Proba-V for total suspended matter and turbidity retrieval in coastal areas (PROBA4COAST)

Start-End 01/12/2015 -  30/11/2017
Programme STEREO 3
Contract SR/00/326
Objective Proba4Coast will evaluate Proba-V imagery for deriving concentrations of Total Suspended Matter (TSM) and Turbidity (T) in near shore waters. The high revisit time, coastal coverage and the good image quality, provides opportunities to expand its current use from the typical land applications to coastal water applications. Proba-V TSM/T products are particularly relevant in combination with TSM/T products from other sensors (to obtain increased temporal coverage).

The project has two main objectives: (1) evaluation of the feasibility of a Proba-V TSM/T product and (2) the intercalibration with MODIS TSM/T products. For the second objective the use of a sediment transport model is explored to take into account TSM/T variations between image acquisitions

Method The project is structured in 4 major tasks:
(1) the theoretical study,
(2) the application and direct validation,
(3) the indirect validation and intercalibration and finally (4) the user evaluation.

(1) The theoretical study

First, the effects of the Proba-V bandwidth, spectral response shape and SNR on TSM/T retrieval will be evaluated theoretically and compared with MODIS and Landsat-8. For this evaluation both in situ measurements (water reflectances and corresponding TSM/T concentrations from SeaSWIR dataset, Belair dataset) and radiative transfer simulations will be used.

(2) The application and direct validation

A TSM/T algorithm will be selected appropriate for the Proba-V sensor. One of the proposed algorithms is the single band semi-analytical algorithm described in Nechad et al. (2009, 2010) which has been selected because of its strong theoretical background, its hyperspectral (sensor generic) character and successful applications in the literature. A very recent publication (Dogliotti et al., 2015) showed that the T algorithm can even be used in very different regions and is almost insensitive to the sediment type.

Both Red and NIR bands will be tested (Red might saturate for high turbidity, NIR might not be suited for low turbidity).

The TSM algorithm will be applied to atmospherically corrected PROBA-V Images where two A/C approaches will be considered: the PROBA-V standard A/C and a marine A/C in order to study the tradeoff between operational complexity and possible improvements in TSM product quality.
In the direct validation the retrieved reflectance values in RED and NIR bands are compared with in-situ normalized water leaving radiances (Lwn), measured by AERONET-OC CIMEL instruments in the Belgian coastal zone. Secondly the Proba-V TSM/T products will be compared with in situ data from Buoys (CEFAS smart Buoys and VLIZ Buoy) and dedicated sampling. Finally an analysis of the spatial patterns will be performed by visual check and by intercomparison with a model. Details of the model are provided in the following section.

(3) The indirect validation and intercalibration

MODIS (-A/-T) data will be processed using SeaDAS for the same period. This will deliver water reflectance products which will be converted to TSM/T. Indirect validation and intercalibration will be performed by fitting a relationship between MODIS and Proba-V TSM/T products. Here temporal variations of TSM/T will be taken into account because these areas are highly dynamic and TSM/T may have varied considerably between the two image acquisitions (varies in the order of 30min. with Terra and 2h with Aqua). These variation of TSM in time between the image acquisitions will be taken into account by modelling.

(4) The user evaluation

The proposed products will be evaluated against the user requirements. VITO has intensive contacts with the dredging sector, which is the main user envisaged here. Some of the products will be shown, the accuracies (for the study area known, taken into account possibly lower accuracies for other regions) will be discussed and a small archive study will be performed to evaluate the number of additional historic images w.r.t. MODIS


• Atmospheric correction module for Proba-V
• TSM algorithm for Proba-V
• Validation report
• Peer reviewed paper


• TSM and turbidity maps based on Proba-V
• Possibly inclusion of algorithms into operational processing chain at VITO and service for the dredging sector
Website link
Project Leader: KNAEPS Els VITO - Remote Sensing
Team Member: MONBALIU Jaak KULeuven - Department of Civil Engineering
Sensors used
Related presentations BEODay 2016 - 18. PROBA4COAST: Coastal turbidity monitoring with Proba-V
Related publications
Datasets used