Canopy_Damage_2017_Fires_Tile_Service (Map Service)
View In:
ArcGIS JavaScript
ArcGIS.com
WMTS
Current Version: 10.81
Service Description: Under a grant from NASA, Sonoma County Ag + Open Space created these canopy damage maps, which depict the percent of the woody canopy (greater than 7 feet in height) damaged by the 2017 Nuns, Tubbs, and Pocket fires. The canopy damage maps reflect the state of the landscape in June, 2018, when 1-foot resolution 4-band imagery of the fires was collected by Quantum Spatial. That imagery is available as an image service and as TIFs.
Methods
The canopy damage maps were created using object-based image analysis (OBIA) in Trimble Ecognition. The analysis utilized as its inputs the 2013 normalized digital surface model (nDSM), which depicts canopy height in 2013, the 2013 6-inch resolution, 4-band orthophotos, and the June 2018, 1-foot resolution, 4-band orthophotos. Segmentation and decision rules utilized the nDSM, the NIR bands of both images, NDVI from both images, and the visible atmospherically resistant index (VARI) from the 2018 image.
The Ecognition ruleset combined image segmentation within fine-scale vegetation map polygons with assignment of ‘canopy damage percent’ and ‘canopy damage class’ for each segment. See Table 1 (next page) for a full list of attributes in the damage maps. The ruleset implemented the following steps to assign percent burned and damage class labels:
- Identified areas not classified as forest or shrub in the 2013 vegetation map – all analysis was limited to forest and shrub areas within the 2017 fire perimeters.
- Performed segmentation within existing vegetation map polygons using the post-fire image to divide vegetation polygons up based on canopy damage (using 2018 NDVI to drive segmentation). For the purposes of these bullets we will refer to these segments as super-objects.
- For the super-objects, in areas where the 2013 nDSM showed a canopy height >= 7 feet, create very small sub-objects to classify as burned or unburned.
- Classify sub-objects as burned or unburned using a number of spectral indices and heuristic rules.
- Label each super-object with its canopy damage percent based on the burned versus unburned composition of its sub-objects.
- Bin the super-objects into one of 6 canopy damage classes based on the object’s canopy damage percent.
Accuracy assessment was conducted for the automated damage maps. Accuracy assessment was fuzzy, allowing a correct map percent damage call to be within +/- 10% of the photointerpreted accuracy assessment percent damage. Initial accuracy assessment indicated an overall accuracy of 80%.
Analysts performed a round of manual editing that resulted in changes to 6,875 of the 46,835 total polygons. This round of manual editing increased accuracy from 80% to 85%. Table 1 shows the attributes of the canopy damage maps. The accuracy assessment revealed that low-vigor forest types, especially Quercus douglasii and Pinus sabiniana, had the lowest overall accuracies, with systematic overmapping of damage in these types.
Table 1. Attributes for the damage maps
Fieldname | Alias | Description |
OBJECTID | OBJECTID | Internal index |
CANOPY_DAMAGE_CLASS | Canopy Damage Class | Canopy damage to canopy over 7 feet tall in 5 classes: 0-5%, 5-20%, 20-40%, 40-60%, 60-80%, 80-100% |
CANOPY_DAMAGE_PERCENT | Canopy Damage Percent | Percent damage to canopy over 7 feet tall |
PRE_FIRE_MAP_CLASS | Pre-fire Map Class | Fine scale map class from the 2013 fine scale vegetation map |
PRE_FIRE_RELATIVE_COVER | Pre-fire Relative Cover | Relative conifer v. hardwood cover from the 2013 fine scale vegetation map |
PRE_FIRE_LIFEFORM | Pre-fire Lifeform | Lifeform from the 2013 fine scale vegetation map |
PRE_FIRE_FOREST_LIFEFORM | Pre-fire Forest Lifeform | Forest lifeform from the 2013 fine scale vegetation map |
FIRE_NAME | Fire name | Name of fire (Nuns, Tubbs, Pocket) |
CANOPY_AREA | Sq. Ft. of Canopy | Approximate area of the segment that is over 7 feet tall (our cutoff for ‘canopy’) |
CANOPY_PERCENT | Percent of Canopy | Approximate percent of the segment that is over 7 feet tall (our cutoff for ‘canopy’) |
OIDCOPY_VEGMAP | OID Copy Fine Scale Veg Map | The OID_COPY from the fine scale veg map; used for table joins to the veg map |
OID_COPY | OID Copy | Unique index for the canopy damage maps |
Layers
Legend
All Layers and Tables
Layers:
Tables:
Description: Under a grant from NASA, Sonoma County Ag + Open Space created these canopy damage maps, which depict the percent of the woody canopy (greater than 7 feet in height) damaged by the 2017 Nuns, Tubbs, and Pocket fires. The canopy damage maps reflect the state of the landscape in June, 2018, when 1-foot resolution 4-band imagery of the fires was collected by Quantum Spatial. That imagery is available as an image service and as TIFs.
