Machine learning approaches for spatial modeling of agricultural droughts in the south-east region of Queensland Australia

Machine learning-based prediction of cadmium pollution in topsoil and identification of critical driving factors in a mining area

Mining activities have resulted in a substantial accumulation of cadmium (Cd) in agricultural soils, particularly in southern China. Long-term Cd exposure can cause plant growth inhibition and various diseases. Rapid identification of the extent of soil Cd pollution and its driving factors are essential for soil management and risk assessment. However, traditional geostatistical methods are difficult to simulate the complex nonlinear relationships between soil Cd and potential features. In this study, sequential extraction and hotspot analyses indicated that Cd accumulation increased significantly near mining sites and exhibited high mobility. The concentration of Cd was estimated using three machine learning models based on 3169 topsoil samples, seven quantitative variables (soil pH, Fe, Ca, Mn, TOC, Al/Si and ba value) and three quantitative variables (soil parent rock, terrain and soil type). The random forest model achieved marginally better performance than the other models, with an R2 of 0.78. Importance analysis revealed that soil pH and Ca and Mn contents were the most significant factors affecting Cd accumulation and migration. Conversely, due to the essence of controlling Cd migration being soil property, soil type, terrain, and soil parent materials had little impact on the spatial distribution of soil Cd under the influence of mining activities. Our results provide a better understanding of the geochemical behavior of soil Cd in mining areas, which could be helpful for environmental management departments in controlling the diffusion of Cd pollution and capturing key targets for soil remediation.

Cadmium accumulation in paddy soils affected by geological weathering and mining: Spatial distribution patterns, bioaccumulation prediction, and safe land usage

The abnormal enrichment of cadmium (Cd) in soil caused by rock weathering and mining activities is an issue in southern China. Although the soil Cd content in these regions is extremely high, the bioavailability of Cd in the soils differs significantly. The carbonate area (CBA) and tin-mining area (TIA) in Hezhou City were investigated to determine the primary features of soil Cd mobility in these regions and improve environmental management. Lateral and vertical spatial distributions revealed different accumulation and migration mechanisms of soil Cd in the CBA and TIA. Further analyses revealed that mining activities and geological weathering resulted in different soil geochemical parameters, thus yielding significantly lower levels of Cd in rice grains in the CBA than in the TIA. The random forest (RF) model predicted the bioaccumulation factor (BAF) (R2 = 0.69) better than the support vector machine (SVM) model (R2 = 0.68). Subsequently, a novel land management scheme was proposed based on soil Cd and the prediction of Cd in rice to optimize the spatial resources of agricultural land and ensure the safety of rice for consumption. This study provides a novel approach for land management in Cd-contaminated areas.

Multi-hazard exposure mapping under climate crisis using random forest algorithm for the Kalimantan Islands, Indonesia

Numerous natural disasters that threaten people's lives and property occur in Indonesia. Climate change-induced temperature increases are expected to affect the frequency of natural hazards in the future and pose more risks. This study examines the consequences of droughts and forest fires on the Indonesian island of Kalimantan. We first create maps showing the eleven contributing factors that have the greatest impact on forest fires and droughts related to the climate, topography, anthropogenic, and vegetation. Next, we used RF to create single and multi-risk maps for forest fires and droughts in Kalimantan Island. Finally, using the Coupled Model Intercomparison Project (CMIP6) integrated evaluation model, a future climate scenario was applied to predict multiple risk maps for RCP-SSP2-4.5 and RCP-SSP5-8.5 in 2040-2059 and 2080-2099. The probability of a 22.6% drought and a 21.7% forest fire were anticipated to have an influence on the study's findings, and 2.6% of the sites looked at were predicted to be affected by both hazards. Both RCP-SSP2-4.5 and RCP-SSP5-8.5 have an increase in these hazards projected for them. Researchers and stakeholders may use these findings to assess risks under various mitigation strategies and estimate the spatial behavior of such forest fire and drought occurrences.

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.

Geospatial assessment of agricultural drought vulnerability using integrated three-dimensional model in the upper Dwarakeshwar river basin in West Bengal, India

