St. Johns River Water Management District
Sandra Fox, MS, GISP Environmental Scientist, Bureau of Watershed Management and Modeling, Division of Water Supply Co-authors: • Keith Paterson and Nick Kules, Dewberry • Kimberli Ponzio, Steven J. Miller, Richard Guilfoyle, Bill Van Sickle, James Walters, and Sherry Brandt-Williams, SJRWMD • Al Karlin, Southwest Florida Water Management District (SWFWMD)
St. Johns River Water Management District
2 talks….. 1) Reprocessing existing point cloud data (at the request of session organizer….) project area happens to be a wetland dominated area
2) LiDAR acquisition in wetlands
St. Johns River Water Management District
Why reprocess existing point cloud data? • DEM (deliverable) has problems •
Not picked up on during evaluation
•
Accuracy statistics don’t always tell you what you need to know….
• AND - - - Point cloud shows evidence of being worth reprocessing • Considerable financial savings over reflying the same area
St. Johns River Water Management District
Where is project located?
St. Johns River Water Management District
Where is project located? LiDAR collected in 2012
St. Johns River Water Management District
Where is project located? LiDAR collected in 2012 Wellsurveyed
St. Johns River Water Management District
Where is project located? LiDAR collected in 2012 Wellsurveyed New survey 2013
St. Johns River Water Management District
Elevation (feet)
St Johns Marsh Conservation Area (SJMCA) story
Distance along transect
St. Johns River Water Management District
Elevation (feet)
St Johns Marsh Conservation Area (SJMCA) story N = 221 (all transects) Median difference = - 0.42 ft Arithmetic mean difference = - 0.45 ft
DEM elevation is below survey ground - - -
Distance along transect
St. Johns River Water Management District
Elevation (feet)
St Johns Marsh Conservation Area (SJMCA) story N = 221 (all transects) Median difference = - 0.42 ft Arithmetic mean difference = - 0.45 ft
DEM elevation is below survey ground - - -
Distance along transect
St. Johns River Water Management District
Elevation (feet)
St Johns Marsh Conservation Area (SJMCA) story N = 221 (all transects) Median difference = - 0.42 ft Arithmetic mean difference = - 0.45 ft
DEM elevation is below survey ground - - -
Distance along transect
St. Johns River Water Management District
Consequence of using the original DEM as delivered?
St. Johns River Water Management District
POINT CLOUD DATA (cross section view)
z
x
z
x y
Solution: reprocess with algorithm better fit for data
St. Johns River Water Management District
BEFORE
AFTER
St. Johns River Water Management District
Looking a little closer - - - -
St. Johns River Water Management District
Looking a little closer - - - -
St. Johns River Water Management District
Looking a little closer - - - -
St. Johns River Water Management District
In 2-D modeling, this elevation difference is interpreted as an extensive “overland weir”, causing flow direction to be wrong
St. Johns River Water Management District
In 2-D modeling, this RMSEz (ft) elevation difference is interpreted as an extensive “overland weir”, causing flow direction to be wrong
Stats on point cloud data
TerraSolid TerraMatch software, MeasureMatch tool
Nick Kules
St. Johns River Water Management District
Why is reprocessing worth doing? • Estimate price for acquiring new LiDAR data for this area – $250,000
• Cost of this project (2015 – 2017) – $57,459 , plus staff time (~ < 0.1 FTE per year)
St. Johns River Water Management District
Reprocessed area
What about this area?
St. Johns River Water Management District
Let’s talk about wetlands
St. Johns River Water Management District
Let’s talk about wetlands
St. Johns River Water Management District
Characteristics of wetlands relevant to LiDAR acquisition • Dense vegetation • Wetlands are frequently inundated • Southern half of that LiDAR collect (that was reprocessed) was inundated
St. Johns River Water Management District
Characteristics of wetlands relevant to LiDAR acquisition • Dense vegetation • Wetlands are frequently inundated • Southern half of that LiDAR collect (that was reprocessed) was inundated
St. Johns River Water Management District
Characteristics of wetlands relevant to LiDAR acquisition • Dense vegetation • Wetlands are frequently inundated • Is one of these easier to “manage”? – YES – dense vegetation – Project specifications, point density, flight line overlap, etc.
St. Johns River Water Management District
Characteristics of wetlands relevant to LiDAR acquisition • Dense vegetation • Wetlands are frequently inundated • Is one of these easier to “manage”? – YES – dense vegetation – Project specifications, point density, flight line overlap, etc.
St. Johns River Water Management District
FY18 acquisition “Puzzle Lake”
USGS 3DEP
FY17 acquisition “Fort Drum”
Is it possible to avoid collecting LiDAR when there’s water on the wetlands?
