We began this paper with an example of STARE data which demonstrated how effective VHF coherent backscatter radars were at imaging magnetospheric substorms. Similar can be achieved at HF, but at substorm onset there is often a loss in the backscatter signal. This either occurs because there is a loss of the scattering targets (a conductivity increase results in the decrease in the electric field and a quenching of the instability which generates the plasma irregularities) or the HF rays become unable to reach the scattering region (e.g. D-region absorption or a change in the propagation path). Milan et al. (1999) has described these mechanisms in more detail. It is difficult to distinguish between which mechanism is most significant from case to case. Ideally information is required about the electron density along the ray path and in the scattering region. Milan et al. (1999) and Yeoman et al. (2000) achieved this in particular case studies by utilising EISCAT data.
The present authors take advantage of the STARE radar which runs on a continual basis and thereby provides a reference for CUTLASS observations. Interpretation of the STARE radar data is not straightforward, but an intercomparison of radar signatures at different wavelengths performed over a large data base may help bring new inferences to understand the behaviour of one or both radar systems in response to substorm activity in the field of view.
The STARE Finland and the CUTLASS Finland radar are both located near the Finnish town of Hankasalmi and the STARE Finland radar field-of-view comprises a sub-region of the CUTLASS Finland field-of-view. This presents a setup for the intercomparison of radar substorm features at HF and VHF. This is a study which has been initiated by Hager (2000). Data from the interval June 1997 to June 1999 was studied. Over 300 substorms were identified, from this a initial sample of 60 cases were taken to look at the radar data from both STARE and CUTLASS more closely.
The combined radar data set presents many different possibilities for
research. The initial question addressed was the prevalence of radar
scatter loss at onset. This is a phenomenon which chiefly affects the
HF radar data. Of the 60 cases examined, 48 exhibited a loss of scatter
associated with the substorm, in 28 of these the loss of scatter occurred
appreciably before substorm onset. Of the 12 other cases, there was either
no appreciable change in the scatter characteristics or the appearance of
new scatter is observed. The creation of scatter only occurred for cases where
the substorm onset timing was greater than an hour before local magnetic
midnight. For the 60 examples, STARE would typically observe scatter
after onset and in an overlapping region to the CUTLASS Finland field of
view of approximately 100 km in range.
In Figure 5 the same interval is plotted but with the addition of all occurrences of groundscatter. During the first loss in scatter during the growth phase an extensive region of groundscatter is present. Directly at onset this groundscatter disappears to be replaced by a very high power region of ground scatter near 1000 km range.
A different beam of the SuperDARN data is plotted in Figure 6 (beam 9) the
gross features are similar. This beam lies along the image meridian and
therefore provides a more straightforward comparison to the IMAGE data
(middle panel). For this interval only the STARE Norwegian radar was operative
and data from beam 4 is plotted in the lowest part of the Figure.
The STARE - CUTLASS comparison illustrates the anticorrelation between scatter at onset. The fields of view overlap partially, albeit the CUTLASS observations are F-region and the STARE are E-region. However, the presence of STARE data indicates that there is still appreciable electric field and so conditions to produce F-region irregularities persist. This is the situation observed in the majority of cases. Some cases do exist when E-region scatter is also lost. This must indicate very strong precipitation which increases the E-region conductivity and the Electric fields are reduced to levels below the threshold of excitation. - see Yeoman et al. (2000). These are not the most typical cases. The example here indicates the loss in scatter is because the HF rays no longer reach the scattering region, either absorption or propagation path altered. The observation of ground-scatter indicates there is not complete absorption. Intense regions of ground scatter (e.g. the patch at onset at 1000 km range) help to pin point (with further analysis) the location of precipitation at substorm onset. Using a pair of SuperDARN radars looking at the same substorm also helps to differentiate these cases.