The Taurus Tunable Filter Field Galaxy Survey: Sample selection and narrowband number counts

Recent evidence suggests a decline in the volume-averaged star formation rate (SFR) with the advance of cosmic time since z similar to 1. It is not clear, however, the extent to which the selection of such samples influences the measurement of this quantity. Using the Taurus Tunable Filter (TTF) we have obtained an emission-line sample of faint star-forming galaxies over comparable look-back times : the TTF Field Galaxy Survey. By selecting through emission lines, we are screening galaxies through a quantity that scales directly with star formation activity for a given choice of initial mass function. The scanning narrowband technique furnishes a galaxy sample that differs from traditional broadband-selected surveys in both its volume-limited nature and selection of galaxies through emission-line flux. Three discrete wavelength intervals are covered, centered at H alpha redshifts z = 0.08, 0.24, and 0.39. Galaxy characteristics are presented and comparisons made with existing surveys of both broadband and emission-line selection. Little overlap is found in a direct comparison between the TTF Field Galaxy Survey and a traditional galaxy redshift survey, as a result of the respective volume and flux limitations of each. When the number counts of emission-line objects are compared with those expected on the basis of existing H alpha surveys, we find an excess of similar to3 times at the faintest limits. While these detections are yet to be confirmed independently, inspection of the stronger subsample of galaxies detected in both the line and continuum (line-on-continuum subsample; 13%) is sufficient to support an excess population. The faintest objects are galaxies with little or no continuum, rendering them undetectable by conventional redshift surveys. This increase in the emission-line field population implies higher star formation densities over z less than or similar to 0.4. However, further study in the form of multiobject spectroscopic follow-up is necessary to quantify this and confirm the faintest detections in the sample.