title: "Introduction to Raster Package"

```{r, echo=FALSE, message=FALSE, results='hide', purl=FALSE}

<div class="extraswell">

<button data-toggle="collapse" class="btn btn-link" data-target="#pres">View Presentation </button> [Open presentation in a new tab](presentations/PS_06_raster.html){target="_blank"}

<div id="pres" class="collapse">

<div class="embed-responsive embed-responsive-16by9">

<iframe class="embed-responsive-item" src="presentations/PS_06_raster.html" allowfullscreen></iframe>

_Click on presentation and then use the space bar to advance to the next slide or escape key to show an overview._

</div>

</div>

</div>

| [<i class="fas fa-code fa-2x" aria-hidden="true"></i><br> R Script](`r output_nocomment`) | [<i class="fa fa-file-code-o fa-2x"></i> <br> Commented R Script](`r output`) | [<i class="far fa-file-alt fa-2x"></i> <br> Rmd Script](`r fullinput`)|

|:--:|:-:|:-:|

```{r message=F,warning=FALSE, results='hide'}

Raster package includes access to some useful (vector and raster) datasets with `getData()`:

* Elevation (SRTM 90m resolution raster)

* World Climate (Tmin, Tmax, Precip, BioClim rasters)

* Countries from CIA factsheet (vector!)

* Global Administrative boundaries (vector!)

`getData()` steps for GADM:

1. _Select Dataset_: ‘GADM’ returns the global administrative boundaries.

2. _Select country_: Country name of the boundaries using its ISO A3 country code

3. _Specify level_: Level of of administrative subdivision (0=country, 1=first level subdivision).

Administrative areas in this database are countries and lower level subdivisions.

<img src="05_assets/gadm25.png" alt="alt text" width="70%">

Divided by country (see website for full dataset). Explore country list:

Download data for South Africa

```{r, eval=F}

Or use the version in the DataScienceData

```{r, eval=F}

Danger: `plot()` works, but can be slow for complex polygons. If you want to speed it up, you can plot a simplified version as follows:

Plot a subsetted region:

<div class="well">

Use the method above to download and plot the boundaries for a country of your choice.

<button data-toggle="collapse" class="btn btn-primary btn-sm round" data-target="#demo1">Show Solution</button>

<div id="demo1" class="collapse">

```{r, purl=F}

</div>

</div>

Spatial data structure dividing region ('grid') into rectangles (’cells’ or ’pixels’) storing one or more values each.

<small> Some examples from the [Raster vignette](http://cran.r-project.org/web/packages/raster/vignettes/Raster.pdf) by Robert J. Hijmans. </small>

* `rasterLayer`: 1 band

* `rasterStack`: Multiple Bands

* `rasterBrick`: Multiple Bands of _same_ thing.

```{r, echo=TRUE}

```{r, echo=T, results=T}

But it is an empty raster

Use `values()` function:

<div class="well">

Create and then plot a new raster with:

1. 100 rows

2. 50 columns

3. Fill it with random values (`rnorm()`)

<button data-toggle="collapse" class="btn btn-primary btn-sm round" data-target="#demo2">Show Solution</button>

<div id="demo2" class="collapse">

```{r, purl=F}

</div>

</div>

Raster memory usage

Plotting is easy (but slow) with `plot`.

rasterVis package has `gplot()` for plotting raster data in the `ggplot()` framework.

Adjust `maxpixels` for faster plotting of large datasets.

Can use all the `ggplot` color ramps, etc.

Raster package uses standard [coordinate reference system (CRS)](http://www.spatialreference.org).

For example, see the projection format for the [_standard_ WGS84](http://www.spatialreference.org/ref/epsg/4326/).

Use `projectRaster()` to _warp_ to a different projection.

`method=` `ngb` (for categorical) or `bilinear` (continuous)

Mean monthly climate and derived variables interpolated from weather stations on a 30 arc-second (~1km) grid.

