Mulitspectra Analysis of Structural Lineaments in Cork Basin

My GIS project will be analysing the structural lineaments that can be seen from using Raster data sets from the Landsat-8 Satellite. The data was downloaded from the EarthViewer program, which contained all 11 bands of light that the satellite is capable of observing. The structural lineament delineation selection criteria for MS imagery are as follows: day-light scenes with a cloud cover percentage of 10 % or lower, and a horizontal spatial Root Mean Square Error(RMSE) < 12m. RMSE is a measure of the error in horizontal position between features in the image and their true actual ground location. Each MS scene which meets the structural lineament delineation selection criteria must be atmospherically corrected from Top of Atmosphere (ToA) to Bottom of Atmosphere (BoA) reflectance (Young, et al., 2017). The datasets used in this project already met these requirements as it has been pre-processed.

The data accessed was pre-processed "On-Demand Level 2" Landsat-8 imagery. From doing background research, I discovered that the calculation of band 5/1 was regarded as the most suitable band for doing this so this is what I chose to use (Mwaniki, et al., 2015). Using horizontal convolution, in the Raster Analysis toolbox of ArcGIS Pro, it was possible to further define the structural lineaments that are present in the Cork Basin. Using this tool makes the lineaments much more visible as it detects straight lines and highlights them in a new raster layer. This can be seen to be an effective method of analysis when it is compared to the GSI map of structural lineaments that was created for Ireland. I find this to be an interesting analysis because it is possible to observe new structural lineaments that have not been previously mapped before from a primary source. The convolution tool is used to designate new values to edges that are seen from the raster data. Once these processes have been completed, comparing them to the GSI map can be used to see how well they are portrayed in a remote sensing environment, because the GSI data is primarily sourced data, taken from ground measurements.

It is possible to map the lineations that were delineated using horizontal convolution against the already existing GSI map to see how they compare. As is visible in the map below, many of the lineaments created using raster analysis match up with the already mapped lineaments from the GSI. This gives us confirmation that this is an effective technique of analysis. Many of the large scale features that were identified match up perfectly with the GSI map. Because it is quite difficult to see the lineations initially, I performed an analysis technique which sharpens the image along horizontal axes to make the structural features more visible. The use of different colours also helps to distinguish this. There is potential for errors even after these techniques are applied but it makes them much more accurate. Straight edges such as roads etc. can be disregarded when we compare it to the basemap. The generated lines match up very well in some areas, but there is also some outliers. More extensive analysis and higher resolution imagery could be used to disregard these and match them up with the GSI map.

Structural lineations in bedrock, such as fault lines are often associated with hydrogeological flow directions. To see if my created lineations match up with these, I downloaded the hydrogeological data from the GSI. I then overlayed these to see the relationship between flow direction towards aquifers and these faults. With this analysis, it is then possible to distinguish the different types of lineations present, because all are not relative to hydrological processes. Water tends to choose fault features to flow through because of their high permeability, this lets the water flow more freely than consolidated bedrock.

SRTM data was obtained from the data.gov.ie website. SRTM data is collected at 3 arc second intervals in latitude & longitude (about every 90m). It shows detailed topographic representations of the area surveyed. I then converted this to an aspect-slope data set to make it easier to view. This is another method that can be used to view structural features using ArcGIS. The features in this map resemble the features created using the 5/1 Landsat-8 bands, further confirming that the method of observation was accurate. Overlaying the lineations with the topographical data set shows a good correlation of fault lines and topographic data.

The aquifer positions are very relative to the lineations because they often fall in areas surrounded by faults. The largest aquifer in the image above is positioned directly in the basin between all these faults. The water flow would run down the faults into the aquifer constantly recharging it.

Scale can be an issue when analysing raster data sets because of the resolution they are in. The landsat-8 satellite uses data at a 30m resolution, if it were possible, ideally a smaller resolution would be used because it would show more features at a more accurate scale. Scale is also a factor even when observing the lineations because if they were viewed at a smaller, more local scale they would be harder to see on the map. They are large features on a county-wide scale so if we were to analyse these at a smaller scale they would be difficult to distinguish. The issue of scale here is helped by resolution, because of the 30m resolution of landsat-8, these issues occur when we zoom into the map more because it becomes distorted and difficult to see any large scale features.

To conclude, this method of analysis of structural features is a viable option for observing but would require higher resolution data sets, such as Lidar, to allow us to have complete confidence in it. This is nonetheless a very useful way to observe structural features and water flow in an area from a remote setting as it does not require any field data to accurately calculate the flow directions and aquifer locations in the area. This type of observation would be extremely useful in developing countries where structural mapping may not have been extensively carried out, to determine good water sources. Once the lithology of the bedrock is confirmed, this analysis can then be used to calculate flow directions and aquifer locations. It could even then be related to sedimentary pollution, as often, structural lineations contain exposed minerals and heavy metals that can enter the groundwater flow through dissolution and become contaminants to larger scale, above ground water features.

References:

Mwaniki, M. W., Moeller, M. S. & Schellmann, G., 2015. A comparison of Landsat 8 (OLI) and Landsat 7 (ETM+) in mapping geology and visualising lineaments: A case study of central region Kenya. Berlin, The International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences.

Young, N., Anderson, R., Chignell, S., Vorster, A., Lawrence, R. and Evangelista, P. (2017). A survival guide to Landsat preprocessing. Ecology, 98(4), pp.920-932.

USGS, 2019. USGS. [Online] Available at: https://earthexplorer.usgs.gov/

GSI, 2019. GSI. [Online] Available at http://gsi.ie/

data.gov.ie, 2019. Irish Government [Online] Available at http://data.gov.ie/