Services Directory Login

Kilombero_catchment_land_cover_2018_WTL1 (Map Service)


View In:    ArcGIS JavaScript   ArcGIS.com   WMTS  

Current Version: 10.81

Service Description:
Summary: Land cover map for the year 2018 of the Kilombero catchment in Tanzania at 20m resolution.

Recommended citation: Cisneros, P. & Ramírez, M. (2019). Land cover, habitat and wildlife corridor mapping in the Kilombero catchment (Unpublished Master's Thesis). University of Copenhagen, Copenhagen, Denmark.

Satellite Data: Sentinel 2 composite (Nov 2016-Nov 2018) | Sentinel 1 composite (Feb 2017-Nov 2018)

Spatial resolution: 20 mts

Classification method (before further manual minor improvements based on expert opinions)
Random Forest
Overall accuracy = 75%
Kappa accuracy = 70%

Identifying 8 classes: Tree Cover (1), which will be reclassified into Moist Forest, Dry Forest and Tree Plantation* classes in another Random Forest classification;  Shrub cover / Natural Mosaic (2); Grassland  (3); Croplands (4); Wetlands (5); Urban Areas (6); Open Water (7); and Clouds (8), which is removed and substituted by the other classes in a later step of the process. Roads layer was obtained from the World Food Programme Geonode portal. It was published on May 2018. Available at https://geonode.wfp.org/layers/ogcserver.gis.wfs.org%3Ageonode%3Atza_trs_road[Accessed 11 Feb 2019]

* The biggest areas of Tree Plantations, Large-scale Crops, and Railways have been done manually based on the classification results and other maps available.



Map Name: Map

Legend

All Layers and Tables

Layers: Tables: Description:
Summary: Land cover map for the year 2018 of the Kilombero catchment in Tanzania at 20m resolution.

Recommended citation: Cisneros, P. & Ramírez, M. (2019). Land cover, habitat and wildlife corridor mapping in the Kilombero catchment (Unpublished Master's Thesis). University of Copenhagen, Copenhagen, Denmark.

Satellite Data: Sentinel 2 composite (Nov 2016-Nov 2018) | Sentinel 1 composite (Feb 2017-Nov 2018)

Spatial resolution: 20 mts

Classification method (before further manual minor improvements based on expert opinions)
Random Forest
Overall accuracy = 75%
Kappa accuracy = 70%

Identifying 8 classes: Tree Cover (1), which will be reclassified into Moist Forest, Dry Forest and Tree Plantation* classes in another Random Forest classification;  Shrub cover / Natural Mosaic (2); Grassland  (3); Croplands (4); Wetlands (5); Urban Areas (6); Open Water (7); and Clouds (8), which is removed and substituted by the other classes in a later step of the process. Roads layer was obtained from the World Food Programme Geonode portal. It was published on May 2018. Available at https://geonode.wfp.org/layers/ogcserver.gis.wfs.org%3Ageonode%3Atza_trs_road[Accessed 11 Feb 2019]

* The biggest areas of Tree Plantations, Large-scale Crops, and Railways have been done manually based on the classification results and other maps available.



Copyright Text: Cisneros, P. & Ramírez, M. (2019). Land cover, habitat and wildlife corridor mapping in the Kilombero catchment (Unpublished Master's Thesis). University of Copenhagen, Copenhagen, Denmark.

Spatial Reference:
102100

Single Fused Map Cache: true

Capabilities: Map,TilesOnly,Tilemap

Tile Info:
Initial Extent:
Full Extent:
Min Scale: 9.2394397E7
Max Scale: 2343.0

Min LOD: 8
Max LOD: 14

Units: esriMeters

Supported Image Format Types: Mixed

Export Tiles Allowed: false
Max Export Tiles Count: 100000

Document Info: