Radar measurement of ionospheric scintillation in the polar region

Abstract

AbstractThis paper considers several estimators that use radar data to measure the S4 scintillation index that characterizes the severity of amplitude scintillation that may occur during RF propagation through ionospheric irregularities. S4 is defined to be the standard deviation of the fluctuations in received power normalized by division by the mean power. Estimates of S4 are based on radar returns obtained during track of targets which may themselves have intrinsic radar cross‐section fluctuations. Key to this work is the consideration of thresholding, which is used in many radars to remove (from further processing) signals whose SNR is considered too low. We consider several estimators here. The “direct” estimator attempts to estimate S4 through the direct calculation of the mean and standard deviation of the SNR from a number of radar returns. The maximum likelihood (ML) estimator uses multiple hypothesis testing and the assumption of Nakagami‐m statistics to estimate the scintillation index that best fits the radar returns from some number of pulses. The ML estimator has perfect knowledge of the number of radar returns that are below the threshold. The direct estimator is accurate for the case where there is no threshold and there are many returns or samples from which to estimate S4. However, the direct estimator is flawed (especially for strong scintillation) if deep fades that fall below the radar threshold are ignored. The modified ML estimator here is based on the ML technique but is useful if the count of missed returns is unavailable. We apply the modified ML estimator to several years of radar tracks of large calibration satellites to obtain the statistics of UHF scintillation as viewed from the early warning radar at Thule, Greenland. One‐way S4 was measured from 5000 low Earth orbit tracks during the 3 year period after solar maximum in May 2000. The data are analyzed to quantify the exceedance or the level of scintillation experienced at various probabilities. For example, from Thule during the period from December 2001 to February 2002, the probability exceeds 5% that the one‐way scintillation index at UHF is greater than 0.7. On a daily basis this translates into 1.2 h a day of strong scintillation. Unfortunately, there are insufficient data to obtain equivalent information for the winter of solar maximum when the scintillation is likely to be most severe.