Signal processing and data analysis in middle atmosphere radar

Abstract

Use of signal‐processing and data analysis methods in radar studies of middle atmospheric motions is examined. Detection of weak Doppler‐shifted signals from thin atmospheric regions in a noise background requires an optimum receiver. It is noted that the optimum receiver is implemented in successive stages using matched‐filter and correlator structures for locating targets distributed in range and a filter bank for estimating their Doppler shifts. Effectiveness of coherent integration in reducing high data rates, and as an efficient but crude low‐pass filter, is discussed. Use of noiselike binary phase codes and decoding in improving range resolution is outlined. Spectral analysis of radar signals using a time‐averaged periodogram and the autoregressive model is briefly reviewed. Methods of reducing undesirable artifacts in spectral analysis of nonstationary signal components, using windows and trend removal, are delineated. Estimation of signal parameters through spectral moments and least squares estimation is summarized. Finally, simple ad hoc data‐editing methods for suppressing outliers and interference are outlined.