Open Access
Add to BibSonomy
Abstract
Diurnal monitoring of the Secchi-disk depth (SDD) of eutrophic lakes is the basic requirement to ensure domestic, industrial, and agricultural water use in surrounding cities. The retrieval of SDD in high frequency and longer observation period is the basic monitoring requirement to guarantee water environmental quality. Taking Lake Taihu as an example, the diurnal high-frequency observation (10 mins) data of the geostationary meteorological satellite sensor AHI/Himawari-8 were examined in this study. The results showed that the AHI normalized water-leaving radiance (Lwn) product derived by the Shortwave-infrared atmospheric correction (SWIR-AC) algorithm was consistent with the in situ data, with determination coefficient (R2) all larger than 0.86 and the mean absolute percentage deviation (MAPD) of 19.76%, 12.83%, 19.03% and 36.46% for the 460 nm, 510 nm, 640 nm and 860 nm bands, respectively. 510 nm and 640 nm bands showed more better consistency with in situ data in Lake Taihu. Therefore, an empirical SDD algorithm was established based on the AHI green (510 nm) and red (640 nm) bands. The SDD algorithm was verified by in situ data showed good performance with R2 of 0.81, RMSE of 5.91 cm, and MAPD of 20.67%. Based on the AHI data and established algorithm, diurnal high-frequency variation of the SDD in the Lake Taihu was investigated and the environmental factor (wind speed, turbidity degree, and photosynthetically active radiance) corresponding to diurnal SDD variation were discussed. This study should be helpful for studying diurnal high-dynamics physical-biogeochemical processes in eutrophication lake waters.
© 2023 Optica Publishing Group under the terms of the Optica Open Access Publishing Agreement
1. Introduction
Secchi-disk depth (SDD) refers to the deepest depth of the Secchi disk that can be observed by the human eye in the vertical direction in the waters and is used to quantitatively describe the lake water [1–3]. Lakes have geographical characteristics of a small area and shallow water depth and play the role of ensuring the production and domestic water demand of urban and rural areas in the basin, while simultaneously containing and degrading pollutants [4–6]. Under the pressure of huge water demand and pollution, the eutrophication of the lake has been serious. Eutrophic lakes are characterized by high concentrations of planktonic algae, high concentrations of suspended particulate matter (SPM), and low SDD values [6–9]. The geostationary satellite provides diurnal multiple repeated observation SDD data [1,4,7]. Using Chl-a and SPM products retrieved by geostationary ocean color image (GOCI), Ding et al. [1] discussed the diurnal SDD change of the eastern China Sea using the bio-optical SDD algorithm [2] and Superpixels algorithm. From GOCI observations between 2011 and 2021, wind force influenced the inner-annual SDD variations in Jiaozhou Bay [10]. Diurnal observation frequency and time length provided important data support in eutrophication lake governance [3–5]. Although the diurnal hourly temporal resolution data from geostationary satellites (such as GOCI-I/COMS and GOCI-II/GK2B) provided lake SDD repeat observations, this is still not enough for lake waters that are vulnerable to external forces to produce drastic horizontal transport and vertical exchange [11–13]. In addition, high-frequency observations were needed for the diurnal phytoplankton bloom [12]. Therefore, it is of great scientific significance to determine the longest effective observation time of lake Secchi-disk depth change with real-time high-frequency observations in a day [14–16].
The Himawari-8 /AHI satellite can observe the full disk area (-60°S–60°N, 80°E–160°W) with high temporal resolution (10 mins) and spatial resolution of 2 km. Although AHI has low spatial resolution and signal-noise-ratio (SNR) compared with traditional ocean color satellites, it still has high-frequency observation advantage in capturing the water environment change in highly dynamic waters [11,17]. Moreover, in summer with good solar illumination, diurnal high-frequency AHI observation can provide effective water color products at dawn and dusk [11,12]. In the Yangtze River estuary, AHI can obtain the concentration of SPM in the diurnal tidal period, and it can obtain more continuous diurnal change data than the hourly observation of GOCI [11]. The area distribution of floating algae extracted by AHI and GOCI, MODIS and Landsat in Lake Taihu has good consistency, and AHI has significant advantages in diurnal high-frequency observation [12]. For observing the highly dynamic waters off the west coast of Australia, the SPM data retrieved by AHI has significant advantages [17]. Thus, AHI high-frequency observation data are helpful for monitoring diurnal variation of water quality in the coastal oceans and inland waters.