The amount of agricultural drought vulnerability in an underdeveloped rain-fed agro-based economy at the local, regional, and national level is most prominent factor for measurement. The desiccation of rain in agricultural sector becomes apprehensive to intercontinental food supply chain. So, adequate investigation and development of sustainable agricultural methodology are key factors to sustain the food security of a territory. In this research, delineation of agricultural drought vulnerability (ADV) status has been carried out by multidimensional mixed-method index approach using remote sensing and geographic information system. An integrated three-dimensional model is utilized to enrich this study. The three indices of this model include exposure index (EI), sensitivity index (SI), and adaptive capacity index (ACI). The ACI has been constructed by combining the environmental adaptive capacity (EAC), social adaptive capacity (SAC), and economic adaptive capacity (EcAC) index. The 40 parameters for ADV modeling are picked up by analyzing meteorological, geo-environmental, social, and remote sensing data. There are six exposure parameters, seven sensitivity parameters, twelve environmental adaptive capacity parameters, six social adaptive capacity parameters, and nine economic adaptive capacity parameters. Each index has been computed by assigning the weights based on their relative importance by using the analytic hierarchy process (AHP) approach. Final results were classified into five vulnerability zones, e.g., very low, low, moderate, high, and very high covering an area 362.32 km2, 186.68 km2, 568.69 km2, 547.05 km2, and 266.89 km2 respectively. Results have been validated with long-term Aman paddy yield data (2004 to 2014) through the yield anomaly index (YAI). Finally, the model ADV is a good model fit (R square = 0.894) and all the relationships were found significant, when SI, EI, and ACI are considered its predictors. While SI (B = 0.391, p < 0.001) and EI (B = 0.223, p < 0.001) are positively associated with ADV, ACI is negatively associated with ADV (B =  - 0.721, p < 0.001). This regional agricultural drought vulnerability model can be useful to identify drought-responsive areas and improve drought mitigation measures.

Impact of river flow modification on wetland hydrological and morphological characters

A good number of researchers investigated the impact of flow modification on hydrological, ecological, and geomorphological conditions in a river. A few works also focused on hydrological modification on wetland with some parameters but as far the knowledge is concerned, linking river flow modification to wetland hydrological and morphological transformation following an integrated modeling approach is often lacking. The current study aimed to explore the degree of hydrological alteration in the river and its effect on downstream riparian wetlands by adopting advanced modeling approaches. After damming, maximally 67 to 95% hydrological alteration was recorded for maximum, minimum, and average discharges. Wavelet transformation analysis figured out a strong power spectrum after 2012 (damming year). Due to attenuation of flow, the active inundation area was reduced by 66.2%. After damming, 524.03 km2 (48.9% of total pre-dam wetland) was completely obliterated. Hydrological strength (HS) modeling also reported areas under high HS declined by 14% after post-dam condition. Wetland hydrological security state (WSS) and HS matrix, a new approach, are used to explore wetland characteristics of inundation connectivity and hydrological security state. WSS was defined based on lateral hydrological connectivity. HS under critical and stress WWS zones deteriorated in the post-dam period. The morphological transformation was also well recognized showing an increase in area under the patch, edge, and a decrease in the area under the large core area. All these findings established a clear linkage between river flow modification and wetland transformation, and they provided a good clue for managing wetlands.

A comprehensive drought monitoring method integrating multi-source data

Droughts are the most expensive natural disasters on the planet. As a result of climate change and human activities, the incidence and impact of drought have grown in China. Timely and effective monitoring of drought is crucial for water resource management, drought mitigation, and national food security. In this study, we constructed a comprehensive drought index (YCDI) suitable for the Yellow River Basin using principal component analysis and the entropy weight-AHP method, which integrated a standardized precipitation evapotranspiration index (SPEI), self-calibrating Palmer drought severity index (scPDSI), vegetation condition index (VCI), and standardized water storage index (SWSI). SWSI is calculated by the terrestrial water storage anomaly (TWSA), which can more comprehensively reflect the impact of surface water resources on drought (as compared with soil moisture-based indexes). The study results showed that: (1) compared with single drought index, YCDI has stronger ability to monitor drought process. In terms of time scale and drought degree, the monitoring results based on YCDI were similar with data presented in the China Flood and Drought Bulletin and Meteorological Drought Yearbook, reaching ~87% and ~69%, respectively. The correlation between drought intensity and crop harvest area was 0.56. (2) By the combined analysis of the Mann-Kendall test and Moving T test, it was found that the abrupt change of YCDI index at the time of 2009, mainly due to the precipitation in 2009 reached the lowest value in the past 30 years in northern China and extreme high temperature weather. (3) The YCDI of Henan and Shandong provinces in the middle and lower reaches of the basin decreased more significantly, with the maximum value reaching 0.097/yr, while the index in the upper reaches showed an increasing trend with the maximum rate of 0.096/yr. (4) The frequency of mild drought, moderate drought, severe drought and extreme drought in the Yellow River basin during the study period was 15.84%, 12.52%, 4.03% and 0.97%, respectively. Among them, the highest frequency of droughts occurred in Ningxia, Inner Mongolia and central Shaanxi provinces. Drought causation in the Yellow River basin is more influenced by human activities than climate change in the middle and lower reaches, while climate change is the main factor in the upper reaches. Overall, YCDI is a reliable indicator for monitoring the spatial and temporal evolution of drought in the Yellow River basin, and it can be used for monitoring soil moisture changes and vegetation dynamics, which can provide scientific guidance for regional drought governance.

