How the the "spill over" and other corrections were made

Spill over calibration:

• Mass spectra from the whole mission were combined and the shapes and positions of peaks for all four major ion species were determined at a wide range of countrates and at all energies for which mass spectra were available. The peak characteristics were interpolated to other energies where required.

• The fractions of each mass peak which spilled over into the higher mass windows and also the fraction of the peak which was outside its own window was estimated and tabulated into "spill over calibration tables" for all masses, energies and a representative range of "Fast Event Counter" countrates.

• The spill over is not known very precisely since it is the tail of the mass peak at the point where another mass peak is located. Interpolating to energies etc. where there were no adequate mass spectra also reduces the precision there. It is estimated that the spill over is generally known to within about +/- 50%.

• In order to correct the countrate of one ion species the countrates of all lighter species are collected. Then the fraction of each of them which would spill over into the window of the target species is determined from the "spill over calibration tables" at the specific energy and interpolated between the appropriate "Fast Event Counter" countrates. This combined spill over is then subtracted from the countrate of the target species. The background determined from the "Background Counting Rate" value is also subtracted.

• The "spill over calibration tables" are also used to determine the fraction of the peak of the target species which was outside its own window. This is then added back before converting the countrate to flux.

• Estimating the uncertainty is rather more complicated than usual. Each of the component species of the combined spill over has its associated "counting statistics" uncertainty. This is calculated in the usual way as the uncertainty of the countrate of the spilled over species scaled by the fraction which spills over. The fraction itself is known only to within +/- 50% so this must be included in the combined uncertainty. It is roughly equivalent to an additional uncertainty of +/- half of the spill over.

• Very high countrates, usually of H⁺, can saturate the MCP, reduce its gain and depress the "Processed Event Counter" value. This reduces the accuaracy of the onboard deadtime correction which uses the "Processed Event Counter" and "Fast Event Counter" values. The correct deadtime correction can be predicted theoretically from the "Fast Event Counter" value and the theoretical prediction has been substituted for the onboard deadtime.

• Magnetometer data from the MFE key parameter files were used whenever available. When they were not available the less reliable magnetometer data from the spacecraft housekeeping were used instead.