In this study, taking eutrophication Lake Taihu as an example, the capacity of AHI to monitor the diurnal variation SDD was investigated. The atmospheric correction and a regional empirical SDD algorithm were developed and validated by cruises data, and the performance of SDD retrieval at dawn and dusk, as well as the impact of environmental factor on SDD diurnal variation were discussed the SWIR-AC algorithm and a regional empirical SDD algorithm were studied for AHI data to discuss the advantage of diurnal high-frequency observation and uncertainty of water-leaving radiance accuracy at dawn and dusk time. The accuracy of atmospheric correction data and SDD data retrieved by AHI was validated through cruise surveys. Finally, the study discusses the uncertainty of SDD retrieval at dawn and dusk times, and the diurnal environmental factor of diurnal SDD changes are discussed in this paper.
2. Materials and methods
2.1. Study areas
Lake Taihu (30°05′–32°08′N, 119°08′–121°55′E) is a typical shallow water huff and puff lake located in the lower reaches of the Yangtze River basin (http://www.tba.gov.cn/), with a surface area of ∼ 2400 km2 (Fig. 1). The lake is high in the west and southwest and low in the east. The average depth of the lake is ∼1.9 m, and the deepest reaches ∼3.34 m [5]. The main rivers directly injected into the lake include the Taipu River, Shaoxi River, Nanxi River, and Xilicao River. Influenced by human activities, such as urban water use, industrial water use, and agricultural water use in large cities (Suzhou City, Wuxi City, Changzhou City, and Huzhou City), the eutrophication degree of Lake Taihu is worsening, especially algae blooms [4,5]. Driven by meteorological factors, the lake is prone to horizontal transport and vertical mixing, leading to a significant diurnal change in the optical properties [3,4].
Fig. 1. The spatial distribution of Lake Taihu and the sample stations in August 2022 (a) (Including 2 DCOS stations from 07:30 to 17:30 (UTC + 08:00) and 10 NCOS stations) (the bathymetry data is from https://www.ncei.noaa.gov/access/metadata/landing-pag/), and two cruises in situ data from Hangzhou Bay (HZB) in February and August 2016 (b).
2.2. Cruise data
The cruise survey of Lake Taihu was carried out from August 28 to September 3, 2022, with a total of 10 groups of normal cruise observation stations (NCOS) and 2 groups of diurnal continuous observation stations (DCOS). The DCOS was observed for two days in the diurnal hourly change from 07:30-17:30 (Beijing Time, UTC + 08:00) (Fig. 1(a)). During the cruise survey, 40 groups of normalized water-leaving radiance (Lwn) (Fig. 2(a)), 51 groups of SDD, and relevant environmental data (such as wind speed, surface water turbidity, and surface water temperature) were obtained. These data were used to build an SDD algorithm and verify the accuracy of atmospheric correction products of AHI images and AHI retrieving SDD products. The Hangzhou Bay (HZB) cruise data (21 groups) were used to verify the accuracy of AHI Lwn products in February and August 2016 (Fig. 1(b)) [11]. In general, we can consider the pixel where the value is located as an abnormal value affected by the green tide if the value and the average of the variable exceed 2 times the standard deviation. To avoid the measurement uncertainty of Lwn and SDD data covered by the green tide, the survey station data covered by the green tide was directly removed in the study. In addition, the matchup between AHI and in situ data is as follows: (a) the time difference should be controlled within ±10 mins, and (b) the spatial distance should be controlled within $3pixel\textrm{s} \times 3pixel$s (6 km${\times} 6$ km), and the AHI pixels mean value was been used to matchup in situ data.
