Open Access
Add to BibSonomy
Abstract
In large-scale water quality evaluation, traditional field-measured data lack spatial-temporal representativeness, and the role of conventional remote sensing parameters (SST, Chla, TSM, etc.) is controversial. By calculating and grading the hue angle of a water body, a Forel-Ule index (FUI) can be obtained, which provides a comprehensive statement of water condition. Using MODIS imagery, hue angles are extracted with better accuracy than the literature's method. It is found that FUI changes in the Bohai Sea have correlated consistently with water quality. The decreasing trend of non-excellent water quality areas in the Bohai Sea was highly correlated with FUI (R2 = 0.701) during the government-dominated land-based pollution reduction program (2012-2021). FUI can monitor and evaluate seawater quality.
© 2023 Optica Publishing Group under the terms of the Optica Open Access Publishing Agreement
1. Introduction
Water color is a vital component of hydrology, and its in situ observation dates back more than a century, utilizing the Forel-Ule scale to classify natural water color from dark blue to red - brown into 21 levels. This scale is used to record the color of global marine and inland water bodies [1]. Even if people did not understand the physical and chemical causes of seawater color changes in the past, they intuitively understood that the color of seawater was directly tied to the substances and organisms it contained. The close relationship between ocean color and chlorophyll concentration has been confirmed by bio-optical radiative transfer models, and ocean color is one of the World Meteorological Organization's fundamental climate variables because it reflects the biological activity of the ocean's surface layer. In addition, scientists used historical observations of ocean color to reconstruct the historical trends of chlorophyll concentration in the global ocean surface layer from 1889 to 2000 [2].
Table 1. Linear coefficients for calculating chromaticity values based on MODIS visible band data
The first coastal zone water color scanner (CZCS) with an ocean color sensor was launched by NASA's Nimbus-7 satellite in the late 1970s, and further ocean color remote sensing investigations started [3]. Subsequently, a number of countries launched the ocean color series satellites, including Seastar/SeaWiFS, Envisat/MERIS, Aqua&Terra/MODIS, SNPP&JPSS1/VIIRS, COMS/GOCI, HY-1/COCTS, FY-3/MERSI, and others. Chlorophyll a (Chla), total suspended matter (TSM), and colored dissolved organic matter (CDOM), often referred to as the three primary elements of ocean color [4], are closely related to the color of ocean water and serve as the fundamental parameters for detecting ocean water in remote sensing [5,6], along with other parameters like transparency, net primary productivity, particulate organic carbon [7–9].
Early researchers employed spectrum information from satellite multispectral sensors or field measurements to get the X, Y, and Z tristimulus values in the Commission International de I’ Éclairage (CIE) standard colorimetry system, and then calculated the colorimetric information to infer the status of water bodies [10–12]. With the advancement of optical technology, researchers are able to observe more continuous spectrum data of water and conduct more extensive research in this area. Wernand et al. [13]. created 21 color solutions in the Forel-Ule scale, measured the transmission spectra of the solutions, and determined the chromaticity coordinates of 21 water colors. In 2013, Novoa et al. [1] further rectified the chromaticity coordinates of 21 water colors in the Forel-Ule scale, estimated the hue angle for each color, and produced the Forel-Ule index (FUI) with hue angle lookup table. In the same year, Wernand et al. [14] developed the FUI extraction algorithm for water using MERIS images. Researchers have proved over the past decade that remote sensing sensors (SeaWiFS, MODIS, MERIS, OLCI, CZCS, MSI, ETM+, OLI, etc.) with a variety of band settings can retrieve the FUI and hue angle with great precision. However, sensors with limited bands and wide bandwidths, such as ETM, OLI, etc., have a much greater influence on the hue angle and the extraction accuracy [15,16]. A further benefit of FUI and hue angle remote sensing extraction is that they convey a “radiometric message” of the water body and are not dependent on the characteristics of the water body in a particular location. As a result, FUI has become an effective and simple method for estimating different aspects of the water environment, such as turbidity, transparency, CDOM, Chla, and TSM [17–23], as well as evaluating the nutrient status of water bodies [24–28].
The extraction of hue angles (or FUI) from MODIS imagery is usually performed with a 7-band [15] or RGB 3-band method [19], but no report has been published on how MODIS's 10 visible bands can be fully applied to extract hue angles and accuracy analysis. Furthermore, when evaluating water quality using FUI, it primarily focuses on inverting water environment elements and assessing the nutrient status, and no research has been conducted on traditional marine water quality evaluation (Bulletin of Marine Ecology and Environment Status of China [29]), based on some chemical parameters, such as dissolved oxygen, nutrient salts, etc.). Based on 825 sets of field-measured spectral data in China's coastal waters, this paper extracts hue angles from 10 visible bands (412, 443, 469, 488, 531, 547, 555, 645, 667, and 678 nm) obtained from MODIS and performs accuracy analysis in order to determine and analyze the annual and monthly variations in FUI in the Bohai Sea. Using the results of the evaluation of seawater quality published in the Bulletin of Marine Ecology and Environment Status of China, the correlation between FUI and water quality conditions in the Bohai Sea is discussed. A study of the causal factors contributing to monthly and interannual variations in FUI and the feasibility of applying FUI to evaluate near-shore water quality is presented.
