This lesson is being piloted (Beta version)

Intro to Raster Data in R

Overview

Teaching: 20 min
Exercises: 10 min
Questions
  • What is a raster dataset?

  • How do I work with and plot raster data in R?

  • How can I handle missing or bad data values for a raster?

Objectives
  • Describe the fundamental attributes of a raster dataset.

  • Explore raster attributes and metadata using R.

  • Import rasters into R using the raster package.

  • Plot a raster file in R using the ggplot2 package.

  • Describe the difference between single- and multi-band rasters.

Things You’ll Need To Complete This Episode

See the lesson homepage for detailed information about the software, data, and other prerequisites you will need to work through the examples in this episode.

In this episode, we will introduce the fundamental principles, packages and metadata/raster attributes that are needed to work with raster data in R. We will discuss some of the core metadata elements that we need to understand to work with rasters in R, including CRS and resolution. We will also explore missing and bad data values as stored in a raster and how R handles these elements.

We will continue to work with the dplyr and ggplot2 packages that were introduced in the Introduction to R for Geospatial Data lesson. We will use two additional packages in this episode to work with raster data - the raster and rgdal packages. Make sure that you have these packages loaded.

library(raster)
library(rgdal)

Introduce the Data

If not already discussed, introduce the datasets that will be used in this lesson. A brief introduction to the datasets can be found on the Geospatial workshop homepage.

For more detailed information about the datasets, check out the Geospatial workshop data page.

View Raster File Attributes

We will be working with a series of GeoTIFF files in this lesson. The GeoTIFF format contains a set of embedded tags with metadata about the raster data. We can use the function GDALinfo() to get information about our raster data before we read that data into R. It is ideal to do this before importing your data.

GDALinfo("data/NEON-DS-Airborne-Remote-Sensing/HARV/DSM/HARV_dsmCrop.tif")
rows        1367 
columns     1697 
bands       1 
lower left origin.x        731453 
lower left origin.y        4712471 
res.x       1 
res.y       1 
ysign       -1 
oblique.x   0 
oblique.y   0 
driver      GTiff 
projection  +proj=utm +zone=18 +datum=WGS84 +units=m +no_defs 
file        data/NEON-DS-Airborne-Remote-Sensing/HARV/DSM/HARV_dsmCrop.tif 
apparent band summary:
   GDType hasNoDataValue NoDataValue blockSize1 blockSize2
1 Float64           TRUE       -9999          1       1697
apparent band statistics:
    Bmin   Bmax    Bmean      Bsd
1 305.07 416.07 359.8531 17.83169
Metadata:
AREA_OR_POINT=Area 

If you wish to store this information in R, you can do the following:

HARV_dsmCrop_info <- capture.output(
  GDALinfo("data/NEON-DS-Airborne-Remote-Sensing/HARV/DSM/HARV_dsmCrop.tif")
)

Each line of text that was printed to the console is now stored as an element of the character vector HARV_dsmCrop_info. We will be exploring this data throughout this episode. By the end of this episode, you will be able to explain and understand the output above.

Open a Raster in R

Now that we’ve previewed the metadata for our GeoTIFF, let’s import this raster dataset into R and explore its metadata more closely. We can use the raster() function to open a raster in R.

Data Tip - Object names

To improve code readability, file and object names should be used that make it clear what is in the file. The data for this episode were collected from Harvard Forest so we’ll use a naming convention of datatype_HARV.

First we will load our raster file into R and view the data structure.

DSM_HARV <- 
  raster("data/NEON-DS-Airborne-Remote-Sensing/HARV/DSM/HARV_dsmCrop.tif")

DSM_HARV
class       : RasterLayer 
dimensions  : 1367, 1697, 2319799  (nrow, ncol, ncell)
resolution  : 1, 1  (x, y)
extent      : 731453, 733150, 4712471, 4713838  (xmin, xmax, ymin, ymax)
coord. ref. : +proj=utm +zone=18 +datum=WGS84 +units=m +no_defs +ellps=WGS84 +towgs84=0,0,0 
data source : /home/travis/build/datacarpentry/r-raster-vector-geospatial/_episodes_rmd/data/NEON-DS-Airborne-Remote-Sensing/HARV/DSM/HARV_dsmCrop.tif 
names       : HARV_dsmCrop 
values      : 305.07, 416.07  (min, max)