Coronavirus disease vulnerability map using a geographic information system (GIS) from 16 April to 16 May 2020

In recent months, the world has been affected by the infectious coronavirus disease and Iran is one of the most affected countries. The Iranian government's health facilities for an urgent investigation of all provinces do not exist simultaneously. There is no management tool to identify the vulnerabilities of Iranian provinces in prioritizing health services. The aim of this study was to prepare a coronavirus vulnerability map of Iranian provinces using geographic information system (GIS) to monitor the disease. For this purpose, four criteria affecting coronavirus, including population density, percentage of older people, temperature, and humidity, were prepared in the GIS. A multiscale geographically weighted regression (MGWR) model was used to determine the vulnerability of coronavirus in Iran. An adaptive neuro-fuzzy inference system (ANFIS) model was used to predict vulnerability in the next two months. Results indicated that, population density and older people have a more significant impact on coronavirus in Iran. Based on MGWR models, Tehran, Mazandaran, Gilan, and Alborz provinces were more vulnerable to coronavirus in February and March. The ANFIS model findings showed that West Azerbaijan, Zanjan, Fars, Yazd, Semnan, Sistan and Baluchistan, and Tehran provinces were more vulnerable in April and May.

Generating High-Resolution and Long-Term SPEI Dataset over Southwest China through Downscaling EEAD Product by Machine Learning

Drought is an event of shortages in the water supply, whether atmospheric, surface water or ground water. Prolonged droughts have negative impacts on ecosystems, agriculture, society, and the economy. Although existing drought index products are widely utilized in drought monitoring, the coarse spatial resolution greatly limits their applications on regional or local scales. Machine learning driven by remote sensing observations offers an opportunity to monitor regional scale droughts. However, the limited time range of remote sensing observations such as vegetation index (VI) resulted in a substantial gap in generating high resolution drought index products before 2000. This study generated spatiotemporally continuous Standardized Precipitation Evapotranspiration Index (SPEI) data spanning from 1901–2018 in southwestern China by machine learning. It indicated that four Classification and Regression Tree (CART) approaches, decision trees (DT), random forest (RF), gradient boosted regression trees (GBRT) and extra trees (ET), can provide valid local drought information by downscaling the Estación Experimental de Aula Dei (EEAD) data. The in-situ SPEI dataset produced by the Penman–Monteith method was used as a benchmark to evaluate the temporal and spatial performance of the downscaled SPEI. In addition, the necessity of VI in SPEI downscaling was also assessed. The results showed that: (1) the ET-based product has the best performance (R2 = 0.889, MAE = 0.232, RMSE = 0.432); (2) the VI provides no significant improvement for SPEI re-construction; (3) topography exerts an obvious influence on the downscaling process, and (4) the downscaled SPEI shows more consistency with the in-situ SPEI compared with EEAD SPEI. The proposed method can be easily extended to other areas without in-situ data and enhance the ability of long-term drought monitoring.

Can Forest-Related Adaptive Capacity Reduce Landslide Risk Attributable to Climate Change?—Case of Republic of Korea

Recent cases of climate disasters such as the European floods in 2021 and Korea’s longest rainy season in 2020 strongly imply the importance of adaptation to climate change. In this study, we performed a numerical prediction on how much climate change adaptation factors related to forest policy can reduce climate disasters such as landslides. We focused on the landslide in Korea and applied a machine learning model reflecting adaptive indicators in the representative concentration pathway 8.5 climate scenario. The changes in the landslide probability were estimated using the Random Forest model, which estimated the landslide probability in the baseline period (2011) with excellent performance, and the spatial adaptation indicators used in this study contributed approximately 20%. The future landslide risk predicting indicated a significant increase in the Very High and High risk areas, especially in 2092. The application of the forest-related adaptation indices based on the policy scenario showed that in 2050, the effect was not pronounced, but in 2092, when the risk of landslides was much higher, the effect increased significantly. In particular, the effect was remarkable in the Seoul metropolitan and southern coastal regions. Even with the same adaptive capacity, it exerted a larger effect on the enhanced disasters. Our results suggest that the enhancement of adaptive capacity can reduce landslide risk up to 70% in a Very High risk region. In conclusion, it implies an importance to respond to the intensifying climate disasters, and abundant follow-up studies are expected to appear in the future.

A state-of-the-art review on modeling the biochar effect: Guidelines for beginners

Biochar has been widely advocated due to its special properties and sustainability for agriculture soil amendment. The influencing mechanism of biochar on soil properties is a key aspect of quantifying and predicting its benefits and trade-offs. The contribution of biochar to both environmental and agricultural benefits has been deeply discussed and extensively reviewed, but few reviews have focused on modeling biochar effects. An overview of recent advances in biochar modeling is illustrated and approaches classified in this paper. Applications of a machine learning model, a deterministic model, and a numerical model to biochar are categorized and summarized. A discussion of the advantages and disadvantages of each model and a comparison among them are also provided. Finally, this paper gives many suggestions on narrowing the knowledge gap to advance biochar modeling. Further study of biochar modeling in management planning and design and application of the model results in agricultural systems will help accelerate the expansion of biochar's application scale and encourage the efficient utilization of waste in agricultural systems.

