TABLE OF CONTENTS:
The GRB Coordinates Network (GCN) is composed of 3 parts.
The first part
(the original portion of GCN) was called the BATSE COordinates DIstribution
NEtwork (BACODINE). It monitors the BATSE real-time telemetry from CGRO,
calculates RA,Dec locations for GRBs that it detects and distributes them
to sites around the world in real time (a few seconds). With the de-orbiting of the CGRO satellite, this portion of GCN is no longer functioning.
The second part consists of collecting RA,Dec locations of GRBs detected
by other spacecraft and distributing that information to the same sites.
When the second part was incorporated, the original BACODINE name was
changed to the more general name, GCN.
The third part consists of receiving and automatically distributing
to the GRB community prose-style e-mail messages
about follow-up observations on various GRBs -- the GCN Circulars and Reports.
1.1) Details of the BATSE-derived (BACODINE) portion:
To increase the probability that Gamma-Ray Burst (GRB) counterparts can be detected, we have built a system of hardware and software that calculates positions of GRBs detected by the BATSE instrument on CGRO and distributes those coordinates to instruments around the world within a time short enough that most GRBs are still bursting. The system is called the BATSE COordinates DIstribution NEtwork (BACODINE). It became possible to do this because of the failure of the on-board CGRO tape recorders. All the data from CGRO is transmitted to ground in real-time instead of delayed transmission from the tape recorder playback.
BATSE detects ~0.8 GRBs per day. Folding in multiplicative factors for the fraction of the sky that any ground-based instrument can see (0.35), the fraction when it is night (0.50, for the IR to UV band passes and 1.0 for some radio and TeV bandpasses), the fraction of clear-sky weather (0.75), and the "program efficiency" (0.60, see discussion below); the rate of making follow-up observations is 0.063 GRBs per day or once every 16 nights. [If factors for New Moon (0.50), for the brightest GRBs (0.50), and for the longest GRBs (0.50) are included; then the rate is 0.007 per day or once every 4 months for optical instruments and 1 month for radio.]
BATSE has two types of data from the 8 Large Area Detectors (LAD). The high time/spectral resolution data are buffered on-board the spacecraft for delayed transmission to the ground, but there is also 1.024-sec rough spectral resolution data which are transmitted continuously in real time. Figure 1 is a timeline of the sequence of steps in the flow of this real-time data from BATSE to the GCN (BACODINE) processing system. The count rates for the 8 LADs in 4 energy intervals are accumulated for two 1.024-sec intervals and then transmitted to a TDRS satellite over the next 2.048 seconds. The next two 1.024-sec counting rate samples are accumulated while the transmission of the previous 2 samples is in progress, and this process repeats continuously. For those portions of the orbit (85%) where CGRO can get direct line-of-sight transmission of its high-gain antenna to any of the three TDRS satellites, the data are relayed to the NASA White Sands Ground Station in New Mexico where they are retransmitted to DOMSAT and then transmitted back down to the Goddard Space Flight Center Data Capture Facility (see Figure 2). There is an additional 1.0 seconds of delay due to four hops of ground-to-geosync-orbit light-travel time and buffering within the White Sands facility. Once received at GSFC, the entire 2.048 sec of data is processed to yield GRB coordinates within 0.1 sec. The fastest method of coordinates distribution (see Table 1) takes an additional 0.3 seconds. If the GRB started at the beginning of the first 1.024-sec count-rate sample, then the total time delay between when the gamma rays interacted with the BATSE LADs and when the coordinates are available at an instrument to make follow-up observations is 5.50 seconds. If the GRB started at the end of the second 1.024-sec sample, then the time delay is 3.45 seconds. More than half of the GRBs are longer than 5.5 seconds, thus allowing follow-up observations to be made while the burst is still occurring. (See the distribution of GRB durations.)
While there are several programs running on several computers to do the processing of the BATSE data, the sequence of processing can be broken down into five basic steps. (1) A program monitors the telemetry stream continuously extracting the count-rates for the 8 LADs in the 4 energy intervals (25-50, 50-100, 100-300, >300 keV). It uses these rates to determine the current background rates. It also extracts some general purpose housekeeping information (spacecraft clock, RA,Dec orientation of the SC, etc). (2) It monitors the "burst-in-progress" flag generated by the BATSE flight processor and when set true, it (3) takes the current count rates, subtracts the previously accumulated background rates to get the source-only rates, finds the 3 brightest detectors, and (4) solves the set of 3 simultaneous equations of the dot-product of the unknown burst direction and the detector normals of the 3 bright detectors. (5) The burst direction is then sent to a list of instruments that are capable of making follow-up observations.
The distribution of GCN sites (as of ~1995) is shown the two figures (a world map and a US map).
