Landscape Mosaic#

The Landscape Mosaic analysis classifies each pixel based on the proportional composition of three land cover classes within a moving window. The result describes the local landscape context of each pixel in terms of the dominant land cover mixture, producing up to 103 compositional classes subsequently remapped to 19 aggregated classes. The methodology is described in detail in the Landscape Mosaic sheet.

Landscape Mosaic Classes#

The 19-class aggregation is based on the proportion of the three input land cover classes within the moving window. By default, the three classes are interpreted as:

Pixel Value

Default Interpretation

Color

1

Agriculture

Blue

2

Natural

Green

3

Developed

Red

Note

The class assignment is purely conventional. Pixel values 1, 2 and 3 can represent any three mutually exclusive land cover types defined by the user (e.g. Forest/Non-forest/Water, Urban/Rural/Natural). The labels Agriculture, Natural and Developed are used throughout this documentation for consistency with the GuidosToolbox convention, but the analysis is valid for any three-class input map.

LM triangle

The Landscape Mosaic triangle with the 19 classes and their proportions to the three land cover types Agriculture, Natural, and Developed.#

Each of the 19 aggregated classes is defined by the combination of proportions of the three input classes within the moving window:

Landscape Mosaic 19-class aggregation scheme.#

N

Code

Description

AGR [%]

NAT [%]

DEV [%]

1

A

Agriculture dominant

[80-100[

[0-10[

[0-10[

2

D

Developed dominant

[0-10[

[0-10[

[80-100[

3

N

Natural dominant

[0-10[

[80-100[

[0-10[

4

Ad

Agriculture with Developed

[60-90[

[0-10[

[10-60[

5

An

Agriculture with Natural

[60-90[

[10-40[

[0-10[

6

Dn

Developed with Natural

[0-10[

[10-40[

[60-90[

7

Da

Developed with Agriculture

[10-40[

[0-10[

[60-90[

8

Na

Natural with Agriculture

[10-40[

[60-90[

[0-10[

9

Nd

Natural with Developed

[0-10[

[60-90[

[10-40[

10

Adn

Agriculture dominant mixed

[60-80[

[10-40[

[10-40[

11

Dan

Developed dominant mixed

[10-40[

[10-40[

[60-80[

12

Nad

Natural dominant mixed

[10-40[

[60-80[

[10-40[

13

ad

Agriculture-Developed transition

[30-60[

[0-10[

[30-60[

14

an

Agriculture-Natural transition

[30-60[

[30-60[

[0-10[

15

dn

Developed-Natural transition

[0-10[

[30-60[

[30-60[

16

adn

Mixed transition

[10-60[

[10-60[

[10-60[

17

NN

Pure Natural (100%)

[0]

[100]

[0]

18

AA

Pure Agriculture (100%)

[100]

[0]

[0]

19

DD

Pure Developed (100%)

[0]

[0]

[100]

Usage#

import pyguidos as pg

result = pg.landmos(
    in_tiff="my_landcover.tif",
    window_size=31,
    outdir="output/",
    statists=True,
    stat_files=True,
    out_colors='bgr',
    return_array=False,
    verb=False
)

Parameters#

Parameter

Type

Default

Description

in_tiff

str or Path

–

Path to input GeoTIFF

window_size

int

–

Moving window size in pixels, odd integer >= 3

outdir

str or Path

None

Output directory

statists

bool

True

Compute statistics

stat_files

bool

True

Write statistics to files

out_colors

str

'bgr'

Color scheme for the 103-class output colormap

return_array

bool

False

Return output array

verb

bool

False

Print progress messages

Output Files#

File

Description

<name>_lm_<window_size>_103class.tif

103-class landscape mosaic result GeoTIFF

<name>_lm_<window_size>.tif

19-class remapped result GeoTIFF

<name>_lm_<window_size>.txt

Statistics report

<name>_lm_<window_size>.csv

Per-value pixel counts and frequencies

<name>_lm_<window_size>_heatmap.csv

Ternary diagram data table

<name>_lm_<window_size>_heatmap.png

Ternary diagram heatmap

Result Object#

landmos() returns a LandMosResult dataclass:

result = pg.landmos("my_landcover.tif", window_size=33)

# Access statistics
print(result.stats.keys())
# dict_keys(['output paths', 'input stats', 'output stats'])

# Input pixel counts
print(result.stats["input stats"])
# {'class1 pxl': 15000, 'class2 pxl': 20000, 'class3 pxl': 10000,
#  'foreground pxl': 45000, 'missing pxl': 0}

# Pixel counts for both 103-class and 19-class outputs
print(result.stats["output stats"].keys())
# dict_keys(['pxl numb 103cl', 'pxl numb 19cl'])

# Output file paths
print(result.stats["output paths"])
# {'path tif': 'output/my_landcover_lm_33_103class.tif',
#  'path txt': 'output/my_landcover_lm_33.txt',
#  'path csv': 'output/my_landcover_lm_33.csv',
#  'path csv hm': 'output/my_landcover_lm_33_heatmap.csv',
#  'path png': 'output/my_landcover_lm_33_heatmap.png'}

Note

result.stats["output paths"] is None when stat_files=False. All other keys are always populated regardless of stat_files.

Computing Statistics Separately#

If you already have a Landscape Mosaic output GeoTIFF, you can compute statistics without rerunning the analysis:

stats = pg.landmos_stats(
    lm_tiff="output/my_landcover_lm_33_103class.tif",
    outfile=True,
    outdir="output/",
    source_tiff="my_landcover.tif"
)

Note

landmos_stats() requires the input GeoTIFF to be an pyGuidos (or GTB) output raster file. . See Input Format for details.

References#

  • Riitters K H, Wickham J D, Wade T G, 2009. An indicator of forest dynamics using a shifting landscape mosaic. Ecological Indicators 9: 107-117. DOI: 10.1016/j.ecolind.2008.02.003.

  • Vogt P, Wickham J, Barredo J I, Riitters K, 2024. Revisiting the Landscape Mosaic model. PLoS ONE 19(5): e0304215. DOI: 10.1371/journal.pone.0304215.