Novel forecasting approaches using combination of machine learning and statistical models for flood susceptibility mapping

Landslide susceptibility assessment based on frequency ratio and semi-supervised heterogeneous ensemble learning model

Epistemic uncertainty in data-driven landslide susceptibility assessment often tends to be increased by the limited accuracy of an individual model, as well as uncertainties associated with the selection of non-landslide samples. To address these issues, this paper centers on the landslide disaster in Ji'an City, China, and proposes a heterogeneous ensemble learning method incorporating frequency ratio (FR) and semi-supervised sample expansion. Based on the superimposed results of 12 environmental factor frequency ratios (FFR), non-landslide samples were selected and input into light gradient boosting machine (LightGBM), random forest (RF), and convolutional neural network (CNN) models for prediction along with historical landslide samples. The predicted probability values are integrated by four heterogeneous ensemble strategies to expand samples from high-confidence results. The model's performance is evaluated using the area under the receiver operating characteristic curve (AUC), partition frequency ratio (PFR), and other verification methods. The results demonstrate that the negative sample based on FFR sampling is more accurate than the random sampling method, and the FR-SSELR model based on frequency ratio sampling and semi-supervised ensemble strategy exhibits the highest performance (AUC = 0.971, ACC = 0.941). A more reasonable landslide susceptibility map was drawn based on this model, with the lowest percentage of landslides in the low and very low susceptibility zones (sum of PFR = 0.194), as well as the highest percentage of landslides in the high and very high susceptibility zones (sum of PFR = 6.800). Furthermore, the FR-SSELR model improved economic benefits by 3.82-14.2%, offering valuable guidance for decision-making regarding landslide management and the sustainability of Ji'an City.

A novel flood risk management approach based on future climate and land use change scenarios

Climate change and increasing urbanization are two primary factors responsible for the increased risk of serious flooding around the world. The prediction and monitoring of the effects of land use/land cover (LULC) and climate change on flood risk are critical steps in the development of appropriate strategies to reduce potential damage. This study aimed to develop a new approach by combining machine learning (namely the XGBoost, CatBoost, LightGBM, and ExtraTree models) and hydraulic modeling to predict the effects of climate change and LULC change on land that is at risk of flooding. For the years 2005, 2020, 2035, and 2050, machine learning was used to model and predict flood susceptibility under different scenarios of LULC, while hydraulic modeling was used to model and predict flood depth and flood velocity, based on the RCP 8.5 climate change scenario. The two elements were used to build a flood risk assessment, integrating socioeconomic data such as LULC, population density, poverty rate, number of women, number of schools, and cultivated area. Flood risk was then computed, using the analytical hierarchy process, by combining flood hazard, exposure, and vulnerability. The results showed that the area at high and very high flood risk increased rapidly, as did the areas of high/very high exposure, and high/very high vulnerability. They also showed how flood risk had increased rapidly from 2005 to 2020 and would continue to do so in 2035 and 2050, due to the dynamics of climate change and LULC change, population growth, the number of women, and the number of schools - particularly in the flood zone. The results highlight the relationships between flood risk and environmental and socio-economic changes and suggest that flood risk management strategies should also be integrated in future analyses. The map built in this study shows past and future flood risk, providing insights into the spatial distribution of urban area in flood zones and can be used to facilitate the development of priority measures, flood mitigation being most important.

Development of High-Precision Urban Flood-Monitoring Technology for Sustainable Smart Cities

Owing to rapid climate change, large-scale floods have occurred yearly in cities worldwide, causing serious damage. We propose a real-time urban flood-monitoring technology as an urban disaster prevention technology for sustainable and secure smart cities. Our method takes advantage of the characteristic that water flow is regularly detected at a certain distance with a constant Doppler velocity within the radar observation area. Therefore, a pure flow energy detection algorithm in this technology can accurately and immediately detect water flow due to flooding by effectively removing dynamic obstacles such as cars, people, and animals that cause changes in observation distance, and static obstacles that do not cause Doppler velocities. Specifically, in this method, the pure flow energy is detected by generating a two-dimensional range-Doppler relation map using 1 s periodic radar observation data and performing statistical analysis on the energy detected on the successive maps. Experiments to verify the proposed technology are conducted indoors and in real river basins. As a result of conducting experiments in a narrow indoor space that could be considered an urban underpass or underground facility, it was found that this method can detect flooding situations with centimeter-level accuracy by measuring water level and flow velocity in real time from the time of flood occurrence. And the experimental results in various river environments showed that our technology could accurately detect changes in distance and flow speed from the river surface. We also confirmed that this method could effectively eliminate moving obstacles within the observation range and detect only pure flow energy. Finally, we expect that our method will be able to build a high-density urban flood-monitoring network and a high-precision digital flood twin.

A new approach based on biology-inspired metaheuristic algorithms in combination with random forest to enhance the flood susceptibility mapping

Flash floods are one of the worst natural disasters, causing massive economic losses and many deaths. Creating a flood susceptibility map (FSM) that pinpoints the areas most at risk of flooding is a crucial non-structural solution for managing floods. This study aimed to assess the efficacy of combinations of the random forest (RF) model with three biology-inspired metaheuristic algorithms, namely invasive weed optimization (IWO), slime mould algorithm (SMA), and satin bowerbird optimization (SBO), for flood susceptibility mapping in Estahban town, Iran. Initially, synthetic-aperture radar (SAR) (Sentinel-1) and optical (Landsat-8) satellite images were integrated to monitor the flooded areas during the July 2022 monsoon in the study area. A dataset of 509 flood occurrence points was created to identify flood-prone areas using remote sensing techniques, considering the monitored flood areas. The dataset also included twelve flood-related criteria: topography, land cover, and climate. The holdout method was employed for modeling, with a ratio of 70:30 used for the train/test split. Data pre-processing techniques were conducted to improve model performance, including determining criteria importance and addressing multicollinearity issues using certainty factor (CF), multicollinearity, and information gain ratio (IGR) methods. Then FSM was prepared using RF, RF-IWO, RF-SBO, and RF-SMA models. The findings of this research revealed that the RF-IWO model was the best predictive model of flood susceptibility modeling, with root-mean-square-error (RMSE) (0.211 and 0.0.27), mean-absolute-error (MAE) (0.103 and 0.15), and coefficient-of-determination (R2) (0.821 and 0.707) in the training and testing phases, respectively. Receiver operating characteristic (ROC) curve analysis of FSM revealed that the most accurate models were the RF-IWO (area under the curve (AUC) = 0.983), RF-SBO (AUC = 0.979), RF-SMA (AUC = 0.963), and RF (AUC = 0.959), respectively. Integrating biology-inspired computing algorithms with machine learning algorithms presents a novel approach to enhancing the accuracy of FSMs.

An integrated model chain for future flood risk prediction under land-use changes

Flood is a very destructive natural disaster in the world that is strongly influenced by land-use change. Therefore, a comprehensive flood risk modeling considering the change in land-use is essential for understanding, predicting, and mitigating flood risk. However, most existing single modeling ignored the derivative effect of land-use change, which may reduce the reality of results. To further address the issue, this study presented an integrated model chain by coupling the Markov-FLUS model, the multiple linear regression and the improved TOPSIS model. By applying it in Guangdong Province, the future land-use simulation, spatialization of hazard-bearing bodies, and determination of flood risk were realized. The results show that the coupled model chain allows for good prediction of flood risk under different scenarios, which could be expressed by flood risk composite index (FRSI). In the natural growth scenario, the flood risk will show markedly increasing trend from 2020 to 2030 (FRSI = 2.06), with the high and highest risk zones will expand significantly. Spatially, these increased high flood risk zones mainly distributed on the periphery of existing built-up lands. On the contrary, the flood risk in ecological protection scenario tends to stabilize (FRSI = 1.98), which may be a reference for alternative development paths. These dynamic information identified by this model chain provides a deeper insight into the spatiotemporal characteristics of future high flood risk areas, which can facilitate reasonable flood mitigation measures to be developed at the most critical locations in the region. In further applications, more efficient spatialization models and climate factor are suggested to be introduced.

Application of genetic algorithm in optimization parallel ensemble-based machine learning algorithms to flood susceptibility mapping using radar satellite imagery

Floods are the natural disaster that occurs most frequently due to the weather and causes the most widespread destruction. The purpose of the proposed research is to analyze flood susceptibility mapping (FSM) in the Sulaymaniyah province of Iraq. This study employed a genetic algorithm (GA) to fine-tune parallel ensemble-based machine learning algorithms (random forest (RF) and bootstrap aggregation (Bagging)). Four machine learning algorithms (RF, Bagging, RF-GA, and Bagging-GA) were used to build FSM in the study area. To provide inputs into parallel ensemble-based machine learning algorithms, we gathered and processed data from meteorological (Rainfall), satellite image (flood inventory, normalized difference vegetation index (NDVI), aspect, land cover, altitude, stream power index (SPI), plan curvature, topographic wetness index (TWI), slope) and geographic sources (geology). For this research, Sentinel-1 synthetic aperture radar (SAR) satellite images were utilized to locate flooded areas and create an inventory map of floods. To train and validate the model, we employed 70 % and 30 % of 160 selected flood locations, respectively. Multicollinearity, frequency ratio (FR), and Geodetector methods were used for data preprocessing. Four metrics were utilized to assess the FSM performance: the root mean square error (RMSE), the area under the receiver-operator characteristic curve (AUC-ROC), the Taylor diagram, and the seed cell area index (SCAI). The results exhibited that all the suggested models have high accuracy of prediction, but the performance of Bagging-GA (RMSE (Train = 0.1793, Test = 0.4543)) was slightly better than RF-GA (RMSE (Train = 0.1803, Test = 0.4563)), Bagging (RMSE (Train = 0.2191, Test = 0.4566)), and RF (RMSE (Train = 0.2529, Test = 0.4724)). According to the ROC index, the Bagging-GA model (AUC = 0.935) was the most accurate in flood susceptibility modeling, followed by the RF-GA (AUC = 0.904), the Bagging (AUC = 0.872), and the RF (AUC = 0.847) models. The study's identification of high-risk flood zones and the most significant factors contributing to flooding make it a helpful resource for flood management.

