Search

Dam failure may lead to significant casualties among downstream residents. Therefore, it is crucial to study a reliable method to quantitatively assess the loss of life (LOL) caused by dam failure for emergency response to dam failure incidents. Based on a statistical analysis of typical dam failure accidents in China and the research on the formation mechanism of LOL, the study quantified the factors influencing LOL using fuzzy theory and constructed a quantitative database for the LOL. Then, it proposed an innovative algorithm integrating the grey wolf optimization (GWO) algorithm and the random forest (RF) model. Finally, a data-driven assessment model for the LOL caused by dam failure was developed by combining the gray correlation analysis of the factors. The performance of the GWO-RF model was validated using a dataset of the LOL caused. The proposed model was used to assess the LOL in typical dam failure events. The results indicate that the model has higher accuracy, with an average absolute error of approximately 945 persons, significantly lower than 2529 persons in the Graham method. Thus, it can effectively estimate the LOL caused by dam failure. This study developed a novel method for quantitatively assessing the LOL caused by dam failure, which could also serve as a reference for modeling disaster consequences in other fields. [ABSTRACT FROM AUTHOR]

China's unsustainable agricultural production activities have caused severe agricultural non-point source pollution (ANPSP), particularly nitrogen source pollution (NSP), which impacts farmland ecological environments. Understanding NSP is important for managing ANPSP and improving environmental water quality. Thus, inventory analysis methods were used to assess county-level total nitrogen (TN) loads in large-scale agricultural areas, analyze spatial and temporal variations in TN loads, examine the major sources of TN loads and identify the driving factors of DITN using Pearson correlation and random forest (RF) methods. The results showed that (1) TN loads and discharge intensity decreased by 4.59% and 12.7%, respectively, increasing and then decreasing via market regulations, policy incentives, and policy constraints. The eastern and northern Henan Province plain areas were highly discharged. (2) The major pollution sources varied, with the total contributions of mineral fertilizers and livestock and poultry breeding to TN being 73.53% and 81.78%, respectively. Farmland solid waste contributed minimally but increased yearly from 10.84 to 19.74%. (3) The discharge intensity of TN was highly significantly negatively correlated with the urbanization rate, economic development and slope; neither was significantly correlated with arable land. The RF results showed that total mineral fertilizer consumption and total livestock and poultry production were the major factors controlling the NSP in Henan Province. China's agricultural policies have slightly reduced pollutant discharge intensities, but the situation remains grim. This study provides a scientific basis for the development of differentiated TN control measures in Henan Province, which is important for improving pollution prevention and control capabilities. [ABSTRACT FROM AUTHOR]

This study investigates monthly streamflow modeling at Kale and Durucasu stations in the Black Sea Region of Turkey using remote sensing data. The analysis incorporates key meteorological variables, including air temperature, relative humidity, soil wetness, wind speed, and precipitation. The study also investigates the accuracy of multivariate adaptive regression (MARS) with Kmeans clustering (MARS-Kmeans) by comparing it with single MARS, M5 model tree (M5Tree), random forest regression (RF), multilayer perceptron neural network (MLP). In the first modeling stage, principal component regression is applied to diverse input combinations, both with and without lagged streamflow (Q), resulting in twenty-three and twenty input combinations, respectively. Results demonstrate the critical role of including lagged Q for improved model accuracy, as models without lagged Q exhibit significant performance degradation. The second stage involves a comparative analysis of the MARS-Kmeans model with other machine-learning models, utilizing the best-input combination. MARS-Kmeans, incorporating three clusters, consistently outperforms other models, showcasing superior accuracy in predicting monthly streamflow. [ABSTRACT FROM AUTHOR]

Identifying the current pollution status, future trends, and driving factors is a crucial prerequisite for protecting and managing soil environmental quality in light of the challenging soil heavy metal pollution situation. In this study, a typical industrial city in southeast China was selected as the study area. The environmental capacity of 12 heavy metals was first measured for the evaluation of soil heavy metal pollution, and then spatiotemporal change was simulated using indicator Kriging to warn the pollution risk. Finally, the driving factors of heavy metal pollution were analyzed using random forest and geographical detectors. The results revealed that the average environmental capacity index was 0.768, at a medium capacity level. The environmental capacity of each heavy metal showed a decreasing trend from 2016 to 2035. In 2035, the dynamic environmental capacity of each heavy metal element will have decreased by more than 75% if no preventative and remedial measures are taken. The distance from rivers, fine particulate matter, organic matter content, population density, and distance from industrial enterprises were the main factors contributing to heavy metal pollution in soil. The interaction effect of two environmental factors will mostly lead to the intensification of pollution. Based on the results, this study discussed strategies for controlling and remediating the polluted land. These findings could support the source investigation and early warning of soil heavy metal pollution, which are crucial to manag soil pollution and alleviating environmental quality issues. [ABSTRACT FROM AUTHOR]

