Rainfall-Induced Landslide Prediction Using Machine Learning Models: The Case of Ngororero District, Rwanda

Landslides fall under natural, unpredictable and most distractive disasters. Hence, early warning systems of such disasters can alert people and save lives. Some of the recent early warning models make use of Internet of Things to monitor the environmental parameters to predict the disasters. Some other models use machine learning techniques (MLT) to analyse rainfall data along with some internal parameters to predict these hazards. The prediction capability of the existing models and systems are limited in terms of their accuracy. In this research paper, two prediction modelling approaches, namely random forest (RF) and logistic regression (LR), are proposed. These approaches use rainfall datasets as well as various other internal and external parameters for landslide prediction and hence improve the accuracy. Moreover, the prediction performance of these approaches is further improved using antecedent cumulative rainfall data. These models are evaluated using the receiver operating characteristics, area under the curve (ROC-AUC) and false negative rate (FNR) to measure the landslide cases that were not reported. When antecedent rainfall data is included in the prediction, both models (RF and LR) performed better with an AUC of 0.995 and 0.997, respectively. The results proved that there is a good correlation between antecedent precipitation and landslide occurrence rather than between one-day rainfall and landslide occurrence. In terms of incorrect predictions, RF and LR improved FNR to 10.58% and 5.77% respectively. It is also noted that among the various internal factors used for prediction, slope angle has the highest impact than other factors. Comparing both the models, LR model’s performance is better in terms of FNR and it could be preferred for landslide prediction and early warning. LR model’s incorrect prediction rate FNR = 9.61% without including antecedent precipitation data and 3.84% including antecedent precipitation data.

Download Full-text

Satellite rainfall products outperform ground observations for landslide prediction in India

Hydrology and Earth System Sciences ◽

10.5194/hess-25-3267-2021 ◽

2021 ◽

Vol 25 (6) ◽

pp. 3267-3279

Author(s):

Maria Teresa Brunetti ◽

Massimo Melillo ◽

Stefano Luigi Gariano ◽

Luca Ciabatta ◽

Luca Brocca ◽

...

Keyword(s):

Early Warning ◽

Daily Rainfall ◽

Early Warning Systems ◽

Rain Gauge ◽

Warning Systems ◽

Landslide Prediction ◽

Landslide Occurrence ◽

Monitoring Process ◽

Satellite Rainfall ◽

The Impact

Abstract. Landslides are among the most dangerous natural hazards, particularly in developing countries, where ground observations for operative early warning systems are lacking. In these areas, remote sensing can represent an important detection and monitoring process to predict landslide occurrence in space and time, particularly satellite rainfall products that have improved in terms of accuracy and resolution in recent times. Surprisingly, only a few studies have investigated the capability and effectiveness of these products in landslide prediction in reducing the impact of this hazard on the population. We have performed a comparative study of ground- and satellite-based rainfall products for landslide prediction in India by using empirical rainfall thresholds derived from the analysis of historical landslide events. Specifically, we have tested Global Precipitation Measurement (GPM) and SM2RAIN-ASCAT satellite rainfall products, and their merging, at daily and hourly temporal resolution, and Indian Meteorological Department (IMD) daily rain gauge observations. A catalogue of 197 rainfall-induced landslides that occurred throughout India in the 13-year period between April 2007 and October 2019 has been used. Results indicate that satellite rainfall products outperform ground observations thanks to their better spatial (0.1∘ vs. 0.25∘) and temporal (hourly vs. daily) resolutions. The better performance is obtained through the merged GPM and SM2RAIN-ASCAT products, even though improvements in reproducing the daily rainfall (e.g. overestimation of the number of rainy days) are likely needed. These findings open a new avenue for using such satellite products in landslide early warning systems, particularly in poorly gauged areas.

