Introduction to Geographical Information Systems

class: center, middle, inverse, title-slide

# Introduction to Geographical Information Systems
## Unit 655
### Benjamin Nowak, VetAgro Sup
### May-June 2022

---

background-color: #e76f51
background-size: cover
class: center, bottom, inverse

# **What is GIS?**

### A computer system to analyze and display spatial data

---

# Projection

- A 3D spherical surface cannot be converted to a 2D map without distortions
 - *Such as trying to flatten the skin of an orange (Picture: [Manson and Kernik](https://open.lib.umn.edu/mapping/chapter/3-scale-and-projections/))*

- It is also necessary to reflect the relief of the Earth's surface

---

# World projection

.pull-left[
####Mercator projection
*Deformation increase with latitudes, with Greenland and Antarctica far bigger than reality*

<center><img src="fig/gis/mercator.jfif" width=70%></img></center>
]

.pull-right[
####Waterman "butterfly"
*One of the projection with the least distortion*
 
<center><img src="fig/gis/waterman.jfif" width=100%></img></center>
]

**Pictures:** [Wikipedia](https://en.wikipedia.org/wiki/Azimuthal_equidistant_projection)

---

# World projection

- Comparison of Tissot's indicatrix of deformation

.pull-left[
####Mercator projection

<center><img src="fig/gis/mercator_tissot.png" width=70%></img></center>
]
.pull-right[
####Waterman "butterfly"

<center><img src="fig/gis/waterman_tissot.png" width=100%></img></center>
]

**Pictures:** [Wikipedia](https://en.wikipedia.org/wiki/Azimuthal_equidistant_projection)

---

# Local projection

- Use of local projections to reduce distorsions

---

# Local projection

- Use of local projections to reduce distorsions

---

# Layers

- With GIS, maps are made of a superposition of layers

---

# Some basics for map creation

.pull-left[
 
<center><img src="fig/gis/final_map.png" width=100%></img></center>
]
.pull-right[
- Provide some context (municipalities borders, waterways...)
 - *You may use a small map to localize your study area at bigger scale*

- Maps need legend, scale bar and north arrow

- Add title, ideally with a comment

- Do not forget to quote data source
]

---

# Not (only) for pretty maps...

- Several geoprocessing operations may be performed for a single layer or between layers

---

background-color: #e76f51
background-size: cover
class: center, bottom, inverse

# **Layers in GIS**

---

# Two main type of layers

.pull-left[
 
 #### Raster
 - Regular grid of cells or pixels
 - One or multiple values per cell
 
<center><img src="fig/gis/raster.PNG" width=80%></img></center>

]

---

# Examples of raster data
- **Aerial photography:** Pixel values correspond to ground color

<center>Comparison of Lempdes (63) between today and 1950</center>
<center>Source: <a href="https://remonterletemps.ign.fr/"> remonterletemps.ign.fr </a></center>

---

# Examples of raster data

- **Remote sensing:** Multispectral images

.pull-left[
- Three outcomes for incoming solar radiation: absorption, reflection or transmission

- For each wavelength, reflectance (R) is calculated as follows: 
`$$R = \frac{\phi_r}{\phi_i}$$`
]

.pull-right[
<img src="fig/gis/rad.png" width=100%></img>
<center>Dynamics of the luminous flux incoming on a leaf</center>
]

---

# Examples of raster data

- **Remote sensing:** Multispectral images

.pull-left[
- Several satellites provide reflectance data at 10 to 30m resolution, such as [Landsat](https://landsat.gsfc.nasa.gov/) or [Sentinel-2](https://sentinel.esa.int/web/sentinel/missions/sentinel-2)

- Each raster may be processed alone or combined as a vegetation index like the Normalized Difference Vegetation Index (NDVI)
`$$NDVI = \frac{R_{NIR}-R_{Red}}{R_{NIR}+R_{Red}}$$`
]

.pull-right[
<img src="fig/gis/NDVI.gif" width=100%></img>
<center>Timelapse of NDVI evolution</center>
]

---

# Examples of raster data

- **Digital Elevation Model (DEM):** Elevation data

.pull-left[
- A DEM is a raster of ground surface topography

- *Each cell gives the average elevation at a given position*

]

.pull-right[
<img src="fig/gis/elevation.png" width=100%></img>
]

---

# Examples of raster data

- **Digital Elevation Model (DEM):** Elevation data

.pull-left[
- A DEM is a raster of ground surface topography

- GIS can convert these data into raster of slope, exposure...

- ***On the right:*** *Each cell gives the main exposure at a given position*

]

.pull-right[
<img src="fig/gis/aspect.png" width=100%></img>
]

---

# Formats of raster layers

- Raster geographic files are like **image files**

- Thus the format of a raster file is often that of an image file (.tif, .ecw, .jpg, .png...)

