Coastal and inland waters cover a wide range of optically complex conditions, from clear blue alpine lakes, to harmful coastal blooms, and sediment plumes. They are typically highly dynamic in both spatial and temporal dimensions. These regions are often where human activities benefit most from, and have the greatest impact on, aquatic environments. As such, there is great need for multi-sensor remote sensing approaches that can cope with this diversity of conditions and provide relevant measurements for marine management. Here we present a framework for establishing optical water types for the selection of water quality product algorithms for a global range of coastal and inland waters imaged by both the Sentinel-2 MSI and Sentinel-3 OLCI instruments. The framework is flexible, allowing for characterisation of the water types from both in situ and solely satellite derived radiometric spectra.

Optimisation of the number of optical water types (OWT) is performed using a parallelised fuzzy c-means classifier, allowing for individual spectra to be classified against multiple class means and supporting smoother blending of regional algorithms. The OWT classifier is applied to an extensive library of images from both the Sentinel-2 MSI and Sentinel-3 OLCI instruments. These images are selected such that they represent a diverse set of blooms, variable sediment conditions, and other forms of optical complexity (including glacial outflow and high CDOM). Each image is classified in turn using an iterative approach to continually add poorly scoring spectra to a global library, and then reassessing the OWT classification. Spectral scoring is performed using a combination of spectral angle and symmetric mean absolute percentage error estimation, with the image-wide scores used a metric to monitor the convergence of, and bring an end to, the iterative procedure. Suitable algorithms for the retrievals of required parameters (including turbidity, chlorophyll concentration, total suspended sediment) are subsequently selected and applied according to the spectral shape of each OWT.

This methodology has been applied to the assessment of water quality parameters across a number of projects, for which case study examples will be shown shown. Applications covered by these projects include measurement of coastal and inland water quality (H2020; EOMORES), turbidity estimation for UK national government monitoring programmes (DEFRA; ORSECT), and for aquaculture management and site selection (H2020; TAPAS).

The European Parliament and The Council of the European Union have been established the Water Framework Directive (2000/60/EC) for all European Union member states to achieve at least “good” ecological status of water bodies evenly in all European waters, which are larger than 50 hectares. MultiSpectral Instrument (MSI) onboard European Space Agency satellite Sentinel-2 has suitable 10, 20, 60 meters spatial resolution to monitor most of the Estonian lakes and coastal areas as required by the Water Framework Directive. Estonian lakes and coastal areas are optically complex waters, with high amount of CDOM and chlorophyll a. With both satellites, Sentinel-2A and Sentinel-2B, the temporal resolution 2-3 days gives the possibility to include and analyze more data for testing and developing new applications for Sentinel-2 MSI data. The aim of the study is to analyze the suitability of Sentinel-2 MSI data to monitor water quality in inland and coastal waters. As Sentinel-2 MSI is mainly designed for vegetation applications, it is important to find the most suitable atmospheric correction (AC) processor for water applications. Therefore, the first part of this study consists of testing various AC processors (ACOLITE, Polymer, C2RCC, Sen2Cor) and comparing the results with in situ radiometric measurements. Second part is comparing and developing chlorophyll a algorithms for Sentinel-2 MSI bands to estimate ecological status class of water by chlorophyll a in Estonian lakes. The study showed, that the spectral resolution of MSI is suitable for estimating chlorophyll a in water bodies, but the problems can be caused by low SNR ratio and AC, which often failed in narrow and small lakes possibly due to adjacency effect. Due to that, deriving satellite-based chlorophyll a from empirical band ratio algorithms and estimating ecological status of water is not possible in every case. The comparison of different approaches to derive chlorophyll a indicated, that the algorithms applicable on L1 data are preferred, as the AC over these complex cases are often prone to errors. Initial results show that Sentinel-2 MSI MultiSpectral Instrument could provide complementary information to in situ data to support Water Framework Directive monitoring requirements.

The aim of the CoastVal project is to develop and deploy a buoy platform dedicated to validation of satellite ocean colour data products for coastal areas in high latitude waters. This ESA-funded project is part of the Sentinel 2 and 3 Validation Team activities, contributing to the overall calibration and validation activities for both missions. The project has also been discussed and presented to the Fiducial Reference Measurements for Satellite Ocean Colour (FRM4SOC) project, which aims to determine accurate radiometric calibration and characterization of individual satellite sensors, as the most critical component toward achieving the goal of consistent, long-term multi-mission Ocean Colour products.

A buoy-based platform dedicated to validation of satellite ocean colour products for coastal areas in high latitude waters is presented. The platform consists of a suite of 5 hyperspectral radiometers, including an in-air downwelling irradiance sensor and a pair of upwelling radiance and downwelling irradiance sensors positioned at two depths with fixed separation. The platform also includes ancillary sensors such as chlorophyll, turbidity as well as temperature, conductivity, dissolved oxygen and pH to characterise the water column.

