Recently, the USGS, the Florida Division of Emergency Management (FDEM) and others released the “Florida Statewide Lidar” or “Florida Peninsular” collection, covering a large area in the state with the most coastline in the lower 48. Concurrently, the National Geodetic Survey (NGS) mapped the Keys, including the Dry Tortugas, while the National Park Service and the USGS mapped the entirety of the Everglades for the first time using lidar. With the release of this new data, the Sea Level Rise Viewer team collected, reviewed, modified, and mapped new inundation extents for the entire state of Florida. Along with Florida, Georgia, CNMI, Guam and Catalina Island, CA were also updated. This post will take a look at this process, including the creation of an Everglades-wide DEM mapped at a horizontal resolution of 3 meters suitable for sea level rise mapping. With this update also came new data and products: we’ve started mapping in ½ foot increments and creating spatial metadata to go along with our DEMs. The funding for all of this work came in part from the Bipartisan Infrastructure Law, in a multi-year effort to provide coastal communities with higher accuracy elevation data and mapping.
Background on Our Cycle and Specifications
The Sea Level Rise Viewer (Viewer) is a web mapping tool on NOAA’s Digital Coast that shows the impact of coastal flooding and sea level rise at a community level. To do so, it uses high-accuracy elevation data from lidar surveys and sea surface information from NOAA’s tide gauges and VDatum software. The Viewer is continuously updated with new data as they become available to reflect changes in the dynamic coastal zone.
When selecting a region to update, we consider the age and resolution of the existing DEMs underlying the Viewer, along with the availability of new data. Most areas are mapped at a 3-meter horizontal resolution, but a few states are still at 5-meter, which dates back to the initial version of the tool (GA was at 5-meter until this update). We also work with the National Hurricane Center (NHC), as they use our DEMs to model storm surge in the various oceanic basins. With an unfortunately busy several years, Florida was in need of updated modeling.
When incorporating new data, our approach is to refine, striving to make the data more accurate, or ‘less wrong’. In recent years, the quality of contractor-provided breaklines and DEMs has improved significantly, capturing and flattening even small lakes, rivers, and streams, which simplifies our work. However, we still thoroughly review the data and make necessary edits to ensure we capture the potential inundation pathways from sea level rise. These edits may involve ‘burning-in’ culverts that allow water to flow beneath roadways, smoothing interpolation artifacts after nearshore buildings have been removed, and incorporating flood control structures that may be too narrow to be accurately represented in a 3-meter DEM, or otherwise weren’t captured by the lidar. These were all relevant, particularly in South Florida, during this update.
New Collections Become Available
Peninsular:
In Florida, a significant lidar collection effort took place in 2018-19 across an expansive 34,873 square miles. ‘The Florida Peninsular’ collect was supported by FDEM collaborating with the USGS, the Florida Department of Transportation (FDOT) and representatives from all five Florida Water Management Districts.
The specifications for the project aimed to meet or surpass the USGS Quality Level 1 (QL1) standards, with a nominal density of 10 pulses per square meter along with correspondingly high vertical accuracy standards. To collect the data, 11 lidar equipped aircraft were deployed nearly simultaneously across the state.
High-quality elevation data is vital for inundation mapping, enabling identification of subtle terrain changes and vulnerable regions. It allows for accurate delineation of potential inundation areas, considering factors like slopes, drainage patterns, and barriers.
In addition to meeting the QL1 standards, the project included extra features to make it more useful. Mapping potential flooding from rising sea levels requires understanding how water bodies and low-lying areas are connected. To address this, the project added “connectors” to the breakline vectors. These connectors show areas that may be linked hydrologically, even if they appear disconnected from the airborne lidar due to obstacles like roads. The connectors were designed to seamlessly align with other water features, such as ocean, river, and lake polygons. After review, we were able to assimilate them into the DEMs through semi-automated processes – capturing many more flooding pathways than would have been possible otherwise.
Everglades:
Accurate lidar surveys in dense marsh areas like the Florida Everglades pose unique challenges. Thick vegetation or water inhibit signal returns, leaving void areas. The marshy topography, with subtle elevation changes, further complicates data collection and processing – identifying real ground returns from dense vegetation is no small feat, especially with a dearth of traditional survey marks. This first-ever collection was therefore a noteworthy accomplishment.
