Open AccessArticle

Chromaticity-Based Discrimination of Algal Bloom from Inland and Coastal Waters Using In Situ Hyperspectral Remote Sensing Reflectance

Dongzhi Zhao

^†

Qinshun Luo

^† and

Zhongfeng Qiu

School of Marine Sciences (SMS), Nanjing University of Information Science and Technology (NUIST), Nanjing 210044, China

Author to whom correspondence should be addressed.

^†

These authors contributed equally to this work.

Water 2024, 16(16), 2276; https://doi.org/10.3390/w16162276

Submission received: 30 May 2024 / Revised: 6 August 2024 / Accepted: 8 August 2024 / Published: 13 August 2024

(This article belongs to the Section Biodiversity and Functionality of Aquatic Ecosystems)

Download

Browse Figures

Versions Notes

Abstract

The rapid growth of phytoplankton and microalgae has presented considerable environmental and societal challenges to the sustainable development of human society. Given the inherent limitations of satellite-based algal bloom detection techniques that rely on chlorophyll and fluorescence methods, this study proposes a method that employs hyperspectral data to calculate water chromatic indices (WCIs), including hue, saturation (S), dominant wavelength (λd), and integrated apparent visual wavelength (IAVW), to identify algal blooms. A global in situ hyperspectral dataset was constructed, comprising 13,110 entries, of which 9595 were for normal waters and 3515 for algal bloom waters. The findings of our investigation indicate statistically significant discrepancies in chromaticity parameters between normal and algal bloom waters, with a p-value of 0.05. It has been demonstrated that different algal blooms exhibit distinct chromatic characteristics. For algae of the same type, the chromaticity parameters increase exponentially with chlorophyll concentration for hue and λd, while S shows low correlation and IAVW displays a good linear relationship with chlorophyll concentration. The application of this method to the Bohai Sea (coastal) and Taihu Lake (inland water) for the extraction of algal blooms revealed a clear separation in chromaticity parameters between normal and algal bloom waters. Moreover, the method can be applied to satellite data, offering an alternative approach for the detection of algal blooms based on satellite data. The indices can serve as ground truth values for colorimetric indices and provide a benchmark for the validation of satellite chromatic products.

Keywords:

chromatic indices; improved apparent visual wavelength; normal water; algal bloom-dominated waters; hyperspectral remote sensing reflectance

1. Introduction

The rapid proliferation of phytoplankton and macroalgae in aquatic ecosystems worldwide, nourished by nutrient discharge resulting from human activities and climate change, has engendered significant environmental and societal challenges for the sustainable development of human society [1,2]. This phenomenon, which is commonly referred to as an algal bloom, has been observed to exert a significant impact on several different areas, including socioeconomic aspects, ecosystems, the quality of human drinking water, and human health [3,4,5]. Algal blooms in inland water bodies are typically identified as Cyanobacteria Bloom, while those leading to abnormal visible discoloration in coastal areas are referred to as red tide [4], green tide [6], brown tide [7,8,9], and golden tide [10], with those harboring toxins denoted as Harmful Algal Bloom (HAB) [11].

At present, the detection of algal blooms is predominantly based on the use of chlorophyll as a proxy indicator. The principal methodology is based on the utilization of the band ratio technique within the visible light spectrum [12,13]. The absorption and scattering properties of watercolor components (chlorophyll, suspended matter, CDOM (colored dissolved organic matter)) differ significantly between algal bloom waters and normal first- and second-class waters [14,15,16]. Consequently, various pigment types are present in different algal species during red tides, including chlorophyll pigments such as chlorophyll a/b/c/d, carotenoids, and other auxiliary pigments [17]. Furthermore, different algal species exhibit varying concentrations of pigments, which can result in the failure of chlorophyll algorithms for watercolor. The efficacy of red tide detection is markedly diminished [18,19].

To address this challenge, advanced chlorophyll fluorescence algorithms operating in the near-infrared range have been employed for the detection of red tides [20,21,22]. However, due to the inherent limitations of satellite band design, the majority of current satellites utilize multispectral band configurations, including instruments such as MODIS, MERIS, Sentinel-2, SGLI, and others [23,24,25]. The utilization of fixed fluorescence bands by these satellites limits their capacity to accurately detect the diverse fluorescence behaviors exhibited by various complex red tide algal species [3,20,26].

The phenomenon of algal blooms is characterized by discoloration, which serves as a primary sensory indicator of their occurrence in the field [8]. The advent of sophisticated satellite remote sensing technology has led to the development of innovative techniques, namely True Color RGB and Natural Color RGB, which are derived from the RGB bands of satellite payloads. These techniques have become the primary tools for monitoring personnel in the identification and detection of algal blooms [27]. The standard chromaticity techniques [28] have been utilized to differentiate color into its constituent components, namely red, green, and blue (expressed as chromaticity coordinates X, Y, and Z, respectively). Additionally, the radiometric spectrum of color generated by the various combinations of color-producing agents (CPAs) present in aquatic environments has been determined. A radiometric color model has been employed to establish a relationship between the color of optically complex (non-Case I) waters and the organic and inorganic color-producing agents responsible for that color. This model has been applied to inland water bodies [29]. In contrast to the Forel-Ule (FU) scale, which has been employed extensively by oceanographers and limnologists since 1890, the optical properties of the FU scale and its capacity to encompass the color of natural waters as perceived by the human eye have been subject to scrutiny [30,31]. The standard 1931 CIE Color Matching Functions (CMFs) are employed to convert the light spectrum into three chromaticity coordinates: x, y, and z. These coordinates can be compressed into a single value, known as the hue angle (α), which represents the true color of natural waters [32]. Once the values of X, Y, and Z are known, chromaticity diagrams facilitate the determination of a dominant wavelength (λd), as well as the associated spectral purity (p) and saturation (S), for the radiometric color spectrum of water columns containing known concentrations and optical cross-section spectra of CPA [29].

The apparent visual wavelength (AVW) represents a one-dimensional geophysical metric that is associated with the intrinsic spectral shape. Notably, this index is highly effective in measuring spectral shape. Should any slight weight be added to either side of the spectrum (for example, a shift in color due to absorption or backscatter contribution), the balance point (AVW) will shift unless offset by a proportional counterbalanced weight elsewhere in the spectrum. This parameter is primarily employed in watercolor analysis and is utilized across a range of satellite images, including MODIS, SeaWiFS, VIIRS, and others [33]. Turner et al. [34] posit that integrative, continuous indices such as AVW can serve as efficacious indicators for the assessment of nearshore biogeochemical variability.

Our object is the investigation of the discoloration of Water Chromatic Indices (WCIs) between normal water (primarily refers to inland and coastal waters before algal bloom occurrences, typically categorized as Case I and Case II water [35]) and algae bloom water (algal-dominated water). The objective is to utilize WCIs for the characterization of watercolor changes to facilitate the separation of algal blooms. The specific objectives are as follows: The objective is to construct a dataset of algal bloom waters, to study the chromaticity differences between normal waters and algal bloom waters, to investigate the chromaticity differences among different algal species, and to examine the chromaticity differences of the same algal species at varying chlorophyll concentrations. The aforementioned efforts would provide an alternative approach to the detection of algal blooms via remote sensing, one that is distinct from those based on chlorophyll analysis.

2. Data and Methods

2.1. Data Resources

Nineteen bio-optical datasets were collated from a range of water bodies across the globe. These included both hyperspectral Rrs and reflectance data, which were collected in situ and in vivo. Further details regarding the datasets are provided in Table 1. The bio-optical datasets are presented in the following manner:

2.2. Case Study Area

The Bohai Sea, located in the northern region of China, represents the northernmost inner sea within the country. The sea is encircled by the Liaodong Peninsula, the North China Plain, and the Shandong Peninsula, with an area of 78,000 km². The geographic coordinates of the Bohai Sea range from approximately 37° to 39° N latitude and 119° to 121° E longitude. It is known for being influenced by over 40 rivers, including the Liao River, Shuangtaizi River, Daling River, Xiaoling River, Luan River, Hai River, Yellow River, and Xiaoqing River. It is noteworthy that the Liao River basin alone encompasses a drainage area of 138,993 km², with an annual runoff of 4 billion cubic meters. The Bohai Sea, a nearly enclosed, shallow sea with an average depth of 17 m, exhibits limited self-purification and exchange capabilities with the open sea. The coastal zone encompassing the Bohai Sea encompasses 20% of the national population. The Bohai Sea has been subjected to eutrophication, primarily as a result of terrestrial pollutant discharge, marine aquaculture, port traffic, and offshore oil development. This has led to elevated levels of inorganic nitrogen, inorganic phosphorus, and oil. As a consequence, this has led to a high incidence of red tide disasters [52].

Taihu Lake is situated in the lower reaches of the Yangtze River and encompasses a drainage basin of 36,895 km². It covers an extensive area of approximately 2338 square kilometers. As the third largest freshwater lake in China, Taihu Lake is situated in the center of the basin and is subject to a range of ecological challenges, particularly eutrophication and cyanobacterial blooms. The total length of the rivers in the Taihu Lake Basin is approximately 120,000 km². The basin is traversed by over 200 rivers, the majority of which are connected to Taihu Lake.

2.3. Methods

2.3.1. Separation of the Spectra of Algal Bloom Waters from the Inland and Coastal Water

From the collected datasets, the hyperspectral Rrs spectra ranging from 360 to 830 nm with a spectral resolution of greater than 5 nm were selected for subsequent analysis. In consideration of the disparate spectral ranges encompassed by the bio-optical datasets, this article is particularly oriented towards water bodies undergoing algal blooms with fluorescence peaks. The selected spectral curves must fall within the near-infrared band. In instances where spectral curves are absent in the shortwave UV (ultraviolet) to blue band, only the 390 nm to NIR (near-infrared) and 400 nm to NIR bands are selected for analysis.

In addition to the spectra of algal bloom water measured in both field and culture settings, the collected bio-optical dataset comprises two distinct types of spectral data: normal water and algal bloom water. To enable further research, it is essential to separate these datasets. Furthermore, the separated algal bloom spectral data must be categorized into two regional types: inland water and coastal water. This entails the classification of the algal bloom water spectrum for inland water and the red tide water spectrum for marine water. The categorization was performed using the Cyanobacteria Index (CI) method and the fluorescence height method, respectively.

Because inland water bodies are predominantly occupied by blue-green algae, Timothy T. Wynne and colleagues employed a method for calculating the value of Cl to classify cyanobacteria [53]. In this study, Rrs represents the hyperspectral data. It is hypothesized that Cl values below 0.002 [54] correspond to normal water bodies, while values equal to or exceeding 0.002 are indicative of water bodies undergoing algal blooms.

In the case of coastal water, the classification of water bodies as normal is determined based on the fluorescence peaks of the spectral data, by the methodology proposed by Dongzhi Zhao [20,26]. Subsequently, the collected spectral data are classified and extracted. In the case of algal bloom water bodies, the algal bloom spectral data are employed, which have been obtained from previous studies within the relevant literature. In addition, laboratory culture measurements for different types of algal blooms are also utilized.

A comparison is made between the difference in remote sensing reflectance of the redlight portion and the remote sensing reflectance at the 675 nm wavelength position. This provides a foundation for determining whether it aligns with the characteristics of a typical water body. A positive ratio indicates the presence of a normal water spectrum, whereas a negative ratio indicates the presence of an algal bloom.

2.3.2. Chromatic Indices Unified to the Wavelength Range of 360–830 nm

The International Commission on Illumination (CIE) has developed a universally recognized objective system of colorimetry that enables the derivation of Y, representing luminance or brightness, and two chromaticity parameters, x and y, representing hue and saturation, respectively [55]. The system is based on color-matching functions, also known as tristimulus functions, which have been derived for an average human subject and are considered to be reasonably accurate and reproducible. The calculation of XYZ is achieved through the utilization of disparate methodologies, contingent upon the spectral range of the dataset in question [56,57]. To ensure conformity with the wavelength range of the CMF and the range of colors perceptible to the human eye, the wavelength range was standardized to 360–830 nm to extract chromaticity parameters. Standard chromaticity techniques [57] were employed to partition Rrs into chromaticity coordinates X, Y, and Z. The aforementioned coordinates were employed to ascertain the chromatic indices of algal bloom water and normal waters, specifically the hue angle, saturation (S), and dominant wavelength (λd) (calculated by the look up table in Appendix A, Table A1). Furthermore, the apparent visual wavelength (AVW), which represents a one-dimensional geophysical metric of color and is inherently correlated with spectral shape, was also employed in this study [33].

Given that the tristimulus value (XYZ) of the water spectra has already been obtained, it is normalized and expressed in chromaticity coordinates (

C h r_{x}, C h r_{y}

C h r_{x} = \frac{X}{X + Y + Z} \begin{matrix} , & C h r_{y} = \frac{Y}{X + Y + Z} \end{matrix}

(1)

The white point has the coordinates

x_{w} = y_{w} = 1 / 3

. In the (

C h r_{x}, C h r_{y}

) chromaticity plane, the coordinates are then transformed into polar coordinates concerning the white point (

C h r_{x} - x_{w}, C h r_{y} - y_{w}

). This transformation allows us to derive the hue angle (a). The white point W represents the achromatic color or “white point” for a ‘white’ upwelling radiance spectrum. The hue angle (a) is measured between the vector to a point with coordinates (

C h r_{x} - x_{w}, C h r_{y} - y_{w}

) and the positive x-axis (at

C h r_{y} - y_{w} = 0

), with higher angles indicating an anti-clockwise direction (refer to Figure 2 in [58]). It is important to note that our definition of (a) differs from that of Wernand et al. [56,58].

a = 270 - \arctan (C h r_{x} - x_{w} / C h r_{y} - y_{w}) * 180 / π

(2)

As shown in Figure 1, in the paper we defined the hue angle (a) that lies between the vector to a white point with coordinates (

C h r_{x} - x_{w}, C h r_{y} - y_{w}

) and the negative y-axis (at

C h r_{x} - x_{w}

To address the inconsistency in the wavelength ranges of the spectral data employed for analysis, a linear model calibration was conducted on XYZ. The calibration was conducted within the wavelength ranges of 390–830 nm and 400–830 nm for XYZ.

2.3.3. Apparent Visual Wavelength (AVW)

The apparent visual wavelength (AVW) is a one-dimensional geophysical metric that is inherently correlated to the spectral shape of Rrs [33] and has been applied in the construction of an optical water classification index [34]. By employing the full visible-range spectrum (400–700 nm) in the AVW calculation, this product guarantees the incorporation of any diagnostic signals present in the Rrs signal. Furthermore, it offers the possibility of describing and analyzing spectral trends in Rrs using a single variable. The algorithm calculates the weighted harmonic mean of the remote sensing reflectance (Rrs) wavelengths and subsequently outputs the AVW in units of nanometers (nm).

The AVW algorithm has been enhanced by extending the wavelength range of the input spectral data to encompass the range of 360–830 nm. This enhanced iteration is henceforth designated the improved apparent visual wavelength (IAVW).

To address the inconsistency in the wavelength ranges of the spectral data employed for analysis, a linear model calibration was conducted on IAVW. The calibration was conducted within the wavelength ranges of 390–830 nm and 400–830 nm for IAVW.

2.3.4. Statistical Analysis Methods

To ascertain the significance of the data, this study employs a one-sample t-test as part of its statistical analysis. The mean values of the chromatic parameters for normal water are employed as the standard mean, and the mean values of the chromatic parameters for algal bloom water are tested against this standard. Furthermore, exponential, polynomial, and linear function fitting methods are employed to align the functions with the scatter plots.

3. Result

3.1. The Constructed Dataset of Normal Water and Algal Bloom Water

By the aforementioned spectral data reconstruction rules, a total of 9595 spectral curves were obtained. Figure 2 illustrates the global distribution of constructed normal water bodies and algal bloom water bodies, with detailed information provided in Table 2. Furthermore, Table 3 presents comprehensive data on the 21 distinct types of algae or macro spectral curves that were collected. The spectral libraries for both types are presented in Figure 3.

3.2. Wavelength Range Unification Impact to XYZ, Hue Angle, Saturation, λd

The application of scatter plots of chromaticity parameters calculated for varying ranges revealed that the discrepancies in hue angle, saturation, and λd values derived from different wavelength ranges based on tristimulus values are insignificant and can be disregarded (Figure 4).

3.3. Wavelength Range Unification Impact on AVW

The application of scatter plots of AVW calculated for varying ranges revealed a notable degree of variation in AVW values across distinct wavelength ranges (Figure 5).

The unequivocal definition of the full spectrum of Rrs(λ) in terms of a single number permits the examination of trends in spectral shape in both the spatial and temporal domains. Given the nature of the calculation, any discrepancies in the spectra are promptly reflected as a change in the weighted mean. Consequently, AVW can be employed as a straightforward diagnostic instrument for the surveillance of the directionality and magnitude of shifts in the spectral time series in spectrally stable aquatic environments [33]. Initially, AVW was employed in the wavelength range of 400–700 nm. Nevertheless, subsequent calculations were conducted over the wavelength range of 400–800 nm, to reduce the impact of noise and artifacts present in the UV and NIR ends of the spectrum. This approach was employed in the optical classification study of an urbanized estuary using hyperspectral remote sensing reflectance, as described by Turner [34] et al. In this article, the spectral calculation range was extended to 360–830 nm in the ultraviolet and infrared ends, respectively, to maintain consistency with the spectral range of CMF and to ensure adaptability to the bio-optical properties of algal bloom water.

