Give Me a Break – Line, Part Deux

Alright, so this is the follow on to a previous GeoZone installment on breaklines where I sorta set the scene, covered the background, and hit the first of my major breakline issues (see list).

Breakline Overview

Here is a quick list of some of the more common issues, in my own biased ranking of error potential:

  1. Lack of water feature coverage – streams or islands not break-lined
  2. Breaklines not following shorelines – poor(?) placement of line nodes
  3. Incorrect water elevations – the water surface floating above the land or dropping down to the water
  4. Connectivity of water bodies – streams not connected to water bodies

I was going to try to cover issues 2-4 from one ‘scene’ (1:12,000 scale) in one data set (the same one looked at for #1), but issue #4 required a slightly different data set and I thought I would throw in some other examples; oh well, I tried.

Issue #2 – Breaklines Not Following Shorelines

As I mentioned several times in the first installment – perfection is not an option. This particular ‘issue’ is a great example of that; in reality, no breakline is going to perfectly match the actual shoreline, maybe a node or two hits the mark, but even that is just a coincidence. In most, if not all, data that we receive there are breakline QA calls (errors) of this type.

What typically ‘triggers’ the error (since there are really thousands) is the difference between the breakline node assigned elevation and the elevation of the lidar surface itself. If, for example, the node of a breakline on a flat pond is assigned zero but the bare earth lidar surface at that same point is 3.5 m, then the difference causes a red flag since the node is probably placed inland of where it should be. The solution – move the node away from the shore. As a result more often than not, the breaklines will be biased water-ward of the actual shoreline.

Like all manual generated or semi-automated processes, scale is an important factor. No one expects that every node will be placed directly on the shoreline zoomed in to 1:100. The scale of the breaklines should be detailed during the project kick-off and review should follow the same scale. But, here is the problem – the QA is typically run in an automated routine and scale is really not factored in. A tolerance for the vertical difference is the primary way to suppress red flags on minor differences.

As a side note then, I suggest that, for the above reason, breaklines have a specified nodal spacing. This would not require placement of nodes at specific spacing, but rather to resample the breaklines to a specified node spacing – something like 5 x the nominal point spacing or so. One could even go so far as to define tolerances based on land cover or land use.

But are minor deviations in breaklines a problem? In most cases, no they are not (see Figure 1). Where it becomes an issue is in areas with infrastructure at or near the shoreline (Figure 2). In this problem case, it appears that the breaklines were developed using a contour as the nodes are spaced in a non-human manner and don’t follow the intensity or imagery, which are the most common ways to produce the hydro breaklines. If you get breaklines that have this character (Figure 1 and 2), there is apt to be lots of breakline issues.

Image of breaklines with incorrect shoreline placement that do not create big problems.
Figure 1. Breaklines with incorrect shoreline placement (arrows and many more instances) that do not create big problems. Note the very jagged breakline behavior – it is likely they were generated from contours not imagery or intensity.

The easy answer here is that this breakline should be corrected. The tougher big-picture issue is that the cost of breakline development/corrections can spiral upwards for little real world improvement when node placement is off, but in a non-critical location. I don’t pretend to have the answer – but it is something to think about when specifying or using data sets with hydro breaklines.

Image incorrect shoreline placement that does matter.
Figure 2. Incorrect shoreline placement (arrow) that does matter. Note how the intensity was not followed in the generation of the breakline. It is likely a contour was created to define the breakline.

Issue #3 – Incorrect Water Elevations

Remember that breaklines are generally biased water-ward, which is the better option; however, this can cause ‘floating or digging’ water (Figure 3). In reality every breakline with Issue #2 will trigger an Issue #3; however, the magnitude of difference between breakline elevation and the water surface will either cause a problem (big difference) or look fine (negligible).

Floating water is when the breakline defined water is at a substantially higher elevation than the measured surrounding shoreline; digging water is when the breakline water elevation is significantly lower than the actual shoreline. In my opinion, ‘substantial’ or ‘significant’ is around 20 to 30 cm. In our work these are more of an aesthetic problem, but it could pose a flow problem if the DEM is being used for Hydrologic & Hydraulic modeling or similar.

Image showing examples of floating and digging water.
Figure 3. Examples of floating and digging water. Digging water is common and often results in high land to water slopes. Floating water is less prevalent but it is very obvious when it occurs.

Generally, digging water is not a problem; but on beaches it can cause steep, unnatural drop-offs that may look like ‘real’ slope change (Figure 4). Using hydro enforced DEMs on beaches for erosion studies should be avoided since geomorphology is important. Floating water may actually be more desirable in this case and indicates that the breakline is offshore of the beach. In most cases, however, floating water is more problematic since water flow is opposite of typical drainage and, frankly, just looks wrong.

Image showing breaklines on a beach.
Figure 4. Breaklines on a beach; both have Issue 2 problem, which also begets an associated Issue 3 problem.

Digging or floating water problems (i.e., a significant difference) can be unavoidable in coastal collections if a tide window is not specified; and even if it a tide window is used there may be some places where it occurs. The solutions include moving breaklines slightly or changing the node elevations. If I had to choose I would change node elevations and favor a slight digging in inland areas (rivers) and floating on coastlines.

Issue #4 – Connectivity of Water Bodies

This is a problem mainly in breaklines used in hydro-enforcement where water flow is being inferred and often occurs where culverts are present. It is less of a problem in hydro-flattened DEMs where the emphasis is on presentation, not modeling. Nevertheless, we use hydro-flattened DEMs in SLR modeling so it does help when water bodies are connected.

Breaklines for hydro-enforcement are a bit more complex and includes single line, stream bank, and pond/lake breaklines (Figure 5) as well as rules on how large a drainage area or stream/pond has to be to receive the different type of breaklines (See FAQ #11 in USGS Document for lots more info). It can also include ‘hydro connector’ lines that come into play when culverts are modeled differently than with single line streams.

Image showing hydro-enforced breaklines and connectivity issues.
Figure 5. Hydro-enforced breaklines and connectivity issues.

I have used a different data set to show this issue, one that was specified with hydro-enforced breaklines (Figure 5). Note that the break in connectivity at the example location includes a culvert running under the road, as is commonly the case. In this example, the culvert is almost as large as the small drainage feature and would not limit flow much but, as provided in the DEM, this stream would not flow – even though it has centerline which suggests conductivity. Unfortunately, the specifications on the collection (it was an early one) did not include hydro connector lines at culverts, so the issue is not in the breakline processing, but rather a function of the specifications. This problem is real and unfortunately quite (very?) common – you need to ask for it or at least understand the limitations of modeling ‘real’ flow (Figure 6).

Image showing flow blockage.
Figure 6. Left example shows flow blockage if culvert not included in hydro-enforcement. Right example shows more conventional flow where culvert is included. Images from USGS Lidar FAQ.

So, those are the most common issues we run into when using breaklines to derive products or generate information. It is not a complete list, but rather the most noticeable ones to output. A good slide show and more examples of breaklines were prepared for ILMF 2010 by Merrick – it is worth a look. Hope the provided examples help both high-light some things to look out for and understand the reasonable limits that can be expected in most data sets.

One comment

Leave a Reply. Comments are moderated.

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s