Search

Pumice rafts are considered to be a long-range drifting agent that promotes material exchange and the dispersal of marine species. Large ones can also interfere with vessel navigation and have a negative impact on the social economy and marine ecosystems. Synoptic observations from the Multispectral Instrument (MSI) on-board Sentinel-2, with a spatial resolution of up to 10 m, provide an excellent means to monitor and track pumice rafts. In this study, the use of a Spectral-Feature-Based Extraction (SFBE) algorithm to automatically discriminate and extract pumice on the ocean surface from submarine volcano eruptions was proposed. Specifically, a Pumice Raft Index (PRI) was developed based on the spectral signatures of pumice in MSI imagery to identify potential pumice features. After pre-processing, the PRI image was then subjected to a series of per-pixel and object-based processes to rule out false-positive detections, including shallow water, striped edges, mudflats, and cloud edges. The SFBE algorithm showed excellent performance in extracting pumice rafts and was successfully applied to extract pumice rafts near the Fiji Yasawa islands in 2019 and Hunga Tonga island in 2022, with an overall pumice extraction accuracy of 95.5% and a proportion of pixels mis-extracted as pumice of <3%. The robustness of the algorithm has also been tested and proved through applying it to data and comparing its output to results from previous studies. The timely and accurate detection of pumice using the algorithm proposed here is expected to provide important information to aid in response actions and ecological assessments, and will lead to a better understanding of the fate of pumice. [ABSTRACT FROM AUTHOR]

Large volumes of crude oil accidentally released into the sea may cause irreversible adverse impacts on marine and coastal environments. Large swath optical imagery, acquired using platforms such as the moderate-resolution imaging spectroradiometer (MODIS), is frequently used for massive oil spill detection, attributing to its large coverage and short global revisit, providing rich data for oil spill monitoring. The aim of this study was to develop a suitable approach for massive oil spill detection in sun glint optical imagery. Specifically, preprocessing procedures were conducted to mitigate the inhomogeneous light field over the spilled area caused by sun glint, enhance the target boundary contrast, and maintain the internal homogeneity within the target. The image was then segmented into super-pixels based on a simple linear clustering method with similar characteristics of color, brightness, and texture. The neighborhood super-pixels were merged into target objects through the region adjacency graph method based on the Euclidean distance of their colors with an adaptive termination threshold. Oil slicks from the generated bright/dark objects were discriminated through a decision tree with parameters based on spectral and spatial characteristics. The proposed approach was applied to oil spill detection in MODIS images acquired during the Montara oil spill in 2009, with an overall extraction precision of 0.8, recall of 0.838, and F1-score of 0.818. Such an approach is expected to provide timely and accurate oil spill detection for disaster emergency response and ecological impact assessment. [ABSTRACT FROM AUTHOR]

This study evaluates the performance of three typical convolutional neural network based deep learning algorithms for oil spill detection using medium-resolution optical satellite imagery from Sentinel-2 MSI, Landsat-8 OLI, and Landsat-9 OLI2. Oil slick training and validation dataset were created through a semi-automatic labeling approach, based on chronic and accidental oil spill cases reported worldwide. The research enhances UNet, BiSeNetV2, and DeepLabV3+ architectures by integrating attention mechanisms including the Squeeze-and-Excitation module (SE), Convolutional Block Attention Module (CBAM), and a Simple, parameter-free Attention Module (SimAM), analyzing the optimal model for oil spill detection. Notably, UNet integrated with CBAM, especially with sun glint as a feature, significantly outperformed others, achieving a micro-average F1 score of 88.8 %. This research highlights deep learning's potential in optical remote sensing for oil spill detection, stressing its escalating relevance with the growing deployment of medium- to high-resolution optical satellites. • Confirmed cases worldwide evaluate CNNs performance in optical oil spill detection • UNet with CBAM attention module shows strong robustness in detecting oil spills • Incorporating sun glint as a feature input reduces false positive detections [ABSTRACT FROM AUTHOR]

Optical remote sensing is one of the most commonly used techniques to detect oil in the surface ocean. This is because oil has optical properties that are different from water to modulate oil–water spatial and spectral contrasts. However, understanding these contrasts is challenging because of variable results from laboratory and field experiments as well as from different observing conditions and spatial/spectral resolutions of remote sensing imagery. Here, through reviewing published oil–water spectral contrasts and analyzing remotely sensed spectra collected by several satellite and airborne sensors (MERIS, MODIS, MISR, Landsat, and AVIRIS) from the Deepwater Horizon oil spill, we provide the interpretation of the spatial/spectral contrasts of various oil slicks and discuss the challenges in such interpretations. In addition to oil thickness, several other factors also affect oil–water spatial/spectral contrasts, including sun glint strength, oil emulsification state, optical properties of oil covered water, and spatial/spectral resolutions of remote sensing imagery. In the absence of high spatial- and spectral-resolution imagery, a multistep scheme may be used to classify oil type (emulsion and non-emulsion) and to estimate relative oil thickness for each type based on the known optical properties of oil, yet such a scheme requires further research to improve and validate. [ABSTRACT FROM AUTHOR]

