Improving the accuracy and stability of daily runoff prediction is crucial for effective water resource management and flood control. This study proposed a novel stacking ensemble learning model based on attention mechanism for the daily runoff prediction. The proposed model has a two-layer structure with the base model and the meta model. Three machine learning models, namely random forest (RF), adaptive boosting (AdaBoost), and extreme gradient boosting (XGB) are used as the base models. The attention mechanism is used as the meta model to integrate the output of the base model to obtain predictions. The proposed model is applied to predict the daily inflow to Fuchun River Reservoir in the Qiantang River basin. The results show that the proposed model outperforms the base models and other ensemble models in terms of prediction accuracy. Compared with the XGB and weighted averaging ensemble (WAE) models, the proposed model has a 10.22% and 8.54% increase in Nash–Sutcliffe efficiency (NSE), an 18.52% and 16.38% reduction in root mean square error (RMSE), a 28.17% and 18.66% reduction in mean absolute error (MAE), and a 4.54% and 4.19% increase in correlation coefficient (r). The proposed model significantly outperforms the base model and simple stacking model indicated by both the Friedman test and the Nemenyi test. Thus, the proposed model can produce reasonable and accurate prediction of the reservoir inflow, which is of great strategic significance and application value in formulating the rational allocation and optimal operation of water resources and improving the breadth and depth of hydrological forecasting integrated services.

Flash floods are the most dangerous kinds of floods because they combine the destructive power of a flood with incredible speed. They occur when heavy rainfall exceeds the ability of the ground to absorb it. The main aim of this study is to generate flash flood maps using Analytical Hierarchy Process (AHP) and Frequency Ratio (FR) models in the river’s floodplain between the Jhelum River and Chenab rivers. A total of eight flash flood-causative physical parameters are considered for this study. Six parameters are based on remote sensing images of the Advanced Land Observation Satellite (ALOS), Digital Elevation Model (DEM), and Sentinel-2 Satellite, which include slope, elevation, distance from the stream, drainage density, flow accumulation, and land use/land cover (LULC), respectively. The other two parameters are soil and geology, which consist of different rock and soil formations, respectively. In the case of AHP, each of the criteria is allotted an estimated weight according to its significant importance in the occurrence of flash floods. In the end, all the parameters were integrated using weighted overlay analysis in which the influence value of drainage density was given the highest weight. The analysis shows that a distance of 2500 m from the river has values of FR ranging from 0.54, 0.56, 1.21, 1.26, and 0.48, respectively. The output zones were categorized into very low, low, moderate, high, and very high risk, covering 7354, 5147, 3665, 2592, and 1343 km², respectively. Finally, the results show that the very high flood areas cover 1343 km², or 6.68% of the total area. The Mangla, Marala, and Trimmu valleys were identified as high-risk zones of the study area, which have been damaged drastically many times by flash floods. It provides policy guidelines for risk managers, emergency and disaster response services, urban and infrastructure planners, hydrologists, and climate scientists.

Drought has been identified as one of the significant complicated natural disasters and exacerbated by land degradation and climate change. Hence, monitoring drought and evaluating its spatiotemporal dynamics are essential to manage regional drought conditions and protecting the natural environment. In this study, various single remote sensing-based drought Indices including Soil Moisture Condition Index (SMCI), Precipitation Condition Index (PCI), Temperature Condition Index (TCI) and Vegetation Condition Index (VCI), and combined RS-based drought Indices including Optimized Meteorological Drought Index (OMDI) and Synthesized Drought Index (SDI) have been used to investigate the spatiotemporal variations of meteorological and agricultural droughts between 2000 and 2021 in Iran. The in situ drought indices, including the Standardized Precipitation Index (SPI) and Standardized Precipitation Evapotranspiration Index (SPEI) series of 1, 3, 6, 12, and 24 months were utilized to verify remote sensing-based drought indices and evaluate their applicability for analyzing drought conditions. The results indicated that the correlation coefficients of the in situ drought indices with the combined drought indices are higher than the RS-based single drought indexes. Generally, single-factor drought indexes, including VCI, TCI, PCI, and SMCI have specific characteristics. The PCI and SMCI have an acceptable correlation with the short-term SPI and SPEI and are more applicable to monitoring short-term drought conditions. Further, the TCI has better performance in monitoring long-term drought conditions in Iran. This research concluded that the central, eastern, and southeastern parts of Iran mainly were experiencing exceptional and extreme drought conditions as the worst agricultural and meteorological drought conditions observed in the years 2008 and 2021 in the region during the last 20 years. The results also showed that, in 2019 and 2020, most areas of Iran had higher OMDI and SDI values and the severity of drought has decreased in these years. Particularly, this research provides an essential reference for reasonably choosing RS-based drought indices for monitoring meteorological and agricultural droughts from a local to global scale.

