Browse/search for people

Publication - Professor Dawei Han

    Estimation of soil moisture using modified Antecedent Precipitation Index with application in landslide predictions


    Zhao, B, Dai, Q, Han, D, Dai, H, Mao, J, Zhuo, L & Rong, G, 2019, ‘Estimation of soil moisture using modified Antecedent Precipitation Index with application in landslide predictions’. Landslides.


    Soil moisture plays a key role in land-atmosphere interaction systems. Although it can be estimated through in-situ measurements, satellite remote sensing and hydrological modelling, using indicators to index soil moisture conditions is another useful way. In this study, one of these indicators, The Antecedent Precipitation Index (API) is explored. Modifications were proposed to the conventional version of API by introducing two parameters to make it more in line with the physical process. First, the recession coefficient is allowed to vary with the change of air temperature, which could take into account the variation of the evapotranspiration process. Second, the API value is restricted by the maximum value of API, accounting for the maximum water holding capacity of the soil. The modified API was then calibrated and validated by comparing with the in-situ measured soil moisture. The better correlation between these two datasets demonstrates that the modified API could better indicate soil moisture conditions, compared with the conventional API. The capability of the modified API to index soil moisture conditions was further explored by applying it to landslide predictions in the Emilia Romagna region, northern Italy. Here the recent 3-day rainfall vs the antecedent soil wetness thresholds (RS thresholds) were constructed, in which the soil wetness is indexed by the modified API. The validation of RS thresholds was carried out with the use of the contingency matrix and Receiver Operating Characteristic (ROC) curves. By comparing the prediction performance between RS thresholds and rainfall thresholds, it is found that RS threshold could provide better prediction capabilities in terms of higher hit rate and lower false alarm rate. The positive results indicate that the modified API could provide superior performance of indexing soil moisture conditions, demonstrating the effectiveness of the proposed modifications.

    Full details in the University publications repository