The information above includes a report of min and max values, but no other data range statistics. Similar to other R data structures like vectors and data frame columns, descriptive statistics for raster data can be retrieved like

summary(DSM_HARV)
Warning in .local(object, ...): summary is an estimate based on a sample of 1e+05 cells (4.31% of all cells)
        HARV_dsmCrop
Min.        305.6900
1st Qu.     345.5600
Median      359.7800
3rd Qu.     374.4325
Max.        414.1200
NA's          0.0000

but note the warning - unless you force R to calculate these statistics using every cell in the raster, it will take a random sample of 100,000 cells and calculate from that instead. To force calculation on more, or even all values, you can use the parameter maxsamp:

summary(DSM_HARV, maxsamp = ncell(DSM_HARV))
        HARV_dsmCrop
Min.          305.07
1st Qu.       345.59
Median        359.67
3rd Qu.       374.28
Max.          416.07
NA's            0.00

You may not see major differences in summary stats as maxsamp increases, except with very large rasters.

To visualise this data in R using ggplot2, we need to convert it to a dataframe. We learned about dataframes in an earlier lesson. The raster package has an built-in function for conversion to a plotable dataframe.

DSM_HARV_df <- as.data.frame(DSM_HARV, xy = TRUE)

Now when we view the structure of our data, we will see a standard dataframe format.

str(DSM_HARV_df)
'data.frame':	2319799 obs. of  3 variables:
 $ x           : num  731454 731454 731456 731456 731458 ...
 $ y           : num  4713838 4713838 4713838 4713838 4713838 ...
 $ HARV_dsmCrop: num  409 408 407 407 409 ...

We can use ggplot() to plot this data. We will set the color scale to scale_fill_viridis_c which is a color-blindness friendly color scale. We will also use the coord_quickmap() function to use an approximate Mercator projection for our plots. This approximation is suitable for small areas that are not too close to the poles. Other coordinate systems are available in ggplot2 if needed, you can learn about them at their help page ?coord_map.

ggplot() +
    geom_raster(data = DSM_HARV_df , aes(x = x, y = y, fill = HARV_dsmCrop)) +
    scale_fill_viridis_c() +
    coord_quickmap()

Raster plot with ggplot2 using the viridis color scale

Data Tip

More information about the Viridis palette used above at R Viridis package documentation.

This map shows the elevation of our study site in Harvard Forest. From the legend, we can see that the maximum elevation is ~400, but we can’t tell whether this is 400 feet or 400 meters because the legend doesn’t show us the units. We can look at the metadata of our object to see what the units are. Much of the metadata that we’re interested in is part of the CRS. We introduced the concept of a CRS in an earlier lesson.

Now we will see how features of the CRS appear in our data file and what meanings they have.

View Raster Coordinate Reference System (CRS) in R

We can view the CRS string associated with our R object using thecrs() function.

crs(DSM_HARV)
CRS arguments:
 +proj=utm +zone=18 +datum=WGS84 +units=m +no_defs +ellps=WGS84
+towgs84=0,0,0 

Challenge

What units are our data in?

Answers

+units=m tells us that our data is in meters.

Understanding CRS in Proj4 Format

The CRS for our data is given to us by R in proj4 format. Let’s break down the pieces of proj4 string. The string contains all of the individual CRS elements that R or another GIS might need. Each element is specified with a + sign, similar to how a .csv file is delimited or broken up by a ,. After each + we see the CRS element being defined. For example projection (proj=) and datum (datum=).

UTM Proj4 String

Our projection string for DSM_HARV specifies the UTM projection as follows:

+proj=utm +zone=18 +datum=WGS84 +units=m +no_defs +ellps=WGS84 +towgs84=0,0,0

Note that the zone is unique to the UTM projection. Not all CRSs will have a zone. Image source: Chrismurf at English Wikipedia, via Wikimedia Commons (CC-BY).

The UTM zones across the continental United States. From: https://upload.wikimedia.org/wikipedia/commons/8/8d/Utm-zones-USA.svg

Calculate Raster Min and Max Values

It is useful to know the minimum or maximum values of a raster dataset. In this case, given we are working with elevation data, these values represent the min/max elevation range at our site.