3.4. The Chromatic Indices of Normal Water and Algal Bloom Waters

The chromatic diagram of normal water bodies at a global scale in the CIE 1931 chromatic space exhibits a distribution that resembles the shape of an eyebrow, as depicted in Figure 6a. The specific ranges for the water chromatic indices (hereafter abbreviated as WCIs) are as follows: Hue angle (HUE). The range is 53.2 to 232.0 degrees, with a value of 0.013 to 0.20 for S and a dominant wavelength of λd. 481.3–581.1 nm; improved apparent visual wavelength (IAVW). The range in question is 444.3–638.3 nm. The respective histograms for each index are represented by the blue bars in Figure 7.

In contrast, the chromatic diagram of algal bloom waters in the CIE 1931 chromaticity space tends to cluster in the yellow-green domain, as illustrated in Figure 6b. The particular ranges for the WCIs in algal bloom waters are as follows: Hue angle (HUE). The range is 56.1 to 241.1 degrees, with a value of 0.014 to 0.192 for the coefficient of scattering (S), and the dominant wavelength (λd) is 482.5–585.4 nm; improved apparent visual wavelength (IAVW). The range in question is 512.3–742.9 nm. The histogram for each index is represented by the red bars in Figure 7.

Figure 8a–e depict the correlation between the indices of WCIs for typical water bodies. Similarly, the scatter plot in Figure 8f–j demonstrates the relationship between the indices of WCIs for waters exhibiting algal blooms.

Figure 8a illustrates a gradual decline in the saturation (S) of normal water bodies, ranging from 0.23 to 0.1–0.025, as the hue angle increases from 25 degrees to 120 degrees. This is evident in the analysis of Figure 8a. Subsequently, as the hue angle continues to increase from 120 degrees to 240 degrees, the saturation (S) gradually increases to 0.05–0.15. Similarly, Figure 8e of Figure 8 illustrates that the saturation (S) of water bodies experiencing algal blooms increases from 0.01 to 0.16 as the hue angle rises gradually from 150 degrees to 220 degrees.

Figure 8b reveals a clear correlation between the hue angle of normal water bodies and the rise in improved apparent visual wavelength (IAVW). The data demonstrate a gradual increase in hue angle, ranging from 40 degrees to 230 degrees. It is noteworthy that the relationship between the hue angle and IAVW exhibits a sigmoid function shape. In contrast, water bodies experiencing algal blooms display concentrated distributions of hue and IAVW within the intervals of 160–210 degrees and 550–630 nm, respectively. It is noteworthy that certain data points representing algal species fall outside the specified range in Figure 8f.

Figure 8c reveals a comparable relationship between saturation and dominant wavelength in normal water bodies, analogous to the relationship between saturation and hue angle. This relationship demonstrates an initial decrease, followed by an increase, with an increase in the dominant wavelength. Furthermore, there is a gradual decrease in saturation from 0.28 to a range of 0.02–0.1 as the hue angle decreases, followed by an increase to a range of 0.05–0.13. In contrast, the saturation and dominant wavelength for water bodies affected by algal blooms are found within the ranges of 480–570 nm and 0.02–0.15, respectively. Furthermore, there is a gradual increase in saturation with the rise in dominant wavelength, which ranges from 0.02 to 0.15 in Figure 8g.

Figure 8d reveals a striking resemblance between the relationship between improved apparent visual wavelength (IAVW) and dominant wavelength in normal water bodies and that between IAVW and hue angle. In general, the dataset exhibits a sigmoid function shape, with IAVW gradually increasing from 420 nm to 650 nm, while the dominant wavelength rises from 470 nm to 580 nm. Furthermore, in water bodies affected by algal blooms, the dominant wavelength and IAVW are observed to be concentrated within the ranges of 480 nm to 580 nm and 520 nm to 650 nm, respectively, in Figure 8h.

Figure 8e illustrates that the range of hue angle (HUE) and dominant wavelength (λd) for normal water bodies is observed to span from 50 to 250 degrees and from 480 to 580 nm, respectively. In contrast, the range of hue angle (HUE) and dominant wavelength (λd) for water bodies affected by algal blooms is predominantly observed to span from 150 to 230 degrees and from 540 to 590 nm, respectively, in Figure 8i.

A one-sample t-test was employed to ascertain significant differences in four chromaticity parameters between normal and algal bloom waters. The mean chromaticity parameters of normal waters were employed as the test values for comparison with the mean values of algal bloom waters. The t-statistics for IAVW, H, S, and λd were 139.268 nm, 150.157 degrees, 24.657, and 204.378 nm, respectively. At the 0.05 level of significance, the chromaticity parameters of algal bloom waters were found to differ significantly from those of normal waters, thereby confirming the existence of notable differences in chromaticity parameters between the two water types.

3.5. The Chromatic Indices of the Different Algal Species

The diversity of phytoplankton and macroalgae in aquatic ecosystems gives rise to color variations in algal bloom waters. In this section, spectral data for algal blooms of 21 different species with varying chlorophyll concentrations are presented in Table 3. The discrepancies in their Water Chromatic Indices (WCIs) are elucidated from a colorimetric standpoint, thereby facilitating the identification and differentiation of algae species.

The article presents the spectra of different algae, namely Platymonas sp., Ceratium fura sp., Dinoflagellates, Gymmodinium sp., Pyramimonas sp., Chlorella sp., Nitzschia closterium, Noctiluca scintillans, Coscinodiscus Concinnus, Skeletonema costatum, Spirulina, Alexandrium, Chaetoceros, Heterosigma akashiwo, Aureococcus anophagefferens, Dicrateria zhanjiangensis Hu., and Marine Cyanobacteria. Figure 9 depicts the chromatic diagrams of various algal species. Figure 10a–d illustrate the scatter plot of WCIs between the aforementioned species, while Figure 10e–h depict the range of WCIs for each species. Table 4 provides comprehensive data on the range of WCIs for various algal species. The results demonstrate that WCIs vary significantly across different algal species and do not exhibit chromatic characteristics typical of normal water.

3.6. The Chromatic Indices of the Same Algae with Different Chlorophyll Concentrations

The OC algorithm, which is typically employed in the context of normal water, is wholly inapplicable when attempting to detect algal blooms with elevated chlorophyll concentrations. Conversely, the fluorescence height method is ineffective when applied to species-specific algal blooms. As discussed in the preceding section, WCIs are more effective when used to quantify the chromatic characteristics and movement of color shifts in the CIE chromatic coordinates. It is of great importance to develop an effective relationship between chlorophyll a and WCIs, as this will facilitate the establishment of chromatic methods for the discrimination of algal blooms using remote sensing, which is currently a challenging task.

In this article, we employ a subset of dinoflagellates data (see Figure 11b) [39] to examine the relationship between spectral curves and chromaticity parameters derived from measurements of varying chlorophyll concentrations within the same algae species. The scatter plots of hue angle and λd vs. chlorophyll exhibit no statistically significant differences (see Figure 12b,c). However, a common trend is observed: when the hue angle exceeds 190° (indicated by the red vertical line in Figure 12) and λd is greater than 560 nm (also indicated by the red vertical line in Figure 12), an exponential upward trend is observed in the relationship between these two parameters and chlorophyll concentration. Conversely, the correlation coefficient between S and chlorophyll is markedly weak, with an R2 value of 0.083.

The AVW offers a distinctive degree of flexibility in integrating the influence of UV and NIR into the classification and indexing of spectral signatures. However, its utility is primarily evident in the categorization of regions exhibiting a pronounced CDOM contribution and the impact of NIR on the red-edge phenomenon, such as the occurrence of harmful algae blooms and the formation of dense sediment plumes [33]. In the context of algal bloom waters, the correlation coefficient between IAVW and chlorophyll concentration, calculated over the 360–830 nm range, is notably high, achieving an R² value of 0.53. Figure 11a’s three-dimensional plot provides further insight into the relationship between chlorophyll concentration, IAVW, and λd, thereby demonstrating the efficacy of these watercolor indices in differentiating algal bloom waters from normal waters.

4. Discussion

4.1. IAVW (360–830 nm) and AVW (400–700 nm)

The results of the three data subsets, comprising normal water, algal bloom water, and a combination of both, were also employed in the calculations, as illustrated in Figure 13a–c. The R² values for IAVW and AVW in the normal water data subset were found to be 0.95. However, the dynamic range of the two differed significantly. The minimum value of AVW was 479.6482, the maximum value was 600.1183, and the average value was 531.2516. Conversely, the minimum value of IAVW was 444.2851, the maximum value was 638.293, and the average value was 528.0795. The correlation coefficient for the subset of algal bloom water exhibited the lowest correlation, with a maximum value of AVW at 629.9297, a minimum value of 514.8006, and an average value of 555.0417. Meanwhile, the maximum value of IAVW was 742.8639, the minimum value was 512.336, and the average value was 584.1249.