2. Study area and data
2.1 Study area
The Bohai Sea (Fig. 1), consisting of Bohai Bay, Laizhou Bay, Liaodong Bay, and the middle portion of Bohai Sea, is the only semi-enclosed inland sea in China, located between the Liaodong Peninsula and the Jiaodong Peninsula and connected to the North Yellow Sea only in the east, with many rivers along the coast, where three of China's seven largest rivers, the Yellow River, the Hai River, and the Liao River, drain into the Sea. Over the past four decades of China's reform and opening up, together with the expansion and building of 13 coastal cities surrounding the Bohai Sea, land-based pollutants have steadily increased and the ecological quality of the Bohai Sea has been severely compromised. Since the 21st century, the Chinese government has developed and executed a number of policy regimes for the management of environmental pollution in the Bohai Sea, including the Bohai Integrated Remediation Plan, Plan for Bohai Blue Sea Action, General Plan for Bohai Environmental Protection (2008-2020), Opinions of the State Oceanic Administration on Strengthening the Protection of the Bohai Sea's Ecological Environment, and Integrated Administration Action Plan for Bohai Sea [30]. The ecological environment of the Bohai Sea has been somewhat improved, however the efficacy of remediation remains unstable.
2.2 Data
2.2.1 In situ data
In situ 825 remote sensing reflectance spectra were acquired in China's coastal waters from 2005 to 2021 using an ASD company spectrometer with the above-water measuring approach. The measurements and data processing were based on NASA ocean optics protocol [31].
Using the MODIS spectral response function, the in situ hyperspectral remote sensing reflectance was used to calculate the equivalent multispectral remote sensing reflectance. The exact formulation is as follows:
2.2.2 Satellite data
For the calculation of FUI, satellite images from Terra and Aqua MODIS were downloaded from the NASA website . The FUI were calculated using the remote sensing reflectance at 412, 443, 469, 488, 531, 547, 555, 645, 667, and 678 nm from the monthly and annual level 3 MODIS data from 2002 to 2021. Notably, the satellite images from January to July 2002 are from Terra MODIS, while the rest are from Aqua MODIS.
3. Methods
3.1 FUI calculation
To quantitatively depict color, the CIE devised the CIE-XYZ standard colorimetry system [32]. Using X, Y, and Z as the three tristimulus values of the spectrum (three primary colors) in the colorimetry system, which can be derived from the relative spectral energy distribution of the light source, the CIE-specified color matching function, and the remote sensing reflectance of water body.
Following are the formulae for computing X, Y, and Z with in situ hyperspectral remote sensing reflectance:
And the equations for estimating X, Y, and Z using multispectral remote sensing reflectance (e.g. satellite data) are as follows [15]:
where n is the number of bands of the satellite sensor, x i, y i, z i are the band linear summation coefficients used to generate the CIE color tristimulus values from satellite data, and they are constants (Table 1).The CIE specifies a two-dimensional chromaticity diagram to depict colors, and the two-dimensional coordinates x and y on the chromaticity diagram are derived using the three tristimulus values X, Y, and Z. The following equations exist:
Two values of x and y can determine a color, then the CIE-xy chromaticity diagram [25] can show all the colors in the visible range, and each color has a chromaticity coordinate (x, y). Using the bivariate arctangent function (arctan2) and the following equation, the hue angle can be found from the chromaticity coordinates (x, y):
Due to the multispectral satellite sensor's band dispersion and position setting, the human eye sees real color from hyperspectral differently than the sensor does [15]. To eliminate color discrepancies due by band dispersion and satellite sensor settings, a systematic deviation delta Δα is defined as the multispectral hue angle (α multi) minus the hyperspectral hue angle (α hyper):
The FUI was then calculated based on the hue angle after rectification using the 21-level FUI lookup table constructed from the chromaticity coordinates of the Forel-Ule scales [1,25,33] (Fig. 2). A larger FUI indicates that the water body is more turbid and the color is poorer, whereas a smaller FUI indicates that the water body is cleaner and the color is better.
Fig. 2. The FUI colors and subdivision of the CIE chromaticity diagram from 1 to 21 are shown as blue dots at the right part of the figure [1]. On the left, the hue angle matching range and 21-level FUI lookup table are shown [25,33].
