Despite significant declines in external phosphorus loads, Lake of the Woods continues to experience severe recurring cyanobacterial harmful algal blooms (cHABs) covering as much as 80% of the lake surface area. Satellite-derived bloom indices were used to assess the status, trends, and drivers of cHAB conditions for the period 2002 to 2021 in support of developing ecosystem objectives and response indicators for the lake. Areas of greatest potential concern, with the most prolonged bloom occurrences, were in the southeast of the lake. Significant decreases in bloom indices suggest the lake may now be responding to historical nutrient reductions. The greatest rates of decrease were within the main water flow paths, with little change in the more isolated embayments, suggesting flushing plays a key role in regulating regional bloom severity. Significant inter-annual variability in bloom phenology was observed, with blooms peaking later in recent years, which may be in response to climate-induced changes in the lake and watershed. The absence of a direct relationship between external phosphorus loads and annual bloom severity reflects the complexity of the lake’s response to eutrophication and the potential roles of other drivers including climate and a strong legacy effect of sedimentary nutrients. A case study of the 2017 bloom season captures the compounding interaction of meteorological variability and seasonal nutrient delivery in regulating the bloom response. Results highlight the need for greater understanding of seasonal and regional variability of bloom drivers to aid in forecasting the lake’s recovery under both nutrient management and climate change scenarios.

Early detection and comprehensive monitoring of inland water algal blooms is fundamental to their effective management and mitigation of potential ecosystem and public health impacts. With the spatial and temporal limitations of in situ sampling, algal bloom monitoring capabilities have been enhanced greatly by advancements in satellite Earth Observation (EO). Three turbid, eutrophic Canadian lakes (Lake Winnipeg (LW); Lake Erie (LE); Lake of the Woods (LoW)) have been the focus of Environment and Climate Change Canada (ECCC) research and monitoring initiatives due to concerns over persistent degraded water quality from recurring algal blooms. ECCC's EOLakeWatch was developed to deliver a suite of useful, easily interpretable, and accessible EO-derived products to support algal bloom monitoring on these three lakes. Algal bloom indices, describing bloom spatial extent, intensity, duration, and severity were derived using the European Space Agency's OLCI (Ocean and Land Colour Instrument) sensor for observations from 2016 to present and its predecessor MERIS (Medium Resolution Imaging Spectrometer) for 2002 to 2011. Results document widespread blooms on each lake, with maximum spatial extent of 21,641 km(2) (representing 88.1% of the lake area) on LW, 3070 km(2) (79.5%) on LoW and 5257 km(2) (19.7%) on LE. Bloom intensity showed seasonal and inter-annual variability on all three lakes, with a suggestion that LoW may be responding to reduced nutrient loads with a recent decrease in bloom intensity. Annual bloom duration on LW and LoW was on average 44 and 47 days respectively, while on LE blooms were significantly shorter in duration at an average of 24 days. Variance among the derived bloom indices was shown to be significant (i.e. the most extensive bloom was not necessarily the longest or most intensive), demonstrating the need for the indices to be used collectively, or for any single comprehensive bloom indicator to capture the variability of all individual metrics. Bloom indices are processed in a fully automated operational capacity, distributed in near-real-time through a web portal and collated into end-user-friendly annual algal bloom reports for each lake. These products go a long way to address existing monitoring gaps, delivering prompt, consistent measures of lake-wide algal bloom conditions required to provide stakeholders with early warning of bloom risks, identify areas of potential concern, quantify spatio-temporal trends, further understand bloom dynamics and drivers, as well as guide and determine the effectiveness of implemented management actions.