Raster statistics are often calculated and embedded in a GeoTIFF for us. We can view these values:

minValue(DSM_HARV)
[1] 305.07
maxValue(DSM_HARV)
[1] 416.07

Data Tip - Set min and max values

If the minimum and maximum values haven’t already been calculated, we can calculate them using the setMinMax() function.

DSM_HARV <- setMinMax(DSM_HARV)

We can see that the elevation at our site ranges from 305.0700073m to 416.0699768m.

Raster Bands

The Digital Surface Model object (DSM_HARV) that we’ve been working with is a single band raster. This means that there is only one dataset stored in the raster: surface elevation in meters for one time period.

Multi-band raster image

A raster dataset can contain one or more bands. We can use the raster() function to import one single band from a single or multi-band raster. We can view the number of bands in a raster using the nlayers() function.

nlayers(DSM_HARV)
[1] 1

However, raster data can also be multi-band, meaning that one raster file contains data for more than one variable or time period for each cell. By default the raster() function only imports the first band in a raster regardless of whether it has one or more bands. Jump to a later episode in this series for information on working with multi-band rasters: Work with Multi-band Rasters in R.

Dealing with Missing Data

Raster data often has a NoDataValue associated with it. This is a value assigned to pixels where data is missing or no data were collected.

By default the shape of a raster is always rectangular. So if we have a dataset that has a shape that isn’t rectangular, some pixels at the edge of the raster will have NoDataValues. This often happens when the data were collected by an airplane which only flew over some part of a defined region.

In the image below, the pixels that are black have NoDataValues. The camera did not collect data in these areas.

plot of chunk demonstrate-no-data-black-ggplot

In the next image, the black edges have been assigned NoDataValue. R doesn’t render pixels that contain a specified NoDataValue. R assigns missing data with the NoDataValue as NA.

The difference here shows up as ragged edges on the plot, rather than black spaces where there is no data.

plot of chunk demonstrate-no-data-ggplot

If your raster already has NA values set correctly but you aren’t sure where they are, you can deliberately plot them in a particular colour. This can be useful when checking a dataset’s coverage. For instance, sometimes data can be missing where a sensor could not ‘see’ its target data, and you may wish to locate that missing data and fill it in.

To highlight NA values in ggplot, alter the scale_fill_*() layer to contain a colour instruction for NA values, like scale_fill_viridis_c(na.value = 'deeppink')

plot of chunk napink

The value that is conventionally used to take note of missing data (the NoDataValue value) varies by the raster data type. For floating-point rasters, the figure -3.4e+38 is a common default, and for integers, -9999 is common. Some disciplines have specific conventions that vary from these common values.

In some cases, other NA values may be more appropriate. An NA value should be a) outside the range of valid values, and b) a value that fits the data type in use. For instance, if your data ranges continuously from -20 to 100, 0 is not an acceptable NA value! Or, for categories that number 1-15, 0 might be fine for NA, but using -.000003 will force you to save the GeoTIFF on disk as a floating point raster, resulting in a bigger file.

If we are lucky, our GeoTIFF file has a tag that tells us what is the NoDataValue. If we are less lucky, we can find that information in the raster’s metadata. If a NoDataValue was stored in the GeoTIFF tag, when R opens up the raster, it will assign each instance of the value to NA. Values of NA will be ignored by R as demonstrated above.

Challenge

Use the output from the GDALinfo() function to find out what NoDataValue is used for our DSM_HARV dataset.

Answers

GDALinfo("data/NEON-DS-Airborne-Remote-Sensing/HARV/DSM/HARV_dsmCrop.tif")
rows        1367 
columns     1697 
bands       1 
lower left origin.x        731453 
lower left origin.y        4712471 
res.x       1 
res.y       1 
ysign       -1 
oblique.x   0 
oblique.y   0 
driver      GTiff 
projection  +proj=utm +zone=18 +datum=WGS84 +units=m +no_defs 
file        data/NEON-DS-Airborne-Remote-Sensing/HARV/DSM/HARV_dsmCrop.tif 
apparent band summary:
   GDType hasNoDataValue NoDataValue blockSize1 blockSize2
1 Float64           TRUE       -9999          1       1697
apparent band statistics:
    Bmin   Bmax    Bmean      Bsd
1 305.07 416.07 359.8531 17.83169
Metadata:
AREA_OR_POINT=Area 

NoDataValue are encoded as -9999.

