Search Results (35)

The location of housing has a significant influence on its pricing. Generally, spatial self-correlation and spatial heterogeneity phenomena affect housing price data. Additionally, time is a crucial factor in housing price modeling, as it helps understand market trends and fluctuations. Currency market fluctuations also directly affect housing prices. Therefore, in addition to the physical features of the property, such as the area of the residential unit and building age, the rate of exchange (dollar price) is added to the independent variable set. This study used the real estate transaction records from Iran’s registration system, covering February, May, August, and November in 2017–2019. Initially, 7464 transactions were collected, but after preprocessing, the dataset was refined to 7161 records. Unlike feedforward neural networks, the generalized regression neural network does not converge to local minimums, so in this research, the Geographically, Temporally, and Characteristically Weighted Generalized Regression Neural Network (GTCW-GRNN) for housing price modeling was developed. In addition to being able to model the spatial–time heterogeneity available in observations, this algorithm is accurate and faster than MLR, GWR, GRNN, and GCW-GRNN. The average index of the adjusted coefficient of determination in other methods, including the MLR, GWR, GTWR, GRNN, GCW-GRNN, and the proposed GTCW-GRNN in different modes of using Euclidean or travel distance and fixed or adaptive kernel was equal to 0.760, 0.797, 0.854, 0.777, 0.774, and 0.813, respectively, which showed the success of the proposed GTCW-GRNN algorithm. The results showed the importance of the variable of the dollar and the area of housing significantly. Full article

This study aimed to enhance the accuracy of forest fire susceptibility mapping (FSM) by innovatively applying recursive feature elimination (RFE) with an ensemble of machine learning models, specifically Support Vector Machine (SVM) and Random Forest (RF), to identify key fire factors. The fire zones were derived from MODIS satellite imagery from 2012 to 2017. Further validation of these data has been provided by field surveys and reviews of land records in rangelands and forests; a total of 326 fire points were determined in this study. Seventeen factors involving topography, geomorphology, meteorology, hydrology, and human factors were identified as being effective primary factors in triggering and spreading fires in the selected mountainous case study area. As a first step, the RFE models RF, Extra Trees, Gradient Boosting, and AdaBoost were used to identify important fire factors among all selected primary factors. The SVM and RF models were applied once on all factors and secondly on those derived from the RFE model as the key factors in FSM. Training and testing data were divided tenfold, and the model’s performance was evaluated using cross-validation. Various metrics, including recall, precision, F1 score, accuracy, area under the curve (AUC), Matthew’s correlation coefficient (MCC), and Kappa, were employed to measure the performance of the models. The assessments demonstrate that leveraging RFE models enhances the FSM results by identifying key factors and excluding unnecessary ones. Notably, the SVM model exhibits significant improvement, achieving an increase of over 10.97% in accuracy and 8.61% in AUC metrics. This improvement underscores the effectiveness of the RFE approach in enhancing the predictive performance of the SVM model. Full article

Identifying the underlying factors derived from geospatial and remote sensing data that contribute to forest fires is of paramount importance. It aids experts in pinpointing areas and periods most susceptible to these incidents. In this study, we employ the geographically and temporally weighted regression (GTWR) method in conjunction with a refined continuous invasive weed optimization (CIWO) algorithm to assess certain spatially relevant drivers of forest fires, encompassing both biophysical and anthropogenic influences. Our proposed approach demonstrates theoretical utility in addressing the spatial regression problem by meticulously accounting for the autocorrelation and non-stationarity inherent in spatial data. We leverage tricube and Gaussian kernels to weight the GTWR for two distinct temporal datasets, yielding coefficients of determination (R2) amounting to 0.99 and 0.97, respectively. In contrast, traditional geographically weighted regression (GWR) using the tricube kernel achieved R2 values of 0.87 and 0.88, while the Gaussian kernel yielded R2 values of 0.8138 and 0.82 for the same datasets. This investigation underscores the substantial impact of both biophysical and anthropogenic factors on forest fires within the study areas. Full article