Since the early 2000s Lake Erie has seen a dramatic increase in phytoplankton biomass, manifested in particular by the rise in the severity of cyanobacteria blooms and the prevalence of potentially toxic taxa such as Microcystis. Satellite remote sensing has provided a unique capacity for the synoptic detection of these blooms, enabling spatial and temporal trends in their extent and severity to be documented. Algorithms for satellite detection of Lake Erie algal blooms often rely on a single consistent relationship between algal or cyanobacterial biomass and spectral indices such as the Maximum Chlorophyll Index (MCI) or Cyanobacteria Index (CI). Blooms, however, are known to vary significantly in community composition over space and time. A suite of phytoplankton and optical property measurements during the western Lake Erie algal bloom of 2017 showed highly diverse bloom composition with variable absorption and backscatter properties. Elevated backscattering coefficients were observed in the Maumee Bay, likely due to phytoplankton cell morphology and buoyancy regulating gas vacuoles, compared with typically Planktothrix dominated blooms in Sandusky Bay. MCI and CI calibrated to historical chlorophyll observations and applied to Sentinel 3's OLCI sensor accurately captured the 2017 bloom in Maumee Bay but underestimated the Sandusky Bay bloom by nearly 80%. The phycoerythrin-rich picocyanobacteria Aphanothece and Synechococcus were found in abundance throughout the western and central basins, resulting in substantial biomass underestimations using blue to green ratio-based algorithms. Potential misrepresentation of bloom severity resulting from phytoplankton optical properties should be considered in assessments of bloom conditions on Lake Erie.

Lake Winnipeg has experienced dramatic increases in nutrient loading and phytoplankton biomass over the last few decades, accompanied by a marked shift in community composition towards the dominance of cyanobacteria. Comprehensive lake-wide observations of algal blooms are critical to assessing the lake's health status, its response to nutrient management practices, and an improved understanding of the processes driving blooms. We present an analysis of the spatial and temporal variability of algal blooms on Lake Winnipeg using satellite-derived chlorophyll and indices for algal bloom intensity, spatial extent, severity, and duration over the period of ESA's MERIS mission (2002–2011). Imagery documented extensive blooms covering as much as 93% of the lake surface. Bloom conditions were analysed in the context of in-lake and watershed processes to gain further insight on the drivers of bloom events. Day to day bloom variability was driven primarily by intermittent wind mixing events, with quiescent periods leading to the formation of dense surface blooms. Seasonal bloom distribution was consistent with light limitation in the south basin and lake circulation transporting bloom material towards the north-east shore. Inter-annual variability in average bloom severity was related to both total phosphorus (TP) loadings and summer lake surface temperatures. Results provide a valuable historical time series of bloom conditions to which ongoing observations from Sentinel-3's OLCI sensor can be added for longer term monitoring and change detection.

Mid-winter limnological surveys of Lake Erie captured extremes in ice extent ranging from expansive ice cover in 2010 and 2011 to nearly ice-free waters in 2012. Consistent with a warming climate, ice cover on the Great Lakes is in decline, thus the ice-free condition encountered may foreshadow the lakes future winter state. Here, we show that pronounced changes in annual ice cover are accompanied by equally important shifts in phytoplankton and bacterial community structure. Expansive ice cover supported phytoplankton blooms of filamentous diatoms. By comparison, ice free conditions promoted the growth of smaller sized cells that attained lower total biomass. We propose that isothermal mixing and elevated turbidity in the absence of ice cover resulted in light limitation of the phytoplankton during winter. Additional insights into microbial community dynamics were gleaned from short 16S rRNA tag (Itag) Illumina sequencing. UniFrac analysis of Itag sequences showed clear separation of microbial communities related to presence or absence of ice cover. Whereas the ecological implications of the changing bacterial community are unclear at this time, it is likely that the observed shift from a phytoplankton community dominated by filamentous diatoms to smaller cells will have far reaching ecosystem effects including food web disruptions.

