This work presents a comprehensive study on fish detection in underwater environments using sonar images from the Caltech Fish Counting Dataset (CFC). We use the CFC dataset, initially designed for tracking purposes, to optimize and evaluate the performance of YOLO v7 and v8 models in fish detection. Our findings demonstrate a high performance of these deep learning models to accurately detect fish species in sonar images.
In our evaluation, YOLO v7 achieved an average precision of 68.3% (AP50) and 62.15% (AP75), while YOLO v8 demonstrated even a better performance with an average precision of 72.47% (AP50) and 66.21% (AP75) across the test dataset of 334,017 images. These high precision results underscore the effectiveness of these models in fish detection tasks under various underwater conditions.
With a dataset of 162,680 training images and 334,017 test images, our evaluation provides valuable insights into the models' performance and generalization across diverse underwater conditions. This study contributes to the advancement of underwater fish detection by showcasing the suitability of the CFC dataset and the efficacy of YOLO v7 and v8 models. These insights can pave the way for further advancements in fish detection, supporting conservation efforts and sustainable fisheries management.

Thermal anomalies detected by Earth observation satellites have been widely used to identify active fires. Fire Radiative Power (FRP) can be estimated from the radiance at medium wave infrared (3-5 μm) wavelengths, measured by multiple polar-orbiting and geostationary satellite sensors, and represents the instantaneous radiative energy that is released from actively burning fires. FRP has been used to support mapping of burned scars, by identifying core areas, and estimating trace gas and aerosol rate of emissions, hence strengthen monitoring of wildfires activities and their impact on environment and ecosystems.

Algorithms to operationally generate FRP products from Earth observation satellite acquisitions in near real-time account for background window statistics, corrections, adjustments and tests to eliminate false alarms, in order to distinguishing fire pixels from non-fire pixels. Nevertheless, a high percentage of thermal anomalies are wrongly classified as possible fire pixels. Source of misclassification could likely be the presence of real thermal anomalies of Earth surface, corresponding to pixels exhibiting significantly higher released radiative energy than background window area (e.g. industrial areas, solar photovoltaic plant).

This research study aims at presenting a methodological approach to identify thermal anomalies hotspots, misclassified as fire pixels. FRP products over Italy National territory, generated for the period 2022-2023 from SLSTR, MODIS and VIIRS satellite sensors and distributed by Copernicus, EUMETSAT and NASA FIRMS, have been collected and analysed.

A total of about 75000 FRP fire pixels have been first spatially and temporally intersected with EFFIS Burned Areas Database, distributed under the Copernicus Emergency Management Service, in order to identify misclassified fire pixels. Later, zonal statistics has been performed in order to evaluate fractional land cover within each fire pixel. Thermal anomalies hotspots misclassified as fire pixels have been identified using a cluster analysis in order to partition a data set into discrete subsets, based on defined distance measures like the spatial distance of the pixel centroids, the temporal frequencies of the pixels and fractional land cover of selected classes.

Results demonstrate that misclassified large surfaces, like industrial areas, can be identified from both spatial and temporal patterns, while other FRP false alarms are smaller in size and exhibit uneven temporal frequencies. Limitations of the proposed approach are discussed, recommending possible future improvements.

Results show the capability of the presented approach to identify thermal anomalies hotspots, misclassified as fire pixels, in order to generate static masks for FRP products post-processing, improving the capacity of FRP products in providing prompt and accurate information for operational services addressing the monitoring of wildfires and their impact on environment and ecosystems.

Cities may have local weather and climate that are significantly different from their surrounding rural areas due to different physical characteristics of urban surfaces and emissions of substances, the latter being modulated by the rhythm of the urban ecosystem. Although the urban heat island is the most recognizable phenomenon caused by these features, the altered radiative, thermal, moisture, and aerodynamic properties of the urban surface relative to the rural one also influence clouds formation as well as their characteristics. The study of the latter is of considerable interest as they are an important element of the energy and water budget of the surface. Regardless of the cloud’s importance, establishing and quantifying urban effects on cloudiness is a significant challenge, as their formation is not limited to processes confined within the urban area, but depends on processes spanning different spatial scales. The inhomogeneous surface of the rural surrounding and the complex landscape are common to many cities and that can further hinder the detection of the urban signal in the cloud features. Also, clouds can be blown by the prevailing wind, so those over the city may originate from quite a distant area. By using in-situ measurements as well as data from remote sensing instruments (ceilometers) located in the city centre and its outskirts, we attempt to detect and quantify the urban impact on cloudiness over the city of Sofia in this work. To achieve this, we analysed 10 years of data from automatic weather stations and ceilometers. The former allows us to calculate lifting condensation level, which has been used as a proxy for the cloud base height. The latter parameter is routinely determined by the ceilometers, which also allow us to evaluate cloud persistence. Urban effects were sought in both cloud parameters.

