U.S. National Lidar – is there a business case?

Lidar Status Quo in the United States

Coverage of high accuracy public domain elevation as of September 2011
Coverage of high accuracy public domain elevation as of September 2011. Click image to enlarge.

If you use elevation data for your work, you probably know that lidar availability in the United States is patchy, inconsistent, and just plain confusing. You can see what sort of patchwork it is on the new United States Interagency Elevation Inventory. In many cases, adjacent data collections have differing specifications of resolution and accuracy, or are processed differently such that they can’t be combined for a larger analysis.

For you coastal fans, you can throw differences in tidal constraints during collection into the mix. One of the activities I’ve been involved with over the last two years is an assessment of the business case for a national lidar program that would provide a consistent baseline of lidar elevation data for everyone to use. It’s been an interesting experience and I thought I’d share some of what I found useful and also what didn’t work out so well. Note that these are my views and not necessarily those of the government or NOAA. The real credit goes to lead agency USGS, the other funders of the study (FEMA, NCRS, NGA), and Dewberry for conducting the study. NOAA also contributed to the elevation inventory work. Dewberry currently has the full 800+ page report available online for those that want the details, but you can also get the corporate overview and the proposed next steps.

National Enhanced Elevation Assessment (NEEA)

Why do a study? Well, there are government agencies at all levels that make use of elevation data and many of them need better data than what’s available in the NED. While the NED suits the needs it was originally intended for and is updated regularly with better data, much of it is still considered old, low accuracy, and low resolution.

Example of the difference between NED data and lidar data in Oregon. Much greater detail can be seen in the lidar data. The purple profile (not shown) illustrated the NED values were higher than the lidar values.

Many of the federal agencies (plus the states represented by NSGIC) get together to coordinate on elevation needs as part of the NDEP and those agencies realized that the current ad hoc approach to elevation collection wasn’t providing a uniform coverage to a common specification. However, if any large new program was going to be created, the American taxpayer would need proof that the benefits outweighed the costs, particularly for the tasks agencies were already required to do. Gathering information from all those involved (government, industry,NGO) would require a lot of manpower and expertise that wasn’t readily available at the agencies, hence it was contracted out. One of the related issues of the ad hoc system is the perception that there is a lot of overlap and redundancy in the data collection currently. Having an independent group examine that issue was a good idea and it was included in the study.

What Didn’t Work?

Many of the newer uses of elevation data go beyond a simple bare-earth model (in a bare-earth model only elevations on the ground are kept and trees, buildings, etc. are removed). The recognition of those needs really drove the focus of the study toward lidar as the technology of choice. Other technologies, primarily IfSAR, were also included but weren’t expected to be able to provide information such as canopy structure. That’s all on the land side. On the water side, only lidar was considered. This was a severe limitation in coverage. Though bathymetric lidar fills an important niche in data collection, it rarely is able to cover all the area needed for any given issue that needs to be addressed. Expanding the study to include the bathymetric needs more completely by, for example, including the extent of U.S territory and acoustic technologies would result in program costs where the water side costs dwarfed the land side. Personally, I favored cutting off the study at the water’s edge because after that, there was no logical cutoff geography. In hindsight (always 20-20), including some shallow bathymetry did reveal that it was an area of interest to far more agencies than I had expected. For better or worse, the study is weak on the water side and that was known from the start.

Perhaps the biggest problem encountered in doing a benefit/cost study like this is that the benefits are frequently very hard to put a dollar value on. Better elevation data could mean that predictions of flood areas during a storm are better, leading to improved emergency response and evacuation of the right areas. All good things, but putting the value on it nationally is very difficult. All the agencies had problems along these lines. In the end, many of the uses had a zero valuation even though we think there are large benefits. Each of those uses would probably require its own study to even figure out the right order of magnitude for the benefits. So, not only are the benefits under reported, it also means that the business uses that appear to be the big benefits in the study may not actually be the biggest ones.

What Did Work?

So, after all that long-winded weaseling above, what were the interesting bits? Counting up over 600 business uses for higher quality elevation data and explaining them was a huge undertaking. Often when I’m asked by people above my pay grade regarding the use of elevation, I’m momentarily at a loss for words because I don’t know what to pick first. Now I know why. It’s a framework dataset and has uses for many different problems. I’d encourage you to at least skim the document to get a flavor. There are certainly a lot that I’d never thought about. Reducing crop loss in precision agriculture by identifying the drainage problems is just one of many on the private industry side. One that surprised me was the potential to build vehicles that use elevation and slope data with transmission-control technology to change gears in anticipation of the road ahead, saving between 4% and 12% in fuel costs. That might sound small, but it is 4% of a very big number. Even 1% would save American drivers over $6 billion per year. There are also lots of examples of the business cases you might already expect, such as estimating sea level change impacts (we use lidar data to drive our coastal flooding impacts viewer), creating DFIRMs, or calculating fire fuel loading.

Lots of different possible programs were examined, varying the resolution, the accuracy, and the repeat cycle. The table shows how the different quality levels were defined and each business case had to determine what quality they needed and how frequently. The benefit/cost of several scenarios were very similar and in the range of 4.5:1, meaning that for every dollar spent on the program, we expect $4.50 back in benefits to the public. Given the inability to quantify a lot of the benefits, this is almost certainly a conservative number. The general recommendation was for coverage every eight years of quality level 2 (see quality levels table below) data for all the U.S. except Alaska and quality level 5 for Alaska. The second best option was the same quality but every fifteen years.

Matrix of quality levels, resolution, and accuracy.
What about that question of duplication in the current way we collect data? Turns out that there isn’t much duplication and we’re doing a good job coordinating. That doesn’t mean perfect. With the amount of activities most agencies engage in, it is nearly impossible for their representative to know all the elevation activities happening. The elevation collection might be just one piece of a much larger research endeavor and the investigator may not even realize there is an elevation representative for the agency. This is generally the small area collections though.

How much of the United States have we collected in the last 10 to 12 years? Around 28% of the area excluding Alaska has some sort of lidar coverage. Alaska generally doesn’t make sense for lidar coverage due to the collection difficulties, but about 15% had been covered with IfSAR when the study was being done and more is being collected. In case you’re wondering where you might find some of that 28% coverage, it is scattered around quite a bit, which is another problem a national program would address. Some good places to look are Digital Coast’s data viewer (primarily for data in coastal states), Earth Explorer (the USGS replacement for CLICK), and OpenTopography. There are also many states that house and distribute their data in their state atlases.

Next Steps

USGS is leading the effort to move this forward. In the current fiscal climate any additional money for a new program is a tough sell, even if you can show benefits to the taxpayer outweigh the costs. The first step will be to see if the primary agencies can work together to support it out of their yearly appropriations. I’m not in a position to guess whether that will work (though some of you might be).

So, what do you guys think? Would a federal program to collect elevation nationally be a good thing? Would it be ineffective given all the usual rules that government has to follow? If you do believe a national dataset freely available to all makes sense, how would you do it? If you don’t, what’s your alternative?

One comment

  1. As a private person living in a floodplain I would love to have a map showing the elevations around me. When someone fills it would be nice to give the county a map showing what the land was before. I am having a problem with the county making a person remove 600 loads of dirt from the floodway below the 10 year flood.
    Ed Lewis


Leave a Reply. Comments are moderated.

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

WordPress.com Logo

You are commenting using your WordPress.com 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