DWARF GALAXY STARBURST STATISTICS IN THE LOCAL VOLUME

An unresolved question in galaxy evolution is whether the star formation histories (SFHs) of low-mass systems are preferentially dominated by starbursts or modes that are more quiescent and continuous. Here, we quantify the prevalence of global starbursts in dwarf galaxies at the present epoch and infer their characteristic durations and amplitudes. The analysis is based on the Ha component of the 11 Mpc H alpha UV Galaxy Survey (11HUGS), which provides H alpha and Galaxy Evolution Explorer UV imaging for an approximately volume-limited sample of similar to 300 star-forming galaxies within 11 Mpc. We first examine the completeness properties of the sample, and then directly tally the number of bursting dwarfs and compute the fraction of star formation that is concentrated in such systems. To identify starbursting dwarfs, we use an integrated H alpha equivalent width (EW) threshold of 100 angstrom, which corresponds to a stellar birthrate of similar to 2.5, and also explore the use of empirical starburst definitions based on sigma thresholds of the observed logarithmic EW distributions. Our results are robust to the exact choice of the threshold, and are consistent with a picture where dwarfs that are currently experiencing massive global bursts are just the similar to 6% tip of a low-mass galaxy iceberg. Moreover, bursts are only responsible for about a quarter of the total star formation in the overall dwarf population, so the majority of stars in low-mass systems are not formed in this mode today. Spirals and irregulars devoid of H alpha emission are rare, indicating that the complete cessation of star formation generally does not occur in such galaxies and is not characteristic of the interburst state, at least for the more luminous systems with M-B < -15. The starburst statistics presented here directly constrain the duty cycle and the average burst amplitude under the simplest assumptions where all dwarf irregulars share a common SFH and undergo similar burst cycles with equal probability. Uncertainties in such assumptions are discussed in the context of previous work.