Currently, the algorithm used to calculate the GRB direction for the BATSE Original, Final, and MAXBC Notices assumes "ideal response" physics for the LADs. This means (1) that the count rates in the LADs are proportional to the cosine of the angle between the detector normal vector and the GRB direction vector, and (2) that there is no distortion of the count rates in each detector due to detection of gamma rays scattered the the Earth's atmosphere or from the spacecraft. These two approximations yields an uncertainty for the BATSE (Original/Final/MAXBC) burst position of about a 20 deg diameter error circle (typical worst case).
1.2) The Other (non-BATSE) Sources of GRB Locations:
As mentioned above in the introductory section, the original BACODINE system has been expanded to include GRB location information from other sources. It was this expansion that motivated the name change from BACODINE to GCN (GRB Coordinates Network). Currently, the other sources are: HETE, INTEGRAL, IPN, RXTE (PCA and ASM), and ALEXIS (although ALEXIS is not strickly GRBs). (Other, now discontinued, missions: BeppoSAX, CGRO-COMPTEL, and NEAR.) The details of these contributions are descussed below.
1.3) Brief description of the GCN Circulars portion:
This part allows the GRB community to submit messages to a central queue where they are automatically distributed (e-mail) to the entire GRB community. These are prose-style messages (as opposed to the "TOKEN: value" style of the GCN e-mail messages) from follow-up observers reporting on their results (detections or nulls) or for coordinating with others. Click for the detail of the GCN Circulars.
1.4) Brief description of the GCN Reports portion:
Reports are also prose-style write-up on burst observation, but they are issued at a later time (not wuick like the Circulars), allowing the observers to do the full analysis and corrections to the measurements. They are distributed (e-mail) to the entire GRB community. Click for the detail of the GCN Reports.
There are currently 14 different types of location information for GRBs in real- or near real-time. They span a range of delay times after the start of the burst and they span a range of the sizes of the error boxes. Each type has its own separate dis/enable flag in the "sites.cfg" file controlling the distribution of that type to each site. If sites want different filtering (ALL, VIS, NIGHT, etc) for the different notice types, they can have multiple entries in the "sites.cfg" file. The comparison is shown in Table 0, and they are described in more detain in the sections below.
| SOURCE | TIME DELAY | ERROR BOX SIZE | RATE | COMMENTS |
|---|---|---|---|---|
| IPN_POS | 0.5-1.5 days | 5-20' dia | 3/month | Small FOV |
| INTEGRAL_WAKEUP | X sec | X' | 1/month | Small FOV. |
| INTEGRAL_REFINED | X sec | X' | 1/month | Small FOV. |
| INTEGRAL_OFFLINE | X sec | X' | 1/month | Small FOV. |
| RXTE-ASM | 1-2 hours | 4'x15-150' | ~8/year(2) | Small FOV telescopes |
| RXTE-PCA | 3-5 hours | 6-40' dia | ~6/year(2) | Small FOV telescopes |
| Swift-BAT_POS | 13-40 sec(1) | 1-5' dia | 2/week | Fast and Small |
| Swift-XRT_POS | 30-80 sec(1) | 5" dia | 2/week | Fast and Small |
| Swift-UVOT_POS | 0.2-9 hrs(1) | 2" dia | 1/week | Fast and Small |
| SuperAGILE | 20 sec | 20' dia | X/month | Small FOV |
Note (1): The time_delys for the 3 Swift-based positions are the values
after the ~135-day Verification phase (when the mission is in the full automatic
distribution mode of operations. Prior to that (during the Verif phase),
the delays will be in the 1-24 hr range because there will be humans-in-the-loop
to verify that the positions are accurate (and that the triggers are true GRBs).
Note (2): The RXTE-ASM and -PCA detection rate has dropped to essentially 0.
