LSMetricsDescription

Brief description

The package

LandScape metrics (LSMetrics) is a free and open source package developed calculates multiple landscape metrics for raster data.

The package uses input raster maps with integer values only, in which each cell represents an area considered to be homogeneous, such as a land use or vegetation type. The maps can be either binary (1 = habitat, 0 = non-habitat) or multi-class (e.g. land use and land cover maps). The majority of landscape metrics are calculated using binary class raster maps, except for the landscape diversity indices, which only make sense for multi-class raster maps. LSMetrics may also transform multiple class maps into binary ones before the calculation of metrics.

Once a GRASS project is created and raster maps are imported into it, using r.import or r.in.gdal (or other r.in.*) modules, for instance, LSMetrics may be run in two ways (see the Figure below). The first is calling a python application and opening the GUI; the second is building a Python script (or opening a Python shell inside GRASS GIS prompt) and calling each landscape metric as a Python function. Both methods allow the users to run multiple metrics with various parameters and scales, for multiple maps, in a single run.

The output maps consist in raster maps for each chosen landscape metric, as well as text files with statistics of some of the metrics at the patch or class level, to be analyzed with any statistical software.

Metrics

LSMetrics currently perform the following calculations:

Preparation of inputs:

• Transformation of land use maps in binary class maps

Metrics of structural connectivity:

• Patch size
• Fragment size
• Structural connectivity
• Proportion of habitat

Metrics of functional connectivity:

• Functionally connected area
• Functional connectivity
• Complete functionally connected area

Edge-based metrics:

• Classification in core/edge/matrix
• Classification in landscape elements: edge/core/stepping stones/corridors/branches/matrix
• Binary maps: edge/non-edge
• Binary maps: core/non-core
• Proportion of edge area
• Proportion of core area
• Area of clumps of edge and core areas

Landscape diversity (through the r.diversity GRASS addon):

• Shannon
• Simpson
• Pielou
• Rényi

For more information on the metrics calculated and details on implementation, look at the publication:

Niebuhr, B. B. S.; Martello, F.; Ribeiro, J. W.; Vancine, M. H.; Muylaert, R. L.; Campos, V. E. W.; Santos, J. S.; Tonetti, V. R.; Ribeiro, M. C. Landscape Metrics (LSMetrics): a spatially explicit tool for calculating connectivity and other ecologically-scaled landscape metrics. In preparation.

The repository

The LSMetrics repository is organized in 7 folders:

• _LSMetrics_v1_0_0: Here the main pieces of the LSMetrics code are located:

  1. LSMetrics_v1_0_0.py: main script.
  2. test_LSMetrics.py: a Python script with the list of functions of LSMetrics and their usage as Python functions (outside GUI).
  3. r_diversity.py: The r.diversity GRASS addon as python code, used to calculate landscape diversity indices in LSMetrics.

• previous_versions: Old versions of the code.

• grassdb_test: raster maps for testing. This includes:

  1. APA_Sao_Joao_RJ_cut_SIRGAS_UTM23S.tif: A land use map in Rio de Janeiro state, Brazil, inside the Golden Lion Tamarin occurrence region. The map was classified based on LANDSAT 7 satellite images.
  2. SP_RioClaro_use_raster.tif: A land use raster of the municipality of Rio Claro, State of São Paulo, Brazil.