3.2 Hue angle correction
The absence of a reference correction equation is due to the fact that few researchers used MODIS images in all visible bands to determine hue angle in earlier research. On the basis of 825 in situ measurements of remote sensing reflectance in the coastal waters of China, a correction equation for deviation is developed. This work established a polynomial equation to model Δα for MODIS data at wavelengths of all visible bands.
The in situ hyperspectral R rs(λ) and the MODIS equivalent multispectral R rs(λ i) were utilized to determine the hyperspectral hue angle (α hyper) and the multispectral hue angle (α multi)), respectively. In accordance with the method given by Van der Woerd and Wernand (2015) [15], the deviation delta Δα was modeled with a fifth-order polynomial fit (Fig. 3) using the in situ 825 remote sensing reflectance spectra. The MODIS data has the following polynomial equation (R2 = 0.81):
Fig. 3. Diagram showing the Δα correction polynomial equation fitting for hue angle generated from MODIS bands at 412, 443, 469, 488, 531, 547, 555, 645, 667, 678 nm.
Prior to rectification (Fig. 4(a)), the hue angles α hyper and α multi show obvious systematic variation, with a slope of 0.84 and an intercept of 30.575. There is no persistent divergence after correction (Fig. 4(b)), and the slope and intercept are 0.994 and 1.042, respectively. It implies that the correction equation can generate a more effective corrective effect, which is required for the FUI calculation.
Fig. 4. Scatterplots showing the comparison of linear coefficient method before and after hue angle correction.
3.3 FUI accuracy analysis
FUI is a discrete grading of the hue angle. Therefore, FUI accuracy is completely dependent on hue angle accuracy. As a matter of fact, the hue angle is a quantifiable parameter that makes it possible to estimate the color of water bodies as perceived by humans. By measuring the water reflecting spectrum in the visible range with a hyperspectral radiometer and calculating the hue angle (section 3.1), we can determine the color of the water.
Hue angle is determined by the type and concentration of light-sensitive substances present in water, but extraction does not rely on these substances, but on the ability of the hyperspectral radiometer (or human eyes) to detect the reflected spectrum within the visible range, a process that is distinct from the inversion of water color elements (Chla, TSM, CDOM) by using the reflected spectrum. Hue angle provides a better indication of water column “radiometric information”. The extraction process for in-situ hyperspectral data is universal and does not differ by region.
Nevertheless, the accuracy of hue angles extracted from multispectral satellite data is influenced by two main factors. On the one hand, the quality of the satellite data, such as the sensor signal-to-noise ratio, the reasonableness of the band settings, the accuracy of the atmospheric correction algorithm, etc. These are mainly related to sensor performance and will not be discussed here. On the other hand, there is a systematic bias between hue angles extracted from hyperspectral data and multispectral satellite imagery. A correction equation is required to remove this bias. The correction accuracy is related to the multispectral satellite's band continuity. It is also related to the representativeness of the sample data used for modeling.
In order to validate the effect of band continuity on the accuracy of hue angle extraction, 7 visible bands of MODIS [15] were utilized for correction equation modeling, and the processing was the same as in Section 3.2. As shown in Fig. 5(a), when compared to 10-band method, the accuracy of hue angle extraction decreases. The representativeness of the sample data used in modelling also impacts the accuracy of hue angle extraction. A decrease in accuracy can be seen in Fig. 5(b).
Fig. 5. Scatterplots show the hue angle derivation accuracy after different Δα correction equations from MODS sensors by comparing with hyperspectral integration. The Δα correction equation for MODIS data at seven visible bands (412, 443, 488, 531, 555, 667, and 678 nm) is from (a) self-fitting based on 825 in situ measurements in China's coastal waters, and (b) the paper of Van der Woerd and Wernand (2015) [15].
Overall, the correlation between the bias-corrected hue angle of the multispectral data and the hyperspectral extracted hue angle is greater than 0.97, with the mean absolute percentage error (MAPE) all below 2% and the root mean square error (RMSE) all below 4.5°, indicating that the hue angle extraction using multispectral satellite data has good generalizability and stability.
The 10-band hue angle extraction method developed in this study, however, is highly effective in Chinese nearshore waters and has improved accuracy over the 7-band hue angle extraction method in the literature. There has been an increase in correlation coefficient (R2) from 0.978 to 0.994, a decrease in MAPE from 1.96% to 0.85%, and a decrease in RMSE from 4.475° to 2.189°. This study focuses on Chinese nearshore waters and does not include sample data from oceanic clean water bodies, so the applicability of the model to oceanic waters needs to be verified.
4. Results
4.1 Monthly spatiotemporal variation of FUI in the Bohai Sea
The spatiotemporal variance of monthly FUI in the Bohai Sea is remarkable (Fig. 6). The coastal waters of Liaodong Bay, Bohai Bay, and Laizhou Bay have a high FUI, turbid water, and poor water color, but the seas of Qinhuangdao and other offshore waters have a low FUI, relatively clean water, and decent water color. In terms of time, the FUI was greater in January, February, March, November, and December, and the water color condition was worse. The FUI was lower from April to October, and the water color condition was better.