Satellite remote sensing methods adopting wavelengths in the red and near infra-red have been shown to be superior to the standard blue to green ratio based approaches in the detection of algal blooms under turbid, eutrophic conditions. Here, the MERIS Maximum Chlorophyll Index (MCI) has been explored as a tool for monitoring algal blooms in North America's inland waters where waters range from optically complex, turbid, eutrophic conditions, to low chlorophyll and oligotrophic conditions. Assessment of the MERIS MCI product is made for intense blooms of cyanobacteria in Lake of the Woods, algal blooms in turbid waters of Lake Erie, and low chlorophyll conditions in Lake Ontario. The MCI product is shown to be a versatile tool in monitoring intense surficial algal blooms with chlorophyll concentrations in the 10–300mgm−3 range, while limited in its application to low-biomass conditions as observed in Lake Ontario. Wavelength shifts in the position of the MCI peak for different chlorophyll concentration ranges, as well as variations in the inherent optical properties of water colouring constituents, are anticipated to account for regional variations in MCI–chlorophyll relationships and potentially hinder a universally applicable quantitative MCI product.

Satellite-derived estimates of chlorophyll concentrations based on colour ratio algorithms traditionally fail in turbid waters such as those found in Lake Erie, resulting in chlorophyll concentrations often orders of magnitude in error and spatial distributions mirroring that of known suspended sediment distributions. Methods are presented here that were used to simultaneously extract algal and mineral suspended particulate matter for Lake Erie from the red and near-infrared bands of NASA's MODIS-Aqua sensor. Results produced spatially and temporally distinct seasonal cycles in agreement with bio-geo-physical processes on the lake. Derived imagery was used to monitor seasonal cycles of both algal and mineral particulate matter on the lake and determine areas of persistently elevated concentrations that may highlight regions of potential water quality concern.

Lake of the Woods (LoW) is an international (USA/Canada) inland water body under significant water quality pressures from recurring cyanobacteria blooms. Its remote location combined with the hydrologically complex nature of its waters makes adequate in situ monitoring of the lake difficult. This work aimed to test the potential of Envisat's Medium Resolution Imaging Spectrometer (MERIS) full-resolution imagery for monitoring algal blooms in the lake. A full assessment of MERIS L1 and L2 chlorophyll and chlorophyll-related products was carried out over LoW during an intense surface algal bloom in September 2009. The Case 2 regional model and fluorescence line height/maximum chlorophyll index (MCI) plug-ins for BEAM were assessed for their ability to accurately distinguish the bloom. Results suggest that none of the Case-2-specific algorithms effectively extract chlorophyll concentrations over LoW, whereas the greatest potential is seen within the MCI product. Adjacency effects in near-shore waters are shown to be significant, although the improved contrast between ocean and land processor (ICOL) does not appear to notably improve water constituent retrievals in these waters. Images of L2 MCI are shown to adequately identify the bloom and are used to track the evolution of the bloom across the lake. Evidence is presented for the effects of variable depth distributions of cyanobacteria on the surface signal seen by the sensor; imagery suggests that day-to-day variations in wind-induced mixing have a profound impact on surface algal biomass as detected by remote sensing.

This paper explores the use of Moderate Resolution Imaging Spectroradiometer (MODIS) wavebands in the red/near-infra-red for estimating concentrations of suspended particulate matter (SPM) in the moderately turbid, optically complex waters of Lake Erie. Observations show that at wavelengths shorter than 550 nm, more than 50% of the absorption signal is accounted for by dissolved organic matter and phytoplankton, confirming that algorithms incorporating these wavelengths may not be appropriate for these waters. Single band and band ratios at wavelengths greater than 667 nm are tested for their suitability for monitoring SPM concentrations in these waters. A simplified regional semi-analytical model is utilized which is independent of variations in dissolved organic matter and chlorophyll absorption, enabling estimates of SPM concentrations from MODIS water-leaving radiance at 748 nm with an average root mean square (RMS) error of 40%. Knowledge of the vertical distribution of particles enables estimates of total water column suspended loads which are then related to wind re-suspension events. The method is applied to MODIS water-leaving radiance at 748 nm to produce a time series of surface and total water column suspended loads in Lake Erie for the period 2003–2007.