Download Full-text

Flood Early Warning Systems Using Machine Learning Techniques: The Case of the Tomebamba Catchment at the Southern Andes of Ecuador

Hydrology ◽

10.3390/hydrology8040183 ◽

2021 ◽

Vol 8 (4) ◽

pp. 183

Author(s):

Paul Muñoz ◽

Johanna Orellana-Alvear ◽

Jörg Bendix ◽

Jan Feyen ◽

Rolando Célleri

Keyword(s):

Machine Learning ◽

Early Warning ◽

Early Warning Systems ◽

Data Representation ◽

Machine Learning Techniques ◽

Lead Times ◽

Warning Systems ◽

Tropical Andes ◽

K Nearest Neighbors ◽

Learning Techniques

Worldwide, machine learning (ML) is increasingly being used for developing flood early warning systems (FEWSs). However, previous studies have not focused on establishing a methodology for determining the most efficient ML technique. We assessed FEWSs with three river states, No-alert, Pre-alert and Alert for flooding, for lead times between 1 to 12 h using the most common ML techniques, such as multi-layer perceptron (MLP), logistic regression (LR), K-nearest neighbors (KNN), naive Bayes (NB), and random forest (RF). The Tomebamba catchment in the tropical Andes of Ecuador was selected as a case study. For all lead times, MLP models achieve the highest performance followed by LR, with f1-macro (log-loss) scores of 0.82 (0.09) and 0.46 (0.20) for the 1 h and 12 h cases, respectively. The ranking was highly variable for the remaining ML techniques. According to the g-mean, LR models correctly forecast and show more stability at all states, while the MLP models perform better in the Pre-alert and Alert states. The proposed methodology for selecting the optimal ML technique for a FEWS can be extrapolated to other case studies. Future efforts are recommended to enhance the input data representation and develop communication applications to boost the awareness of society of floods.

Download Full-text

Landslide precipitation thresholds in Rwanda

Landslides ◽

10.1007/s10346-020-01457-9 ◽

2020 ◽

Vol 17 (10) ◽

pp. 2469-2481

Author(s):

Judith Uwihirwe ◽

Markus Hrachowitz ◽

Thom A. Bogaard

Keyword(s):

System Development ◽

Explanatory Power ◽

Probabilistic Approach ◽

Area Under The Curve ◽

Radial Distance ◽

Warning System ◽

High Rate ◽

Operating Characteristics ◽

Landslide Occurrence ◽

Antecedent Precipitation

Abstract Regional empirical-statistical thresholds indicating the precipitation conditions initiating landslides are of crucial importance for landslide early warning system development. The objectives of this research were to use landslide and precipitation data in an empirical-statistical approach to (1) identify precipitation-related variables with the highest explanatory power for landslide occurrence and (2) define both trigger and trigger-cause based thresholds for landslides in Rwanda, Central-East Africa. Receiver operating characteristics (ROC) and area under the curve (AUC) metrics were used to test the suitability of a suite of precipitation-related explanatory variables. A Bayesian probabilistic approach, maximum true skill statistics and the minimum radial distance were used to determine the most informative threshold levels above which landslide are high likely to occur. The results indicated that the event precipitation volumes E, cumulative 1-day rainfall (RD1) that coincide with the day of landslide occurrence and 10-day antecedent precipitation are variables with the highest discriminatory power to distinguish landslide from no landslide conditions. The highest landslide prediction capability in terms of true positive alarms was obtained from single rainfall variables based on trigger-based thresholds. However, that predictive capability was constrained by the high rate of false positive alarms and thus the elevated probability to neglect the contribution of additional causal factors that lead to the occurrence of landslides and which can partly be accounted for by the antecedent precipitation indices. Further combination of different variables into trigger-cause pairs and the use of suitable thresholds in bilinear format improved the prediction capacity of the real trigger-based thresholds.

