Search

Phytoplankton blooms threaten aquatic ecosystems worldwide, with implications going beyond their apparent ecological aspects. Management solutions are needed to control the appearance of phytoplankton blooms and alleviate their impacts. Such solutions are supported by scientific results, many of which derive from modeling approaches.Data-driven models are now routinely deployed for the short-term (day to weeks) forecasting of phytoplankton dynamics. Nonetheless, such data-oriented efforts are often plagued by two issues, i.e., the lack of sufficient data and interpretability. On one hand, insufficient data result in overfitting, which produces poorly generalizable models that are unreliable under extrapolating conditions. On the other hand, the lack of interpretability hinders the contribution of such models in decision-making, since acting upon model predictions relies heavily on understanding of the model hypothesis.These two challenges motivated the present work, which aspired to investigate the suitability of multi-spectral satellite imagery as a source of phytoplankton-related data for the development of credible and accountable data-driven models. To this end, first, satellite-derived chlorophyll-a times series were created using Sentinel-2 and Landsat 8 imagery and a physics-based modular inversion and processing system. Then, two machine learning algorithms, i.e. a Random Forest (RF) and a Gaussian Process (GP) regression algorithm, were trained to map hydrometeorological drivers to the satellite-derived chlorophyll-a time series.The two algorithms were benchmarked against each other and against a naïve alternative, i.e., the persistence method, in terms of accuracy, uncertainty, and interpretability in three cases: (a) the mesotrophic Mulargia reservoir in Italy, (b) the hypereutrophic Harsha Lake in the USA, and (c) Lake Hume in Australia, a reservoir facing an increasing number of algal bloom events over the last 10 years.Results indicate that both machine learning models forecasted surface phytoplankton dynamics more accurately compared to their naïve alternative up to ten days ahead in the future. It should be noted though that forecasting accuracy deteriorated with increasing forecasting windows, mostly due to the uncertainty of meteorological forecasts.When the machine learning methods were compared to each other, the RF-based models were marginally better compared to their GP counterparts; they produced slightly more accurate and more certain chlorophyll-a predictions. RF-based models are also preferable in terms of interpretability. Their predictions unveiled specific patterns in hydrometeorological data that could explain phytoplankton dynamics in each case. On the contrary, it remained obscure how chlorophyll-a predictions were made by the GP regression models.More importantly this work offers evidence supporting that multi-spectral satellite data allow for the development of theory-guided, data-driven models for the forecasting of phytoplankton dynamics in lakes and reservoirs.

Traditional data sources are not sufficient for measuring the United Nations Sustainable Development Goals. New and non-traditional sources of data are required. Citizen science is an emerging example of a non-traditional data source that is already making a contribution. In this Perspective, we present a roadmap that outlines how citizen science can be integrated into the formal Sustainable Development Goals reporting mechanisms. Success will require leadership from the United Nations, innovation from National Statistical Offices and focus from the citizen-science community to identify the indicators for which citizen science can make a real contribution.

Underwater image restoration attracts significant attention due to its importance in unveiling the underwater world. This paper elaborates on a novel method that achieves state-of-the-art results for underwater image restoration based on the unsupervised image-to-image translation framework. We design our method by leveraging from contrastive learning and generative adversarial networks to maximize mutual information between raw and restored images. Additionally, we release a large-scale real underwater image dataset to support both paired and unpaired training modules. Extensive experiments with comparisons to recent approaches further demonstrate the superiority of our proposed method., Accepted to IGARSS 2021 as oral presentation. Code is available at https://github.com/JunlinHan/CWR

Underwater image restoration is of significant importance in unveiling the underwater world. Numerous techniques and algorithms have been developed in the past decades. However, due to fundamental difficulties associated with imaging/sensing, lighting, and refractive geometric distortions, in capturing clear underwater images, no comprehensive evaluations have been conducted of underwater image restoration. To address this gap, we have constructed a large-scale real underwater image dataset, dubbed `HICRD' (Heron Island Coral Reef Dataset), for the purpose of benchmarking existing methods and supporting the development of new deep-learning based methods. We employ accurate water parameter (diffuse attenuation coefficient) in generating reference images. There are 2000 reference restored images and 6003 original underwater images in the unpaired training set. Further, we present a novel method for underwater image restoration based on unsupervised image-to-image translation framework. Our proposed method leveraged contrastive learning and generative adversarial networks to maximize the mutual information between raw and restored images. Extensive experiments with comparisons to recent approaches further demonstrate the superiority of our proposed method. Our code and dataset are publicly available at GitHub., Comment: In submission, code/dataset are at https://github.com/JunlinHan/CWR. arXiv admin note: text overlap with arXiv:2103.09697