Toward an Integrated Disaster Management Approach: How Artificial Intelligence Can Boost Disaster Management

Technical and methodological enhancement of hazards and disaster research is identified as a critical question in disaster management. Artificial intelligence (AI) applications, such as tracking and mapping, geospatial analysis, remote sensing techniques, robotics, drone technology, machine learning, telecom and network services, accident and hot spot analysis, smart city urban planning, transportation planning, and environmental impact analysis, are the technological components of societal change, having significant implications for research on the societal response to hazards and disasters. Social science researchers have used various technologies and methods to examine hazards and disasters through disciplinary, multidisciplinary, and interdisciplinary lenses. They have employed both quantitative and qualitative data collection and data analysis strategies. This study provides an overview of the current applications of AI in disaster management during its four phases and how AI is vital to all disaster management phases, leading to a faster, more concise, equipped response. Integrating a geographic information system (GIS) and remote sensing (RS) into disaster management enables higher planning, analysis, situational awareness, and recovery operations. GIS and RS are commonly recognized as key support tools for disaster management. Visualization capabilities, satellite images, and artificial intelligence analysis can assist governments in making quick decisions after natural disasters.

Drought Vulnerability Assessment Using Geospatial Techniques in Southern Queensland, Australia

In Australia, droughts are recurring events that tremendously affect environmental, agricultural and socio-economic activities. Southern Queensland is one of the most drought-prone regions in Australia. Consequently, a comprehensive drought vulnerability mapping is essential to generate a drought vulnerability map that can help develop and implement drought mitigation strategies. The study aimed to prepare a comprehensive drought vulnerability map that combines drought categories using geospatial techniques and to assess the spatial extent of the vulnerability of droughts in southern Queensland. A total of 14 drought-influencing criteria were selected for three drought categories, specifically, meteorological, hydrological and agricultural. The specific criteria spatial layers were prepared and weighted using the fuzzy analytical hierarchy process. Individual categories of drought vulnerability maps were prepared from their specific indices. Finally, the overall drought vulnerability map was generated by combining the indices using spatial analysis. Results revealed that approximately 79.60% of the southern Queensland region is moderately to extremely vulnerable to drought. The findings of this study were validated successfully through the receiver operating characteristics curve (ROC) and the area under the curve (AUC) approach using previous historical drought records. Results can be helpful for decision makers to develop and apply proactive drought mitigation strategies.

Determining and forecasting drought susceptibility in southwestern Iran using multi-criteria decision-making (MCDM) coupled with CA-Markov model

Forecasting drought and determining relevant data to predict drought are an important topic for decision-makers and planners. It is critical to predicting drought in the south of Fars province, an important agricultural center in Iran located in arid and semi-arid climates. The purpose of this study was to generate a drought map in 2019 using 12 parameters: altitude, aridity index, erosion, groundwater depth, land use, PET (Potential evapotranspiration), precipitation days, precipitation, slope, soil texture, soil salinity, and distance to river, and predict drought maps in 2030 and 2040 using the cellular automata (CA)-Markov model spatially. The fuzzy method was first used to homogenize the data. Next, by evaluating each parameter, the weight of each parameter was calculated using the analytic hierarchy process (AHP), and a map of drought-prone areas was generated. The results of the fuzzy-AHP method showed that the eastern and southeastern regions of the study area were prone to drought. The four most predictive parameters in causing drought, i.e., aridity index, PET, precipitation, and soil texture, were selected using the Best search method and were then chosen as the input to determine drought mapping using the fuzzy and AHP methods. Finally, the CA-Markov model was used to predict future drought maps, and the results showed that in 2030 and 2040 the drought situation in the east and south of the study area would intensify.

Prediction of meteorological drought by using hybrid support vector regression optimized with HHO versus PSO algorithms

Drought is considered one of the costliest natural disasters that result in water scarcity and crop damage almost every year. Drought monitoring and forecasting are essential for the efficient management of water resources and sustainability in agriculture. However, the design of a consistent drought prediction model based on the dynamic relationship of the drought index with its antecedent values remains a challenging task. In the present research, the SVR (support vector regression) model was hybridized with two different optimization algorithms namely; Particle Swarm Optimization (PSO) and Harris Hawks Optimization (HHO) for reliable prediction of effective drought index (EDI) 1 month ahead, at different locations of Uttarakhand State of India. The inputs of the models were selected through partial autocorrelation function (PACF) analysis. The output produced by the SVR-HHO and SVR-PSO models was compared with the EDI estimated from observed data using five statistical indicators, i.e., RMSE (Root Mean Square Error), MAE (Mean Absolute Error), COC (Coefficient of Correlation), NSE (Nash-Sutcliffe Efficiency), WI (Willmott Index), and graphical inspection of radar-chart, time-variation plot, box-whisker plot, and Taylor diagram. Appraisal of results indicates that the SVR-HHO model (RMSE = 0.535-0.965, MAE = 0.363-0.622, NSE = 0.558-0.860, COC = 0.760-0.930, and WI = 0.862-0.959) outperformed the SVR-PSO model (RMSE = 0.546-0.967, MAE = 0.372-0.625, NSE = 0.556-0.855, COC = 0.758-0.929, and WI = 0.861-0.956) in predicting EDI. Visual inspection of model performances also showed a better performance of SVR-HHO compared to SVR-PSO in replicating the median, inter-quartile range, spread, and pattern of the EDI estimated from observed rainfall. The results indicate that the hybrid SVR-HHO approach can be utilized for reliable EDI predictions in the study area.