Bad Data Values in Rasters

Bad data values are different from NoDataValues. Bad data values are values that fall outside of the applicable range of a dataset.

Examples of Bad Data Values:

Find Bad Data Values

Sometimes a raster’s metadata will tell us the range of expected values for a raster. Values outside of this range are suspect and we need to consider that when we analyze the data. Sometimes, we need to use some common sense and scientific insight as we examine the data - just as we would for field data to identify questionable values.

Plotting data with appropriate highlighting can help reveal patterns in bad values and may suggest a solution. Below, reclassification is used to highlight elevation values over 400m with a contrasting colour.

plot of chunk demo-bad-data-highlighting

Create A Histogram of Raster Values

We can explore the distribution of values contained within our raster using the geom_histogram() function which produces a histogram. Histograms are often useful in identifying outliers and bad data values in our raster data.

ggplot() +
    geom_histogram(data = DSM_HARV_df, aes(HARV_dsmCrop))
`stat_bin()` using `bins = 30`. Pick better value with `binwidth`.

plot of chunk view-raster-histogram

Notice that a warning message is thrown when R creates the histogram.

stat_bin() using bins = 30. Pick better value with binwidth.

This warning is caused by a default setting in geom_histogram enforcing that there are 30 bins for the data. We can define the number of bins we want in the histogram by using the bins value in the geom_histogram() function.

ggplot() +
    geom_histogram(data = DSM_HARV_df, aes(HARV_dsmCrop), bins = 40)

plot of chunk view-raster-histogram2

Note that the shape of this histogram looks similar to the previous one that was created using the default of 30 bins. The distribution of elevation values for our Digital Surface Model (DSM) looks reasonable. It is likely there are no bad data values in this particular raster.

Challenge: Explore Raster Metadata

Use GDALinfo() to determine the following about the NEON-DS-Airborne-Remote-Sensing/HARV/DSM/HARV_DSMhill.tif file:

  1. Does this file have the same CRS as DSM_HARV?
  2. What is the NoDataValue?
  3. What is resolution of the raster data?
  4. How large would a 5x5 pixel area be on the Earth’s surface?
  5. Is the file a multi- or single-band raster?

Notice: this file is a hillshade. We will learn about hillshades in the Working with Multi-band Rasters in R episode.

Answers

GDALinfo("data/NEON-DS-Airborne-Remote-Sensing/HARV/DSM/HARV_DSMhill.tif")
rows        1367 
columns     1697 
bands       1 
lower left origin.x        731453 
lower left origin.y        4712471 
res.x       1 
res.y       1 
ysign       -1 
oblique.x   0 
oblique.y   0 
driver      GTiff 
projection  +proj=utm +zone=18 +datum=WGS84 +units=m +no_defs 
file        data/NEON-DS-Airborne-Remote-Sensing/HARV/DSM/HARV_DSMhill.tif 
apparent band summary:
   GDType hasNoDataValue NoDataValue blockSize1 blockSize2
1 Float64           TRUE       -9999          1       1697
apparent band statistics:
        Bmin      Bmax     Bmean       Bsd
1 -0.7136298 0.9999997 0.3125525 0.4812939
Metadata:
AREA_OR_POINT=Area 
  1. If this file has the same CRS as DSM_HARV? Yes: UTM Zone 18, WGS84, meters.
  2. What format NoDataValues take? -9999
  3. The resolution of the raster data? 1x1
  4. How large a 5x5 pixel area would be? 5mx5m How? We are given resolution of 1x1 and units in meters, therefore resolution of 5x5 means 5x5m.
  5. Is the file a multi- or single-band raster? Single.

More Resources

Key Points

  • The GeoTIFF file format includes metadata about the raster data.

  • To plot raster data with the ggplot2 package, we need to convert it to a dataframe.

  • R stores CRS information in the Proj4 format.

  • Be careful when dealing with missing or bad data values.