Similarity of the temporal pattern of soil moisture across soil profile in karst catchments of southwestern China

Knowledge of dynamics of profile soil moisture is of importance for soil moisture prediction and modeling. However, information on the relationships between soil moisture at the surface and subsurface layers is relatively inadequate. The objective of this study was to examine the similarities in the overall and scale specific temporal patterns of soil moisture across soil profile. Soil moisture was measured at five soil depths using time-domain reflectometry in two sampling plots over 242 days in a karst depression. Spearman rank correlation and wavelet coherency analyzes were used to determine overall and scale-dependent similarities of temporal patterns of soil moisture, respectively. Results showed that soil moisture in farmland was higher than grassland especially in 10-50 cm depths. The surface depth had higher variability for both farmland and grassland. Spearman rank correlation coefficients between any two soil depths were significant, indicating strong depth persistence in the overall temporal pattern of soil moisture. The similarity decreased as the distance between depths increased. Wavelet coherency analysis indicated that soil moisture temporal pattern had a higher depth persistence in the farmland than grassland. Soil depth had a larger influence than land uses on the similarity in soil moisture temporal pattern, and the relative humidity and precipitation had larger effect than air temperature controlling the change of soil moisture at larger scales. These scale specific similarity in the temporal pattern of soil moisture can be used to guide estimating deeper soil moisture and to improve sampling strategies for soil moisture prediction with a higher prediction accuracy.