(1) Measurement of Lwn data
A portable spectrometer (spectral range 350-2500 nm; ASD Inc., USA) was used to measure in situ Lwn data. To obtain diurnal real-time Lwn change, the measurement time (Beijing time, UTC + 08:00) was from 07:30 am to 17:30 pm under suitable solar illumination conditions. Upward radiance above the sea surface (${L_{water}}$) was measured with a viewing zenith angle and azimuth angle (referring to the solar incident plane) of approximately 40° and 135° direction, respectively [13,18]. The corresponding downward sky radiance (${L_{sky}}$) was measured to estimate the water surface reflecting radiance. Moreover, the downward irradiance (${E_s}$) above the sea surface was estimated based on the sensor-measured upward radiance from the reference plate [11]. Based on the measured ${L_{water}}$, ${L_{sky}}$, and ${E_s}$, Lwn was calculated as follows:
where ${F_0}$ is the extraterrestrial solar irradiance [13]. ${\beta _s}$ is reflectance coefficient and the value of 0.028 is according to the conclusion by Mobley [19] and He et al. [13].(2) Measurement of SDD and environmental factors
The SDD is measured using a black and white Secchi disk with diameter of 30 cm. To avoid the measurement inaccuracy caused by direct sunlight and ship shadow, the SDD measurement must be performed in the backlight of the ship’s main deck [1–3]. The specific measurement procedure consists of sinking the disk to a depth where it cannot be seen, and then reading the marked value of the rope on the lake water, which is recorded as the SDD value in the water. This is repeated three times, and the mean value of the observed transparency is recorded as the true value of this observation. In this study, a total of 40 groups of in situ SDD data were obtained in Lake Taihu.
Corresponding to the measured Lwn data and SDD data of the cruise station, the turbidity degree (TD, unit: NTU) was measured by using the RBRconcerto3 device (RBR Inc., Canada). We need to correct the unit of turbidity in distilled water before measuring the turbidity using the probe device. The probe also needs to use distilled water to clean after measuring at each station. Each station RBRconcerto3 device will measure 10 ± 5 mins, and the mean value of this observation time was used in this study. In situ wind speed data were collected from a nearby meteorological station.
In addition, the AHI 10 min PAR product (ftp://ftp.ptree.jaxa.jp/pub/himawari/L2/PAR/) and ERA5 1- hPa wind ground data (https://cds.climate.copernicus.eu/) were used to study the relationship between AHI SDD and environmental factors. AHI PAR products showed good consistency with cruise data from the collaborative marine Atlas project (CMAP) (https://simonscmap.com/catalog/datasets/), and history diurnal PAR observation data from AERONET Lulin station (https://aeronet.gsfc.nasa.gov/) [14]. The European Centre for Medium-Range Weather Forecasts (ECMWF) (https://www.ecmwf.int/) is responsible for the quality control of ERA5 1-hPh wind data. Lwn and SDD products derived by AHI were downward interpolated to the same spatial resolution as AHI PAR ($5km \times 5km$) and ERA5 wind ground (0.25°${\times} $0.25°) products. In addition, the spatial matchup between AHI SDD and AHI PAR, ERA5 Wind speed followed the principle of nearest distance.
2.3. Satellite data and atmospheric correction
The first 6 bands (460, 510, 640, 860, 1600, and 2300 nm) of the AHI were used for atmospheric correction to retrieve the Lwn product by using the SWIR-AC algorithm (Table 1), and the details are as follows.
(1) AHI/Himawari-8 data and atmospheric correction
Table 1. The band parameters of AHI/Himawari-8 in this study.
The total radiance (${L_t}$) from the top of the atmosphere (TOA) measured by the AHI can be expressed as follows:
All the radiance terms were converted to reflectance according to the following definition:
Then, Eq. (5) was obtained from (3) and (4) as follows:
According to band equivalent Rayleigh scattering optical thickness and extraterrestrial solar irradiance [20] (Formula (8), and Table 2), Rayleigh scattering correction at each AHI first 4 bands was conducted based on the general Rayleigh-scattering look-up table generated by He et al. [21].
Table 2. Himawari-8/AHI first 6 bands equivalent parameters.
Assuming that water-leaving radiance (Lw) at the two SWIR bands (1600 and 2300 nm) was negligible, we obtained the aerosol scattering reflectance at the two SWIR bands, which was equal to the Rayleigh-corrected reflectance [17,21]. Finally, the exponential extrapolation approximation was applied to obtain aerosol scattering reflectance at the VIS and NIR bands as follows:
2.4. SDD retrieval algorithm
The red band is more sensitive to SPM and SDD in high turbidity degree waters, while the green band is more sensitive to SDD change in low turbidity degree waters [3], and the AHI green (510 nm) and red (640 nm) bands have high accuracy derived by the SWIR-AC algorithm. Therefore, based on 40 groups of in situ SDD data and Lwn curves from the cruise survey, this study built the Lake Taihu SDD algorithm applicable to the green (510 nm) and red (640 nm) bands of the AHI data, and 30 groups of SDD data have been used for verification of the AHI retrieval SDD products. There was good consistency between the modeled SDD data and in situ Lwn data (R2 = 0.84, RMSE = 0.84 cm, MAPD = 4.14%), and a regional empirical algorithm based on the cruise-measured Lwn and SDD data was established (Fig. 2) as follows:
3. Statistical analysis and accuracy assessment
The mean absolute percentage deviation (MAPD) [1,2] is a very intuitive interpretation in terms of relative error and was used as a loss function to evaluate the overall relative percentage error between two datasets as follows:
In addition, the statistical parameter (MAPD, RMSE, R2) were coded in Python 3.7 software.