| SOURCE | TIME DELAY | ERROR BOX SIZE | RATE | COMMENTS |
|---|---|---|---|---|
| AGILE-GRID | XX sec | 10-30' dia | X/month | Small FOV |
| Fermi-GBM | 20 sec | 4-10 deg dia | 15/month | Large FOV |
| Fermi-LAT | 100 sec | 10-30' dia | 1/month | Small FOV |
| SOURCE | TIME DELAY | ERROR BOX SIZE | RATE | COMMENTS |
|---|---|---|---|---|
| Original | 5 sec | 5-20deg dia | 1/day | Dedicated, automated, wide FOV. |
| Final | 37 sec | 5-18deg dia | 1/day | Dedicated, automated, wide FOV. |
| Light Curve | 5 min | n/a | 1/day | Useful for burst assessment. |
| MAXBC | 10 min | 5-20deg dia | 1/day | Fast response, wide FOV. |
| LOCBURST | 15-35 min | 4-8deg dia | 7/month | Medium response, medium FOV. |
| MSFC Admin | 1-3 days | 2-5deg dia | 1/day | Final report |
| IPN_SEG | 0.5-3 days | 4'x4-8deg | 3/month | Small FOV with tiling. |
| COMPTEL | 20-30 min | 3-5deg dia | 6/year | Medium response, medium FOV. |
| SAX-WFC | 2-3 hours | 6-20' dia | ~8/year | Small FOV telescopes |
| SAX-NFI | 12-48 hours | 100" dia | ~4/year | Small FOV telescopes |
| MILAGRO | 30 sec | 0.5 deg | 5/week | Fast and Small |
| HETE_SC_ALERT | 10-20 sec | n/a | 8/month | Small FOV. |
| HETE_SC_UPDATE | 10-60 sec | X-X' | 2/month | Small FOV. |
| HETE_SC_LAST | 20-60 sec | X-X' | 1/month | Small FOV. |
| HETE_GNDANA | 1-3 hr | X-X' | 2/month | Small FOV. |
| SOURCE | TIME DELAY | ERROR BOX SIZE | RATE | COMMENTS |
|---|---|---|---|---|
| INTEGRAL_SPIACS | n/a | n/a | n/a | The date/timestamp of a rate-trigger increase in the SPI Anti-Coincidence Shield. |
| INTEGRAL_POINTDIR | n/a | n/a | n/a | The s/c pointing direction after the next slew. |
| ALEXIS (extreme UV) | 12-24-48 hours | 0.38deg radius | 20/year | Small FOV telescopes. |
| XTE-ASM hard x-ray TRANS | 1-2 hours | 4'x15-150' | 1-8/year | Small FOV telescopes. |
| Swift-BAT hard x-ray TRANS | 20 sec | 1-4'' | 1-2/month | Small FOV telescopes. |
2a) BATSE GRB LOCATIONS (no longer available):
2a.1) Original
This is the original form of GRB locations distributed to the then BACODINE sites. (Now this, as are all the other notice types, is distributed to the GCN sites.) The positions in these location notices are derived using the "ideal physics" response functions for the LADs and therefore have the largest location uncertainties. However, they are also available to the follow-up community with the shortest time delays from the start of the GRB -- typically less than 5 sec, which is shorter than the duration of more than half the GRBs. Click for more information on Original Coordinates Notices.
2a.2) Final
Unlike the Original location which uses only the first 1 or 2 seconds of data in the GRB, this Final location is based on the integral of the light curve for up to 32 seconds of the GRB. The Final Notices report the peak intensity and the fluence during the 32 seconds. These two parameters allow the user to assess the "importance" of the burst. With more counting statistics, there is slight improvement in the location uncertainty (i.e. it is somewhat systematics dominated). Click for more information on Final Coordinates Notices.
2a.3) MAXBC
The MAXBC locations are derived from the 16 "Maximum Burst Channels" that are determined by the on-board BATSE flight software. The 16 rates are available to the GCN (BACODINE) system at T+10 minutes after the trigger. They are useful for those situations when the initial GRB BATSE trigger happened within a telemetry gap, making the generation of the Original Notice impossible. The location uncertainty is comparable to the Original locations. Click for more information on MAXBC Coordinates Notices.
2a.4) LOCBURST
This source of GRB location information improves the location uncertainty by a factor of 2-5x over the Original and MAXBC locations. It uses the LOCBURST algorithm developed by the BATSE team at MSFC. The GCN (BACODINE) system captures the BATSE telemetry data during a GRB, automatically ftp's it to MSFC, pages an "on call" person, who then applies the Locburst algorithm plus the advantages of a human-in-the-loop for data better selection, calculates a location, sends it back to the GCN (BACODINE) system. The GCN (BACODINE) system monitors for these incoming Locburst locations and distributes them. This whole process takes only 15-35 minutes! Click for more information on LOCBURST Coordinates Notices.
2a.5) MSFC Admin
These reports come out 1-3 days after the burst. They are based on the full data set(s) availalble, with the cleanest data chosen, and have the smallest error boxes (of the BATSE-derived Notices). These are the best BATSE-derived locations available until the periodic BATSE Burst Catalog is issued. These GRB locations are not distributed via the normal distribution methods. They are available only in the GRB Locations Web Page Table.
2a.6) Light Curves
The GCN program produces several forms of light curve information for the BATSE triggers (GRBs and non-GRBs alike). Currently, these light curves are only available through the GRB Web Page Table. There is a column that has links to the various forms and formats of the light curves: (1) there are the planewave light curves in the 4 DISCLA energy windows and the IPN light curve in the 20-100 keV window, and (2) the formats are ascii text, postscript, and soon jpeg. Soon, GCN will also have the ability to e-mail the text and postscript forms of the light curves directly to those sites that request this form of information.
2b) THE OTHER SPACECRAFT/INSTRUMENTS GRB LOCATIONS:
2b.1) CGRO-COMPTEL GRB LOCATIONS (no longer available)
The COMPTEL instrument also on CGRO also can localize GRB -- particularly the brighter GRBs and those with harder spectrums. Click for a description of the CGRO-COMPTEL GRB location notices. And click to see the Table of Recent COMPTEL Locations.