Baseline determination and operational continental scale monitoring of water quality are required for reporting on marine and inland water progress to Sustainable Development Goals (SDG). This study aims to improve our knowledge of the optical complexity of Australian waters. A workflow was developed to cluster the modelled spectral response of a range of in situ bio-optical observations collected in Australian coastal and continental waters into distinct optical water types (OWTs). Following clustering and merging, most of the modelled spectra and modelled specific inherent optical properties (SIOP) sets were clustered in 11 OWTs, ranging from clear blue coastal waters to very turbid inland lakes. The resulting OWTs were used to classify Sentinel-2 MSI surface reflectance observations extracted over relatively permanent water bodies in three drainage regions in Eastern Australia. The satellite data classification demonstrated clear limnological and seasonal differences in water types within and between the drainage divisions congruent with general limnological, topographical, and climatological factors. Locations of unclassified observations can be used to inform where in situ bio-optical data acquisition may be targeted to capture a more comprehensive characterization of all Australian waters. This can contribute to global initiatives like the SDGs and increases the diversity of natural water in global databases.

Water managers need tools to assist in the management of ever increasing algal bloom problems over wide spatial areas to complement sparse and declining in situ monitoring networks. Optical methods employing satellite data offer rapid and widespread coverage for early detection of bloom events. The advent of the Analysis Ready Data (ARD) and Open Data Cube concepts offer the means to lower the technical challenges confronting managers, allowing them to adopt satellite tools. Exploiting Landsat ARD integrated into the Digital Earth Australia data cube, we developed a prototype algal bloom alerting tool for the state of New South Wales, Australia. A visualization portal allows managers to gain insights into bloom status across the state as a whole and to further investigate spatial patterns in bloom alerts at an individual water body basis. To complement this we also proposed an algal alert system for trial based on chlorophyll and TSM levels which requires further testing. The system was able to retrieve the status of 444 water bodies across the state and outputs from the visualization system are presented. Time series of image acquisitions during an intense bloom in one lake are used to demonstrate the potential of the system. We discuss the implications for further development and operationalisation including the potential for augmentation with alternative algorithms and incorporation of other sensor ARD data to improve both temporal and spectral resolutions.

Remote sensing observations show optical signatures to conform to the physical oceanographic patterns in the Tasman Sea. To test the link between physical oceanographic processes and bio-optical properties we investigated an in situ bio-optical dataset collected in the Tasman Sea. Analysis of in situ observations showed the presence of four different water masses in the Tasman Sea, formed by the relatively warm and saline East Australia Current (EAC) water, a mesoscale cold core eddy on the continental slope, cooler Tasman Sea water on the shelf and river plume water. The distribution of suspended substances and their inherent optical properties in these water masses were distinctly different. Light absorption and attenuation budgets indicate varying optical complexity between the water masses. Specific inherent optical properties of suspended particulate and dissolved substances in each group were different as they were influenced by physical and biogeochemical processes specific to that water mass. Remote sensing reflectance signature varied in response to changing bio-optical properties between the water masses; thus providing the link between physical oceanographic processes, bio-optical properties and the optical signature. Findings presented here extend our knowledge of the Tasman Sea, its optical environment and the role of physical oceanographic processes in influencing the inherent optical properties and remote sensing signature in this complex oceanographic region.

Satellite imagery is increasingly being used to provide estimates of bathymetry in near-coastal (shallow) areas of the planet, as a more cost-effective alternative to traditional methods. In this paper, the relative accuracy of radiative-transfer and photogrammetric stereo methods applied to World View 2 imagery are examined, using LiDAR bathymetry and towed video as ground truth, and it is demonstrated, with a case study, that these methods are complementary; where one method might have limited accuracy, the other method often has improved accuracy. The depths of uniform, highly-reflective (sand) sea bed are better estimated with a radiative transfer-based method, while areas where there is high visual contrast in the scene, as identified by using a local standard deviation measure, are better estimated using the photogrammetric technique. In this paper, it is shown that a hybrid method can give a potential improvement in accuracy of more than 50% (from 2.84 m to 1.38 m RMSE in the ideal case) compared to either of the two methods alone. Metrics are developed that can be used to characterize regions of the scene where each technique is superior, realizing an improved overall depth accuracy over either method alone of between 16.9% and 19.7% (demonstrating a realised RMSE of 2.36 m).

This chapter addresses the demand of environmental agencies and water industries for tools enabling them to prevent and mitigate events of rapid deterioration of environmental assets such as contamination of air, soils and water, declining biodiversity, desertification of landscapes. Getting access to reliable early warning signals may avoid excessive ecological and economic costs.

