Abstract The frequency distribution for several characteristics of a solar flare obeys a power law only above a certain threshold, below which there is an apparent loss of small scale events presumably caused by limited instrumental sensitivity and the corresponding event selection bias. It is also possible that this deviation in the power law can have a physical origin in the source. We propose two fitting models incorporating a power law distribution with a low count rate cutoff plus a noise component for the frequency distribution of the hard X-ray peak count rate of all solar flare samples obtained with HXRBS/SMM and BATSE/CGRO observations. Our new fitting method produces the same power-law index as previously developed methods, a low cutoff of the power-law function and its corresponding noise level, which is consistent with measurements of the actual noise level of the hard X-ray count rate. We found that the fitted low cutoff appears to be related to the noise level, i.e., flares are only recognized when their peak count rate is 3σ greater than noise. Therefore, the fitted low cutoff, which is smaller than the aforementioned threshold, might be attributed to selection bias, and probably not to the actual count rate cutoff in flares at smaller scales. Whether or not the actual low cutoff physically exists needs to be checked by future observations with increased sensitivities.

Keywords Sun: flares — Sun: X-rays, gamma-rays — methods: statistical

null