The platform is also supplied with a tilt and heading sensor to record buoy attitude during data acquisition and to correct for variations with respect to solar zenith angle and azimuth. The tilt sensor measures pitch and roll at high frequency compared to the hyperspectral sensors so that radiometric data acquired outside a certain tilt range can be removed from the inversion dataset. The radiometric sensors, along with ancillary measurements such as depth, water properties, and pitch and roll of the buoy platform, provided much larger volumes of data than is generally collected on an autonomous basis, and required significant software development and data gathering capabilities during this project. Significant work has also been performed to experimentally measure and account for a range of effects, from tidal currents to buoy and instrument self-shading.

The platform has been purposely designed to minimise the uncertainty in the radiance observations while also providing a robust system to withstand high currents and rough sea states. The project is ongoing until Q3 2018, with a successful test deployment of the buoy and sensors in December 2017. The final deployment of the system at the mouth of Dublin Bay took place in August 2018, with all instruments successfully streaming data and the processing and quality control well underway. The deployment of the system and the analysis required to produce radiometric products for comparisons with the satellite observations will be presented.

Passive optical approaches to detection of algal blooms and other water quality issues rely on the fundamental physically-based principle that the spectral reflectance signal from algal dominated inland waters contains information which can be related to quantitative descriptions of algal amount, such as chlorophyll and the detection of accessory pigments (e.g. phycocyanin). We developed a spectrometer which could reliably measure spectral reflectance across the visible spectrum from inland water bodies at a relatively low cost. We have evaluated the spectral performance of a prototype sensor in ~15 Australian waterbodies; the low-cost device measures spectral reflectances comparable to those made using a more expensive scientific grade spectroradiometer. We have further demonstrated continuous monitoring using a deployable version of the sensor in remote locations with wireless data upload. Preliminary results show the capacity to independently measure TSS, CDOM, and chlorophyll, with accuracies similar to that of much more expensive research instruments.

Permanently mounted at key sites, the device allows for the remote detection of the occurrence of surface cyanobacterial blooms and other water quality phenomena to form part of a low maintenance sensor network across inland water systems. Preliminary analyses of the reflectance/time series, and automated photographs of the water surface, suggests that the system can detect blooms as they pass across the instrument‘s field of view, as well as measuring changing surface Chl-a concentrations due to cyanobacteria floating to the surface. Effective detection of such events is critical to the development of an early-warning system. The system provides high density match up data for validation of satellite observations, as well as data for parametrisation of hydrodynamic/growth models for forecasting algal bloom occurrence.

The presentation will outline the development and validation of the low-cost sensor concept and its deployment in applications for algal bloom alerting. We will illustrate the rich temporal datasets that result and the development of data processing workflows and dashboards.