A two-dimensional model has been developed in order to improve understanding of the processes which interact to maintain the sediment concentrations at an isolated turbidity maximum in the Irish Sea throughout the year. The model comprises two interchangeable populations of particles with different diameters, one of fine cohesive material, the other made up of floes. Both populations are slow settling, and subject to horizontal diffusion, resuspension, settling, aggregation and disaggregation.The equations used to describe the processes of aggregation and the break-up of floes in response to sediment concentration and turbulent shear have been developed by the tuning of the model to observations. Due to high turbulent shear at the turbidity maximum, the particles are predominantly fine, while the sediment in the surrounding water is made up of larger floes. Diffusion of small particles out of the turbidity maximum balanced by the diffusion of aggregated material towards it provides a mechanism for its maintenance. The modelled sediment concentrations at the turbidity maximum can be reproduced year-on-year, with no loss due to diffusive processes. This is achievable with a limited, exhaustible source of material which is not replenished once resuspended.Through comparison with satellite imagery the correlation of the modelled sediment concentrations with observations is investigated, both spatially and temporally. The seasonal cycle is reproduced well by the model with winter and summer concentrations matching those observed. Spatially the model also performs well in both turbid regions and those where the surface sediment load is low. (c) 2007 Elsevier Ltd. All rights reserved.

A method is described for estimating near surface suspended particle size over whole shelf regions using visible band satellite data. The technique can be applied to the mainly mineral, flocculated particles commonly found in tidally mixed shelf seas and estuaries. It is based on estimating light scattering per unit concentration (b∗) from simultaneous measurements of water colour (expressed as a reflectance ratio) and brightness (expressed as the absolute value of reflectance at a specific wavelength), both of which can be measured from space. A test of the method, using in situ data, produces predictions of b∗ which are in good agreement (R2=0.79) with direct measurements. An empirical relationship has been established between b∗ and median particle size by volume, DV, using in situ measurements of particle size measured with a laser diffraction (LISST) instrument. This, together with the algorithm for b∗ has been applied to two SeaWiFs images of the Irish Sea, one in winter, the other towards the end of summer. The maps show a decrease in particle size in regions of most intense tidal energy, in support of theories concerning the maintenance of turbidity maxima in the absence of a source of particles at these locations. The satellite observations of particle size are used to test the hypothesis that the maximum size is controlled by turbulence through the Kolmogorov microscale. A positive, statistically significant, correlation is found between median particle size and the turbulent microscale in both summer and winter. However, for a given turbulence level, particles are larger in summer than in winter, providing evidence for the importance of biological binding. These results mean that local knowledge of turbulence can be used to improve our estimation of suspended sediment load, underwater light attenuation and hence primary productivity from visible band satellite images of shelf seas.

Small mineral particles suspended in the sea are excellent at reflecting light and show up well in visible band satellite images. In order to make quantitative estimates of the particle concentration, and its effect on the penetration of sunlight into the sea, it is necessary to know how the absorption, scattering and backscattering coefficients of these inorganic particles change with concentration, the nature of the particles, and with wavelength. In this paper, observations from the literature are supplemented with a data set from the Irish Sea. The concentration-specific absorption coefficient of mineral particles am∗ is generally found to decrease exponentially with wavelength towards (in our data) a constant non-zero value in the red. Specific scattering coefficients show a tendency to decrease from the open ocean into energetic shelf seas and estuaries, but then to increase again within shelf seas as turbulent energy increases. The variation of specific scattering with turbulent energy in the Irish Sea is consistent with particle size scaling with the Kolmogorov microscale. Colour ratios (the ratio of two reflection coefficients) are less sensitive to variations in scattering, and we suggest that a combination of satellite measurements of brightness and colour in water with high mineral suspended sediment content will produce (1) a better estimate of concentration and (2) information on the variation of specific scattering.

