Search

Citizen science projects engage the public in monitoring the environment and can collect useful data. One example is the CrowdWater project, in which stream levels are observed and compared to reference photos taken at an earlier time to obtain stream level class data. However, crowd-based observations are uncertain and require data quality control. Therefore, we used a deep learning model to estimate the water-level class for photos taken by citizen scientists at different times for the same stream and compared different options for model training. The models had a root mean square error (R) of 0.5 classes or better for all but four of the 385 sites for which the model was trained. Low water levels were in general predicted better than high water levels (R of 0.6 vs 1.0 classes). The study thus highlights the potential of human–computer interaction for data collection and quality control in citizen science projects. [ABSTRACT FROM AUTHOR]

Particulate organic matter (POM) plays a major role in freshwater ecosystems by serving as a bridge for the conversion of various nutrients. The composition and sources of POM in inland lakes are complex, making it difficult to estimate its concentration accurately via remote sensing. Therefore, a classification-based method based on the sources and composition of POM is proposed for estimating POM concentrations in inland lakes. In this study, 379 samples were collected from ten lakes in the Yangtze River Delta (YRD) at different times. A water-type classification method based on OLCI R r s (560) , R r s (620) , R r s (681) , and R r s (709) was developed for POM estimation based on biological and optical characteristics. Water type 1 is relatively clear, and POM may originate from aquatic vegetation or sediment. Water type 2 was dominated by inorganic suspended matter, and POM mainly originated from the attachment and entrainment of inorganic minerals. Water type 3 is an algae-dominated water body, and POM is mainly derived from fresh algal particles and the microbial degradation of phytoplankton. Therefore, specific POM estimation algorithms were developed for each water type. OLCI R r s (490) , R r s (665) , R r s (673) , and R r s (754) were used for water type 1; R r s (620) , R r s (865) , and R r s (885) were adopted for water type 2; and R r s (620) , R r s (665) , and R r s (709) were selected for water type 3. Using an independent dataset to evaluate the estimation accuracy of the developed algorithm, the results show that the estimation performance of this algorithm is significantly improved compared to the two other algorithms used; the mean absolute percentage errors (MAPE) decreased from 72.56% and 52.21% to 32.61%, and the root mean square errors (RMSE) decreased from 3.05 mg/L and 2.24 mg/L to 1.75 mg/L. A random error analysis of the atmospheric correction demonstrated that this algorithm is robust and can still perform well within a random error of 30%. Finally, this method was successfully applied to map the POM concentrations in the YRD using OLCI images acquired on November 12, 2020. [ABSTRACT FROM AUTHOR]

Eutrophication is a severe environmental problem in the most inland lakes. Taihu Lake, the third largest freshwater lake in China, is one of the most eutrophic lakes. Increased phosphorus concentration is thought to be the primary cause of water eutrophication. Proper total phosphorus (TP) monitoring is essential to effectively manage eutrophication in the Taihu Lake. In this study, 216 sample points with 172 modelling points and 44 verification points acquired from 5 field experiments in Taihu Lake, a semi-analytical model was established for estimating the TP concentration based on the absorption coefficient, with the mean absolute percentage error (MAPE) and root mean square error (RMSE) as 16.8% and 0.04 mg L−1, respectively. The application of the model on Sentinel-3 Ocean and Land Colour Instrument (OLCI) data proved the suitability of the model for satellite image with the MAPE and RMSE as 28.56% and 0.05 mg L−1. The OLCI derived TP concentration in the Taihu Lake shown the unique temporal and spatial distribution: TP concentration in Zhushan Bay, Meiliang Bay, Gonghu Bay, and the western coastal regions was significantly higher than the eastern, the southern, and the centre areas. The concentration of TP in Taihu Lake was the highest in summer and the lowest in winter. Furthermore, the versatility of this method was validated by applying in the Hongze Lake and Qiandao Lake, which is a water body dominated by inorganic suspended matter and a clean water body in China with the MAPE and RMSE as 24.10%, 19.35% and 0.004 mg L−1, 0.014 mg L−1. The results of this study could provide a reference for future remote sensing estimation of TP concentration and inland water eutrophication researches. [ABSTRACT FROM AUTHOR]