Fig. 6. Spatiotemporal variation of monthly FUI in the Bohai Sea from January to December over the past two decades
Further statistical analysis (Table 2) revealed that the FUI changes in the range of 5-17 in the Bohai Sea, the percentage of area with higher FUI decreases and then increases from January to December each year (Lower FUI, in contrast, has the reverse behavior), i.e., the percentage of turbid water decreases and then increases from January to December each year (the opposite is true for clear water), reflecting the process of the Bohai Sea ocean color changing from turbid to clear and then to turbid. The sea area percentage of FUI ≥ 10 is greater than 60% in January, February, March, November, and December, with the highest FUI in January indicating the most turbid water and the worst ocean color condition. From April through November (excluding October, 65%), the area percentage of FUI ≤ 9 is greater than 75%, with the lowest FUI in July indicating the clearest water and best ocean color condition.
Table 2. Area percentage of 21-level FUI in the Bohai Sea from January to December for the past two decades (%), with data in bold indicating that the area percentage surpasses 10%.
The monthly FUI of the Bohai Sea varies from 6 to 11 (Fig. 7), and the FUI basically shows a “V"-shaped change from January to December each year, i.e., the FUI value changes from large to small and then to large, which also reflects the process of the Bohai Sea water color changing from turbid to clear and then to turbid from January to December. Furthermore, the overall lowering trend of FUI in the Bohai Sea from 2002 to 2021 implies that the water color condition in the Bohai Sea is improving.
Fig. 7. The monthly mean FUI in the Bohai Sea from January 2002 to December 2021, and the red line illustrates the declining trend of FUI.
4.2 Interannual spatiotemporal variation of FUI in the Bohai Sea
Bohai Sea annual FUI exhibits spatiotemporal features (Fig. 8). The coastline waters of Liaodong Bay, Laizhou Bay, and Bohai Bay are always muddy due to land-based runoff from the Liaohe River, Yellow River, and Haihe River, but the near-shore waters of Qinhuangdao and the central Bohai Sea are always clean. The annual FUI fluctuates, but not significantly.
Bohai Sea FUI averages 8.49 to 9.74. (Fig. 9). The first ten years (2002–2011) saw two troughs in 2005 and 2009, when the mean annual FUI was 9.29 and 8.99, and two peaks in 2004 and 2006, at 9.69 and 9.74, respectively. The following 10 years (2012-2021), the highest mean annual FUI was 9.49 in 2013, then steadily fell, with a tiny recovery in 2016, a substantial decrease in 2018, and the lowest mean in 2019 at 8.49, followed by a small increase.
As shown in Fig. 10, other than a few years (2011, 2018, 2019), the percentage of sea area where the FUI of the Bohai Sea has not changed varies between 40% and 65%. In 2003-2008, 2010-2014, and 2016, the percentage of sea area with FUI more than the historical average in the Bohai Sea was higher than the area less than the historical average, and the ocean color condition was worse than the historical average, with the worst water color condition in 2006. The percentage of area with ΔFUI < 0, ΔFUI = 0, ΔFUI > 0 was 3.69, 45.14, and 51.17 (Fig. 11), respectively. While the percentage of sea area with FUI less than the historical average was more than the percentage of sea area with FUI greater than the historical average in 2002, 2009, 2015, and 2017-2021, and the ocean color condition is better than the historical average, with 2019 having the best ocean color condition, the percentages of sea area with ΔFUI < 0, ΔFUI = 0, ΔFUI > 0 were 64.75, 34.13, and 1.12, respectively.
Fig. 10. Spatiotemporal distribution of FUI change (ΔFUI) in the Bohai Sea from 2002 to 2021 compared to the historical annual average FUI.
Fig. 11. The percentage of Bohai Sea area where ocean color has improved (FUI < 0), stable (FUI = 0), or worsened (FUI > 0) compared to the historical annual average FUI from 2002 to 2021.
5. Discussion
FUI is related with both natural and human influences. The natural influences include sea surface wind and waves, tidal dynamics, and river inlet, whereas the artificial factors include reclamation projects, marine dumping, and land-based discharges. In particular, surface wind and waves influence the concentration of suspended particulate matter by local resuspension, resulting in an increase in FUI. The tidal current influences the distribution of suspended particles by advection transport and local resuspension, resulting in a change in FUI. Large quantities of suspended particle matter are emitted by reclamation projects and ocean dumping activities, leading to a rise in FUI. Land-based discharges typically enter the ocean via rivers, which contain a great deal of suspended particulate matter, nutrient salts, and other pollutants, resulting in high levels of FUI in estuaries and coastal seas. In addition, the huge input of nutrients from land-based sources can cause eutrophication in near-shore seas, resulting in an increase in the abundance of planktonic algae and even red tide catastrophes, which alter the FUI. Therefore, the input of suspended particulate matter from land-based sources, the resuspension of bottom sediments, and the growth of planktonic algae are strongly associated with the alteration of ocean color.