The problem of estimating earthquake risk is one of the primary themes for researchers and investigators in the field of geosciences. The combined assessment of spatial probability and the determination of earthquake risk at large scales is challenging. To the best of the authors’ knowledge, there no updated earthquake-hazard-and-risk assessments for the Eurasia region have been published since 1999. Considering that Eurasia is characterized by a seismically active Alpine–Himalayan fault zone and the Pacific Ring of Fire, which are frequently affected by devastating events, a continental-scale risk assessment for Eurasia is necessary to check the global applicability of developed methods and to update the earthquake-hazard, -vulnerability, and -risk maps. The current study proposes an integrated deep-transfer-learning approach called the gated recurrent unit–simple recurrent unit (GRU–SRU) to estimate earthquake risk in Eurasia. In this regard, the GRU model estimates the spatial probability, while the SRU model evaluates the vulnerability. To this end, spatial probability assessment (SPA), and earthquake-vulnerability assessment (EVA) results were integrated to generate risk A, while the earthquake-hazard assessment (EHA) and EVA were considered to generate risk B. This research concludes that in the case of earthquake-risk assessment (ERA), the results obtained for Risk B were better than those for risk A. Using this approach, we also evaluated the stability of the factors and interpreted the interaction values to form a spatial prediction. The accuracy of our proposed integrated approach was examined by means of a comparison between the obtained deep learning (DL)-based results and the maps generated by the Global Earthquake Model (GEM). The accuracy of the SPA was 93.17%, while that of the EVA was 89.33%. Full article

Among all the natural hazards, earthquake prediction is an arduous task. Although many studies have been published on earthquake hazard assessment (EHA), very few have been published on the use of artificial intelligence (AI) in spatial probability assessment (SPA). There is a great deal of complexity observed in the SPA modeling process due to the involvement of seismological to geophysical factors. Recent studies have shown that the insertion of certain integrated factors such as ground shaking, seismic gap, and tectonic contacts in the AI model improves accuracy to a great extent. Because of the black-box nature of AI models, this paper explores the use of an explainable artificial intelligence (XAI) model in SPA. This study aims to develop a hybrid Inception v3-ensemble extreme gradient boosting (XGBoost) model and shapely additive explanations (SHAP). The model would efficiently interpret and recognize factors’ behavior and their weighted contribution. The work explains the specific factors responsible for and their importance in SPA. The earthquake inventory data were collected from the US Geological Survey (USGS) for the past 22 years ranging the magnitudes from 5 M_w and above. Landsat-8 satellite imagery and digital elevation model (DEM) data were also incorporated in the analysis. Results revealed that the SHAP outputs align with the hybrid Inception v3-XGBoost model (87.9% accuracy) explanations, thus indicating the necessity to add new factors such as seismic gaps and tectonic contacts, where the absence of these factors makes the prediction model performs poorly. According to SHAP interpretations, peak ground accelerations (PGA), magnitude variation, seismic gap, and epicenter density are the most critical factors for SPA. The recent Turkey earthquakes (M_w 7.8, 7.5, and 6.7) due to the active east Anatolian fault validate the obtained AI-based earthquake SPA results. The conclusions drawn from the explainable algorithm depicted the importance of relevant, irrelevant, and new futuristic factors in AI-based SPA modeling. Full article

The improvement in computer vision, sensor quality, and remote sensing data availability makes satellite imagery increasingly useful for studying human settlements. Several challenges remain to be overcome for some types of settlements, particularly for internally displaced populations (IDPs) and refugee camps. Refugee-dwelling footprints and detailed information derived from satellite imagery are critical for a variety of applications, including humanitarian aid during disasters or conflicts. Nevertheless, extracting dwellings remains difficult due to their differing sizes, shapes, and location variations. In this study, we use U-Net and residual U-Net to deal with dwelling classification in a refugee camp in northern Cameroon, Africa. Specifically, two semantic segmentation networks are adapted and applied. A limited number of randomly divided sample patches is used to train and test the networks based on a single image of the WorldView-3 satellite. Our accuracy assessment was conducted using four different dwelling categories for classification purposes, using metrics such as Precision, Recall, F1, and Kappa coefficient. As a result, F1 ranges from 81% to over 99% and approximately 88.1% to 99.5% based on the U-Net and the residual U-Net, respectively. Full article