Phytoplankton carbon, an important biogeochemical and ecological parameter, plays a critical role in the carbon cycle and in global warming reduction. Estimation of phytoplankton carbon in inland waters on a large scale using remote sensing is useful for understanding, evaluating, and monitoring the carbon dynamics, and, in particular, for determining the spatial–temporal variation of primary production in inland waters. In a correlation analysis of the phytoplankton carbon concentration and water components, the result revealed no significant correlation between the chlorophyll-a concentration and phytoplankton carbon concentration in inland waters. However, the absorption peak height of particles at 675 nm, which is defined as the absorption at 675 nm subtracted by that at 660 nm, was found to be closely correlated with the phytoplankton carbon concentration. Thus, the absorption peak height of particles at 675 nm could be used as an indicator of the phytoplankton carbon concentration. A semianalytical method based on the remote-sensing reflectance in Sentinel-3 Ocean and Land Color Instrument (OLCI) bands 8, 9, and 17 was developed to derive the absorption peak of particles at a wavelength of 675 nm. Finally, an algorithm for estimating the phytoplankton carbon concentration in inland waters using OLCI bands 8, 9, and 17 was constructed. From 2013 to 2018, eight field campaigns were conducted in inland lakes in different seasons, and the optical properties, optically active water components, and phytoplankton carbon concentrations were obtained. An assessment of its accuracy using an independent data set demonstrated that the algorithm performance is acceptable (mean absolute percentage error, 48.6%, and root mean square error, 0.36 mg/L). As a demonstration, the algorithm was successfully applied to map the phytoplankton carbon concentration in Taihu Lake and Chaohu Lake, China, using OLCI images acquired on December 5, 2017, and August 5, 2018 and December 8, 2017, and August 7, 2017, and the spatial variation of the phytoplankton carbon concentration in Taihu Lake and Chaohu Lake was thoroughly examined. A semianalytical remote-sensing method, instead of a traditional time-consuming approach, was developed in this article, which can be used to estimate phytoplankton carbon concentration quickly, and provides an effective way to map the phytoplankton carbon concentration in inland water on a large scale. [ABSTRACT FROM AUTHOR]

The proportion of cyanobacterial biomass (PCB) can indicate the dominance of cyanobacteria in water, and it provides important information regarding the phytoplankton composition. Variations in the PCB can clarify the cyanobacteria accumulation process. Therefore, using remote sensing to obtain the spatial–temporal distribution of the PCB in inland lakes is very important for understanding the cyanobacterial bloom process. A total of 357 water samples was collected from three Chinese eutrophic inland lakes: Taihu Lake, Dianchi Lake, and Chaohu Lake, over eight field campaigns from 2013 to 2018. A normalized index of the PCB (NIPCB) was proposed using the absorption of chlorophyll-a (Chl-a) at 665 nm ($a_{\mathrm {chl}}$ (665)) and the absorption of phycocyanin (PC) at 620 nm ($a_{\mathrm {pc}}$ (620)) to indicate the PCB. A classification method using the PC index (PCI) was developed to identify the water types using the different algae composition features, and the relationship between the NIPCB and PCB in two water types was extensively examined. Finally, a semianalytical algorithm was developed to estimate the PCB based on the NIPCB index and the OLCI bands. The result of the estimated accuracy assessment using the independent validating data set demonstrated that the algorithm performance is satisfactory with mean absolute percentage error (MAPE) of 31.01% and root mean square error (RMSE) of 0.20. The MUMM method was then applied to conduct an OLCI image atmospheric correction with a satisfactory performance with the MAPE calculated between the measured remote sensing reflectance and the retrieved remote sensing from a satellite image of less than 25%. Consequently, the developed algorithm was successfully applied to three OLCI images of Dianchi Lake on April 12, 2017, Taihu Lake on May 18, 2017, and Chaohu Lake on September 8, 2018. The spatial distribution of the PCB in these three inland lakes is reasonable, and it demonstrated that the developed algorithm could be used to effectively monitor PCB at a large scale and provide important assistance in predicting the cyanobacterial blooms in eutrophic inland lakes by using remote sensing. [ABSTRACT FROM AUTHOR]