See [worldclim.org](http://www.worldclim.org/methods)

<small>

Variable Description

- -

BIO1 Annual Mean Temperature

BIO2 Mean Diurnal Range (Mean of monthly (max temp – min temp))

BIO3 Isothermality (BIO2/BIO7) (* 100)

BIO4 Temperature Seasonality (standard deviation *100)

BIO5 Max Temperature of Warmest Month

BIO6 Min Temperature of Coldest Month

BIO7 Temperature Annual Range (BIO5-BIO6)

BIO8 Mean Temperature of Wettest Quarter

BIO9 Mean Temperature of Driest Quarter

BIO10 Mean Temperature of Warmest Quarter

BIO11 Mean Temperature of Coldest Quarter

BIO12 Annual Precipitation

BIO13 Precipitation of Wettest Month

BIO14 Precipitation of Driest Month

BIO15 Precipitation Seasonality (Coefficient of Variation)

BIO16 Precipitation of Wettest Quarter

BIO17 Precipitation of Driest Quarter

BIO18 Precipitation of Warmest Quarter

BIO19 Precipitation of Coldest Quarter

</small>

Download the data:

```{r, eval=F}

`res` is resolution (0.5, 2.5, 5, and 10 minutes of a degree)

Instead of downloading the full dataset, we'll use the copy in the `DataScienceData` package:

Note the min/max of the raster. What are the units? Always check metadata, the [WorldClim temperature dataset](http://www.worldclim.org/formats) has a `gain` of 0.1, meaning that it must be multipled by 0.1 to convert back to degrees Celsius. We'll set the temperature variables (see table above) to 0.1 and leave the others at 1.

Or use `rasterVis` methods with gplot

Let's dig a little deeper into the data object:

Use `[[1:3]]` to select raster layers from raster stack.

<div class="well">

Create a new raster by aggregating to the minimum (`min`) value of `r1` within a 10 pixel window

<button data-toggle="collapse" class="btn btn-primary btn-sm round" data-target="#demo3">Show Solution</button>

<div id="demo3" class="collapse">

```{r, purl=F}

</div>

</div>

<div class="well">

Plot the focal standard deviation of `r1` over a 3x3 window.

<button data-toggle="collapse" class="btn btn-primary btn-sm round" data-target="#demo4">Show Solution</button>

<div id="demo4" class="collapse">

```{r, purl=F}

</div>

</div>

the `raster` package has many options for _raster algebra_, including `+`, `-`, `*`, `/`, logical operators such as `>`, `>=`, `<`, `==`, `!` and functions such as `abs`, `round`, `ceiling`, `floor`, `trunc`, `sqrt`, `log`, `log10`, `exp`, `cos`, `sin`, `max`, `min`, `range`, `prod`, `sum`, `any`, `all`.

So, for example, you can

* points

* lines

* polygons

* extent (rectangle)

* cell numbers

Extract all intersecting values OR apply a summarizing function with `fun`.

`sampleRandom()` generates random points and automatically extracts the raster values for those points. Also check out `?sampleStratified` and `sampleRegular()`.

Generate 100 random points and the associated climate variables at those points.

```{r}

Often you will have some locations (points) for which you want data from a raster* object. You can use the `extract` function here with the `pts` object (we'll pretend it's a new point dataset for which you want climate variables).

Use `gather()` to reshape the climate data for easy plotting with ggplot.

And plot density plots (like histograms).

Extract values along a transect.

Calculate mean annual temperature averaged by province (polygons).

1. Download the Maximum Temperature dataset using `getData()`

2. Set the gain to 0.1 (to convert to degrees Celcius)

2. Crop it to the country you downloaded (or ZA?)

2. Calculate the overall range for each variable with `cellStats()`

3. Calculate the focal median with an 11x11 window with `focal()`

4. Create a transect across the region and extract the temperature data.

Default layer names can be problematic/undesirable.

See `?sprintf` for details

Report quantiles for each layer in a raster* object

Reformat to 'long' format.

Or the same data in a levelplot:

Things to consider:

* RAM limitations

* Disk space and temporary files

* Use of external programs (e.g. GDAL)

* Use of external GIS viewer (e.g. QGIS)