- But, in addition to the image data, the rasters have **positioning information**, which can be stored in a separate file or in the same file in a GeoTIFF format (.tif or .tiff)

---

# Two main type of layers
.pull-left[
 
 #### Raster
 - Regular grid of cells or pixels
 - One or multiple values per cell
 
<center><img src="fig/gis/raster.PNG" width=80%></img></center>

]

.pull-right[

#### Vector
- Set of geometric shapes (points, lines or polygons)
- With an attribute table describing each item

]

---

# Examples of vector data

- With vectors, the elementary spatial entities are not cells but geometric shapes: **points**, **lines** or **polygons**

---

# Examples of vector data

- Variables of different types (numerical, factorial, text...) can be stored in **the attribute table** to define each item

- *One table per vector file, with one line per item*

---

# Examples of vector data

- Vectors are useful to make **choropleth map**, in which spatial units are colored or shaded to indicate the mean values of an indicator

.pull-left[
- First choropleth map: [Popular education in France](https://gallica.bnf.fr/ark:/12148/btv1b530830640) (Dupin, 1826)

- ***Legend:*** *From 1 pupil per 10 inhabitants to 1 pupil per 268 inhabitants from lightest to darkest*
]

.pull-right[
<img src="fig/gis/chloro.png" width=100%></img>
]
---

# Examples of vector data

- Choropleth maps are well suited to show statistical values by administrative entity
<img src="fig/gis/cattle.png" width=100%></img>

---

# Examples of vector data

- The principal of choropleth maps may also be applied to point vectors
  - *To produce a yield map, the attribute table provides estimated weight, moisture, working width...*

---

# Examples of vector data

- With **a cartogram**, it is the size of the entities that is proportional to the mapped indicator

---

# Formats of vector layers

- There are several formats for vector files
  - *GeoJSON (.geojson, .json), GeoPackage (.gpkg), .kml and .kmz...*

- The **shapefile** format is a popular vector data format developed and regulated by Esri

- Each shapefile is composed of several files
  - *.shp, .shx, .dbf (table of attribute), .prj (optional file with CRS)...*
  
  - *All files must be stored in the same folder, with the same name*

---

background-color: #e76f51
background-size: cover
class: center, bottom, inverse

# **Some examples of GIS software**

---

# Specialized softwares

.pull-left[
#### Free, open source
 
<center><img src="fig/gis/qgis.png" width=90%></img></center>
]

.pull-right[
#### Commercial
<center><img src="fig/gis/arcgis.png" width=70%></img></center>
]

---

# Non-specialized softwares

---

# Cloud-based softwares

---

background-color: #e76f51
background-size: cover
class: center, bottom, inverse

# **An example of GIS application**

### Adoption of cover crops in France

---

# Goal of the study

- Use of multispectral imagery to detect cover crops

Nowak *et al.* (2022)

---

# Methodology

- What is the best time to detect cover crops?

---

# Methodology

- Border of the fields: Land Parcel Identification System of the Common Agricultural Policy
  - *Filter for spring crops only*

<center><img src="fig/gis/rpg.PNG" width=80%></img></center>
[**Picture:** Geoportail](https://www.geoportail.gouv.fr/)

---

# Methodology

- NDVI extraction based on [Sentinel-2](https://sentinel.esa.int/web/sentinel/missions/sentinel-2) images with the [Google Earh Engine](https://earthengine.google.com/)

<center><img src="fig/gis/data_extraction.PNG" width=100%></img></center>
[Nowak *et al.* (2021)](https://iopscience.iop.org/article/10.1088/1748-9326/ac007c#:~:text=At%20country%20scale%2C%20winter%20soil,at%20least%2050%25%20of%20vegetation.)

---

# Results

- Map of the adoption of winter cover crops in France
  - *~ 40% of fields with cover crops before spring crops*

<center><img src="fig/gis/cover_map.png" width=60%></img></center>
[Nowak *et al.* (2022)](https://iopscience.iop.org/article/10.1088/1748-9326/ac007c#:~:text=At%20country%20scale%2C%20winter%20soil,at%20least%2050%25%20of%20vegetation.)

---

# Results

- Analysis of adoption by crop
 
<center><img src="fig/gis/species.PNG" width=60%></img></center>
[Nowak *et al.* (2022)](https://iopscience.iop.org/article/10.1088/1748-9326/ac007c#:~:text=At%20country%20scale%2C%20winter%20soil,at%20least%2050%25%20of%20vegetation.)

---

# Data sources

- Population of the Puy-de-Dôme (France) department: [Global Human Settlement Layer](https://ghsl.jrc.ec.europa.eu/), European Comission

- Infrastructures: [BD TOPO](https://geoservices.ign.fr/documentation/donnees/vecteur/bdtopo), Institut national de l'information géographique et forestière (IGN)

- Map visualization, with lot of different ressources available for France; [Géoportail](https://www.geoportail.gouv.fr/)

---

# Useful resources

- [Learn how to make cartogram with R](https://www.r-graph-gallery.com/cartogram.html), The R Graph Gallery

- [Making maps with R](https://bjnnowak.github.io/Lessons/2nd_session_R#1), Benjamin Nowak

- [Handling Digital Elevation Model with Google Earth Engine](https://bjnnowak.netlify.app/2021/09/27/gee-terrain-features/), Benjamin Nowak

---

background-color: #264653
background-size: cover
class: center, bottom, inverse

### [Back to Summary](https://bjnnowak.github.io/gis/)