The correlation coefficient of the two sets has been enhanced, with AVW exhibiting a maximum value of 629.9297, a minimum value of 479.6482, and an average value of 537.6301. Similarly, IAVW has a maximum value of 742.8639, a minimum value of 444.2851, and an average value of 543.1062. The results indicate that AVW is more suitable for Class I and Class II water bodies. This is because the calculation does not employ the spectral data of the 360–400 nm and 700–830 nm ranges within the shortwave band, which may result in notable discrepancies when dealing with atypical water bodies, such as those impacted by algal blooms. The enhanced AVW (IAVW) displays a more expansive dynamic range and more accurately depicts the spectral attributes of both typical and algal bloom waters. Furthermore, the discrepancy between the three data subsets (IAVW minus AVW) produces disparate outcomes, which are depicted in Figure 13d–f.

4.2. Case Analysis

The chromatic indices that are characteristic of normal and algal bloom waters, derived from global spectral data, encompass a portion of the same area. Consequently, this section employs two particular case studies to distinguish between normal and algal bloom waters. To illustrate the methodology, the Bohai Sea is used as a case study for coastal waters, while Taihu Lake is used as a case study for inland waters.

4.2.1. Case of Coastal Water in the Bohai Sea

The data points are illustrated in Figure 2b, and the detailed information can be seen in Table 5. The Bohai Sea is susceptible to the formation of red tides, with its water body falling under the second category. The WCIs results, calculated from the spectral data gathered in the field, indicate that the normal water body is situated within the blue-green-yellow domain of the chromaticity space. The hue angle ranges from 63.69 to 219.45 degrees, the S value from 0.0454 to 0.1787, the λd from 485.1 nm to 575.7 nm, and the IAVW from 456.81 nm to 601.63 nm. The chromatic diagram for the typical water body is presented in Figure 14a, while the corresponding spectra are shown in Figure 14b. Four species of algae have been identified among the bloom-forming species. One of the identified species is an unidentified algae, which can be distinguished from the in situ measured spectra using the fluorescence peaks of the spectral data. The other three species are Noctiluca scintillans, Aureococcus anophagefferens, and Ceratium fura sp. The spectra of these algae species are illustrated in Figure 14c. They are situated within the yellow-green domain of the chromatic diagram, exhibiting hue angles between 151.6 and 213.3 degrees, S values between 0.0381 and 0.1086, λd between 518.7 nm and 573.0 nm, and IAVW between 528.7 nm and 613.3 nm. The chromatic diagram is presented in Figure 14a, while the corresponding spectra are provided in Figure 14c. Figure 15e–h provide detailed information regarding WCIs in normal water and algal bloom water.

As illustrated in the chromatic diagram of the two categories of water, the intersection occurs predominantly within the green and yellow light domains. This can be attributed to the fact that water in the Bohai Sea belonging to Case 2 is predominantly discharged with a high concentration of sediment by rivers, which results in the water appearing reddish in color. However, the scatter plot of hue and IAVW in Figure 15a demonstrates that the chromatic diagram of normal water follows a typical sigmoid function (illustrated by the blue dots). Although the chromaticity angle employed in this article is calculated in a clockwise direction, the resulting value is analogous to that obtained by Vandermeulen [33] et al. Conversely, the red tide dots on the chromatic diagram are situated towards the lower end of the distribution curve (illustrated by the red dots). Similarly, the scatter plots of λd and IAVW exhibit a comparable pattern. Nevertheless, it is challenging to differentiate between the two in the scatter plots of hue angles vs. S and S vs. λd.

The parameter λd is employed to quantify the chromatic characteristics of snowmelt and glacial water, spanning the wavelength range of 480–550 nm (corresponding to the blue-turquoise-green spectrum). It is only the water introduced by snowmelt or groundwater that falls within the restricted range of 550–570 nm, encompassing green, yellowish, and brownish hues [29]. John R. Gardner employed λd as an indicator of river color, which exhibits distinct seasonal patterns that are synchronous with flow regimes and has demonstrated significant color shifts over the past 35 years. This provides the inaugural map of river color and new insights into the macrosystem ecology of river. However, the use of a single colorimetric index often proves inadequate for distinguishing internal differences within the same water. The combination of these two indices among the WCIs can markedly enhance the discrimination of the constituents in the waters (see Figure 15a,b).

From the fitted curves of IAVW, chromaticity angle H, and λd for normal and algal bloom waters (Figure 15a,b), it is evident that there is a clear boundary between the two sets of curves. This indicates a clear distinction in chromaticity parameters between normal and algal bloom waters in the Bohai Sea. Furthermore, Figure 15c,d demonstrate that for normal waters, the fitted curves of chromaticity angle H, λd, and saturation exhibit a positive correlation, whereas, for algal bloom waters, these parameters display a negative correlation. These notable discrepancies indicate that chromaticity parameters are capable of accurately differentiating between normal and algal bloom waters in the Bohai Sea.

4.2.2. Case of Inland Water in Taihu Lake

The comprehensive spectral data of Taihu Lake are presented in Table 6 of this research. From the chromatic parameters of the water bodies in Taihu Lake, it can be discerned that there is a significant distinction between normal and algal blooms, particularly in hue angle, IAVW, and dominant wavelength (λd). As illustrated in Figure 16a, the red points may be taken to indicate the region in which an algal bloom is occurring. The blue points may be considered indicative of normal water bodies. From Figure 17e–h, it can be observed that the ranges of the chromatic parameters of normal and algal bloom water bodies are as follows: the IAVW ranges of normal water bodies are from 510 nm to 610 nm, while it is from 550 nm to 750 nm for algal bloom water bodies; the hue angle of normal water bodies is from 130 degrees to 205 degrees, while it is from 170 degrees to 210 degrees for algal bloom bodies; and the saturation (S) of normal water bodies is from 0. The values for the normal water bodies range from 0.04 to 0.12, while those for the algal blooms range from 0.03 to 0.18. The dominant wavelength (λd) for normal water bodies is between 510 and 570 nanometers, while that for the algal blooms is between 550 and 570 nanometers. It can be observed that Figure 17a–d exhibit superior discrimination between the normal water bodies and the algal bloom water bodies about the AVW, hue angle, and λd in Taihu Lake.

Figure 17a–d illustrate the scatter plot fitted curves of chromaticity parameters for normal and algal bloom waters in Taihu Lake, exhibiting patterns analogous to those observed in the Bohai Sea. This indicates that chromaticity parameters are an effective means of distinguishing between normal and algal bloom waters in Taihu.

5. Conclusions

By the tenets of colorimetry, the existing colorimetric index has been augmented with the incorporation of the dominant wavelength AVW, thereby enhancing the calculation conditions. This has led to the creation of a comprehensive set of water chromatic indices (WCIs) that are both unified and versatile. The WCIs comprise hue, S, λd, and IAVW. The integration of the colorimetric visual matching function (CMF) and the spectral response of the water constituents of algal blooms in the ultraviolet and near-infrared bands were considered. The WCIs are calculated over the spectral range of 360 to 830 nm. The use of water chromatic indices is of particular benefit when differentiating between algal blooms. While an expansion of the wavelength range has a limited impact on chromaticity parameters, it notably influences IAVW, resulting in significant effects.

At the global level, there are inconsistencies in the chromatic classification of normal and algal bloom waters, with some degree of overlap observed in certain regions. In specific bodies of water such as Taihu Lake and the Bohai Sea, waters exhibiting algal blooms display notable differences from background waters, rendering them readily distinguishable. The application of chromatic parameters allows for effective differentiation in such cases.

In local water bodies such as Taihu Lake and the Bohai Sea, algal bloom waters show clear distinctions from background waters, allowing for effective separability. These chromatic parameters can be leveraged for precise differentiation.

The chromatic differences observed in different algal bloom species, including those in inland waters, marine red tides, and macroalgae, were significant. To quantify the distribution of these phenomena, WCIs (water chromatic indices) were employed to characterize their position within a chromatic coordinate system. Notable differences in chromaticity parameters have been observed among various algal species, which may serve as a basis for species identification. Furthermore, within a single algal species, variations in chromaticity are observed at different concentrations. Furthermore, these differences in chromaticity are closely associated with cell counts.

In the chromatic coordinate, the concentration of chlorophyll was found to covary not only with hue but also with S. The results of the study of dinoflagellates demonstrated that hue and λd (another parameter) increased exponentially with an increase in chlorophyll concentration. Furthermore, the correlation between S and chlorophyll was found to be relatively weak, whereas a different parameter, IAVW, demonstrated a strong linear relationship. Nevertheless, this type of relationship displayed evident species-specific characteristics among algae.

