Analysis Ready Data & Intermediate Products: streamlining access - PowerPoint PPT Presentation

www.geoger.co.uk | @ajggeoger | http://uk.linkedin.com/in/ajggeoger/ Analysis Ready Data & Intermediate Products: streamlining access to data for land applications Alastair Graham | Geoger Ltd

jncc.defra.gov.uk www.geoger.co.uk www.specto-natura.co.uk www.apmgeo.co.uk

Overview • Analysis Ready Data • Why Intermediate Products? • The Processing System • UK Intermediate Products • User-Driven Development • What Now?

ARD

Analysis Ready Data (Sentinel-2) • Georectified • Stacked bands • Atmospheric correction to sref • Standardisation of pixel size? • Standard format • Cloud masked?

Why Intermediate Products?

Intermediate Products • Use common components • Highly automated • Improved quality • Improved consistency • Reduced cost • Reduced duplication of effort • Exploits economies of scale • Multiple use

Intermediate Products (EO) • Relates to basic surface property/indicator • Needed by a range of user communities • Standard, stable & transparent specification • Deliver spatially & temporally consistent results • Produced centrally within a qualified processing system with quality control • Available to the whole user community

Copernicus

Defra CoE/JNCC • NDVI, NDWI • Mosaicing and stacking • Monthly composites • Timeseries

The Processing System

Processing chain Analysis Ready NDVI Bespoke code Sentinel Data composites written in Bash, calling GDAL and pktools NDWI composites

Processing chain • HWL land mask for GB Preprocessed ARD – (EPSG: 27700) Extract bands • Admin land mask for NI – (EPSG: 29903) Calculate Int Prod • Date of max value recorded Clip to land mask for each pixel MVC Spatial metadata Put on S3

Data reduction • Input data – 114 Sentinel-2A scenes – Total data volume of ~5 TB – Download time ~12 days • Output products – 12 composite images for 2 indicators (+ meta data layers) – Total data volume 243 Gb

UK Intermediate Products

Results: single month

Results: coverage

Results: timeseries

Results: spatio-temporal

User-Driven Development

Users • There has been a wide range of interest • Standard workflows, no duplication of effort Organisation Relevance The Crown Estate (Rural) Land management The Crown Estate (Forestry) Forest estate management Organisation Relevance NT Conservation uses SNH Upland habitat mapping FAO Food security issues NE Living Maps project OS/RPA RPA updates EA Geomatics Support Enforcement Officers Eunomia Waste, environmental economics Forestry Commission National Forest Inventory Natural England Coastal habitats, regional monitoring Forest Research Forestry EODIP project Climate Wine Consulting Viticulture - Plant health and NRW Phase 1 mapping productivity, site suitability

What Now?

What Now? • Additional application examples • Additional Intermediate Products • Inclusion of other EO datasets • Optimisation of the code/workflow • ARD improvements

Thank you a.graham@geoger.co.uk