Patch metrics and landscape patterns of forest disturbances at the beginning of the 20th Century

Summary:

The database consists of a compressed .CSV file containing structural information of forest disturbance patches identified between 2002 and 2014 using the Global Forest Change Tree Cover Loss Year dataset version 1.6 (Hansen et al, 2013) available at https://earthenginepartners.appspot.com/science-2013-global-forest/download_v1.6.html. Each row in the database represents a patch (249,149,911 in total). The columns (15) represent the structural metrics calculated for each patch, as well as the landscape patterns identified using kmeans cluster analysis. 

The methods used for building this database are published in the paper: Acil N, Sadler JP, Senf C, Suvanto S & Pugh TAM (2024) Landscape patterns in stand-replacing disturbances across the world’s forests. Nature Sustainability. https://doi.org/10.1038/s41893-024-01450-3

Aggregated global maps of the patch metrics can be visualised in Google Earth Engine and accessed in the asset "http://projects/ee-treemort-disturbances-nacil/assets/PatchMetrics_Means_nonLU_2002-2014/". 

Some of the scripts associated with this project are hosted in GitHub and Google Earth Engine.

Additional scripts and data will be made available upon request.

Database structure: 

Patch metrics

Occurrence: 

Patch form and year were retrieved from the Global Forest Change tree cover loss year dataset version 1.6 (Hansen et al, 2013).

Column name
Description
Unit
Format
Valid values

PID
Patch unique identifier in the format Tile_Year_PatchNumber (e.g. 01U_02_00000001).

Characters

X_INT_deg
Longitude of the patch's internal centroid
Degrees
Float
[-180-180]

Y_INT_deg
Latitude of the patch's internal centroid
Degrees
Float
[-90-90]

YEAR_maj
Year of patch majority occurrence. 

Integer
[2-14]

YEAR_n
Number of years over which the patch exhibited continuous growth.

Integer
>0

Metrics: 

These patch and landscape metrics were calculated from the patch delineated.

Column name
Description
Unit
Format
Valid values

AREA_G_ha
Patch geodesic area
Hectares
Float
>0

PERIM_G_m
Patch geodesic perimeter
Meters
Float
>0

PARA
Perimeter-area ratio

Float
>0

SHAPE
Shape index

Float
>=1

ELONG
Elongation index

Float
[0-1[

FRAC
Fractal dimension index

Float
[1-2]

NN5000_T0_n
Number of patches assigned the same year within 5 km radius.

Integer
>0

NN5000_AREA_T0_perc
Percent of the total area disturbed over the period 2001-2018 within 5 km radius from the focal patch centroid.
%
Float
[0-100]

Clusters:

Cluster identification was performed using AREA_G_ha, YEAR_n, SHAPE, ELONG, NN5000_T0_n and NN5000_AREA_T0_perc. 

Column name
Description
Unit
Format
Valid values

CLUSTER_CODE
Code assigned to each cluster

Integer
[1-4]

CLUSTER_LABEL
Name given to the cluster identified. 

Character

Small-isolated

Clustered

Complex

Large-multiyear