An Earth's crust contains 0.06 to 0.09% fluoride and its constituent, which affects the plant's growth and causes public health problems. This study was proposed for the Perambalur district, located in the southern part of India, and soils with high moisture availability to increase fluoride mobilization. Sentinel-1 data with an operational frequency of 5.4 GHz is used to perform a semi-empirical model. A modified approach is based on the relation between the salinity due to fluoride and the total salts. It makes sense that each saline-induced components provide a different loss angle value based on its concentration. Mostly, ion mobility in clay soil is limited by the interaction of the electrical potential and surface charges on the mineral and is controlled by soil moisture and permeability behavior. Based on the preceding, the salinity model DSDM is developed to distinguish saline soil from non-saline soil using conductance loss. Furthermore, the impact of fluoride contents is approximated from laboratory-based dielectric components (real and imaginary parts) of soil samples with high electrical conductivity under high and low fluoride conditions, making a loss angle of 4°1'. The dielectric behavior of the fluoride salt is investigated by relating the loss angle and real and imaginary parts of dielectric constants. An imaginary part of the dielectric component was sensitive to changes in dielectric loss and might be useful for predicting fluoride across vast areas over time. Finally, a fluoride value is estimated based on linear regression that considers conductance loss and field-observed fluoride values. The statistical results (R² = 0.86, RMSE = 1.90, and Bias = 0.35) obtained between the estimated and simulated values reveal that the C-band SAR data could delineate the fluoride levels over variable clay soil and soil with varying vegetation growth.

The variability in the river flow in relation to rainfall changes plays a pivotal role in understanding the occurrences of disaster in the Ganga-Brahmaputra basin. In the recent decade, findings are indicating an increase in flash floods and droughts due to the impacts of climate change. The present study is focused on the recent flood disaster in the Ganga-Brahmaputra basin mainly affecting the states of Bihar, and Assam in India during the months of July, August, September, and October 2022. SAR data-based flood inundation analysis was performed using the Google Earth Engine platform. The composite area for July to October under flood was estimated to be 11048.52 km² for Assam and 4362.71 km² for Bihar. Extensive area was inundated in Assam due to increased rainfall intensity during the monsoon season with values ranging from 384.92 to 476.07mm as compared to Bihar with 272.56 to 386mm. MODIS Land Cover Type Yearly Global dataset was used to calculate the impact of flood on agricultural area and urban areas which revealed that the highest impact was on Bihar with an area of 1563.14 km², 463.81 km² respectively, and Assam (161.19 km², 51.91 km²). JRC Global Human Settlement Population Density layer was employed to evaluate the flood impacts which revealed that a total of 3.7 million of the population was affected. The highest impact in Bihar can be attributed to its vast population size and extensive settlements near riverbed areas whereas in Assam despite the higher inundated area the impact was less on agricultural land and urban area and more on other land cover classes. Following the same analysis, the trend of rainfall variability was also examined based on the Chirps precipitation data for the months of July to October from the year 1993-2022. The rainfall anomaly data was used to estimate the millimeter change per year. Bihar recorded the least precipitation variation in the month of August with major values ranging between 0.5 to -0.5mm year^-1 however the month of October presents the highest variability over the past 30 years with an increase of more than 1.5mm year^-1 after 2020 which has previously been non-significant. The major years when a prominent increase was observed are 1993, 1998, 2003-2005, and 2008. In recent years, the rainfall trend in Bihar is shifting towards the end of the monsoon with higher precipitation anomaly in October as compared to July. However, in Assam, the monsoon period has not recorded significant change in the anomaly from July to September with values falling between -0.5 to 1.0mm year^-1. In October, the region recorded an increase of more than 1.0mm year^-1 in the years 2000-2004, 2014, 2018, 2020, and 2022 with a maximum increase of 2.0mm year^-1 in 2022. The trend analysis of precipitation change will provide an approach to identify the areas at risk with increasing or decreasing anomaly of 2mm year^-1 in terms of flood and drought disasters which will provide necessary information regarding the reduction, mitigation, and management of disasters.

