IAFE   05512
INSTITUTO DE ASTRONOMIA Y FISICA DEL ESPACIO
Unidad Ejecutora - UE
congresos y reuniones científicas
Título:
Assessing turbidity and total suspended matter algorithms using field reflectance measurements in the Río de la Plata turbid estuary
Autor/es:
DOGLIOTTI, A. I.; RUDDICK, K.; DOXARAN, D.; RAYMAEKERS, D.; LAMARO, A.; IBAÑEZ, G.; KNAEPS, E.
Lugar:
Matera
Reunión:
Workshop; 6th EARSeL Workshop on Remote Sensing of the Coastal Zone; 2013
Resumen:
The Río de la Plata estuary is located in the eastern coast of southern South America at approximately 35 °S and is one of the largest estuarine systems of the world, which. The system carries a large amount of sediments and nutrients to the ocean, which in turn influence the biological productivity, sedimentological processes and underwater visibility. During two weeks (nine days) water leaving reflectance, turbidity (T) and total suspended matter (TSM) concentration were measured at the Río de la Plata. The objective of the present study is to test the performance of existing one‐band and band‐ratio T and TSM ocean color remote sensing algorithms using field data collected during the field campaigns. Algorithms for retrieval of TSM were tested, including a band ratio of MODIS red (620‐670) and near infrared (841‐876) band, and a one‐band algorithm for the MODIS 841‐867 band. A similar one‐band algorithm was used to estimate T. Variability of the water leaving reflectance in the short wave infrared (SWIR) region at 1020 and 1071 nm was analyzed and its relation to TSM and T was investigated. A significant correlation was found between estimated and measured TSM (r~0.8). The one‐band algorithm showed a systematic underestimation of ‐20% (mean relative error) and a relative uncertainty of 22% (mean absolute relative error), while the band‐ratio showed a systematic overestimation of 20% and a relative uncertainty of 25%. Better results were found for the T algorithm; it showed a higher correlation (r=0.87) and lower mean relative and absolute differences (6% and 11%, respectively). Even though the water leaving reflectance in the SWIR was not negligible, there was no significant correlation between TSM and T and reflectance in the 1020 and 1071nm bands, probably due to the low sensitivity of the SWIR bands to the moderately turbid waters found in this study (30‐160 mg l‐1). Limits of the optimal range for TSM retrieval in this spectral region are discussed. These results suggest the potential use for the Río de la Plata of existing algorithms that were calibrated using field data from other turbid regions of the world. The mapping of TSM and T using MODIS imagery then depends on a suitable atmospheric correction for these extremely turbid waters.