Landslides triggered in mountainous areas can have catastrophic consequences, threaten human life, and cause billions of dollars in economic losses. Hence, it is imperative to map the areas susceptible to landslides to minimize their risk. Around Abbottabad, a large city in northern Pakistan, a large number of landslides can be found. This study aimed to map the landslide susceptibility over these regions in Pakistan by using three Machine Learning (ML) techniques, specifically Linear Regression (LiR), Logistic Regression (LoR), and Support Vector Machine (SVM). Several influencing factors were used to identify the potential landslide areas, including elevation, slope degree, slope aspect, general curvature, plan curvature, profile curvature, landcover classification system, Normalized Difference Water Index (NDWI), Normalized Difference Vegetation Index (NDVI), soil, lithology, fault density, topographic roughness index, and road density. The weights of these factors were calculated using ML techniques. The weightage overlay tool is adopted to map the final output. According to three ML models, lithology, NDWI, slope, and LCCS significantly impact landslide occurrence. The area under the ROC curve (AUC) is applied to validate the performance of models, and the results show the AUC value of LiR (88%) is better than SVM (86%) and LoR (85%) models. ML models and final susceptibility map gives good accuracy, which can be reliable for the results. The study’s outcome provides baselines for policymakers to propose adequate protection and mitigation measures against the landslides in the region, and any other researcher can adopt this methodology to map the landslide susceptibility in another area having similar characteristics.

Land use and land cover (LULC) change is a main driver of global environmental change and has destructive effects on the structure and function of the ecosystem. This study attempts to detect temporal and spatial changes in LULC patterns of the Chalus watershed during the last two decades using multitemporal Landsat images and predict the future LULC changes and patterns of the Chalus watershed for the year 2040. A hybrid method between segment-based and pixel-based classification was applied for each Landsat image in 2001, 2014, and 2021 to produce LULC maps of the Chalus watershed. In this study, the transition potential maps and the transition probability matrices between LULC types were provided by the support vector machine algorithm and the Markov chain model, respectively, to project the 2021 and 2040 LULC maps. The achieved K-index values that compared the simulated LULC map with the actual LULC map of the year 2021 resulted in a Kstandard = 0.9160, Kno = 0.9379, Klocation = 0.9318, and KlocationStrata = 0.9320, showing good agreement between the actual and simulated LULC map. Analysis of the historical LULC changes depicted that during 2001–2021, the significant increase of agricultural land (14317 ha) and barren area (9063 ha), and the sharp decline of grassland (26215 ha), and forest cover (5989 ha) were the major LULC changes in the Chalus watershed. The model predicted that forest cover will continue to decrease from 29.46 % (50720.2667 ha) in 2021 to 25.67 % of area (44207.78694 ha) in 2040, as well as, unceasing expansion of barren area, agricultural land, and built-up area will be expected by 2040. Therefore, understanding the spatiotemporal dynamics of LULC change is extremely important to implement essential measures and minimize the destructive consequences of these changes.

The development of remote sensing images in recent years has made it possible to identify materials in inaccessible environments and study natural materials on a large scale. But hyperspectral images (HSIs) are a rich source of information with their unique features in various applications. However, several problems reduce the accuracy of HSI classification; for example, the extracted features are not effective, noise, the correlation of bands, and most importantly, the limited labeled samples. To improve accuracy in the case of limited training samples, we propose a multiscale dual-branch residual spectral–spatial network with attention to the HSI classification model named MDBRSSN in this article. First, due to the correlation and redundancy between HSI bands, a principal component analysis operation is applied to preprocess the raw HSI data. Then, in MDBRSSN, a dual-branch structure is designed to extract the useful spectral–spatial features of HSI. The advanced feature, multiscale abstract information extracted by the convolution neural network, is applied to image processing, which can improve complex hyperspectral data classification accuracy. In addition, the attention mechanisms applied separately to each branch enable MDBRSSN to optimize and refine the extracted feature maps. Such an MDBRSSN framework can learn and fuse deeper hierarchical spectral–spatial features with fewer training samples. The purpose of designing the MDBRSSN model is to have high classification accuracy compared to state-of-the-art methods when the training samples are limited, which is proved by the results of the experiments in this article on four datasets. In Salinas, Pavia University, Indian Pines, and Houston 2013, the proposed model obtained 99.64%, 98.93%, 98.17%, and 96.57% overall accuracy using only 1%, 1%, 5%, and 5% of labeled data for training, respectively, which are much better compared to the state-of-the-art methods.