Figure 1.
Study lakes and in situ sampling points. (a) Lake Hongze, (b) Lake Chaohu, and (c) Lake Taihu. Dots, triangles, crosses, and forks represent the in situ sampling locations.
Figure 1.
Study lakes and in situ sampling points. (a) Lake Hongze, (b) Lake Chaohu, and (c) Lake Taihu. Dots, triangles, crosses, and forks represent the in situ sampling locations.
Figure 2.
The remote sensing reflectance (Rrs) measured in (a) Lake Hongze on 24 October 2014, (b) Lake Chaohu on 11 October 2015 and 15 January 2016, and (c) Lake Taihu on 11 May 2017 and 27 May 2017. The color of lines correspond to the sampling date.
Figure 2.
The remote sensing reflectance (Rrs) measured in (a) Lake Hongze on 24 October 2014, (b) Lake Chaohu on 11 October 2015 and 15 January 2016, and (c) Lake Taihu on 11 May 2017 and 27 May 2017. The color of lines correspond to the sampling date.
Figure 3.
Average (solid line) and standard deviation (shadow) of Rrs for water types: turbid water (a), in-water algae (b), and floating bloom (c).
Figure 3.
Average (solid line) and standard deviation (shadow) of Rrs for water types: turbid water (a), in-water algae (b), and floating bloom (c).
Figure 4.
Scatter plots of Rrs retrieved by the water-AC algorithms (SWIR (a), EXP (b), DSF (c), and MUMM (d)) versus in situ Rrs.
Figure 4.
Scatter plots of Rrs retrieved by the water-AC algorithms (SWIR (a), EXP (b), DSF (c), and MUMM (d)) versus in situ Rrs.
Figure 5.
Scatter plots of Rrs retrieved by land-AC algorithms (FLAASH (a), 6SV (b), and QUAC (c)) versus in situ Rrs.
Figure 5.
Scatter plots of Rrs retrieved by land-AC algorithms (FLAASH (a), 6SV (b), and QUAC (c)) versus in situ Rrs.
Figure 6.
Frequency of Rrs(λ) retrieved using EXP algorithm (blue line) and 6SV algorithm (red line) from OLI measurements in Lake Hongze on 24 October 2014, Lake Chaohu on 15 October 2015, January 11, 2016, and Lake Taihu on 11 May 2017 and 27 May 2017, respectively.
Figure 6.
Frequency of Rrs(λ) retrieved using EXP algorithm (blue line) and 6SV algorithm (red line) from OLI measurements in Lake Hongze on 24 October 2014, Lake Chaohu on 15 October 2015, January 11, 2016, and Lake Taihu on 11 May 2017 and 27 May 2017, respectively.
Figure 7.
OLI true color image presenting, Rrs(λ) for Lake Hongze on 24 October 2014, Lake Chaohu on 15 October 2015, 11 January 2016, and Lake Taihu on 11 May 2017 and 27 May 2017, respectively. Rrs(λ) was derived by EXP and 6SV algorithm.
Figure 7.
OLI true color image presenting, Rrs(λ) for Lake Hongze on 24 October 2014, Lake Chaohu on 15 October 2015, 11 January 2016, and Lake Taihu on 11 May 2017 and 27 May 2017, respectively. Rrs(λ) was derived by EXP and 6SV algorithm.
Figure 8.
Scatter plots of SPM calibration between in situ measurement data and SPM model (a), validation between measured and simulated Rrs-based derived measurements (b).
Figure 8.
Scatter plots of SPM calibration between in situ measurement data and SPM model (a), validation between measured and simulated Rrs-based derived measurements (b).
Figure 9.
OLI true color image, and distribution of estimated SPM patterns in Lake Chaohu on 15 October 2015, 11 January 2016. OLI-estimated SPM derived by EXP (a,c) and 6SV (b,d), and frequency of OLI-estimated SPM on 15 October 2015 (f) and 11 January 2016 (g), blue line was driven by EXP algorithm and red line was driven by 6SV algorithm. The comparison of in situ-measured SPM and OLI-estimated SPM derived by EXP and 6SV (e).
Figure 9.