5.1 Impact of the monsoon on FUI monthly variation
The monthly FUI of the Bohai Sea from January to December exhibits a “V"-shaped fluctuation, which is mostly attributable to the effect of natural causes on ocean color conditions. The Bohai Sea exhibits obvious monsoon characteristics. As a result of the winter northwest wind direction, which is stable, the wind is powerful, the sea gales occur on average 2 to 4 times a month, each time lasting approximately 2 to 5 days, the prevailing period between October and March is about six months. Summer winds are mainly southeast winds, the direction of which is unstable, the wind strength is weak, the prevailing period occurs from May to August. While April and September are transitional periods. This corresponds to the change in water color in the Bohai Sea from January to December, from turbid to clear to turbid again, i.e. “strong winds increase FUI, weak winds decrease FUI”.
In general, severe precipitation in summer causes a sudden increase in land-based runoff, resulting in a rise in FUI in the sea adjacent to the estuary. However, the effects tend to be relatively weak compared to the resuspension of particulate matter resulting from strong winds in winter. The impact is generally short-lived and predominantly located in the sea adjacent to the estuary.
5.2 Impact of pollution control on FUI interannual variation
Changes in the annual FUI of the Bohai Sea from 2002 to 2021 may be mostly attributable to the effect of human influences on ocean color conditions. In the first decade, the FUI fluctuated significantly, with a sharp fall from 2007 to 2009, coinciding with the Beijing Olympics in 2008. The “Green Olympics” was one of the three key themes of the 2008 Olympic Games in Beijing, and it permeated the whole planning and hosting process. At that time, the Chinese government enacted stringent pollution reduction measures for Beijing and its nearby provinces and cities, substantially reducing the impact of land-based discharges into the sea and resulting in improved ocean color conditions. According to the Bulletin of Marine Ecology and Environment Status of China, the area of non-excellent water quality (not meeting the standard Grade I and Grade II seawater quality) in the Bohai Sea from 2007 to 2010 was 17040, 13810, 12580 and 16990 km2. The trend of FUI (9.38, 9.32, 8.99 and 9.59) indicates that value in 2009 reached a trough, with better water quality and color levels.
The Chinese government's concern for the environment caused a steady reduction in FUI in the Bohai Sea during the next decade. Since 2013, the Chinese government has fiercely promoted ecological civilization and adhered to the development theory that lucid waters and lush mountains are invaluable assets, resulting in a substantial reduction of pollution discharge. The Bohai Sea's ecological environment has been significantly improved as a result of the major efforts for Bohai comprehensive governance, which was fully implemented in 2018. This has led to a steep decline in FUI in 2018 and 2019 as well as a notable improvement in the Bohai Sea's ocean color condition. Nevertheless, the biological environment of the Bohai Sea confronts long-term and complicated issues, and a little increase in FUI in 2020–2021 indicates that governance is not yet solid. An analysis of the Bulletin of Marine Ecology and Environment Status of China [29] for the last ten years indicates that the water quality of the Bohai Sea has generally improved, with a decreasing trend in the area of non-excellent water quality and a significant positive correlation with the annual average value of FUI (Fig. 12, R2 = 0.701). This essentially indicates that the lower the FUI, the better the color condition of the water, and the smaller the area of non-excellent water quality sea area, the better the water quality condition.
Fig. 12. Scatterplots of the sea area with water quality inferior to grade II and annual FUI of the Bohai Sea from 2012 to 2021. China's seawater quality standards define five categories of water quality: I, II, III, IV, and inferior to IV, among which meet Grade I, II seawater quality standards for excellent water quality, and inferior to Grade II water quality standards for non-excellent water quality.
5.3 Analysis of other possible influencing factors
Does climate change, etc., affect the interannual variability of FUI in addition to policy factors such as pollution prevention? The interannual variation in FUI, SST, rainfall, wind speed and coastal reclamation area is shown in Fig. 13. It can be found that in the past 20 years, due to global climate change, the overall SST of the Bohai Sea has shown an increasing trend. In contrast, the reclamation area presents a reversed V-shaped pattern, reflecting the government's policy shift from economic priority to environmental protection. There is no significant inter-annual variation in rainfall and wind speed.
Fig. 13. The annual mean FUI, SST, rainfall, wind speed, and reclamation in the Bohai Sea for the past 20 years. SST data is from Aqua MODIS Global Mapped 11 µm Daytime Sea Surface Temperature (SST) Data, version R2019.0. Rainfall data is from IMERG Final Precipitation L3 1 month 0.1 degree x 0.1 degree V06 (GPM_3IMERGM) [34]. Wind speed data is from M2TMNXFLX [35], a time-averaged 2-dimensional monthly mean data collection in Modern-Era Retrospective analysis for Research and Applications version 2 (MERRA-2), and reclamation data is obtained from shoreline change information extracted by human-machine interaction.
