Detail

Publication date: 1 de June, 2021

2010 – Large-scale parallel Monte Carlo simulations for ocean colour applications

This project focuses on large-scale Monte Carlo (MC) simulations of the radiative transfer process in the Ocean Colour (OC) application domain, with the aim of quantifying uncertainties due to environmental effects and measurement protocols on in-situ radiometric products derived from in-water optical profiles performed in coastal regions. Uncertainty sources of interest include light focusing and defocusing effects induced by facets of sea-surface waves, as well as perturbations due to sensor tilt during the deployment of free-fall optical profiling systems. A parallel MC code has been developed by the project members to compute high-resolution 2D representations of in-water light distributions by tracing a large number of photons and tracking their trajectories.
Coordinated by: Tamito Kajiyama (CITI, Davide D´Alimonte (CENTRIA), Jose Cardoso Cunha (CITI)
Large-scale simulations are necessary to reduce MC intrinsic statistical noise below a level that allows a thorough investigation of uncertainty budgets. MC simulations are planned to contribute to the quality assurance of in-situ radiometric products and enhance the exploitation of OC radiometric measurements for the validation of primary remote sensing (RS) products and for the development of bio-optical algorithms to determine high-level RS products. The project also represents a real-world case study for developing new HPC solutions, such as load balancing schemes for improving parallel efficiency and predictive performance models for optimizing the utilization of available HPC resources.
(A total of 25000 CPU-hours was awarded to the present project for consumption of the Milipeia supercomputer at the University of Coimbra.)

Team

José Cardoso e Cunha, Tamito Kajiyama,

Sname LSMC 2010
State Concluded
Startdate 01/05/2010
Enddate 31/10/2010