5.6 Future Needs and Directions

There are a number of uncertainties in the MMRD calibration of novae, most of which have been discussed in Sec. 5.2 and Sec. 5.4. Although none of these uncertainties is debilitating, all must be addressed in future efforts to improve the nova distance scale. We describe two large projects below that would improve the calibration of the MMRD relation for distance scale work.

Galactic Calibration: The systematic difference between the Galactic and M31 MMRD relations needs to be understood better. Following Ford and Ciardullo's (1988) suggestion, it would be extremely helpful to have a new Galactic calibration based solely on novae whose shell geometry is well determined.

M31 Calibration: Problems with the existing M31 nova database include possible transformation errors from one set of observations to another, systematic photometric errors for novae observed photographically against the bright M31 bulge, sensitivity to reddening, and poor temporal sampling of nova light curves. A new CCD survey of the M31 bulge would solve many of these problems. Using the I bandpass would allow observing at all lunar phases (thus improving the continuity of the observations), and would lower the effect of internal reddening in M31 (if any). It is also possible that the effect of the ``reddening pulse'' near maximum light (van den Bergh and Younger 1987) may be smaller at longer wavelengths. The use of a detector with linear response would provide an effective means of standardizing the observations to a common photometric system, and removing the effects of the strongly varying background of the M31 bulge. If two 0.5-1.0-m telescopes (separated by ~ 180° in longitude) were used, high quality data on a number of novae could be obtained in only 2-3 years. The only problem with using longer wavelength (I) data for the calibration is that the contrast of novae against an underlying old (red) population would be decreased; this could be critical for observations of more distant galaxies. However this decrease in contrast would be offset by the higher rate of detected photons in the red, and possibly by the use of active optics techniques (for galaxies beyond the Local Group).

Clearly nova surveys are feasible in a number of galaxies beyond the Local Group using large format CCDs. Such surveys would be particularly valuable to tie the Cepheid distance scale to that of novae, and to add more calibrators to the Tully-Fisher relation. Certainly the M81 group is within reach, and, in sub-arcsec seeing, galaxies in the Leo Group may also be observable. Since the nova rate in ellipticals and spiral bulges appears to be substantially higher than in disks (Ciardullo et al. 1987), it follows that surveys in galaxies of morphological type Sb and earlier would be most productive.

As pointed out by Ciardullo and collaborators (Ciardullo et al. 1983; Ciardullo et al. 1987; Ciardullo et al. 1990a, b), there are several strong arguments for observing extragalactic novae in the light of H. This emission line remains bright for a much longer time than does continuum emission (e.g., Fig. 1 of Ciardullo et al. 1990a), and the contrast of a nova against the underlying background light is excellent when observed through a narrowband filter. Unfortunately, the available evidence (Ciardullo et al. 1990a) suggests that there is no correlation between the rate of decline and maximum brightness in H light. However, Ciardullo et al. have discussed how the rate of fading of intrinsically faint novae may still be obtained from ``surrogate'' H observations, because the late stages of nova light curves have similar rates of decline in B and H.

An alternate technique for using H observations in distance scale work involves the measurement of the nova rate (normalized per unit luminosity). Ciardullo et al. (1990b) argue that this quantity may be constant over a wide range of galaxy types, provided that the luminosity of the old stellar component is used to normalize the nova rate. From Figs. 2 and 3 of Ciardullo et al. (1990a), it is apparent that most novae in M31 remain brighter than M_H = -6.5 for at least 15 days. It follows that, if observations of a galaxy were obtained more frequently than every 15 days, and if the limiting magnitude of the observations were deep enough to reach M_H = -6.5, then the measured nova rate would sample all novae, and hence would be proportional to the old star luminosity of the region observed. A distance modulus could therefore be derived based solely on counting statistics, and the precision of the result would improve over the years as additional novae are observed. Calculations show that observations of almost complete samples of novae in the Virgo cluster are feasible in the 0."5-0."6 seeing conditions that are not uncommon on Mauna Kea. For this technique to be applied, more H data on M31 novae would be required - both to improve our estimate of the M31 nova rate per unit luminosity (upon which the distance determinations rest), and also to calibrate the optimal temporal spacing more accurately. Such a survey is well within the reach of a small telescope.