In the present study, the AGST series is investigated using a powerful technique called MF-DFA .
The seasonal cycles from the raw data [T.sub.i] are removed by computing the AGST anomaly [DELTA][T.sub.i] = [T.sub.i] - [<[T.sub.i]>.sub.d], where [<[T.sub.i]>.sub.d] denotes the average value for a given calendar day.
Four representative weather, stations are selected to study temporal scaling properties for daily AGST records.
The daily AGST records at four representative geographically well-separated locations over China were recorded during the time frame ranging 1951-2009.
Figures 3(e), 3(f), 3(g), and 3(h) show the generalized Hurst exponents h(q) with varying moments (q = 5,4, 3,2,1, -1, -2, -3, -4, -5) of the daily AGST records for the four sites, respectively.
As shown in Figures 4(a), 4(b), 4(c), and 4(d), the fractal behaviors of the four AGST time series are not from broad probability density function since long-range correlations are destroyed by the shuffling procedure and the slopes [h.sub.shuf](q) of MF-DFA4, with different orders for the shuffled data over the four selected stations being the same and equal to about 0.5 as random series.