This study investigates the use of single-band reflectance at visible wavelengths for the derivation of suspended sediment concentrations in the Irish Sea. A reasonably strong relationship was observed between irradiance reflectance at 665 nm (R665) and mineral suspended sediment (MSS) concentrations. Variability in the Reflectance–MSS relationship was found to be the consequence of changes in the mass-specific scattering coefficient (bMSS*) brought about by differences in particle properties such as grain size and composition. A systematic increase in the slope of the Reflectance–MSS relationship was observed with increasing bMSS*. A reflectance model is presented that highlights the dependence of reflectance on bMSS* and suggests that the errors in predicted MSS concentrations can be reduced from 56% to as little as 12% with prior knowledge of the scattering properties of the sediments under study. This paper highlights the need for a complete understanding of the scattering properties of particles in order to accurately estimate MSS concentrations from reflectance measurements. It is suggested that in order to obtain quantitative estimates of MSS in moderately turbid waters from space, it may be necessary to pre-determine scattering efficiencies, bMSS*, for the area of interest.

A study was conducted in the Irish Sea with the aim of deriving an algorithm for the retrieval of suspended sediment concentrations from ocean colour imagery obtained from the Sea-viewing Wide Field-of-view Sensor (SeaWiFS). In situ observations of the diffuse attenuation coefficient, Kd , and irradiance reflectance, R, were obtained at wavelengths coincident with the SeaWiFS visible wavebands using a Profiling Reflectance Radiometer (PRR600, Biospherical Instruments Inc., San Diego). Results showed that surface reflectance at 665 nm (R665 ), rather than variations in the intrinsic colour of the ocean (using colour ratios), was the most widely applicable method of obtaining suspended sediment concentrations from ocean colour imagery in this region. The derived algorithm enabled the estimation of mineral suspended sediment (MSS) concentrations from ocean colour in the Irish Sea, accurate to within 1 mg l−1 (see equation below). Furthermore, the application of this algorithm to a SeaWiFS image of the Irish Sea accurately reproduced known regions of high turbidity with realistic MSS concentrations. MSS=0.0441R 2 665 + 1.1392R 665 + 1.7459 (R 2=0.9105, n=124, RMS error=0.907) Specific absorption and scattering coefficients were derived for all optically active in-water constituents, namely yellow substance (YS), mineral suspended sediments (MSS) and phytoplankton pigments (C). An optical model based on the empirically derived absorption and scattering coefficients reproduced the observed relationship between MSS and R 665. Model results highlighted the relative insensitivity of reflectance at 665 nm to variations in the concentrations of other in-water constituents, suggesting that the algorithm may be applicable to the Irish Sea throughout the year.

This paper describes a novel method of deriving surface salinity from remotely sensed ocean colour. The method is based on two important observations of optical properties in regions of freshwater influence (ROFI). The first is the strong effect that a form of dissolved organic matter (yellow substance) has on ocean colour when present in relatively high concentrations. The second is the close relationship between salinity and yellow substance originating from fresh water runoff. In this paper these relationships are demonstrated for the Clyde Sea, Scotland, and applied to SeaWiFS imagery for the derivation of yellow substance and surface salinity. The empirical relationships demonstrated in this study allow the satisfactory prediction of yellow substance and salinity in the Clyde Sea from remotely sensed ocean colour. The r.m.s. difference between the observed and predicted parameters are 0.19 m(-1) and 1.1 for yellow substance and salinity, respectively, over a range of salinity from 16 to 34.Salinity maps created from the satellite-retrieved ocean colour identify features in the surface salinity distribution that to date have only been observed by in situ instrumentation. For example, the salinity distribution derived in this study is suggestive of an anticlockwise circulation in the Clyde Basin, driven both by the input of fresh water from the River Clyde and the intrusion of more saline water from the North Channel. The movement of deep water in the Clyde Sea has been well documented (Proc. R. Soc. Edinb. 90B (1986) 67), however, the residual circulation of surface water remains uncertain. The ability to obtain synoptic views of salinity such as those presented in this paper therefore provides great potential in furthering the understanding of Shelf Sea and coastal dynamics. (C) 2003 Elsevier Science B.V. All rights reserved.