Despite the collection of 9595 spectral data points, the paucity of hyperspectral datasets that meet the requisite standards, particularly those regarding bio-optical datasets of algal bloom water bodies, represents a significant challenge. It is therefore necessary to further expand these datasets.

In situ monitoring offers insights into the spectral alterations and chromaticity parameter threshold fluctuations of particular algal blooms (red tides) throughout their growth and decay cycles. Integrating this method into satellite imagery allows for the assessment of its efficacy and introduces a novel approach for the utilization of satellite observations in the detection of algal bloom distribution and the identification of algal species.

Author Contributions

D.Z.: Conceptualization, Methodology, Data curation, Writing, Validation, Formal analysis. Q.L.: Data curation, Investigation, Writing, Software, Writing, Formal analysis. Z.Q.: Methodology, Investigation, Supervision, Funding acquisition, Formal analysis. All authors have read and agreed to the published version of the manuscript.

Funding

This research was funded by the National Natural Science Foundation of Zhongfeng Qiu grant number [41976165]. And The APC was funded by [41976165].

Data Availability Statement

Data will be made available on request.

Acknowledgments

Thanks to Jun Chen, Tingwei Cui, Hongtao Duan, Chuanming Hu, Ronghua Ma, Qianguo Xing, Yunlin Zhang, and Deyu An for providing in situ observations.

Conflicts of Interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Appendix A

Table A1. A lookup table of the dominant wavelength λd.

Dominate Wavelength (λd)	X	Y	Z	Hue Angle
360	0.0001299	3.920 × 10⁻⁶	0.0006061	25.686
360.1	0.0001314	3.963 × 10⁻⁶	0.0006132	25.687
360.2	0.0001329	4.007 × 10⁻⁶	0.0006203	25.688
360.3	0.0001345	4.053 × 10⁻⁶	0.0006276	25.6883
360.4	0.0001361	4.098 × 10⁻⁶	0.0006349	25.689
360.5	0.0001376	4.1450 × 10⁻⁶	0.0006424	25.690
360.6	0.0001392	4.193 × 10⁻⁶	0.0006499	25.691
…	…	…	…	…
829.7	1.2747 × 10⁻⁶	4.618 × 10⁻⁷	1.801 × 10⁻¹⁰⁹	279.546
829.8	1.266 × 10⁻⁶	4.585 × 10⁻⁷	1.453 × 10⁻¹⁰⁹	279.556
829.9	1.258 × 10⁻⁶	4.553 × 10⁻⁷	8.628 × 10⁻¹¹⁰	279.570