Hydrology as a driver for LiDAR acquisition
How did SJRWMD make this happen? a) Assemble a team b) Convince the decision makers c) Instruct the vendor
d)Model – when is wetland area acceptably dry? e)Model – when is that likely to happen? f) Notify vendor frequently of hydrological status in order to mobilize acquisition team/equipment
29
St. Johns River Water Management District
Model – when is wetland area acceptably dry? New FY18 LiDAR acquisition Footprint Three USGS stage recorders: • Lake Harney • Christmas • Cocoa
WSIS (2012 Water Supply Impact Study) Hydroperiod Tool (HT) Results Note: HT method received favorable comments from National Research Council - - -
(a bit of serendipity - - - - )
In short - - - - this is a well-studied area
St. Johns River Water Management District
Model – when is wetland area acceptably dry? 50% exceedance looks like this (all areas in the polygon are classed as wetlands) BLUE = INUNDATED
Reference: Hydroperiod Tool Results WSIS Chapter 10, Appendix D. Fox, Kinser, Keenan, Montague and Hydorn. 2012
This is what 70% EXC looks like
St. Johns River Water Management District
Model – when is wetland area acceptably dry? 50% exceedance looks like this (all areas in the polygon are classed as wetlands) BLUE = INUNDATED
Reference: Hydroperiod Tool Results WSIS Chapter 10, Appendix D. Fox, Kinser, Keenan, Montague and Hydorn. 2012
This is what 70% EXC looks like
St. Johns River Water Management District
Model – when is wetland area acceptably dry? Prediction
Hydroperiod Tool Model output
Historic lows occur late in the spring……
St. Johns River Water Management District
Model – when is that likely to happen? (Ponzio)
Report to vendor
St. Johns River Water Management District
FY18 “Puzzle Lake” acquisition update
St. Johns River Water Management District
Accuracy assessment of the 2017 Fort Drum LiDAR acquisition – Project specs - including timing - make a big difference!
Thanks to Richard Guilfoyle and team for extra effort on survey.
St. Johns River Water Management District
Accuracy assessment of the 2017 Fort Drum LiDAR acquisition – Project specs - including timing - make a big difference!
Mean near zero
Thanks to Richard Guilfoyle and team for extra effort on survey.
St. Johns River Water Management District
Accuracy assessment of the 2017 Fort Drum LiDAR acquisition – Project specs - including timing - make a big difference!
Mean near zero Balanced error
Thanks to Richard Guilfoyle and team for extra effort on survey.
St. Johns River Water Management District
Accuracy assessment of the 2017 Fort Drum LiDAR acquisition – Project specs - including timing - make a big difference!
Mean near zero Balanced error
“to do” – evaluate accuracy in different wetland classes Thanks to Richard Guilfoyle and team for extra effort on survey.
St. Johns River Water Management District
Summary 1: Reprocessing existing LiDAR data If the point cloud warrants….
St. Johns River Water Management District
Improving LiDAR-derived DEMs; DEM evaluation Yes Decision tree
Survey data
No – DEM biased high (vegetation)
No – DEM biased low
Elevation NAVD88 (ft.)
DEM
Is the DEM acceptably representative of “true ground”?
SURVEY DEM
St. Johns River Water Management District
Improving LiDAR-derived DEMs; DEM evaluation Yes Decision tree
DEM
Surve y data
Is the DEM acceptably representative of “true ground”?
No – DEM biased high No – DEM biased low
Feedback to industry – improve future LiDAR acquisition
Yes
Point cloud data
Are there any returns representative of “true ground”?
No No Point Cloud data
REPROCESS POINT CLOUD
NEW DEM
Why aren’t there any? “Z-bump” Correction
NEW DEM
St. Johns River Water Management District
Improving LiDAR-derived DEMs; DEM evaluation Yes Decision tree
DEM
Surve y data
Is the DEM acceptably representative of “true ground”?
No – DEM biased high No – DEM biased low
Feedback to industry – improve future LiDAR acquisition
Yes
Point cloud data
Are there any returns representative of “true ground”?
No No Point Cloud data
REPROCESS POINT CLOUD
NEW DEM
Why aren’t there any?
Or ----- Fly new “Z-bump” LiDAR….NEW DEM Correction $$$
St. Johns River Water Management District
Summary 2: LiDAR acquisition in wetlands • Negotiate for highest point density and flight overlap possible…. In this respect, LiDAR is getting better and better - -
• Insist on accuracy assessment in wetlands • Fly when it’s dry
St. Johns River Water Management District
Thank you!
[email protected]