Download Full-text

Enhancing the reliability of landslide early warning systems by machine learning

Landslides ◽

10.1007/s10346-020-01453-z ◽

2020 ◽

Vol 17 (9) ◽

pp. 2231-2246

Author(s):

Hemalatha Thirugnanam ◽

Maneesha Vinodini Ramesh ◽

Venkat P. Rangan

Keyword(s):

Machine Learning ◽

Early Warning ◽

Early Warning Systems ◽

Warning Systems ◽

Landslide Early Warning

Download Full-text

The Influence of Rainfall Amounts Not Exceeding Critical Rainfall on Landslide Occurrence in Japan

10.21203/rs.3.rs-1063753/v1 ◽

2021 ◽

Author(s):

Soichi Kaihara ◽

Noriko Tadakuma ◽

Hitoshi Saito ◽

Hiroaki Nakaya

Keyword(s):

Early Warning Systems ◽

Rainfall Data ◽

Warning Systems ◽

Negative Data ◽

True Negative ◽

Rainfall Events ◽

Landslide Occurrence ◽

The Past ◽

Critical Rainfall ◽

Landslide Early Warning

Abstract Critical rainfall events are used in landslide early-warning systems to predict the occurrence and severity of disasters. In this study, past landslide disasters in Japan were identified for which the critical rainfall set for each 1-km grid was exceeded using historical landslide records, radar-based rainfall data over a 1-km grid, and standard rainfall data collected over the past 17 years. It was determined that nearly equal numbers of rainfall events were identified with higher and lower rainfall amounts than the critical rainfall. The probability that a series of rainfall events would cause a landslide was approximately 1.15% when the critical rainfall was exceeded and 0.09% otherwise, a difference of approximately 10 times. It was also found that even if critical rainfall was not exceeded, in the case of debris flow and slope failures, there was rainfall that exceeded the standard rainfall one or two days before. In the case of landslides, there was rainfall that exceeded the critical rainfall one or two weeks before, and if the critical rainfall was exceeded in another rainfall event, a landslide could occur. The operational evaluation of Japanese LEWSs has a recall value of 0.486 as the accuracy of occurrence prediction, which was related to the fact that almost half of the rainfall events occurred in nonexceedance of the reference rainfall. The specificity was 0.935, known as the accuracy of nonoccurrence prediction, which was also greatly influenced by the TN (true negative) data of nonexceeding rainfall events, which accounted for most of the data.

Download Full-text

Flood Early Warning Systems using Machine Learning Techniques. Case the Tomebamba Catchment at the Southern Andes of Ecuador

10.20944/preprints202111.0510.v1 ◽

2021 ◽

Author(s):

Paul Muñoz ◽

Johanna Orellana-Alvear ◽

Jörg Bendix ◽

Jan Feyen ◽

Rolando Célleri

Keyword(s):

Machine Learning ◽

Early Warning ◽

Early Warning Systems ◽

Data Representation ◽

Machine Learning Techniques ◽

Lead Times ◽

Warning Systems ◽

Tropical Andes ◽

K Nearest Neighbors ◽

Learning Techniques

Flood Early Warning Systems (FEWSs) using Machine Learning (ML) has gained worldwide popularity. However, determining the most efficient ML technique is still a bottleneck. We assessed FEWSs with three river states, No-alert, Pre-alert, and Alert for flooding, for lead times between 1 to 12 hours using the most common ML techniques, such as Multi-Layer Perceptron (MLP), Logistic Regression (LR), K-Nearest Neighbors (KNN), Naive Bayes (NB), and Random Forest (RF). The Tomebamba catchment in the tropical Andes of Ecuador was selected as case study. For all lead times, MLP models achieve the highest performance followed by LR, with f1-macro (log-loss) scores of 0.82 (0.09) and 0.46 (0.20) for the 1- and 12-hour cases, respectively. The ranking was highly variable for the remaining ML techniques. According to the g-mean, LR models correctly forecast and show more stability at all states, while the MLP models perform better in the Pre-alert and Alert states. Future efforts are recommended to enhance the input data representation and develop communication applications to boost the awareness of the society for floods.

Download Full-text

A Study on the Slope Failure Monitoring of a Model Slope by the Application of a Displacement Sensor

Journal of Sensors ◽

10.1155/2019/7570517 ◽

2019 ◽

Vol 2019 ◽

pp. 1-9 ◽

Cited By ~ 1

Author(s):

Sungyong Park ◽

Hyuntaek Lim ◽

Bibek Tamang ◽

Jihuan Jin ◽

Seungjoo Lee ◽

...