Abstract An oil platform in the Mississippi Canyon 20 (MC-20) site was damaged by Hurricane Ivan in September 2004. In this study, we use medium- to high-resolution (10–30 m) optical remote sensing imagery to systematically assess oil spills near this site for the period between 2004 and 2016. Image analysis detects no surface oil in 2004, but ~40% of the cloud-free images in 2005 show oil slicks, and this number increases to ~70% in 2006–2011, and >80% since 2012. For all cloud-free images from 2005 through 2016 (including those without oil slicks), delineated oil slicks show an average oil coverage of 14.9 km2/image, with an estimated oil discharge rate of 48 to ~1700 barrels/day, and a cumulative oil-contaminated area of 1900 km2 around the MC-20 site. Additional analysis suggests that the detected oil slick distribution can be largely explained by surface currents, winds, and density fronts. Highlights • Over 80% of the cloud-free images show oil slicks near the MC-20 site in 2012–2016. • Average oil coverage area is 14.9 km2 for all cloud-free images in 2005–2016. • A cumulative oil-contaminated footprint of 1900 km2 around the site is detected. • Average oil discharge rate is estimated to range from 48 to ~1700 barrels/day. • River plume dynamics largely influences the variability in oil slick distributions. [ABSTRACT FROM AUTHOR]

The critical angle at which the brightness of oil slicks and oil-free seawater is reversed occurs under sunglint and is often shown as an area of uncertainty due to different roughness and surface Fresnel reflection parameters. Consequently, differentiating oil slicks from the seawater in these areas using optical sensors is a challenge. Polarized optical remote sensing techniques provide complementary information for intensity imagery with different physical properties and, thus, possess the ability to resolve this difficult problem. In the polarized reflectance model, the degree of linear polarization (DOLP) of sunglint depends on accurately knowing the Stokes parameter for the reflected light, and varies with the refractive index of the surface layer and viewing angles. For the polarized detection of oil slicks, the highest sensitivity of the DOLP to the refractive index is located within the specular reflection direction where the sum of the solar and sensor zenith angles is 82.6°. The modeled results clearly indicate that the DOLP of oil slicks is weaker in comparison with oil-free seawater under sunglint. Using measurements from the space-borne Polarization and Anisotropy of Reflectances for Atmospheric Sciences coupled with Observations from a Lidar (PARASOL) over the Deepwater Horizon oil spill in the Gulf of Mexico, we illustrate that the PARASOL-derived DOLP difference between the oil spill and seawater is obvious and is in accordance with the modeled results. These preliminary results suggest that the potential of multiangle measurement and feasibility of deriving refractive index of ocean surface from space-borne sensors need further researches. [ABSTRACT FROM AUTHOR]

Different oil spill pollution types could be produced in oil transport and weathering processes. Investigation of these pollution types is beneficial for oil spill recovery and processing. Optical remote sensing techniques play an important role in marine oil spill monitoring and have the ability to identify different oil spill pollution types. Recently, research on oil spill optical remote sensing has made much progress in detecting targets, identifying spectral response characteristics, and formulating theories. Floating black oil, oil slicks, and oil-water mixture in marine oil spill accidents are the main targets to be investigated by optical remote sensors. The visible spectral response differences of these targets are the base of oil spill optical remote sensing research. Bi-directional reflectance distribution function, light interference, absorption, and scattering of targets produce different spectra. Therefore, oil spill optical remote sensing could be used to identify the main oil spill pollution types and estimate oil spill volume. [ABSTRACT FROM AUTHOR]