2b.2 RXTE-ASM GRB LOCATIONS
The RXTE-ASM instrument detects several GRBs per year in its FOV. The locations for these are automatically processed and made available to the follow-up community through the GCN. Click for a description of the RXTE-ASM GRB location notices. And click to see the Table of Recent RXTE-PCA/-ASM Locations.
2b.3) RXTE-PCA GRB LOCATIONS
The RXTE-PCA team has started a program to make ToO observations of bright GRBs detected by BATSE. Click for a description of the RXTE-PCA GRB location notices. And click to see the Table of Recent RXTE-PCA/-ASM Locations.
2b.4) BeppoSAX-WFC/-NFI GRB LOCATIONS (no longer available)
The BeppoSAX-WFC/-NFI teams have a program to detect GRBs in the WFC FOV and quickly analyse the WFC data to yield a location. And if operational constraints permit, then a TOO is invoked with the NFI instruments. Click for a description of the BeppoSAX-PCA/-NFI GRB location notices. And click to see the Table of Recent SAX-WFC/-NFI Locations.
2b.5) IPN_POSITION GRB LOCATIONS
The GCN system also include automated processing of the NEAR-XGRS instrument telemetry data and the Wind-KONUS instrument telemetry data on a daily basis. This processing looks for GRBs in the data stream, and when a burst is found then the lightcurve is auotmatically sent to the K.Hurly's automated processing at UC Berkeley for comparison with the Ulysses burst detections. Using these 3 s/c (plus optional location information), small error box locations can be obtained. Click for a description of the IPN POSITION GRB location notices.
2b.6) IPN_SEGMENT GRB LOCATIONS (no longer available)
The GCN system has greatly speeded up the process of producing IPN information by (a) automatically sending the BATSE light curve data to Kevin Hurley (UCB) where is it automatically combined with the Ulysses data. Given the current situation of having only two spacrecraft with a long baseline, only annuli are produced. However, these annuli are combined with the BATSE-Original or BATSE-LOCBURST locations to provide "arc segments" which are then distributed by GCN. Click for a description of the IPN SEGMENT GRB location notices. And click to see the Table of Recent GCN/IPN_SEG Locations.
2b.7) HETE2 GRB LOCATIONS
In the near future the GCN system will capture and distribute transient locations detected by the HETE (MIT) spacecraft. Because HETE GRB error boxes will be small (0.3 deg, X-ray Monitor), traditional narrow-FOV telescopes with much fainter sensitivity can be used to make follow-up observations. Note: The Pagasus failed to deploy the HETE spacecraft, but when HETE2 flys (launch in the Fall of 2000) GCN will distribute those transient locations as well.
(Oct 2000): HETE2 has been launched and is undergoing a series of on-orbit commissioning checks. Once completed the burst positions will be made available through the GCN system. See the HETE (GCN) page.
2b.8) ALEXIS Extreme-UV TRANSIENT LOCATIONS (no longer available)
Click for a description of the ALEXIS location notices on the extreme-UV (66-93 eV) transients. And click to see the Table of Recent ALEXIS extreme-UV Transient Locations.
2b.9) Swift GRB LOCATIONS, LIGHTCURVES, SPECTRA, and IMAGES
The Swift mission opens up a new era in GRB research. This is because of the combined (a) rapid availability, plus (b) small error boxes. The Swift mission also provides data products not previously available from prior missions: (a) spectra and (b) images, plus (c) lightcurves. See the Swift (GCN) page. And click to see the Table of Recent Swift Locations.
2b.10) MILAGRO GRB LOCATIONS
The MILAGRO instrument detects hi-energy (>100 GeV) air shows from gamma-rays and cosmic rays and localizes the direction of these showers. See the MILAGRO (GCN) page. And click to see the Table of Recent MILAGRO Notices.
2b.11) GRB COUNTERPART LOCATIONS
Follow-up observers can submit to GCN locations of GRB counterparts that they have detected. Once vetted (same as Circulars process), these are distributed to those sites enabled to receive this Notice type. These counterpart locations are useful to the automated sites because these Notices can be read by the automated sites whereas the Circulars (where counterparts are also published) can not be read by automated programs. See the COUNTERPART page.
2c) TEST LOCATIONS
While not strictly a source of GRB locations, the Test Notices are another "source" of information in that they can be enabled or disabled for each site just like the above sources of GRB locations. They can be used by sites for end-to-end testing of their communication link and their instrument. There are several mission-based types of test notices. Click for the details.
2c) POINTING DIRECTIONS
Various missions have pre-planned knowable pointing directions. This informations is cptured within the GCN system and made available tot he community so that their telescopes can follow along (ie be pointed at the location of the mission so that their slew time is minimized should aburst notice for that mission be issued. And further, they can be taking data so as to have data prior to and during the prompt phase of the burst. Click for the details.