4. Results
4.1. Validation of AHI-retrieved Lwn product
The accuracy of AHI-retrieved Lwn products derived by the SWIR-AC algorithm was validated by using in situ data from Lake Taihu and Yangtze River estuary surveys. Overall, The AHI first 4 band Lwn products retrieved from the SWIR-AC algorithm obtained satisfactory results validated by cruise in situ Lwn data (Fig. 3). The results showed that the AHI normalized water-leaving radiance (Lwn) product derived by the SWIR-AC algorithm was consistent with the in situ data, with determination coefficient (R2) all larger than 0.86. Specifically, the RMSE values were 0.95,1.00, 1.07, and 0.47 for the 460, 510, 640, and 860 nm band, respectively, and the mean absolute percentage deviation (MAPD) of 19.76%, 12.83%, 19.03% and 36.46% for the 460, 510, 640 and 860 nm bands, respectively. 510 nm and 640 nm bands showed better agreement with in situ data, 460 nm and 860 nm bands showed relatively low agreement with in situ data in Lake Taihu. Exponential extrapolation approximation of aerosol scattering reflectance at VIS and NIR bands from two SWIR bands (1600 nm and 2300 nm) inevitably accumulated uncertainty errors to VIS and NIR bands in AHI atmospheric correction [11,13,21].
Fig. 3. Validation of AHI’s first 4 band products derived by the SWIR-AC algorithm using in situ survey data in Lake Taihu and Yangtze River estuary.
Comparing the consistency of the DCOS station measurement in situ Lwn data and AHI-retrieved Lwn data on August 28, 2022, and September 1, 2022, the Lwn spectra between the AHI-retrieved and in situ measured values were matched in the diurnal hourly changes, as shown in Fig. 4 for DCOS stations, indicating the application of the established atmospheric correction algorithm for AHI data, with R2 values larger than 0.81 for all four AHI bands. Overall, the results of AHI’s first 4 bands of atmospheric correction verified by in situ data have obtained satisfactory results. In addition, diurnal hourly AHI Lwn data on the fixed station clearly showed the characteristics of rapid changes in lake optical properties. Diurnal hourly AHI Lwn at station-a showed a peak (10:30) and a trough (13:30). At Station-B, the diurnal hourly AHI Lwn data showed a high at dawn and dusk, but a low at noon.
Fig. 4. The diurnal change consistency of DCOS stations (A and B) Lwn products derived by AHI and in situ observation, respectively.
The diurnal hourly pattern of AHI-retrieved Lwn products showed an hourly continuous change pattern in the spatial distribution of Lake Taihu on May 5, 2022 (Fig. 5). AHI-retrieved Lwn products showed a rapidly increasing trend after 07:30 pm. AHI first 4 bands showed great differences under the weak solar illumination conditions at dawn and dusk (Fig. 4, Fig. 5). Referred to the abnormal change of Lwn under the high solar zenith angle condition, AHI 640 nm and 860 nm bands have relatively high sensitivity to weak solar illumination at dawn and dusk, while 460 nm and 510 nm bands have relatively low sensitivity.
Fig. 5. Diurnal hourly spatial pattern of AHI Lwn 460, 510, 640, and 860 nm products derived by the SWIR-AC algorithm in Lake Taihu on May 5, 2022.
4.2. Validation of AHI-retrieved SDD product
The accuracy of the AHI-retrieved SDD products derived by the AHI Lwn510 and Lwn640 bands was validated by using in situ SDD data. Figure 6 shows the comparison between AHI-derived and cruise-measured SDD values. The AHI SDD algorithm verified by in situ data obtained satisfactory results, indicating the application of this algorithm for AHI SDD retrieval, with high consistency (R2 = 0.81), RMSE of 5.91 cm, and MAPD of 20.67%.