Motivated by the lack of research on land cover and dust activity in the Middle East, this study seeks to increase the understanding of the sensitivity of dust centers to climatic and surface conditions in this specific region. In this regard, we explore vegetation cover and dust emission interactions using 16-day long-term Normalized Difference Vegetation Index (NDVI) data and daily Aerosol Optical Depth (AOD) data from Moderate Resolution Imaging Spectroradiometer (MODIS) and conduct spatiotemporal and statistical analyses. Eight major dust hotspots were identified based on long-term AOD data (2000–2019). Despite the relatively uniform climate conditions prevailing throughout the region during the study period, there is considerable spatial variability in interannual relationships between AOD and NDVI. Three subsets of periods (2000–2006, 2007–2013, 2014–2019) were examined to assess periodic spatiotemporal changes. In the second period (2007–2013), AOD increased significantly (6% to 32%) across the studied hotspots, simultaneously with a decrease in NDVI (−0.9% to −14.3%) except in Yemen−Oman. Interannual changes over 20 years showed a strong relationship between reduced vegetation cover and increased dust intensity. The correlation between NDVI and AOD (−0.63) for the cumulative region confirms the significant effect of vegetation canopy on annual dust fluctuations. According to the results, changes in vegetation cover have an essential role in dust storm fluctuations. Therefore, this factor must be regarded along with wind speed and other climate factors in Middle East dust hotspots related to research and management efforts. Full article

This work evaluates the performance of three machine learning (ML) techniques, namely logistic regression (LGR), linear regression (LR), and support vector machines (SVM), and two multi-criteria decision-making (MCDM) techniques, namely analytical hierarchy process (AHP) and the technique for order of preference by similarity to ideal solution (TOPSIS), for mapping landslide susceptibility in the Chitral district, northern Pakistan. Moreover, we create landslide inventory maps from LANDSAT-8 satellite images through the change vector analysis (CVA) change detection method. The change detection yields more than 500 landslide spots. After some manual post-processing correction, the landslide inventory spots are randomly split into two sets with a 70/30 ratio for training and validating the performance of the ML techniques. Sixteen topographical, hydrological, and geological landslide-related factors of the study area are prepared as GIS layers. They are used to produce landslide susceptibility maps (LSMs) with weighted overlay techniques using different weights of landslide-related factors. The accuracy assessment shows that the ML techniques outperform the MCDM methods, while SVM yields the highest accuracy of 88% for the resulting LSM. Full article

The effects of the spatial resolution of remote sensing (RS) data on wildfire susceptibility prediction are not fully understood. In this study, we evaluate the effects of coarse (Landsat 8 and SRTM) and medium (Sentinel-2 and ALOS) spatial resolution data on wildfire susceptibility prediction using random forest (RF) and support vector machine (SVM) models. In addition, we investigate the fusion of the predictions from the different spatial resolutions using the Dempster–Shafer theory (DST) and 14 wildfire conditioning factors. Seven factors are derived separately from the coarse and medium spatial resolution datasets for the whole forest area of the Guilan Province, Iran. All conditional factors are used to train and test the SVM and RF models in the Google Earth Engine (GEE) software environment, along with an inventory dataset from comprehensive global positioning system (GPS)-based field survey points of wildfire locations. These locations are evaluated and combined with coarse resolution satellite data, namely the thermal anomalies product of the moderate resolution imaging spectroradiometer (MODIS) for the period 2009 to 2019. We assess the performance of the models using four-fold cross-validation by the receiver operating characteristic (ROC) curve method. The area under the curve (AUC) achieved from the ROC curve yields 92.15% and 91.98% accuracy for the respective SVM and RF models for the coarse RS data. In comparison, the AUC for the medium RS data is 92.5% and 93.37%, respectively. Remarkably, the highest AUC value of 94.71% is achieved for the RF model where coarse and medium resolution datasets are combined through DST. Full article