References

Zingone, A.; Enevoldsen, H.O. The diversity of harmful algal blooms: A challenge for science and management. Ocean. Coast. Manag. 2000, 43, 725–748. [Google Scholar] [CrossRef]
Zhao, D.; Wen, S.; Song, L. Red Tide Disaster Risk Assessment Theory and Zoning Method; The Ocean Press: Beijing, China, 2013; Volume 5. [Google Scholar]
Zhao, D. The Occurrence Law of Red Tide Disasters in Typical Sea Areas of China; Ocean Press: Beijing, China, 2010. [Google Scholar]
Zhao, D.Z.; Zhao, L.; Zhang, F.S.; Zhang, X.Y. Temporal Occurrence and Spatial Distribution of Red Tide Events in China’s Coastal Waters. Hum. Ecol. Risk Assess. 2004, 10, 945–957. [Google Scholar] [CrossRef]
Huisman, J.; Codd, G.A.; Paerl, H.W.; Ibelings, B.W.; Verspagen, J.M.; Visser, P.M. Cyanobacterial blooms. Nat. Rev. Microbiol. 2018, 16, 471–483. [Google Scholar] [CrossRef] [PubMed]
Liu, J.L.; Zhao, X.H.; Kang, X.Y.; Zhuang, M.M.; Ding, X.W.; Zhao, L.J.; Wen, Q.L.; Zhu, Y.; Gu, K.; Bao, Q.J.; et al. Good news: We can identify Ulva species that erupted in the Yellow Sea more easily and cheaply now. Conserv. Genet. Resour. 2020, 12, 447–449. [Google Scholar] [CrossRef]
Gobler, C.J.; Sunda, W.G. Brown Tides. In Ecology of Harmful Algae; Springer: Berlin/Heidelberg, Germany, 2006; pp. 111–123. [Google Scholar] [CrossRef]
Gobler, C.J.; Sunda, W.G. Ecosystem disruptive algal blooms of the brown tide species, Aureococcus anophagefferens and Aureoumbra lagunensis. Harmful Algae 2012, 14, 36–45. [Google Scholar] [CrossRef]
Zhang, Q.C.; Qiu, L.M.; Yu, R.C.; Kong, F.Z.; Wang, Y.F.; Yan, T.; Gobler, C.J.; Zhou, M.J. Emergence of brown tides caused by Aureococcus anophagefferens Hargraves et Sieburth in China. Harmful Algae 2012, 19, 117–124. [Google Scholar] [CrossRef]
Zhuang, M.M.; Liu, J.L.; Ding, X.W.; He, J.Z.; Zhao, S.; Wu, L.J.; Gao, S.; Zhao, C.Y.; Liu, D.Y.; Zhang, J.H.; et al. Sargassum blooms in the East China Sea and Yellow Sea: Formation and management. Mar. Pollut. Bull. 2021, 162, 111845. [Google Scholar] [CrossRef]
Anderson, D.M.; Fensin, E.; Gobler, C.J.; Hoeglund, A.E.; Hubbard, K.A.; Kulis, D.M.; Landsberg, J.H.; Lefebvre, K.A.; Provoost, P.; Richlen, M.L.; et al. Marine harmful algal blooms (HABs) in the United States: History, current status and future trends. Harmful Algae 2021, 102, 101975. [Google Scholar] [CrossRef]
Pinkerton, M.H.; Richardson, K.M.; Boyd, P.W.; Gall, M.P.; Zeldis, J.; Oliver, M.D.; Murphy, R.J. Intercomparison of ocean colour band-ratio algorithms for chlorophyll concentration in the Subtropical Front east of New Zealand. Remote Sens. Environ. 2005, 97, 382–402. [Google Scholar] [CrossRef]
Guo, Y.; Li, H.; Wei, X.; Huang, Z.; Xue, K.; Ma, R.; Hu, Z. Uncertainty evaluation of water color parameters inversion in optically complex lakes. Environ. Sci. Technol. 2023, 46, 15–26. [Google Scholar]
O’Reilly, J.E.; Werdell, P.J. Chlorophyll Algorithms for Ocean Color Sensors—Oc4, Oc5 & Oc6. Remote Sens. Environ. 2019, 229, 32–47. [Google Scholar] [CrossRef] [PubMed]
Merzlyak, M.N.; Chivkunova, O.B.; Maslova, I.P.; Naqvi, K.R.; Solovchenko, A.E.; Klyachko-Gurvich, G.L. Light absorption and scattering by cell suspensions of some cyanobacteria and microalgae. Russ. J. Plant Physiol. 2008, 55, 420–425. [Google Scholar] [CrossRef]
Wang, M.J.; Wang, Z.Y.; Huang, C.J. Scattering Characteristics of Marine Mixed Suspended Particles to Blue and Green Lasers. Spectrosc. Spectr. Anal. 2022, 42, 1749–1754. [Google Scholar] [CrossRef]
Wang, G.; Moisan, J. Remote Sensing of Phytoplankton Pigments. In Plankton Communities; IntechOpen: London, UK, 2022. [Google Scholar]
Le, C.F.; Li, Y.M.; Zha, Y.; Sun, D.Y.; Huang, C.C.; Lu, H. A four-band semi-analytical model for estimating chlorophyll a in highly turbid lakes: The case of Taihu Lake, China. Remote Sens. Environ. 2009, 113, 1175–1182. [Google Scholar] [CrossRef]
Luo, J.; Qin, L.; Mao, P.; Xiong, Y.; Zhao, W.; Gao, H.; Qiu, G. Research Progress in the Retrieval Algorithms for Chlorophyll-a, a Key Element of Water Quality Monitoring by Remote Sensing. Remote Sens. Technol. Appl. 2021, 36, 473–488. [Google Scholar]
Zhao, D.; Zhang, F.; Du, F.; Zhao, L.; Guo, H. Study on Sun-Induced Chlorophyll Fluorescence Peaks (SICF) Characteristics in Water with Different Algal Species. J. Remote Sens. 2005, 9, 265–270. [Google Scholar]
Gower, J.F.R. A simpler picture of satellite chlorophyll fluorescence. Remote Sens. Lett. 2014, 5, 583–589. [Google Scholar] [CrossRef]
Lavigne, H.; Ruddick, K.; Vanhellemont, Q. Monitoring of high biomass Phaeocystis globosa blooms in the Southern North Sea by in situ and future spaceborne hyperspectral radiometry. Remote Sens. Environ. 2022, 282, 113270. [Google Scholar] [CrossRef]
Li, X.; Huang, P.; Sun, W.; Fan, C.; You, L. Study of Chlorophyll-a Inversion Algorithm for Hengshan Reservoir Using Sentinel-2 Images. Hydro Sci. Cold Zone Eng. 2023, 6, 93–96. [Google Scholar]
Moses, W.J.; Gitelson, A.A.; Berdnikov, S.; Povazhnyy, V. Estimation of chlorophyll-a concentration in case II waters using MODIS and MERIS data-successes and challenges. Environ. Res. Lett. 2009, 4, 045005. [Google Scholar] [CrossRef]
Zhao, D.; Yang, J.; Zhao, L.; Chen, Y.; Wang, L.; Shen, H. MODIS Bio-optical algorithms in Liaodong Bay(I) Algorithms assessment. Mar. Environ. Sci. 2007, 26, 401–407. [Google Scholar]
Zhao, D.; Zhang, F.; Yang, J.; Guo, H.; Zhao, L. Band Optimization for Characterizing the Sun-Induced Chlorophyll Fluorescence Height in Red Tide Waters. Acta Oceanol. Sin. 2005, 27, 146–153. [Google Scholar]
Wynne, T.; Meredith, A.; Briggs, T.; Litaker, W.; Stumpf, R. Harmful Algal Bloom Forecasting Branch Ocean Color Satellite Imagery Processing Guidelines; National Oceanic and Atmospheric Administration: Washington, DC, USA, 2018; pp. 252–296.
Bukata, R.P.; Pozdnyakov, D.V.; Jerome, J.H.; Tanis, F.J. Validation of a radiometric color model applicable to optically complex water bodies. Remote Sens. Environ. 2001, 77, 165–172. [Google Scholar] [CrossRef]
Wernand, M.R.; Van Der Woerd, H.J. Spectral analysis of the Forel-Ule ocean colour comparator scale. J. Eur. Opt. Soc. Rapid Publ. 2010, 5, 10014s. [Google Scholar] [CrossRef]
Novoa, S.; Wernand, M.R.; Van der Woerd, H.J. The Forel-Ule scale revisited spectrally: Preparation protocol, transmission measurements, and chromaticity. J. Eur. Opt. Soc. Rapid Publ. 2013, 8, 13057. [Google Scholar] [CrossRef]
Vandermeulen, R.A.; Mannino, A.; Craig, S.E.; Werdell, P.J. 150 shades of green: Using the full spectrum of remote sensing reflectance to elucidate color shifts in the ocean. Remote Sens. Environ. 2020, 247, 111900. [Google Scholar] [CrossRef]
Turner, K.J.; Tzortziou, M.; Grunert, B.K.; Goes, J.; Sherman, J. Optical classification of an urbanized estuary using hyperspectral remote sensing reflectance. Opt. Express 2022, 30, 41590–41612. [Google Scholar] [CrossRef] [PubMed]
Stuart, V. Remote Sensing of Ocean Colour in Coastal, and Other Optically-Complex, Waters; International Ocean Colour Coordinating Group (IOCCG): Dartmouth, NS, Canada, 2000.
Lehmann, M.K.; Gurlin, D.; Pahlevan, N.; Alikas, K. GLORIA—A globally representative hyperspectral in situ dataset for optical sensing of water quality. Sci. Data 2023, 10, 100, Erratum in Sci. Data. 2023, 10, 191. [Google Scholar] [CrossRef]
Lavigne, H.; Dogliotti, A.; Doxaran, D.; Shen, F.; Castagna, A.; Beck, M.; Vanhellemont, Q.; Sun, X.; Gossn, J.I.; Renosh, P.R.; et al. The HYPERMAQ dataset: Bio-optical properties of moderately to extremely turbid waters. Earth Syst. Sci. Data 2022, 14, 4935–4947. [Google Scholar] [CrossRef]
Knaeps, E.; Doxaran, D.; Dogliotti, A.; Nechad, B.; Ruddick, K.; Raymaekers, D.; Sterckx, S. The SeaSWIR dataset. Earth Syst. Sci. Data 2018, 10, 1439–1449. [Google Scholar] [CrossRef]
Maier, P.M.; Keller, S. Estimating Chlorophyll a Concentrations of Several Inland Waters with Hyperspectral Data and Machine Learning Models. ISPRS Ann. Photogramm. Remote Sens. Spat. Inf. Sci. 2019, IV-2/W5, 609–614. [Google Scholar] [CrossRef]
Garaba, S.; Wernand, M.; Zielinski, O. Quality control of automated hyperspectral remote sensing measurements from a seaborne platform. Ocean. Sci. Discuss. 2011, 8, 613–638. [Google Scholar]
Legleiter, C.J.; King, T.V.; Carpenter, K.D.; Hall, N.C.; Mumford, A.C.; Slonecker, T.; Graham, J.L.; Stengel, V.G.; Simon, N.; Rosen, B.H. Spectral mixture analysis for surveillance of harmful algal blooms (SMASH): A field-, laboratory-, and satellite-based approach to identifying cyanobacteria genera from remotely sensed data. Remote Sens. Environ. 2022, 279, 113089. [Google Scholar] [CrossRef]
Castagna, A.; Amadei Martínez, L.; Bogorad, M.; Daveloose, I.; Dasseville, R.; Dierssen, H.M.; Beck, M.; Mortelmans, J.; Lavigne, H.; Dogliotti, A.; et al. Optical and biogeochemical properties of diverse Belgian inland and coastal waters. Earth Syst. Sci. Data 2022, 14, 2697–2719. [Google Scholar] [CrossRef]
Tilstone, G.H.; Pardo, S.; Simis, S.G.; Qin, P.; Selmes, N.; Dessailly, D.; Kwiatkowska, E. Consistency between Satellite Ocean Colour Products under High Coloured Dissolved Organic Matter Absorption in the Baltic Sea. Remote Sens. 2021, 14, 89. [Google Scholar] [CrossRef]
Zhao, D.; Liu, Y. Chinese Polluted Water Body Spectral Characteristics; The Ocean Press: Beijing, China, 2001. [Google Scholar]
Xing, Q.; Hu, C. Mapping macroalgal blooms in the Yellow Sea and East China Sea using HJ-1 and Landsat data: Application of a virtual baseline reflectance height technique. Remote Sens. Environ. 2016, 178, 113–126. [Google Scholar] [CrossRef]
An, D.; Xing, Q.; Wei, Z.; Li, L. Analysis of Spectral Features of Representative Floating Macroalgae in the Yellow Sea. J. Oceanol. Limnol. 2018, 49, 1054–1060. [Google Scholar]
Zhao, D.; Zhang, F.; Du, F.; Zhao, L.; Guo, H. A Study on relation of chlorophyll-a concentration with the reflectance peak near 700 nm in algae-dominated waters and sensitivity of fluorescence algorithms for detecting algal bloom. Int. J. Remote Sens. 2010, 31, 39–48. [Google Scholar] [CrossRef]
Jiang, L.; Wang, L.; Zhao, D.; Wang, X.; Wang, X. Spectral variability of particulate backscattering of nanoplankton and microplankton—Focusing on skeletonema costatum and prorocentrum micans. Mar. Environ. Sci. 2013, 32, 884–888. [Google Scholar]
Jiang, L.; Wang, L.; Zhao, D.; Wang, X. Study on the Backscattering Characteristics of Amphidinium carterae Hulburt. Spectrosc. Spectr. Anal. 2013, 33, 1892–1896. [Google Scholar]
Jiang, L.; Wang, L.; Zhang, X.; Chen, Y.; Xiong, D. A Semianalytical Model Using MODIS Data to Estimate Cell Density of Red Tide Algae (Aureococcus anophagefferens). Adv. Meteorol. 2016, 2016, 1780986. [Google Scholar] [CrossRef]
Slonecker, T.; Bufford, B.; Graham, J.; Carpenter, K.; Opstal, D.; Simon, N.; Hall, N. Hyperspectral reflectance characteristics of cyanobacteria. Adv. Remote Sens. 2021, 10, 66–77. [Google Scholar] [CrossRef]
Zhao, Y.; Li, K.-Q.; Sun, S.; Chen, K.; Tan, G.-S.; Zhang, J.; Wang, X.-L. Numerical simulation and eutrophication assessment of spatiotemporal distribution of nitrogen, phosphorus, and chlorophyll concentrations in the Bohai Sea. Oceanol. Limnol. Sin. 2023, 55, 118–134. [Google Scholar] [CrossRef]
Wynne, T.T.; Stumpf, R.P.; Litaker, R.W.; Hood, R.R. Cyanobacterial bloom phenology in Saginaw Bay from MODIS and a comparative look with western Lake Erie. Harmful Algae 2021, 103, 101999. [Google Scholar] [CrossRef] [PubMed]
Williamson, S.J.; Cummins, H.Z. Light and Color in Nature and Art; John Wiley and Sons: Hoboken, NJ, USA, 1985; Volume 10, pp. 123–124. [Google Scholar]
Bukata, R.P.; Jerome, J.H.; Kondratyev, A.S.; Pozdnyakov, D.V. Optical Properties and Remote Sensing of Inland and Coastal Waters; CRC Press: Boca Raton, FL, USA, 1995. [Google Scholar]
Ohta, N.; Robertson, A. Colorimetry: Fundamentals and Applications; John Wiley and Sons: Hoboken, NJ, USA, 2005. [Google Scholar]
Woerd, H.J.; Wernand, M.R. True colour classification of natural waters with medium-spectral resolution satellites: SeaWiFS, MODIS, MERIS, and OLCI. Sensors 2015, 15, 25663–25680. [Google Scholar] [CrossRef] [PubMed]
Ohta, N.; Robertson, A. Colorimetry Fundamentals and Application; Noboru Ohta Rochester Institute of Technology: Rochester, NY, USA, 2005. [Google Scholar]
Gardner, J.R.; Yang, X.; Topp, S.N.; Ross, M.R.; Altenau, E.H.; Pavelsky, T.M. The Color of Rivers. Geophys. Res. Lett. 2021, 48, e2020GL088946. [Google Scholar] [CrossRef]
Li, J. MODIS observations of water color of the largest 10 lakes in China between 2000 and 2012. Int. J. Digit. Earth 2016, 9, 788–805. [Google Scholar] [CrossRef]

