An Improved Scheme for Correcting Remote Spectral Surface Reflectance Simultaneously for Terrestrial BRDF and Water-Surface Sunglint in Coastal Environments

Global spectroscopic missions, such as NASA's Surface Biology and Geology, will observe coastal environments and must account for the optical properties of both land and sea. Specifically, they must consider reflectance effects that arise from interactions between surface structures and variable observing geometries that are unique to terrestrial and aquatic domains. Over vegetated surfaces, Bidirectional Reflectance Distribution Function (BDRF) effects must be corrected to consistently map vegetation properties. At the water surface interface, sunglint effects must be corrected to estimate water column and benthic properties. Current analyses focus on vegetation or aquatic domains and do not easily address environments with mixed water and vegetation cover. Wetland environments within global spectroscopic data sets will pose a challenge to correction methods when scientific applications in coastal regions require continuous reflectance correction over both surface types. Here, we present the first simultaneous treatment of terrestrial BRDF and aquatic sunglint in a wetland environment. We evaluate existing vegetation-BRDF correction methods and pair them with sunglint correction of aquatic pixels. We test multiple sunglint correction strategies to produce continuous corrected products over both vegetation and water surfaces. We show that the addition of sunglint correction in wetland environments significantly reduces error between overlapping image regions. Comparisons between airborne and in situ data demonstrate sub-percent error in remote estimates of water-leaving reflectance. Our results demonstrate the importance of both BRDF and sunglint corrections in wetland environments. A unified treatment will be critical for global-scale hyperspectral spectroscopy missions. Plain Language Summary Upcoming imaging satellite missions, such as NASA's Surface Biology and Geology (SBG), will provide powerful global data sets. One expected challenge in dealing with the volume of data is that SBG will capture the full diversity of Earth's environments. Coastal environments are specifically challenging because coastal data processing must account for data quality issues that are unique to both water and land. In this study, we propose a workflow to correct for effects that influence data quality of land and water simultaneously. We test our method in coastal environments from Mississippi and California. These locations provide a range of vegetation as well as a range in water composition to ensure our method is robust to these environmental variables. We also test multiple correction strategies to ensure that our workflow is robust in a coastal setting and compare our resulting data to equivalent in situ data collected in the field. Our results demonstrate the importance of corrections for both water and land data quality effects, which will be a critical perspective when managing a global imaging satellite mission.

Automated summary

Global spectroscopic missions need to consider the optical properties of both land and sea when observing coastal environments. This study presents a method for simultaneously correcting terrestrial and aquatic optical effects in wetland environments, demonstrating the importance of these corrections for accurate remote sensing.