Enhancing flood susceptibility modeling using multi-temporal SAR images, CHIRPS data, and hybrid machine learning algorithms

Flood susceptibility maps are useful tool for planners and emergency management professionals in the early warning and mitigation stages of floods. In this study, Sentinel-1 dB radar images, which provide Synthetic-Aperture Radar (SAR) data were used to delineate flooded and non-flooded locations. 12 input parameters, including elevation, lithology, drainage density, rainfall, Normalized Difference Vegetation Index (NDVI), curvature, ground slope, Stream Power Index (SPI), Topographic Wetness Index (TWI), soil, Land Use Land Cover (LULC), and distance from the river, were selected for model development. The importance of each input parameter on flood occurrences was assessed via the Mutual Information (MI) technique. Several machine learning models, including Radial Basis Function (RBF), and three hybrid models of Bagging (BA-RBF), Random Committee (RC-RBF), and Random Subspace (RSS-RBF), were developed to delineate flood susceptibility areas at Goorganrood watershed, Iran. The performance of each model was evaluated using several error indicators, including correlation coefficient (r), Nash Sutcliffe Efficiency (NSE), Mean Absolute Error (MSE), Root Mean Square Error (RMSE), and the Area Under the Receiver Operating Characteristic (ROC) Curve (AUC). The results showed that the hybrid techniques enhanced the modeling performance of the standalone model, and generally, all hybrid models are more accurate than the standalone model. Although all developed models have performed well, RC-RBF outperforms all of them (AUC = 0.997), followed by BA-RBF (AUC = 0.996), RSS-RBF (AUC = 0.992), and RBF (AUC = 0.975). Generally, about 12 % of the study area has high and very high susceptibility to future flood occurrences.

A comparative assessment of flood susceptibility modelling of GIS-based TOPSIS, VIKOR, and EDAS techniques in the Sub-Himalayan foothills region of Eastern India

In the Sub-Himalayan foothills region of eastern India, floods are considered the most powerful annually occurring natural disaster, which cause severe losses to the socio-economic life of the inhabitants. Therefore, the present study integrated geographic information system (GIS) and three comprehensive and systematic multicriteria decision-making (MCDM) techniques such as Technique for Order Preference by Similarity to Ideal Solution (TOPSIS), Vise Kriterijumska Optimizacijaik Ompromisno Resenje (VIKOR), and Evaluation Based on Distance from Average Solution (EDAS) in Koch Bihar district for comparative assessment of the flood-susceptible zones. The multi-dimensional 21 indicators were considered, and multicollinearity statistics were employed to erase the issues regarding highly correlated parameters (i.e., MFI and long-term annual rainfall). Results of MCDM models depicted that the riparian areas and riverine "chars" (islands) are the most susceptible sectors, accounting for around 40% of the total area. The microlevel assessment revealed that flooding was most susceptible in the Tufanganj-I, Tufanganj-II, and Mathabhanga-I blocks, while Haldibari, Sitalkuchi, and Sitai blocks were less susceptible. Spearman's rank (r_s) tests among the three MCDM models revealed that TOPSIS-EDAS persisted in a high correlation (r_s = 0.714) in contrast to the relationships between VIKOR-EDAS (r_s = 0.651) and TOPSIS-VIKOR (r_s = 0.639). The model's efficiency was statistically judged by applying the receiver operating characteristic-area under the curve (ROC-AUC), mean absolute error (MAE), mean square error (MSE), and root mean square error (RMSE) techniques to recognize the better-suited models for mapping the flood susceptibility. The performance of all techniques is found good enough (ROC-AUC =  > 0.700 and MAE, MSE and RMSE =  < 0.300). However, TOPSIS and VIKOR have manifested an excellent outcome and are highly recommended for identifying flood susceptibility in such active flood-prone areas. Thus, this kind of study addresses the role of GIS in the construction of the flood susceptibility of the region and the performance of the respective models in a very lucid manner.

Flood vulnerability mapping and urban sprawl suitability using FR, LR, and SVM models

Floods are among the most destructive disasters because they cause immense damage to human life, property (land and buildings), and resources. They also slow down a country's economy. Due to the dynamic and complex nature of floods, it is difficult to predict the areas that are prone to flooding. In this study, an attempt was made to create a suitability map for future urban development based on flood vulnerability maps for the catchment area of Taif, Saudi Arabia. Three models were used for this purpose, including bivariate (FR), multivariate (LR), and machine learning (SVM) were used. Thirteen parameters were used as flood-contributing parameters. The inventory map was constructed using field surveys, historical data, analysis of RADAR (Sentinel-1A), and Google Earth imagery collected between 2013 and 2020. In general, 70% flood locations were randomly selected from the flood inventory map to generate the flood susceptibility model, and the remaining 30% of the flood locations were used for model validation. The flood susceptibility map was classified into five zones: very low, low, moderate, high, and very high. The AUC value used to predict the performance of the models showed that the accuracy reached 89.5, 92.0, and 96.2% for the models FR, LR, and SVM, respectively. Accordingly, the flood susceptibility map produced by the SVM model is accurate and was used to produce a flood vulnerability map with the help of urban and road density maps. Then slope and elevation maps were integrated with the flood vulnerability model to produce the final suitability map, which was classified into three zones: isolated zone, low suitability, and high suitability areas. The results showed that the highly suitable areas are located in the east and northeast of the Taif Basin, where the flood risk is low and very low. The results of this work will improve the land use planning of engineers and authorities and take possible measures to reduce the flood hazards in the area.

A novel framework for addressing uncertainties in machine learning-based geospatial approaches for flood prediction

Globally, many studies on machine learning (ML)-based flood susceptibility modeling have been carried out in recent years. While majority of those models produce reasonably accurate flood predictions, the outcomes are subject to uncertainty since flood susceptibility models (FSMs) may produce varying spatial predictions. However, there have not been many attempts to address these uncertainties because identifying spatial agreement in flood projections is a complex process. This study presents a framework for reducing spatial disagreement among four standalone and hybridized ML-based FSMs: random forest (RF), k-nearest neighbor (KNN), multilayer perceptron (MLP), and hybridized genetic algorithm-gaussian radial basis function-support vector regression (GA-RBF-SVR). Besides, an optimized model was developed combining the outcomes of those four models. The southwest coastal region of Bangladesh was selected as the case area. A comparable percentage of flood potential area (approximately 60% of the total land areas) was produced by all ML-based models. Despite achieving high prediction accuracy, spatial discrepancy in the model outcomes was observed, with pixel-wise correlation coefficients across different models ranging from 0.62 to 0.91. The optimized model exhibited high prediction accuracy and improved spatial agreement by reducing the number of classification errors. The framework presented in this study might aid in the formulation of risk-based development plans and enhancement of current early warning systems.

Image-driven hydrological parameter coupled identification of flood plain wetland conservation and restoration sites

A good many works focus on wetland vulnerability; some works also explore restoration sites at a very limited spatial extent. But the satellite image-driven hydrological data-based approach adopted in this work is absolutely new. Moreover, existing work only focused on identifying restoration sites in the present context, but for devising long-term sustainable planning, predicted hydrological parameters based on possible restoration sites may be an effective tool. Considering this, the present work focused on exploring hydrological data (water presence frequency (WPF), hydro-period (HP) and water depth (WD)) from time-series satellite images. This exploration may resolve the hydrological data scarcity of wetland over the wider geographical areas. Using these parameters, we developed wetland restoration and conservation sites for different historical years (2008, 2018) and predicted years (2028) using ensemble machine learning (EML) models. From the analysis, it was found that water depth, hydro-period and WPF became poorer over the period, and the trend may seem to continue in predicted years. Among the applied EML models, Random Subspace (RS) predicted wetland restoration and conservation sites precisely over others. The predicted area under high-priority restoration sites is 34% in 2018, which was 14% in 2008. In 2028, 12% more areas may fall in this priority level. Wetland away from main streams (mainly ortho-fluvial wetland) and fringe wetland parts should be given more priority for restoration. These present and predicted information will effectively help to frame sustainable wetland restoration planning.