According to the correlation analysis (Fig. 14), all other elements were not significantly correlated with FUI, except for sea surface temperature, which was weakly correlated. FUI is most likely to be affected by changes in planktonic algae concentration, which have decreased over the last decade [36]. A small change in temperature, however, may not have a significant impact on planktonic algae in the semi-enclosed Bohai Sea. The key factor is the reduction in nutrient input from land-based pollution reduction. Rainfall and wind variation over an annual period are not characterized by significant variability, and they are not related to changes in the FUI. However, reclamation activities generally cause transient increases in FUI in the surrounding waters. These effects are short-lived and small in scope, and have no significant impact on the interannual changes in FUI across the Bohai Sea.
Fig. 14. Scatterplots of the (a) FUI and SST, (b) FUI and wind speed, (c) FUI and rainfall, and (d) FUI and reclamation.
6. Conclusions
- (1) In this study, a fifth-order polynomial equation (R2 = 0.81) was developed to calculate hue angle and FUI at 10 visible wavelengths of 412, 443, 469, 488, 531, 547, 555, 645, 667, and 678 nm using in situ 825 remote sensing reflectance spectra collected in Chinese nearshore waters. The hue angle in this paper has improved accuracy over the 7-band hue angle extraction method in the literature. There has been an increase in correlation coefficient (R2) from 0.978 to 0.994, a decrease in MAPE from 1.96% to 0.85%, and a decrease in RMSE from 4.475° to 2.189°.
- (2) Significant spatiotemporal variation is observed in the monthly FUI of the Bohai Sea. The FUI of Liaodong Bay, Bohai Bay, and Laizhou Bay is high, the water is muddy, and the water color of these bays is poor. However, the FUI of Qinhuangdao and other offshore waters is low, the water is reasonably clear, and the water color condition is excellent. A fluctuation in the FUI is observed from 2002 to 2021, corresponding to the movement of the Bohai Sea's water from turbid to clear to turbid between the months of January and December.
The annual average FUI for the first 10 years (2002–2011) fluctuated greatly, and maintained a high level, with the poorest water color condition and the highest annual average FUI of 9.74 in 2006, which was much higher than the historical average. In the ten years that followed (2012–2021), the annual average FUI declined and the water quality improved, with the lowest annual average FUI value of 8.49 in 2019 signifying the best ocean color condition.
- (3) The monthly variation in FUI is related to the Bohai Sea monsoon characteristics. During the winter season, winds have higher speeds and longer durations, prevailing between October and March. During the summer season, winds have lower speeds and longer durations from May to August, while April and September are transitional months. In the Bohai Sea, water color changed from turbid to clear to turbid again from January to December, i.e. “strong winds increase FUI, weak winds decrease FUI”.
There is a close correlation between FUI interannual variation and the Chinese government's efforts to reduce land-based emissions. The correlation between FUI and SST, rainfall, wind speed, and coastal reclamation, however, was not significant. As part of the Beijing Olympics (2007-2009), strict pollution abatement measures were implemented and in accordance with the trend of the annual average of FUI, the area of the Bohai Sea with non-excellent water quality has decreased year by year. Over the past decade, the Chinese government has placed an emphasis on ecological civilization, which has set strict requirements for reducing emissions from land-based sources. Using the Bulletin of Marine Ecology and Environment Status of China from 2012 to 2021, it is evident that the Bohai Sea's water quality has generally improved since 2012. The area of non-excellent water quality is decreasing, which is evidently correlated with the annual average value of FUI. The FUI can be used as a powerful tool for monitoring and assessing nearshore water quality.
Funding
Ministry of Science and Technology of the People's Republic of China (2018YFC1407605, 2019YFC1407904); National Natural Science Foundation of China (42076186).
Acknowledgments
We acknowledge Ocean Biology Processing Group (OBPG) at NASA’s Goddard Space Flight Center for providing MODIS satellite data. We would like to thank our colleagues from National Marine Environmental Monitoring Center, for their help in field spectral measurements, the results of which were used in this work.
Disclosures
The authors declare no conflicts of interest.
Data availability
Data underlying the results presented in this paper are not publicly available at this time but may be obtained from the authors upon reasonable request.