Discovering hidden spatial patterns and their associations with controlling factors for potentially toxic elements in topsoil using hot spot analysis and K-means clustering analysis

The understanding of sources and controlling factors of potentially toxic elements (PTEs) in soils plays an important role in the improvement of environmental management. With the rapid growth of data volume, effective methods are required for data analytics for the large geochemical data sets. In recent years, spatial machine learning technologies have been proven to have the potential to reveal hidden spatial patterns in order to extract geochemical information. In this study, two spatial clustering techniques of Getis-Ord G_i* statistic and K-means clustering analysis were performed on 15 PTEs in 6,862 topsoil samples from the Tellus datasets of Northern Ireland to investigate the hidden spatial patterns and association with their controlling factors. The spatial clustering patterns of hot spots (high values) and cold spots (low values) for the 15 PTEs were revealed, showing clear association with geological features, especially peat and basalt. Peat was associated with high concentrations of Bi, Pb, Sb and Sn, while basalt was associated with high concentrations of Co, Cr, Cu, Mn, Ni, V and Zn. The high concentrations of As, Ba, Mo and U were associated with mixture of various lithologies, indicating the complicated influences on them. In addition, three hidden patterns in the 6,862 soil samples were revealed by K-means clustering analysis. The soil samples in the first and second clusters were overlaid on the peatland and basalt formation, respectively, while the samples in the third cluster were overlaid on the mixture of the other lithologies. These hidden patterns of soil samples were consistent with the spatial clustering patterns for PTEs, highlighting the dominant control of peat and basalt in the topsoil of Northern Ireland. This study demonstrates the power of spatial machine learning techniques in identifying hidden spatial patterns, providing evidences to extract geochemical knowledge in environmental studies.

Machine Learning on the COVID-19 Pandemic, Human Mobility and Air Quality: A Review

The ongoing COVID-19 global pandemic is touching every facet of human lives (e.g., public health, education, economy, transportation, and the environment). This novel pandemic and non-pharmaceutical interventions of lockdown and confinement implemented citywide, regionally or nationally are affecting virus transmission, people's travel patterns, and air quality. Many studies have been conducted to predict the diffusion of the COVID-19 disease, assess the impacts of the pandemic on human mobility and on air quality, and assess the impacts of lockdown measures on viral spread with a range of Machine Learning (ML) techniques. This literature review aims to analyze the results from past research to understand the interactions among the COVID-19 pandemic, lockdown measures, human mobility, and air quality. The critical review of prior studies indicates that urban form, people's socioeconomic and physical conditions, social cohesion, and social distancing measures significantly affect human mobility and COVID-19 viral transmission. During the COVID-19 pandemic, many people are inclined to use private transportation for necessary travel to mitigate coronavirus-related health problems. This review study also noticed that COVID-19 related lockdown measures significantly improve air quality by reducing the concentration of air pollutants, which in turn improves the COVID-19 situation by reducing respiratory-related sickness and deaths. It is argued that ML is a powerful, effective, and robust analytic paradigm to handle complex and wicked problems such as a global pandemic. This study also explores the spatio-temporal aspects of lockdown and confinement measures on coronavirus diffusion, human mobility, and air quality. Additionally, we discuss policy implications, which will be helpful for policy makers to take prompt actions to moderate the severity of the pandemic and improve urban environments by adopting data-driven analytic methods.

Delineation of groundwater quality locations suitable for target end-use purposes through deep neural network models

Groundwater is the main source of water for beverages, and its quality varies depending on extraction location; this is particularly the case in regions with complex geology, topography, and multiple forms of land use. Thus, it is important to determine a suitable groundwater extraction location based on intended water use and the related water quality standards. In this study, deep neural network (DNN) models and GIS data relating to groundwater quality were applied to estimate potential maps of Gangwon Province in South Korea, where groundwater is frequently extracted for drinking purposes. These maps specify areas where the groundwater quality is conducive for being used as mineral water and water for brewing coffee (hereafter referred as "coffee water"). Sensitivity analysis identified how inputs were sensitive to model estimation and showed that land-use variables were the most sensitive. The importance of each variable quantified how good or bad its region is for the desired groundwater. The overall features of importance were similar between mineral water and coffee water. However, with differences in hydrogeological units, carbonate rock was a variable of high positive importance for mineral water; metamorphic rock was its equivalent for coffee water. Our results offer a potential map of desired groundwater quality in the absence of a detailed understanding of the underlying hydrochemical processes governing groundwater quality. Additionally, the development of such a potential mapping model can help to determine the appropriate development area of groundwater for their respective purposes.