Methods
The canopy damage maps were created using object-based image analysis (OBIA) in Trimble Ecognition. The analysis utilized as its inputs the 2013 normalized digital surface model (nDSM), which depicts canopy height in 2013, the 2013 6-inch resolution, 4-band orthophotos, and the June 2018, 1-foot resolution, 4-band orthophotos. Segmentation and decision rules utilized the nDSM, the NIR bands of both images, NDVI from both images, and the visible atmospherically resistant index (VARI) from the 2018 image.
The Ecognition ruleset combined image segmentation within fine-scale vegetation map polygons with assignment of ‘canopy damage percent’ and ‘canopy damage class’ for each segment. See Table 1 (next page) for a full list of attributes in the damage maps. The ruleset implemented the following steps to assign percent burned and damage class labels:
- Identified areas not classified as forest or shrub in the 2013 vegetation map – all analysis was limited to forest and shrub areas within the 2017 fire perimeters.
- Performed segmentation within existing vegetation map polygons using the post-fire image to divide vegetation polygons up based on canopy damage (using 2018 NDVI to drive segmentation). For the purposes of these bullets we will refer to these segments as super-objects.
- For the super-objects, in areas where the 2013 nDSM showed a canopy height >= 7 feet, create very small sub-objects to classify as burned or unburned.
- Classify sub-objects as burned or unburned using a number of spectral indices and heuristic rules.
- Label each super-object with its canopy damage percent based on the burned versus unburned composition of its sub-objects.
- Bin the super-objects into one of 6 canopy damage classes based on the object’s canopy damage percent.
Accuracy assessment was conducted for the automated damage maps. Accuracy assessment was fuzzy, allowing a correct map percent damage call to be within +/- 10% of the photointerpreted accuracy assessment percent damage. Initial accuracy assessment indicated an overall accuracy of 80%.
Analysts performed a round of manual editing that resulted in changes to 6,875 of the 46,835 total polygons. This round of manual editing increased accuracy from 80% to 85%. Table 1 shows the attributes of the canopy damage maps. The accuracy assessment revealed that low-vigor forest types, especially Quercus douglasii and Pinus sabiniana, had the lowest overall accuracies, with systematic overmapping of damage in these types.
Table 1. Attributes for the damage maps
Fieldname | Alias | Description |
OBJECTID | OBJECTID | Internal index |
CANOPY_DAMAGE_CLASS | Canopy Damage Class | Canopy damage to canopy over 7 feet tall in 5 classes: 0-5%, 5-20%, 20-40%, 40-60%, 60-80%, 80-100% |
CANOPY_DAMAGE_PERCENT | Canopy Damage Percent | Percent damage to canopy over 7 feet tall |
PRE_FIRE_MAP_CLASS | Pre-fire Map Class | Fine scale map class from the 2013 fine scale vegetation map |
PRE_FIRE_RELATIVE_COVER | Pre-fire Relative Cover | Relative conifer v. hardwood cover from the 2013 fine scale vegetation map |
PRE_FIRE_LIFEFORM | Pre-fire Lifeform | Lifeform from the 2013 fine scale vegetation map |
PRE_FIRE_FOREST_LIFEFORM | Pre-fire Forest Lifeform | Forest lifeform from the 2013 fine scale vegetation map |