Flood susceptibility evaluation through deep learning optimizer ensembles and GIS techniques

It is difficult to predict and model with an accurate model the floods, that are one of the most destructive risks across the earth's surface. The main objective of this research is to show the prediction power of three ensemble algorithms with respect to flood susceptibility estimation. These algorithms are: Iterative Classifier Optimizer - Alternating Decision Tree - Frequency Ratio (ICO-ADT-FR), Iterative Classifier Optimizer - Deep Learning Neural Network - Frequency Ratio (ICO-DLNN-FR) and Iterative Classifier Optimizer - Multilayer Perceptron - Frequency Ratio (ICO-MLP-FR). The first stage of the manuscript consisted of the collection and processing of the geodatabase needed in the present study. The geodatabase comprises a number of 14 flood predictors and 132 known flood locations. The Correlation-based Feature Selection (CFS) method was used in order to assess the prediction capacity of the 14 predictors in terms of flood susceptibility estimation. The training and validation of the three ensemble models constitute the next stage of the scientific workflow. Several statistical metrics and ROC curve method were involved in the evaluation of the model's performance and accuracy. According to ROC curves all the models achieved high performances since their AUC had values above 0.89. ICO-DLNN-FR proved to be the most accurate model (AUC = 0.959). The outcomes of the study can be used to guide future flood risk management and sustainable land-use planning in the designated area.

Data-Driven Community Flood Resilience Prediction

Climate change and the development of urban centers within flood-prone areas have significantly increased flood-related disasters worldwide. However, most flood risk categorization and prediction efforts have been focused on the hydrologic features of flood hazards, often not considering subsequent long-term losses and recovery trajectories (i.e., community’s flood resilience). In this study, a two-stage Machine Learning (ML)-based framework is developed to accurately categorize and predict communities’ flood resilience and their response to future flood hazards. This framework is a step towards developing comprehensive, proactive flood disaster management planning to further ensure functioning urban centers and mitigate the risk of future catastrophic flood events. In this framework, resilience indices are synthesized considering resilience goals (i.e., robustness and rapidity) using unsupervised ML, coupled with climate information, to develop a supervised ML prediction algorithm. To showcase the utility of the framework, it was applied on historical flood disaster records collected by the US National Weather Services. These disaster records were subsequently used to develop the resilience indices, which were then coupled with the associated historical climate data, resulting in high-accuracy predictions and, thus, utility in flood resilience management studies. To further demonstrate the utilization of the framework, a spatial analysis was developed to quantify communities’ flood resilience and vulnerability across the selected spatial domain. The framework presented in this study is employable in climate studies and patio-temporal vulnerability identification. Such a framework can also empower decision makers to develop effective data-driven climate resilience strategies.

Image Classification and Land Cover Mapping Using Sentinel-2 Imagery: Optimization of SVM Parameters

Land use/cover (LU/LC) classification provides proxies of the natural and social processes related to urban development, providing stakeholders with crucial information. Remotely sensed images combined with supervised classification are common to define land use, but high-performance classifiers remain difficult to achieve, due to the presence of model hyperparameters. Conventional approaches rely on manual adjustment, which is time consuming and often unsatisfying. Therefore, the goal of this study has been to optimize the parameters of the support vector machine (SVM) algorithm for the generation of land use/cover maps from Sentinel-2 satellite imagery in selected humid and arid (three study sites each) climatic regions of Iran. For supervised SVM classification, we optimized two important parameters (gamma in kernel function and penalty parameter) of the LU/LC classification. Using the radial basis function (RBF) of the SVM classification method, we examined seven values for both parameters ranging from 0.001 to 1000. For both climate types, the penalty parameters (PP) showed a direct relationship with overall accuracy (OA). Statistical results confirmed that in humid study regions, LU/LC maps produced with a penalty parameter >100 were more accurate. However, for regions with arid climates, LU/LC maps with a penalty parameter >0.1 were more accurate. Mapping accuracy for both climate types was sensitive to the penalty parameter. In contrast, variations of the gamma values in the kernel function had no effect on the accuracy of the LU/LC maps in either of the climate zones. These new findings on SVM image classification are directly applicable to LU/LC for planning and environmental and natural resource management.

Combination of Hyperspectral and Machine Learning to Invert Soil Electrical Conductivity

An accurate estimation of soil electrical conductivity (EC) using hyperspectral techniques is of great significance for understanding the spatial distribution of solutes and soil salinization. Although spectral transformation has been widely used in data pre-processing, the performance of different pre-processing techniques (or combination methods) on different models of the same data set is still ambiguous. Moreover, extremely randomized trees (ERT) and light gradient boosting machine (LightGBM) models are new learning algorithms with good generalization performance (soil moisture and above-ground biomass), but are less studied in estimating soil salinity in the visible and near-infrared spectra. In this study, 130 soil EC data, soil measured hyperspectral data, topographic factors, conventional salinity indices such as Salinity Index 1, and two-band (2D) salinity indices such as ratio indices, were introduced. The five spectral pre-processing methods of standard normal variate (SNV), standard normal variate and detrend (SNV-DT), inverse (1/OR) (OR is original spectrum), inverse-log (Log(1/OR) and fractional order derivative (FOD) (range 0–2, with intervals of 0.25) were performed. A gradient boosting machine (GBM) was used to select sensitive spectral parameters. Models (extreme gradient boosting (XGBoost), LightGBM, random forest (RF), ERT, classification and regression tree (CART), and ridge regression (RR)) were used for inversion soil EC and model validation. The results reveal that the two-dimensional correlation coefficient highlighted EC more effectively than the one-dimensional. Under SNV and the second order derivative, the two-dimensional correlation coefficient increased by 0.286 and 0.258 compared to the one-dimension, respectively. The 13 characteristic factors of slope, NDI, SI-T, RI, profile curvature, DOA, plane curvature, SI (conventional), elevation, Int2, aspect, S1 and TWI provided 90% of the cumulative importance for EC using GBM. Among the six machine models, the ERT model performed the best for simulation (R2 = 0.98) and validation (R2 = 0.96). The ERT model showed the best performance among the EC estimation models from the reference data. The kriging map based on the ERT simulation showed a close relationship with the measured data. Our study selected the effective pre-processing methods (SNV and the 2 order derivative) using one- and two-dimensional correlation, 13 important factors and the ERT model for EC hyperspectral inversion. This provides a theoretical support for the quantitative monitoring of soil salinization on a larger scale using remote sensing techniques.

Machine Learning and Remote Sensing Application for Extreme Climate Evaluation: Example of Flood Susceptibility in the Hue Province, Central Vietnam Region

Floods are the most frequent natural hazard globally and incidences have been increasing in recent years as a result of human activity and global warming, making significant impacts on people’s livelihoods and wider socio-economic activities. In terms of the management of the environment and water resources, precise identification is required of areas susceptible to flooding to support planners in implementing effective prevention strategies. The objective of this study is to develop a novel hybrid approach based on Bald Eagle Search (BES), Support Vector Machine (SVM), Random Forest (RF), Bagging (BA) and Multi-Layer Perceptron (MLP) to generate a flood susceptibility map in Thua Thien Hue province, Vietnam. In total, 1621 flood points and 14 predictor variables were used in this study. These data were divided into 60% for model training, 20% for model validation and 20% for testing. In addition, various statistical indices were used to evaluate the performance of the model, such as Root Mean Square Error (RMSE), Receiver Operation Characteristics (ROC), and Mean Absolute Error (MAE). The results show that BES, for the first time, successfully improved the performance of individual models in building a flood susceptibility map in Thua Thien Hue, Vietnam, namely SVM, RF, BA and MLP, with high accuracy (AUC > 0.9). Among the models proposed, BA-BES was most effective with AUC = 0.998, followed by RF-BES (AUC = 0.998), MLP-BES (AUC = 0.998), and SVM-BES (AUC = 0.99). The findings of this research can support the decisions of local and regional authorities in Vietnam and other countries regarding the construction of appropriate strategies to reduce damage to property and human life, particularly in the context of climate change.

Urban Flood-Risk Assessment: Integration of Decision-Making and Machine Learning

Urban flood-risk mapping is an important tool for the mitigation of flooding in view of continuing urbanization and climate change. However, many developing countries lack sufficiently detailed data to produce reliable risk maps with existing methods. Thus, improved methods are needed that can help managers and decision makers to combine existing data with more soft semi-subjective data, such as citizen observations of flood-prone and vulnerable areas in view of existing settlements. Thus, we present an innovative approach using the semi-subjective Analytic Hierarchy Process (AHP), which integrates both subjective and objective assessments, to help organize the problem framework. This approach involves measuring the consistency of decision makers’ judgments, generating pairwise comparisons for choosing a solution, and considering criteria and sub-criteria to evaluate possible options. An urban flood-risk map was created according to the vulnerabilities and hazards of different urban areas using classification and regression-tree models, and the map can serve both as a first stage in advancing flood-risk mitigation approaches and in allocating warning and forecasting systems. The findings show that machine-learning methods are efficient in urban flood zoning. Using the city Rasht in Iran, it is shown that distance to rivers, urban drainage density, and distance to vulnerable areas are the most significant parameters that influence flood hazards. Similarly, for urban flood vulnerability, population density, land use, dwelling quality, household income, distance to cultural heritage, and distance to medical centers and hospitals are the most important factors. The integrated technique for both objective and semi-subjective data as outlined in the present study shows credible results that can be obtained without complicated modeling and costly field surveys. The proposed method is especially helpful in areas with little data to describe and display flood hazards to managers and decision makers.

