The picture above shows typical PMSE (polar mesosphere summer echoes) measured by ESRAD. These very strong radar echoes appear in summer, between about May and August. They are most persistent in June and July (example in the upper panel) and more sporadic in May and August (lower panel). They are caused by highly structured plasma density fluctuations concentrated in thin layers, sometimes, but not always, together with corresponding structure in the neutral density (This is known from measurements with sounding rockets - see Luebken et al., Geophys. Res. Lett. 1993) . It is not known how the plasma structuring is produced - breaking of gravity waves propagating up from below and aerosols, such as sub-visual cloud particles, are most likely be involved.
Noctilucent clouds are thin clouds which form at about the same height (about 82 km) and during about the same season as PMSE. They are seen from the ground when the sun is just below the horizon - they are then lit from underneath and the light reflected at a low angle picks out the many waves which cross the cloud layer. The clouds form because temperatures are extremely low (as low as -170 C) at these heights in summer. Noctilucent cloud observations are regularly collected by networks of amateur observers. Noctilucent clouds seen together with the PMSE in the lower panel above, are shown below.
Noctilucent clouds seem to be increasing in occurrence (Gadsden, Journal of Atmospheric and Terrestrial Physics vol 52, p247, 1990) and some scientists believe this may be due to increased greenhouse gases in the atmosphere. The clouds also show an 11-year variation roughly opposite in phase to the Sun\'s activity cycle, ie thay are less often seen at solar maximum. It is not yet known how closely PMSE are related to noctilucent clouds (sometimes they have been seen together, sometimes not) , nor whether PMSE also vary with the solar cycle nor whether they are increasing on a longer time scale.
Long-term studies of PMSE and their connection (or not) to noctilucent clouds are being made with the Esrange MST radar, cameras to record the clouds, and the University of Bonn lidar which has been located at Esrange from late 1996.
The first questions to be addressed are the reasons for short-term (daily an multi-day) variability in PMSE (see our publications). Daily and multi-day variation of PMSE occurrence at Esrange, measured during 1997 and 1998, are illustrated below.
The figure above shows how PMSE vary during the day - they are seldom present between 20-23 UT, and are more common at higher altitudes after noon, compared to the hours before noon (local solar noon is at about 10:40 UT).
The figure above shows how PMSE vary during the season. The upper panel shows the percentage of hours each day when the ESRAD radar detected PMSE in 1998 (the blue and magenta lines corespond to measurements made with different radar modes) - there are strong fluctuations with periods of a few days superimposed on the seasonal growth and decay . The fluctuations with periods 4-6 days are selected by filtering and shown in the lower panel - during much of the time these are found to anti-correlate with temperature fluctuations in 5-day planetary waves detected in meteorological data.