Hydroacoustic remote sensing of human impacted tidal and shallow water marine environments

Abstract

Coastal areas are among the most ecologically diverse, economically valuable and fragile ecosystems in nature: human pressure on coasts is higher than on any other natural environment and it keeps growing; climate changes are also adversely affecting the coastal marine habitats. Tidal environments, with their peculiar combination of terrestrial, palustrine, and aquatic habitats are considered of extreme importance for biodiversity. Nevertheless, they are vanishing faster than any other complex ecosystem on Earth. In the past decades, national and international regulations were issued for mapping and protecting coastal and marine habitats. Therefore, the scientific community has been involved in finding efficient and effective methods for producing seafloor habitat maps, which could be used, then, as basis for habitat monitoring, protection, restoration, and for managing the sustainable development of coastal areas. Among the possible different tools for habitat mapping, remote sensing and in-situ techniques represent a useful combination for characterizing large areas. Hydroacoustic sources have been used since decades for seafloor mapping and still represent the best tool for investigating the nature and the dynamics of underwater habitats. The large amount of different available acoustic systems and the variety of possible processing approaches led to a lack in standards about the optimal use of such systems for habitat mapping, thus in the last 10 years many programs were launched to bridge such gap. Nevertheless, many aspects are still unclear. Seafloor roughness is among the most relevant parameters that influence the acoustic signal. Roughness is a product of multiple factors, including sediment composition and distribution (“sediment roughness”), seafloor morphology (“topographic roughness”), and biological communities (“benthic roughness”). Direct human disturbance (e.g. from dredging and dumping activities) further complicate the natural seafloor roughness. The present research made use of multiple hydroacoustic sources (singlebeam and multibeam echosounders, sidescan sonars, ADCP) coupled with in-situ techniques, in order to study the role of the different roughness components in influencing the backscatter, both when automatic and expert-based classification strategies are adopted. The results were used for assessing the impact of a deep-water terminal construction (JadeWeserPort, in the Jade Channel) on sediments and biocommunities. Five research sites within three areas (the backbarrier system of Norderney Island, the Jade Channel, and the seafloor in vicinity of the Helgoland Island) were selected in the German Bight (southern North Sea), in order to represent intertidal and subtidal environments characterized by the presence of both crisp and fuzzy habitat boundaries, hard and soft grounds, and affected by different degrees of human disturbance. The results show that automatic classification tools allow to highlight the roughness factors that play the primary role in influencing the backscatter at specific sites. The multiple generations of dredging marks present in the Jade Channel dominated the classification outcomes (topographic roughness >> sediment and benthic roughness). As a result, the acoustic classes corresponded to the main morphological regions, rather than to the sediment bottom characteristics. For the habitat mapping of the strongly impacted area nearby the JadeWeserPort, an expert-based approach was therefore preferred. The impact of the port construction produced a redistribution of sediments and the development of new benthic assemblages (mainly made by opportunistic and highly mobile species), which replaced the existing ones. On the less disturbed seafloor in the nearby of Helgoland, the crisp boundary between the soft and hard ground regions was correctly identified by the automatic classification tools (topographic roughness >> sediment and benthic roughness). In addition, the acoustic data showed that in the soft ground area characterized by homogeneous sediment composition and by the absence of seafloor morphologies, the presence of a dense colony of soft, centimetre-sized brittle stars (devoid of rigid structure) was able to modify the backscatter values (benthic roughness >> topographic and sediment roughness). Such fuzzy boundary was also successfully mapped on the automatic classification outcomes. However, within the hard ground regions the detection of two different population of bryozoan resulted to be difficult (topographic and sediment roughness >> benthic roughness). In the Norderney intertidal and subtidal areas, the complex interplay of all the roughness components made it difficult to use automatic classification tools (topographic ≈ sediment ≈ benthic roughness), which were at most able to identify the main tidal-flat drainage system. On the contrary, the expert-based approach allowed the segmentation of the sidescan-sonar backscatter in several morpho-sedimentary and benthic classes. A comparison with an existing seafloor classification based on satellite data (RapidEye) showed the superior ability of sidescan sonar in resolving seafloor structures, sediment classes, and shellfish beds. The outcomes of the present research offered a contribution to the development of recommendations and guidelines for mapping and monitoring the German coastal waters (WIMO Project – Scientific monitoring concepts for the German Bight).