Understanding the Land Use Land Cover (LULC) changes driven by urbanization, socioeconomic growth, deforestation, agricultural practices, mining activities, etc. is vital for assessing the dynamics of hydrometeorological parameters of a river basin. Thus, a decadal evaluation of spatio-temporal changes in LULC for the years 2001, 2010 and 2020 was carried out with an objective to understand the impact on its hydrometeorological conditions of the upper Subarnarekha River Basin (SRB). Satellite based multi-temporal data such as Climate Hazards Group InfraRed Precipitation with Station (CHIRPS) and Famine Early Warning Systems Network (FEWS NET) Land Data Assimilation System (FLDAS) temperature data (2001-2020) was used to analyse the climatic implications. Additionally, conventional groundwater level data for the same period were collected from the India-Water Resource Information System (WRIS). The study examined the impact of changes in Land Use and Land Cover (LULC) and climatic conditions on the fluctuations of groundwater levels, particularly in terms of pre and post-monsoon depths within the Upper Subarnarekha River basin, India. Theil Sen’s Median trend in conjunction with Mann-Kendall (MK) test was also applied in this study. Results indicate that the southern region of the basin exhibits higher values for both precipitation and temperature. A considerable increase in built-up area with concurrent fluctuations in the groundwater level was observed in the upper SRB region. Our study explored the relationship between climatic parameters and groundwater levels in select urbanized areas of the region. The investigation revealed a strong link between rainfall and groundwater level. Mann-Kendall test indicated a non-significant upward trend in rainfall at a rate of 9.83 mm/year, while temperature showed an incessantly significant increasing trend. Inferences on variable temporal behaviour of the basin over the reference period of study further points to the need of monitoring the changes in hydrometeorological health of the basin. It can be critical to ensuring long-term water resources sustainability through appropriate planning and management in the rapidly changing geo-environment of river basins like SRB.

Food security and agriculture are essential for sustainable development, urbanization, and the well-being of population growth, while climate change in the Middle East and North Africa (MENA) region is causing erratic rainfall and rising temperatures, resulting in changes in agricultural land use. The purpose of this study is to examine land changes in the Fez region of Morocco over six years (2017-2022). The study utilized Sentinel-2 satellite multispectral images, spectral indices (NDVI, NBI, and NHFD), climate data and drought index to achieve the research objective. Using ground truth data, feature selection bands and spectral indices two machine learning algorithms: Random Forest (RF) and Gradient Tree Boost algorithms(GTB) were trained and tested via the Google Earth Engine (GEE) platform for the classification of LULC into four classes (Built-up Area, Water, Agricultural land and barren land), and ArcGIS Pro software was used for analysis. The overall performance of the algorithms was assessed, and the result shows that RF outperformed the GTB algorithm with 90% and 98 in the Kaffa coefficient and accuracy respectively. The final LULC analysis result shows that agricultural land use has fallen by 20% over the last five years, while builtup area has expanded by 10%, water decreased by 4% and barren land has increased by 6%. Correlation research between the change and the drought index revealed that climate change has an impact on agricultural land use in the region.

Despite the growing interest in seabed mapping in the context of sea level rise and storm intensification, only 10% of the global bathymetry has been sampled using reliable technologies, such as sonar or lidar, primarily due to the high cost of associated campaigns (waterborne and airborne). Consequently, satellite-derived bathymetry has thrived rapidly in recent decades, given its affordability for the user, and its ongoing gain in radiometric, spatial, spectral and temporal resolutions.

As a flagship of the very high spatial resolution sensor, the Pléiades Neo (PNEO) multispectral imagery, acquired by PNEO 3 and 4 sensors, leverages 6 bands : 1 purple (so-called deep blue), 3 visible, 1 red edge, and 1 infrared, provided with a spatial resolution of 1.2 m. This new sensor thus outperforms the Pléiades-1 multispectral imagery endowed with 4 bands (3 visible and 1 infrared) at 2 m pixel size.

