HAM presented preliminary estimates of the rates at which starbursts inject mass, metals, and energy into superwinds at the present epoch. These estimates were based on the following assumptions (see HAM for details): First, all galaxies in the local universe with large IR luminosities (LIR > 1044 erg s-1), large IR excesses (LIR / LOPT > 2), and/or warm far-IR colors (flux density at 60 µm greater than 50% of the flux density at 100 µm) drive superwinds. Second, The superwind outflow rates are proportional to the far-IR luminosity, with the constants of proportionality determined from well-studied examples like M 82 and NGC 253.
The recent survey of optical line emission associated with a well-defined sample of nearly 50 IR-selected edge-on starbursting disk galaxies by L92 and LH have put these two assumptions on a much firmer observational foundation. They find that such indicators of a superwind as excess optical line-emission along the minor axis, broad emission-line widths and velocity shears along the minor axis, and 'shock-type' optical emission-line ratios (e.g., strong [O I] and [S II] emission relative to H) along the minor axis all correlate with such quantities as the IR luminosity, the IR/optical luminosity ratio, and the IR color-temperature. They conclude that superwinds are probably ubiquitous in galaxies meeting the IR criteria listed in the preceding paragraph.
Integrating over the local far-IR luminosity function for galaxies meeting the above criteria (Bothun, Lonsdale, & Rice 1989) then yields the superwind injection rates per unit volume in the local universe. The resulting injection rates per unit volume can be normalized by multiplying them by the age of the universe, and then dividing the result by the local space density of galaxies. This calculation then implies that superwinds have carried out approximately 5 × 108 M in metals and 1059 ergs in kinetic plus thermal energy per average (Schechter L*) galaxy over the history of the universe. For reference, note that these two quantities are approximately equal to the mass of metals contained inside an average galaxy and the gravitational binding energy of an average galaxy respectively.
These values pertain to the conservative assumption that the superwind rate has not evolved with look-back time. Since the star-formation rate was presumably higher in the early universe, the true time-integrated values are likely to be about an order-of-magnitude higher (cf. Ostriker & Cowie 1981; HAM). On the other hand, it is unclear what fraction of the energy and mass injected into a superwind by a starburst ultimately escapes the galaxy and is delivered to the intergalactic medium (e.g., less energetic or more strongly confined superwinds may conceivably cool and turn into galactic fountains (cf. Norman & Ikeuchi 1989). In any case, it seems quite likely that superwinds do play an important role in enriching and heating the intergalactic medium (as we will discuss more fully below).