Modeling the spatio-temporal dissolved organic carbon concentration in Barra Bonita reservoir using OLI/Landsat-8 images
MetadataShow full item record
Through exchange of heat between the water and the atmosphere inland waters affect climate at the regional scale and play an important role in the global carbon cycle. Therefore, there is a need to develop methods and models for mapping inland water carbon content to understand the role of lakes in the global carbon cycle. The colored dissolved organic matter (CDOM) has a strong correlation with dissolved organic carbon (DOC) and can be studied using remote sensed images. In this work, we developed an empirical model to estimate the DOC concentration by using the absorption coefficient of CDOM (a(CDOM)). The a(CDOM) was estimated through band ratio index and validated with in situ data. The empirically adjusted model to estimate the DOC was applied to a series of OLI/Landsat-8 images. The results showed a good relationship between the a(CDOM) at 412 nm (a(CDOM412)) and the ratio between OLI band 1 and OLI band 3, but the validation results showed a normalized root mean square error (NRMSE) of about 37.89%. The a(CDOM412) obtained in laboratory was used to establish a relationship between a(CDOM412) and DOC. The DOC spatial distribution was then obtained and the concentration varied from 22 to 52 mg.l(-1) during the year of 2014.