Abstract The topographic wetness index (TWI) is a common factor for describing the control of the relative soil moisture spatial distribution. However, the application accuracy of the TWI remains unfavorable. Thus, this study aimed to enhance the precision of the TWI by combining the triangular facet network (TFN) with flow-path network (FPN) from a digital elevation model (DEM). First, the aspect and slope were derived from the vector TFN. Then, the FPN was designed to estimate the flow accumulation. Finally, the TWI was obtained by combining the slope and specific catchment area (SCA) which was calculated by integrating the flow accumulation with the aspect. Experiments were conducted in two respects: (1) to quantitatively assess the error in the vector method under the DEMs of 5 m, 10 m, 20 m, 30 m, 60 m, and 90 m constructed by twelve mathematical models (three groups of parameters each for four surface types: ellipsoid, inverse ellipsoid, saddle, and plane); (2) to qualitatively assess the error under two real-world DEMs of the small Black Brook watershed (BBW) and the mountainous plateau area. The experimental results demonstrated that the TWI value derived from the proposed method was generally closer to the theoretical value with the lowest root mean standard error (RMSE) and mean absolute error (MAE) than the results of other commonly used algorithms. The results of Kruskal-Wallis test demonstrated that the vector method exhibits the best accuracy and stability in TWI calculation under the study scenario. The proposed method can maintain consistent results on different surfaces at various resolutions. Moreover, the spatial distribution of the TWI calculated from the vector method has advantages.
Wu et al. (Thu,) studied this question.