Assessing gully erosion susceptibility dynamics using information value and hazard index methods: A case study of Agulu-Nanka watershed, Southeast Nigeria

Abstract

Gully erosion is one of the major causes of land degradation leading to the loss of productive lands globally. The Benue Trough in Nigeria is plagued with the problem of gully erosion, with a myriad of natural and anthropogenic factors considered as contributory factors. In this study, quantitative and qualitative methods were combined to evaluate gully susceptibility in the Agulu-Nanka watershed located in the lower part of Benue Trough between 2000 and 2018. Landsat 7 and 8 images, Digital Elevation Model (DEM), geological, soil, and rainfall erosivity maps were used to generate the gully erosion controlling factors including vegetation cover, lithology, soil, land use/cover, slope, curvature, aspect and rainfall erosivity. Additionally, high resolution image was used to develop gully inventory map. The information value method was used evaluate the influence of various classes of gully erosion controlling factors on gully distribution in the watershed. This enabled objective derivation of weights for the various classes of these factors. Information value analysis reveal that low percentage vegetation cover (<10 %), Ameki Group lithological unit, high slope gradient (17 − 43°) and Dystric Fluvisols soil type have the highest influence on gully erosion in the watershed. Since the information value method does not consider interrelationship between factors, the rank sum technique was used to derive factor weights based on the relationship between the factors. To develop the gully susceptibility map for the Agulu-Nanka watershed for the years 2000 and 2018, the erosion controlling factors were integrated using factor class weights derived through information value and factor weights derived through rank sum method. The analysis of gully erosion hazard maps show that approximately 48 % of the watershed was under low erosion risk over the period of 18 years. Areas under moderate, high and severe erosion risk decreased from 27.45 %, 9.45 % and 0.27 % of the watershed in year 2000, to 21.30 %, 8.41 %, and 0.20 % of the watershed respectively in 2018. On the other hand, the areas under negligible erosion risk increased from 14.69 % in year 2000, to 21.26 % of the watershed in 2018. The validation of the gully erosion susceptibility map for years 2000 and 2018 using the Receiver Operating Curve (ROC) showed ROC-AUC values of 0.81 % and 0.88 % respectively suggesting good model results. While the result of study indicates that restoration efforts through afforestation is yielding some result, the minimal reduction in the size of areas under severe and high gully erosion susceptibilities indicate that additional mitigation measures need to be combined with intensified afforestation efforts to effectively combat gully erosion in the watershed.