Comparing the consistency of DCOS station measured in situ SDD data and AHI retrieved SDD data on August 28, 2022 and September 1, 2022, the SDD spectra between the AHI retrieved and in situ measured values were consistent in the diurnal hourly changes as shown in Fig. 7 for DCOS stations, indicating the application of the established regional empirical for AHI data, with R2 values greater than 0.85 for two DCOS stations.
Fig. 7. The diurnal change consistency of DCOS stations (A and B) SDD products derived from AHI and in situ observations, respectively.
4.3. Diurnal change in the AHI-retrieved SDD product
Compared to the hourly GOCI-II observations (23:30-08:30, UTC), the AHI can provide continuous ∼60 times observations SDD products of diurnal variation (Fig. 8, Fig. 9). The SDD products observed by the AHI and GOCI-II showed similar tendencies in tempo-spatial distribution pattern; however, compared to the diurnal SDD products from minute-level continuous observations by the AHI, the GOCI-II hourly observations had an inevitable time gap that could not observe high-frequency dynamics of SDD within 1 h.
Fig. 8. Diurnal 10-minute level spatial change of SDD products derived by AHI, 1-hourly level SDD products derived by GOCI-II and fixed site (120.1°E, 31.2°N) diurnal changes in Lake Taihu on May 5, 2022.
Fig. 9. Same as Fig. 8, but for the results on August 3, 2022, and with another fixed site (120.14°E, 31.3°N).
The SDD products of AHI and GOCI-II showed good consistency in the diurnal DCOS stations, and the continuous observation time of AHI SDD products could reach ∼12 h in the diurnal change (Fig. 8, Fig. 9). In addition, AHI has more significant advantages than GOCI-II in the details of diurnal minute-level change observations. With higher temporal resolution and longer observation length, the AHI makes daily high-frequency monitoring and research on the Secchi-disk depth of eutrophic lakes closer to real-time. Therefore, the high temporal resolution and long observation period have significant advantages in monitoring SDD variations in highly dynamic lake waters.
5. Discussion
5.1. Uncertainties in SDD retrieval at dawn and dusk
The accuracy and uncertainty of the atmospheric correction of watercolor under the condition of weak solar illumination at dawn and dusk is a major limitation of the current geostationary satellite retrieval of diurnal watercolor products [13,16]. Li et al. retrieved open ocean color products under high solar zenith angles (80°∼88°) using the Neural Net atmospheric correction algorithm [16]. Although the SWIR-AC algorithm has realized the high-frequency observation in the tidal period in Hangzhou Bay, the uncertainty of the atmospheric correction and the SDD products under the condition of the large solar zenith (≥70) at dawn and dusk has not been thoroughly investigated [11,16,22,23]. The high-frequency AHI data of diurnal 10-min frequency observations promoted a more refined diurnal effective observation time length under the appropriate solar illumination (Fig. 10) [1,11]. Diurnal changes in suspended particulate matter in the Yangtze River estuary have been studied in tidal period observations (∼12 h), and there are still effective Lwn and SPM data in the larger SOZ (70°-75°) at dawn and dusk [11]. In this study, the DCOS station showed effectively continuous Lwn and SDD changes under low solar zenith angle conditions (SOZ ≤ 70°) but huge fluctuations under high low solar zenith angle conditions (SOZ > 70°) (Fig. 11). The influence of Earth curvature on the large low solar zenith angle needs to be considered for high accuracy atmospheric correction; specifically, Earth curvature influences up to 1%, 3%, and 12% for solar zenith angles at 75°, 80°, and 85°, respectively [21]. The atmospheric correction products can be retrieved under a relatively high solar zenith angle (70°-75°) by using a traditional algorithm, and the secondary SDD products can also be computed under weak illumination conditions at dawn (Fig. 12). In addition, geostationary ocean color remote sensing will unavoidably encounter observations at high solar zenith angles and high viewing zenith angles at the edge of the observable area when the effects of Earth curvature cannot be ignored [21]. Based on the time series SDD products from AHI, the characteristics of the SDD dynamic along the diurnal longest effective observation time and minute high-frequency variations can be studied, which should be our next step.
Fig. 10. Diurnal longest effective observation time length under good solar illumination (SOZ ≤ 70) along 180°E (-60°S—60°N) at the spring equinox, summer solstice, autumn equinox, and winter solstice.
