Lidar has become a huge benefit for those of us looking at coastal change. Quantifying change based on a beach profile every mile or so has been replaced by tens of millions of points every mile, increasing the resolution of the change analysis and efficacy in doing miles of shoreline dramatically. The downside is that it may have made it a bit too easy to become complacent with the results – they look so darn good. Notice I did not add accuracy to the advantages of lidar – that is what it comes down to and what I am going to focus on.
It is important that when a municipality/county/state uses lidar to quantify how much sand they lost, so they can recoup some losses or plan on renourishment, they understand what the value represents. It can be easy to overlook some subtleties because, regardless of technique, the change will likely be a large number.
I was looking back at some data from Post-Sandy lidar surveys and some coastal change surfaces and it finally struck me, after a bit of puzzlement over the lack of change I expected, that lidar derived surfaces make seemingly correct results easy to accept from afar, but when you look closely you see it is only half the story. The unfortunate fact is that you can only compare two surfaces where there are two surfaces.
The generic profiles (Figure 1) pretty much sum up what I was seeing, and it really is not anything earth-shattering. If the data has been collected by a topographic lidar sensor – the most common – the data will be limited to the swash zone and above (Zone A – Figure 1) and will depend on tide levels during flights (I think tide considerations make it into each of my blogs). The tough part is that the need to fly quickly post storm can lead to unfavorable tide conditions – and thus change calculations can be further limited. Only change in the zone (A) where both surfaces are present will be computed, which can be substantially smaller than what you were expecting, and is what I consider “Beach Change Volumes.”
OK, so data from topographic lidar sensors will tell you a bit more, extending the analysis out to Zone B (Figure 1), if the data are looked at separately. That is to say, if you calculate the volumes of both surfaces (red and green in Figure 1) above the tide line (blue line) and subtract those volumes you get a slightly better picture, not complete, but better and what I would consider “Beach Volume Changes.” The important distinction is that the values have a qualifier that describes the measurement – i.e., 10,000 C. yds lost above MTL.
That leaves Zone C; this is the value the municipality/county/state is looking for. Capturing the data to provide the value will take (absolutely!) fantastic tide conditions – (twice!), topo-bathy lidar data, or beach profiles. The distance from B to C is obviously very location dependent, but just getting the full swash zone (call it B+) can provide a significantly better estimate of the Beach Volume Change.
Let’s look at real data to put some relative values on these generic profile examples. I have chosen a new data set (check it out – this is a great data set to show what lidar can do!) to compare to existing data in New Jersey. I have corrected the lidar data to mean tide level (MTL) and am looking an 800 m stretch of beach that extends from the primary dune to the approximate breaker line (red box in Figure 2).
I am fortunate to have two topo-bathy lidar data sets (USACE 2005 and NGS 2013) that extend seaward enough to say we are getting close (?) to capturing Zone C (Figure 3). For the sake of discussion, I first pretended that these data sets were topo only and they were collected at MTL to do volume comparisons (Zones A and B). Then I used all the data (topo-bathy) to generate the best results I could (Zone C). Volumes were then computed (Table 1) to look at the relative differences between the compartments in Figure 1.
[table id=1 /]
So, the real-life profiles (Figure 3) look a bit different than the faked/generic profiles (Figure 1) but you can see the same basic issues detailed in the generic graphic. Most notably that the loss is really highest at or near the shoreline (where we quickly lose data). Also note that the profiles don’t converge and that there is really more sediment missing from these loss values, such that Zone C is really not complete.
What Does It All Mean?
Some ‘take aways’ – when comparing only the Zone A data, the “Beach Change Volume,” the results probably provide like a 65% estimate of the loss depending on tide; this technique should be avoided. You can get a better estimate if you include what you have from Area B – “Beach Volume Change” – maybe up to 70-75% of loss. But Zone C will be illusive without great tide coordination or topo-bathy lidar. Or, and we have yet to talk about them, profiles. Even though the profiles represent a 1D look at this beach system – using the data from them (even they don’t go any further in this case than the lidar – and probably won’t unless you have a very tall swimmer) was pretty consistent with the full surface method (Table 1). In fact, a couple of profiles can provide a decent way to gauge how well the volumes compare (much as I have done) allowing for some better estimates to be made. I guess it may not be time to throw the profiles out yet.
One last note, and I have seen this several times, is that topo data flown directly after storms can be done to meet very strict time requirements, thus capturing the storm’s affects before other stuff happens. The downside is that tide windows can go, well, out the window and getting data below MHW can be difficult. Providing decent volume estimates from this data can be even more tenuous if using direct comparison of surfaces, e.g., “Beach Change Volumes.” For example, picture the tide at 0.5m for the 2013 data (approximate MHW in Figure 3) and the loss of information that would occur. Under this scenario, the “Beach Change Volume” is -45,000 c.yds or 47% of the Zone C value. Using the Zone B technique the “Beach Volume Change” above ≈MHW is 59,375 c. yds (62%) – a significant difference from only using Zone A.
I hope this makes sense and will help those digging in and looking at changes from lidar data alone. It is great data, but like most things, has limitations; knowing them is 80% of the battle.