Surface temperature at a volcanic system is the result of a mixture of two heat sources originating one from above and another from below the ground surface. The atmosphere significantly affects the surface's temperature of the volcano, and removal of the atmospheric effect is required for meaningful interpretations prior to any characterization of anomalous volcanic activity. Time series of meteorological data were acquired from the closest weather station to Nisyros volcano, and surface volcanic temperature from the southern part of Nisyros caldera were analyzed. Time series analysis took place at different time scales through a decomposition method. The decomposition of the time series into long-, seasonal-, and short-term components was performed by the application of the Kolmogorov-Zurbenko (KZ) filter to separate the noise that appears on the obtained data. The Time Series Regression (TSR) model was applied in all time scales. The long- and the short-term components of the surface volcanic temperature show no significant correlation with the atmospheric variables, while the seasonal term indicates a very strong association with an R2 equal with 99%. After the removal of the atmospheric effect, our model indicates that there is a normal activity of the seasonal heat emission in Nisyros with no anomalous signals in the temperature residuals during our survey. This methodology may facilitate the process of filtering a seasonal atmospheric effect and better comprehend the heat emission of a volcano-hydrothermal system from temperature surface readings.