Fig. 11. Fixed site (120.1°E, 31.2°N) diurnal longest effective observation length, and the uncertainty of atmospheric correction products and SDD products under the weak illumination of larger solar zenith angle (SOZ ≤ 70°) condition at dawn and dusk time.
Fig. 12. Spatial distribution pattern of SOZ, Lwn460, Lwn510, Lwn640, Lwn860, and SDD products under weak illumination conditions at dawn on May 4, 2022.
5.2. Environmental factor-driven analysis
In shallow eutrophic lake water, Secchi-disk depth is mainly affected by phytoplankton physiological activities and hydrodynamic transport and exchange in the horizontal and vertical directions [3,5,24]. The horizontal transport of algal blooms is closely related to the direction and speed of the wind field, which greatly changes the Secchi-disk depth within a short time [1,7]. The wind speed has a close negative correlation with vertical water mixing in shallow lakes [5,25]. The change in wind speed is closely related to the exchange strength between the upper layer and bottom layer in shallow lakes [25–27], and lake bottom sediment is suspended due to the increase in wind speed, which further leads to a rapid increase in the degree of water turbidity and a decrease in Secchi-disk depth [1,28]. SDD and wind speed showed a negative correlation with an R2 of 0.66 in Lake Taihu (Fig. 13(a)). In addition, lake algal bloom water will also diffuse and be transported in the horizontal direction driven by the wind field, which will lead to a rapid reduction in the Secchi-disk depth of the surrounding waters [27,29,30]. SDD and water turbidity degree showed a negative correlation with an R2 of 0.62 in Lake Taihu (Fig. 13(b)). The diurnal photosynthetically active radiance (PAR) determines the solar energy that can be used by the phytoplankton population, and the ultimate cause of diel vertical migration is a better light supply for all populations [4,14,15]. As Secchi-disk depth to PAR was identified as an important variable controlling thermal structure, PAR and SDD showed a significant linear relationship with an R2 of 0.79 in small arctic lakes of southwest Greenland [31]. SDD and PAR showed a positive correlation with an R2 of 0.69 in Lake Taihu (Fig. 13(c)). In eutrophic lake waters, the higher the PAR value, the higher the concentration of plankton algae, and the higher the water turbidity degree, but the lower the Secchi-disk depth [14,15,28,32].
Fig. 13. The impact of environmental factors (wind speed (a), turbidity (b), and PAR (c)) on SDD change in Lake Taihu. The red line is the linear regression.
6. Conclusion
Taking Lake Taihu as an example, the capacity of the geostationary meteorological satellite sensor AHI/Himawari-8 along the diurnal effective time SDD dynamics with a 10-min high frequency in eutrophication waters was examined. The results of AHI’s first 4 bands of atmospheric correction verified by in situ data have obtained satisfactory results. The diurnal hourly variation of AHI-retrieved Lwn products showed an hourly continuous change pattern in the spatial distribution of Lake Taihu. The AHI SDD algorithm verified by in situ data obtained satisfactory results, indicating the application of this algorithm for AHI SDD retrieval, with high consistency (R2 = 0.81) and low deviation (RMSE = 5.91 cm, MAPD = 20.67%). Compared to the hourly GOCI-II observations (23:30-08:30, UTC), the AHI can provide continuous ∼60 times observations of diurnal change. Although the AHI demonstrated uncertainty under weak solar illumination conditions at dawn and dusk, the high-frequency AHI data can obtain the longest diurnal effective observation time as much as possible. At last, the relationship between diurnal SDD variation and environmental factors (wind speed, water turbidity, and PAR) were been discussed. The relatively low spatial resolution of AHI makes it suitable for large lakes such as Lake Taihu. However, when monitoring small lakes or rivers, it is necessary to use satellite data with higher spatial resolution because AHI has insufficient spatial resolution. This study should help to clarify diurnal high-dynamics physical-biogeochemical processes in eutrophication lake waters.
Funding
The Postgraduate Research & Practice Innovation Program of Jiangsu Province (KYCX21_0452); National Key Research and Development Program of China (2022YFC3104900, 2022YFC3104901); National Natural Science Foundation of China (41825014, 42176177); Graduate Research Innovation Fund Project of Yunnan University (KC-22221838).
Acknowledgments
We thank the Japan Meteorological Agency for providing the Himawari-8 AHI data. We also thank the members of the State Key Laboratory of Satellite Ocean Environment Dynamics, Second Institute of Oceanography, Ministry of Natural Resources (SOED/SIO/MNR) satellite ground station, satellite data processing & sharing center, and marine satellite data online analysis platform for their help with data collection and processing.
