Issues and Answers in Quality Control of LIDAR

Issues and Answers in Quality Control of LIDAR DEMs for North Carolina DFIRMs Gary W. Thompson, RLS North Carolina Geodetic Survey David F. Maune, Ph. D. , C. P. Dewberry & Davis LLC, Fairfax, VA

Hurricane Floyd — 1999 l l Revealed limitations in the State’s flood hazard data and maps Many maps compiled in the 1970 s by approximate methods; no detailed H&H Most of NC needed to be remapped digitally, consistent with FEMA’s Map Modernization Plan Over 50 counties needed re-mapping immediately with new DFIRMs

DFIRM Components Base + = Topography + Flood Data DFIRM

Cooperating Technical State (CTS) l North Carolina, FEMA’s first CTS, is responsible for: 4 Re-surveying the State 4 Conducting flood hazard analyses 4 Producing updated DFIRMs l l North Carolina Geodetic Survey (NCGS) serves as the State’s technical lead Dewberry & Davis LLC serves as FEMA’s Map Coordination Contractor (MCC)

Phases I, II and III

Photogrammetry or LIDAR? l l l The North Carolina advertisement did not specify technologies to be used Focus was on high-resolution and high-accuracy digital elevation data suitable for semi-automated H&H modeling All firms proposed using LIDAR to generate the TINs and DEMs; but some proposed using photogrammetry to generate breaklines

Winning Teams l Watershed Concepts team includes: 4 Earth. Data International (LIDAR) 4 ESP Associates (ground surveying) l Greenhorne & O’Mara team includes: 4 3 Di Eagle. Scan (LIDAR) 4 Mc. Kim & Creed (ground surveying) 4 Hobbs, Upchurch & Assoc. (ground surveying)

Delivery Order No. 1 l Task 1: LIDAR Data Acquisition 4 Vertical RMSE = 20 cm in coastal areas and 25 cm inland (equivalent contour interval of 2. 16’ and 2. 70’), the highest accuracy realistically achievable 4 This was a compromise from FEMA’s 15 -cm LIDAR standard, considered unrealistic based on prior studies 4 Daily calibration at local test site l Task 2: Generation of Bare-Earth ASCII files (randomly spaced)

LIDAR Laser Sensor l l Laser scanner with mirror measures scan angles and distances for up to 50, 000 pulses per second Airborne GPS measures position Inertial Measuring Unit (IMU) measures roll, pitch, heading Record first/last returns

Issue: How best to perform LIDAR system calibration Courtesy of Earth. Data International

Issue: How best to post-process LIDAR (These are “raw” images) Courtesy U. S. Army Topographic Engineering Center

Bare-earth data (post processed for vegetation/building removal) Courtesy U. S. Army Topographic Engineering Center

Delivery Order No. 1 (continued) l l l Task 3: Generation of Triangulated Irregular Network (TIN) and breaklines Task 4: Development of 5 m x 5 m DEMs in ESRI GRID Float Format Task 5: Development of DEMs in Three Additional File Formats Task 6: Preparation of Project Report Task 7: Production of Optional Digital Orthophoto Images

Digital Elevation Models (DEMs) l l l DEMs typically have uniform “post spacing” where x/y coordinates are evenly divisible by 5 m, 10 m, 30 m, etc. Interpolated from TIN data; e. g. , LIDAR. Neither TIN nor DEM points are clearly defined on the ground.

TINs — Superior for 3 -D Surface Modeling; e. g. , H&H Modeling l l l A TIN is a set of adjacent, non -overlapping triangles computed from irregularly spaced mass points with x, y coordinates and z values, plus breaklines. Mass points can come from LIDAR or other source. Best breaklines come from photogrammetry, then digital orthophotos.

Hydraulic Models Require “Representative” Cross Sections l l l Cross sections are carefully selected to be representative of reaches that are as long as possible, without permitting excessive conveyance change between sections. Typically between 500’ and 2, 500’ apart. In addition to surveyed cross sections, others can be “cut” from the LIDAR data.

Issue: How best to generate Cross Sections

Issue: How best to generate Breaklines Watershed Concepts l l Surveyed cross sections at bridges Hydro-enforced stream centerlines Digital orthophoto breaklines at stream shorelines LIDAR models stream banks and overbank areas Greenhorne & O’Mara l l Surveyed cross sections at bridges Hydro-enforced stream centerlines Photogrammetric breaklines at tops and bottoms of stream banks LIDAR models overbank areas

Issue: How best to handle “obscured areas” and “artifacts”

Issue: How best to compute RMSEz of bare-earth TINs/DEMs Since TIN/DEM points not clearly defined: l Survey a minimum of 20 checkpoints in all 5 major land cover categories representative of the floodplain l Choose checkpoints on flat or uniformly sloping terrain; interpolate LIDAR points l Use no checkpoints in vegetation known to be too dense for LIDAR penetration l Discard 5% of “outliers”

Issue: Check points in such areas skew RMSE calculations l l LIDAR has fewer areas than photogrammetry where the terrain is obscured. One “bad” checkpoint in such areas will over -ride 1, 000 “good” checkpoints elsewhere, and thus skew the results.

LIDAR Advantages Compared with Photogrammetry l l LIDAR needs only a single line-of-sight to measure through/between trees High-altitude LIDAR data are more accurate than from photogrammetry LIDAR generates higher-density TINs/DEMs at lower costs LIDAR acquires data both day and night (but not through clouds)

LIDAR Disadvantages Compared with Photogrammetry l l l LIDAR returns on water are unreliable LIDAR is ill-suited for breaklines; e. g. , 5 -m point spacing could “jump” across a breakline LIDAR is new technology; standards have not yet been developed Contour lines are not as smooth Streams are not automatically hydroenforced, must be done manually

LIDAR contours not hydro-enforced (same problem with TINs/DEMs)

Conclusions l l This project will demonstrate the do’s and don’ts of LIDAR for H&H modeling and serve as a model for years to come This project will also be used to update FEMA standards

Issues and Answers in Quality Control of LIDAR DEMs for North Carolina DFIRMs QUESTIONS ? ? ?