Cost Distance & Viewshed Analysis

Introduction:

This project employs a series of spatial analysis processes in order to help guide a search for a lost hiker in the Willamette National Forest in eastern Lane County. In particular, we implement a ‘cost-surface’ analysis (with respect to topographical relief and land cover) on a digital elevation model of the Willamette Valley and Central Cascades in order to derive a least-cost path – our best guess for where we might locate the lost hiker.

The end goal of this analysis is compare the calculated least-cost path with the product of a view-shed analysis that regards two possible ‘lookout’ locations. This comparison will help us determine the best place to set up a lookout facility and will help guide efforts to search for our friend.

Methods:

[Q1] Initially, we are given a digital elevation model that plots elevation values across the Central Cascades and Southern Willamette Valley in 10 by 10 meter cells. We take note of the resolution of the data model for two reasons: (1) it is a direct indication of how precise our analysis is, since a smaller cell size will result in a more detailed, robust analysis, and (2)  so that we know the resolution that we should look for in other datasets if we wish to perform some calculation between this model and another, and so we know when aggregation will take place.

For this layer, the cell size is rather easy to determine because it is documented in the raster’s file name. In a situation where we aren’t so lucky to have the file labeled with its cell resolution, we can find the cell size in the Layer Properties tab in ArcMap, where it lists the cell size as ~ 32.8 units.

We use the digital elevation model to derive a raster dataset displaying slope values across our approximated search area, and add this raster to another dataset marking nearby bodies of water. Before we performed the add calculation on the rasters, we first need to reclassify their values.

[Q2] We have no attribute in our water dataset that will allow for the production of a clean cost surface (before we reclassified the water, raster values were populated with the feature ID, which holds no significance in the environment we’re analyzing). In order to give this raster layer significance, we assign meaningful value to the cells in this new raster. In our case, we are assuming that our friend will probably not be swimming across lakes and rivers on her hike, so we assign a relatively large value to all cells of land covered by water – a value of 100 is assigned to any areas covered by water to contrast with the reclassified slope dataset that now spans integer values 1 – 9. The rest of this particular raster surface will have values of 0 so that the re-assignment will not interfere with the reclassified slope values.

This reclassification (when added to our slope dataset) will result in a raster with no null values and will allow a cost surface to be calculated leaving the bodies of water with disproportionately large values of cost in the hopes that a cost path will be generated that avoids all bodies of water (if possible).

Accumulated Cost Surface: Basemap

Cartographer: Julie Stringham Data Sources: University of Oregon GIS Server, ESRI (Basemap) Projection: NAD 83, Lambert Conformal Conic

[Q4] The resulting cost-surface has an attribute table that is populated with values that denote the accumulated cost distance at each cell with respect to the hiker’s origin. These values span from 0 (at the origin) t0 about 235,000 at the far reaches of the search area and are the result of keeping an accumulated total of the friction values (defined by slope and water) when traversing from the origin to every cell in our extent.

Utilizing the cost-surface, we are able to calculate a least-cost path. In light of our analysis, this path can be used to focus our search efforts, as it denotes the trail that is most accessible to the hiker with respect to her starting point, and her destination.

Least-cost Path

Cartographer: Julie Stringham Data Sources: University of Oregon GIS Server, ESRI (Basemap) Projection: NAD 83, Lambert Conformal Conic

This map is an extension of our basemap, this time, showing the least cost path that was generated from the cost-surface analysis shown in the previous map. The path has been placed over the cost surface analysis to show the relationship between the calculations, as well as its adherence to topography.

Evaluation of Methods and Findings:

[Q3] In our analysis we use a broad definition of landcover and slope to generate our cost surface. We should keep in mind that more specific landcover categories would be helpful in determining a cost surface (for example, a dense forest would theoretically take more time to navigate through than a meadow). Known hiking trails, roads, or other infrastructure that might attract a lost hiker should also be taken into account, but is not a part of this analysis due to lack of data accessibility.

Proposed Lookout Facility Viewshed Evaluations

Cartographer: Julie Stringham Data Sources: University of Oregon GIS Server, ESRI (Basemap) Projection: NAD 83, Lambert Conformal Conic

 

 This map is the result of an observation point (viewshed analysis) performed on the digital elevation model. The DEM was replaced with a topographic basemap to give a more sensible understanding of the natural landscape.

[Q5] Despite the confines of our data, we can still make an educated guess as to which proposed lookout station would serve us better. From the map above, we may deduce that the lookout station located on the ridge of  the Olallie Mountain, the east-most proposed lookout facility, should be used. It is clear from our Observation Point analysis that the view-shed of this lookout facility encompasses a significant segment of the approximated least-cost path. The map product makes our claim apparent; we can see a much larger section of the path intersecting the green view-shed than the orange (belonging to the west-most lookout point).

[Q6] In general, we conclude that search efforts should be focused east south east of the hiker’s starting point. From the topography, cost-surface and least-cost paths shown in the maps above, it is seen that a particular mountain ridge (located southeast of the starting point) draws our attention since it encounters the least-cost path.

If we wanted to provide a more concise analysis, we may want to approximate the travel time of our friend. A time value may be used to better prioritize search areas, and a simple calculation may be computed to approximate the distance to which the hiker may have traveled from the starting point, if we are able to approximate her hiking pace. This would be an ideal way to locate a section of the least-cost path to focus our search efforts on.