Improving water reflectance retrieval from MODIS imagery in the highly turbid waters of La Plata river
Dogliotti, A.I.; Ruddick, K.; Nechad, B.; Lasta, C. (2011). Improving water reflectance retrieval from MODIS imagery in the highly turbid waters of La Plata river, in: Proceedings of VI International Conference «Current problems in optics of natural waters» (ONW’2011), St. Petersburg, Russia, September 6-9, 2011. pp. 8
In: (2011). Proceedings of VI International Conference «Current problems in optics of natural waters» (ONW’2011), St. Petersburg, Russia, September 6-9, 2011. [S.n.]: [s.l.]. , more
|
| Available in | Authors |
|
Document type: Conference paper
|
| Authors | | Top |
- Dogliotti, A.I., more
- Ruddick, K., more
- Nechad, B., more
- Lasta, C.
|
|
|
| Abstract |
The accurate retrieval of marine reflectance from remotely sensed data depends on the effective removal of the contribution of the atmosphere to the total signal that reaches the sensor at the top of the atmosphere, i.e. the atmospheric correction process. This is particularly important in highly turbid waters where assumptions made by traditional atmospheric correction algorithms for open ocean waters are often invalid. We have analyzed three different atmospheric correction algorithms in the challenging turbid waters of La Plata River estuary located in the western Atlantic coast at ~35° S. Three algorithms were applied to Moderate Resolution Imaging Spectroradiometer (MODIS) images of the region, the standard near-infrared (NIR) algorithm and two algorithms that use the short wave infrared (SWIR) bands. The two SWIR algorithms differ in the way the aerosol model is selected. The standard NIR atmospheric correction completely failed in retrieving water reflectance in the river plume waters mainly due to sensor saturation and an incorrect estimation of the marine contribution in the NIR. The standard SWIR approach showed better results, but unphysical correlations between marine features and atmospheric products, such as aerosol reflectance, in the most turbid part of the estuary were clearly identified. The use of an iterative SWIR-based atmospheric correction approach that accounts for non-zero water reflectance in the SWIR bands seems to be a good alternative for retrieving accurate marine reflectance. The difference in the derived water reflectance between the two SWIR approaches showed a spectral dependence, being higher in the shorter wavelengths and lower in the NIR. A comparison between MODIS-derived turbidity values from the different atmospheric correction approaches and in situ data showed no significant differences mainly because the one-band turbidity algorithm applied uses the 859 nm band where differences between the approaches are lower. |
|