The contribution of the novel bands of the PNEO to the bathymetry retrieval was innovatively quantified over an optically-challenging body of coastal seawater (0.2 m^-1 of vertical light attenuation). The importance of the level of the radiometric correction was tested based on the bathymetric lidar bathymetry predicted by a neural network (1 hidden layer and three neurons).

A PNEO 4 imagery was collected over the megatidal Bay of Saint-Malo (Brittany, France) on December 7, 2022. Following the orthorectification, the multispectral imagery was processed for the radiometric correction using the PNEO 4-specific spectral sensitivity, yielding five outputs: digital numbers (DN), top-of-atmosphere (TOA) radiance, TOA reflectance, bottom-of-atmosphere (BOA) maritime-modelled reflectance, and BOA tropospheric-modelled reflectance. The lidar response dataset, ranging from 0 to 20 m depth, was statistically stratified at the rate of 90 random samples per bathymetric slice of 1 m, every one divided into calibration, validation and test sub-samples.

The best predictions, reaching R²test of 0.81, were obtained for the full PNEO 4 dataset when uncorrected for the radiometry (namely, DN), corrected at both the TOA radiance and reflectance. For both BOA full-dataset products, the results were slightly less satisfactory: R²test of 0.75 (maritime) and 0.76 (tropospheric).

Taking the reference of the blue-green-red-infrared (simulating Pléiades-1 imagery), gains in R²test attained 0.05 for DN and TOA radiance datasets when the deep blue band replaced the blue one; 0.07 for maritime BOA reflectance when both deep blue and blue bands were integrated; and even 0.11 for that BOA reflectance when all PNEO 4 bands were used as predictors.

Fuel moisture content (FMC) is a crucial factor that influences fire behavior, rendering its precise estimation indispensable for effective fire risk assessment and management. However, despite the widespread availability of remotely sensed imagery, which offers valuable insights into live fuel moisture content (LFMC) estimation, it remains a significant challenge, especially given the dynamic nature of live forest fuels.

The aim of this study was to establish a robust method for estimating and monitoring LFMC by employing spatio-temporal modelling with a universal kriging approach, integrating remote sensing data and field measurements. This research was conducted in the Sierra Morena region of Andalusia, Spain, focusing on Cistus ladanifer shrub patches, well-known for their high fire risk. A total of 38 sampling plots were established to monitor LFMC over a 15-month period, with destructive sampling techniques used to determine LFMC in the laboratory.

The universal kriging model was enriched by incorporating variables derived from Sentinel-2 and MODIS products to estimate and validate the moisture content, resulting in an RMSE (Root Mean Squared Error) score of 12%. These findings have practical implications for forest fuel modeling, fire risk evaluation, and operational decision-making concerning fire prevention and management not only in the study area but also in potentially similar regions.

Antarctic coastal polynyas are important sea ice production regions and the process of sea ice production is accompanied by the formation dense water, which further affects the global circulation of ocean currents. Therefore, it is important to study the ice production of the Antarctic coastal polynyas. The study area of this paper is the Amery Ice Shelf region along the Antarctic coast because polynyas in Amery Ice Shelf region are major source areas of sea ice production in Antarctica. What’s more, there was a collapse event occurred on the Amery Ice Shelf from September 20 to 25 in 2019. The study period is 2013-2020 during winter time (from March to October). In this study, the thin ice thickness is retrieved based on AMSR2 monthly brightness temperatures (TBs) data by applying empirical formulas, and then two polynyas (Cape Darnley Polynya and Mackenzie Bay Polynya) are extracted in the Amery Ice Shelf region based on thin ice thickness. It is found that the area of the two polynyas fluctuates greatly, with maximal area frequently occurring in March. Then, sea ice production of the two polynyas is calculated in conjunction with ERA5 reanalysis meteorological data. The annual and intra-annual analyses of sea ice production in this study reveals that the inter-annual sea ice production of the two polynyas fluctuated considerably and reached a maximum in 2018, while the maximum intra-annual sea ice production of the two polynyas frequently occurred in March or April. As for the collapse event occurred on the Amery Ice Shelf in 2019, daily sea ice production of the two polynyas for a total of 40 days before and after the collapse event is calculated. Based on the result, it is assumed that the collapse event may have exacerbated the volatility and instability of sea ice production.