Multifractal fingerprinting of fine resolution daily gridded rainfall of Kerala meteorological subdivision, India using detrended fluctuation analysis

Abstract

This paper performs the Multiracial Detrended Fluctuation Analysis (MFDFA) of fine resolution (0.25ox0.25o) daily gridded rainfall of Kerala Meteorological subdivision, India. First the MFDFA method is applied for long term rainfall datasets of different grid points for 1901-2013 period. The results show that rainfall time series of majority of the grid points (~66 %) exhibits short term persistence with strong multifractal characteristics. Except at three locations, the asymmetry index of rainfall time series displays right skewed spectra indicating local high fluctuations and fine structured series. Further, MFDF analysis was performed after splitting the time series into two with respect to the global climatic shift of 1976/77. The results of pre and post climatic shift period shows that 34 grids (out of 65) exhibited an increase multifractal spectral width while 41 grids exhibited a increase in persistence after the shift. The study further showed that the persistence property changed from short to long term in 23 % grids, which also indicated a change in pattern of rainfall within the subdivision. The results showed that a strong correlation 0.994 exists between Hurst exponent and Holder exponent, reinforcing the universal association between the two multifractal exponents.