Accessing the Impact of Meteorological Variables on Machine Learning Flood Susceptibility Mapping

Machine learning (ML) algorithms have emerged as competent tools for identifying areas that are susceptible to flooding. The primary variables considered in most of these works include terrain models, lithology, river networks and land use. While several recent studies include average annual rainfall and/or temperature, other meteorological information such as snow accumulation and short-term intense rain events that may influence the hydrology of the area under investigation have not been considered. Notably, in Canada, most inland flooding occurs during the freshet, due to the melting of an accumulated snowpack coupled with heavy rainfall. Therefore, in this study the impact of several climate variables along with various hydro-geomorphological (HG) variables were tested to determine the impact of their inclusion. Three tests were run: only HG variables, the addition of annual average temperature and precipitation (HG-PT), and the inclusion of six other meteorological datasets (HG-8M) on five study areas across Canada. In HG-PT, both precipitation and temperature were selected as important in every study area, while in HG-8M a minimum of three meteorological datasets were considered important in each study area. Notably, as the meteorological variables were added, many of the initial HG variables were dropped from the selection set. The accuracy, F1, true skill and Area Under the Curve (AUC) were marginally improved when the meteorological data was added to the a parallel random forest algorithm (parRF). When the model is applied to new data, the estimated accuracy of the prediction is higher in HG-8M, indicating that inclusion of relevant, local meteorological datasets improves the result.

Likelihood of Transformation to Green Infrastructure Using Ensemble Machine Learning Techniques in Jinan, China

Rapid urbanization influences green infrastructure (GI) development in cities. The government plans to optimize GI in urban areas, which requires understanding GI spatiotemporal trends in urban areas and driving forces influencing their pattern. Traditional GIS-based methods, used to determine the greening potential of vacant land in urban areas, are incapable of predicting future scenarios based on the past trend. Therefore, we propose a heterogeneous ensemble technique to determine the spatial pattern of GI development in Jinan, China, based on driving biophysical and socioeconomic factors. Data-driven artificial neural networks (ANN) and random forests (RF) are selected as base learners, while support vector machine (SVM) is used as a meta classifier. Results showed that the stacking model ANN-RF-SVM achieved the best test accuracy (AUC 0.941) compared to the individual ANN, RF, and SVM algorithms. Land surface temperature, distance to water bodies, population density, and rainfall are found to be the most influencing factors regarding vacant land conversion to GI in Jinan.

The Evaluation of Color Spaces for Large Woody Debris Detection in Rivers Using XGBoost Algorithm

Large woody debris (LWD) strongly influences river systems, especially in forested and mountainous catchments. In Taiwan, LWD are mainly from typhoons and extreme torrential events. To effectively manage the LWD, it is necessary to conduct regular surveys on river systems. Simple, low cost, and accurate tools are therefore necessary. The proposed methodology applies image processing and machine learning (XGBoost classifier) to quantify LWD distribution, location, and volume in river channels. XGBoost algorithm was selected due to its scalability and faster execution speeds. Nishueibei River, located in Taitung County, was used as the area of investigation. Unmanned aerial vehicles (UAVs) were used to capture the terrain and LWD. Structure from Motion (SfM) was used to build high-resolution orthophotos and digital elevation models (DEM), after which machine learning and different color spaces were used to recognize LWD. Finally, the volume of LWD in the river was estimated. The findings show that RGB color space as LWD recognition factor suffers serious collinearity problems, and it is easy to lose some LWD information; thus, it is not suitable for LWD recognition. On the contrary, the combination of different factors in different color spaces enhances the results, and most of the factors are related to the YCbCr color space. The CbCr factor in the YCbCr color space was best for identifying LWD. LWD volume was then estimated from the identified LWD using manual, field, and automatic measurements. The results indicate that the manual measurement method was the best (R2 = 0.88) to identify field LWD volume. Moreover, automatic measurement (R2 = 0.72) can also obtain LWD volume to save time and workforce.

Poplar's Waterlogging Resistance Modeling and Evaluating: Exploring and Perfecting the Feasibility of Machine Learning Methods in Plant Science

Floods, as one of the most common disasters in the natural environment, have caused huge losses to human life and property. Predicting the flood resistance of poplar can effectively help researchers select seedlings scientifically and resist floods precisely. Using machine learning algorithms, models of poplar's waterlogging tolerance were established and evaluated. First of all, the evaluation indexes of poplar's waterlogging tolerance were analyzed and determined. Then, significance testing, correlation analysis, and three feature selection algorithms (Hierarchical clustering, Lasso, and Stepwise regression) were used to screen photosynthesis, chlorophyll fluorescence, and environmental parameters. Based on this, four machine learning methods, BP neural network regression (BPR), extreme learning machine regression (ELMR), support vector regression (SVR), and random forest regression (RFR) were used to predict the flood resistance of poplar. The results show that random forest regression (RFR) and support vector regression (SVR) have high precision. On the test set, the coefficient of determination (R²) is 0.8351 and 0.6864, the root mean square error (RMSE) is 0.2016 and 0.2780, and the mean absolute error (MAE) is 0.1782 and 0.2031, respectively. Therefore, random forest regression (RFR) and support vector regression (SVR) can be given priority to predict poplar flood resistance.

A Novel Integration of IF-DEMATEL and TOPSIS for the Classifier Selection Problem in Assistive Technology Adoption for People with Dementia

The classifier selection problem in Assistive Technology Adoption refers to selecting the classification algorithms that have the best performance in predicting the adoption of technology, and is often addressed through measuring different single performance indicators. Satisfactory classifier selection can help in reducing time and costs involved in the technology adoption process. As there are multiple criteria from different domains and several candidate classification algorithms, the classifier selection process is now a problem that can be addressed using Multiple-Criteria Decision-Making (MCDM) methods. This paper proposes a novel approach to address the classifier selection problem by integrating Intuitionistic Fuzzy Sets (IFS), Decision Making Trial and Evaluation Laboratory (DEMATEL), and the Technique for Order of Preference by Similarity to Ideal Solution (TOPSIS). The step-by-step procedure behind this application is as follows. First, IF-DEMATEL was used for estimating the criteria and sub-criteria weights considering uncertainty. This method was also employed to evaluate the interrelations among classifier selection criteria. Finally, a modified TOPSIS was applied to generate an overall suitability index per classifier so that the most effective ones can be selected. The proposed approach was validated using a real-world case study concerning the adoption of a mobile-based reminding solution by People with Dementia (PwD). The outputs allow public health managers to accurately identify whether PwD can adopt an assistive technology which results in (i) reduced cost overruns due to wrong classification, (ii) improved quality of life of adopters, and (iii) rapid deployment of intervention alternatives for non-adopters.

Hybrid Models Incorporating Bivariate Statistics and Machine Learning Methods for Flash Flood Susceptibility Assessment Based on Remote Sensing Datasets

Flash floods are considered to be one of the most destructive natural hazards, and they are difficult to accurately model and predict. In this study, three hybrid models were proposed, evaluated, and used for flood susceptibility prediction in the Dadu River Basin. These three hybrid models integrate a bivariate statistical method of the fuzzy membership value (FMV) and three machine learning methods of support vector machine (SVM), classification and regression trees (CART), and convolutional neural network (CNN). Firstly, a geospatial database was prepared comprising nine flood conditioning factors, 485 flood locations, and 485 non-flood locations. Then, the database was used to train and test the three hybrid models. Subsequently, the receiver operating characteristic (ROC) curve, seed cell area index (SCAI), and classification accuracy were used to evaluate the performances of the models. The results reveal the following: (1) The ROC curve highlights the fact that the CNN-FMV hybrid model had the best fitting and prediction performance, and the area under the curve (AUC) values of the success rate and the prediction rate were 0.935 and 0.912, respectively. (2) Based on the results of the three model performance evaluation methods, all three hybrid models had better prediction capabilities than their respective single machine learning models. Compared with their single machine learning models, the AUC values of the SVM-FMV, CART-FMV, and CNN-FMV were 0.032, 0.005, and 0.055 higher; their SCAI values were 0.05, 0.03, and 0.02 lower; and their classification accuracies were 4.48%, 1.38%, and 5.86% higher, respectively. (3) Based on the results of the flood susceptibility indices, between 13.21% and 22.03% of the study area was characterized by high and very high flood susceptibilities. The three hybrid models proposed in this study, especially CNN-FMV, have a high potential for application in flood susceptibility assessment in specific areas in future studies.

How far spatial resolution affects the ensemble machine learning based flood susceptibility prediction in data sparse region

Although the effect of digital elevation model (DEM) and its spatial resolution on flood simulation modeling has been well studied, the effect of coarse and finer resolution image and DEM data on machine learning ensemble flood susceptibility prediction has not been investigated, particularly in data sparse conditions. The present work was, therefore, to investigate the performance of the resolution effects, such as coarse (Landsat and SRTM) and high (Sentinel-2 and ALOS PALSAR) resolution data on the flood susceptible models. Another motive of this study was to construct very high precision and robust flood susceptible models using standalone and ensemble machine learning algorithms. In the present study, fifteen flood conditioning parameters were generated from both coarse and high resolution datasets. Then, the ANN-multilayer perceptron (MLP), random forest (RF), bagging (B)-MLP, B-gaussian processes (B-GP) and B-SMOreg algorithms were used to integrate the flood conditioning parameters for generating the flood susceptible models. Furthermore, the influence of flood conditioning parameters on the modelling of flood susceptibility was investigated by proposing an ROC based sensitivity analysis. The validation of flood susceptibility models is also another challenge. In the present study, we proposed an index of flood vulnerability model to validate flood susceptibility models along with conventional statistical techniques, such as the ROC curve. Results showed that the coarse resolution based flood susceptibility MLP model has appeared as the best model (area under curve: 0.94) and it has predicted 11.65 % of the area as very high flood susceptible zones (FSz), followed by RF, B-MLP, B-GP, and B-SMOreg. Similarly, the high resolution based flood susceptibility model using MLP has predicted 19.34 % of areas as very high flood susceptible zones, followed by RF (14.32 %),B-MLP (14.88 %), B-GP, and B-SMOreg. On the other hand, ROC based sensitivity analysis showed that elevation influences flood susceptibility largely for coarse and high resolution based models, followed by drainage densityand flow accumulation. In addition, the accuracy assessment using the IFV model revealed that the MLP model outperformed all other models in the case of a high resolution imageThe coarser resolution image's performance level is acceptable but quite low. So, the study recommended the use of high resolution images for developing a machine learning algorithm based flood susceptibility model. As the study has clearly identified the areas of higher flood susceptibility and the dominant influencing factors for flooding, this could be used as a good database for flood management.

