Spatial analysis with

Introduction to GIS for ecologists

November 2025

Romain Frelat

Data scientist
@ FRB-CESAB    

Objectives





  • Feel more comfortable to work with spatial data
  • Load spatial data in and visualize it
  • Extract information from remote sensing or model datasets
  • Take advantage of the numerous databases that are available online

Vies-à-Vies

Observations in 2021 from GBIF

Schedule

Time Activities
9.00 – 9.30 am
Introduction
9.30 – 12.00 am
Tutorials on handling vectors data type:
points, lines, and polygons
12.00 – 1.30 pm Lunch break
1.30 – 4.00 pm
Tutorials on handling raster data type,
high dimensioinal rasters, and make a grid

It’s a practical workshop, you will work in front of the screen, so don’t forget to take breaks !

What is Geographic Information System (GIS)?

Geographic Information System

A geographic information system (GIS) consists of integrated computer hardware and software that store, manage, analyze, edit, output, and visualize geographic data. (Wikipedia)

  • softwares to handle spatial data
  • data stored with their coordinates
  • defined by their projection systems


Two key concepts in GIS:

  • Coordinate Reference Systems (CRS)
  • data type: raster vs vector

Coordinate Reference Systems

All maps are wrong

There are many many projections

The most commun projection systems

One system to code them all: EPSG1

The projection system will depend on the objective and the spatial extent.

Local

  • France:
    Lambert 93, conformal conic projection EPSG:2154
  • Europe:
    Lambert azimuthal equal-area EPSG:3035
  • Other: Universal Transverse Mercator (UTM), e.g. 32N for Western Europe EPSG:25832

World

When and why must you care about projection systems?


Crossing multiple spatial data from different sources

  • Keep track of the projection system of your observation data
  • Make sure all datasets share the same projection system or project them


Calculating distances, or areas1

  • Use appropriate projection system (e.g. equal-area for areas or equidistant for distances)
  • The projection system will define the unit

Two world views: Raster vs Vector

Vector vs Raster data type

Vector

  • Points, Lines or Polygons
  • Conserve shape and spatial accuracy
  • Good for GPS observations, patches, boundaries, …
  • Rivers
  • Protected areas

adapted from https://sist.pages.in2p3.fr/anf23-geomatique/

Rivers ?

Vegetation index (e.g. NDVI)?

Elevation ?

Protected areas ?

Land use-land cover ?

Raster

  • Matrix, regular grid with equidistant cells
  • Spatial resolution = cell size
  • Format for satellite data, model output, …
  • Vegetation index (e.g. NDVI)
  • Elevation

Vector vs Raster file format

Vector

  • Geopackage(.gpkg)
    open format based on SQLite
  • ESRI shapefile (.shp)
    historical format from ESRI.
    Multiple files (dbf, shx, shp, prj)
  • GPS Exchange Format (.gpx)
    XML file designed for GPS data
  • geojson (.geojson)
    plain text file, used in web API but limited to WGS84
  • Google Earth (.kml)

adapted from https://sist.pages.in2p3.fr/anf23-geomatique/

Raster

  • GeoTIFF (.tif)
    the standard format. Can store multiple “bands”.
  • NetCDF (.nc)
    for storing multi-dimensional, array-oriented variables
  • ASCII Grid (.asc)
    simple text files
  • GRIdded Binary (.grb), WMO standard
  • Imagine (.img) ERDAS file format

Where can we find spatial data?

Data sources



France

Europe

World


Topics

Climate

Land cover

Soil


Elevation and bathymetry

Biomes and Ecoregions

Population density


Keeping track of new spatial data

MetaGIS


Let’s help each other


Feel free to add new dataset


Share it with colleagues : terrameta

as a GIS tool

Is a good GIS software?

Desktop GIS (GUI)
ArcGIS, Quantum GIS		 Scripting
, Python1
Home disciplines Geography Computing, Statistics
Software focus Graphical User Interface Command line
Reproducibility Minimal Maximal


packages

In these tutorials, we will see how to use terra, sf, and mapview.

Other interesting packages:

  • mapsf and tmap: creates nice looking maps
  • exactextractr : fast extraction on large raster

Recently deprecated packages (end of 2023) : sp, raster, rgdal, rgeos, maptools

Summary of key functions in terra

Vector

# read
vector <- vect("my_polygons.gpkg")

# write
writeVector(vector, "my_file.gpkg")

# calculate distance or perimeter
perim(vector)

# calculate area for polygons
expanse(vector)

# calculate distance between objects
distance(vector)

# create buffers around objects
buffer(vector, set_dist)

# find spatially interesecting features
intersect(vector1, vector2)

# transform vector to raster format
rasterize(vector)

<>

# coordinate reference system
crs()

# extent of the spatial object 
ext() # (xmin, xmax, ymin, ymax)

# getting the dimensions of the object
dim()

# making a simple map
plot()

# projection to different CRS
project()

# extract cell values at vector locations
extract(vector, raster)

# remove area outside extent of interest
crop()
mask()

Raster

# read
raster <- rast("my_raster.tif")

# write
writeRaster(raster, "my_file.tif")

# resolution of the raster
res(raster)

# decrease the resolution
aggregate(raster, fact = 2)

# change the grid 
resample(raster1, raster2)

# transform into vector
as.line(raster)
as.polygon(raster)

Your turn !


https://frbcesab.github.io/spatial-r/