OLI true color image, and distribution of estimated SPM patterns in Lake Chaohu on 15 October 2015, 11 January 2016. OLI-estimated SPM derived by EXP (a,c) and 6SV (b,d), and frequency of OLI-estimated SPM on 15 October 2015 (f) and 11 January 2016 (g), blue line was driven by EXP algorithm and red line was driven by 6SV algorithm. The comparison of in situ-measured SPM and OLI-estimated SPM derived by EXP and 6SV (e).
Figure 10.
Distribution of RMSE (right) and MAPE (left) of bands on different water types. Blue is TW (turbid water), orange is IW (in-water algae) and black is FB (floating bloom). The different patterns of column represent different water-AC algorithms.
Figure 10.
Distribution of RMSE (right) and MAPE (left) of bands on different water types. Blue is TW (turbid water), orange is IW (in-water algae) and black is FB (floating bloom). The different patterns of column represent different water-AC algorithms.
Figure 11.
Distribution of RMSE (right) and MAPE (left) of bands on different water types used for land-AC algorithms. Blue is TW (turbid water), orange is IW (in-water algae), and black is FB (floating bloom). The different patterns of column represent different land-AC algorithms.
Figure 11.
Distribution of RMSE (right) and MAPE (left) of bands on different water types used for land-AC algorithms. Blue is TW (turbid water), orange is IW (in-water algae), and black is FB (floating bloom). The different patterns of column represent different land-AC algorithms.
Figure 12.
OLI-RGB images and Rayleigh-corrected reflectance (ρrc) images of two SWIR bands in Lake Taihu (a,c,d) and Lake Chaohu (b,e,f). The Rayleigh-corrected reflectance derived by SeaDAS processing. (g) and (h) were line graphs of ρrc value from figure (a) and (b) (100 × 100 pixel area), respectively.
Figure 12.
OLI-RGB images and Rayleigh-corrected reflectance (ρrc) images of two SWIR bands in Lake Taihu (a,c,d) and Lake Chaohu (b,e,f). The Rayleigh-corrected reflectance derived by SeaDAS processing. (g) and (h) were line graphs of ρrc value from figure (a) and (b) (100 × 100 pixel area), respectively.
Figure 13.
The frequency of seasonal variation of aerosol types in Lake Taihu from 2005 to 2018. The HA, MA, SA, and HS represent highly absorbing, moderately absorbing, slightly absorbing, and highly scattering fine-mode aerosols, respectively.
Figure 13.
The frequency of seasonal variation of aerosol types in Lake Taihu from 2005 to 2018. The HA, MA, SA, and HS represent highly absorbing, moderately absorbing, slightly absorbing, and highly scattering fine-mode aerosols, respectively.
Table 1.
Bands of the OLI (Operational Land Imager) on Landsat-8, with band range, band center, ground sampling distance (GSD), and signal-to-noise ratio (SNR) at reference radiance [
27].
Table 1.
Bands of the OLI (Operational Land Imager) on Landsat-8, with band range, band center, ground sampling distance (GSD), and signal-to-noise ratio (SNR) at reference radiance [
27].
Band | Band Range (nm) | Band Center (nm) | GSD (m) | SNR at Reference L |
---|
Band1 Coastal/Aerosol | 433–453 | 443 | 30 | 232 |
Band 2 Blue | 450–515 | 482 | 30 | 355 |
Band 3 Green | 525–600 | 561 | 30 | 296 |
Band 4 Red | 630–680 | 655 | 30 | 222 |
Band 5 NIR | 845–885 | 865 | 30 | 199 |
Band 6 SWIR 1 | 1560–1660 | 1609 | 30 | 261 |
Band 7 SWIR 2 | 2100–2300 | 2201 | 30 | 326 |
Band 8 Pan | 500–680 | 590 | 15 | 146 |
Band 9 Cirrus | 1360–1390 | 1375 | 30 | 162 |
Table 2.
The match-up dates of in situ measurement and OLI data acquisition over lakes Hongze, Chaohu, and Taihu.
Table 2.