Data-driven flood hazard zonation of Italy

We present Flood-SHE, a data-driven, statistically-based procedure for the delineation of areas expected to be inundated by river floods. We applied Flood-SHE in the 23 River Basin Authorities (RBAs) in Italy using information on the presence or absence of inundations obtained from existing flood zonings as the dependent variable, and six hydro-morphometric variables computed from a 10 m × 10 m DEM as covariates. We trained 96 models for each RBA using 32 combinations of the hydro-morphometric covariates for the three return periods, for a total of 2208 models, which we validated using 32 model sets for each of the covariate combinations and return periods, for a total of 3072 validation models. In all the RBAs, Flood-SHE delineated accurately potentially inundated areas that matched closely the corresponding flood zonings defined by physically-based hydro-dynamic flood routing and inundation models. Flood-SHE delineated larger to much larger areas as potentially subject of being inundated than the physically-based models, depending on the quality of the flood information. Analysis of the sites with flood human consequences revealed that the new data-driven inundation zones are good predictors of flood risk to the population of Italy. Our experiment confirmed that a small number of hydro-morphometric terrain variables is sufficient to delineate accurate inundation zonings in a variety of physiographical settings, opening to the possibility of using Flood-SHE in other areas. We expect the new data-driven inundation zonings to be useful where flood zonings built on hydrological modelling are not available, and to decide where improved flood hazard zoning is needed.

A multi-hazard map-based flooding, gully erosion, forest fires, and earthquakes in Iran

We used three state-of-the-art machine learning techniques (boosted regression tree, random forest, and support vector machine) to produce a multi-hazard (MHR) map illustrating areas susceptible to flooding, gully erosion, forest fires, and earthquakes in Kohgiluyeh and Boyer-Ahmad Province, Iran. The earthquake hazard map was derived from a probabilistic seismic hazard analysis. The mean decrease Gini (MDG) method was implemented to determine the relative importance of effective factors on the spatial occurrence of each of the four hazards. Area under the curve (AUC) plots, based on a validation dataset, were created for the maps generated using the three algorithms to compare the results. The random forest model had the highest predictive accuracy, with AUC values of 0.994, 0.982, and 0.885 for gully erosion, flooding, and forest fires, respectively. Approximately 41%, 40%, 28%, and 3% of the study area are at risk of forest fires, earthquakes, floods, and gully erosion, respectively.

An Integrated Approach for Post-Disaster Flood Management Via the Use of Cutting-Edge Technologies and UAVs: A Review

Rapid advances that improve flood management have facilitated the disaster response by providing first aid services, finding safe routes, maintaining communication and developing flood maps. Different technologies such as image processing, satellite imagery, synthetic imagery and integrated approaches have been extensively analysed in the literature for disaster operations. There is a need to review cutting-edge technologies for flood management. This paper presents a review of the latest advancements in the flood management domain based on image processing, artificial intelligence and integrated approaches with a focus on post-disaster. It answers the following research questions: (1) What are the latest developments in image processing for flood management in a post-disaster scenario? (2) What are the latest techniques for flood management based on artificial intelligence in a post-disaster scenario? (3) What are the existing gaps in the selected technologies for post-disaster? (4) How can the authorities improve the existing post-disaster management operation with cutting-edge technologies? A novel framework has been proposed to optimise flood management with the application of a holistic approach.

Combining machine learning models through multiple data division methods for PM2.5 forecasting in Northern Xinjiang, China

In this study, daily average PM_2.5 forecasting models were developed and applied in the Northern Xinjiang, China, through combining the back propagation artificial neural network (BPANN) and multiple linear regression (MLR) with another BPANN model. The meteorological (daily average precipitation, pressure, relative humidity, temperature, and wind speed, daily maximum wind speed and sunshine hours on the same day) and air pollutant data (daily PM_2.5, PM₁₀, SO₂, CO, NO₂, and O₃ concentrations on the previous day) in January and August of each year from 2015 to 2019 were used as candidate inputs. The optimal member and combining models were evaluated through the leave-one-out cross-validation (LOOCV), fivefold cross-validation, and hold-out methods. Twelve member models with optimal or sub-optimal performance were further used to develop the combining models. The performances of the BPANN and MLR member models were different using three data division methods. The models were evaluated more comprehensively through the LOOCV. The performances of the combining models were generally better than the member models. For both member and combining models, the PM_2.5 forecasting model performance in August was generally better than in January. The correlation coefficient (R) for the validation set of the optimal combination model was about 0.87 in January and 0.946 in August. These results showed that combining linear and nonlinear models through multiple data division methods would be an effective tool to forecast PM_2.5 concentrations.

Flood susceptibility mapping by integrating frequency ratio and index of entropy with multilayer perceptron and classification and regression tree

Episodes of frequent flooding continue to increase, often causing serious damage and tools to identify areas affected by such disasters have become indispensable in today's society. Using the latest techniques can make very accurate flood predictions. In this study, we introduce four effective methods to evaluate the flood susceptibility of Poyang County, in China, by integrating two independent models of frequency ratio and index of entropy with multilayer perceptron and classification and regression tree models. The flood locations of the study area were identified through the flood inventory process, and 12 flood conditioning factors were used in the training and validation processes. According to the results of the linear support vector machine, elevation, slope angle, and soil have the highest predictive ability. The experimental results of the four hybrid models demonstrate that between 20% and 50% of the study area has high and very high flood susceptibility. The multilayer perceptron-probability density hybrid model is the most effective among the six comparative methods.

Nonlinear relationship of greening and shifts from greening to browning in vegetation with nature and human factors along the Silk Road Economic Belt

Understanding the trends in vegetation evolution and their driving factors is critical to revealing changes in ecosystem's structure and function. However, less is known about the nonlinear relationship between greening trends and greening to browning shifts in vegetation and its driving factors. Based on ensemble empirical mode decomposition (EEMD) and Boosted Regression Tree (BRT) methods, we investigated the nonlinear relationships of the greening trends and greening to browning shifts in vegetation to climate change, elevation, and human activities along the Silk Road Economic Belt (SREB). Results indicated that: (1) For vegetation with greening trends, although the total importance of climate-related variables was greater than human activities (59.57% and 35.33%, respectively), Land use changes (14.91%) and CO₂ (9.41%) were more important than the other nature-related variables. In addition, volumetric soil moisture (8.72%), precipitation of warmest quarter (6.50%), and daily snow water equivalent (5.75%) were also very important. For vegetation with greening to browning shifts, the total contributions of climate-related variables were also larger than human activities (69.10% and 19.36%, respectively). Moreover, elevation (11.54%), vapor pressure deficit (9.85%), annual mean temperature (6.84%), min temperature of the coldest month (5.98%), soil moisture (5.16%), and annual precipitation (5.11%) might be the main reasons. (2) Increased cropland and grassland, along with the land conversions from shrub land/sparse vegetated to forest could promote vegetation greening. Increased rates of CO₂, volumetric soil moisture, warmest quarter precipitation, and daily snow water equivalent could also promote it. For vegetation with greening to browning shifts, low-altitude areas, increased vapor pressure deficit, decreased cold temperature, and drought could promote vegetation shift greening to browning. The nonlinear analysis can correctly reveal the actual trends in vegetation and its responses to driving factors, which will further provide adaptation strategies to protect ecology.

Performance Evaluation of Sentinel-2 and Landsat 8 OLI Data for Land Cover/Use Classification Using a Comparison between Machine Learning Algorithms

With the development of remote sensing algorithms and increased access to satellite data, generating up-to-date, accurate land use/land cover (LULC) maps has become increasingly feasible for evaluating and managing changes in land cover as created by changes to ecosystem and land use. The main objective of our study is to evaluate the performance of Support Vector Machine (SVM), Artificial Neural Network (ANN), Maximum Likelihood Classification (MLC), Minimum Distance (MD), and Mahalanobis (MH) algorithms and compare them in order to generate a LULC map using data from Sentinel 2 and Landsat 8 satellites. Further, we also investigate the effect of a penalty parameter on SVM results. Our study uses different kernel functions and hidden layers for SVM and ANN algorithms, respectively. We generated the training and validation datasets from Google Earth images and GPS data prior to pre-processing satellite data. In the next phase, we classified the images using training data and algorithms. Ultimately, to evaluate outcomes, we used the validation data to generate a confusion matrix of the classified images. Our results showed that with optimal tuning parameters, the SVM classifier yielded the highest overall accuracy (OA) of 94%, performing better for both satellite data compared to other methods. In addition, for our scenes, Sentinel 2 date was slightly more accurate compared to Landsat 8. The parametric algorithms MD and MLC provided the lowest accuracy of 80.85% and 74.68% for the data from Sentinel 2 and Landsat 8. In contrast, our evaluation using the SVM tuning parameters showed that the linear kernel with the penalty parameter 150 for Sentinel 2 and the penalty parameter 200 for Landsat 8 yielded the highest accuracies. Further, ANN classification showed that increasing the hidden layers drastically reduces classification accuracy for both datasets, reducing zero for three hidden layers.

An Overview of Flood Risk Analysis Methods

Scientific papers present a wide range of methods of flood analysis and forecasting. Floods are a phenomenon with significant socio-economic implications, for which many researchers try to identify the most appropriate methodologies to analyze their temporal and spatial development. This research aims to create an overview of flood analysis and forecasting methods. The study is based on the need to select and group papers into well-defined methodological categories. The article provides an overview of recent developments in the analysis of flood methodologies and shows current research directions based on this overview. The study was performed taking into account the information included in the Web of Science Core Collection, which brought together 1326 articles. The research concludes with a discussion on the relevance, ease of application, and usefulness of the methodologies.

Spatial Prediction of Future Flood Risk: An Approach to the Effects of Climate Change

Preparation of a flood probability map serves as the first step in a flood management program. This research develops a probability flood map for floods resulting from climate change in the future. Two models of Flexible Discrimination Analysis (FDA) and Artificial Neural Network (ANN) were used. Two optimistic (RCP2.6) and pessimistic (RCP8.5) climate change scenarios were considered for mapping future rainfall. Moreover, to produce probability flood occurrence maps, 263 locations of past flood events were used as dependent variables. The number of 13 factors conditioning floods was taken as independent variables in modeling. Of the total 263 flood locations, 80% (210 locations) and 20% (53 locations) were considered model training and validation. The Receiver Operating Characteristic (ROC) curve and other statistical criteria were used to validate the models. Based on assessments of the validated models, FDA, with a ROC-AUC = 0.918, standard error (SE = 0.038), and an accuracy of 0.86% compared to the ANN model with a ROC-AUC = 0.897, has the highest accuracy in preparing the flood probability map in the study area. The modeling results also showed that the factors of distance from the River, altitude, slope, and rainfall have the greatest impact on floods in the study area. Both models’ future flood susceptibility maps showed that the highest area is related to the very low class. The lowest area is related to the high class.

An AI-Application-Oriented In-Class Teaching Evaluation Model by Using Statistical Modeling and Ensemble Learning

In-class teaching evaluation, which is utilized to assess the process and effect of both teachers’ teaching and students’ learning in a classroom environment, plays an increasingly crucial role in supervising and promoting education quality. With the rapid development of artificial intelligence (AI) technology, the concept of smart education has been constantly improved and gradually penetrated into all aspects of education application. Considering the dominant position of classroom teaching in elementary and undergraduate education, the introduction of AI technology into in-class teaching evaluation has become a research hotspot. In this paper, we propose a statistical modeling and ensemble learning-based comprehensive model, which is oriented towards in-class teaching evaluation by using AI technologies such as computer vision (CV) and intelligent speech recognition (ISR). Firstly, we present an index system including a set of teaching evaluation indicators combining traditional assessment scales with new values derived from CV and ISR-based AI analysis. Next, we design a comprehensive in-class teaching evaluation model by using both the analytic hierarchy process-entropy weight (AHP-EW) and AdaBoost-based ensemble learning (AdaBoost-EL) methods. Experiments not only demonstrate that the two modules in the model are respectively applicable to the calculation of indicators with different characteristics, but also verify the performance of the proposed model for AI-based in-class teaching evaluation. In this comprehensive in-class evaluation model, for students’ concentration and participation, ensemble learning module is chosen with less root mean square error (RMSE) of 8.318 and 9.375. In addition, teachers’ media usage and teachers’ type evaluated by statistical modeling module approach higher accuracy with 0.905 and 0.815. Instead, the ensemble learning approaches the accuracy of 0.73 in evaluating teachers’ style, which performs better than the statistical modeling module with the accuracy of 0.69.

Flash Flood Susceptibility Modeling Using New Approaches of Hybrid and Ensemble Tree-Based Machine Learning Algorithms

Flash flooding is considered one of the most dynamic natural disasters for which measures need to be taken to minimize economic damages, adverse effects, and consequences by mapping flood susceptibility. Identifying areas prone to flash flooding is a crucial step in flash flood hazard management. In the present study, the Kalvan watershed in Markazi Province, Iran, was chosen to evaluate the flash flood susceptibility modeling. Thus, to detect flash flood-prone zones in this study area, five machine learning (ML) algorithms were tested. These included boosted regression tree (BRT), random forest (RF), parallel random forest (PRF), regularized random forest (RRF), and extremely randomized trees (ERT). Fifteen climatic and geo-environmental variables were used as inputs of the flash flood susceptibility models. The results showed that ERT was the most optimal model with an area under curve (AUC) value of 0.82. The rest of the models’ AUC values, i.e., RRF, PRF, RF, and BRT, were 0.80, 0.79, 0.78, and 0.75, respectively. In the ERT model, the areal coverage for very high to moderate flash flood susceptible area was 582.56 km2 (28.33%), and the rest of the portion was associated with very low to low susceptibility zones. It is concluded that topographical and hydrological parameters, e.g., altitude, slope, rainfall, and the river’s distance, were the most effective parameters. The results of this study will play a vital role in the planning and implementation of flood mitigation strategies in the region.

Predicting flood insurance claims with hydrologic and socioeconomic demographics via machine learning: Exploring the roles of topography, minority populations, and political dissimilarity

Current research on flooding risk often focuses on understanding hazards, de-emphasizing the complex pathways of exposure and vulnerability. We investigated the use of both hydrologic and social demographic data for flood exposure mapping with Random Forest (RF) regression and classification algorithms trained to predict both parcel- and tract-level flood insurance claims within New York State, US. Topographic characteristics best described flood claim frequency, but RF prediction skill was improved at both spatial scales when socioeconomic data was incorporated. Substantial improvements occurred at the tract-level when the percentage of minority residents, housing stock value and age, and the political dissimilarity index of voting precincts were used to predict insurance claims. Census tracts with higher numbers of claims and greater densities of low-lying tax parcels tended to have low proportions of minority residents, newer houses, and less political similarity to state level government. We compared this data-driven approach and a physically-based pluvial flood routing model for prediction of the spatial extents of flooding claims in two nearby catchments of differing land use. The floodplain we defined with physically based modeling agreed well with existing federal flood insurance rate maps, but underestimated the spatial extents of historical claim generating areas. In contrast, RF classification incorporating hydrologic and socioeconomic demographic data likely overestimated the flood-exposed areas. Our research indicates that quantitative incorporation of social data can improve flooding exposure estimates.

On the Operational Flood Forecasting Practices Using Low-Quality Data Input of a Distributed Hydrological Model

Low-quality input data (such as sparse rainfall gauges, low spatial resolution soil type and land use maps) have limited the application of physically-based distributed hydrological models in operational practices in many data-sparse regions. It is necessary to quantify the uncertainty in the deterministic forecast results of distributed models. In this paper, the TOPographic Kinematic Approximation and Integration (TOPKAPI) distributed model was used for deterministic forecasts with low-quality input data, and then the Hydrologic Uncertainty Processor (HUP) was used to provide the probabilistic forecast results for operational practices. Results showed that the deterministic forecasts by TOPKAPI performed poorly in some flood seasons, such as the years 1997, 2001 and 2008, despite which the overall accuracy of the whole study period 1996–2008 could be acceptable and generally reproduced the hydrological behaviors of the catchment (Lushi basin, China). The HUP model can not only provide probabilistic forecasts (e.g., 90% predictive uncertainty bounds), but also provides deterministic forecasts in terms of 50% percentiles. The 50% percentiles obviously improved the forecast accuracy of selected flood events at the leading time of one hour. Besides, the HUP performance decayed with the leading time increasing (6, 12 h). This work revealed that deterministic model outputs had large uncertainties in flood forecasts, and the HUP model may provide an alternative for operational flood forecasting practices in those areas with low-quality data.

Daily Water Level Prediction of Zrebar Lake (Iran): A Comparison between M5P, Random Forest, Random Tree and Reduced Error Pruning Trees Algorithms

Zrebar Lake is one of the largest freshwater lakes in Iran and it plays an important role in the ecosystem of the environment, while its desiccation has a negative impact on the surrounded ecosystem. Despite this, this lake provides an interesting recreation setting in terms of ecotourism. The prediction and forecasting of the water level of the lake through simple but practical methods can provide a reliable tool for future lake water resource management. In the present study, we predict the daily water level of Zrebar Lake in Iran through well-known decision tree-based algorithms, including the M5 pruned (M5P), random forest (RF), random tree (RT) and reduced error pruning tree (REPT). We used five different water input combinations to find the most effective one. For our modeling, we chose 70% of the dataset for training (from 2011 to 2015) and 30% for model evaluation (from 2015 to 2017). We evaluated the models’ performances using different quantitative (root mean square error (RMSE), mean absolute error (MAE), coefficient of determination (R2), percent bias (PBIAS) and ratio of the root mean square error to the standard deviation of measured data (RSR)) and visual frameworks (Taylor diagram and box plot). Our results showed that water level with a one-day lag time had the highest effect on the result and, by increasing the lag time, its effect on the result was decreased. This result indicated that all the developed models had a good prediction capability, but the M5P model outperformed the others, followed by RF and RT equally and then REPT. Our results showed that these algorithms can predict water level accurately only with a one-day lag time in water level as an input and they are cost-effective tools for future predictions.

Landslide Detection and Susceptibility Modeling on Cameron Highlands (Malaysia): A Comparison between Random Forest, Logistic Regression and Logistic Model Tree Algorithms

We used remote sensing techniques and machine learning to detect and map landslides, and landslide susceptibility in the Cameron Highlands, Malaysia. We located 152 landslides using a combination of interferometry synthetic aperture radar (InSAR), Google Earth (GE), and field surveys. Of the total slide locations, 80% (122 landslides) were utilized for training the selected algorithms, and the remaining 20% (30 landslides) were applied for validation purposes. We employed 17 conditioning factors, including slope angle, aspect, elevation, curvature, profile curvature, stream power index (SPI), topographic wetness index (TWI), lithology, soil type, land cover, normalized difference vegetation index (NDVI), distance to river, distance to fault, distance to road, river density, fault density, and road density, which were produced from satellite imageries, geological map, soil maps, and a digital elevation model (DEM). We used these factors to produce landslide susceptibility maps using logistic regression (LR), logistic model tree (LMT), and random forest (RF) models. To assess prediction accuracy of the models we employed the following statistical measures: negative predictive value (NPV), sensitivity, positive predictive value (PPV), specificity, root-mean-squared error (RMSE), accuracy, and area under the receiver operating characteristic (ROC) curve (AUC). Our results indicated that the AUC was 92%, 90%, and 88% for the LMT, LR, and RF algorithms, respectively. To assess model performance, we also applied non-parametric statistical tests of Friedman and Wilcoxon, where the results revealed that there were no practical differences among the used models in the study area. While landslide mapping in tropical environment such as Cameron Highlands remains difficult, the remote sensing (RS) along with machine learning techniques, such as the LMT model, show promise for landslide susceptibility mapping in the study area.

Landslide Susceptibility Mapping Using Machine Learning Algorithms and Remote Sensing Data in a Tropical Environment

We used AdaBoost (AB), alternating decision tree (ADTree), and their combination as an ensemble model (AB-ADTree) to spatially predict landslides in the Cameron Highlands, Malaysia. The models were trained with a database of 152 landslides compiled using Synthetic Aperture Radar Interferometry, Google Earth images, and field surveys, and 17 conditioning factors (slope, aspect, elevation, distance to road, distance to river, proximity to fault, road density, river density, normalized difference vegetation index, rainfall, land cover, lithology, soil types, curvature, profile curvature, stream power index, and topographic wetness index). We carried out the validation process using the area under the receiver operating characteristic curve (AUC) and several parametric and non-parametric performance metrics, including positive predictive value, negative predictive value, sensitivity, specificity, accuracy, root mean square error, and the Friedman and Wilcoxon sign rank tests. The AB model (AUC = 0.96) performed better than the ensemble AB-ADTree model (AUC = 0.94) and successfully outperformed the ADTree model (AUC = 0.59) in predicting landslide susceptibility. Our findings provide insights into the development of more efficient and accurate landslide predictive models that can be used by decision makers and land-use managers to mitigate landslide hazards.

Novel hybrid models between bivariate statistics, artificial neural networks and boosting algorithms for flood susceptibility assessment

Across the world, the flood magnitude is expected to increase as well as the damage caused by their occurrence. In this case, the prediction of areas which are highly susceptible to these phenomena becomes very important for the authorities. The present study is focused on the evaluation of flood potential within Trotuș river basin in Romania using six ensemble models created by the combination of Analytical Hierarchy Process (AHP), Certainty Factor (CF) and Weights of Evidence (WOE) on one hand, and Gradient Boosting Trees (GBT) and Multilayer Perceptron (MLP) on the other hand. A number of 12 flood predictors, 172 flood locations and 172 non-flood locations were used. A percentage of 70% of flood and non-flood locations were used as input in models. From the input data, 70% were used as training sample and 30% as validating sample. The highest accuracy was obtained by the MLP-CF model in terms of both training (0.899) and testing (0.889) samples. A percentage between 21.88% and 36.33% of study area is covered with high and very high flood potential. The results validation, performed through the ROC Curve method, highlights that the MLP-CF model provided the most accurate results.

Integration of GIS-Based Multicriteria Decision Analysis and Analytic Hierarchy Process to Assess Flood Hazard on the Al-Shamal Train Pathway in Al-Qurayyat Region, Kingdom of Saudi Arabia

Understanding the dynamics of floods in dry environments and predicting an accurate flood hazard map considering multiple standards and conflicting objectives is of great political and planning importance in the Kingdom of Saudi Arabia’s vision for the year 2030, in order to reduce losses in lives, property, and infrastructure. The objectives of this study are (1) to develop a flood vulnerability map identifying flood-prone areas along the Al-Shamal train railway pathway; (2) to forecast the vulnerability of urban areas, agricultural land, and infrastructure to possible future floods hazard; and (3) to introduce strategic solutions and recommendations to mitigate and protect such areas from the negative impacts of floods. In order to achieve these objectives, multicriteria decision analysis based on geographic information systems (GIS-MCDA) is used to build a flood hazard map of the study area. The analytic hierarchy process (AHP) is applied to extract the weights of eight criteria which affect the areas which are prone to flooding hazards, including flow accumulation, distance from the wadi network, slope, rainfall density, drainage density, and rainfall speed. Furthermore, the receiver operating characteristic (ROC Curve) method is used to validate the presented flood hazard model. The results of the study reveal that there are five degrees of flooding hazard along the Al-Shamal train path, ranging from very high to very low. The high and very high hazard zones comprise 19.2 km along the path, which constitutes about 26.45% of the total path length, and are concentrated at the intersections of the Al-Shamal train pathway with the Bayer and Al-Makhrouk wadis. Moderate, low, and very low flood severity areas constitute nearly 53.39 km, representing 73.55% of the total length (72.59 km) of the track. These areas are concentrated at the intersection of the Al-Shamal train track with the Haseidah Al-Gharbiyeh and Hsaidah Umm Al-Nakhleh wadis. Urban and agricultural areas that are vulnerable to high and very high flooding hazards are shown to have areas of 29.23 km2 (22.12%) and 59.87 km2 (46.39%), respectively.

Novel Ensemble Landslide Predictive Models Based on the Hyperpipes Algorithm: A Case Study in the Nam Dam Commune, Vietnam

Development of landslide predictive models with strong prediction power has become a major focus of many researchers. This study describes the first application of the Hyperpipes (HP) algorithm for the development of the five novel ensemble models that combine the HP algorithm and the AdaBoost (AB), Bagging (B), Dagging, Decorate, and Real AdaBoost (RAB) ensemble techniques for mapping the spatial variability of landslide susceptibility in the Nam Dan commune, Ha Giang province, Vietnam. Information on 76 historical landslides and ten geo-environmental factors (slope degree, slope aspect, elevation, topographic wetness index, curvature, weathering crust, geology, river density, fault density, and distance from roads) were used for the construction of the training and validation datasets that are the prerequisites for building and testing the proposed models. Using different performance metrics (i.e., the area under the receiver operating characteristic curve (AUC), negative predictive value, positive predictive value, accuracy, sensitivity, specificity, root mean square error, and Kappa), we verified the proficiency of all five ensemble learning techniques in increasing the fitness and predictive powers of the base HP model. Based on the AUC values derived from the models, the ensemble ABHP model that yielded an AUC value of 0.922 was identified as the most efficient model for mapping the landslide susceptibility in the Nam Dan commune, followed by RABHP (AUC = 0.919), BHP (AUC = 0.909), Dagging-HP (AUC = 0.897), Decorate-HP (AUC = 0.865), and the single HP model (AUC = 0.856), respectively. The novel ensemble models proposed for the Nam Dan commune and the resultant susceptibility maps can aid land-use planners in the development of efficient mitigation strategies in response to destructive landslides.

Shallow Landslide Susceptibility Mapping: A Comparison between Logistic Model Tree, Logistic Regression, Naïve Bayes Tree, Artificial Neural Network, and Support Vector Machine Algorithms

Shallow landslides damage buildings and other infrastructure, disrupt agriculture practices, and can cause social upheaval and loss of life. As a result, many scientists study the phenomenon, and some of them have focused on producing landslide susceptibility maps that can be used by land-use managers to reduce injury and damage. This paper contributes to this effort by comparing the power and effectiveness of five machine learning, benchmark algorithms-Logistic Model Tree, Logistic Regression, Naïve Bayes Tree, Artificial Neural Network, and Support Vector Machine-in creating a reliable shallow landslide susceptibility map for Bijar City in Kurdistan province, Iran. Twenty conditioning factors were applied to 111 shallow landslides and tested using the One-R attribute evaluation (ORAE) technique for modeling and validation processes. The performance of the models was assessed by statistical-based indexes including sensitivity, specificity, accuracy, mean absolute error (MAE), root mean square error (RMSE), and area under the receiver operatic characteristic curve (AUC). Results indicate that all the five machine learning models performed well for shallow landslide susceptibility assessment, but the Logistic Model Tree model (AUC = 0.932) had the highest goodness-of-fit and prediction accuracy, followed by the Logistic Regression (AUC = 0.932), Naïve Bayes Tree (AUC = 0.864), ANN (AUC = 0.860), and Support Vector Machine (AUC = 0.834) models. Therefore, we recommend the use of the Logistic Model Tree model in shallow landslide mapping programs in semi-arid regions to help decision makers, planners, land-use managers, and government agencies mitigate the hazard and risk.

Shallow Landslide Susceptibility Mapping by Random Forest Base Classifier and Its Ensembles in a Semi-Arid Region of Iran

We generated high-quality shallow landslide susceptibility maps for Bijar County, Kurdistan Province, Iran, using Random Forest (RAF), an ensemble computational intelligence method and three meta classifiers—Bagging (BA, BA-RAF), Random Subspace (RS, RS-RAF), and Rotation Forest (RF, RF-RAF). Modeling and validation were done on 111 shallow landslide locations using 20 conditioning factors tested by the Information Gain Ratio (IGR) technique. We assessed model performance with statistically based indexes, including sensitivity, specificity, accuracy, kappa, root mean square error (RMSE), and area under the receiver operatic characteristic curve (AUC). All four machine learning models that we tested yielded excellent goodness-of-fit and prediction accuracy, but the RF-RAF ensemble model (AUC = 0.936) outperformed the BA-RAF, RS-RAF (AUC = 0.907), and RAF (AUC = 0.812) models. The results also show that the Random Forest model significantly improved the predictive capability of the RAF-based classifier and, therefore, can be considered as a useful and an effective tool in regional shallow landslide susceptibility mapping.

Comparative assessment of the flash-flood potential within small mountain catchments using bivariate statistics and their novel hybrid integration with machine learning models

The present study is carried out in the context of the continuous increase, worldwide, of the number of flash-floods phenomena. Also, there is an evident increase of the size of the damages caused by these hazards. Bâsca Chiojdului River Basin is one of the most affected areas in Romania by flash-flood phenomena. Therefore, Flash-Flood Potential Index (FFPI) was defined and calculated across the Bâsca Chiojdului river basin by using one bivariate statistical method (Statistical Index) and its novel ensemble with the following machine learning models: Logistic Regression, Classification and Regression Trees, Multilayer Perceptron, Random Forest and Support Vector Machine and Decision Tree CART. In a first stage, the areas with torrentiality were digitized based on orthophotomaps and field observations. These regions, together with an equal number of non-torrential pixels, were further divided into training surfaces (70%) and validating surfaces (30%). The next step of the analysis consisted of the selection of flash-flood conditioning factors based on the multicollinearity investigation and predictive ability estimation through Information Gain method. Eight factors, from a total of ten flash-floods predictors, were selected in order to be included in the FFPI calculation process. By applying the models represented by Statistical Index and its ensemble with the machine learning algorithms, the weight of each conditioning factor and of each factor class/category in the FFPI equations was established. Once the weight values were derived, the FFPI values across the Bâsca Chiojdului river basin were calculated by overlaying the flash-flood predictors in GIS environment. According to the results obtained, the central part of Bâsca Chiojdului river basin has the highest susceptibility to flash-flood phenomena. Thus, around 30% of the study site has high and very high values of FFPI. The results validation was carried out by applying the Prediction Rate and Success Rate. The methods revealed the fact that the Multilayer Perceptron - Statistical Index (MLP-SI) ensemble has the highest efficiency among the 3 methods.

Flash-flood hazard assessment using ensembles and Bayesian-based machine learning models: Application of the simulated annealing feature selection method

Flash-floods are increasingly recognized as a frequent natural hazard worldwide. Iran has been among the mostdevastated regions affected by the major floods. While the temporal flash-flood forecasting models are mainly developed for warning systems, the models for assessing hazardous areas can greatly contribute to adaptation and mitigation policy-making and disaster risk reduction. Former researches in the flash-flood hazard mapping have heightened the urge for the advancement of more accurate models. Thus, the current research proposes the state-of-the-art ensemble models of boosted generalized linear model (GLMBoost) and random forest (RF), and Bayesian generalized linear model (BayesGLM) methods for higher performance modeling. Furthermore, a pre-processing method, namely simulated annealing (SA), is used to eliminate redundant variables from the modeling process. Results of the modeling based on the hit and miss analysis indicates high performance for both models (accuracy = 90-92%, Kappa = 79-84%, Success ratio = 94-96%, Threat score = 80-84%, and Heidke skill score = 79-84%). The variables of distance from the stream, vegetation, drainage density, land use, and elevation have shown more contribution among others for modeling the flash-flood. The results of this study can significantly facilitate mapping the hazardous areas and further assist watershed managers to control and remediate induced damages of flood in the data-scarce regions.

Mapping of Groundwater Spring Potential in Karst Aquifer System Using Novel Ensemble Bivariate and Multivariate Models

Groundwater is an important natural resource in arid and semi-arid environments, where discharge from karst springs is utilized as the principal water supply for human use. The occurrence of karst springs over large areas is often poorly documented, and interpolation strategies are often utilized to map the distribution and discharge potential of springs. This study develops a novel method to delineate karst spring zones on the basis of various hydrogeological factors. A case study of the Bojnourd Region, Iran, where spring discharge measurements are available for 359 sites, is used to demonstrate application of the new approach. Spatial mapping is achieved using ensemble modelling, which is based on certainty factors (CF) and logistic regression (LR). Maps of the CF and LR components of groundwater potential were generated individually, and then, combined to prepare an ensemble map of the study area. The accuracy (A) of the ensemble map was then assessed using area under the receiver operating characteristic curve. Results of this analysis show that LR (A = 78%) outperformed CF (A = 67%) in terms of the comparison between model predictions and known occurrences of karst springs (i.e., calibration data). However, combining the CF and LR results through ensemble modelling produced superior accuracy (A = 85%) in terms of spring potential mapping. By combining CF and LR statistical models through ensemble modelling, weaknesses in CF and LR methods are offset, and therefore, we recommend this ensemble approach for similar karst mapping projects. The methodology developed here offers an efficient method for assessing spring discharge and karst spring potentials over regional scales.

GIS-based flood hazard mapping using relative frequency ratio method: A case study of Panjkora River Basin, eastern Hindu Kush, Pakistan

Flood is the most devastating and prevalent disaster among all-natural disasters. Every year, flood claims hundreds of human lives and causes damage to the worldwide economy and environment. Consequently, the identification of flood-vulnerable areas is important for comprehensive flood risk management. The main objective of this study is to delineate flood-prone areas in the Panjkora River Basin (PRB), eastern Hindu Kush, Pakistan. An initial extensive field survey and interpretation of Landsat-7 and Google Earth images identified 154 flood locations that were inundated in 2010 floods. Of the total, 70% of flood locations were randomly used for building a model and 30% were used for validation of the model. Eight flood parameters including slope, elevation, land use, Normalized Difference Vegetation Index (NDVI), topographic wetness index (TWI), drainage density, and rainfall were used to map the flood-prone areas in the study region. The relative frequency ratio was used to determine the correlation between each class of flood parameter and flood occurrences. All of the factors were resampled into a pixel size of 30×30 m and were reclassified through the natural break method. Finally, a final hazard map was prepared and reclassified into five classes, i.e., very low, low, moderate, high, very high susceptibility. The results of the model were found reliable with area under curve values for success and prediction rate of 82.04% and 84.74%, respectively. The findings of this study can play a key role in flood hazard management in the target region; they can be used by the local disaster management authority, researchers, planners, local government, and line agencies dealing with flood risk management.

GIS-Based Gully Erosion Susceptibility Mapping: A Comparison of Computational Ensemble Data Mining Models

Gully erosion destroys agricultural and domestic grazing land in many countries, especially those with arid and semi-arid climates and easily eroded rocks and soils. It also generates large amounts of sediment that can adversely impact downstream river channels. The main objective of this research is to accurately detect and predict areas prone to gully erosion. In this paper, we couple hybrid models of a commonly used base classifier (reduced pruning error tree, REPTree) with AdaBoost (AB), bagging (Bag), and random subspace (RS) algorithms to create gully erosion susceptibility maps for a sub-basin of the Shoor River watershed in northwestern Iran. We compare the performance of these models in terms of their ability to predict gully erosion and discuss their potential use in other arid and semi-arid areas. Our database comprises 242 gully erosion locations, which we randomly divided into training and testing sets with a ratio of 70/30. Based on expert knowledge and analysis of aerial photographs and satellite images, we selected 12 conditioning factors for gully erosion. We used multi-collinearity statistical techniques in the modeling process, and checked model performance using statistical indexes including precision, recall, F-measure, Matthew correlation coefficient (MCC), receiver operatic characteristic curve (ROC), precision–recall graph (PRC), Kappa, root mean square error (RMSE), relative absolute error (PRSE), mean absolute error (MAE), and relative absolute error (RAE). Results show that rainfall, elevation, and river density are the most important factors for gully erosion susceptibility mapping in the study area. All three hybrid models that we tested significantly enhanced and improved the predictive power of REPTree (AUC=0.800), but the RS-REPTree (AUC= 0.860) ensemble model outperformed the Bag-REPTree (AUC= 0.841) and the AB-REPTree (AUC= 0.805) models. We suggest that decision makers, planners, and environmental engineers employ the RS-REPTree hybrid model to better manage gully erosion-prone areas in Iran.