Post-processing using deep learning algorithms can be conducted to improve accuracy of hydrologic predictions and quantify their uncertainty. In this paper, a revised version of the Local uncertainty estimation model (LUEM-R) and the circular block bootstrap (CBB) method have been used to improve the accuracy of the Variable Infiltration Capacity's (VIC) streamflow simulation and quantify its uncertainty. We used the simulated and observed streamflow at the gauge located in North CAPE Fear basin, USA from the Dayflow dataset. In the LUEM-R method, a combination of Gaussian Mixture Model and Long Short-Term Memory (LSTM) networks, which are able to capture the dependencies in time series, were used to construct the upper and lower prediction limits (PLs) for the 90% confidence level. In the CBB method, a circular block resampling technique was used to account for the dependencies in the time series (CBB-LSTM). The improved streamflow, calculated as the mean of the LSTM outputs for 200 bootstrap realizations, showed a very high correlation with observations, with coefficient of determination values of 0.97 and 0.87 for the training and testing periods, compared to the 0.77 and 0.74 values for the initial VIC simulations. For the CBB-LSTM method, the 90% PLs were constructed by fitting the best distribution at each time step, accordingly. The PLs bracketed 90% and 70% of observations in the training and testing periods, while being significantly narrower than the LUEM-R bands, which contained 92 and 91 percent of observations in the training and testing periods. Ordinary bootstrapping was also conducted using the Random Forest model (OB-RF). Comparison of the results indicates the superiority of CBB-LSTM method to the OB-RF in improving the accuracy and quantifying the uncertainties in hydrological model simulations. Overall, the CBB-LSTM was successful in improving the deterministic accuracy of VIC model simulations and quantifying its uncertainty. Also, the LUEM-R method could be efficiently utilized in quantification of uncertainty of VIC model simulations. [ABSTRACT FROM AUTHOR]

We use five established, but conceptually different artificial intelligence algorithms for analysing clogging and quantifying colloid transport at pore scale: artificial neural networks, decision tree, random forest, linear regression, and support vector regression. We test how these algorithm can predict clogging by interpolating physics based simulation data. Our training and test data set is based on results from Lattice Boltzmann simulations reproducing the physics of colloid transport through a typical pore throat present in glass beads or medium sized sand. We perform hyperparameter optimization through cross validation for all algorithms. The tree based methods have the highest Nash–Sutcliffe efficiencies among all tested algorithms with values mostly above 0.9 for the independent test data. The event of clogging can be predicted even with 100% accuracy. Our results indicate a non-linear, rather categorial nature of the (simulation) data. This is in contrast to the typical use of neural network algorithms for simulation data while tree based methods are often applied to observational data. We partly link this to the small size of our dataset. Our application of artificial intelligence in porous media research shows that time-consuming Lattice Boltzmann simulations can be easily supplemented and extended at small computational costs while predictability of clogging and quantitative effects of process specific parameters on colloidal transport are given with high reliability. [ABSTRACT FROM AUTHOR]

Landslides are among the most destructive natural hazards causing loss of life, destruction of infrastrucures and damage to properties, especially in hilly and mountaneous areas all over the world. To properly plan and manage these regions, it is crucial to accurately identify areas susceptible to landslides. In this study, we developed and applied novel hybrid machine learning models, RF-FPA and MAB-FPA, which combine forest by penalizing attribute (FPA), with two ensemble techniques, random forest (RF) and multiboost (MAB), to construct landslide susceptibility maps of Dien Bien province in Vietnam. For the development of hybrid models, we have used data of 665 landslide events and 12 landslide conditioning factors namely distance to rivers, elevations, distance to roads, normalized difference vegetation index, faults, curvature, slope, flow accumulation, stream power index, geology, topographic wetness index (TWI), and aspect. Standard statistical measures including area under the receiver operating characteristic (AUC-ROC) curve were used to evaluate predictive performance of the developed models. Results showed that both proposed novel hybrid models performed well in the correct identification of landslide susceptible areas, with RF-FPA (AUC = 0.840) slightly outperforming MAB-FPA (AUC = 0.814). The performance of the proposed hybrid models was also significantly better than that of single base classifier FPA (AUC = 0.807). In conclusion, while all the studied models performed well, the novel RF-FPA model is a more promissing tool for accurate mapping and correct prediction of landslide susceptibile areas compared to MAB-FPA amd FPA models. This approach of integrating RF with FPA can also be applied to spatially predict landslides in other areas, considering local geo-environmental conditions for proper management of landslide-prone areas. [ABSTRACT FROM AUTHOR]

Monitoring of river water is necessary to reveal its quality and pollution level so that we can protect human health and the environment. The present study explored the water quality of the Narmada River in India. To evaluate the water quality of the Narmada River, water samples were collected from 13 sites during the pre- and post-monsoon seasons, and were analyzed for different physicochemical parameters. The results from the analysis were used for the development of the entropy-river water quality index (ERWQI). The ERWQI was used to estimate the Narmada river water quality for two different uses: drinking after disinfection (ERWQId) and bathing (ERWQIb). The machine-learning-based classification models, namely the Logistic regression (LR), Support Vector (SV), K-Nearest Neighbor (KNN), Random Forest (RF), and Gradient Boosting (GB) models were examined to predict and classify ERWQI. The precision, recall, F1 score, and confusion matrix were used to evaluate the performance of the model. The findings of this study identified the LR model as the most accurate classification model with the highest accuracy score for both the ERWQId and ERWQIb. Moreover, this study also revealed that the water quality of the Narmada River was unsuitable for drinking after disinfection and hence, before any further use it requires treatment through conventional or an advanced techniques. However, the ERWQIb of the Narmada River was categorized as excellent to fair. This study has broad implications for the classification of river water quality and can provide some very useful information to monitoring agencies and policymakers. [ABSTRACT FROM AUTHOR]

