Introducing: Nearmap for Post-catastrophe Response
Post-catastrophe imagery and AI-derived property damage and condition data unite to help insurers process customer claims more efficiently.
Post-catastrophe imagery and AI-derived property damage and condition data unite to help insurers process customer claims more efficiently.
We’re commonly asked at Nearmap about our digital elevation models (DEM), and how they can be used to understand and model runoff and surface water flow. This is a hydrology task, and it is useful knowledge for urban and regional planning. Understanding runoff around sensitive areas close to waterways and catchment areas is necessary if we want to manage the health of our water systems, as well as our overall ecosystem.
To demonstrate how terrain models can inform runoff modeling, let’s take a look at importing a DEM from Nearmap into ArcGIS Pro for use with Spatial Analyst and the hydrology toolset.
Terrain models can be exported from MapBrowser by anyone with a Nearmap 3D Export subscription. The exported file contains a raster dataset where each pixel represents a bare earth elevation. Bare earth elevations are modeled to remove any above ground objects including buildings, cars, and trees.
When exporting from MapBrowser, you can select an option for pixel size (1 ft, 2 ft, 3 ft). These options are based on a 0.5 ft source resolution, and as the pixel size increases, you get an averaging of elevations and therefore a relative smoothing of the surface.
When displaying a sample from an agricultural area in ArcGIS and symbolizing with the ‘magma’ theme, we can see a relatively flat area with a hill in the lighter color and the elevation dropping off to darker colors towards the creek line visible in the image. I’ve drawn in some arrows that show the expected slope.
To prepare our dataset, it’s a good idea to ‘fill sinks’ which removes small imperfections introduced by the modeling process. Typically, Interpolation algorithms result in dips between areas of high confidence. The ArcGIS Hydrology Fill tool fills in the dips to create a surface corrected for hydrology use. In theory, water should flow cleanly down a slope and this will reduce noise in later calculations.
We can now run the flow direction tool which will calculate a direction of flow for every pixel based on those around it. For this we will use the D8 algorithm.
The D8 algorithm (O'Callaghan and Mark, 1984) is a simple grid-based method that approximates the primary flow direction from each pixel to one of its eight neighbors. Pixels are classified with direction defined by the following values:
If we symbolize with unique colors, the output at scale looks noisy, but there are some clear signs of areas of direction correlation:
But we can improve this even further by converting the output direction values to degrees and displaying as vector arrows. It’s a little clearer when we zoom in:
Our flow direction can now be used as an input to calculate flow accumulation. The flow accumulation tool takes the output produced by the flow direction and adds the number of connected cells so that every downward cell is the sum of all upstream cells that flow into it. Ridgelines would have very few upstream cells connected to them, so they result in a low score, while valleys and concave areas have a high number of connected upstream cells, resulting in high scores.
This method allows us to locate even very subtle areas of flow accumulation that would not otherwise be visible, and we can symbolize these ‘streams’ to be even more visible by running ‘Focal Statistics’ on the output and using a 3x3 cell that takes the maximum value. This in effect widens the number of pixels with high values.
We can view these streams as white pixels (below), and you can see the streams detected in the image underneath. The subtler, smaller streams are often confirmed by variations in the vegetation.
What’s interesting about this technique is that we can see streams running along the rows of this agricultural field as we might expect, but there are visible areas where water accumulates and can be seen running off across the rows towards other larger streams. This sort of runoff is less intuitive and may not be identified without detailed analysis.
Identifying areas of high flow accumulation gives us some indication about potential runoff and therefore areas that could be suitable for protection or mitigation strategies.
In the image above, we see where these smaller streams aggregate into natural waterways. If needed, water quality could be improved in these areas by implementing runoff management practices, such as planting trees and creating a buffer zone that can filter runoff and slow down the speed of water.
Another technique would be to alter the drainage characteristics of the land surface by implementing contours that divert runoff away from faster-flowing areas. This slows the flow of water, which allows for increased infiltration.
This is just one way of looking at how continually updated high resolution 3D data can better inform engineering planning and design strategies.
For more examples of how to use Nearmap data in third-party GIS software, please visit our Help Center.
