Lab 6

Lab 1 | Lab 2 | Lab 3 | Lab 4 | Lab 5 | Lab 6 | Lab 7 | Lab 8 | Lab 9 | Lab 10 | Final Project

Lab 6 Information

The purpose of this week's lab was to delineate a watershed within our area of choice, and then calculate stream order for each stream or river within our watershed.

First, I downloaded a DEM from the USGS data download website. I chose a 10-meter DEM of north-central Vermont, near Montpelier. I spent a couple summers near there doing avian point count surveys, and it was a blast!

After downloading the raster and adding it to a blank map project, I used the "Project Raster" tool, using bilinear projection to project the raster to NAD 1983 - UTM Zone 18N. Then, I used the "Fill" tool to fill any holes in the raster to make sure no errors occurred later on when calculating flow direction. After the fill tool, I ran the "Flow Direction" tool to calculate what of eight directions water would flow in if it was poured over the raster, and then the "Flow Accumulation" tool to calculate how many cells within the raster flowed into each subsequent cell.

I manually searched for an outflow to one of the valleys in the area, and drew a point within the raster cell that looked like it was the pour point for the valley. Then I ran the "watershed" tool with my flow direction and pour point layers as inputs to create a watershed boundary for my river, which I then converted to a polygon using the "Raster to Polygon" tool. With the new watershed polygon, I clipped my flow accumulation layer to the watershed boundary using the "Extract by Mask" tool.

Next, I attempted to define what was actually a stream or river where water actually flows, not just a dry valley or ephemeral stream. To do this, I used the USGS topomap as a basemap and manually adjusted the flow accumulation layer (made semi-transparent) symbology break values until the stream extents matched up with that was shown as a stream on the basemap. My break point ended up being 15000, which was a lot lower than I expected. However, Vermont is a really wet place, and just about any valley holds water. From there, I reclassified the data above and below the break threshold as 0 (no streams) and 1 (streams) using the "Reclassify" tool.

Now, it was time to classify the stream order of the streams in the watershed. I ran the "Stream Order" tool, with my reclassed watershed layer and my flow direction layer as inputs. The tool gave me four stream orders, with water running from 1st to 4th order. To calculate total stream length for each of the orders, I used the "Select by Attribute" tool and selected all segments from each of the stream orders from the stream order polygon I created, and then ran "Summary Statistics" on them to calculate the sum of each. Then I calculated the percent total of each stream order:

Considering Vermont is so wet, 52% of the streams being 1st order isn't surprising.

Then, I explored my map using the 3D Analyst tools, including "Create Contour", "Steepest Path", and "Interpolate Line", the results of which are below:

Lab 6 Answers

1. When projecting elevation data, what is the appropriate choice for the resampling technique? Bilinear? Nearest Neighbor? Cubic?

   • • Bilinear or cubic is fine. Avoid nearest neighbor.

2. If a cell in your flow direction raster has a value of 8, what does that mean?

   • • That means water within that cell flows southwest.

3. If a cell in your flow accumulation raster has a value of 8, what does that mean?

   • • That means that if water were poured on the raster, there would be eight cells that flow into that cell.

4. Where are the higher elevations in the image above? The white areas? The dark areas? In between? Why?

   • • The highest elevations are the darkest areas, because this layer is a flow accumulation layer - meaning that the lowest values (dark) are the cells that accumulate the least flow (i.e. you don't see water flowing up to a mountain top!).

5. Imagine you are working with a 2m lidar DEM and a 10m USGS DEM of the same watershed. You want to create a stream network from the flow accumulation raster. You have determined reasonable ‘threshold’ value to use to reclassify the flow accumulation into ‘stream’ and ‘not stream’ binary cells using the 10m USGS DEM. You can use this threshold value on the 2m lidar DEM to get the same stream network results. True or False? Why?

   • • No, you wouldn't be able to use the same threshold value, because the flow accumulation raster counts how many cells flow into each cell, and rasters with different resolutions have differing numbers of cells. We'd expect the 2m flow accumulation raster to have higher values than the 10m raster.

Additional Downloads

High resolution (PDF) download available here