A soil layer on a hillslope acts as a quasi-steady state system generating rainfall-stormflow responses, which are controlled by pressure propagation in it after the wet zones become mutually connected as a hydraulic continuum given sufficient rainfall volume. That continuum can be simulated by a simple model consisting of a tank with a drainage hole. However, flow paths in the continuum comprising vertical unsaturated flow and downslope groundwater flow are characterized by large heterogeneity and diversity. This contrast might derive from an evolutionary process of the soil layer on a steep hillslope subject to intense tectonic activities. An aspect of the stormflow responses from the pressure propagation might impose a paradigm shift of the stormflow prediction strategy.
We specifically examined the leaky coaxial cable (LCX) of a cable-like communication antenna which is already installed widely in Japan, and proposed a new technique that enables monitoring of heavy rains near the ground. We conducted digital signal processing for detecting rainfall from signals of the electric field fluctuation around LCX. Signals were contaminated by high-level electric noise from various sources outdoors, making it difficult to detect rainfall with 1 min rainfall intensity less than 30 mm/h from the noisy signals using the conventional noise-filtering algorithm based on frequency filtering and statistical filtering. As described herein, we noted that the time variation of the rainfall intensity has multi-fractality. We tried detecting rainfall by the multi-fractal analysis of the signals. Dividing the sig-nals into sections for 15 min, we quantitatively evaluated the multi-fractality of each section based on wavelet transform modulus maxima (WTMM) method. Results show that the multi-fractality of the sec-tions with rainfall tend to be stronger than that without rainfall. Results also show that rainfall with inten-sity less than 30 mm/h can be detected from signals of the electric field fluctuation of LCX.