Skip to content

Latest commit

 

History

History
93 lines (68 loc) · 3.47 KB

spatial-data-raster-R.md

File metadata and controls

93 lines (68 loc) · 3.47 KB
Raw
layout element title language
page
notes
Spatial Data Raster
R

Remember to download and put into data subdirectory:

  • [LiDAR rasters and plot locations]({{ site.baseurl }}/data/neon-airborne.zip)

Load the following into browser window:

Introduction to raster data

  • There are two common types of spatial data, raster and vector
  • Raster data stores data that is continous across space
  • Like climate variables, satellite imagery, and elevation
  • It is stored in a gridded format

![Aerial photo of a green landscape. A section of the landscape is expanded to show that it is composed of green pixels. This is expanded to show that underlying the green pixels is a matrix of numerical values. The matrix is then shown as green pixels again to represent plotting the raster.]({{ site.baseurl }}/materials/raster_concept.png)

  • In the grid each pixel contains a value
  • So it is basically a matrix of numbers of one value at each position in the matrix
  • So for elevation we would have a matrix of heights above sea level

Importing and exploring

  • We import raster data using the read_stars() function from the stars package
  • We'll start by importing some elevation data collected from an airplane using an instrument called LIDAR
  • One of the values that LIDAR can generate is a Digital Terrain Model or DTM, which is the elevation of the ground

![Drawing of trees on undulating terrain. A brown line along the top of the terrain indicates the Digital Terrain Model]({{ site.baseurl }}/materials/digital-terrain-model.png)

library(stars)
dtm_harv <- read_stars("data/harv/harv_dtmcrop.tif")
  • Looking at this object provides information on the data it contains
dtm_harv
  • This is "metadata" or "data about "data"
  • It is is important because it provides the context of spatial data this raster matrix so that R knows how to work with it
    • refsys
    • units
    • min, max, mean

Plotting spatial data with ggplot

  • Spatial data can be plotted using either base R or ggplot

  • We'll use ggplot since that's the data visualization tool we're using for this course

  • Useful for making nice maps combined with other figures

  • There is a special geom for plotting stars raster data geom_stars

  • Since it is raster data it doesn't require an aesthetic

library(ggplot2)

ggplot() +
  geom_stars(data = dtm_harv)

  • For spatial data we're going to put the data in the geom calls instead of ggplot() because we are often trying to combine data of different types from different objects into a single map

  • We can change the color ramp by using scale functions

  • This is equivalent to when we used scale to change the axes, but now we're changing the color ramp instead

  • One good color ramp is "viridis"

  • To use this color ramp we add `scale_fill_viridis_c() to our ggplot object

  • We use fill because we are coloring the inside of each raster pixel (like the inside of a bar plot or histogram)

  • The _c at the end indicates that it is a "continuous" scale

ggplot() +
  geom_stars(data = dtm_harv) +
  scale_fill_viridis_c()
  • If we had discrete data, e.g., on soil types, we would use _d instead

Do Task 1 of [Canopy Height from Space]({{ site.baseurl }}/exercises/Neon-canopy-height-from-space-R).