O-O COLLISION CROSS-SECTION AND LONG-TERM F-REGION O DENSITY VARIATIONS DEDUCED FROM THE IONOSPHERIC ENERGY BUDGET()

Recent estimates of the O+-O collision cross section Q(O+-O) have sugg ested that the commonly used value of Banks [1966b] is too low by a fa ctor of 1.2-1.9. On the other hand, past deductions of the F region ne utral atomic oxygen density [O] from incoherent scatter (IS) radar stu dies have been quite compatible with the original Banks value, and any serious adjustment of that value would cause serious disagreement bet ween these radar [O] results and those of the mass spectrometer incohe rent scatter (MSIS) model. We have derived the mean daytime value for the product of Q(O+-O) and [O] for 153 days during the period 1970-197 5 from Millstone Hill IS radar measurements of ionospheric temperature and density. If we assume that MSIS model [O] is correct on average o ver this period, we find that Q(O+-O) is only 75% of the Banks value a nd 40 to 60% lower than other values recently proposed. We show that a ll previous derivations of [O] from IS radar energy-budget studies are compatible with our value, and thus a discrepancy exists in Q(O+-O) d etermination by different methods. We could make our Q(O+-O) results c ompatible with the larger values noted above by allowing a neutral tem perature gradient to exist at high altitudes, where none is currently expected, or by increasing the heat input to the ion gas. We note that both of these effects could be produced by a sufficient reservoir of hot oxygen in the topside of the ionosphere. Independent of the way in which we choose a value of Q(O+-O) to make the radar and MSIS [O] val ues agree on average over our 6-year span of data, the radar record sh ows [O] to have episodic departures from MSIS predictions reaching amp litudes of up to 50% and lasting for periods of several months. This a rgues for caution in analyses which depend crucially upon the validity of MSIS densities during campaign-like experimental studies.