Resolving near-coastal remote sensing signal into contributions by bottom, water column, glint and the adjacency effect

All waterbodies are changing in the variable climate conditions. Monitoring of these changes over large areas is possible only by using remote sensing. However, remote sensing of waterbodies is hampered by the nearby land as part of the signal measured above the waterbodies originates from the nearby land not from the water itself. This problem, called the adjacency effect, is detectable up to kilometres from the shore. In the case of majority of lakes on Earth it means that every water pixel is affected by the nearby land. Moreover, the signal measured near the shores may contain effects from the bottom (if water is shallow), signal from emerging vegetation, sun and sky glint. In the case of marine remote sensing these problematic areas are usually masked out in remote sensing studies as too complicated to resolve. This is not an option in lake remote sensing as just of few tens of lakes (out of 117 million) are large enough to contain pixels free from the adjacency and other coastal effects. Moreover, up to 99% of carbon is processed in the near-shore waters and never reaches the deep ocean carbon pool. Thus, many important processes take place in the near-coastal waters that are currently masked out from remote sensing imagery as too complex to resolve. This limits the use of remote sensing in environmental research and monitoring. Our aim is to study the very nearshore waters in order to resolve the contribution of the adjacency effect, lake/sea bottom, sun and sky glint and the water column itself and develop algorithms for removing the adjacency effect and glint from Sentinel imagery.