Keyword(s):

Real Time ◽

Early Warning ◽

Slope Failure ◽

Early Warning Systems ◽

Displacement Sensor ◽

Scale Model ◽

Economic Losses ◽

Real Time Monitoring ◽

Landslide Occurrence ◽

Model Slope

Many causalities and economic losses are caused by natural disasters, such as landslides and slope failures, every year. This suggests that there is a need for an early warning system to mitigate casualties and economic losses. Most of the studies on early warning systems have been carried out by predicting landslide-prone areas, but studies related to the prediction of landslide occurrence time points by the real-time monitoring of slope displacement are still insufficient. In this study, a displacement sensor and an Internet of Things (IoT) monitoring system were combined together, to monitor slope failure through cutting experiments of a real-scale model slope. Real-time monitoring of the slope movement was performed simultaneously via a low-cost, efficient, and easy-to-use IoT system. Based on the obtained displacement data, an inverse displacement analysis was performed. Finally, a slope instrumentation standard was proposed based on the slope of the inverse displacement for early evacuation before slope failure.

Download Full-text

Fluctuating Rainfall, Persistent Food Crisis—Use of Rainfall Data in the Kenyan Drought Early Warning System

Atmosphere ◽

10.3390/atmos11121328 ◽

2020 ◽

Vol 11 (12) ◽

pp. 1328

Author(s):

Sofie Sandström ◽

Sirkku Juhola ◽

Aleksi Räsänen

Keyword(s):

Early Warning ◽

Food Systems ◽

Qualitative Content Analysis ◽

Warning System ◽

Meteorological Drought ◽

Early Warning Systems ◽

Rainfall Data ◽

Humanitarian Crises ◽

Drought Early Warning ◽

Systems Framework

Early warning systems (EWSs) have been developed to trigger timely action to disasters, yet persistent humanitarian crises resulting from hazards such as drought indicate that these systems need improvements. We focus our research on the county of Turkana in Kenya, where drought repeatedly results in humanitarian crises, especially with regard to food insecurity. Focusing on the key elements of the Kenyan EWS, we ask two questions: firstly, what indicators, especially meteorological drought indicators, are used in the national biannual assessments conducted by the Kenyan National Drought Management Authority and monthly drought bulletins for Turkana? Secondly, are there differences in the methodology used for analysis of meteorological indicators in the different documents? Firstly, by utilizing a food systems framework, we conduct qualitative content analysis of the use of indicators in the documents; secondly, we analyze rainfall data and its use. The EWS relies primarily on food availability indicators, with less focus for food access and utilization. The biannual assessments and the country bulletins use different sets of rainfall data and different methodologies for establishing the climate normal, leading to discrepancies in the output of the EWS. We recommend further steps to be taken towards standardization of methodologies and cooperation between various institutions to ensure streamlining of approaches.

Download Full-text

Using Machine Learning to Advance Early Warning Systems: Promise and Pitfalls

Teachers College Record ◽

10.1177/016146812012201403 ◽

2020 ◽

Vol 122 (14) ◽

pp. 1-30

Author(s):

James Soland ◽

Benjamin Domingue ◽

David Lang

Keyword(s):