This study focuses on the seasonal and spatial characterization of inherent optical properties and biogeochemical concentrations in the Van Diemen Gulf and Darwin Harbour, two neighboring tropical coastal environments of Northern Australia that exhibit shallow depths (~20 m), large (>3 m) semi-diurnal tides, and a monsoonal climate. To gain insight in the functioning of these optically complex coastal ecosystems, a total of 23 physical, biogeochemical, and optical parameters were sampled at 63 stations during three field campaigns covering the 2012 wet and dry seasons, and the 2013 dry season. The total light absorption budget in the Van Diemen Gulf was dominated by non-algal particles (aNAP; >45%) during the dry season (May–October) and colored dissolved organic matter (aCDOM; 60%) during the wet season (November–April). The combined absorption by aNAP and aCDOM generally exceeded ~80% of the total absorption budget from 400 to 620 nm, with phytoplankton, aPhy, accounting for

In order to retrieve bathymetry, substratum type and the concentrations of the optically active constituents of the water column, an integrated physics based mapping approach was applied to airborne hyperspectral data of Moreton Bay, Australia. The remotely sensed data were sub-optimal due to high and mid-level cloud covers. Critical to the correct interpretation of the resultant coastal bathymetry map was the development of a quality control procedure based on additional outputs of the integrated physics based mapping approach and the characteristics of the instrument. These two outputs were: an optical closure term which defines differences between the image and model based remote sensing signal; and an estimate of the relative contribution of the substratum signal to the remote sensing signal. This quality control procedure was able to identify those pixels with a reliable retrieval of depth and to detect thin and thick clouds and their shadows, which were subsequently masked out from further analysis. The derived coastal bathymetry in depths ranging 4–13 m for the mapped area was within ± 15% of boat-based multi-beam acoustic mapping survey of the same area. The agreement between the imaging spectrometry and the acoustic datasets varies as a function of the contribution of the bottom visibility to the remote sensing signal. As expected, there was greater agreement in shallower clear water (± 0.67 m) than quasi-optically deep water (± 1.35 m). The quantitative identification and screening of the optically deep waters and the quasi-optically deep waters led to improved precision in the depth retrieval. These results suggest that the physics based mapping approach adopted in this study performs well for retrieving water column depths in coastal waters in water depths ranging 4–13 m for the area and conditions studied, even with sub-optimal imagery. Crown copyright © 2008 Published by Elsevier Inc.

Understanding, monitoring and modelling attributes of seagrass biodiversity, such as species composition, richness, abundance, spatial patterns, and disturbance dynamics, requires spatial information. This work assessed the accuracy of commonly available airborne hyper-spectral and satellite multi-spectral image data sets for mapping seagrass species composition, horizontal horizontal-projected foliage cover and above-ground dry-weight biomass. The work was carried out on the Eastern Banks in Moreton Bay, Australia, an area of shallow and clear coastal waters, containing a range of seagrass species, cover and biomass levels. Two types of satellite image data were used: Quickbird-2 multi-spectral and Landsat-5 Thematic Mapper multi-spectral. Airborne hyper-spectral image data were acquired from a CASI-2 sensor using a pixel size of 4.0 m. The mapping was constrained to depths shallower than 3.0 m, based on past modelling of the separability of seagrass reflectance signatures at increasing water depths. Our results demonstrated that mapping of seagrass cover, species and biomass to high accuracy levels (> 80%) was not possible across all image types. For each parameter mapped, airborne hyper-spectral data produced the highest overall accuracies (46%), followed by Quickbird-2 and then Landsat-5 Thematic Mapper. The low accuracy levels were attributed to the mapping methods and difficulties in matching locations on image and field data sets. Accurate mapping of seagrass cover, species composition and biomass, using simple approaches, requires further work using high-spatial resolution (< 5 m) and/or hyper-spectral image data. Further work is required to determine if and how the seagrass maps produced in this work are suitable for measuring attributes of seagrass biodiversity, and using these data for modelling floral and fauna biodiversity properties of seagrass environments, and for scaling-up seagrass ecosystem models.