Estimation of Daily Stage–Discharge Relationship by Using Data-Driven Techniques of a Perennial River, India

Modeling the stage-discharge relationship in river flow is crucial in controlling floods, planning sustainable development, managing water resources and economic development, and sustaining the ecosystem. In the present study, two data-driven techniques, namely wavelet-based artificial neural networks (WANN) and a support vector machine with linear and radial basis kernel functions (SVM-LF and SVM-RF), were employed for daily discharge (Q) estimation. The hydrological data of daily stage (H) and discharge (Q) from June to October for 10 years (2004–2013) at the Govindpur station, situated in the Burhabalang river basin, Orissa, were considered for analysis. For model construction, an optimum number of inputs (lags) was extracted using the partial autocorrelation function (PACF) at a 5% level of significance. The outcomes of the WANN, SVM-LF, and SVM-RF models were appraised over the observed value of Q based on performance indicators, viz., root mean square error (RMSE), Nash–Sutcliffe efficiency (NSE), Pearson’s correlation coefficient (PCC), and Willmott index (WI), and through visual inspection (time variation, scatter plot, and Taylor diagram). Results of the evaluation showed that the SVM-RF model (RMSE = 104.426 m3/s, NSE = 0.925, PCC = 0.964, WI = 0.979) outperformed the WANN and SVM-LF models with the combination of three inputs, i.e., current stage, one-day antecedent stage, and discharge, during the testing period. In addition, the SVM-RF model was found to be more reliable and robust than the other models and having important implications for water resources management at the study site.

Flash Flood Susceptibility Modeling and Magnitude Index Using Machine Learning and Geohydrological Models: A Modified Hybrid Approach

In an arid region, flash floods (FF), as a response to climate changes, are the most hazardous causing massive destruction and losses to farms, human lives and infrastructure. A first step towards securing lives and infrastructure is the susceptibility mapping and predicting of occurrence sites of FF. Several studies have been applied using an ensemble machine learning model (EMLM) but measuring FF magnitude using a hybrid approach that integrates machine learning (MCL) and geohydrological models have not been widely applied. This study aims to modify a hybrid approach by testing three machine learning models. These are boosted regression tree (BRT), classification and regression trees (CART), and naive Bayes tree (NBT) for FF susceptibility mapping at the northern part of the United Arab Emirates (NUAE). This is followed by applying a group of accuracy metrics (precision, recall and F1 score) and the receiving operating characteristics (ROC) curve. The result demonstrated that the BRT has the highest performance for FF susceptibility mapping followed by the CART and NBT. After that, the produced FF map using the BRT was then modified by dividing it into seven basins, and a set of new FF conditioning parameters namely alluvial plain width, basin gradient and mean slope for each basin was calculated for measuring FF magnitude. The results showed that the mountainous and narrower basins (e.g., RAK, Masafi, Fujairah, and Rol Dadnah) have the highest probability occurrence of FF and FF magnitude, while the wider alluvial plains (e.g., Al Dhaid) have the lowest probability occurrence of FF and FF magnitude. The proposed approach is an effective approach to improve the susceptibility mapping of FF, landslides, land subsidence, and groundwater potentiality obtained using ensemble machine learning, which is used widely in the literature.

Using large-scale climate drivers to forecast meteorological drought condition in growing season across the Australian wheatbelt

Recurring drought has caused large crop yield losses in Australia during past decades. Long-term drought forecasting is of great importance for the development of risk management strategies. Recently, large-scale climate drivers (e.g. El Niño-Southern Oscillation) have been demonstrated as useful in the application of drought forecasting. Machine learning-based models that use climate drivers as input are commonly adopted to provide drought forecasts as these models are easy to develop and require less information compared to physical-based models. However, few machine learning-based models have been developed to forecast drought conditions during growing season across all Australian cropping areas. In this study, we developed a growing season (Apr.-Nov.) meteorological drought forecasting model for each climate gauging location across the Australian wheatbelt based on multiple lagged (past) large-scale climate indices and the Random Forest (RF) algorithm. The Standardized Precipitation Index (SPI) was used as the response variable to measure the degree of meteorological drought. Results showed that the RF model could provide satisfactory drought forecasts in the eastern areas of the wheatbelt with Pearson's correlation coefficient r > 0.5 and normalized Root Mean Square Error (nRMSE) < 23%. Forecasted drought maps matched well with observed drought maps for three representative periods. We identified NINO3.4 sea surface temperature and Multivariate ENSO Index as the most influential indices dominating growing season drought conditions across the wheatbelt. In addition, lagged impacts of large-scale climate drivers on growing season drought conditions were long-lasting and the indices in previous year could also potentially affect drought conditions during current year. As large-scale climate indices are readily available and can be rapidly used to feed data driven models, we believe the proposed meteorological drought forecasting models can be easily extended to other regions to provide drought outlooks which can help mitigate adverse drought impacts.