Machine Learning ◽

High School ◽

At Risk ◽

Early Warning ◽

Early Warning Systems ◽

Machine Learning Techniques ◽

Dropping Out ◽

Learning Methods ◽

Machine Learning Methods ◽

Learning Techniques

Background/Context Early warning indicators (EWI) are often used by states and districts to identify students who are not on track to finish high school, and provide supports/interventions to increase the odds the student will graduate. While EWI are diverse in terms of the academic behaviors they capture, research suggests that indicators like course failures, chronic absenteeism, and suspensions can help identify students in need of additional supports. In parallel with the expansion of administrative data that have made early versions of EWI possible, new machine learning methods have been developed. These methods are data-driven and often designed to sift through thousands of variables with the purpose of identifying the best predictors of a given outcome. While applications of machine learning techniques to identify students at-risk of high school dropout have obvious appeal, few studies consider the benefits and limitations of applying those models in an EWI context, especially as they relate to questions of fairness and equity. Focus of Study In this study, we will provide applied examples of how machine learning can be used to support EWI selection. The purpose is to articulate the broad risks and benefits of using machine learning methods to identify students who may be at risk of dropping out. We focus on dropping out given its salience in the EWI literature, but also anticipate generating insights that will be germane to EWI used for a variety of outcomes. Research Design We explore these issues by using several hypothetical examples of how ML techniques might be used to identify EWI. For example, we show results from decision tree algorithms used to identify predictors of dropout that use simulated data. Conclusions/Recommendations Generally, we argue that machine learning techniques have several potential benefits in the EWI context. For example, some related methods can help create clear decision rules for which students are a dropout risk, and their predictive accuracy can be higher than for more traditional, regression-based models. At the same time, these methods often require additional statistical and data management expertise to be used appropriately. Further, the black-box nature of machine learning algorithms could invite their users to interpret results through the lens of preexisting biases about students and educational settings.

Download Full-text

Identification of hydro-meteorological thresholds for rainfall-induced landslide prediction with clustering techniques

10.5194/egusphere-egu21-356 ◽

2021 ◽

Author(s):

Pasquale Marino ◽

Carlo Giudicianni ◽

Giovanni francesco Santonastaso ◽

Roberto Greco

Keyword(s):

Soil Cover ◽

Early Warning Systems ◽

Shallow Landslides ◽

Landslide Prediction ◽

Clustering Techniques ◽

Antecedent Precipitation ◽

Limestone Bedrock ◽

Definition Of ◽

Hydrologic Conditions ◽

Triggering Rainfall

Operational early warning systems for rainfall-induced landslides (LEWS) usually rely on simple empirical thresholds based on the statistical analysis of either triggering rainfall characteristics, e.g. intensity and duration (Guzzetti et al., 2007). The main pro of this simplified approach is that it requires only rainfall records, at the desired spatial and temporal resolution, and a database of landslides with known time and location. The effect of the hydrologic conditions of the slopes prior the onset of the triggering rainfall is usually neglected, limiting the performance of the LEWS, which often give rise to false and missing alarms. To address this issue, antecedent precipitation is sometimes included in the definition of the threshold, but the identification of the antecedent precipitation duration is doubtful, as this approach neglects non-linear hydrological processes affecting slope response. Hydro-meteorological thresholds, linking a variable accounting for the antecedent hydrologic conditions with a characteristic of the triggering rainfall, have been recently proposed (Bogaard and Greco, 2018).In this study, hydro-meteorological thresholds for landslide prediction are identified for a slope in southern Italy, characterized by an unsaturated pyroclastic soil cover laying upon fractured limestone bedrock and subject to rainfall-induced shallow landslides. To this aim, a synthetic 1000 years long hourly point rainfall record is generated with the Neyman-Scott rectangular pulse stochastic model, calibrated thanks to available measured rainfall. The response of the slope to the synthetic rainfall record is simulated by means of a physically-based model, which couples unsaturated flow in the soil cover with a temporary perched aquifer in the limestone bedrock, and allows estimating all the terms of slope water balance (Greco et al., 2018). The stability of the slope is eventually assessed under the infinite slope hypothesis, allowing the identification of the occurrence of landslides.The obtained synthetic dataset of rainfall and hydrologic variables has been exploited for the definition of hydro-meteorological thresholds. All the combinations of hydrologic variables with triggering rainfall height have been analyzed with several clustering techniques, so to identify the most effective combinations for landslide predictions.&#160;References:Bogaard TA, Greco R (2018). Invited perspectives: Hydrological perspectives on precipitation intensity-duration thresholds for landslide initiation: proposing hydro-meteorological thresholds, Nat Hazards Earth Syst Sci, 18: 31&#8211;39.Greco R, Marino P, Santonastaso GF, Damiano E (2018). Interaction between Perched Epikarst Aquifer and Unsaturated Soil Cover in the Initiation of Shallow Landslides in Pyroclastic Soils, Water, 10: 948.Guzzetti F, Peruccacci S, Rossi M, Stark CP (2007). Rainfall thresholds for the initiation of landslides in central and southern Europe, Meteorol Atmos Phys, 98: 239&#8211;267.

Download Full-text