Investigation of the possible relation between precipitation and cosmic ray flux using wavelet analysis

Abstract

The galactic cosmic rays (GCR) are charged particles, primarily protons and helium nuclei. In the present work, by applying data analysis techniques, we have investigated the proposal that GCR can in uence precipitation. Wavelet transforms have been applied for the rst time to study the problem. This is because the wavelet is the best tool available today to analyse non-stationary and non-linear time series. The method suggests a common 9:8 year cycle in the cosmic ray ux (CRF) and precipitation time series during the considered period [1979-2008] at Thule(76.5 N). Finally, the point process method on wavelet maps has been applied at Thule. The method depicts that CRF leads the precipitation by about 2 and 9 months at Thule. For the present analysis, six geographically diverse locations across the globe have been selected. The six stations are Thule (76.5 N), Climax (39.4 N), Huancayo (12 S), Namibia (19.12 S), Potchefstroom (26.4 S) and Hermanus (34.25 S); all of them are neutron monitor stations. For the present study, precipitation data have been retrieved from Global Precipitation Climatology Project (GPCP). GPCP blends estimates based on various data sources to produce a global gridded data set, taking advantage of the strengths of each data type. A reliable set of cosmic ray ux data at the six neutron monitor stations mentioned above is available from National Geophysical Data Centre (NGDC). For cloud cover, the data has been extracted from the International Satellite Cloud Climatology Project (ISCCP) archives. In the present analysis, the correlation between the GCR ux and precipitation has been found to be signi cantly high at higher latitudes. However, no signi cant correlation between low cloud cover(LCC) and GCR was detected at any of the six stations. The LCC data was taken from ISCCP. The absence of correlation between the GCR and LCC has been attributed by Svensmark(2003) to calibration problems. Thus, we are not in a position to come to a de nite conclusion about the absence or presence of correlation between LCC and GCR. Using the Fourier power spectrum, a signi cant 10 year period has been detected in both cosmic ray ux and precipitation at Thule, Potchefstroom and Hermanus. Such a cycle is absent in the precipitation data at Huancayo and Namibia, in the tropics. One of the ways to explain the physical mechanism underlying a GCR-precipitation connection can be stated as follows. GCR being charged particles a ects the earth's global electrical circuit, and may thus stimulate the formation of charged cloud con- densation nuclei (CCN) in the atmosphere. Therefore, higher GCR uxes would lead to more charged CCN in the atmosphere. It has been noticed1 that charged CCN are more capable of attracting neutral or oppositely charged ambient parti- cles in the atmosphere. The higher charged-CCN concentrations at times of high GCR uxes would lead to increased ice nucleation which eventually enhances the precipitation release in cold clouds2. This mechanism is expected to be stronger at the higher latitudes due to low geomagnetic cuto value. Thus, the concentration of GCR is higher at higher latitudes, as in case of Thule, where cold clouds are common and ice nucleation in cold clouds helps in precipitation. Thus the higher correlation coe cient and the presence of a common 9.8 year cycle in both CRF and precipitation time series at Thule, using power spectrum and wavelet power spectrum method, supports the above mechanism. Kniveton and Todd in 2001 proposed a relation between the CRF, precipitation and precipitation e ciency. Using data from 1979 to 1999 they found evidence of statistically strong relationships between the three variables over ocean surfaces at mid to high latitudes. Our work con rms their conclusion. In addition, we found non-stationrity and non-linearity in the precipitation time series, and used the wavelet transform in our analysis. Moreover, the 9:8 year cycle in precipitation and CRF time series has been found. Here, we nd that CRF leads precipitation by 2 months at Thule.