The match-up dates of in situ measurement and OLI data acquisition over lakes Hongze, Chaohu, and Taihu.
| OLI Image | Acquisition Date | in situ Number |
---|
Lake Hongze | LC81210372014297 | 24 October 2014 | 10 |
Lake Chaohu | LC81210382015284 | 11 October 2015 | 15 |
| LC81210382016015 | 15 January 2016 | 16 |
Lake Taihu | LC81190382017131 | 11 May 2017 | 11 |
| LC81190382017147 | 27 May 2017 | 22 |
Table 3.
Variations of bio-optical properties of the three study lakes (Lake Hongze, Lake Chaohu, and Lake Taihu). The unit for Chla is μg/L. The units for SPOM and SPIM are mg/L.
Table 3.
Variations of bio-optical properties of the three study lakes (Lake Hongze, Lake Chaohu, and Lake Taihu). The unit for Chla is μg/L. The units for SPOM and SPIM are mg/L.
Lake | Parameters | Minimum | Maximum | Mean | SD |
---|
Lake Hongze (n = 10) | Chla | 7.35 | 19.21 | 11.62 | 3.33 |
SPOM | 6.00 | 11.33 | 8.67 | 1.61 |
SPIM | 28.00 | 47.33 | 35.33 | 6.68 |
SPOM/SPIM | 0.16 | 0.40 | 0.25 | 0.07 |
ag(440) | 1.19 | 1.59 | 1.38 | 0.13 |
aph(665) | 0.16 | 0.25 | 0.19 | 0.03 |
Lake Chaohu (n = 31) | Chla | 9.86 | 687.14 | 140.68 | 184.01 |
SPOM | 8.00 | 216.00 | 38.53 | 48.08 |
SPIM | 8.00 | 93.00 | 40.94 | 27.93 |
SPOM/SPIM | 0.15 | 9.82 | 1.39 | 1.94 |
ag(440) | 0.79 | 1.75 | 1.22 | 0.27 |
aph(665) | 0.13 | 8.48 | 1.31 | 1.81 |
Lake Taihu (n = 33) | Chla | 19.96 | 1022.53 | 206.13 | 267.93 |
SPOM | 4.00 | 321.33 | 72.04 | 85.48 |
SPIM | 13.33 | 88.00 | 44.14 | 16.80 |
SPOM/SPIM | 0.12 | 6.51 | 1.67 | 1.76 |
ag(440) | 0.46 | 3.28 | 1.16 | 0.60 |
aph(665) | 0.34 | 26.95 | 3.48 | 5.80 |
Table 4.
Variations in bio-optical properties of three water types (turbid water, in-water algae, and floating bloom). The unit for Chla is μg/L. The units for SPOM and SPIM are mg/L.
Table 4.
Variations in bio-optical properties of three water types (turbid water, in-water algae, and floating bloom). The unit for Chla is μg/L. The units for SPOM and SPIM are mg/L.
Lake | Parameters | Minimum | Maximum | Mean | SD |
---|
Turbid water (n = 20) | Chla | 7.35 | 91.72 | 28.10 | 22.57 |
SPOM | 6.00 | 34.00 | 14.77 | 8.09 |
SPIM | 28.00 | 93.00 | 52.44 | 20.34 |
SPOM/SPIM | 0.12 | 0.47 | 0.28 | 0.10 |
ag(440) | 0.89 | 1.75 | 1.36 | 0.24 |
aph(665) | 0.13 | 0.92 | 0.29 | 0.21 |
In-water algae (n = 38) | Chla | 19.96 | 687.14 | 104.39 | 125.78 |
SPOM | 4.00 | 226.67 | 44.29 | 44.35 |
SPIM | 8.00 | 68.00 | 28.18 | 17.52 |
SPOM/SPIM | 0.23 | 8.67 | 1.57 | 1.96 |
ag(440) | 0.46 | 1.75 | 1.07 | 0.31 |
aph(665) | 0.31 | 16.43 | 1.89 | 3.01 |
Floating bloom (n = 16) | Chla | 103.86 | 1022.53 | 198.04 | 203.85 |
SPOM | 25.33 | 321.33 | 117.27 | 69.46 |
SPIM | 16.00 | 88.00 | 44.91 | 20.17 |
SPOM/SPIM | 0.40 | 9.82 | 2.61 | 2.50 |
ag(440) | 0.37 | 3.28 | 2.62 | 1.04 |
aph(665) | 0.32 | 26.95 | 6.59 | 8.07 |
Table 5.
