====== Introduction to geostatistics with R ======

April 2023. Online. Instructor: Guillaume Larocque (glaroc@gmail.com). 

===== Presentation =====

[[https://prezi.com/view/yjD0wcl7coF4wBJpeNTS/|Prezi presentation]]

===== R Code =====
{{ ::qcbs_geostats_workshop_files_2023.zip |Click here to download the zip file containing the scripts and the datasets that will be used during the workshop.}}

Please extract this zip archive to an easily accessible folder on your computer. 

===== Ressources =====

==== Books ====
  * [[https://www.routledge.com/Spatial-Data-Science-With-Applications-in-R/Pebesma-Bivand/p/book/9781138311183|Spatial Data Science - With applications in R.]] - Provides a good overview of spatial statistical tools in R. Contains a good chapter on geostatistics. The online version of the book can be found [[https://r-spatial.org/book/|here]]
  * [[http://www.amazon.ca/Geostatistics-Natural-Resources-Evaluation-Goovaerts/dp/0195115384|Geostatistics for Natural Resources Evaluation, Pierre Goovaerts.]] - Very good book, which provides a more in-depth coverage of geostatistical theory. 
  * [[http://www.lu.lv/materiali/biblioteka/es/pilnieteksti/vide/A%20Practical%20Guide%20to%20Geostatistical%20Mapping%20of%20Environmental%20Variables.pdf|A Practical Guide to Geostatistical Mapping of Environmental Variables (PDF)]]
  * [[http://cg.ensmp.fr/bibliotheque/public/MATHERON_Ouvrage_00167.pdf|Georges Matheron - The theory of regionalized variables and its application (PDF).]] - Original text estbalishing the basic foundations of geostatistics. 
  * {{:gisonr.pdf|Analyses spatiales sous R}} - Document préparé par Nicolas Casajus pour un cours offert à l'UQAR à l'hiver 2013. 
==== Papers of interest ====
  * [[http://www.esajournals.org/doi/abs/10.2307/2937096|Geostatistical Tools for Modeling and Interpreting Ecological Spatial Dependence]]
  * [[http://link.springer.com/article/10.1023/A%3A1012734519752|GIS and geostatistics: Essential partners for spatial analysis]]
  * [[http://www.ecoscience.ulaval.ca/en/paper/spatial-autocorrelation-and-statistical-tests-in-ecology-fait-parti-du-numero-thematique-sur-les-analyses-spatiales-en-ecologie-part-of-the-special-feature-on-spatial-analyses-in-ecology|Spatial autocorrelation and statistical tests in ecology]]
  * [[http://www.esajournals.org/doi/pdf/10.1890/0012-9658(2003)084%5B1045:SCIPCI%5D2.0.CO%3B2|Spatial covariance in plant communities: intergrating ordination, geostatistics and variance testing]]
  * [[http://link.springer.com/article/10.1007/s10651-008-0090-z|Coregionalization analysis with a drift for multi-scale assessment of spatial relationships between ecological variables]]

==== R packages ====
  * [[http://cran.r-project.org/web/packages/gstat/|Gstat package]] - Gstat R package. 
  * [[http://cran.r-project.org/web/packages/geoR/|geoR package geoR]] - Another package for geostatistics in R. 
  * [[https://geodacenter.github.io/rgeoda/|rgeoda]] - R package for spatial data analysis. 
  * [[https://github.com/r-spatial/qgisprocess|qgisprocess]] - Package that allows you to run QGIS scripts within R. 
  * [[https://cran.r-project.org/web/packages/gdalcubes/index.html|gdalcubes]] - R Package for working with spatial-temporal raster data in the form of cubes. Particularly useful to access STAC catalog data in combination with the [[https://cran.r-project.org/web/packages/rstac/index.html|rstac]] package. 

==== Educational software ====

{{::e_z_kriging.zip|EZ-Kriging}} Small Windows software that facilitates understanding of kriging equations. 


==== Courses ====
[[https://www.teluq.ca/site/etudes/offre/cours/TELUQ/SCI%201031/|Visualisation et analyse de données spatiales avec R. ]]. Course offered at Teluq by Elise Filotas. The course material can be found [[https://sci1031.github.io/sci1031/|here.]]

===== How to create a mask in QGIS =====


==== Digitize a polygon defining the boundary of your study site ====
You can digitize the boundary of your study site from a Google Satellite layer by following the following steps: 
  - Open QGIS. 
  - Install and activate the QuickMapServices Plugin=
  - Add a Google Satellite layer by using the QuickMapServices extension (Web/QuickMapServices). You might need to enable additional providers in the Plugin settings
  - Note that the CRS of the canvas is now WGS 84 / Pseudo mercator and units are in meters. 
  - If you have other vector layer that can help you define the limits of your study area, you can add them to the canvas by clicking on the 'add vector layer' icon.
  - Digitize the zone defining your study area. You first need to add a new polygon shapefile layer under Layer>New>New Shapefile layer. Specify the proper CRS depending on the extent of your study area and specify a name for your layer. It is always preferable to work with a reference system that uses meters as units, such as UTM (medium extent), MTM (small extent in Quebec) or Lambert (large extent). Refrain from using a latitude/longitude system. 
  - Activate Edit mote (Layer>Toggle editing) and digitize the polygon(s) by clicking on the icon {{::capture_du_2013-09-25_10_58_40.png|}} and specify '1' as the ID. 

==== Convert the polygon to raster format ====
  - Find the Rasterize function under Raster>Conversion. Specify the polygon that you just created as the input layer, and choose a file name with a .tif extension as the output file. Choose 'Raster resolution in map units per pixel'. You then need to choose a resolution in the reference system's units used for the input polygon. For example, if you specified UTM as the CRS of the polygon, the resolution will be specified in meters. Choose a resolution that will yield a raster with a manageable size, with a maximum of a few hundred lines and columns (up to 2000-3000max). For example, if your study area is 2km by 3 km, you can specify 5 m. as the resolution, which will yield a raster of 400 pixels by 600 pixels. 
  - After you have clicked on OK, this raster will be saved on your computer as a .tif file. All pixels within the study area will have a value of 1, and pixels outside will have a value of 0. 

==== Import the mask in R ====

  - You can now import the mask in R: 

<file>
mask<-read_stars(file.choose())
</file>

and choose the raster .tif file. 

Then, you can do

<file>
mask[mask==0]=NA
</file>

to change all the zero values to NA (nulls).

Your mask is now ready to use!