This study estimates intra-daily PM10 concentrations at 213 inland and coastal monitoring sites in Türkiye from 2008 to 2019 using satellite-based aerosol optical depth (AOD) from the Moderate Resolution Imaging Spectroradiometer (MODIS). An estimation model based on the random forest (RF) approach was developed using the AOD data from the Terra satellite, the meteorological data, and aerosol diagnostics from the Modern-Era Retrospective analysis for Research and Applications, version 2 (MERRA-2), and other auxiliary variables. First, the correlation between the matched PM10 concentrations and MODIS AOD was investigated simply with the quadrant regression (QR) approach. Next, the feature selection procedure was applied to obtain the most significant predictive variables for the estimation model. Then, the spatial and temporal performances of the developed RF model were intensely discussed. Finally, a bias analysis based on the most influential input parameters was also performed to examine the potential errors in the estimated PM10 concentrations. As a result, the RF model showed moderately good performance, with a correlation coefficient (R) of 0.72 and low root mean square error (RMSE) over the entire country, which was better than the results of previous studies in the region. Moreover, the model better estimated PM10 concentrations at individual monitoring sites (with R up to 0.90), particularly in coastal regions. However, overfitting occurred in areas with low populations and few monitoring stations. Additionally, the RF model's performance varied slightly across different seasons, such as autumn (R ≅ 0.69), spring (R ≅ 0.65), winter (R ≅ 0.64), and summer (R ≅ 0.60), and it did not adequately estimate intra-daily PM10 concentrations at the seasonal scale. Furthermore, the bias analyses indicated that higher PM10 and dust mass concentrations, u and v wind components were significant parameters that caused bias in the estimations. Finally, this study provides valuable information for further applications in PM10 patterns and represents the first step toward constructing a high-resolution satellite-based air quality monitoring network in Türkiye. [ABSTRACT FROM AUTHOR]

The southwestern foothills of the Himalayan Mountain range have been experiencing a surge of catastrophic landslides in the last two decades, as a tragic result of the adverse effects of climate change. This research is about the landslide risk assessment (LRA) which has not been explored yet in the landslide-prone district Muzaffarabad, Pakistan. Landslide susceptibility (spatial probability) was analyzed using random forest model while landslide hazard (temporal probability) was analyzed using Poisson probability model. A random forest-based landslide susceptibility map depicts an accuracy of 0.90. A landslide hazard map was generated by multiplying the temporal probability with the spatial probability and classified as well. Semi-quantitative danger pixels and a fuzzy set theory approach for LRA have been adopted to estimate future landslide risks in the region. The pixel-based LRA approach indicates that 14, 18 and 20 km2 area of settlement while, the fuzzy set theory-based approach depicts that 15, 19 and 21 km2 area of the settlement are under very high landslide risk for 1-, 3-, and 5- year return period respectively. Both approaches produced risk maps that designated various risk zones with almost the same area coverage and results. The LRA maps were classified into five classes including very high (1.99%, 2.33%, 2.80%), high (2.16%, 2.53%, 3.04%), moderate (8.02%, 9.79%, 11.22%), low (17.76%, 22.94%, 23.20%), and very low (70.08%, 62.40%, 59.74%) risk zones for 1, 3 and 5 years return period respectively. This research will assist planners and scientists in developing high-precision management strategies for landslide-affected natural resources, especially in the context of the increasing impact of geomorphic hazards on climate change. [ABSTRACT FROM AUTHOR]

Dynamic traffic and complex roadside environments always cause fine variations in traffic pollutants with many uncertainties in nonmotorized lanes located close to motorways; thus, reliable methods for identifying pollution risks are urgently needed so that measures can be taken to reduce these slow-moving risks. Focusing on the nonmotorized lanes along an expressway in Fuzhou, China, in this study, we established a cycling platform instrumented by portable detectors to collect fine particle (PM2.5), coarse particle (PM10), and black carbon (BC) concentrations at a high spatiotemporal resolution; then, wavelet transform (WT) and random forest (RF) methods were combined to reveal the fine-scale distribution of different particulate matter types. The results indicated that WT was able to accurately decompose the total measurement value ( C t ) of each particulate matter into immediate vehicle-emitted ( C v ) and background-contributed ( C b ) values, thereby successfully identifying the spatiotemporal variations in traffic-induced pollution hotspots rather than background-disguised hotspots. Furthermore, the RF results were substantially better than the land-use regression results with regards to the fine-scale prediction of each particle in nonmotorized lanes. Although the RF predictions of C t and C v particles differed, traffic pollution hotspots could still be captured by the results. Compared to the measurements, the spatial distributions of the PM2.5 and PM10 predictions presented R2 values larger than 0.96, higher than those of BC (R2 = 0.77); this was the result of the different impacts of the same predictors, especially their differentiated determinants such as barometric pressure, relative humidity and air temperature. This study highlights the potential of using WT and RF methods to reveal fine-scale variations in roadside traffic pollution, which is beneficial for preventing and controlling air pollution in road microenvironments. [ABSTRACT FROM AUTHOR]

This study examines the use of snow avalanche susceptibility maps (SASMs) to identify areas prone to avalanches and develop measures to mitigate the risk in the Province of Sondrio, Italy. Various machine learning classifiers such as Random Forest, Gradient Boosting Machines, and AdaBoost, as well as newer classifiers like XGBoost, LightGBM, and NGBoost, were used with 17 conditioning factors and 1880 snow avalanche samples. The XGBoost classifier was found to have the best performance and McNemar's test results indicated that certain classifier pairs, such as RF-AdaBoost, RF-XGBoost, and XGBoost-LightGBM, produced significant predictions while others did not. The XGBoost classifier found that 19.31% of Sondrio was very susceptible to avalanches. Instead of providing a global explanation of the classifier models, the study employs a local eXplainable Artificial Intelligence (XAI) approach called SHapley Additive eXplanations (SHAP) to give insight into how each conditioning factor contributes to the likelihood of snow avalanches. According to the SHAP values, the three most important factors in the XGBoost classifier model for determining the likelihood of snow avalanches are elevation, maximum temperature, and slope. The model shows that as elevation increases, the likelihood of avalanches also increases. On the other hand, a higher maximum temperature is found to decrease the likelihood of an avalanche. Slope is found to have a positive effect on the likelihood of an avalanche, meaning that steeper slopes increase the likelihood of an avalanche. This study also analyzes the avalanche susceptibility of ski resorts in the province and found that the majority of them are located in low and moderately susceptible areas, but some are in highly susceptible areas. The study used SHAP force plots to examine the local factors that contribute to the likelihood of avalanches in these specific ski resorts. The results show that ski resorts with elevations greater than 2000 m and slopes greater than 30 degrees, such as Livigno, Santa Caterina-Valfurva and Passo dello Stelvio, have a higher susceptibility to avalanches due to higher positive SHAP values. Conversely, ski resorts with elevations less than 2000 m and slopes less than 30 degrees, such as Aprica and Bormio, have a lower susceptibility to avalanches because of negative SHAP values. This study provides a valuable tool for creating new strategies to reduce the harm and damage caused by slow avalanches in the region. [ABSTRACT FROM AUTHOR]

