Satellites and Bedrock - Part I

Looking at the ways in which satellite images from Sentinel could be used to examine underlying bedrock types. In this first of a three part series, we look at ways of accessing bedrock data from GSI.

Geological Survey Ireland is producing wonderful data resources, impossible to resist when it comes to doing some analysis. In this post, welook at bedrock data, what you would see after sweeping away all the soil and loose covering from the ground. Bedrock is stuff like sandstone, limestone, igneous rock….

I was interested to see how overlying vegetation depends on the bedrock; specifically if the colours from satellite images taken at different times of year could be used to take a guess at what lies beneath.

This requires two suites of information, the bedrock data from GSI and the satellite data from Copernicus-Sentinel. The first we’ll download directly and store locally from this resource, the second we’ll access using the R package, sen2r (but we’ll also end up downloading and storing image files locally, but sen2R will help scope them out).

This was originally written as one post, but then it got so long I decided to split it into three parts:

• Part I: getting the bedrock data from GSI
• Part II: getting the satellite data from Sentinel
  
• Part III: putting the first two together

ref <- st_read("data/Shapefiles/Bedrock_Polygons_ITM_2018.shp", quiet = TRUE)
my_proj <- "+proj=utm +zone=29 +datum=WGS84 +units=m +no_defs"
ref <- st_transform(ref, crs = my_proj)

The bedrock file arrives in the form of a shapefile. It contains bunch of attributes: location, decription, area, as well as the polygon shapes themselves. And some attributes I’m at a loss to understand (M, Y, K, C anyone?). We end up with a total of 25,958 fields covering the Republic of Ireland.

Below we show the geometry for this shapefile:

Geometry set for 25958 features 
Geometry type: MULTIPOLYGON
Dimension:     XY
Bounding box:  xmin: 384779.9 ymin: 5696646 xmax: 701597.5 ymax: 6145919
CRS:           +proj=utm +zone=29 +datum=WGS84 +units=m +no_defs
First 5 geometries:

And here is a table showing the attributes for a sample of a few fields. Note, we’ve done a wee text-trawl through the description attribute to try and categorise the rocks in to a manageable number of types (quartz, sandstone, limestone, schist, shale, granite, or other).

class	NEWCODE	DESCRIPT	UNIT_NAME	M	Y	K	C	LABEL	SHAPE_AREA
limestone	CDTOBE	Calcareous shale, limestone conglomerate	Tober Colleen Formation	20	20	0	40	TC	1,943K
schist	OTMAUL3	Dark grey semi-pelitic, psammitic schist	Ballybeg Member	15	30	0	10	MNbb	821K
schist	LCRQ	Psammitic schist, some marble beds	Lower Crana Quartzite Formation	30	40	0	0	LC	41.17
sandstone	DUOHSF	Flaser-bedded sandstone & minor mudstone	Old Head Sandstone Formation	0	50	0	0	OH	1,766K
limestone	CDRATH	Pale-grey massive mud-grade limestone	Rathronan Formation	10	0	0	20	RR	249K
quartz	MRBENQ	Pale quartzites, grits, graphitic top	Bennabeola Quartzite Formation	0	70	0	0	BX	119.39
other	SMBRDF	Fine to conglomeratic graded greywacke	Broadford Formation	10	50	0	30	BF	11K
sandstone	S*GCRA	Thin-bedded sandstone & mudstone	Glencraff Formation	10	30	0	10	GC	58K
other	IOMGAB	Hornblende metagabbros & metanorites	Metagabbro and Related Lithologies	0	40	0	70	Mg	33.45
sandstone	DUCARR	Red, brown conglomerate & sandstone	Carrigmaclea Formation	40	70	0	30	CI	1,364K

Plotting this, and just looking at the NorthWest, we get: