S 2015 | Vienna | 10.-12. November 2015 LUCAS pilot study on provision of harmonized land cover and land use statistics from national data sources LUCAS „Bottom-up Approach“ (Eurostat Grant) Stephan Arnold Agricultural land use and areal statistics Federal Statistical Office
28
Embed
EFGS 2015 | Vienna | 10.-12. November 2015 LUCAS pilot study on provision of harmonized land cover and land use statistics from national data sources LUCAS.
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
EFGS 2015 | Vienna | 10.-12. November 2015
LUCAS pilot study on provision of harmonized land cover
and land use statistics from national data sources
LUCAS „Bottom-up Approach“ (Eurostat Grant)
Stephan ArnoldAgricultural land use and areal statistics
Federal Statistical Office
EFGS 2015 | Vienna | 10.-12. November 2015
Destatis - Stephan Arnold - LUCAS pilot study 2013-2014
Content LUCAS field survey
Eurostat grant LUCAS pilot study
National approach, method, results
Conclusion
Outlook
Folie 2
EFGS 2015 | Vienna | 10.-12. November 2015
Destatis - Stephan Arnold - LUCAS pilot study 2013-2014
LUCAS survey LUCAS – Land Use Cover Area Frame Statistical Survey
Europe-wide capture of land cover and land use data, field fotos
Point sample field survey, oriented on regular 2x2 km grid
fixed reference years in a 3-year cycle (2006 – 2009 – 2012 – 2015 – 2018)
Folie 3
EFGS 2015 | Vienna | 10.-12. November 2015
Destatis - Stephan Arnold - LUCAS pilot study 2013-2014
-Grant: project target
Semantic harmonisation between national and European LC/LU nomenclatures
Consistence between national LC/LU statistics and estimates on European level
Reduce costs and efforts of LUCAS field survey
Integration of national initiatives into pan-European datasets (bottom-up approach), synergies with national initiatives (spatial data = micro data and/or statistical tables)
Folie 4
EFGS 2015 | Vienna | 10.-12. November 2015
Destatis - Stephan Arnold - LUCAS pilot study 2013-2014
Grant details Project period
18 months (2013-2014) Expected results
Report on methodology Statistical table
Financing Eurostat Grant
(10 % own contribution) Grant holder
PT, NO, NL, DE, PL, IT, EL
Folie 5
EFGS 2015 | Vienna | 10.-12. November 2015
Destatis - Stephan Arnold - LUCAS pilot study 2013-2014
LUCAS pilot study in DE
Folie 6
EFGS 2015 | Vienna | 10.-12. November 2015
Destatis - Stephan Arnold - LUCAS pilot study 2013-2014 Folie 7
Minimum mapping units
Reference spatial unit
Reference time window
Semantics, definitions
Interpretation !
EFGS 2015 | Vienna | 10.-12. November 2015
Destatis - Stephan Arnold - LUCAS pilot study 2013-2014
National approach in DE Information content assessment of different
data sources according to required Eurostat LC/LU-Codes
Destatis - Stephan Arnold - LUCAS pilot study 2013-2014
Challenges and constraints
Varying criteria within target classes (thematic content vs minimum mapping unit)
Consistence of semtantic transformation
Thematically unsplitable source classes
Reliability, quality and completness of input data sets
Folie 9
EFGS 2015 | Vienna | 10.-12. November 2015
Destatis - Stephan Arnold - LUCAS pilot study 2013-2014
Workflow ofarea size calculation
Folie 10
II. Semantic Transformation
I. Source code area sizes
III. Connection of area sizes
V. Analysis & Plausibility
IV. Target code area sizes
VI. Total area sums
EFGS 2015 | Vienna | 10.-12. November 2015
Destatis - Stephan Arnold - LUCAS pilot study 2013-2014
Working steps
Folie 11
… for the derivation of LC/LU-Statistics: Semantic Transformation (Key code matching) between
source feature types and target classes
Preprocessing of area sizes figures of source feature types
Connection of semantic transformation table with area size figures of source feature types
Calculation of area sizes for target Grant-codes and re-analysis of semantic matching accuracy
Iterative process
EFGS 2015 | Vienna | 10.-12. November 2015
Destatis - Stephan Arnold - LUCAS pilot study 2013-2014
European Level: Target
ALB ALKIS BONational Level: Source
DLM-DE
LUCAS
STT (a) STT (b) STT (c) STT (d)
Semantic Transformation Table (STT)
Folie 12
EFGS 2015 | Vienna | 10.-12. November 2015
Destatis - Stephan Arnold - LUCAS pilot study 2013-2014 Folie 13
Normal Transformation strict interpretation of class definitions,
=> highest possible degree of semantic correctness and plausibility
=> reduced matching classes accepted
Generalized Transformation tolerant interpretation of class definitions,
=> highest possible number of unique semantic matches between source and target definitions
=> over/under-estimates accepted
Auxiliary classeswork around for non-matching source feature types
=> 100% of area assigned to a class (required / auxiliary) => leftovers at least partly separated from „no match“
Semantic Transformation Table (STT)
EFGS 2015 | Vienna | 10.-12. November 2015
Destatis - Stephan Arnold - LUCAS pilot study 2013-2014 Folie 14
Auxiliary classes
Land CoverB00/E00 – CROPLAND or GRASSLAND (agricultural areas)B00&E00 – CROPLAND and GRASSLAND (CLC 242)B00&X00 – CROPLAND and natural Vegetation (CLC 243)D00/E00 – SHRUBLAND or Grassland (CLC 324)V00 – Non-forest Vegetation „Landwirtschaftsfläche“ = land under (possible) cultivationY00 – Any VegetationZ00 Unknown
Land UseU100/U600 – Primary production or No other useU200/U300/U400 – Non-residential artificial use (commerce and industry)U500/U800 – Residential or Mixed use (CLC 111, 112)U800 – Mixed use (ALB residential with commercial/public) U900 – Unknown / Various use
Semantic Transformation Table (STT)
EFGS 2015 | Vienna | 10.-12. November 2015
Destatis - Stephan Arnold - LUCAS pilot study 2013-2014
Total area sums for LUCAS LC-Codes
Folie 15
MVP = Minimum common publication standard
Level 1 codes, normal transformation, area size results of uninque matches
Destatis - Stephan Arnold - LUCAS pilot study 2013-2014
Grant-LU-Codes Level 1ALB 2009
[DE]ALB_MVP 2009 [DE]
BO 2010 [DE]
DLM-DE 2009 [DE]
U100 Primary Production 26.842.302 10.919.011 18.358.320 30.484.833U200 Secondary Production 0 0 0 0U300 Tertiary Production 810.026 426.138 0 424.645U400 Transport Networks Logistics
And Utilities 1.893.552 1.785.564 0 206.226U500 Residential use 1.185.303 1.185.303 0 0U600 Other Uses or no socio-
economic use 792.014 728.631 28.785 355.378U1/600 Primary production/No other use 2.376.848 18.574.070 0 86.557U2/3/400 Non-residential artificial use 325.651 325.651 0 683.609U3/400 Tertiary production or