Figure 1. Schematic diagram of chromatic coordinate.

Figure 2. The global distribution map of spectral libraries for normal water and algal bloom water bodies. (The red dots are algal blooms water, and the blue dots are normal water). ((a) The global scale, (b) the Bohai scale).

Figure 3. The library of spectral data for algal blooms and normal water bodies: ((a) normal water bodies (b) algae water bodies).

Figure 4. The scatter plot of the XYZ_{390–830 nm}, XYZ_{400–830 nm}, and XYZ_{360–830 nm}. ((a–c) is the adjustment scatter plot between XYZ_{390–830 nm} and XYZ_{360–830 nm}, (d–f) is the adjustment scatter plot between XYZ_{400–830 nm} and XYZ_{360–830 nm}).

Figure 5. The adjustment of the IAVW 390–830 nm and IAVW400–830 nm. ((a) The IAVW with the wavelength from 390 to 830 nm, (b) the IAVW with the wavelength from 400 to 830 nm).

Figure 6. The chromatic diagram of the normal water bodies and algae bloom water bodies. ((a) normal waters (b) algae bloom waters).

Figure 7. Normal water bodies and algae bloom water bodies WCIs (Water Chromatic Indices) histogram. ((a) Hue angle, (b) saturation, (c) λd, (d) IAVW. Blue bars are normal water and red bars are algal bloom waters).

Figure 8. Scatter plots of each other among the WCIs (Water Chromatic Indices). (Left column: normal water, scatter plot in blue: (a) saturation (S) vs. hue angle, (b) hue angle vs. IAVW, (c) saturation (S) vs. λd, (d) λd vs. IAVW, (e) saturation (S) vs. IAVW; right column: algal bloom water, scatter plot in red: (f) saturation (S) vs. hue angle, (g) hue angle vs. IAVW, (h) saturation (S) vs. λd, (i) λd vs. IAVW, (j) saturation (S) vs. IAVW.

Figure 9. The chromatic diagram of different algal species in CIE 1931 chromatic coordinate.

Figure 10. (a–d) are the scatter plots of different algal species. ((a) The verses of the IAVW and the hue angle, (b) the verses of the IAVW and the λd, (c) the verses of the S and the hue angle, (d) the verses of the S and the λd). (e–h) are the radar charts for different algal species. ((e) IAVW, (f) hue angle, (g) saturation, (h) λd).

Figure 11. (a) The 3-D (three-dimensional) plot of the Dinoflagellates, chlorophyll a, and the dominant wavelength (λd), IAVW. (b) The spectra of the Dinoflagellates. The colors of the spectral lines stand for the colors of the objects.

Figure 12. The scatter plot of the Dinoflagellates. (The chlorophyll a vs. the (a) IAVW, (b) hue angle, (c) λd, (d) saturation).The red curve is the fitted curve and the blue dotted line is the numerical identification line.

Figure 13. (a–c) are the scatter plots between the AVW400–700 nm and IAVW360–830 nm in normal waters, algae bloom waters, and both of the two types of waters. (d–f) are the histogram of the value of IAVW360–830 nm minus AVW400–700 nm. ((a,d) The normal waters, (b,e) the algae bloom waters, (c,f) both of the two types of the waters).

Figure 14. The color discrimination between normal water and algal bloom water in the Bohai Sea based on spectral wavelength range from 360–830 nm. ((a) The chromatic point of the normal and algae bloom waters in Bohai Sea. (b) The spectra of the normal waters in Bohai Sea. (c) The algae bloom waters in the Bohai Sea. (The color of the curve was standard for the value of λd). The colors of the spectral lines stand for the colors of the objects).

Figure 15. (a–d) are the scatter plot of each other among WCIs of the Bohai Sea. (The red dots are algal bloom water, and the blue dots are normal water). (e–h) are the box plot of chromatic indices calculated by spectral data of the Bohai Sea with different wavelength ranges from 360−830 nm. ((e) the box plot of the IAVW, (f) the box plot of the hue angle, (g) the box plot of the saturation, (h) the box plot of the λd.

Figure 16. The color discrimination between normal water and algal bloom water in Taihu Lake based on spectral wavelength range from 360 to 830 nm. ((a) The chromatic point of the normal and algae bloom waters in Bohai Sea. (b) The spectra of the normal waters in Bohai Sea. (c) The algae bloom waters in the Bohai Sea. (The color of the curve was standard for the value of λd). The colors of the spectral lines stand for the colors of the objects).

Figure 17. (a–d) are the scatter plots of each other among WCIs of Taihu Lake. (The red dots are algal bloom water, and the blue dots are normal water). (e–h) are the box plot of chromatic indices calculated by spectral data of the Bohai Sea with different wavelength ranges from 360 to 830 nm. ((e) The box plot of the IAVW, (f) the box plot of the hue angle, (g) the box plot of the saturation, (h) the box plot of the λd.

Table 1. Detailed information about the Dataset.