References
1. S. Novoa, M. Wernand, and H. Van der Woerd, “The Forel-Ule scale revisited spectrally: preparation protocol, transmission measurements and chromaticity,” J. Eur. Opt. Soc. Rapid Publ. 8, 13057 (2013). [CrossRef]
2. M. R. Wernand, H. J. van der Woerd, and W. W. Gieskes, “Trends in ocean colour and chlorophyll concentration from 1889 to 2000, worldwide,” PLoS One 8(6), e63766 (2013). [CrossRef]
3. H. R. Gordon and A. Y. Morel, “Remote assessment of ocean color for interpretation of satellite visible imagery: A review,” Phys. Earth Planet. Inter. 37(4), 292 (1983). [CrossRef]
4. L. Prieur and S. Sathyendranath, “An optical classification of coastal and oceanic waters based on the specific spectral absorption curves of phytoplankton pigments, dissolved organic matter, and other particulate materials 1,” Limnol. Oceanogr. 26(4), 671–689 (1981). [CrossRef]
5. J. E. O’Reilly, S. Maritorena, B. G. Mitchell, D. A. Siegel, K. L. Carder, S. A. Garver, M. Kahru, and C. McClain, “Ocean color chlorophyll algorithms for SeaWiFS,” J. Geophys. Res. 103(C11), 24937–24953 (1998). [CrossRef]
6. J. Tang, X. Wang, Q. Song, T. Li, J. Chen, H. Huang, and J. Ren, “The statistic inversion algorithms of water constituents for the Huanghai Sea and the East China Sea,” Acta Oceanol Sin. 23(4), 617–626 (2004).
7. Z. Lee, S. Shang, C. Hu, K. Du, A. Weidemann, W. Hou, J. Lin, and G. Lin, “Secchi disk depth: A new theory and mechanistic model for underwater visibility,” Remote Sens. Environ. 169, 139–149 (2015). [CrossRef]
8. G. M. Silsbe, M. J. Behrenfeld, K. H. Halsey, A. J. Milligan, and T. K. Westberry, “The CAFE model: A net production model for global ocean phytoplankton,” Global Biogeochem. Cycles 30(12), 1756–1777 (2016). [CrossRef]
9. A. G. Bonelli, H. Loisel, D. S. Jorge, A. Mangin, O. F. d’Andon, and V. Vantrepotte, “A new method to estimate the dissolved organic carbon concentration from remote sensing in the global open ocean,” Remote Sens. Environ. 281, 113227 (2022). [CrossRef]
10. T. Alföldi and J. Munday Jr, “Water quality analysis by digital chromaticity mapping of Landsat data,” Can. J. Remote Sens. 4(2), 108–126 (1978). [CrossRef]
11. R. Bukata, J. Bruton, and J. Jerome, “Use of chromaticity in remote measurements of water quality,” Remote Sens. Environ. 13(2), 161–177 (1983). [CrossRef]
12. R. P. Bukata, D. V. Pozdnyakov, J. H. Jerome, and F. J. Tanis, “Validation of a radiometric color model applicable to optically complex water bodies,” Remote Sens. Environ. 77(2), 165–172 (2001). [CrossRef]
13. M. Wernand and H. Van der Woerd, “Spectral analysis of the Forel-Ule Ocean colour comparator scale,” Eur Opt Soc Rapid Publ.5(2010).
14. M. Wernand, A. Hommersom, and H. J. van der Woerd, “MERIS-based ocean colour classification with the discrete Forel–Ule scale,” Ocean Sci. 9(3), 477–487 (2013). [CrossRef]
15. H. J. van der Woerd and M. R. Wernand, “True colour classification of natural waters with medium-spectral resolution satellites: SeaWiFS, MODIS, MERIS and OLCI,” Sensors 15(10), 25663–25680 (2015). [CrossRef]
16. H. J. Van der Woerd and M. R. Wernand, “Hue-angle product for low to medium spatial resolution optical satellite sensors,” Remote Sens. 10(2), 180 (2018). [CrossRef]
17. S. Garaba, T. Badewien, A. Braun, A.-C. Schulz, and O. Zielinski, “Using ocean colour remote sensing products to estimate turbidity at the Wadden Sea time series station Spiekeroog,” J. Eur Opt Soc Rapid Publ. 9, 14020(2014). [CrossRef]
18. S. P. Garaba, D. Voß, and O. Zielinski, “Physical, bio-optical state and correlations in North–Western European Shelf Seas,” Remote Sens. 6(6), 5042–5066 (2014). [CrossRef]
19. S. Wang, J. Li, Q. Shen, B. Zhang, F. Zhang, and Z. Lu, “MODIS-based radiometric color extraction and classification of inland water with the Forel-Ule scale: A case study of Lake Taihu,” IEEE J. Sel. Top. Appl. Earth Obs. Remote Sens. 8(2), 907–918 (2015). [CrossRef]
