Next Contents Previous

9.5 Future Needs and Directions

Measurements of fluctuation fluxes are accurate in comparison with traditional distance estimators. They are also relatively cheap in terms of telescope time. However, the uncertainties rise rapidly with distance and degraded seeing. Table 2 lists the required observing times and errors in barmI (in percent) that we might encounter observing with a CCD on a 4-meter telescope in the I band. Note that these numbers assume a galaxy with a typical specific frequency of globular clusters (S = 6), and do not account for errors which will creep in when we try to push to fluctuations smaller than 0.2% of the overall signal. Under non-ideal conditions the errors might be a factor of two larger, but they should not be a factor of three larger (so the 32% error measurements may be impossible, but the 16% ones are not). The observation times are based on collecting 20 photons per star with a CCD which has 50% QE in the I band.

Table 2. Observation Times and Uncertainties

Flux Error (percent)

Distance Obs. time - psf FWHM -

(km/s) (min) 1.5" 1.0" 0.5"

700
1000
1400
2000
2800
4000
5600
8000

(Leo)
(Virgo)
(Eridanus)
(Centaurus)
(Great Attractor)
(Perseus)
(Coma)
4
8
15
30
60
120
250
500
4
5
9
16
32
-

-
3
4
5
9
16
32
-
. .
3
3
3
4
5
9
16
32

The most important task for this method is to verify that barMI behaves universally according to the trend with color and to calibrate its zero-point. The first test can be carried out by observing galaxies in tight groups whose distances are identical. The calibration problem is more difficult. The bulge of M31 has been observed as well as some of the other elliptical galaxies in the Local Group whose distances are known, but we must worry about possible systematic differences between calibrator (primarily the bulge of M31) and application (primarily giant ellipticals). An important test will be observations of a number of other spiral bulges such as NGC 7331, NGC 4565, NGC 4594, NGC 3368, and NGC 2841, which can be compared with their elliptical neighbors or with other distance estimates. Preliminary results from observations of barMI in the bulges of NGC 3031, NGC 4565, NGC 4594 show no discrepancy with the current calibration with mean color. It is also possible to use galactic globular clusters as calibrators as well as clusters in the LMC and SMC, but even the most metal rich clusters will have brighter barMI than most galaxies.

There are potentially large improvements in the theoretical calibration of barMI. Improved theoretical isochrones at solar metallicity and better observations in the I band to which bolometric corrections can be tied are necessary. The connection between fluctuations and actual stellar populations in Local Group galaxies needs to be improved. Freedman (1989) has reported observations of the brightest 2 magnitudes of the luminosity function of M32, and with slightly deeper observations barmI could be calculated directly from equation (20).

There is some potential for improving the measurement of barm. The dominant source of error comes from the residual point sources, and there are a number of approaches that may improve the situation. Since the fluctuations are redder than most point sources, the point sources can perhaps be identified in a blue image and excised from the I band image. This leads to difficulties with the unknown colors of the unseen sources, but some improvement should be possible. Improvements in seeing are enormously helpful; they not only make the point sources easier to detect, but they also make the fluctuation amplitude larger. This is an important consideration at distances greater than about 3 times that of Virgo, because it is difficult to obtain a CCD image which is flat and unfringed at the 0.3% level.

As infrared arrays improve, it will be worthwhile observing in the K band. barMK is very bright, barMK approx -5.6 (Luppino and Tonry 1992), and the seeing is typically better in the IR. Thus the fluctuation signal is substantially larger, both in absolute terms and relative to point sources for which (V - K) approx 3. The variability of barMK with stellar population is currently unknown, but barMK in M31 and M32 differ by only 0.3 magnitudes. A significant drawback is that the sky is substantially brighter in the K band, (V - K) = 7-8, but of course the extinction due to dust at K is smaller by about a factor of four relative to the I band.

In the future we can hope to be able to observe with a psf whose FWHM is 0.1 arcsec or better and with CCDs capable of signal-to-noise ratios significantly better than 1000. Such an observation could measure the fluctuations in a galaxy at 20,000 km s-1 distance with sufficient accuracy that it would just be possible to determine peculiar velocities and to resolve the depth of a cluster of galaxies. Until that time we will have to be content with high accuracy in our own supercluster and vicinity.

Next Contents Previous