• A two-step method was proposed to estimate forest LAI using CHRIS/PROBA data. • PVI was found to have the best relationship with LAI. • The forward observation angle 55° had the highest estimation accuracy of five angle. • The combination of multiple observation angles can provide more useful spectral indications to estimate the LAI. Vegetation is a key factor affecting soil erosion, which depends not only on the horizontal structure of vegetation, but also on its vertical structure. The leaf area index (LAI) of plants, as an important structural parameter of the terrestrial ecosystem, can reflect horizontal vegetation coverage, vertical structure of vegetation, litter thickness and underground biomass. All these are the principal aspects of vegetation affecting soil erosion. The multi-angle remote sensing data can provide the stereoscopic and directional information of the surface target, which is helpful to extract the vertical structure parameters of vegetation effectively. Therefore, in order to improve the estimation accuracy of soil erosion, a two-step method was developed in this study to estimate the forest LAI under different vertical structures using the PROSAIL model and multi-angle CHRIS/PROBA data based on the field data. A total of 70 quadrats (10 m × 10 m sized) were deployed in Mount Zijin with undisturbed natural vegetation, considering the vegetation types and accessibility in the middle of August 2017. The pixels were first classified into two categories depending on complexity of vertical structure; then, a specific algorithm was applied to the different classes. For Class 1, the single angle data of 55° had the highest accuracy, indicating that, for forests with a simple vertical structure, the single angle data was sufficient for LAI estimation. For Class 2, the multi-angle data combination of 0°, 55° and −55° had the highest accuracy, indicating that, for forests with complex vertical structure, the multi-angle data can improve the accuracy of LAI estimation. The mean absolute percentage error (MAPE) and the root mean square error (RMSE) were applied to evaluate the accuracy of the LAI estimation algorithm. The results indicated that the two-step method can significantly improve the LAI estimation accuracy of forests with complex vertical structure, with a decrease in MAPE and RMSE of 19.52% and 23.08%, respectively, compared to the traditional empirical algorithm. The developed LAI estimation algorithm can be widely applied to forests with different vertical structures. Lastly, the two-step method was applied to the CHRIS/PROBA image acquired on September 7th, 2015 for mapping the spatial distribution of forest LAI in the western part of Mount Zijin. The result showed that the northern part and the southern part had higher LAI than the middle part of Mount Zijin. [ABSTRACT FROM AUTHOR]

Abstract: With the rapid development of the economy in recent years, massive algal (blue-green algae in particular) blooms have often observed in Chinese eutrophic lakes. The concentration of the cyanobacterial pigment phycocyanin (PC), an accessory pigment unique to freshwater blue-green algae, is often used as a quantitative indicator of blue-green algae in eutrophic inland waters. The purpose of this study was to evaluate the semi-analytic PC retrieval algorithm proposed by Simis et al. and to explore the potential to improve this PC algorithm so that it is more suitable for eutrophic lakes, such as Taihu Lake. In this paper, we recalculated the correction coefficients γ and δ to calculate the absorptions of chlorophyll-a at 665nm and the absorptions of phycocyanin at 620nm in terms of in situ measurements and observed that the values of these coefficients differed from the values used by Simis et al. and Randolph et al. The two coefficients are site dependent due to the different bio-optical properties of lakes. We also observed that the specific PC absorption at 620nm a pc*(620) decreases exponentially with an increase in PC concentrations. Therefore, a non-linear power–function of a pc*(620), instead of a constant value of a pc*(620) as used by Simis et al., was proposed for our improved PC retrieval algorithm in Taihu Lake, yielding a squared correlation coefficient (R2) of 0.55 and a root mean square error (RMSE) of 58.89μg/L. Compared with the original PC retrieval algorithm by Simis et al., the improved retrieval algorithm has generally superior performance. In evaluating the limitation of the PC retrieval algorithms, we observed that the ratio of the total suspended solids to phycocyanin can be used as a primary measure for retrieval performance. Validation in Dianchi Lake and an error analysis proved that the improved PC algorithm has a better universality and is more suitable for eutrophic lakes with higher PC concentrations. [Copyright &y& Elsevier]