There are 6 methods currently available for distributing the GCN GRB coordinates. They are listed in Table 1 and are discussed in more detail below. Click here for a plain text version of Table 1. The following table and seven paragraphs only contain a brief description of the distribution methods/media. For a very detailed description of the contents, formats, and meaning of these various distribution media/methods, please see the technical details section.
| TIME DELAY | METHOD/MEDIA | COMMENTS |
|---|---|---|
| 0.3 sec | Dedicated phone | Continuous phone/modem connection. (no longer available) |
| 0.1-1.0 sec | Socket | Fast & suited for automated instruments. |
| 5-30 sec | E-mail (text) | To any network address (johndoe@machine.domain). |
| 5-30 sec | E-mail (XML) | To any network address (johndoe@machine.domain). |
| 30-90 sec | Dialed phone | Slower but much cheaper than Dedicated. (no longer available) |
| 60-180 sec | Pager | RA,Dec,UT,Intensity displayed on the pager. |
| 60-180 sec | Short Pager | RA & Dec displayed on the pager. |
| 5-180 sec | Subject-only | RA & Dec displayed in the Subject-line. |
| 5-180 sec | SubjHHMM-only | RA, Dec, Time, & Intensity displayed in the Subject-line in RA=HH:MM:SS format. |
Dedicated Phone (no longer available): The fastest method is the dedicated phone line. Around sunset at the instrument site (assuming it's an optical instrument), a phone/modem connection is made between the GCN computer and the computer at the instrument site. This connection is maintained throughout the night and should a burst occur during this time then the coordinates (RA, Dec, UT) are sent over the connection. At 9600 baud it takes 0.3 seconds. Due to a limitation of the US Government phone network here at GSFC, this method is only available within the US (overseas calls require a human operator in the loop to place the call and this is not possible with a computer-initiated call).
Socket: The second fastest method, and much less costly (it's free), is the Internet socket connection. Sockets is a technique to connect two computers over a network. Like the dedicated phone method, the socket connection is made at some initial time and is maintained for long periods of time. The GCN system runs continuously allowing sites to connect and disconnect at their leisure. The time delay for the propagation of the coordinates packet varies due to the distance between the two computers, the number of routers and gateways in between, and the amount of other network traffic. However, we have routinely shown that for a connection between Maryland and California (coast to coast US) the (roundtrip) propagation time is typically less than 1.0 seconds and less than 0.1% of the packets take longer than 2.0 seconds ( a histogram of distribution times). For the details.
E-mail (text): For those sites that do not have automated instruments, the e-mail notice distribution method is suitable. (Examples of the E-mail Notices that are distributed.) It is still quite fast; typically delivery times range from a few seconds to ~30 seconds, when the destination machine and local gateways are on-line. All the same information in the Internet socket packets for a burst notification is in the body of the e-mail. A "TOKEN: value" format was chosen as a reasonable compromise between human-readable and machine-readable formats. The later is useful for sites that have daemons waiting for incoming notices and take further action based on the contents (some sites have exploders and some sites parse the information). For the details.
E-mail (XML): The XML format contains the same information as the "text" format email notices described above. However, the contents are in XML (instead of the TOKEN:value format above). The e-mail notice distribution method is suitable. (Examples of the E-mail XML Notices that are distributed.) The XML VOEvent message can be in either the body of the email or as an attachment (customer's choice). For the details.
Dialed Phone (no longer available): GCN also has the capability to dial a predefined phone number at the time of the burst and make a modem transfer of the relevant information. The protocols are the same as for the "dedicated phone connection" method, it is just that the establishment of the link at the time of the burst adds a time delay to the receipt of the information, however, with a great savings in the telephone bill. Like the Dedicated Phone method described above, Dialed Phone calls are limited to the US (overseas calls require an operator).
Pager/Cellphone: The distribution of GCN GRB information is also available via alpha-numeric pagers and text-message-capable cellphones. (Examples of the Pager messages that are distributed.) Many pager and cellphone service-providors accept e-mail and automatically transmit it to the designated pager/cellphone -- you can get beeped by the Universe. This distribution method is perfect for those sites that do not have Internet access (for either e-mail or sockets). The pager units typically have a limit of 100-400 characters, but this is plenty to display the GRB's RA,Dec location, the UT time of the burst, and the initial intensity (the RA,Dec are current epoch). Current experience shows that the time between the GRB and the "beep" of the pager is 5 sec to 3 minutes depending on company, geography and time of day. The main limitation if geographic coverage. The companies tend to locate their transmitters in large metropolitan areas (cities with more than 50K people) and the range seems to be about 20-30 miles. This was the situation circa Late 1995. More companies have entered the market; and the coverage is a lot better now.