FIRE_NAME | Fire name | Name of fire (Nuns, Tubbs, Pocket) |
CANOPY_AREA | Sq. Ft. of Canopy | Approximate area of the segment that is over 7 feet tall (our cutoff for ‘canopy’) |
CANOPY_PERCENT | Percent of Canopy | Approximate percent of the segment that is over 7 feet tall (our cutoff for ‘canopy’) |
OIDCOPY_VEGMAP | OID Copy Fine Scale Veg Map | The OID_COPY from the fine scale veg map; used for table joins to the veg map |
OID_COPY | OID Copy | Unique index for the canopy damage maps |
NASA, Sonoma County Ag + Open Space, Tukman Geospatial
Spatial Reference: 102100
Single Fused Map Cache: true
Capabilities: Map,TilesOnly
Tile Info:
- Height: 256
- Width: 256
- DPI: 96
- Levels of Detail: (# Levels: 20)
- Level ID: 0 [Start Tile, End Tile]
Resolution: 156543.03392800014
Scale: 5.91657527591555E8
- Level ID: 1 [Start Tile, End Tile]
Resolution: 78271.51696399994
Scale: 2.95828763795777E8
- Level ID: 2 [Start Tile, End Tile]
Resolution: 39135.75848200009
Scale: 1.47914381897889E8
- Level ID: 3 [Start Tile, End Tile]
Resolution: 19567.87924099992
Scale: 7.3957190948944E7
- Level ID: 4 [Start Tile, End Tile]
Resolution: 9783.93962049996
Scale: 3.6978595474472E7
- Level ID: 5 [Start Tile, End Tile]
Resolution: 4891.96981024998
Scale: 1.8489297737236E7
- Level ID: 6 [Start Tile, End Tile]
Resolution: 2445.98490512499
Scale: 9244648.868618
- Level ID: 7 [Start Tile, End Tile]
Resolution: 1222.992452562495
Scale: 4622324.434309
- Level ID: 8 [Start Tile, End Tile]
Resolution: 611.4962262813797
Scale: 2311162.217155
- Level ID: 9 [Start Tile, End Tile]
Resolution: 305.74811314055756
Scale: 1155581.108577
- Level ID: 10 [Start Tile, End Tile]
Resolution: 152.87405657041106
Scale: 577790.554289
- Level ID: 11 [Start Tile, End Tile]
Resolution: 76.43702828507324
Scale: 288895.277144
- Level ID: 12 [Start Tile, End Tile]
Resolution: 38.21851414253662
Scale: 144447.638572
- Level ID: 13 [Start Tile, End Tile]
Resolution: 19.10925707126831
Scale: 72223.819286
- Level ID: 14 [Start Tile, End Tile]
Resolution: 9.554628535634155
Scale: 36111.909643
- Level ID: 15 [Start Tile, End Tile]
Resolution: 4.77731426794937
Scale: 18055.954822
- Level ID: 16 [Start Tile, End Tile]
Resolution: 2.388657133974685
Scale: 9027.977411
- Level ID: 17 [Start Tile, End Tile]
Resolution: 1.1943285668550503
Scale: 4513.988705
- Level ID: 18 [Start Tile, End Tile]
Resolution: 0.5971642835598172
Scale: 2256.994353
- Level ID: 19 [Start Tile, End Tile]
Resolution: 0.29858214164761665
Scale: 1128.497176
- Format: PNG
- Compression Quality: 0
- Origin:
X: -2.0037508342787E7
Y: 2.0037508342787E7
- Spatial Reference:
102100
Initial Extent: XMin: -1.377740740166902E7
YMin: 4606447.456992517
XMax: -1.3533048363884935E7
YMax: 4700481.875912493
Spatial Reference:
102100
Full Extent: XMin: -1.3687175649764894E7
YMin: 4611949.10654559
XMax: -1.3622894666209862E7
YMax: 4696187.110264202
Spatial Reference:
102100
Min Scale: 577790.554289
Max Scale: 4513.988705
Min LOD: 10
Max LOD: 17
Units: esriMeters
Supported Image Format Types: PNG
Export Tiles Allowed: false
Max Export Tiles Count: 100000
Document Info: - Title: Canopy Damage Maps - 2017 North Bay Fires (Tile Service)
- Author: SonomaOpenSpace
- Comments:
- Subject: Canopy Damage Maps - 2017 North Bay Fires
- Category:
- Keywords: Sonoma, Fire, Nuns, Tubbs, Pocket, 2017