Runoff prediction with high accuracy is vital to protect people's properties and lives. In this study, a Short-Term Streamflow Forecasting Framework combining Long Short-Term Memory (LSTM) with Self-Organizing Map (SOM) is proposed, which clusters hydro-meteorological factors through SOM network to cluster the streamflow into several flow modes with different physical characteristics, converting complex global modeling into local modeling. The framework also considers the temporal characteristics of the streamflow, so the data of different flow modes are fitted with multiple LSTM separately. Besides, Random Forest is used in the framework for feature selection. The model is used for streamflow forecasting 1 to 3 day(s) ahead in the Qingxi River Basin in China, and the performance is compared with those of SVR, ANN, LSTM, SOM-SVR, SOM-ANN, FCM-LSTM, and K-means-LSTM. In addition, SOM-LSTM with three and four clusters were compared. The results show that (1) the sub-processes obtained by SOM clustering show some physical characteristics consistent with the actual process. (2) SOM-LSTM significantly improves the prediction accuracy of short-term streamflow forecasting. (3) All the integrated models using different clustering methods can improve the prediction accuracy of a single LSTM, and SOM-LSTM has the most reliable prediction when the lead time is 1–2 day(s). (4) The number of clusters being set to 4 is more suitable for the proposed prediction framework. The overall results show that the SOM-LSTM provides an efficient method for short-term streamflow forecasting. [ABSTRACT FROM AUTHOR]

Interactions across meteorological drought, agricultural drought and irrigation enlarge complexities of simulating and forecasting crop yield under droughts. Besides, contradictory relationships amid multiple objectives considering respective critical thresholds aggravate difficulties in agricultural water resources optimization. To address the above problems, this paper developed a framework of random forest based multi-objective multi-critical control programming with fuzzy boundary intervals (RFMOMCCPFBI) model for agricultural water resources optimization. The MOMCCP is established based on the MOP approach, and it improves upon the MOP in giving lower and upper thresholds of each objective, promising achievement of subjectivities of managers. Moreover, the model enhances the advantages of MCCP in reflecting multiple variations and uncertainties of thresholds by introducing fuzzy boundary intervals. In addition, the crop yield is simulated by random forest approach, which improves calculation efficiency as well as reflect influences of multiple interactive drought and irrigation on crop yield. The RFMOMCCPFBI model is used to Yingke irrigation district to allocate water resources to multiple crops at different growth stages to verify its applications. The results disclose that frequencies of meteorological drought are around 27%, 19% at five growth stages, higher than agricultural drought in future year 2022 to year 2047. Meteorological drought has many times of abrupt changes without significance. Agricultural water-saving potential increases about 20% considering demands of acceptable thresholds of objectives. Water saving potential enlarges with increase of tolerance levels of objective deviations. The model can provide alternatives for managers based on their preferences on objectives. [ABSTRACT FROM AUTHOR]

Reliable precipitation estimates are crucial for understanding hydrological processes and engineering applications. Precipitation data can be collected from different sources, such as gauges, satellites, and reanalysis products. However, each source has its own pros and cons, including uncertainties. Merging multisource data may help improve spatio-temporal precipitation estimates and reduce uncertainties. This study aimed to improve precipitation estimates in the data scarce river basins taking the cases of upper Blue Nile River basin by blending gauge records with IMERG early, CHIRPS, PERSIANN-CCS, and ERA5 products. A machine learning random forest-based merging procedure (RF-MEP) was applied as a merging technique. Daily measured multi-year (2010–2017) precipitation data were used from daily measurements. Several quantitative and qualitative evaluation metrics were applied to assess the performance of the merged product. The performance of individual unmerged rainfall products, namely, CHIRPS, PERSIANN-CCS, IMERG, and ERA-5 was evaluated and compared with the Random Forest based merged product (gauge, satellite and reanalysis products) for modeling the spatial–temporal pattern of precipitation in the Upper Blue Nile Basin. Compared with reference data the merged product exhibited a consistent improvement in terms of Index of Agreement, CC, MAE, NSCE, KGE, and RMSE over the original precipitation products. Moreover, POD, FAR, and CSI exhibited an improvement over unmerged product. Therefore, the merged precipitation product showed a better spatio-temporal representation and accuracy which can be successfully used for scientific studies as well as operational purposes in data scarce river basins. [ABSTRACT FROM AUTHOR]