Short Pager: This is essentially identical to the above "Pager" capability with the exception that only the RA,Dec location (Epoch 1950) of the burst is transmitted to, and displayed on, the pager screen. (Examples of the Short-form Pager messages that are distributed.) This reduced information message is suitable for those people/sites with small-display alpha-numeric pagers.
Subject-only: This is a specialized version of distribution. It puts a short-form synopsis (RA,Dec position) in the Subject-line of an e-mail message. This is suitable for some pagers and cell-phones which only display the From-line and Subject-line of the e-mail (i.e. they do NOT have the ability to display the body of the e-mail). The RA,Dec coordinates are Current Epoch. (Examples of the Subject-only messages that are distributed.) This reduced information message is suitable for those people/sites with reduced-capability pagers and cell-phones (e.g.GSM Information Network in Europe).
SubjHHMM-only: This is a specialized version of distribution. It puts a medium-form synopsis (RA,Dec,Time,Intensity) in the Subject-line of an e-mail message. This is suitable for some pagers and cell-phones which only display the From-line and Subject-line of the e-mail (i.e. they do NOT have the ability to display the body of the e-mail). The RA,Dec coordinates are J2000 Epoch. (Examples of the SubjHHMM-only messages that are distributed.) This reduced information message is suitable for those people/sites with reduced-capability pagers and cell-phones (e.g.GSM Information Network in Europe).
Currently, there are several criteria applied to each GRB Location Notice
to determine if that notice should be sent to a particular site.
These distribution criteria
are independent of the distribution method (sockets, e-mail, pager, etc).
1) The first criterion is the source of the GRB location information.
The sources of Notices are HETE, INTEGRAL, IPN, RXTE-PCA, RXTE-ASM,
BeppoSAX-WFC/-NFI, & ALEXIS).
There are dis/enable flags bits for each source type for each site in the
"sites.cfg" file.
Sites can elect to receive whichever of the sources they choose.
Within some notice types there are also subtypes. There are dis/enable flags
for each subtype as well.
2) The visibility criteria.
2a) The use of the ALL filter, so that no matter what the visibibilty of the
RA,Dec location is with respect to the site's location, it gets the Notice.
2b) For the VIS fitler, the notice position must be visible from the site
(i.e. take the RA,Dec,Lon,Lat,current_UT to yield an Az,El and test to see if El is
greater than +10 degrees above the site's horizon).
2c) For the use of the NIGHT filter,
it must be night at the site (i.e. the Sun has an Elevation < -6 degs).
3) The size of the error box for the burst (ie the location uncertainty.
If the the error is larger than X.X degrees, the site does not get the Notice.
4) The amount of eleapsed time from the start of the burst to the time
when the Notice is available for distribution. If this time is greater
than Y.Y hours, then the site does not get the Notice.
5) There is a "brightness" criterion -- if the initial 1-2sec time
sample rate data from the burst is greater than X above background then send it.
The value of X is chosen by the each site.
Some sites choose an intensity threshold of 0 (i.e. everything) and
some increase
to a value for which the bursts greater than that intensity have a position
accuracy commensurate with their FOV.
6) There is a "trigger type" criteria.
The GCN system can identify the type of BATSE trigger
(e.g. real GRB, electron precipitation event, solar flare, SAA entry/exit, etc).
It is 98% accurate.
For the other sources of Notices, there have already been selection
criteria applied, so they are always locations for true GRBs (or extreme-UV
transients in the case of ALEXIS).
Please see the technical details of these
filtering functions.
There are several motivations for applying one or more of these filtering functions to determine if a site is sent a particular notification. Obviously, the use of the VIS filter function is appropriate since it is not possible for sites to observe RA,Dec locations that are below their horizon. Optical instruments can elect to apply the NIGHT filter function, so as to not get "bothered" by notices during the day when they cannot observe. Even some radio sites elect to use this NIGHT filter because the ionosphere becomes noisy during the daylight hours in some radio bandpasses. The NIGHT filter has the VIS filter included. Sites not wishing to restrict their notices by either the VIS or NIGHT filters can elect to use the ALL filter (they get everything after any TRIGGER_ID or INTENSITY threshold is applied).
Original Notices:
Currently, the algorithm used to calculate the GRB direction assumes "ideal response" physics for the LADs (i.e. no correction for non-cos(theta) and earth scattering). And since only the first 1.024 or 2.048 second samples of the light curve rates are used to calculate the location, the error in the Original locations is generally dominated by counting statistics (not systematics). Table 2 shows the radii of circles which contain the Original location from the location calculated by the BATSE instrument team. They have the luxury of time, cpu power, better deconvolution algorithms, and human brains in the loop to (1) select good data sets, (2) to select clean backgrounds, (3) calculate positions. As the "initial intensity" of the burst increases the error decreases. Two different error radii are given, one describes the circle that contains 68% of the locations and one for 90% containment. Click here for a plain text version of Table 2.
