The INTEGRAL spectrometer SPI: performance of point-source data analysis

The performance of the SPI point-source data analysis system is assessed using a combination of simulations and of observations gathered during the first year of INTEGRAL operations. External error estimates are derived by comparing source positions and fluxes obtained from independent analyses. When the source detection significance provided by the SPIROS imaging reconstruction program increases from ∼10 to ∼100, the errors decrease as the inverse of the detection significance, with values from ∼10 to ∼1 arcmin in positions, and from ∼10 to ∼1 per cent in relative flux. These errors are dominated by Poisson counting noise. Our error estimates are consistent with those provided by the SPIROS program. With higher detection significance, the accuracy is ultimately limited to ∼0.5 arcmin in position and ∼1 per cent in relative flux by other types of errors. In these cases, SPIROS underestimates the true errors as it only takes into account the Poisson counting noise. At the low signal-to-noise ratio end, SPIROS is also too optimistic, and the number of spurious detections is significantly higher than would be expected from statistical noise fluctuations only. The analysis results do not depend significantly on the target off-axis angle, or on the number of pointings considered, provided that this number is larger than ∼15. Realistic source confusion tests are carried out by adding simulated data to the observation of the Crab nebula and pulsar. Reliable flux values can be obtained for sources separated by angles as small as 0.5 deg if their positions are known and kept fixed in the data deconvolution process. However, when SPIROS is searching for best source positions, if two sources are separated by less than ∼2 deg, it may only find a single spurious source located between the real ones (with a flux equal to the sum of the fluxes). Finally, users of the SPIROS program can find a number of important recommendations in the conclusion