Various machine learning (ML) techniques have been recommended and used in the literature to produce landslide susceptibility map (LSM). On the other hand, feature engineering (FE) is an important topic in ML studies, but the concept is ignored by most research. In this study, a novel FE framework, including feature selection, feature transformation, feature binning, and feature weighting, is proposed to produce LSMs using eXtreme gradient boosting (XGBoost), random forest (RF), and support vector machine (SVM). For this purpose, first, thirteen landslide conditioning factors used in data preprocessing were utilized for producing LSM models in the study area, Babadag district of Denizli Province in the Aegean region of Turkey. Second, two irrelevant factors eliminated from the input feature subset using the feature selection in the FE framework. Third, features determined as skewed data were converted into symmetric form by applying feature transformation analysis with log transformation. Then, the remaining factors having continuous values were turned into categorical values using the quantile classifier technique. During the feature weighting phase, four different feature weighting methods, namely, eXtreme Gradient Boosting, random forest (RF), non-negative least squares (NNLS), and Frequency Ratio, were utilized to calculate the weights in each subclass of each landslide-related factor. In addition, the proposed feature subsets were also compared with raw data. At the end of process, the XGBoost model constructed with a FR-selected subset (Overall Accuracy (Acc) = 0.907 and area under curve (AUC) = 0.9822) outperformed both raw (Acc = 0.874; AUC = 0.960) and other methods (i.e., RF–FR and SVM–NNLS). Consequently, the study results revealed that the proposed FE approach could be a useful framework to increase the performance of ML techniques in identifying and extracting relevant features to develop highly optimized and enriched models. [ABSTRACT FROM AUTHOR]

Effective environmental management and contamination remediation require accurate spatial distributions and predictions of potentially toxic elements (PTEs) in the soil. However, no single method has been developed to predict soil PTE accurately. This study evaluated the advanced geostatistical method of empirical Bayesian kriging regression prediction (EBKRP), machine learning algorithm of random forest (RF), and hybridized model (RF-EBKRP) to predict and map PTEs content in greenspace soils. As identified by RF, soil organic carbon, organic matter, total (nitrogen, phosphorus, and potassium), topographic features, and urban functional types were used as significant covariates to improve the prediction accuracy of PTEs in soil. The model prediction performance was evaluated using the root mean square error (RMSE), mean absolute percentage error (MAPE), and coefficient of determination (R2). Results showed that RF performed much better than EBKRP in predicting soil PTEs, with lower prediction errors and a higher R2. The RMSE, MAPE, and R2 values for the RF model were 0.25–85.32 mg/kg, 3.86–25.40%, and 0.77–0.90, respectively, while the values for the EBKRP method were 0.51–99.03 mg/kg, 5.42–32.13%, and 0.40–0.66. Moreover, the RF-EBKRP method produced more accurate spatial predictions and distributions of PTEs than the individual models, with R2 improvements of 122.5% for EBKRP and 15.58% for RF. The better performances of RF-EBKRP are due to its incorporation of various covariates and its ability to handle complex nonlinear relationships between soil PTEs and covariates. In the end, a hybrid RF-EBKRP method is a promising approach to improving the spatial distribution map of soil PTEs. [ABSTRACT FROM AUTHOR]

Agricultural droughts are a prime concern for economies worldwide as they negatively impact the productivity of rain-fed crops, employment, and income per capita. In this study, Standard Precipitation Index (SPI) has been used to evaluate different drought indices for Rajasthan of India. In agricultural, hydrological, and meteorological applications such as irrigation scheduling, crop simulation, water budgeting, reservoir operations, and weather forecasting, the accurate estimation of the drought indices such as the Standardized Precipitation Index (SPI) plays an important role. Thus, the present study was conducted to examine the feasibility and effectiveness of the Random Subspace (RSS) model and its hybridization with the M5 Pruning tree (M5P), Random Forest (RF), and Random Tree (RT) to estimate the SPI at 3, 6, and 12 droughts during 2000–2019. Performances of RSS and hybridized algorithms were assessed and compared using performance indicators (i.e., MAE, RMSE, RAE, RRSE, and R2) and various graphical interpretations. Results indicated that the RSS-M5P provided the most accurate SPI prediction (MAE = 0.497, RMSE = 0.682, RAE = 81.88, RRSE = 87.22, and R2 = 0.507 for SPI-3; MAE = 0.452, RMSE = 0.717, RAE = 69.76, RRSE = 85.24, and R2 = 0.402 for SPI-6. And MAE = 0.294, RMSE = 0.377, RAE = 55.79, RRSE = 59.57, and R2 = 0.783 for SPI-12) compare to RSS alone, RSS-RF, and RSS-RT models for study the drought situation in Jaisalmer Rajasthan. The M5P algorithms have improved the performance of the RSS structure. [ABSTRACT FROM AUTHOR]

For effective water resource management, water budgeting, and optimal release discharge from a reservoir, the accurate prediction of daily inflow is critical. An attempt has been made using artificial intelligence (AI) techniques to enhance water management efficiency in the Haditha-dam reservoir. This case study occasionally suffers from severe drought events and thus causes significant water shortages as well as stopping hydroelectric power stations for several months. Four different approaches were employed for inflow forecasting, namely multiple linear regression (MLR), random forest (RF), extreme learning machine (ELM), and regularized extreme learning machine (RELM). Autocorrelation function (ACF) and partial autocorrelation function (PACF) were used to select the best-lagged variables. The obtained results revealed the superiority of the RELM model compared to other forecast models. The proposed model (RELM) yielded higher prediction accuracy, and its prediction records were similar to the actual values. Moreover, the adopted model achieved a higher correlation of coefficient value (R = 0.955). The regularization approach effectively enhanced the prediction capacity and the generalization ability of the proposed model. On the other hand, the RF model's performance capacity was poor compared to other comparable models due to the overfitting issue. Moreover, the results showed that the PACF (partial autocorrelation function) gave more accurate and realistic predictors than ACF (autocorrelation function) because of its ability to cope with a sudden temporal variation of inflow time series. Overall, the RELM approach provided higher adequacy and tighter confidence in forecasting daily inflow even in noisy data and severe climatic conditions. [ABSTRACT FROM AUTHOR]

Prediction of flow rate in rivers is essential for the planning and management of water resources. This study shows that, based on a Machine Learning approach, accurate models for streamflow prediction can be developed. The Des Moines watershed, which includes both Des Moines River and Raccoon River, was chosen as a case study. Only the daily river discharge was considered for the modeling of 10 stations located on different tributaries of the two rivers. Four machine learning algorithms were applied for the streamflow prediction: Random Subspace, M5P, Random Forest and Bagging. The performance of applied algorithms was assessed using statistical performance indicators and graphical representations. Three stations were selected for the training and testing of the different Machine Learning models. Then, the best model was validated on the other seven stations. Prediction accuracy was also assessed as the forecast horizon increased. Overall, M5P algorithms led to the best predictions, with R2 equal to 0.970 and 0.960 for the stations of East Fork Des Moines River at Dakota City and Des Moines River near Tracy, respectively. Accurate predictions were obtained also on the Raccoon River, with R2 equal to 0.938 and 0.887 for the stations of North Raccoon River near Jefferson and Raccoon River at Van Meter, respectively. [ABSTRACT FROM AUTHOR]

This research aims to generate a landslide susceptibility map (LSM) for the Bhagirathi river basin located in the Tehri Garhwal district of Uttarakhand state in India. For this study, we incorporated and utilized machine learning novel ensemble models, namely: Generalize Linear Model (GLM), Random Forest (RF), Convolutional Neural Network (CNN), GLM-RF, CNN-RF, CNN-GLM and CNN-GLM-RF. The above-specified ensemble models were incorporated for the preparation of LSM. A total of 171 landslide locations were chosen and studied for the preparation of landslide inventory. From the landslide inventory, 70% of landslides were utilized for the training purpose, whereas the rest, 30%, were used for the validation purpose. In the ongoing research study, a total of 17 landslide conditioning factors (LCFs) were utilized, and the analysis for multi-collinearity was performed by tolerance (TOL) as well as variance inflation factor (VIF) methods. These LSM were then validated through receiver operating characteristic curve (ROC), mean absolute error (MAE), root mean square error (RMSE) and Chi-Square methods. Finally, the outcomes of the analysis, as well as validation models, show that the CNN model proves to be the most effective for predicting landslide susceptibility in the study area. The study found that the methods incorporated for the area under consideration can be applied to the regions having similar geo-environmental factors globally. [ABSTRACT FROM AUTHOR]

This study introduces a methodology that integrates SHAP (SHapley Additive exPlanations) analysis with Random Forest (RF) classification to enhance the prediction accuracy of fire severity across diverse forested regions in Türkiye. Leveraging a comprehensive forest fire database spanning from 2018 to 2022 and utilizing the Google Earth Engine (GEE) platform, 436 fire events ranging from 9 to 53,764.1 hectares were automatically detected and mapped using the difference Normalized Burn Ratio (dNBR). Subsequently, a robust fire severity model was developed by incorporating 19 variables, including biophysical, topographic, climatic, and vegetation-related factors. The RF classification achieved noteworthy performance metrics, with an overall accuracy of 0.75, a Kappa value of 0.61, and a macro-average AUC value of 0.88. Furthermore, the integration of SHAP analysis provided insightful contributions to the RF classification model, elucidating the impacts of individual input features. Notably, variables such as NDMI (Normalized Difference Moisture Index), LAI (Leaf Area Index), and LWVI (Land Water Vegetation Index) emerged as significant influencers, followed by WSPD (Wind Speed) and LST (Land Surface Temperature). Additionally, an analysis of fire severity distribution across fuel types (FMT) revealed intricate patterns, underscoring the complex relationships between vegetation composition and fire behavior. The findings of this study have implications for forest management and wildfire risk assessment, offering valuable insights for decision-making processes. Furthermore, the integration of SHAP analysis with RF classification enhances the interpretability and transparency of machine learning-based fire severity prediction models, contributing to the advancement of fire management strategies. [ABSTRACT FROM AUTHOR]

Coronavirus has been identified as one of the deadliest diseases and the WHO has declared it a pandemic and a global health crisis. It has become a massive challenge for humanity. India is also facing its fierceness as it is highly infectious and mutating at a rapid rate. To control its spread, many interventions have been applied in India since the first reported case on January 30, 2020. Several studies have been conducted to assess the impact of climatic and weather conditions on its spread in the last one and half years span. As it is a well-established fact that temperature and humidity could trigger the onset of diseases such as influenza and respiratory disorders, the relationship of meteorological variables with the number of COVID-19 confirmed cases has been anticipated. The association of several meteorological variables has therefore been studied in the past with the number of COVID-19 confirmed cases. The conclusions in those studies are based on the data obtained at an early stage, and the inferences drawn based on those short time series studies may not be valid over a longer period. This study attempted to assess the influence of temperature, humidity, wind speed, dew point, previous day's number of deaths, and government interventions on the number of COVID-19 confirmed cases in 18 districts of India. It is also attempted to identify the important predictors of the number of confirmed COVID-19 cases in those districts. The random forest model and the hybrid model obtained by modelling the random forest model's residuals are used to predict the response variable. It is observed that meteorological variables are useful only to some extent when used with the data on the number of the previous day's deaths and lockdown information in predicting the number of COVID-19 cases. Partial lockdown is more important than complete or no lockdown in predicting the number of confirmed COVID-19 cases. Since the time span of the data in the study is reasonably large, the information is useful to policymakers in balancing the restriction activities and economic losses to individuals and the government. [ABSTRACT FROM AUTHOR]

