Linearity analysis on stationary segments of hydrologic time series

The rainfall-runoff process is widely perceived as being non-linear; however, the degree of non-linearity might not be significant in hydrologic time series. Evidence of non-linearity was reported in the past in the detrended time series of daily precipitation, but found not to be significant in annual series. The objective of this study is to detect non-linearity in monthly hydrologic time series by applying the Hinish tests for Gaussianity and linearity to selected stationary segments of four kinds of such series, namely, streamflow, temperature, precipitation and Palmer's drought severity index. The results indicate that all of the stationary segments of standardized monthly temperature and precipitation series are found to be either Gaussian or linear. Some of the standardized monthly streamflow and Palmer's drought severity index are found to be non-linear.