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

The northeastern region of Bangladesh often experiences flash flooding during the pre-harvesting period of the boro rice crop, which is the major cereal crop in the country. In this study, our objective was to delineate the impact of the 2017 flash flood (that initiated on 27 March 2017) on boro rice using multi-temporal Landsat-8 OLI and MODIS data. Initially, we opted to use Landsat-8 OLI data for mapping the damages; however, during and after the flooding event the acquisition of cloud free images were challenging. Thus, we used this data to map the cultivated boro rice acreage considering the planting to mature stages of the crop. Also, in order to map the extent of the damaged boro area, we utilized MODIS data as their 16-day composites provided cloud free information. Our results indicated that both the cultivated and damaged boro area estimates based on satellite data had strong relationships while compared to the ground-based estimates (i.e., r² values approximately 0.92 for both cases, and RMSE of 18,374 and 9380 ha for cultivated and damaged areas, respectively). Finally, we believe that our study would be critical for planning and ensuring food security for the country.

Understanding the land surface temperature (LST) trends is crucial for policymakers and stakeholders to develop adaptation and mitigation strategies suitable for a sustainable environment coping in the face of climate change. This article presents a systematic review of the studies related to delineating spaceborne sensor based LST trends, including information on the instruments and constellations of satellites (missions) that provide thermal infrared (TIR) and passive microwave (PMW) observations. About 99% of the studies used TIR, where 76% were MODIS (onboard Terra/Aqua) observations. Opportunities, challenges, and research gaps for using the TIR and PMW observations were also explored, with instruments onboard either polar-orbiting or geostationary satellites. We identified that calibrated dataset (e.g., processed, harmonized, and standardized) is extremely limited for each constellation, with multiple satellites and instruments, to make it fully useful for the entire mission period. A few problematic methodological concepts were identified, including using a few images in a longer time series. Using only a few images, acquired on different calendar months in different years, would not provide the true annual trends over the study period because they can be influenced by seasonal variations. To estimate the warming or cooling daytime, nighttime, or diurnal LST trends, the use of MODIS observations could be useful, even though it does not acquire images during the maximum or minimum temperature in a daily cycle. This article indicated further investigations into those research gaps and recommended directions to overcome most of these limitations.

Pan-sharpening is the process of fusing higher spatial resolution panchromatic (PAN) with lower spatial resolution multispectral (MS) imagery to create higher spatial resolution MS images. Here, our overall objective was to pan-sharpen Landsat-8 images and calculate vegetation greenness (i.e., normalized difference vegetation index (NDVI)), canopy structure (i.e., enhanced vegetation index (EVI)), and canopy water content (i.e., normalized difference water index (NDWI))-related variables. Our proposed methods consisted of: (i) evaluating the relationships between PAN band (0.503–0.676 µm) with a spatial resolution of 15 m and individual MS bands of Landsat-8 from blue (i.e., acquiring in the range 0.452–0.512 µm), green (i.e., 0.533–0.590 µm), red (i.e., 0.636–0.673 µm), near infrared (NIR: 0.851–0.879 µm), shortwave infrared-I (SWIR-I: 1.566–1.651 µm), and SWIR-II (2.107–2.294 µm) bands with a spatial resolution of 30 m; (ii) determining the suitable individual MS bands to be enhanced into the spatial resolution of the PAN band; and (iii) calculating several vegetation greenness and canopy moisture indices (i.e., NDVI, EVI, NDWI-I, and NDWI-II) at 15 m spatial resolution and subsequent validation using their equivalent-values at a spatial resolution of 30 m. Our analysis revealed that strong linear relationships existed between the PAN and most of the MS individual bands of interest except NIR. For example, r² values were 0.86–0.89 for blue band; 0.89–0.95 for green band; 0.84–0.96 for red band; 0.71–0.79 for SWIR-I band; and 0.71–0.83 for SWIR-II band. As a result, we performed smoothing filter-based intensity modulation method of pan-sharpening to enhance the spatial resolution of 30 m to 15 m. In calculating the vegetation indices, we used the enhanced MS images and resampled the NIR to 15 m. Finally, we evaluated these indices with their equivalents at 30 m spatial resolution and observed strong relationships (i.e., r² values in the range 0.98–0.99 for NDVI, 0.95–0.98 for EVI, 0.98–1.00 for NDWI).

The Athabasca River watershed plays a dominant role in both the economy and the environment in Alberta, Canada. Natural and anthropogenic factors rapidly changed the landscape of the watershed in recent decades. The dynamic of such changes in the landscape characteristics of the watershed calls for a comprehensive and up-to-date land-use and land-cover (LULC) map, which could serve different user-groups and purposes. The aim of the study herein was to delineate a 2016 LULC map of the Athabasca River watershed using Landsat-8 Operational Land Imager (OLI) images, Moderate Resolution Imaging Spectroradiometer (MODIS)-derived enhanced vegetation index (EVI) images, and other ancillary data. In order to achieve this, firstly, a preliminary LULC map was developed through applying the iterative self-organizing data analysis (ISODATA) clustering technique on 24 scenes of Landsat-8 OLI. Secondly, a Terra MODIS-derived 250-m 16-day composite of 30 EVI images over the growing season was employed to enhance the vegetation classes. Thirdly, several geospatial ancillary datasets were used in the post-classification improvement processes to generate a final 2016 LULC map of the study area, exhibiting 14 LULC classes. Fourthly, an accuracy assessment was carried out to ensure the reliability of the generated final LULC classes. The results, with an overall accuracy and Cohen’s kappa of 74.95% and 68.34%, respectively, showed that coniferous forest (47.30%), deciduous forest (16.76%), mixed forest (6.65%), agriculture (6.37%), water (6.10%), and developed land (3.78%) were the major LULC classes of the watershed. Fifthly, to support the data needs of scientists across various disciplines, data fusion techniques into the LULC map were performed using the Alberta merged wetland inventory 2017 data. The results generated two useful maps applicable for hydro-ecological applications. Such maps depicted two specific categories including different types of burned (approximately 6%) and wetland (approximately 30%) classes. In fact, these maps could serve as important decision support tools for policy-makers and local regulatory authorities in the sustainable management of the Athabasca River watershed.

Here, the objective was to study the local warming trend and its driving factors in the natural subregions of Alberta using a remote-sensing approach. We applied the Mann–Kendall test and Sen’s slope estimator on the day and nighttime MODIS LST time-series images to map and quantify the extent and magnitude of monthly and annual warming trends in the 21 natural subregions of Alberta. We also performed a correlation analysis of LST anomalies (both day and nighttime) of the subregions with the anomalies of the teleconnection patterns, i.e., Pacific North American (PNA), Pacific decadal oscillation (PDO), Arctic oscillation (AO), and sea surface temperature (SST, Niño 3.4 region) indices, to identify the relationship. May was the month that showed the most significant warming trends for both day and night during 2001–2020 in most of the subregions in the Rocky Mountains and Boreal Forest. Subregions of Grassland and Parkland in southern and southeastern parts of Alberta showed trends of cooling during daytime in July and August and a small magnitude of warming in June and August at night. We also found a significant cooling trend in November for both day and night. We identified from the correlation analysis that the PNA pattern had the most influence in the subregions during February to April and October to December for 2001–2020; however, none of the atmospheric oscillations showed any significant relationship with the significant warming/cooling months.