TransNetLogUtilities 31.466 0 0 0U5/800 Residential or Mixed use 0 0 0 2.265.964U800 Mixed use 353.135 0 0 0U900 Unknown / Multiple use 1.102.185 1.768.112 0 1.259.289Total area 35.712.482 35.712.482 18.387.105 35.766.501Area proportion unique match level 1 88,27% 42,13% 51,48% 88,12%Area proportion unique match level 2 86,99% 41,01% 50,77% 88,12%
Total area sums for LUCAS LU-Codes
Folie 16
Level 1, entire country of Germany, old model ALB, normal transformation
EFGS 2015 | Vienna | 10.-12. November 2015
Destatis - Stephan Arnold - LUCAS pilot study 2013-2014
Level 1, federal state of Hessen, new model ALKIS, normal transformation
EFGS 2015 | Vienna | 10.-12. November 2015
Destatis - Stephan Arnold - LUCAS pilot study 2013-2014
Grant-LU-Codes Level 1ALKIS 2011
[HE]%
ALKIS 2011 MVP [HE]
%
U100 Primary Production 26.842.302 82,27% 10.919.011 82,20%U200 Secondary Production 0 0,01% 0 0,00%U300 Tertiary Production 810.026 2,10% 426.138 1,32%U400 Transport Networks Logistics And
Utilities 1.893.552 7,93% 1.785.564 6,57%U500 Residential use 1.185.303 4,31% 1.185.303 4,31%U600 Other Uses or no socio-economic
use 792.014 0,73% 728.631 0,43%U1/600 Primary production / No other use 2.376.848 0,56% 18.574.070 0,43%U2/3/400 Non-residential artificial use 325.651 0,71% 325.651 0,87%U3/400 Tertiary production or TransNetLogUt 31.466 0,08% 0 0,79%U5/800 Residential or Mixed use 0 0,00% 0 0,00%U800 Mixed use 353.135 1,08% 0 1,12%U900 Unknown / Multiple use 1.102.185 0,22% 1.768.112 1,95%Total area 35.712.482 35.712.482Area proportion unique match level 1 97,35% 94,83%Area proportion unique match level 2 97,03% 94,82%
Total area sums for LUCAS LU-Codes
Folie 18
Level 1, federal state of Hessen, new model ALKIS, normal transformation
EFGS 2015 | Vienna | 10.-12. November 2015
Destatis - Stephan Arnold - LUCAS pilot study 2013-2014
Conclusions from LUCAS pilot study Data model change (ALB-ALKIS) in
progress, finalization planned in 2016
No singular source data set is capable to fulfill all LUCAS target classes
Compromise approach for statistical tables:
LC covered by ALB/ALKIS/DLM-DE + BO (with leftover)
LU covered by ALKIS
=> Data mining is possible in principle, plausible total area size has to be met
Folie 19
EFGS 2015 | Vienna | 10.-12. November 2015
Destatis - Stephan Arnold - LUCAS pilot study 2013-2014
Outlook
Folie 20
EFGS 2015 | Vienna | 10.-12. November 2015
Destatis - Stephan Arnold - LUCAS pilot study 2013-2014
Outlook: follow-up project
Folie 21
Project „Cop4Stat_2015plus“ Feasibility study how to use Copernicus-products to
derive LC/LU information for LC/LU statistical estimates
Orientation on user requirements from European level
Separation of LC & LU, arable land / grasland, use intensities, cultivation measures, seasonal patterns, status
Build upon results of LUCAS pilot study: Which existing national data sources can be used to derive pan-European LCL/ statistics?
Data material Sentinel-2 optical imagery (Sentinel-1 RADAR Daten
optional)
High Resolution Layer
Land Cover Model LBM-DE (formerly DLM-DE) as additional data
EFGS 2015 | Vienna | 10.-12. November 2015
Destatis - Stephan Arnold - LUCAS pilot study 2013-2014 Folie 22
Data flow from national to European level
CLC/EEA
ATKISALB/
ALKIS
HRL
FEtNStat
LUCAS/Eurostat
European Level
Länder-Level
Official areal statistics based on cadastral data (ALB/ALKIS)
Testing of alternative data source: remote sensing
DLM-DE
FEtNStat
Federal Level
AgrarStat
Sen-tinel
LB/LN„Stat“
BWI
IACS!?
EFGS 2015 | Vienna | 10.-12. November 2015
Destatis - Stephan Arnold - LUCAS pilot study 2013-2014
EAGLE Concept
Folie 23
EFGS 2015 | Vienna | 10.-12. November 2015
Destatis - Stephan Arnold - LUCAS pilot study 2013-2014
De-composition of landscape
Folie 24
From classification to characterization and description