Soil–Water Conservation, Erosion and Landslide

In the wake of climate change, extreme storm events, catastrophic disasters (including soil erosion, debris and landslide formation, loss of life, etc.) have surged. These disasters are more common in mountainous regions, and could be a result of tectonic, climatic, and/or human activities [1,2,3]. Over the past two decades, more than 300 natural disasters occur annually around the globe, affecting over four billion and cost around USD 2.97 trillion [4,5]. The 2021 state of the environment notes that disasters are continuing to take a heavy toll on life and assets, severely affecting and rolling back the development gains of countries [6]. In addition, Mohammed et al. [7] and the sixth Intergovernmental Panel on Climate Change report [8] note with confidence that human-induced climate change is the dominant driver in sediment related natural disasters. In assessing the influence of climate change on soil erosion and sediment yield, Chen et al. [9] illustrate an increase in these events under the A1B-climate change scenario. This study highlights the importance of incorporating climate change in sediment-related disaster models. One of the most important transboundary rivers in China, the Lancang-Mekong River has been shown to cause major sediment loads in the last decade in Asian Rivers with a mean annual loss of 5350 t ha⁻¹ year⁻¹ [10].

In sight of this, this Special Issue aimed to contribute towards improving our knowledge and understanding on the processes and mechanics of soil erosion and landslides, as these are among the main natural disasters affecting the globe. This is crucial in developing the right tools and models for soil and water conservation, disaster mitigation, and early warning systems. Several novel tools and methodologies are presented in this Special Issue.

Several novel tools and methodologies are presented in this Special Issue, which consists of 19 articles, covering a wide range of topics, including landslide prediction models, soil erosion and sediment yield estimation, flood simulations, dam breach, and rainfall-runoff models.

Wu and Yeh [11] developed an improved landslide probability model from existing models by including long-term landslide inventory and rainfall factors, which can further be used to predict landslides based on future changes in rainfall patterns. In addition to the aforementioned landslide probability model, Wu et al. [12] argue that landslide susceptibility assessments after extreme rainfall events is equally critical. Four methods are evaluated based on 12 landslide related factors which formed the basis for the landslide susceptibility assessment. The methods include Landslide ratio-based logistic regression (LRBLR), Frequency Ration Method (FRM), Instability Index Method (IIM), and Weight of Evidence Method (WEM). Among these, the LRBLR method is shown to be the best in landslide susceptibility assessment. The article by Wu and Lin [13] presents a method applying rainfall analysis, spatiotemporal landslide analysis, and comparison analysis of rainfall induced landslide and earth quakes to evaluate how landslides would be active after specific rainfall events. In classifying landslides, Wan et al. [14] presented a methodology that applied hyperspectral data instead of solely relying on digital elevation models. The model can differentiate bare land to landslides, which has often been a complex task.

In addition to the application of complex models, some authors [15,16] argued that the ability of vegetation in mitigating soil erosion and subsequent large sediment yield under extreme rainfall events and runoff is not well investigated. Different vegetation communities enacted at 1, 11, 15, 25, and 40 years were evaluated. Findings showed root biomass can reach up to ~11 mg/cm³, which reduces slope runoff velocity by 48%, while increasing runoff resistance by 35 times. The results suggest the importance of indigenous knowledge in reducing the impacts of sediment related disasters. Chen, Guo, and Wang [16] compared farmland with revegetated gullies and demonstrated that revegetating gullies could lower soil erodibility by 31–78% and could further improve critical shear stress by up to four times, and stable conditions were possible in approximately 18 years. The importance of riparian vegetation in stabilizing river banks is illustrated by Zhu et al. [17]. The authors show that healthy native alpine swamp can enhance riverbank stability and could further delay the development of tensile cracks.

The article by Liu et al. [18] describes a novel shallow water equation based methodology that could simulate flood routing in complex terrains. Hung et al. [19] used flume tests to study dam breach and the resulting seismic signals induced. The authors conclude overtopping discharge and lateral sliding masses are significant in influencing the evolution of dam breach. The resulting dam breach model from the study is important for dam breach warning, which could save lives in the event of a catastrophic dam breach associated with floods.

Mosavi et al. [20] presents a novel machine learning model (Weighted sub-space random forest) to map susceptibility of water erosion of the soil. In applying the model, 19 factors are applied (some of these are aspect, curvature, slope length, flow accumulation, normalized vegetative index, soil texture, lithology, etc.). Such a tool is crucial in watershed conservation, especially in a world generating enormous data that requires super-fast models. In another instance, Lee et al. [21] evaluates seven machine learning models for time-saving to estimate the rainfall-erosivity factor (R-factor) used in the Universal Soil Loss Equation (USLE). Their findings show that deep neural networks are very efficient estimating the R-factor given monthly precipitation, maximum daily precipitation, and maximum hourly precipitation.

Finally, the article by Lee, Lu, and Huang [21] investigate the interaction between overfall types and scour at bridges. When a bridge pier is in the maximum scour location, it induces more scour due to disturbances caused by the water jet and the pier; hence, more attention is required to protect the pier. This is especially important in areas prone to earthquakes, which may cause the riverbed to uplift.

In conclusion, the Special Issue presents several articles. They are broadly categorized into five themes: (i) those with emphasis on soil erosion and how climate change has worsened natural disasters; (ii) those investigating and developing landslide related models; (iii) articles regarding flood models and dam breach processes; (iv) articles focusing on the application of recent technologies, such as machine learning in addressing sediment related disasters; and (v) studies dedicated at the protection of riverine structures. These articles have indeed contributed immensely towards understanding soil erosion and landslide processes, and the corresponding necessary tools to foster resilience. This further aligns with some of the post 2015 global frameworks, such as the Sendai Framework which articulates the need for improved understanding of disasters in all its dimensions.