20. S. P. Garaba, A. Friedrichs, D. Voß, and O. Zielinski, “Classifying natural waters with the Forel-Ule Colour index system: results, applications, correlations and crowdsourcing,” Int. J. Environ. Res. Public Health 12(12), 16096–16109 (2015). [CrossRef]
21. J. Pitarch, H. J. van der Woerd, R. J. Brewin, and O. Zielinski, “Optical properties of Forel-Ule water types deduced from 15 years of global satellite ocean color observations,” Remote Sens. Environ. 231, 111249 (2019). [CrossRef]
22. S. Wang, Z. Lee, S. Shang, J. Li, B. Zhang, and G. Lin, “Deriving inherent optical properties from classical water color measurements: Forel-Ule index and Secchi disk depth,” Opt. Express 27(5), 7642–7655 (2019). [CrossRef]
23. S. Wang, J. Li, B. Zhang, Z. Lee, E. Spyrakos, L. Feng, C. Liu, H. Zhao, Y. Wu, and L. Zhu, “Changes of water clarity in large lakes and reservoirs across China observed from long-term MODIS,” Remote Sens. Environ. 247, 111949 (2020). [CrossRef]
24. J. Li, S. Wang, Y. Wu, B. Zhang, X. Chen, F. Zhang, Q. Shen, D. Peng, and L. Tian, “MODIS observations of water color of the largest 10 lakes in China between 2000 and 2012,” Int. J. Digit. Earth 9(8), 788–805 (2016). [CrossRef]
25. S. Wang, J. Li, B. Zhang, E. Spyrakos, A. N. Tyler, Q. Shen, F. Zhang, T. Kuster, M. K. Lehmann, and Y. Wu, “Trophic state assessment of global inland waters using a MODIS-derived Forel-Ule index,” Remote Sens. Environ. 217, 444–460 (2018). [CrossRef]
26. Q. Chen, M. Huang, and X. Tang, “Eutrophication assessment of seasonal urban lakes in China Yangtze River Basin using Landsat 8-derived Forel-Ule index: A six-year (2013–2018) observation,” Sci. Total Environ. 745, 135392 (2020). [CrossRef]
27. M. Li, Y. Sun, X. Li, M. Cui, and C. Huang, “An improved eutrophication assessment algorithm of estuaries and coastal waters in Liaodong Bay,” Remote Sens. 13(19), 3867 (2021). [CrossRef]
28. Y. Zhou, B. He, C. Fu, F. Xiao, Q. Feng, H. Liu, X. Zhou, X. Yang, and Y. Du, “An improved Forel–Ule index method for trophic state assessments of inland waters using Landsat 8 and sentinel archives,” GIScience Remote Sens. 58(8), 1316–1334 (2021). [CrossRef]
29. Ministry of Ecology and EnvironmentBulletin of Marine Ecology and Environmental Status of China (2023)http://english.mee.gov.cn/Resources/Reports/bomeaesoc/.
30. C. Yu, R. Zhu, W. Sui, Y. Xu, B. Liang, C. Bao, and M. Ma, “Comparative study on the effectiveness of water environmental pollution control between Bohai Sea and major international bays,” Mar Environ Sci. 40(6), 843–850 (2021) (in Chinese).
31. J. L. Mueller and G. S. Fargion, Ocean optics protocols for satellite ocean color sensor validation, revision 3, National Aeronautics and Space Administration, Goddard Space Flight Center, (2002), Vol. 210004.
32. T. Smith and J. Guild, “The C.I.E. colorimetric standards and their use,” Trans. Opt. Soc. 33(3), 73–134 (1931). [CrossRef]
33. X. Chen, L. Liu, X. Zhang, J. Li, S. Wang, D. Liu, H. Duan, and K. Song, “An assessment of water color for inland water in China using a Landsat 8-derived Forel–Ule index and the Google Earth Engine platform,” IEEE J. Sel. Top. Appl. Earth Obs. Remote Sens. 14, 5773–5785 (2021). [CrossRef]
34. G.J. Huffman, E.F. Stocker, D.T. Bolvin, E.J. Nelkin, and Jackson Tan, “GPM IMERG Final Precipitation L3 1 month 0.1 degree x 0.1 degree V06”, Greenbelt, MD, Goddard Earth Sciences Data and Information Services Center (GES DISC), Accessed: [2023-04-14], 10.5067/GPM/IMERG/3B-MONTH/06(2019).
35. Global Modeling and Assimilation Office (GMAO), “MERRA-2 tavgM_2d_flx_Nx: 2d,Monthly mean,Time-Averaged,Single-Level,Assimilation,Surface Flux Diagnostics V5.12.4”, Greenbelt, MD, USA, Goddard Earth Sciences Data and Information Services Center (GES DISC), Accessed: [2023-04-14], 10.5067/0JRLVL8YV2Y4(2015).
36. Q. Meng, L. Wang, Y. Chen, X. Wang, and X. Wang, “Change of Chlorophyll a Concentration and Its Environmental Response in the Bohai Sea from 2002 to 2021,” Environmental Monitoring of China. 38(6), 228–236 (2022).