Estimation of reservoir inflow is of particular importance in optimal planning and management of water resources, proper allocation of water to consumption sectors, hydrological studies, etc. This study aimed to estimate monthly inflow (Q) to the Maroon Dam reservoir located in Iran utilizing climatic data such as minimum, maximum, and mean air temperatures (Tmin, Tmax, T), reservoir evaporation (E), and rainfall (R). The impact of any of the mentioned variables was analyzed by the entropy-based pre-processing technique. The results of the pre-processing showed that the rainfall is the most important parameter affecting the reservoir inflow. Therefore, three types of input patterns were taken into consideration consisting the antecedent Q-based, antecedent R-based, and combined antecedent Q and R-based input combinations. To estimate the monthly reservoir inflow, a random forest (RF) was firstly employed as the standalone model. Then, two different types of hybrid models were proposed via coupling the RF on complete ensemble empirical mode decomposition (CEEMD) and wavelet analysis (W) in order to implement the coupled CEEMD-RF and W-RF models. It is worthwhile to mentioning that six mother wavelets were used in developing the hybrid W-RF models. Four error metrics including root mean square error (RMSE), mean absolute error (MAE), Kling-Gupta efficiency (KGE), and Willmott index (WI) were used to assess the accuracy of implemented models. The attained results indicated the superiority of proposed hybrid models over the classic RF for estimating the monthly reservoir inflow. The most precise model during the test phase was W-RF(3) utilizing the Sym(2) as the mother wavelet under a lagged Q-based pattern with error measures of RMSE = 15.011 m3/s, MAE = 10.439 m3/s, KGE = 0.832, WI = 0.773. [ABSTRACT FROM AUTHOR]

Reliable water quality prediction can improve environmental flow monitoring and the sustainability of the stream ecosystem. In this study, we compared two machine learning methods to predict water quality parameters, such as total nitrogen (TN), total phosphorus (TP), and turbidity (TUR), for 97 watersheds located in the Southeast Atlantic region of the USA. The modeling framework incorporates multiple climate and watershed variables (characteristics) that often control the water quality indicators in different landscapes. Three techniques, such as stepwise regression (SR), Least Absolute Shrinkage and Selection Operator (LASSO), and genetic algorithm (GA), are implemented to identify appropriate predictors out of 28 climate and catchment-related variables. The selected predictors were then used to develop the Random Forest (RF) and Boosted regression tree (BRT) models for water quality predictions in selected watersheds. The results highlighted that while both algorithms provided reasonable results (based on statistical metrics), the RF algorithm was easier to train and robust to model overfitting. Partial dependence plots highlighted the complex and nonlinear relationships between the individual predictors and the water quality indicators. The thresholds obtained from partial dependence plots showed that the median values of total nitrogen (TN) and total phosphorus (TP) in streams increase significantly when the percentage of urban and agricultural lands is above 40% and 43% of the watershed area, respectively. Furthermore, when soil hydraulic conductivity increases, the reduction in runoff results in decreased Turbidity levels in streams. Therefore, identifying the key watershed characteristics and their critical thresholds can help watershed managers create appropriate regulations for managing and sustaining healthy stream ecosystems. Besides, the forecasting models can improve water quality predictions in ungauged watersheds. [ABSTRACT FROM AUTHOR]

The relation between the increase in the frequency and the effects of extreme events with climate change has been widely demonstrated and the related consequences are a global concern. In this framework, the strong correlation between significant lightning occurrence and intense precipitation events has been also documented. Consequently, the possibility of having a short-term forecasting tool of the lightning activity may help in identifying and monitoring the evolution of severe weather events on very short time ranges. The present paper proposes an application of Random Forest (RF), a popular Machine Learning (ML) algorithm, to perform a nowcasting of Cloud-to-Ground (CG) lightning occurrence over the Italian territory and the surrounding seas during the months of August, September, and October from 2017 to 2019. Results obtained with three different spatial resolutions have been compared, suggesting that, to enhance the skills of the model in identifying the presence or absence of strokes, all the data selected as input should be commonly gridded on the finest available spatial resolution. Moreover, the features' importance analysis performed confirms that meteorological features describing the state of the atmosphere, especially at higher altitudes, have a stronger impact on the final result than topology data, such as Latitude or Digital Elevation Model (DEM). [ABSTRACT FROM AUTHOR]

In this paper, we compare the performance of two data-driven algorithms to deal with an automatic classification problem in geomorphology: Direct Sampling (DS) and Random Forest (RF). The main goal is to provide a semi-automated procedure for the geomorphological mapping of alpine environments, using a manually mapped zone as training dataset and predictor variables to infer the classification of a target zone. The applicability of DS to geomorphological classification was never investigated before. Instead, RF based classification has already been applied in few studies, but only with a limited number of geomorphological classes. The outcomes of both approaches are validated by comparing the eight detected classes with a geomorphological map elaborated on the field and considered as ground truth. Both DS and RF give satisfactory results and provide similar performances in term of accuracy and Cohen's Kappa values. The map obtained with RF presents a noisier spatial distribution of classes than when using DS, because DS takes into account the spatial dependence of the different classes. Results suggest that DS and RF are both suitable techniques for the semi-automated geomorphological mapping in alpine environments at regional scale, opening the way for further improvements. [ABSTRACT FROM AUTHOR]

