PR Additional information

From Wikicap - European Commission
Jump to: navigation, search

Go up to the main pro rata page

general

  • The DGJRC will provide statistical methods for land cover class RC assessment.
  • The DGJRC invites the LPIS custodians to share any methodologies, practices, field cases and experiences. With the permission of the LPIS custodian, the DGJRC will publish a selection of the typical pro rata land cover classes and assessment methods.


Example of systematic pro rata application

Step 1: Retrieve and compile the available data

Description Image

A reference parcel encloses among other lands, area of permanent grassland

with low productivity that is considered suitable for extensive grazing

ficticious RP with pro rata grasslands on top of extract from World_Imagery (http://services.arcgisonline.com/arcgis/services)
e.g. agro-ecological data from a permanent grassland register
a grassland register - extracts the project for the management and use of natural grassland, Turkey (http://ulusalmera-tagem.gov.tr/proje.asp)
e.g. multi-temporal and multi-spectral image data (optical, SAR, LIDAR,....)
e.g. ancillary data, thematic maps, topomaps....
maps - extract from: (top) the raster thematic data of Corine Land Cover (http://land.copernicus.eu/pan-european/corine-land-cover/view); (bottom) World_Terrain_Base (http://services.arcgisonline.com/arcgis/services)

Note: (1) World Imagery sources: Esri, DigitalGlobe, Earthstar Geographics, CNES/Airbus DS, GeoEye, USDA FSA, USGS, Getmapping, Aerogrid, IGN, IGP, swisstopo, and the GIS User Community; (2) World Terrain Base sources: Sources: Esri, USGS, NOAA; (3) Copernicus Core 03 source: SPOT 5 © CNES (2010-2013), distribution Astrium Services/Spot Image S.A.

Step 2: Define pro-rata classes in the eligibility profile

Description Image/Result
  • Observe the physical reality on the ground
  • Detect the type of the intrinsic mix present
  • Retrieve the distinct characteristics of the mix
  • Check scattering pattern, density, physiognomy, phenology, topography, location, edaphic conditions, etc.
  • Verify the visibility and stability of these characteristics in time
  • Decide whether mix is natural in origin or derived from particular agronomic practice
  • Check farmer data, if available
  • Verify the regional variability of the proposed class
identification of reliable land cover classes on top of extract from World_Imagery (http://services.arcgisonline.com/arcgis/services)
  • Provide semantic definition of the pro-rata classes, based on the findings
  • Elaborate detailed description of the class using FAO LCCS or ISO19144-2 (LCML)
  • Derive photointerpretation keys or training data for automated classification
  • Share findings and documentation to stakeholder community (farmers, LPIS operators, OTSC inspectors)
See outcome table below

outcome: comprehensive grassland class definitions

LCCCode LCCLevel LCCOwnLabel LCCOwnDescr LCCLabel MapCode User Description
21669-127116 A6A20B4XXXXXXF1-A21B15 G1 Closed to Open Grassland Closed to Open (100-40)% Medium To Tall Grassland, Single Layer Gp grassland (both managed and natural)
20299-Zt1(4)[Z1](1) A1A11B1XXXXXXF2F4F7G4-A13G11-Zt1Z1 G2 Grassland with Very Open (30-10%) Woody Vegetation Open (50 - 10%) Woody Vegetation with Medium to Tall Herbaceous Layer

Floristic Aspect: Graminoids User-defined classifier (Z1): Clusters of scattered woody vegetation, following regular pattern, cover 10 to 30% of the area

Gr_Wvop _1 closed grassland with woody vegetation, where grassland is predominant. Clusters of woody vegetation cover the area in the range of 10 to 30%
20299-Zt1(4)[Z2](1) A1A11B1XXXXXXF2F4F7G4-A13G11-Zt1Z2 G3 Grassland with Very Open (50-30%) Woody Vegetation Open (50 - 10%) Woody Vegetation with Medium to Tall Herbaceous Layer

Floristic Aspect: Graminoids User-defined classifier (Z2): Clusters of scattered woody vegetation, following regular pattern, cover 30 to 50% of the area

Gr_Wvop_2 closed grassland with woody vegetation, where grassland is predominant. Clusters of woody vegetation cover the area in the range of 30 to 50%

Step 3: Assess the class reduction coefficient (RC)

Description Image

For each of the defined classes:

  • Define the minimum area of interest (AOI) that correctly represent the spatial pattern of the intrinsic mix between agriculture and non-agriculture land cover features
  • Use the precise land cover definition from the eligibility profile
  • Determine the optimal timing of the year to detect the abundance of the vegetation and to spot any relevant agriculture activity

Note: Optimal timing is often when the vegetation reaches its maximum phenological development

Areas of interest on top of extract from World_Imagery (http://services.arcgisonline.com/arcgis/services)

For each of the defined classes:

  • Locate and select a sample of at least 30 such areas of interest randomly spread all over the country/region
  • Derive the abundance of the grassland for each of the areas of interest with the appropriate method:
  • Manual photointerpretation
  • Automated classification (pixel or object based)
  • Applying pre-set scorecards
  • Perform statistical analysis of the resulted RC (non-agri land/grassland)
  • Results should ahdere to a normal (Gaussian) distrbution and standard deviation (StDev) is less than 5%
  • If not AOI or class definition should be refined
RC measurement and analysis - extract from World_Imagery (http://services.arcgisonline.com/arcgis/services); histogram of Synthetic Gaussian Data, Math-Blog (http://math-blog.com/2011/04/25/plotting-and-graphics-in-octave)

Step 4: Categorize class reduction coefficient into its pro-rata category

Description Result
For classes with samples yielding normal distribution and low StDev:
  • Determine the pro-rata categories by:
  • Calculating the mean RC from the sample
  • For mean RC less than 10%, apply no reduction"
  • For any RC above 10%, derive the pro rata rate by substracting the RC from 100, and round to nearest 5 to get the eligibility rate
  • Update the eligibility profile, if needed (in this case user defined Z-attributes in LCCS will be updated)
See outcome table below

outcome: eligibility profile classes

LC Class observed RC [%] Pro-rata rate=100-RC Pro rata category Eligibility rate MapCode User description
Class G1 0 100 Grassland 100% Gp Grassland (both managed and natural)
Class G2 22,3 77,7 Grassland80 80% Gr_Wvop_1 Closed grassland with woody vegetation, where grassland is predominant.

Clusters of woody vegetation cover the area in the range of 10 to 30%

Class G3 38,5 61,5 Grassland60 60% Gr_Wvop_2 Closed grassland with woody vegetation, where grassland is predominant.

Clusters of woody vegetation cover the area in the range of 30 to 50%

Step 5: Establish the reference area for the reference parcel

Description Image
  • Delineate all agriculture land cover (different types of grassland) following the class entries in the eligibility profile
  • Exclude any individual clusters of woody vegetation (not grazed and not accessible)
  • Sum up the resulted polytgon areas per category
  • Calculate the reference area for the RP, following the table template
Mapped pro rata polygons on top of extract from World_Imagery (http://services.arcgisonline.com/arcgis/services)

outcome: reference parcel reference area

polygons LC Class Pro rata category Eligibility rate Geometric area [ha] Eligible hectares [ha]
1 G1 Grassland 100% 23,5 23,5
2+3 G2 Grassland80 80% 35,6 28,5
4+5+6+7 G3 Grassland60 60% 7,8 4,7
1-7 PG n/a n/a 66.9 56,7


Field examples can be found here



Example of sporadic pro rata application

The parcel in the figure has pro rata grassland, but no systematic land cover grassland class has been defined.


Action Illustration

Step 1: Prepare the reference or agricultural parcel for assessment Retrieve ancillary information if available.

The parcel's polygon in the picture delineates an area of 3.2 ha.

a parcel with pro rata grassland on top of extract from World_Imagery (http://services.arcgisonline.com/arcgis/services)
]]

Step 2: Establish the proportion of grassland

In this case:

  • by using a scorecard or other means.
  • that assessment reveals that 77% of the area is actually covered by grass or other herbaceous species that can be grazed.
example of an area frame scorecard; source: Paola Codipietro, FAO

Step 3: Reduce the polygon area according to the corresponding pro rata category to obtain the new reference area / maximum eligible area.

In this case,

  1. the proportion of grassland is 77%
  2. the resulting category is grassland80
  3. the reference area will be set at 0.8 x 3.2 ha = 2.56ha
See the conversion table below

Conversion table

percentage [%] Pro rata category Eligibility rate
90-100 Grassland 100%
70-89.9 Grassland80 80%
50-69.9 Grassland60 60%
<50 Check!!! tbd


A field example can be found here

Go up to the main pro rata page