No. 995

Title:

A time series model for air temperature anomalies

Author(s):

Wakaura, Masatsugu (The Graduate University for Advanced Studies);
Ogata, Yosihiko (The Institute of Statistical Mathematics)

Key words:

Surface air temperature anomalies; Nonstationary AR models; AIC; The Fourier form AR(FFAR) model; a high-pass filtered anomaly data; The daily power spectrum.

Abstract:

Surface air temperature anomalies relative to the seasonal variations are of our great concern from a long-term forecasting viewpoint. We introduce a particular parametric form for a nonstationary autoregressive (AR) model to quantify the anomalies by applying it to the residual data in which the means and variances of the deterministic seasonal cycles are removed from the original temperature data. This model fits significantly better than the ordinary AR model for such residual datasets taken from almost all of the stations in Japan, and still exhibits a significant seasonal structure in their autocorrelation. Then, we apply the model to a high-pass filtered data of the residuals to show the relation between the seasonal structure and a high frequency variability ranging 1-5 years in anomalies, from which we revealed a clue about the climatic system influence on anomalies of surface air temperature in Japan.