This paper proposes a new approach based on an unsupervised deep learning (DL) model for landslide detection. Recently, supervised DL models using convolutional neural networks (CNN) have been widely studied for landslide detection. Even though these models provide robust performance and reliable results, they depend highly on a large labeled dataset for their training step. As an alternative, in this paper, we developed an unsupervised learning model by employing a convolutional auto-encoder (CAE) to deal with the problem of limited labeled data for training. The CAE was used to learn and extract the abstract and high-level features without using training data. To assess the performance of the proposed approach, we used Sentinel-2 imagery and a digital elevation model (DEM) to map landslides in three different case studies in India, China, and Taiwan. Using minimum noise fraction (MNF) transformation, we reduced the multispectral dimension to three features containing more than 80% of scene information. Next, these features were stacked with slope data and NDVI as inputs to the CAE model. The Huber reconstruction loss was used to evaluate the inputs. We achieved reconstruction losses ranging from 0.10 to 0.147 for the MNF features, slope, and NDVI stack for all three study areas. The mini-batch K-means clustering method was used to cluster the features into two to five classes. To evaluate the impact of deep features on landslide detection, we first clustered a stack of MNF features, slope, and NDVI, then the same ones plus with the deep features. For all cases, clustering based on deep features provided the highest precision, recall, F1-score, and mean intersection over the union in landslide detection. Full article

The importance of Land Cover (LC) classification is recognized by an increasing number of scholars who employ LC information in various applications (i.e., address global climate change and achieve sustainable development). However, studying the roles of balancing data, image integration, and performance of different machine learning algorithms in various landscapes has not received as much attention from scientists. Therefore, the present study investigates the performance of three frequently used Machine Learning (ML) algorithms, including Extreme Learning Machines (ELM), Support Vector Machines (SVM), and Random Forest (RF) in LC mapping at six different landscapes. Moreover, the Geometric Synthetic Minority Over-sampling Technique (G-SMOTE) was adopted to deal with the class imbalance problem. In this work, the time-series of Sentinel-1 and Sentinel-2 data were integrated to improve LC mapping accuracy, taking advantage of both data. Moreover, Support Vector Machine-Recursive Feature Elimination (SVM-RFE) was implemented to distinguish the most informative features. Based on the results, the RF integrated with G-SMOTE showed the best result for four landscapes (coastal, cropland, desert, and semi-arid). SVM integrated with G-SMOTE had the highest accuracy in the remaining two landscapes (plain and mountain). Applied ML algorithms showed good performances in various landscapes, ranging Overall Accuracy (OA) from 85% to 93% for RF, 83% to 94% for SVM, and 84% to 92% for ELM. The outcomes exhibit that although applying G-SMOTE may slightly decrease OA values, it generally boosts the results of LC classification accuracies in various landscapes, particularly for minority classes. Full article

In this study, we used Sentinel-1 and Sentinel-2 data to delineate post-earthquake landslides within an object-based image analysis (OBIA). We used our resulting landslide inventory map for training the data-driven model of the frequency ratio (FR) for landslide susceptibility modelling and mapping considering eleven conditioning factors of soil type, slope angle, distance to roads, distance to rivers, rainfall, normalised difference vegetation index (NDVI), aspect, altitude, distance to faults, land cover, and lithology. A fuzzy analytic hierarchy process (FAHP) also was used for the susceptibility mapping using expert knowledge. Then, we integrated the data-driven model of the FR with the knowledge-based model of the FAHP to reduce the associated uncertainty in each approach. We validated our resulting landslide inventory map based on 30% of the global positioning system (GPS) points of an extensive field survey in the study area. The remaining 70% of the GPS points were used to validate the performance of the applied models and the resulting landslide susceptibility maps using the receiver operating characteristic (ROC) curves. Our resulting landslide inventory map got a precision of 94% and the AUCs (area under the curve) of the susceptibility maps showed 83%, 89%, and 96% for the F-AHP, FR, and the integrated model, respectively. The introduced methodology in this study can be used in the application of remote sensing data for landslide inventory and susceptibility mapping in other areas where earthquakes are considered as the main landslide-triggered factor. Full article