Band ratio errors between in situ Rrs and OLI Rrs obtained with atmospheric correction algorithms for water. The bold means the minimal statistical value and the underline means maximum.
Table 5.
Band ratio errors between in situ Rrs and OLI Rrs obtained with atmospheric correction algorithms for water. The bold means the minimal statistical value and the underline means maximum.
Algorithm | | Band Ratio |
---|
443/561 | 482/561 | 655/561 | 865/561 | 443/655 | 482/655 | 865/655 |
---|
SWIR | RMSE | 0.3125 | 0.2180 | 0.0915 | 0.3916 | 0.3538 | 0.2339 | 0.7163 |
MAPE (%) | 69.91 | 37.35 | 10.50 | 227.69 | 59.95 | 30.15 | 201.77 |
Bias (%) | 37.62 | 20.16 | 6.30 | −88.12 | 27.24 | 11.80 | −71.86 |
EXP | RMSE | 0.1758 | 0.1694 | 0.0971 | 0.7781 | 0.2099 | 0.1645 | 1.5130 |
MAPE (%) | 54.48 | 30.93 | 10.65 | 197.08 | 46.11 | 21.59 | 195.52 |
Bias (%) | 24.72 | 21.20 | 6.53 | 42.88 | 15.10 | 12.60 | 63.03 |
DSF | RMSE | 0.19 | 0.1577 | 0.0984 | 0.6372 | 0.1951 | 0.1406 | 1.0799 |
MAPE (%) | 88.91 | 29.31 | 11.31 | 130.87 | 61.72 | 17.23 | 132.07 |
Bias (%) | 85.08 | 26.18 | 8.06 | 120.54 | 59.08 | 15.00 | 118.22 |
MUMM | RMSE | 0.5013 | 0.3888 | 0.1656 | 0.4601 | 0.7234 | 0.5489 | - |
MAPE (%) | 91.19 | 57.73 | 17.54 | 89.43 | 101.08 | 63.13 | - |
Bias (%) | −86.60 | −53.97 | 4.36 | −89.28 | −97.62 | −60.88 | - |
Table 6.
Band ratio errors between in situ Rrs and OLI Rrs obtained with the land-AC algorithms. The bold means the minimal statistical value and the underline means maximum.
Table 6.
Band ratio errors between in situ Rrs and OLI Rrs obtained with the land-AC algorithms. The bold means the minimal statistical value and the underline means maximum.
Algorithm | | Band Ratio |
---|
443/561 | 482/561 | 655/561 | 865/561 | 443/655 | 482/655 | 865/655 |
---|
FLAASH | RMSE | 0.3422 | 0.1876 | 0.0861 | 0.6970 | 0.3994 | 0.1908 | 1.1892 |
MAPE (%) | 131.05 | 34.72 | 9.85 | 237.69 | 108.43 | 24.52 | 229.94 |
Bias (%) | 129.99 | 30.50 | 6.16 | 35.80 | 108.10 | 21.24 | 51.04 |
6SV | RMSE | 0.2363 | 0.1697 | 0.0774 | 0.6360 | 0.2885 | 0.1907 | 1.1338 |
MAPE (%) | 84.97 | 30.49 | 8.26 | 162.83 | 72.14 | 23.71 | 168.78 |
Bias (%) | 76.91 | 24.47 | 3.14 | 39.20 | 65.26 | 19.03 | 54.70 |
QUAC | RMSE | 0.2246 | 0.1849 | 0.0963 | 0.7539 | 0.2445 | 0.1880 | 1.1645 |
MAPE (%) | 96.18 | 34.33 | 11.50 | 158.53 | 71.28 | 24.34 | 149.59 |
Bias (%) | 94.71 | 32.29 | 8.82 | 55.21 | 69.89 | 20.85 | 56.11 |