TITLE:   GCN CIRCULAR
NUMBER:  10777
SUBJECT: Update to the configuration of the Fermi-LAT onboard GRB search
DATE:    10/05/19 00:54:02 GMT
FROM:    Julie McEnery at NASA/GSFC  <julie.e.mcenery@nasa.gov>

Julie McEnery (GSFC), Gregg Thayer (SLAC), J. J. Russell (SLAC), Sylvia Zhu
(GSFC) and Nicola Omodei (Stanford) report on behalf of the Fermi-LAT
collaboration.

On May 14, the Fermi- LAT team made a significant update to the
configuration of the GRB search algorithm onboard the LAT.

There are two search modes in the onboard algorithm: a) one that
searches solely within the LAT data for spatial and temporal clusters of
events and b) another that seeds a search based on the position and time
of a GBM detected burst. The recent changes affect the GBM seeded case.

For most of the mission the configuration of the onboard algorithm had a
very low threshold for the GBM seeded window, as a diagnostic test on
the algorithm performance. This meant that if even 1 event was observed
within 10 deg radius of the GBM seed position then the LAT onboard
flight software would trigger and enter a localization and refinement
stage (essentially always triggering if the GBM position was in the LAT
FoV). The localization and refinement stage lasts for 600 seconds,
during which time no further triggers are allowed. We applied a filter
on the ground before sending public GCN notices to filter out the large
number of false triggers.

The updated configuration performs a LAT search on a GBM seeded position
at 2, 5, 10, 20, 30, 60, 90, and 150s after the GBM trigger. The
threshold is applied onboard so that the LAT remains in trigger mode
unless and until a significant number of events are seen that are
consistent with the seeded position. This provides a capability to
generate a LAT trigger for GRB where the high energy emission is
significantly delayed with respect to the GBM detection. This behaviour
has been seen from several GRB detected at high energies.

To explore the false trigger rate, we ran the new configuration on 80,000
test positions. This resulted in a predicted false trigger rate of
0.33/year. To explore the ability of the new configuration to detect
GRB, we ran it on data containing known LAT bursts. This resulted in
triggers of 5 out of the first 13 LAT detected bursts (080916C, 081024B,
090510, 090902B, 090926). Previously, the LAT had sent out only one
onboard-generated notice (090510). We thus expect the new configuration
to provide onboard detections of 3-5 GRB/year, with localizations in the
range 0.1 to 0.5 degrees.