Water quality monitoring with Sentinel-2 and Landsat-8 satellites during the 2021 volcanic eruption in La Palma (Canary Islands)

In this study, seawater quality was monitored with high-resolution satellite imagery during the 2021 volcanic eruption (September-December) on La Palma Island (Spain), the longest recorded in the history of the island, and the most destructive in the last century in Europe. The Sentinel-2A/B twin satellites and Landsat-8 satellite were jointly used as an optical constellation, which allowed us to successfully characterize the short- and medium-term evolution of the new lava delta and subsequent impact on the seawater. Robust atmospheric and sunglint correction approaches were applied to thoroughly quantify the environmental changes caused on the adjacent coastal waters. The cloud and volcanic ash coverage remained very high over the coast during the event, so restricted information with 14 images (45% of the total scenes) was retrieved from the multi-sensor approach. Nevertheless, the availability of pre-, syn-, and post-eruption satellite products allowed us to map and detect the main water quality variations in the marine environment. On the one hand, during the eruption, a change in the properties of the water quality was observed, with a markedly increased turbidity on the western side of the island near the new lava delta due to the deposition of volcanic ash and material. On the other hand, chlorophyll-a concentration did not significantly increase, algal blooms were not observed, and oligotrophic conditions were not swiftly altered towards eutrophic conditions. This information offered an excellent opportunity to characterize the emplacement of the new lava delta and its impact on the marine environment in La Palma. The present multi-sensor strategy is an excellent opportunity to highlight the potential of remote sensing technology as a relevant and powerful tool for future hazard monitoring and assessment during catastrophes and for a better interpretation of their impact on the marine environment.

Monitoring cyanoHABs and water quality in Laguna Lake (Philippines) with Sentinel-2 satellites during the 2020 Pacific typhoon season

Laguna Lake, the largest freshwater lake in the Philippines, is permanently subject to nutrient-driven eutrophication and pollution and experiences harmful algal blooms (cyanoHABs) periodically with serious socio-economic implications. The aim of this study is to evaluate the suitability of the Sentinel-2 imagery of the European Commission's Copernicus Earth Observation programme for lake monitoring during the 2020 Pacific typhoon season (September-November 2020). The Case-2 Regional CoastColour processor is used to atmospherically correct Level 1 data and generate water quality parameters, such as chlorophyll-a (Chl-a) and total suspended matter (TSM) at 10 m. Results show that Super Typhoon Goni and Typhoon Vamco delivered high suspended sediment loads to the reservoir at concentrations above 170 g/m³ compared to pre-storm situations (0-35 g/m³). The typhoons also affect Chl-a, with a mean concentration of 10 mg/m³ and 30 mg/m³ for pre- and post-typhoons, respectively. In addition, the normalized difference chlorophyll index (NDCI) is used in the Google Earth Engine platform for near-real time monitoring of cyanoHABs at 20 m spatial resolution. Satellite maps are key for detecting the distribution of the blooms due to the patchiness of the green algae species, which usually form scum and elongated slicks in the lake. Maximum records of bloom detection during the study period occur in the Central Bay, one of the lake sections with major aquaculture and fisheries activities. The Sentinel-2 mission improves synoptic mapping of cyanoHABs and enables trends in their extent and severity to be documented. These available methods provide an essential tool for rapid detection after extreme events and for regular water quality monitoring, which will assist and benefit the cost-effective management of Laguna Lake.

Using Copernicus Sentinel-2 and Sentinel-3 data to monitor harmful algal blooms in Southern Chile during the COVID-19 lockdown

During the southern summer of 2020, large phytoplankton blooms were detected using satellite technology in Chile (western Patagonia), where intensive salmonid aquaculture is carried out. Some harvesting sites recorded massive fish mortalities, which were associated with the presence of the dinoflagellate species Cochlodinium sp. The bloom included other phytoplankton species, as Lepidodinium chlorophorum, which persistently changed the colour of the ocean to green. These blooms coincided with the government-managed emergency lockdown due to the COVID-19 pandemic. Local in situ sampling was slowed down. However, imagery from the Copernicus programme allowed operational monitoring. This study shows the benefits of both Sentinel-3 and Sentinel-2 satellites in terms of their spectral, spatial and temporal capabilities for improved algal bloom monitoring. These novel tools, which can foster optimal decision-making, are available for delivering early alerts in situations of natural catastrophes and blockages, such as those occurred during the global COVID-19 lockdown.

Deep Gaussian processes for biogeophysical parameter retrieval and model inversion

Parameter retrieval and model inversion are key problems in remote sensing and Earth observation. Currently, different approximations exist: a direct, yet costly, inversion of radiative transfer models (RTMs); the statistical inversion with in situ data that often results in problems with extrapolation outside the study area; and the most widely adopted hybrid modeling by which statistical models, mostly nonlinear and non-parametric machine learning algorithms, are applied to invert RTM simulations. We will focus on the latter. Among the different existing algorithms, in the last decade kernel based methods, and Gaussian Processes (GPs) in particular, have provided useful and informative solutions to such RTM inversion problems. This is in large part due to the confidence intervals they provide, and their predictive accuracy. However, RTMs are very complex, highly nonlinear, and typically hierarchical models, so that very often a single (shallow) GP model cannot capture complex feature relations for inversion. This motivates the use of deeper hierarchical architectures, while still preserving the desirable properties of GPs. This paper introduces the use of deep Gaussian Processes (DGPs) for bio-geo-physical model inversion. Unlike shallow GP models, DGPs account for complicated (modular, hierarchical) processes, provide an efficient solution that scales well to big datasets, and improve prediction accuracy over their single layer counterpart. In the experimental section, we provide empirical evidence of performance for the estimation of surface temperature and dew point temperature from infrared sounding data, as well as for the prediction of chlorophyll content, inorganic suspended matter, and coloured dissolved matter from multispectral data acquired by the Sentinel-3 OLCI sensor. The presented methodology allows for more expressive forms of GPs in big remote sensing model inversion problems.