCSNe/yr could be found, although this would require more than 1000h of telescope time allocated to the network. Within the same radial distance, a volume-limited search in the future Pan-STARRS PS1 all-sky survey should uncover 12.5 CCSNe/yr in low-metallicity galaxies. Over a period of a few years this would allow a detailed comparison of their properties. We then extend our calculations to determine the total numbers of CCSNe that can potentially be found in magnitude-limited surveys with PS1 (24000/yr, within z≲0.6), PS4 (69000/yr, within z≲0.8) and LSST (160000/yr, within z≲0.9) surveys.