Retrieval of particulate backscattering in coastal/oceanic waters

Accurate retrieval of particulate backscattering coefficients has been a great challenge in optically complex Case 2 waters. Although several models exist for retrieval of the Inherent Optical Properties (IOP’s) such as particulate backscattering (bbp) from the remote sensing reflectance, these models produce bbp with high errors mainly because of improper parameterization or inappropriate definition of the slope and reference parameters. After analyzing the performance of several models for open and coastal waters, we found that GSM model retrieves bbp with accuracy better than the other models. However, its retrieved bbp values are still uncertain in coastal waters mainly because of the above problems. In this study, we introduced new steps in the GSM model to improve the retrieval of bbp in a wide range of waters. One important step involved in the improved model is an appropriate definition of the slope parameter that varies depending on the characteristics and concentration of the sea water constituents. When applied to simulated dataset from Boss-Roesler simulation data sets and NOMAD in situ data set, the improved model outperforms the GSM model. The RMSE for the simulated data sets was found low (0.1266 m bbp(443)) with the improved model when compared to that with the GSM (0.1937 m bbp(443)). For NOMAD in situ data set the RMSE value was reduced from 0.1526 m for the GSM to 0.1256 m for the improved model. Results of the sensitivity analysis showed that the improved model is more robust than the GSM model and can be used to accurately retrieve bbp values from a wide range of waters commonly found in the coastal and offshore domains.