Understanding the spatiotemporal dynamics of total dissolved phosphorus concentration (C TDP) and its regulatory factors is essential to improving our understanding of its impact on inland water eutrophication, but few studies have assessed this in eutrophic inland lakes due to a lack of suitable bio-optical algorithms allowing the use of remote sensing data. We developed a novel semi-analytical algorithm for this purpose and tested it in the eutrophic Lake Taihu, China. Our algorithm produced robust results with a mean absolute square percentage error of 29.65% and root mean square error of 9.54 μg/L. Meanwhile, the new algorithm demonstrates good portability to other waters with different optical properties and could be applied to various image data, including Moderate Resolution Imaging Spectroradiometer (MODIS), Medium Resolution Imaging Spectrometer (MERIS), and Ocean and Land Color Instrument (OLCI). Further analysis based on Geostationary Ocean Color Imager observations from 2011 to 2020 revealed a significant spatiotemporal heterogeneity of C TDP in Lake Taihu. Correlation analysis of the long-term trend between C TDP and driving factors demonstrated that air temperature is the dominant regulating factor in variations of C TDP. This study provides a novel algorithm allowing remote-sensing monitoring of C TDP in eutrophic lakes and can lead to new insights into the role of dissolved phosphorus in water eutrophication. [Display omitted] • A novel algorithm allows remote-sensing monitoring of dissolved phosphorous in eutrophic lakes. • Applying this to Lake Taihu allowed new insights into spatiotemporal dissolved P dynamics. • Dissolved P levels were found to be strongly linked to air temperature over time. • This method has clear benefits for monitoring and mitigation efforts. [ABSTRACT FROM AUTHOR]

The accurate remote estimation of the Secchi disk depth (ZSD) in turbid waters is essential in the monitoring the ecological environment of lakes. Using the field measured ZSD and the remote sensing reflectance (Rrs(λ)) data, a new semi-analytical algorithm (denoted as ZSDZ) for retrieving ZSD was developed from Rrs(λ), and it was applied to Geostationary Ocean Color Imager (GOCI) images in extremely turbid waters. Our results are as follows: (1) the ZSDZ performs well in estimating ZSD in turbid water bodies (0.15 m < ZSD < 2.5 m). By validating with the field measured data that were collected in four turbid inland lakes, the determination coefficient (R2) is determined to be 0.89, with a mean absolute square percentage error (MAPE) of 22.39%, and root mean square error (RMSE) of 0.24 m. (2) The ZSDZ improved the retrieval accuracy of ZSD in turbid waters and outperformed the existing semi-analytical schemes. (3) The developed algorithm and GOCI data are in order to map the hourly variation of ZSD in turbid inland waters, the GOCI-derived results reveal a significant spatiotemporal variation in our study region, which are significantly driven by wind forcing. This study can provide a new approach for estimating water transparency in turbid waters, offering important support for the management of inland waters. [ABSTRACT FROM AUTHOR]