The frequent occurrence of urban waterlogging seriously affects people's lives and the national economy. The use of machine learning (ML) methods to spatially assess urban waterlogging susceptibility is critical for reducing the losses caused by such disasters. It is important to select an equal number of positive and negative samples to train binary ML classifiers for evaluation; in most cases, researchers are only able to obtain a relatively small number of historical waterlogging locations (positive samples), which leads to the selection of a limited number of negative samples, further affecting the trained classifiers' performance. Facing this issue, we proposed an optimized seed spread algorithm (OSSA) that can estimate the potential inundated areas based on the spatial distribution of elevation and natural waters, thereby increasing the number of positive samples. The primary urban area of Guangzhou, China, was selected as the study region, and random forest was selected as the evaluation algorithm. We further employed two ML methods, support vector machine and logistic regression, to verify the quality of the increased positive samples. The results indicate that compared with the original positive samples, the OSSA-based positive samples achieve the highest area under the curve values among the three tested ML methods, indicating that the OSSA can be a suitable approach to increase the number of positive samples for such studies. We believe that this study advances the ML-based waterlogging susceptibility assessments, which could be valuable for developing countries where intensive hydrologic monitoring is lacking. [ABSTRACT FROM AUTHOR]

In this study, we developed and evaluated a hybrid framework for reservoir inflow forecast. This framework is unprecedented, which integrates new quasi-globally available observation-, satellite-, or model-based datasets using machine learing models to forecast inflow at the local scale. Under this framework, we compared random forests, gradient boosting machine, extreme learning machine, M5-cubist, elastic net, as well as their multi-model ensemble using Bayesian model averaging (BMA), and tested contributions from different input datasets, including retrospective forecast (reforecast) from florecast oriented low ocean resolution (FLOR) dynamic climate model, satellite-based hydrologic products, and climate phenomenon information. The performance was evaluated using Kling–Gupta efficiency (KGE) and correlation coefficient (R) in two headwater reservoirs, Harris reservoir in the humid Alabama–Coosa–Tallapoosa river basin and the Navajo reservoir in the arid Upper Colorado River Basin (UCRB). The results showed that for the Harris reservoir, the BMA combining five models with antecedent inflow and satellite-based hydrologic information as model inputs provided the best performance (KGE = 0.66, R = 0.76). For the Navajo reservoir, the gradient boosting machine model with all variables combined as input showed the best performance (KGE = 0.76, R = 0.83). Satellite-based soil moisture and evaporation consistently showed significant contributions to the inflow forecast. Benefits from climate indices and FLOR reforecast varied by locations, with more benefits coming from climate indices than FLOR potential evaporation reforecast at the Navajo reservoir in UCRB. Given the global coverage of the model inputs, our approach can be potentially applicable to improve reservoir inflow forecasts in different regions of the world. [ABSTRACT FROM AUTHOR]

Climate change has caused a rise in temperature extremes, particularly heatwaves, in recent decades. Physical-empirical models are developed in this study using two classical machine learning algorithms, namely decision tree (DT) and random forests, and a novel hybrid technique consists of Ada-Boost Regression and decision tree (ABR-DT) for forecasting annual heatwave days (HWDs) of Iran from synoptic predictors. The daily temperature data of Princeton Meteorological Forcing were extracted for 27 points to estimate the annual number of HWDs, and the National Centers for Environmental Prediction reanalysis data were used as predictors. The major synoptic variables were extracted for four pressure levels (e.g., 300, 500, 850, and 1000 hPa) and three monthly time lags. The Principal Component Analysis was employed to reduce the diverse predictors and their features to the most optimal structure. The grid point-based performance evaluation revealed the superiority of ABR-DT, which showed a correlation coefficient (CC) of 0.860 and meant absolute error (MAE) of 6.929, using only specific humidity and wind component as predictors. The spatial performance indices over eight different climate regions of Iran also showed the better performance of ABR-DT, which improved the CC and MAE of its two alternatives up to 185 and 19%. The study identified the optimal parameter combination as the predictors of heatwaves by examining the effects of numerous weather components. The results proved the proposed hybrid forecasting approach's effectiveness in predicting heatwave days, a devastating hazard, for many regions. [ABSTRACT FROM AUTHOR]

Baseflow plays a critical role in maintaining the aquatic environmental health. However, the driving factors and predictions of baseflow have not been rigorously investigated on a large scale, partly preventing hydrologist from deeply understanding runoff generation. To this end, the Lyne–Hollick digital filter method and the automatic baseflow identification technique were used to estimate the long-term and seasonal baseflow index (BFI) of 619 catchments across Continental United States (CONUS) from 1981 to 2014. Six natural driving factors are selected from the 31 catchment attributes about topography and location, soil, geology, land cover, and climate characteristics. The Random Forest (RF) technique was used to predict the BFI with the selected six driving factors as predictors. Results show that the long-term average BFI was 0.49, and the BFI value was different in four seasons, with the highest value of 0.55 in winter and the lowest value of 0.46 in autumn. The forest fraction, clay proportion and snow fraction were the most powerful factors affecting the long-term average BFI. The RF technique predicts the BFI across the 619 sites in CONUS with a R2 of 0.59 after Leave-One-Location cross-validation, which was more satisfactory than the multiple linear regression method. This study can provide a deep insight into the generation and variation of baseflow and guide the annual baseflow prediction for water resources management. [ABSTRACT FROM AUTHOR]

Furthermore, most of these data include a temporal dimension, often inextricably linked to the spatial one by the underlying physical process that the data represent. In this context, data is measured by instruments or generated by simulations before being processed by software and the resulting information or knowledge being saved in computers. 10.1007/s00477-022-02257-4 22 Tian D, He X, Srivastava P, Kalin L. A hybrid framework for forecasting monthly reservoir inflow based on machine learning techniques with dynamic climate forecasts, satellite-based data, and climate phenomenon information. [Extracted from the article]