One important source of errors on GNSS signals is the ionospheric effect. This layer of the atmosphere is filled with charged particles. Ionospheric effects on the waves are dependent on the amount of TEC. This paper uses the least square harmonic estimation (LS-HE) that is one of the analytical methods in the frequency domain. We used the vertical TEC values obtained from GIM models with global cover provided by the JPL analysis center. We use 15 years of bihourly data gathered from the 152th day in 1998 to the first day of 2014. We first determine the important periodic signals by applying the univariate and multivariate harmonic estimate on the TEC time series. The multivariate analysis revealed the presence of daily periodic signal with its higher harmonics and annual period with its higher harmonics. We then calculate the spectral power of a number of identified signals in all available data range. The result indicate that the higher harmonics of the daily signal (tri and quad diurnal) show their maximum spectral values in the dip equator. This indicates that the earth's magnetic field is one of the cause, to these provide patterns.