Satellite imagery was used to detect the change in seagrass and macroalgal communities of a shallow coastal lake over a period of 14 years. The lake benthic material was classified into sets of spectral classes representing the patterns and texture of the ecosystem, and then linked to environmentally relevant labels through a radiative transfer model. The classification results for 2002 achieved an accuracy of 76% for the least understood areas; other areas were significantly better, but not quantified. Classification results of 1988, 1991, and 1995 were consistent with past surveys and maps. Based on the change detection from 1988 to 2002 Posidonia, Ruppia and Halophila change slightly in the 14 year period from 1988 to 2002. However, Zostera has undergone significant change and adaptation. Early in the time series (between 1988 and 1991) a reduction in Zostera beds was evident, especially in the middle and south of the lake with some areas not returning by 2002. Epiphytic growth over Zostera could be a confounding factor here, but the Landsat sensors do not have sufficient spectral resolution to detect these subtleties. Hyperspectral remote sensing could resolve this issue more clearly.

Consistent and accurate information on inland water quality over wider areas of the Australian continent are required to assess current condition and trends in response to key environmental and climatic impacts. Optical remote sensing offers a method to objectively assess this over multiple spatial scales provided retrieval algorithms are accurate. Here, we present the results of initial research aimed at exploring the optical variability in Australian inland waters and of linear matrix inversion algorithms applied to both in situ reflectance spectra and high resolution satellite data to retrieve water inland water quality parameters. In situ sampling reveals a high degree of optical variability both within and between lakes across the regions sampled with regional patterns evident; sub-tropical and tropical lakes exhibited greater optical complexity than deep lakes in mid-latitude regions. Clustering analysis indicated the presence of 8 different optical water types in the water bodies measured. The ability of the linear matrix inversion algorithm to map water quality, tested on in situ reflectance and WorldView2 image datasets, showed relative accuracy when parameter sets were sufficient to achieve algorithm closure. Improved algorithm parameterization will be required to account for the high degree in spatial and temporal optical variability observed in Australian inland waters.

Accurate baseline data of macrophyte extent is vital in estuarine monitoring. Previous techniques have often been laborious and subjective, while a purely empirical methodology often precludes transferring the method to other systems. The development of objective physics-based inversions models allows for the retrieval of; water depth, substratum composition and concentration of the water constituents from hyperspectral imagery. This paper describes approaches required to apply this method to QuickBird multispectral data from 2003 and 2008 over an estuarine lake. The addition of the inversion models quality control, improved the classification accuracy.

The focus of this chapter lies in describing digi-tal multispectral and hyperspectral remote sensingdevelopments and applications in the mapping andmonitoring of seagrass ecosystems. Multispectralrefers to a sensor that registers light in a limitednumber of relatively broad spectral bands (band-widths of 20–60 nm); hyperspectral (also referred toas imaging spectrometry) is defined for sensors thatmeasure the entire spectrum under consideration incontiguous narrow spectral bands (bandwidths be-tween 2 and 20 nm).Currently, seagrass maps are still predominantlybeingproducedfromtheinterpretationofaerialpho-tographyalthoughitislikelythatairborneandspace-borne remote sensing methods will rapidly take overthisrolegiventheadvantagestheypresentintermsofaccuracy, repeatability, versatility, and informationcontent. Nevertheless, retrospective studies of sea-grass change using the more modern methodologieswillstillneedtomakeuseofresultsgeneratedbythemore traditional methods since aerial photographsare the dominant archival source of historical spatial

Airborne hyperspectral data was collected in April 1999 over 16 square kilometres of coastal waters adjacent to the South Australian Bolivar Wastewater Treatment Plant (near Adelaide). Concurrent in situ measurements of benthic reflectances were collected. A subsequent field-work mission to validate the 1999 image analysis was completed in February 2000. The aim was to map substrate type accurately, differentiating species and canopy density, if possible. The analysis approach was based on the coupling of radiative transfer models using in situ atmospheric measurements and inwater measurements to remove atmospheric and water column effects from the image data. After removal of atmospheric effects, the spectral variation is a function of the water columns constituents and spatial related effects along the image. Water samples were collected at the time of the flights and analysed for the retrieval of inherent optical properties (IOPs). The SpecTool software was used to derive benthic reflectance models based on the IOPs. The atmospheric and in-water radiative transfer corrections applied to the imagery enabled delivery of environmental baseline data of the substrate and a basis for accurate multitemporal data analysis.

Moreton Bay is the Australian EO1-Hyperion coastal site used for Cal/Val activities. Moreton Bay shows spatial gradients in optical depth, bathymetry, and substrate composition. The turbid and humic river inputs, as well as the open ocean flushing, determine the water quality of the bay. Lyngbya toxic algae blooms have become a serious environmental and health concern. The field campaigns, carried out to coincide with Hyperion overpasses, focussed on the retrieval of inherent optical properties, apparent optical properties, substrate reflectance spectra and water quality parameters. Spectra from a 12 January 2001 Hyperion image show very close agreement to in situ upwelling radiance spectra.