Thermal remote sensing has been used in assessing oil spills in the ocean, mostly based on empirical interpretations. This study designs a ground-based experiment to measure brightness temperatures (BTs) of oil-in-water (OW) emulsions with different concentrations and oil-free water as a function of time in 32 consecutive hours. Compared with a previous thermal experiment to measure oil slicks (i.e., non-emulsified oil) with different thicknesses and considering the similarity between oil slicks and water-in-oil (WO) emulsions, it is found that (1) the diurnal response of brightness temperature difference (BTD, between oil samples and oil-free sample) to oil emulsion types (OW or WO) is similar, making it difficult to classify oil emulsion types using thermal remote sensing; (2) in contrast, BTD under thermal balance during the optimal time window appears to be a function of equivalent oil thickness (EOT) (oil volume per area, mm) regardless of oil emulsion type, suggesting that EOT could be estimated from BTD. Application of such experimental results to Landsat imagery over the Deepwater Horizon oil spill in the Gulf of Mexico suggests that although their ability to quantify oil footprint is limited, thermal data show potentials in providing unique information (e.g., estimating EOTs for up to 4 mm without the need of differentiating sub-pixel heterogeneity) to complement optical data in characterizing oil type and oil quantity when the spilled oil is thick (> 0.4 mm). • Laboratory thermal experiment conducted to characterize oil-in-water emulsions. • Oil concentrations in OW emulsions are related to brightness temperature (BT). • BT not useful to detect oil type, but useful to quantify equivalent oil thickness. • With known uncertainties, BT complements other techniques to assess oil spills. [ABSTRACT FROM AUTHOR]

Water in oil (WO) and oil in water (OW) emulsions from marine oil spills have different physical properties, volume concentrations, and spectral characteristics. Identification and quantification of these different types of oil emulsions are important for oil spill response and post-spill assessment. While the spectral characteristics of WO and OW emulsions have been presented in previous studies including Part I of this series, their application to airborne and satellite imagery is further demonstrated here. Using AVIRIS and Landsat observations, we firstly show that false color Red-Green-Blue composite images from Landsat-like sensors (R: 1677 nm, G: 839 nm, B: 660 nm) are effective in differentiating WO and OW emulsions as they show reddish and greenish colors, respectively, in such composite images. This is a consequence of the relative difference in the reflectance of WO and OW emulsions at 1677 and 839 nm, which is not impacted by the presence of medium-strength sunglint or the surface heterogeneity within medium-resolution pixels (e.g., 30 m). Based on image statistics, a decision tree method is proposed to classify oil type, and oil quantification is further attempted, with results partially validated through spectral analysis and spatial coherence test. The numerical mixing experiments using AVIRIS pixels further indicate that the SWIR bands might be used to develop linear unmixing models in the future once the coarse-resolution oiled pixels are first classified to WO and OW types, and 1295 nm is the optimal wavelength to perform spectral unmixing of mixed coarse-resolution pixels. • Principles to differentiate WO and OW emulsions applied to multi-band imagery • False color RGB composite images from multi-band sensors effective for this purpose • A decision-tree method proposed to classify emulsion types and estimate thickness • Linear mixing effects found from multi-band coarse-resolution pixels. • SWIR band at 1295 nm being optimal for unmixing coarse-resolution pixels [ABSTRACT FROM AUTHOR]

Optical identification and quantification of various marine-spilled oils play an important role in oil spill monitoring, assessment, and response. Through weathering processes, oil may become emulsified in two forms of oil-water mixture: water in oil (WO) and oil in water (OW). These two forms of oil emulsion are significantly different in their volume concentration (oil/water ratio), physical properties (viscosity, density, thickness), and optical properties (spectral reflectance (R u (λ), sr−1), and spectral absorption (a (λ), m−1)). In this study, the optical properties of both types of oil emulsion, with different volumetric concentrations, are determined from carefully prepared oil emulsion samples, with the aim of helping to interpret optical remote sensing imagery. The concentrations of stable WO and OW emulsions range from 45% to 95% and from 0.025% to 3%, respectively. They exhibit different R u spectral shapes in the near-infrared and shortwave-infrared wavelengths, with five "-CH" molecular bonds evident in the WO emulsion spectra. R u (600–1400 nm) of OW emulsions increases with volume concentrations from 0% to 3.0%, but R u (600–2400 nm) of the WO emulsions decreases with volume concentrations from 45% to 100%. On the other hand, for a fixed concentration (80%), R u (600–2400 nm) of WO emulsions increases monotonically with thicknesses of up to ~0.4 mm, beyond which R u (600–2400 nm) no longer increases with oil thickness. The difference between the R u spectral shapes of OW and WO emulsions, as well as the statistical relationships between volume concentrations and R u (NIR-SWIR) and between oil thickness and R u (NIR-SWIR), provide the basis for developing optical models to classify oil emulsion types and for quantifying oil volume from remote sensing imagery. The potential of such an application is demonstrated using hyperspectral AVIRIS imagery collected over the Deepwater Horizon (DWH) oil spill in the Gulf of Mexico (GoM). • Oil emulsions are found to form two types: oil in water (OW) and water in oil (WO). • Spectral reflectance and absorption of oil emulsions are determined from experiments. • OW and WO emulsions with various concentrations have different spectral features. • For fixed oil type, oil thickness of up to 0.4 mm can be estimated from reflectance. • The proof-of-concept is demonstrated with AVIRIS data over the DWH oil spill. [ABSTRACT FROM AUTHOR]