*Physics and Chemistry of the Earth, 27*,
589-594, 2002.

## Methodological influences on F-region peak
height trend analyses

### M.J. Jarvis^{1}, M.A. Clilverd^{1}, and
Th.Ulich^{2}

^{1}British Antarctic Survey, Cambridge, England, U.K.

^{2}Sodankyla Geophysical Observatory, Sodankyla, Finland.

**Abstract**

Published estimates of the trend in hmF2 using data from ionosondes
over the last 30-40 years range from +0.8 km y^{-1} to -0.6 km
y^{-1} and are subject to the influence of several factors. These
are considered here based upon an analysis of two southern hemisphere
geomagnetically mid-latitude stations, Argentine Islands and Port Stanley.
The influence of the equation used to calculate hmF2 at these stations can
result in variations of +/- 0.2 km y^{-1}; choice of solar proxy
has a small influence on the end result, where using E10.7 instead of
F10.7 produces changes of -0.04 km y^{-1}; neglecting any trends
in geomagnetic activity can produce variations of +0.03 to +0.2 km
y^{-1} at the two mid-latitude stations considered in this paper;
for datasets of 30-40 years length ringing due to long memory processes
can produce +/- 0.2 km y^{-1} variability; the phase of the
11-year solar cycle, and its harmonics, captured by the datasets can cause
variability of +/- 0.5 km y^{-1}; and the neglect of local time
variations in thermospheric wind conditions could result in +0.2 km
y^{-1} for analysis which only considers local midday data. The
Argentine Islands and Port Stanley datasets show ringing terms that are
still converging towards trend results of -0.25 to -0.30 km
y^{-1}, which are in close agreement with the satellite drag trend
estimates.