Also, listed in the table are the rate of occurence of triggers versus intensity. These are the GRB occurance rates (as a function of initial intensity) that GCN is able to calculate a location solution (i.e. the "program efficiency"). To calculate the rate that a given site is notified, these rates must be multiplied by the appropriate efficiency factors of the site's selected filters.
TABLE 2: GCN/BATSE LOCATION ACCURACY for the Original & MAXBC Notices Initial Intensity
[cnts]Error(90%)
[deg]Error(68%)
[deg]Occurance Rate
[num/mo]1000 24 12 8.6 2000 18 9 4.8 3000 14 7 2.9 4000 12 6 2.0 5000 n/a n/a 1.1 10000 n/a n/a 0.2
Final Notices:
With more counting statistics than the Original Notices due to integrating the light curve beyond the first 1-2 sec up to a total of 32 seconds, there is slight improvement in the location uncertainty (i.e. it is somewhat systematics dominated).
MAXBC Notices:
Again, the "ideal physics" detector response functions are used on the 16 MAXBC rate data, and the location uncertainty is essentially the same as for the above Original type of GRB location notices.
LOCBURST Notices:
In general, the location uncertainty
of these locations will be about 3x better than the Original/Final/MAXBC
Notices.
The statistical error will range from 0.1 to 2.0 deg (for the brightest
90% of the GRBs) plus a systematic error of 2 deg
(to be added in quadrature with the statistical).
The statistical error as a function of intensity is
shown in the following figure (M.Kippen,MSFC).
And the percentage of all LOCBURSTs that have a Total_error less than X-degrees is listed in the Table 2a.
TABLE 2a: FRACTION OF BURSTS vs TOTAL ERROR Fraction of
LOCBURSTsTotal 1-sigma Error
Radius Limit [deg]100% 5.4 90% 2.8 75% 2.3 50% 2.1
In addition to the basic service of distributing GRB coordinates, we also provide follow-up support activities for the groups doing the follow-up observations. Groups that have made follow-up observations can contact us for the more accurate locations that are available well after the burst. These locations come from: (1) the Huntsville BATSE team. Several days after the burst they have access to the full data set (the finer time sampling, the better spectroscopy data, etc) and have the luxury of having humans-in-the-loop for the analysis effort. They also have the full-blown location algorithm with the non-cos(theta) and scattering corrections. (2) The Interplanetary Network (IPN), also contributes to about 15% of the bursts seen by BATSE. The IPN annuli are important because they greatly reduce the amount of sky that must be analyzed looking for a transient. (3) COMPTEL and EGRET can provide error boxes that are much smaller than BATSE, and again this reduces the amount of analysis a site must perform. (4) And on rare occasions, the WATCH instruments provide locations. When more accurate locations become available, we provide these when asked.
Since the start of operations in June 1993, there have been many improvements to the system. Here is a chronology and description of these improvements.
Future sources of GRB locations: To maximize the utility of the GCN system and to minimize the effort by the various follow-up groups, GCN is persuing an agressive campaign to include any and all sources of GRB locations within the GCN system. As each new spacecraft/instrument comes on-line, GCN will do whatever modifications are necessary to incorporate those GRB locations and distribute them to any and all who wish them. (And like the extreme-UV transients detected by the ALEXIS spacecraft, GCN will also expand its operations to include any astrophysical transient (GRB & non-GRB alike) information.)
The list of future missions that will be distributing their GRB locations includes: Fermi and anything new that comes along.
If you want all the gory details and specs on the various data formats and operations, then click here (about 8 pages).
COMPTEL Data Transfer: The same hardware and software that captures and processes the BATSE instrument data also captures the CGRO-COMPTEL instrument data when there is a GRB in the COMPTEL FOV. Using a ring buffer, with 120 seconds of pre-trigger data and 230 seconds of post-trigger, the data are captured and automatically ftp-ed to a COMPTEL computer at U. of New Hampshire. There, the data are automatically analyzed and if the GRB was bright enough for the COMPTEL instrument, a location is available 10-15 minutes after the GRB. This is particularly useful because the COMPTEL error boxes (1-3 deg) are smaller than the BATSE-Original/-Final/-MAXBC locations. They are comparable in size to the BATSE-LOCBURST locations and provide and independent check on the LOCBURST algorithm. This rapid analysis program and the UNH COMPTEL GRB webpage are managed by Alanna Conners (UNH). The COMPTEL Rapid Response Network (RRN) distributes these locations and is managed by Bernie McNamara (NMSU). No longer available since the de-orbit of CGRO.