A wide range of environmental issues are increasingly challenging for global governments, societies and industries. One of the most important one is the protection of freshwater resources, which are limited and face increasing pressures in an era of climate changes. The actual agenda of governments and areas of conflicts demonstrate a paradigm change and a new era: Governmental and industrial investments on information and monitoring programs increasingly reflect the huge economic and social impact of the changing environment. Increasingly, expert knowledge and well-resourced information is used for adequate decision-making. Access to country-wide or transboundary status information and trends of water quality in inland and coastal waters is essential for environmental assessments of drinking reservoirs, impact monitoring of dams or high-level water-resource-management of governmental bodies: Such information can support efficient planning on localized comprehensive measurements and serve as a continuous monitoring instrument.
Consistent and independent water quality information can be derived extremely cost-efficient and with a minimum of logistical effort for millions of inland and coastal waterbodies, using measurements from satellite-based earth observation (EO) sensors. State-of-the-art, globally harmonized information products in conjunction with a very user friendly web access support the need of data efficiently. Physics-based retrieval algorithms as the Modular Inversion and Processing system MIP developed by EOMAP operationally account and correct for a variety of environmental impacts, such as atmospheric conditions, adjacent land cover, water surface and the composition of water constituents as well as considering varying observation geometries and sensor properties.
To address the demand from the stakeholders, EOMAP has developed a fully operational multi-resolution processing system providing users with easy online-access to water quality information for water quality monitoring and for responding to water quality threats for any location worldwide. The system has been implemented and validated in two different showcases.
Within the H2020 funded SPACE-O project, EOMAP applies this technology for two reservoirs, Aposelemis on Crete and Mulargia in Italy, monitoring the main water quality parameters like harmful algae blooms, turbidity, chlorophyll-a concentrations and water temperature. These parameters have weekly or almost daily sampling frequencies under cloud-free conditions. The necessary steps towards operational data processing and delivery of EO based water quality data during the SPACE-O project comprised automated routines for satellite archive access, the operational water quality product processing chain, updated with new developed algorithms and outcome products as well as the automated delivery mechanisms using GeoServer for the development of the SPACE-O Environmental Information Platform. During the two-year-project, EOMAP together with CNR has further improved and developed water quality parameters for inland basins including the trophic state index, the evaporation rate, a better bathymetric model and the separation of inorganic and organic matter.
Spatial resolutions from 30m to up to 2m pixel size allow monitoring of very small water bodies and ponds, and with a temporal frequency of almost daily measurements of the water quality parameters. EOMAP uses for this project its multi-sensor processing algorithm MIP with harmonized water quality data based on Copernicus missions of Sentinel-2A/B combined with additional data from Landsat 7/8, WorldView-2 or PlanetScope satellite data. Validation exercises with in situ data sets from 2013-2018 have proven the consistency for quantifying temporal and spatial patters of the water quality parameters in the water reservoirs.
The same harmonized approach was applied on country-wide level in Georgia within a commercial contract for Stucky Ltd.: Satellite based turbidity and related suspended matter concentrations and their seasonal changes for the major rivers in Georgia were derived, providing data for planning of hydropower and water constructions. The analysis was performed independently from in-situ data for the rivers Adzhariskali, Mtkvari, Rioni and Enguri. Background of this study for the Ministry of Energy in Georgia is the exploration of hydropower energy in Georgia. The major drivers of the reservoir storage capacity over time, reservoir life time (and related re-investment time), and also a major driver of the operations costs are directly related to the sediment regime and sediment trapping. These are defined by the external environmental boundary conditions. However, also here the typical situation at planning stage was the absence of such data, especially as extended areas and river systems had to be investigated.
To improve water quality monitoring at the global level and support the SDGs implementation and monitoring, UNESCO-IHP International Initiative on Water Quality (IIWQ) developed the UNESCO World Water Quality Portal—a pioneering tool to monitor freshwater quality using satellite data. The portal provides data on five key water quality indicators: turbidity and sedimentation distribution, chlorophyll-a, Harmful Algal Blooms, organic absorption and surface temperature. These indicators provide information on the impact of other sectors and land uses such as urban areas, fertilizer use in agriculture, climate change or dam and reservoir management. Hence, it helps better understand the impact of climate and anthropogenic changes on water resources. It also facilitates science-based, informed decision-making for water management and supports SDGs implementation and monitoring.
Environmental managers, politicians and scientists gain a global perspective on water and the environment through the portal. They can discover how earth observation can help monitor the effects of pollution on our water cycle and raise awareness of the importance of environmental monitoring to policy and decision makers, diplomats and the general public.
Current activities concentrate on improvements of reporting tools and the integration into directive monitoring or monitoring of Sustainable Developments Goals, in close cooperation with the responsible stakeholders like, UNEP, UFZ or state water authorities like LUBW in Germany.

Eutrophication is a well-recognized problem in many inland and coastal waters globally. Human activities have accelerated the rate and extent of eutrophication through both point-source discharges and non-point loadings of limiting nutrients, such as nitrogen and phosphorus, into aquatic ecosystems, with dramatic consequences for drinking water sources, fisheries, and recreational water bodies. The known consequences include blooms of cyanobacteria, deteriorated drinking water supplies, degradation of recreational opportunities, loss of biodiversity and hypoxia, and the related economic consequences are considerable. 2018 was a year with an exceptional warm and sunny year in Europe supporting intense cyanobacteria growth in many lakes and in the Baltic Sea.
CyanoAlert is a global service for environmental authorities and the commercial sector, concerned by health risks and quality of water resources, and for the public. CyanoAlert makes use of the wealth of data provided by the Copernicus space component, especially Sentinel-3 and Sentinel-2 data, to deliver a fully automated application for assessing potentially toxin-producing cyanobacteria blooms in water resources globally. The on-going service developments are funded by the H2020 Innovation Action and will continue until February 2020. After this period, a transition to customer funded activities will be made, and a flexible service with user-specific information for monitoring and reporting purposes and an open information service for the general public are foreseen.
The project consortium consists of six partners, three EO development and implementation partners (IPs) and three User partners (UPs), also potential customers of the service after the project period. The IPs are Brockmann Geomatics Sweden AB (Sweden), Brockmann Consult GmbH (Germany) and Odermatt & Brockmann GmbH (Switzerland). The UPs are one public government authority, The Information Office for the Baltic Sea (Sweden), and two research institutes linked to routine monitoring programs or laboratory analysis, Istituto Superiore di Sanità, Italy, and Danube Delta National Institute for Research and Development, Romania.
An overview of the project efforts in terms of user requirements, service design, hardware-, processing- and storage solutions, in situ data collection, validation results, as well as service platform and dissemination system developments will be presented together with the foreseen service offer. In addition, launched pilot service demonstrations will be introduced, which includes a mobile application with focus on crowdsourcing and cyanobacteria bloom Near Real Time warnings.