Atmospheric correction is an essential prerequisite for obtaining accurate inland water color information. An inland water atmospheric correction algorithm, ACbTC (Atmospheric Correction based on Turbidity Classification), was proposed in this study by using OLCI (Ocean and Land Color Instrument) and SLSTR (Sea and Land Surface Temperature Radiometer) synergistic observations for the first time. This method includes two main steps: (1) water turbidity classification by the GRA index (GRAdient of the spectrum index); and (2) atmospheric correction by synergistic use of OLCI and SLSTR images. The algorithm was validated with 72 in situ sampling sites in Lake Erhai, Lake Hongze, and Lake Taihu, and compared with other atmospheric correction methods, i.e., C2RCC (Case 2 Regional Coast Colour processor), MUMM (Management Unit of the North Seas Mathematical Models), FLAASH (Fast Line-of-sight Atmospheric Analysis of Hypercubes), POLYMER (POLYnomial based algorithm applied to MERIS), and BPAC (Bright Pixel Atmospheric Correction). The results show that (1) the GRA index performed better than the proposed turbidity classification indices, i.e., the Diff (spectral difference index) and the Tind (turbid index), in inland lakes by using the reflectance peak at 1020 nm in clean water; (2) the synergistic use of OLCI and SLSTR performed feasibly for atmospheric correction, and the ACbTC algorithm achieved full-band average values of the mean absolute percentage error (MAPE) = 29.55%, mean relative percentage error (MRPE) = 13.98%, and the root mean square of error (RMSE) = 0.0039 sr−1, which were more reliable than C2RCC, MUMM, FLAASH, POLYMER, and BPAC; and (3) the synergistic use of the 17th band (865 nm) on OLCI and the 5th band (1613 nm) on SLSTR are suitable for clean inland lakes, while both the 5th band (1613 nm) and 6th band (2250 nm) on SLSTR are advisable for the turbidity. [ABSTRACT FROM AUTHOR]

• The a ph (681)/b b (681) was an optimal optical indicator of particulate composition. • A semianalytical algorithm was developed for estimating Chla/TSM in inland waters. • The proposed algorithm had better performance than two state-of-the-art algorithms. • The method could be applied to satellite sensors with similar bands. The composition of suspended particles is a key factor in determining the underwater light field, which is of great significance for understanding the variability in the optical properties of water bodies. In this study, the ratio of the phytoplankton absorption coefficient to the backscattering coefficient at wavelength of 681 nm (a ph (681)/b b (681)) was found to be an optimal optical indicator of the ratio of chlorophyll-a to the total suspended matter concentration (Chla/TSM), a parameter indicating the particulate composition. Therefore, a semianalytical algorithm was proposed to estimate Chla/TSM from remote sensing reflectance (R rs (λ)) at 681 nm and 754 nm on Sentinel-3 Ocean and Land Color Instrument (OLCI) images. The validation dataset collected from 11 inland lakes and 3 reservoirs in China and 2 inland lakes in America was used to evaluate the algorithm's performance. The evaluation results demonstrated that the proposed algorithm could have favorable performance in inland waters. Furthermore, comparison with two other state-of-the-art algorithms (Sun_13 and NTD675) showed that this proposed algorithm had higher estimation accuracy, with an overall winning rate (OWR) of 60%, an unbiased mean absolute percentage error (UMAPE) reduction from 72.95% to 46.44%, a root mean square error (RMSE) decline from 1.42 µg/mg to 0.83 µg/mg and a normalized root mean square error (NRMSE) decline from 11.36% to 6.62%. This algorithm was successfully applied to acquire the Chla/TSM tempo-spatial variation using the OLCI images of Lake Taihu from 2016 to 2019. It was found that the algorithm developed based on OLCI images can be applied to satellite sensors with similar bands, such as Medium-Resolution Imaging Spectrometer (MERIS), and Sentinel-2 multispectral instrument (MSI), etc. As a simple and effective algorithm, the proposed algorithm has the potential to monitor changes in Chla/TSM in inland waters on a global scale. [ABSTRACT FROM AUTHOR]

