Log in (or create account for non-CSDMS members)
Forgot username? Search or email:CSDMSweb@colorado.edu
Browse abstracts
‘Seamless over the Strandline’: Joining the Direct-Sample Databases for Marine Substrates (dbSEABED), Shorelines, and Terrestrial Soils
The shoreline is a boundary where survey methods change dramatically, where the time dimension is extremely important, where sediment fluxes are very large, where flotsam is trapped, and where numerical/physical singularities occur as the water depth goes to zero. The shoreline boundary oscillates; and as sea-levels rise what is now shore will become sea. So the issue of likely response of sediment/beach/soil substrates to those changes is important. Models plus data will be required in this, to understand and make forecasts.
This study looked at datasets for Louisiana – onshore soil surveys, shore descriptions, and offshore seabed substrates - and investigated to what extent they were already in harmony, and what challenges remain in trying to make one seamless dataset. Of course, technologies like LIDAR carry out highly detailed imaging that achieves this to an extent. But we are focused on direct samplings of the ground-truthing type on which physical properties, fabrics, chemical compositions, grain types, genesis, can be directly determined.
Difficulties: Onshore surveys have a different data topology, more focused on soil polygons and boreholes; offshore mappings focus on point-samplings, for instance grabs and cores. Descriptions of the soils often focus on layer-profile identities such as “Mollisol”; offshore datasets focus on bulk textures and compositions. Semantic differences exist between the mapping disciplines, even for the same terms. Onshore areas are greatly modified by agriculture and construction; marine areas not so much.
Positives: But also, we discovered several information-integration pathways for merging the sample-based data from the two realms. Exhaustive searching can uncover the necessary data on onshore soils – and riverbed substrates; the marine data is already well integrated. Name geographical locations are linkable with coordinates through online gazetteers. Computational and database methods exist to merge polygon and point data sets. In the semantics, glossaries provide some information to link onshore and offshore descriptions. And machine learning methods can now extend known direct-measurement data right across the complexity of the coastline boundary.