Capability and robustness of novel hybridized models used for drought hazard modeling in southeast Queensland, Australia

Widespread detrimental and long-lasting droughts are having catastrophic impacts around the globe. Researchers, organizations, and policy makers need to work together to obtain precise information, enabling timely and accurate decision making to mitigate drought impacts. In this study, a spatial modeling approach based on an adaptive neuro-fuzzy inference system (ANFIS) and several metaheuristic optimizations (ANFIS-BA, ANFIS-GA, ANFIS-ICA, ANFIS-PSO) was developed to predict the spatial occurrence of drought in a region in southeastern Queensland, Australia. In this approach, data describing the distribution of eight drought-contributing factors were prepared for input into the models to serve as independent variables. Relative departures of rainfall (RDR) and relative departures of soil moisture (RDSM) were analyzed to identify locations where drought conditions have occurred. The set of locations in the study area identified as having experienced drought conditions was randomly divided into two groups, 70% were used for training and 30% for validation. The models employed these data to generate maps that predict the locations that would be expected to experience drought. The prediction accuracy of the model-produced drought maps was scrutinized with two evaluation metrics: area under the receiver operating characteristic curve (AUC) and root mean square error (RMSE). The results demonstrate that the hybridized models (ANFIS-BA (AUC_mean = 83.7%, RMSE_mean = 0.236), ANFIS-GA (AUC_mean = 81.62%, RMSE_mean = 0.247), ANFIS-ICA (AUC_mean = 82.12%, RMSE_mean = 0.247), and ANFIS-PSO (AUC_mean = 81.42%, RMSE_mean = 0.255)) yield better predictive performance than the standalone ANFIS model (AUC_mean = 71.8%, RMSE_mean = 0.344). Furthermore, sensitivity analyses indicated that plant-available water capacity, the percentage of soil comprised of sand, and mean annual precipitation were the most important predictors of drought hazard. The versatility of the new approach for spatial drought modeling and the capacity of ANFIS model hybridization to improve model performance suggests great potential to assist decision makers in their formulations of drought risk, recovery, and response management, and in the development of contingency plans.

A Modeling Comparison of Groundwater Potential Mapping in a Mountain Bedrock Aquifer: QUEST, GARP, and RF Models

In some arid regions, groundwater is the only source of water for human needs, so understanding groundwater potential is essential to ensure its sustainable use. In this study, three machine learning models (Genetic Algorithm for Rule-Set Production (GARP), Quick Unbiased Efficient Statistical Tree (QUEST), and Random Forest (RF)) were applied and verified for spatial prediction of groundwater in a mountain bedrock aquifer in Piranshahr Watershed, Iran. A spring location dataset consisting of 141 springs was prepared by field surveys, and from this three different sample datasets (S1–S3) were randomly generated (70% for training and 30% for validation). A total of 10 groundwater conditioning factors were prepared for modeling, namely slope percent, relative slope position (RSP), plan curvature, altitude, drainage density, slope aspect, topographic wetness index (TWI), terrain ruggedness index (TRI), land use, and lithology. The area under the receiver operating characteristic curve (AUC) and true skill statistic (TSS) were used to evaluate the accuracy of models. The results indicated that all models had excellent goodness-of-fit and predictive performance, but that RF (AUCmean = 0.995, TSSmean = 0.89) and GARP (AUCmean = 0.957, TSSmean = 0.82) outperformed QUEST (AUCmean = 0.949, TSSmean = 0.74). In robustness analysis, RF was slightly more sensitive than GARP and QUEST, making it necessary to consider several random partitioning options for preparing training and validation groups. The outcomes of this study can be useful in sustainable management of groundwater resources in the study region.

Assessing drought vulnerability using geospatial techniques in northwestern part of Bangladesh

Drought is a damaging and costly natural disaster that frequently affects many climatic regions in the world. A multi-criteria-based approach for integrated spatial drought vulnerability mapping that combines all drought categories is required to generate detailed vulnerability information for formulating drought mitigation strategies. This study presents a spatial multi-criteria integrated approach for mapping comprehensive drought vulnerability using geospatial techniques and an analytical hierarchy process (AHP). The developed approach was applied in the northwestern region of Bangladesh to justify its applicability. A total of 17 criteria under 4 drought categories, namely, meteorological, agricultural, hydrological and socio-economic, were selected. Moreover, spatial layers for each criterion were developed. AHP was used to calculate the weights for each criterion and drought types using pair-wise comparison matrices. Individual categories of drought and overall drought vulnerability maps were developed using the weighted overlay technique by integrating the corresponding criteria. The produced maps effectively defined the spatial extents and levels (e.g. normal, mild, moderate, severe and extreme) of drought vulnerability. Results demonstrated that approximately 77% of the total area of the north-western region of Bangladesh was moderately to extremely vulnerable to drought. The output of the developed approach was successfully validated using the receiver operating characteristics and area under the curve techniques. The findings suggest that the proposed approach is highly effective in mapping comprehensive drought vulnerability for formulating strong drought mitigation strategies.