Water clarity is a critical parameter of water, it is typically measured using the setter disc depth (SDD). The accurate estimation of SDD for optically varying waters using remote sensing remains challenging. In this study, a water classification algorithm based on the Landsat 5 TM/Landsat 8 OLI satellite was used to distinguish different water types, in which the waters were divided into two types by using the a d (443)/a p (443) ratio. Water type 1 refers to waters dominated by phytoplankton, while water type 2 refers to waters dominated by non-algal particles. For the different water types, a specific algorithm was developed based on 994 in situ water samples collected from Chinese inland lakes during 42 cruises. First, the R rs (443)/R rs (655) ratio was used for water type 1 SDD estimation, and the band combination of (R rs (443)/R rs (655) - R rs (443)/R rs (560)) was proposed for water type 2. The accuracy assessment based on an independent validation dataset proved that the proposed algorithm performed well, with an R2 of 0.85, mean absolute percentage error (MAPE) of 25.98%, and root mean square error (RMSE) of 0.23 m. To demonstrate the applicability of the algorithm, it was extensively evaluated using data collected from Lake Erie and Lake Huron, and the estimation accuracy remained satisfactory (R2 = 0.87, MAPE = 28.04%, RMSE = 0.76 m). Furthermore, compared with existing empirical and semi-analytical SDD estimation algorithms, the algorithm proposed in this paper showed the best performance, and could be applied to other satellite sensors with similar band settings. Finally, this algorithm was successfully applied to map SDD levels of 107 lakes and reservoirs located in the Middle-Lower Yangtze Plain (MLYP) from 1984 to 2020 at a 30 m spatial resolution, and it was found that 53.27% of the lakes and reservoirs in the MLYP generally show an upward trend in SDD. This research provides a new technological approach for water environment monitoring in regional and even global lakes, and offers a scientific reference for water environment management of lakes in the MLYP. • An optical classification method for estimating water clarity in inland lakes is proposed. • The proposed algorithm had better performance than the other empirical and semi-analytical algorithms. • The proposed algorithm has the ability of mapping inland waters clarity with large-scale. • The 36-years of water clarity levels in the Middle-Lower Yangtze Plain were mapped using Landsat images. [ABSTRACT FROM AUTHOR]

• A K d (490) estimation algorithm was developed based on MODIS/Aqua image in Lake Hongze. • The spatiotemporal patterns of K d (490) were mapped from 2002 to 2016 in Lake Hongze. • The affection of the optical properties, the concentration of water constituents and their PSD characteristics on K d (490) were discussed. • The response of K d (490) spatiotemporal patterns on precipitation and wind speed were discussed from the annual and inter-annual scales. The diffuse attenuation coefficient K d (λ) is an important optical parameter that can be determined remotely. It plays a meaningful role in inland and coastal biogeochemical processes and aids in determining water quality. Four data-gathering expeditions were conducted in Lake Hongze (HZL), China, between 2015 and 2017. Based on the in situ optical and biological data, we constructed and evaluated a band ratio model for retrieving K d (490) of HZL from MODIS/Aqua images between 2002 and 2016. Our results are as follows: (1) The developed algorithm, using the visible bands (R rs (red)/ R rs (green)), performs robustly for K d (490) estimation, with R 2 of 0.82 and p < 0.0001 for the modeling data. By validating with the in situ observed data, the mean absolute percentage error is determined to be 17.82%, and the root means square error is 1.52 m−1. (2) The spatial distribution of K d (490) strong relates with the input river that formed a cone-shaped distribution with the vertex at Huai River Estuary, decreasing from the southeast (RC) to the northern (NE) and western bays in HZL. The largest K d (490) value is in RC, with 7.10 ± 3.03 m−1, followed by NE and WL (wetland) with values of 6.10 ± 2.58 m−1 and 5.32 ± 2.09 m−1, respectively. (3) K d (490) reaches high values in winter and low values in the summer although the variance is quite different in different lake areas. The seasonal distribution of K d (490) in HZL is driven by the runoff produced by precipitation. (4) K d (490) decreased systematically as the East Asian monsoon weakened from 2002 to 2012 that means wind has great influence on K d (490) variation from year scale. However, human activity like sand-mining disrupted this trend, leading to the rise of K d (490) during 2012 and 2016. [ABSTRACT FROM AUTHOR]