We live in a sphere that has unpredictable and multifaceted landscapes that make the risk arising from several incidences that are omnipresent. Floods and landslides are widespread and recurring hazards occurring at an alarming rate in recent years. The importance of this study is to produce multi-hazard exposure maps for flooding and landslides for the federal State of Salzburg, Austria, using the selected machine learning (ML) approach of support vector machine (SVM) and random forest (RF). Multi-hazard exposure maps were established on thirteen influencing factors for flood and landslides such as elevation, slope, aspect, topographic wetness index (TWI), stream power index (SPI), normalized difference vegetation index (NDVI), geology, lithology, rainfall, land cover, distance to roads, distance to faults, and distance to drainage. We classified the inventory data for flood and landslide into training and validation with the widely used splitting ratio, where 70% of the locations are used for training, and 30% are used for validation. The accuracy assessment of the exposure maps was derived through ROC (receiver operating curve) and R-Index (relative density). RF yielded better results for both flood and landslide exposure with 0.87 for flood and 0.90 for landslides compared to 0.87 for flood and 0.89 for landslides using SVM. However, the multi-hazard exposure map for the State of Salzburg derived through RF and SVM provides the planners and managers to plan better for risk regions affected by both floods and landslides. Full article

Landslides are one of the most detrimental geological disasters that intimidate human lives along with severe damages to infrastructures and they mostly occur in the mountainous regions across the globe. Landslide susceptibility mapping (LSM) serves as a key step in assessing potential areas that are prone to landslides and could have an impact on decreasing the possible damages. The application of the fuzzy best-worst multi-criteria decision-making (FBWM) method was applied for LSM in Austria. Further, the role of employing a few numbers of pairwise comparisons on LSM was investigated by comparing the FBWM and Fuzzy Analytical Hierarchical Process (FAHP). For this study, a wide range of data was sourced from the Geological Survey of Austria, the Austrian Land Information System, Humanitarian OpenStreetMap Team, and remotely sensed data were collected. We used nine conditioning factors that were based on the previous studies and geomorphological characteristics of Austria, such as elevation, slope, slope aspect, lithology, rainfall, land cover, distance to drainage, distance to roads, and distance to faults. Based on the evaluation of experts, the slope conditioning factor was chosen as the best criterion (highest impact on LSM) and the distance to roads was considered as the worst criterion (lowest impact on LSM). LSM was generated for the region based on the best and worst criterion. The findings show the robustness of FBWM in landslide susceptibility mapping. Additionally, using fewer pairwise comparisons revealed that the FBWM can obtain higher accuracy as compared to FAHP. The finding of this research can help authorities and decision-makers to provide effective strategies and plans for landslide prevention and mitigation at the national level. Full article

The Zagros forests in Western Iran are valuable ecosystems that have been seriously damaged by human interference (harvesting the wood and forest sub-products, converting the forests to the agricultural lands, and grazing) and natural events (drought events and fire). In this study, we generated accurate land cover (LC), and tree canopy cover percentage (TCC%) maps for the forests of Shirvan County, a part of Zagros forests in Western Iran using Sentinel-2, Google Earth, and field data for protective management. First, we assessed the accuracy of Google Earth data using 300 random field plots in 10 different land cover types. For land cover mapping, we evaluated the performance of four supervised classification algorithms (minimum distance (MD), Mahalanobis distance (MaD), neural network (NN), and support vector machine (SVM)). The accuracy of the land cover maps was assessed using a set of 150 stratified random plots in Google Earth. We mapped the forest canopy cover by using the normalized difference vegetation index (NDVI) map, and field plots. We calculated the Pearson correlation between the NDVI values and the TCC% (obtained from field plots). The linear regression between the NDVI values and the TCC% was used to obtain the predictive model of TCC% based on the NDVI. The results showed that Google Earth data yielded an overall accuracy of 94.4%. The SVM algorithm had the highest accuracy for the classification of Sentinel-2 data with an overall accuracy of 81.33% and a kappa index of 0.76. The results of the forest canopy cover analysis showed a Pearson correlation coefficient of 0.93 between the NDVI and TCC%, which is highly significant. The results also showed that the linear regression model is a good predictive model for TCC% estimation based on the NDVI (r² = 0.864). The results can be used as a baseline for decision-makers to monitor land cover change in the region, whether produced by human activities or natural events and to establish measures for protective management of forests. Full article