Machine Learning to Evaluate Impacts of Flood Protection in Bangladesh, 1983–2014

Impacts of climate change adaptation strategies need to be evaluated using principled methods spanning sectors and longer time frames. We propose machine-learning approaches to study the long-term impacts of flood protection in Bangladesh. Available data include socio-economic survey and events data (death, migration, etc.) from 1983–2014. These multidecadal data, rare in their extent and quality, provide a basis for using machine-learning approaches even though the data were not collected or designed to assess the impact of the flood control investments. We test whether the embankment has affected the welfare of people over time, benefiting those living inside more than those living outside. Machine-learning approaches enable learning patterns in data to help discriminate between two groups: here households living inside vs. outside. They also help identify the most informative indicators of discrimination and provide robust metrics to evaluate the quality of the model. Overall, we find no significant difference between inside/outside populations based on welfare, migration, or mortality indicators. However, we note a significant difference in inward/outward movement with respect to the embankment. While certain data gaps and spatial heterogeneity in sampled populations suggest caution in any conclusive interpretation of the flood protection infrastructure, we do not see higher benefits accruing to those living with higher levels of protection. This has implications for Bangladesh’s planning for future and more extreme climate futures, including the national Delta Plan, and global investments in climate resilient infrastructure to create positive social impacts.

Spatial Modeling of Snow Avalanche Using Machine Learning Models and Geo-Environmental Factors: Comparison of Effectiveness in Two Mountain Regions

Although snow avalanches are among the most destructive natural disasters, and result in losses of life and economic damages in mountainous regions, far too little attention has been paid to the prediction of the snow avalanche hazard using advanced machine learning (ML) models. In this study, the applicability and efficiency of four ML models: support vector machine (SVM), random forest (RF), naïve Bayes (NB) and generalized additive model (GAM), for snow avalanche hazard mapping, were evaluated. Fourteen geomorphometric, topographic and hydrologic factors were selected as predictor variables in the modeling. This study was conducted in the Darvan and Zarrinehroud watersheds of Iran. The goodness-of-fit and predictive performance of the models was evaluated using two statistical measures: the area under the receiver operating characteristic curve (AUROC) and the true skill statistic (TSS). Finally, an ensemble model was developed based upon the results of the individual models. Results show that, among individual models, RF was best, performing well in both the Darvan (AUROC = 0.964, TSS = 0.862) and Zarrinehroud (AUROC = 0.956, TSS = 0.881) watersheds. The accuracy of the ensemble model was slightly better than all individual models for generating the snow avalanche hazard map, as validation analyses showed an AUROC = 0.966 and a TSS = 0.865 in the Darvan watershed, and an AUROC value of 0.958 and a TSS value of 0.877 for the Zarrinehroud watershed. The results indicate that slope length, lithology and relative slope position (RSP) are the most important factors controlling snow avalanche distribution. The methodology developed in this study can improve risk-based decision making, increases the credibility and reliability of snow avalanche hazard predictions and can provide critical information for hazard managers.

A Novel Intelligence Approach of a Sequential Minimal Optimization-Based Support Vector Machine for Landslide Susceptibility Mapping

The main objective of this study is to propose a novel hybrid model of a sequential minimal optimization and support vector machine (SMOSVM) for accurate landslide susceptibility mapping. For this task, one of the landslide prone areas of Vietnam, the Mu Cang Chai District located in Yen Bai Province was selected. In total, 248 landslide locations and 15 landslide-affecting factors were selected for landslide modeling and analysis. Predictive capability of SMOSVM was evaluated and compared with other landslide models, namely a hybrid model of the cascade generalization optimization-based support vector machine (CGSVM), individual models, such as support vector machines (SVM) and naïve Bayes trees (NBT). For validation, different quantitative criteria such as statistical based methods and area under the receiver operating characteristic curve (AUC) technique were used. Results of the study show that the SMOSVM model (AUC = 0.824) has the highest performance for landslide susceptibility mapping, followed by CGSVM (AUC = 0.815), SVM (AUC = 0.804), and NBT (AUC = 0.800) models, respectively. Thus, the proposed novel SMOSVM model is a promising method for better landslide susceptibility mapping and prediction, which can be applied also in other landslide prone areas.