BATSE Data Transfer: Since the GCN hardware and software captures (and processes) the BATSE instrument data, it can send it to the BATSE team in Huntsville, where they can process it in near-real time (20-25 min) with their locburst algorithm to produce GRB locations with improved error boxes. This procedure (see LOCBURST Notices is only performed for the brighter bursts (>2000 c/s). A ring buffer, with 143 seconds of pre-trigger data and 369 seconds of post-trigger, is used to captured the data. After fill the ring-buffer it is automatically ftp-ed to a BATSE computer at MSFC. No longer available since the de-orbit of CGRO.
Combined BATSE Position & IPN Annuli: The GCN system captures and automatically transfers to Kevin Hurley (UC Berkeley) the BATSE light curve for each GRB. This gets the appropriate data to UCB a couple days before the Huntsville BATSE team could do so, thus eliminating a couple days from the total time delay. At Berkeley a daemon program monitors for incoming light curves, and upon receipt waits for the appropriate time segment of data to be downlinked from the Ulysses spacecraft via the DSN. When the Ulysses data has been transferred from JPL, it is searched for a matching GRB and when a significant enough match is found, an IPN annulus is calculated and sent back to GSFC. This annulus is combined with the BATSE-Original/-Final/-MAXBC/-LOCBURST error circle and an a GCN/IPN_SEGMENT Notice is composed and sent to all GCN sites (this is the human-in-the-loop part). No longer available since the de-orbit of CGRO.
Heads-up: GCN also provides a "heads-up" service to the EGRET, OSSE, ALEXIS, TGRS, IPN, COMPTEL-RRN, SAX, and several optical & radio sites so that they can perform quick-look analysis of their instruments' data and/or be ready to observe when smaller error boxes become available at a later time (T+hours). EGRET has detected 5 GRBs and their error boxes are quite small (1-3 deg) (Dingus, 1995). This heads-up notification helps the EGRET team analyze their data in a timely fashion.
Currently, there are two parts of these GCN web pages that are updated automatically in real-time, so that the user community can have the latest information about what is happening with the GRBs and the operations of the GCN system. There is, at most, a 60-sec delay between when the Notices are distributed to the world community and when the web pages are updated. And since these web pages are updated automatically, people should remember to use the reload command on their web browsers, so that they see the current version of the various pages.
Real-time Coordinates Pages:
Currently Active
a) There is page which contains all the Notices from the
RXTE spacecraft (even the ones not detected by CGRO-BATSE).
It is called the RXTE GRB Locations page.
This contains both the ALERT and POSITION notices
from the RXTE-PCA and -ASM instruments.
b) The SWIFT Triggers and Locations page.
All of the latest information about the GRB locations from all 3 Swift instruments
is listed here.
c) The INTEGRAL Triggers and Locations page.
All of the latest information about the GRB locations from the INTEGRAL instruments
is listed here.
Soon to be Active
a) The AGILE Locations page.
b) The FERMI Triggers and Locations page. No longer Active
a) The first is the BATSE Triggers and Locations page. All of the latest and greatest information about the GRB locations from BATSE triggers is listed here. Only the BATSE Triggers are listed (Original, Final, MAXBC, LOCBURST, Light Curves, and ADMIN Reports). And if any of the BATSE GRBs are detected by other instruments (e.g. RXTE-PCA/-ASM, COMPTEL, and the IPN) these locations are also listed.
b) There is page with all the Notices from the COMPTEL instrument on CGRO (even the ones not detected by BATSE). It is called the COMPTEL GRB Locations page. This contains both the Detection (position) and non-detection notices.
c) There is page which contains all the Notices from the BeppoSAX spacecraft. It is called the SAX GRB Locations page. This contains the POSITION Notices from both the SAX-WFC and -NFI instruments.
d) There is page which contains all the Notices from the HETE spacecraft. It is called the HETE GRB Locations page. This contains the POSITION Notices from both the HETE-WXM and -SXC instruments.
e) There is also a real-time page of the ALEXIS Notices and Locations. It is called the ALEXIS UV Transients Locations page.
Partly Real-time Special_Burst Information Page:
The selected burst page contains links to all the information there is to-date on specially selected GRBs. This information includes: Identification (date, time, trigger number, etc), Locations (RA,Dec, radii, annuli, maps, etc), Lightcurves, and the archive of the GCN Circulars follow-up observation notices (optical, radio, etc). The main criteria that makes a burst "special" to be included in this page is that follow-up observations have been made. The 'partially' real-time designation means that some of the information in the pages is updated in real-time and some of it is done manually (usually within minutes to hours).
Real-time Sites.cfg File Page:
The second is the CGN sites.cfg file. This is a copy of the currently used "sites.cfg" file. This allows sites to monitor their configuration entries and verify that they are as requested.
A list of GCN related acronyms and terms.
A list of Frequently Asked Questions.
The GCN contact is: Scott Barthelmy,
scott@lheamail.gsfc.nasa.gov,
(301)-286-3106
This file was last modified on 27-Aug-08.