NRAO Home  >  Green Bank  |  Wiki Topic:    GB > Data > PlanOfRecordC42004 > StandardObservingModes > HowToReducePointingData

Changes | Index | Contents | Search | Statistics | Jump to Topic

Data Reduction Case #9

Pointing Data

Pointing Data is reduced in the following steps:

1. Calibrate the total power data
2. Average the polarizations (optional)
3. Flag any data if necessary
4. Subtract the second beam (which records atmospheric contributions) if applicable
5. Fit a Gaussian to the source profile in each scan
6. Apply the measured offset as the new LPC

Pointing data is taken in continuum total power mode. Each scan can first be calibrated to determine antenna temperature as a function of position on the sky, using the standard calibration:

Here, p is an index for each of the two polarizations, and x is the position of the telescope.

The polarizations may either be averaged, or handled separately.

There will be an antenna temperature measurement for each scan. If there are two beams, the antenna temperature of the secondary beam should be subtracted from that of the first for each scan individually, thereby removing any atmospheric fluctuations in the data. If the second beam passed through the source of interest, as may happen in the Az scans, then this must be considered at the time a Gaussian is fit (see below).

At this point the observer has data which should show a baseline with the Gaussian profile of a source roughly centered in the scan. There may be a residual baseline even after subtracting the atmosphere. Some fraction of the data must be marked to fit a baseline. Which part of the scan are used for the baseline depends on how the data were taken, but it should be selected to exclude the source itself, and it should probably exclude the edges of the scan. The user can then fit a polynomial function to the baseline region, subtract the polynomial fit, and fit a Gaussian function to the source. These two functions (polynomial and Gaussian) may also be fit simultaneously if it is convenient.

Because the wings of the source profile are not really Gaussian, the observer may wish to exclude data with values less than 80% or even 50% of the peak intensity when doing the Gaussian fit.

In multibeam systems, after the reference beam has been subtracted there may be a negative image of the source at a known separation from the center. For example, this happens in GBT K-band data in the Az scans, and the separation is 3'. The negative image can either be simultaneously fit with a second Gaussian, or it can be flagged so as not to interfere with the baseline/Gaussian fit. In the former case, the known separation of the beams should be used as a constraint in order to maximize the utility of the fit.

The pointing offset is the difference between the measured and known positions of the calibration source. The offset should be averaged between like scans, i.e. if there are two azimuth scans, the offsets from those two should be averaged. Then the result is sent back to the M&C system.

If a back and forth motion is used, then the order of operations for the pointing corrections procedure is:

1. Scan in Az, forward
2. Scan in Az, back
3. Apply the averaged Az correction
4. Scan in El, forward
5. Scan in El, back
6. Apply the averaged El correction

Otherwise, it is

1. Scan in Az
2. Apply the Az correction
3. Scan in El
4. Apply the El correction

The application of the Az correction in the middle of the procedure, rather than at the end, ensures that the telescope moves through the center of the source for the El scans. This is particularly important if these scans are to be used for flux calibration purposes, but it also helps get the best S/N so improves the overall fidelity of the measurement.

-- JimBraatz - 04 Aug 2004