Premium
Toward resolving small‐scale structures in ionospheric convection from SuperDARN
Author(s) -
André R.,
Villain J.P.,
Senior C.,
Barthes L.,
Hanuise C.,
Cerisier J.C.,
Thorolfsson A.
Publication year - 1999
Publication title -
radio science
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.371
H-Index - 84
eISSN - 1944-799X
pISSN - 0048-6604
DOI - 10.1029/1999rs900044
Subject(s) - divergence (linguistics) , radar , grid , algorithm , scale (ratio) , computer science , geodesy , physics , computational physics , geology , quantum mechanics , telecommunications , philosophy , linguistics
The combination of radial velocities measured by a pair of Super Dual Auroral Radar Network (SuperDARN) HF coherent radars gives, in their common field of view, the velocity vectors in a plane perpendicular to the magnetic field. The standard merging is based on a natural grid defined by the beam intersections, which provides a resolution varying between 90 and 180 km (depending upon the distance to the radars). This allows the description of structures with a typical scale size ( L ) of the order of 500 km. The present study is devoted to a merging method which takes advantage of individual radar grids to enhance the resolution ( L ≈ 200 km). After a brief description of the standard merging method, we define the high‐resolution grid and discuss the potential problems which have to be overcome. The first problem concerns the localization of the scattering volume, whereas the second one deals with the independence of the velocity vectors. These two limitations have been addressed in previous studies [ André et al ., 1997; Barthes et al ., 1998]. In the method proposed here, several velocity vectors are determined at each grid point, from which the selection is made by using the hypothesis of minimization of the divergence magnitude. The selected map is the one which minimizes the divergence. The performances are tested and compared to the standard merging algorithm through simulated double vortices. Finally, we apply this method to real data, and show, through two examples, its ability to describe small‐scale structures ( L ≈ 200 km).