No.	The Name of the Spectral Library	Reference	Year	Data Number	Wavelength Range (nm)	Resolution (nm)	Data Source
a	GLORIA	[36]	1990–2022	7572	350–900	1	https://doi.pangaea.de/10.1594/PANGAEA.948492 accessed on 1 June 2023
b	HYPERMAQ	[37]	2022	111	350–900	2.5	https://doi.pangaea.de/10.1594/PANGAEA.944313 accessed on 1 June 2023
c	SeaSWIR	[38]	2012–2013	137,200	350–1300, 350–900	1, 2.5	https://doi.pangaea.de/10.1594/PANGAEA.886287 accessed on 1 June 2023
d	SpecWa	[39]	2018–2019	3685	389.35–910.32	0.74	https://dataservices.gfz-potsdam.de/panmetaworks/showshort.php?id=6800b0c8-dd51-11ea-9603-497c92695674 accessed on 1 June 2023
e	NORCOHAB II	[40]	2009	44	320–950	5	https://doi.pangaea.de/10.1594/PANGAEA.753830?format=html#download accessed on 1 June 2023
f	SMASH	[41]	2020	222	325–1075	1	https://www.sciencebase.gov/catalog/item/5fe38f8ed34ea5387deb4923 accessed on 1 June 2023
g	Belgian inland and coastal waters	[42]	2017–2019	14,220	380–850, 380–900	1	https://doi.pangaea.de/10.1594/PANGAEA.940240 accessed on 20 October 2022
h	The Baltic Sea dataset	[43]	2016	5805	320–953.6	3.3	https://zenodo.org/record/5572537 accessed on 16 October 2023
i	Spectrum of Polluted Water in China	[44]	2001	35	393.8–1041.5 398.3–1043.61	2.7, 2.69	accessed on 1 June 2023
j	The Bohai and Huanghai Sea Dataset	in situ bio-optical dataset (2014–2018) measured by Zhongfeng Qiu and Shengqiang Wang	2014–2018	30, 36, 65	350–2500	1	in situ accessed on 1 June 2023
k	Ulva prolifera, Sargassum	[45]	2016, 2018	10, 18	350.11–999.99 347.07–1040.46	0.17	http://dx.doi.org/10.1016/j.rse.2016.02.065 accessed on 7 August 2024
l	Ulva prolifera, Sargassum	[46]	2016, 2018	10, 18	350.11–999.99 347.07–1040.46	0.17	http://qdhys.ijournal.cn/hyyhz/ch/reader/view_abstract.aspx?doi=10.11693/hyhz20171200331 accessed on 20 June 2023
m	Spectrum of Red Tide	[47]	2010	6, 5, 9, 5, 4, 1, 2, 5, 4, 4, 5, 1	396.6–1041.91, 393.8–1041.5, 398.3–1043.61	2.69, 2.7, 2.69	https://www.tandfonline.com/doi/full/10.1080/01431160902882512 accessed on 6 June 2023
n	Bohai Sea 863 Dataset China	Measure by Dongzhi Zhao, 863 Project	2003–2017	31, 54, 15, 45, 22, 26, 51	350–1050, 342.5–844.1, 342.5–2509.9, 320–946, 320–950, 325–1072, 400–900, 350–900	1, 1.6, 1.2, 1, 1, 1, 1	in situ accessed on 6 June 2023
o	Taihu Lake Dataset China	Provided by Hongtao Duan et al.	2021	25	400–1072	1	in situ accessed on 10 July 2023
p	Chaohu Lake Dataset China	Provided by Hongtao Duan et al.	2020	20	400–1072	1	in situ
q	Prorocentrum micans	[48]	2004	5	400–750	1	https://doi.org/10.3964/j.issn.1000-0593(2013)07-1892-05 accessed on 1 September 2023
r	Amphidiniumcarterae Hulburt	[49]	2013	7	400–750	1	https://doi.org/10.3964/j.issn.1000-0593(2013)07-1892-05 accessed on 1 September 2023
s	Skeletonema costatum	[48]	2014	4	400–752	1	https://doi.org/10.3964/j.issn.1000-0593(2013)07-1892-05 accessed on 1 September 2023
t	Aureococcus anophagefferens	[50]	2016	6	400–899	1	http://dx.doi.org/10.1155/2016/1780986 accessed on 1 September 2023
u	Hyperspectral Reflectance Characteristics of Cyanobacteria	[51]	2021	13	400–800	1	https://doi.org/10.4236/ars.2021.103004 accessed on 1 September 2023

Table 2. Database of normal water bodies.

Spectral Range (nm)	360–830 nm	390–830 nm	400–830 nm	Total
Inland Water	1767	76	844	2687
Coastal Water	6885	0	23	6908
Total	8652	76	867	9595

Table 3. Database of algal bloom water bodies.

The Colors of the Tides	Algae Species	Spectral Range (nm)	Data Numbers	Increments (nm)	Measurement Technique
Red Tides	Ceratium fura sp.	400–830	5	1	in vivo
	Dinoflagellates	360–830	34	1	in vivo, in situ
		390–830	2656		in situ
		400–830	418		in situ
	Gymmodinium sp.	400–830	6	1	in vivo
	Nitzschia closterium	400–830	4	1	in vivo
	Noctiluca scintillans	400–830	5	1	in vivo
	Coscinodiscus Concinnus	400–830	1	1	in vivo
	Spirulina	400–830	1	1	in vivo
	Alexandrium	400–830	10	1	in vivo
	Heterosigma akashiwo	400–830	9	1	in vivo
Brown Tides	Aureococcus anophagefferens	400–830	6	1	in situ
Brown Tides	Dicrateria zhanjiangensis Hu.	400–830	10	1	in vivo
Green Tides	Ulva prolifera	360–830	29	1	in situ
	Ulva prolifera	400–830	6	1	in situ
	Pyramimonas sp.	400–830	5	1	in vivo
	Platymonas sp.	400–830	8	1	in vivo
	Chlorella sp.	400–830	4	1	in vivo
Green-Blue Tides	Marine Cyanobacteria	400–830	4	1	in vivo
	Cyanobacteria	360–830	242	1	in situ
	Cyanobacteria	400–830	19	1	in situ
Gloden Tides	Sargassum	360–830	5	1	in situ
		390–830	11		in situ
		400–830	3		in situ
	Chaetoceros	400–830	9	1	in vivo
	Skeletonema costatum	400–830	4	1	in situ

Table 4. The WCIs of the different algal species in the waters.

The Colors of the Tides	Algae Species	H	S	λd	IAVW
Red Tides	Ceratium fura sp.	176.6–213.3	0.055–0.085	551.5–573	539.0–575.7
	Dinoflagellates	121.5–222.4	0.014–0.291	499.8–576.9	512.3–697.7
	Gymmodinium sp.	197.0–222.1	0.079–0.105	565.3–576.8	590.3–609.1
	Nitzschia closterium	64.9–175.0	0.021–0.053	485.5–550	593.2–618.9
	Noctiluca scintillans	151.6–161.8	0.058–0.069	518.7–533.8	528.7–530.5
	Coscinodiscus Concinnus	211.5–211.5	0.077–0.077	572.2–572.2	590.5–590.5
	Spirulina	151.3–151.3	0.059–0.059	518.2–518.2	621.2–621.2
	Alexandrium	201.6–260.0	0.110–0.288	567.7–597.6	613.7–727.1
	Heterosigma akashiwo	180.0–225.5	0.0287–0.146	554.4–578.3	539.9–634.6
Brown Tides	Aureococcus anophagefferens	200.1–212.0	0.075–0.105	566.9–572.4	576.5–613.3
Brown Tides	Dicrateria zhanjiangensis Hu.	65.4–67.4	0.0543–0.068	485.6–486.2	540.9–571.1
Green Tides	Ulva prolifera	167.1–207.0	0.070–0.176	541.2–570.2	600.3–734.6
	Platymonas sp.	71.7–191.7	0.025–0.165	487.4–562.4	554.1–682.7
	Pyramimonas sp.	64.2–118.9	0.0327–0.071	485.2–498.9	541.6–586.5
	Chlorella sp.	56.1–169.6	0.030–0.072	482.5–544.3	547.5–652.1
Green-Blue Tides	Marine Cyanobacteria	78.4–180.6	0.027–0.092	489.1–554.9	590.8–701.0
Green-Blue Tides	Cyanobacteria	157.0–203.2	0.036–0.109	526.5–568.4	529.16–615.6
Golden Tides	Skeletonema costatum	59.7–220.7	0.036–0.117	483.8–576.2	500.5–600.0
	Chaetoceros	217.9–254.2	0.0821–0.270	575–593.1	628.0–709.2
	Sargassum	187.4–241.1	0.029–0.143	559.7–585.4	617.0–742.9

Table 5. The detailed information of the spectral data in the Bohai Sea.

Type of the Water Bodies		Wavelength Range (nm)	Numbers of the Spectral Data	Total
Normal Water Bodies		360–830	206	209
Normal Water Bodies		400–830	3	209
Algal Bloom Water Bodies	Unknown Algae	360–830	22	71
	Unknown Algae	400–830	23
	Noctiluca scintillans	400–830	10
	Aureococcus anophagefferens	400–830	6
	Ceratium fura sp.	400–830	10

Table 6. The detailed information of the spectral data in Taihu Lake.

Type of the Water Bodies	Wavelength Range (nm)	Numbers of the Spectral Data	Total
Normal Water Bodies	360–830	215	232
Normal Water Bodies	400–830	17	232
Algae Bloom Water Bodies	360–830	5	22
Algae Bloom Water Bodies	400–830	17	22

Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content.

© 2024 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https://creativecommons.org/licenses/by/4.0/).

Share and Cite

MDPI and ACS Style

Zhao, D.; Luo, Q.; Qiu, Z. Chromaticity-Based Discrimination of Algal Bloom from Inland and Coastal Waters Using In Situ Hyperspectral Remote Sensing Reflectance. Water 2024, 16, 2276. https://doi.org/10.3390/w16162276

AMA Style

Zhao D, Luo Q, Qiu Z. Chromaticity-Based Discrimination of Algal Bloom from Inland and Coastal Waters Using In Situ Hyperspectral Remote Sensing Reflectance. Water. 2024; 16(16):2276. https://doi.org/10.3390/w16162276

Chicago/Turabian Style

Zhao, Dongzhi, Qinshun Luo, and Zhongfeng Qiu. 2024. "Chromaticity-Based Discrimination of Algal Bloom from Inland and Coastal Waters Using In Situ Hyperspectral Remote Sensing Reflectance" Water 16, no. 16: 2276. https://doi.org/10.3390/w16162276

Note that from the first issue of 2016, this journal uses article numbers instead of page numbers. See further details here.

Article Menu