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. X Ding, F Gong, Q Zhu, J Li, X Wang, R Bai, and Y Xu, “Using geostationary satellite ocean color data and superpixel to map the diurnal dynamics of water transparency in the eastern China seas,” Ecol. Indic. 142, 109219 (2022). [CrossRef]
2. X He, D Pan, and Z Mao, “Water-transparency (Secchi Depth) monitoring in the China Sea with the SeaWiFS satellite sensor,” Proc. SPIE 5568, 112–122 (2004). [CrossRef]
3. D Liu, H Duan, S Loiselle, C Hu, G Zhang, J Li, H Yang, R Thompson, Z Cao, M Shen, R Ma, M Zhang, and W Han, “Observations of water transparency in China’s lakes from space,” Int J. Appl. Earth Obs. 92, 102187 (2020). [CrossRef]
4. H Cao and L Han, “Hourly remote sensing monitoring of harmful algal blooms (HABs) in Taihu Lake based on GOCI images,” Environ. Sci. Pollut R. 28(27), 35958–35970 (2021). [CrossRef]
5. Q Xiao, X Xu, M Zhang, H Duan, Z Hu, W Wang, W Xiao, and X Lee, “Coregulation of nitrous oxide emissions by nitrogen and temperature in China's third largest freshwater lake (Lake Taihu),” Limnol. Oceanogr. 64(3), 1070–1086 (2019). [CrossRef]
6. Y Hu and H Cheng, “Water pollution during China's industrial transition,” Environ. Dev. 8, 57–73 (2013). [CrossRef]
7. S Lei, J Xu, Y Li, C Du, G Liu, Z Zheng, Y Xu, H Lyu, M Mu, S Miao, S Zeng, J Xu, and L Li, “An approach for retrieval of horizontal and vertical distribution of total suspended matter concentration from GOCI data over Lake Hongze,” Sci. Total Environ. 700, 134524 (2020). [CrossRef]
8. L Håkanson, “The relationship between salinity, suspended particulate matter and water clarity in aquatic systems,” Ecol. Res. 21(1), 75–90 (2006). [CrossRef]
9. Y Zhang, Y Zhang, K Shi, Y Zhou, and N Li, “Remote sensing estimation of water clarity for various lakes in China,” Water Res. 192, 116844 (2021). [CrossRef]
10. Y Zhou, D Yu, W Cheng, Y Gai, H Yao, L Yang, and S Pan, “Monitoring multi-temporal and spatial variations of water transparency in the Jiaozhou Bay using GOCI data,” Mar. Pollut. Bull. 180, 113815 (2022). [CrossRef]
11. X Ding, X He, Y Bai, Q Zhu, F Gong, H Li, and J Li, “High-frequency and tidal period observations of suspended particulate matter in coastal waters by AHI/Himawari-8,” Opt. Express 28(19), 27387–27404 (2020). [CrossRef]
12. X Chen, S Shang, Z Lee, L Qi, J Yan, and Y Li, “High-frequency observation of floating algae from AHI on Himawari-8,” Remote Sens. Environ. 227, 151–161 (2019). [CrossRef]
13. X He, Y Bai, D Pan, N Huang, X Dong, J Chen, CTA Chen, and Q Cui, “Using geostationary satellite ocean color data to map the diurnal dynamics of suspended particulate matter in coastal waters,” Remote Sens. Environ. 133, 225–239 (2013). [CrossRef]
14. X Ding, X He, Y Bai, J Li, Y Xu, X Wang, and Q Zhu, “Requirement on the diurnal observation frequency for satellite remote sensing of photosynthetically available radiation,” Opt. Express 30(6), 9021–9034 (2022). [CrossRef]
15. R Semyalo, J Nattabi, and P Larsson, “Diel vertical migration of zooplankton in a eutrophic bay of Lake Victoria,” Hydrobiologia 635(1), 383–394 (2009). [CrossRef]
16. H Li, X He, Y Bai, P Shanmugam, Y Park, J Liu, Q Zhu, F Gong, D Wang, and H Huang, “Atmospheric correction of geostationary satellite ocean color data under high solar zenith angles in open oceans,” Remote Sens. Environ. 249, 112022 (2020). [CrossRef]
17. P Dorji and P Fearns, “Atmospheric correction of geostationary Himawari-8 satellite data for Total Suspended Sedimentation mapping: A case study in the Coastal Waters of Western Australia,” ISPRS-J. Photogramm. Remote Sens. 144, 81–93 (2018). [CrossRef]
18. J Mueller, A Morel, R Frouin, C Davis, R Arnone, K Carder, Z Lee, R Steward, S Hooker, C Mobley, S McLean, B Holben, M Miller, C Pietras, K Knobelspiesse, G Fargion, J Porter, and K Voss, “Ocean Optics Protocols for Satellite Ocean Color Sensor Validation, Revision 4, Volume III: Radiometric Measurements and Data Analysis Protocols” (2003).
