Project overview
Whilst modern survey data (lidar, ground surveys, EO data etc) are available in digital vector or point form for recent monitoring the historic data (1840 – 1970s) are still as paper image data (raster) yet form an important input for designing restoration and nature-based solutions.
Extracting the position of the morphological features (e.g. shoreline, back marsh and channels) from historic data is currently both time-consuming and expensive but vital in terms of providing understanding of longer-term trends, the impact of development and defence on process rates and for in informing restoration design parameters, options and strategies. This feeds into national strategies and plans for coastal defence, managed realignment into former reclaimed tidal areas (for example, the Medmerry coastal realignment) and design of nature-based solutions. It also supports the targets for the national Nature Recovery Network and the Natural Capital and Ecosystem Assessment based approaches now being widely adopted by national agencies as part of the implementation of the Environment Act 2021. These approaches will become more common, driven by climate change mitigation and adaptation needs and facilitated by green finance initiatives carbon capture in wetland situations.
The challenges mean that such mapping is typically undertaken only on piecemeal basis, with project-level collection, to different standards, without common data models (e.g., which lines are extracted and how is the data quality assured). This precludes the use for comparative assessments and national assemblage, yet these are the crucial inputs to major protection policies and shoreline protection measures such as the National Flood and Coastal Erosion Risk Management (FCERM), Shoreline Management Plans and Local Nature Recovery Strategies (LNRS).
The ability to capture and analyse this historic information alongside current data at national levels is constrained by the time-consuming nature of acquiring and pre-processing the raster data (Country series mapping from 1840’s/1870’s) and the largely manual process of extracting vector data from which to measure change (such as changing intertidal widths, cliff toe and cliff top retreat and dune morphological change, intertidal channel extractions). This project will capitalise on knowledge exchange to test advanced artificial intelligence (AI) processing, textual and related image processing of the rasterised historic data and LiDAR terrain data to generate classified categorical vector data of linear features through pattern recognition.
Extracting the position of the morphological features (e.g. shoreline, back marsh and channels) from historic data is currently both time-consuming and expensive but vital in terms of providing understanding of longer-term trends, the impact of development and defence on process rates and for in informing restoration design parameters, options and strategies. This feeds into national strategies and plans for coastal defence, managed realignment into former reclaimed tidal areas (for example, the Medmerry coastal realignment) and design of nature-based solutions. It also supports the targets for the national Nature Recovery Network and the Natural Capital and Ecosystem Assessment based approaches now being widely adopted by national agencies as part of the implementation of the Environment Act 2021. These approaches will become more common, driven by climate change mitigation and adaptation needs and facilitated by green finance initiatives carbon capture in wetland situations.
The challenges mean that such mapping is typically undertaken only on piecemeal basis, with project-level collection, to different standards, without common data models (e.g., which lines are extracted and how is the data quality assured). This precludes the use for comparative assessments and national assemblage, yet these are the crucial inputs to major protection policies and shoreline protection measures such as the National Flood and Coastal Erosion Risk Management (FCERM), Shoreline Management Plans and Local Nature Recovery Strategies (LNRS).
The ability to capture and analyse this historic information alongside current data at national levels is constrained by the time-consuming nature of acquiring and pre-processing the raster data (Country series mapping from 1840’s/1870’s) and the largely manual process of extracting vector data from which to measure change (such as changing intertidal widths, cliff toe and cliff top retreat and dune morphological change, intertidal channel extractions). This project will capitalise on knowledge exchange to test advanced artificial intelligence (AI) processing, textual and related image processing of the rasterised historic data and LiDAR terrain data to generate classified categorical vector data of linear features through pattern recognition.
