6. Summary
Using an objective classification, and a dataset of satellite (OC-CCI
products) and in situ chlorophyll match-up data, we tested the performance
of two standard ocean-colour empirical chlorophyll algorithms
(one based on a blue to green band-ratio, OC4, and the other a band difference,
OCI), a semi-analytical algorithm, and an empirical algorithm
that accounts for the influence of CDOMon the remotely-sensed chlorophyll
estimates.We found that the two empirical algorithms (OC4 and
OCI) had the highest performance, but systematically overestimated
chlorophyll when compared with the in situ data.
By developing a Red Sea ocean-colour model, parameterised where
possible to data from the Red Sea, we adjusted the two ocean-colour
empirical algorithms for chlorophyll estimation and the systematic
over estimation in chlorophyll originally observed was removed. The relationships
of particulate absorption and particulate backscattering with
chlorophyll that are used in the Red Sea model, are similar to
established global relationships, but the amount of CDOM absorption
per unit chlorophyll concentration in the model is higher than standard
global relationships. An enhanced amount of CDOM absorption per unit
chlorophyll in the Red Sea was found to explain the over estimation in
chlorophyll originally observed for the OCI and OC4 algorithms. A series
of algorithms adjusted for the Red Sea have been proposed, designed
for a range of ocean-colour sensors, and are now available for further
testing. Given the unique and understudied marine and atmospheric
environment of the region, uncertainties in the in situ data, and the potential
influence of aeolian dust on atmospheric correction, additional
information is required to scrutinise our findings.
6. SummaryUsing an objective classification, and a dataset of satellite (OC-CCIproducts) and in situ chlorophyll match-up data, we tested the performanceof two standard ocean-colour empirical chlorophyll algorithms(one based on a blue to green band-ratio, OC4, and the other a band difference,OCI), a semi-analytical algorithm, and an empirical algorithmthat accounts for the influence of CDOMon the remotely-sensed chlorophyllestimates.We found that the two empirical algorithms (OC4 andOCI) had the highest performance, but systematically overestimatedchlorophyll when compared with the in situ data.By developing a Red Sea ocean-colour model, parameterised wherepossible to data from the Red Sea, we adjusted the two ocean-colourempirical algorithms for chlorophyll estimation and the systematicover estimation in chlorophyll originally observed was removed. The relationshipsof particulate absorption and particulate backscattering withchlorophyll that are used in the Red Sea model, are similar toestablished global relationships, but the amount of CDOM absorptionper unit chlorophyll concentration in the model is higher than standardglobal relationships. An enhanced amount of CDOM absorption per unitchlorophyll in the Red Sea was found to explain the over estimation inchlorophyll originally observed for the OCI and OC4 algorithms. A seriesof algorithms adjusted for the Red Sea have been proposed, designedfor a range of ocean-colour sensors, and are now available for furthertesting. Given the unique and understudied marine and atmosphericenvironment of the region, uncertainties in the in situ data, and the potentialinfluence of aeolian dust on atmospheric correction, additionalinformation is required to scrutinise our findings.
การแปล กรุณารอสักครู่..