19. C D Mobley, “Estimation of the remote-sensing reflectance from above-surface measurements,” Appl. Opt. 38(36), 7442–7455 (1999). [CrossRef]
20. H R Gordon, “Remote sensing of ocean color: a methodology for dealing with broad spectral bands and significant out-of-band response,” Appl. Opt. 34(36), 8363–8374 (1995). [CrossRef]
21. X He, D Pan, B Yan, and G Fang, “A general purpose exact Rayleigh scattering look-up table for ocean color remote sensing,” Acta Oceanol. Sin. 1, 48–56 (2006). [CrossRef]
22. X He, K Stamnes, Y Bai, W Li, and D Wang, “Effects of Earth curvature on atmospheric correction for ocean color remote sensing,” Remote Sens. Environ. 209, 118–133 (2018). [CrossRef]
23. H Li, X He, P Shanmugam, Y Bai, D Wang, H Huang, Q Zhu, and F Gong, “Radiometric sensitivity and signal detectability of ocean color satellite sensor under high solar zenith angles,” IEEE Trans. Geosci. Remote Sensing 57(11), 8492–8505 (2019). [CrossRef]
24. X Liu, Z Lee, Y Zhang, J Lin, K Shi, Y Zhou, B Qin, and Z Sun, “Remote sensing of secchi depth in highly turbid lake waters and its application with MERIS data,” Remote Sens. 11(19), 2226 (2019). [CrossRef]
25. X Hou, L Feng, H Duan, X Chen, D Sun, and K Shi, “Fifteen-year monitoring of the turbidity dynamics in large lakes and reservoirs in the middle and lower basin of the Yangtze River, China,” Remote Sens. Environ. 190, 107–121 (2017). [CrossRef]
26. P Hunter, A Tyler, N Willby, and D Gilvear, “The spatial dynamics of vertical migration by Microcystis aeruginosa in a eutrophic shallow lake: A case study using high spatial resolution time-series airborne remote sensing,” Limnol. Oceanogr. 53(6), 2391–2406 (2008). [CrossRef]
27. B Qin, G Zhu, L Zhang, L Luo, G Gao, and B Gu, “Estimation of internal nutrient release in large shallow Lake Taihu, China,” Sci. China, Ser. D: Earth Sci. 49(S1), 38–50 (2006). [CrossRef]
28. K Shi, Y Zhang, G Zhu, B Qin, and D Pan, “Deteriorating water clarity in shallow waters: Evidence from long term MODIS and in-situ observations,” Int J. Appl. Earth Obs. 68, 287–297 (2018). [CrossRef]
29. K Shi, Y Zhang, X Liu, M Wang, and B Qin, “Remote sensing of diffuse attenuation coefficient of photosynthetically active radiation in Lake Taihu using MERIS data,” Remote Sens. Environ. 140, 365–377 (2014). [CrossRef]
30. V Fleming-Lehtinen and M Laamanen, “Long-term changes in Secchi depth and the role of phytoplankton in explaining light attenuation in the Baltic Sea,” Estuarine, Coastal Shelf Sci. 102, 1–10 (2012). [CrossRef]
31. J Saros, R Northington, C Osburn, B Burpee, and N John Anderson, “Thermal stratification in small arctic lakes of southwest Greenland affected by water transparency and epilimnetic temperatures,” Limnol. Oceanogr. 61(4), 1530–1542 (2016). [CrossRef]
32. H Liu, Q Zhou, Q Li, S Hu, T Shi, and G Wu, “Determining switching threshold for NIR-SWIR combined atmospheric correction algorithm of ocean color remote sensing,” ISPRS-J. Photogramm. Remote Sens. 153, 59–73 (2019). [CrossRef]
