Conclusion and Future work

A search for astrophysical point sources with a continuous output using the full 1997 data from the AMANDA-B10 detector was performed. No sources were identified, but neutrino flux limits were improved over earlier work by a factor of 3 to 9. A set of selection criteria have been optimized on neutrinos with a hard spectra index of 2, yet this analysis is still sensitive to the softer energy spectra. A suite of software tools have been established to re-optimize the analysis for pointsources, calculate event rates and flux limits.

The analysis depends critically on the Monte Carlo simulation to establish detector sensitivity (effective area $\cdot$ live-time) at all levels of background rejection.. The absolute value of a given analysis variable must be accurately simulated, signal efficiency may be overestimated if the cut parameters are placed in a region of the distribution that changes rapidly. However, inspection of the distributions of variable used in this analysis revealed good agreement between data and simulated background. The agreement was preserved throughout each step in the iteration procedure, creating confidence in the gross behavior of the simulation program. Also, systematic error in the absolute value of the analysis parameters should not affect calculation of detection efficiency because few of the cut parameters are placed in a region of the distribution that rapidly fluctuates. Any discrepancy between experimental and simulated background data is in the direction of underestimating the signal effective area.

The limits presented in this report are valid for sources with continuous output of neutrinos that follow power law energy spectra. Obviously, sources with more complicated energy structure, such as energy spectra with a high energy cut-off (E<<> 1 TeV) or an enhancement at sub-100 GeV energies, or those with periodic or episodic emission do not satisfy the assumptions, and the limits do not apply. It is not unusual for sources of GeV to TeV photons to exhibit strong time variability, and this may be true for neutrino sources as well. Positive observations of point sources by complementary analysis efforts, if any, may point in this direction, but the physics would require fine-tuning to satisfy the Super-Kamiakande and MACRO limits.

List of items that are planned to be done in the future:

$\bullet$ First we will re-process the background MC (after modification have been completed) with the LBNL filter. Greater statistics must be generated for background MC. Without this data, comparisons with the variables used in the optimized cuts at the required background rejection will not be possible.

$\bullet$ Shower reconstruction and iterative reconstruction of the data and signal will be performed. The plan is to use the variables from these fits to further optimize signal to noise and perhaps relax some of the current selection criteria to avoid strong dependence on a particular variable.

$\bullet$ Continue to work on developing a reliable energy variable to remove low energy background while keeping the high energy signal events. If a point source is observed, this analysis will lead to an energy spectrum (given sufficient statistics).

$\bullet$ Study and characterize the events that pass the cuts. One factor that needs to be investigated is the accuracy of the Monte Carlo to the data, and this result of the Monte Carlo event rates being a factor 2 too high.

$\bullet$ Lastly, process the 1998 data, using similar or improved analysis.