Led by the USGS and National Park Service, data collection took place during a period of historically low water levels within the park. The contract team used both a topographic and a topobathymetric lidar sensor to collect terrestrial near-infrared, and submerged green wavelength data. Extensive ground surveys were conducted prior to data collection for testing and calibration (see the Project Report for impressive photos from this campaign). There was even a comparison between the topobathy raster elevations and a plumb-bob survey conducted across the extents of the northern park reaches – underscoring the difficulty of capturing traditional ground control points in this environment.
Final Pieces:
Rounding out the southern reaches of the state, NGS collected topobathy lidar throughout the entire Keys as part of a post-Hurricane Irma supplemental. They also flew Biscayne Bay and the Dry Tortugas. The latter, like the Everglades, is making its first appearance in the Viewer with this round of updates.
Combining These, Then Breaking Them Apart
So this sounds like just a few datasets that needed to be merged, right? Well, the Peninsular collection was released per county (>30), the Everglades were in at least three pieces, while the Miami-Dade area had been part of an altogether different collection, plus all of that NGS data. All told, there were 41 different source DEMs that were wrangled in order to paint one complete picture of the peninsula and Keys. This many disparate datasets creates challenges for data management, assimilation and subsequent mapping. Mapping, at this point, seemed like the easy part! For, along with the data, we needed all of those breaklines, connectors and other ancillary metadata to do things like pull out and flatten all the ocean connected water to one value, and to establish and map confidence in the final products.
We combined these data, ensuring consistent projections and geoids, then clipped and shipped these annoyingly large datasets into manageable chunks, as no one wants to download a single, 50 GB DEM.
Determining where to split is a bit of a balancing act. For our internal, mapping purposes, bigger is generally better – to a point. Bigger input DEMs result in fewer output layers to manage through all of the subsequent steps, culminating in the final Viewer and downloadable products. But as the size increases, so do the amount of CPU cycles to make it through one area. Loathe to find a mistake after a 24 hour script run!
The other downside of huge DEMs hinges on you, the consumer. We try to keep DEMs to somewhere around 5 GB, that way it and any derived products, like inundation depth grids, don’t get unmanageable to pull down for further analysis. To keep things relatively clear, we also look to break things at county boundaries, and, if we are able, those that are demarcated by a river. That way any ocean connected water is easy to piece back together for a seamless look within the Viewer.
Mapping Products
While we have our mapping process published, I’ll summarize. We need two critical pieces to start from. The first, detailed here, is the underlying elevation data. The second, equally integral portion, is a variable surface reflecting local sea heights – a topic for another time.
But there are a few new additions with this round of updates to share. First, we have started to map ½ foot increments, in addition to the whole feet currently in the Viewer. For now, these are only available for download in the regions updated with this cycle – that is, the majority of Florida, Georgia, CNMI, Guam and Catalina, CA. We are working hard to map these additional increments throughout the rest of the country and territories. Once the mapping is done, we will create the map services, modify the Viewer and roll it all out to production. Stay tuned.
We have also taken a cue from projects like the NCEI CUDEM and started to create spatial metadata files to go along with the DEMs. After all, with 41 different source datasets for the Florida peninsula alone, knowing where the data came from is not that straightforward. These are currently available through the Data Access Viewer for the aforementioned regions or by request. We will soon be updating our download page, where they will also have a home.
Other Relevant Links
- Sea Level Rise Viewer, Florida Peninsula Update
- SLR Data Download, including ½ foot increments for FL, GA, CNMI, Guam, Catalina Island
- Articles about the Peninsular Collect:
Digital Coast GeoZone 
Is there a tool for identifying at risk flood towns/regions within Florida, for example Amelia Island vs. Ormond Beach?
LikeLike
James,
While the Sea Level Rise Viewer can help with identifying at risk towns or regions, there are others as well. One that we host at the Office for Coastal Management is the Coastal Flood Exposure Mapper (https://coast.noaa.gov/floodexposure/). Additionally, the University of Florida hosts this state-specific tool: https://sls.geoplan.ufl.edu/viewer/.
Thanks for reading – hope this helps.
LikeLike