On-orbit radiometric characterization of OLI (Landsat-8) for applications in aquatic remote sensing

Abstract

Landsat-8 carries two separate sensors, namely the Operational Land Imager (OLI) and the Thermal Infrared Radiometer Suite (TIRS), that image the earth surface throughout the visible and thermal portions of the spectrum. Compared to Landsat heritage sensors, the OLI has enhanced features, which include its 12-bit radiometric resolution and the addition of a band centered at 443nm. The dramatically improved data quality/quantity expands existing applications of Landsat imagery in aquatic sciences from the retrieval of bio-geochemical properties, such as near-surface concentrations of chlorophyll-a (CHL) and total suspended solids (TSS), to benthic mapping. This study offers analysis of OLI's absolute radiometric performance over bodies of water using benchmark observations, namely the top-of-atmosphere (TOA) ocean color observations and marine in situ radiometric measurements. Sensor-to-sensor comparisons are performed to derive gain factors (g1) from near-concurrent observations in TOA radiance and reflectance domains. The gains in the radiance domain were further validated/adjusted by determining a second set of gains (g2) via analysis of OLI-derived water-leaving radiance, i.e., Lw(λ), against in situ measurements made either at the Ocean Color AErosol RObotic NETwork (AERONET-OC) sites or during field campaigns. The analyses yield the OLI calibration uncertainties that need to be accounted for when studying aquatic environments. It was found that, for the visible and near-infrared channels, the OLI radiometric responses, on average, are well in agreement (<2% discrepancies) with the TOA radiances estimated by ocean color satellites or those predicted by models based on measurements of aquatic and atmospheric properties. However, the TOA radiance at the new 443-nm band is found to be, on average, 3.4% larger than the reference observations. The inter-sensor comparisons in the reflectance domain, however, indicated slightly different results with the OLI responses being low in the blue bands. To enhance the retrieval accuracy of aquatic-science products from OLI datasets, sets of temporally averaged gains (radiance and reflectance) are derived and recommended for use prior to the retrieval of in-water products.