Towards Optimizing the determination of accurate heights with GNSS

1
Towards Optimizing the determination of accurate
heights with GNSS

• Dan Gillins, Ph.D., P.L.S.

• October 9, 2014

2
Current Research Efforts

• Optimizing the determination of accurate heights
  with GNSS (NGS)
• NGS 58/59 guidelines
• OPUS-RS, OPUS-S, OPUS Projects
• GPSGLONASS vs. GPS-only
• Real-time networks
• UAV remote sensing
• Evaluating accuracy of current practice
• Improving data collection and processing steps
• Earthquake hazard mapping
• Megaquake-induced liquefaction (USGS)
• Liquefaction lateral spreading hazard maps
  (USGS)
• O-HELP a web-based GIS tool for assessing
  earthquake hazards in Oregon (CLiP)
• GNSS surveying in forested environments

3
Importance of Research

• Accurate heights are crucial for a multitude of
  scientific studies and engineering projects
• monitoring deformations, engineering layout,
  flood mapping, sea level rise, development of
  nautical charts, topographic mapping, crustal
  movement, subsidence studies
• Geodetic leveling remains the most accurate form
  of obtaining heights
• Requires line-of-sight, slow (expensive), prone
  to errors
• GPS has revolutionized the surveying of geodetic
  networks
• Does not require line-of-sight, easy to use,
  quite accurate
• It is desirable to take advantage of the
  economics of GPS to determine ellipsoidal and
  orthometric heights

4
Testing and Improving NGS 58/59 Height
Modernization Guidelines

• Objectives
• Evaluate NOS NGS 58 and 59 guidelines
• Follow guidelines to establish a control network
  from Salem to Corvallis
• 20 varying benchmarks
• B versus C stability under varying canopies
• Recommend new guidance based on current
  technology
• Use of GLONASS?
• Improved hybrid geiod models (GEOID12A, GEOID14)
• Use of modern GNSS antennasreceivers
• Improved accuracy and availability of GNSS orbits
• Real-time networks
• Various processing tools
• OPUS, OPUS Projects
• StarNet

5
Summer 2014 Height Mod. Survey

• Collect static GPSGLONASS data
• 10 total days of surveying
• 3 days of 5 hour sessions (primary network), 7
  days of 1 hour sessions (secondary network)
• 5 receivers (6 for 3 days)
• 20 marks covering 350 square miles
• 28 total unique sessions
• 264 total baselines observed
• 103 independent baselines obtained after removing
  outliers (avg. length 10.79 km)

6
Phase 2 Static Survey
7
Phase 3 Find ellipsoidal heights following NGS 58
Station dN dE dZ 95 Confidence Ellipsoidal height (cm)
U727 0.0035 -0.0009 0.0081 0.8403
G728 -0.0031 -0.0028 -0.0005 0.9270
NESMITH -0.0005 0.0016 -0.0053 0.9270
BICKFORD -0.0052 -0.0028 -0.0096 1.0664
S714 -0.0074 0.0005 0.0158 0.9270
N99RESET 0.0036 0.0058 -0.0229 0.9746
J99 0.0034 -0.0117 -0.0118 1.3146
Y683 0.0061 -0.0073 -0.0032 1.1725
BEEF -0.0021 -0.0017 0.0173 1.1571
PRICE 0.0207 0.0144 -0.0009 1.3712
G287 0.0036 0.0037 0.0038 1.3323
J54 -0.0487 -0.0617 -0.0114 1.6302
T714 0.0037 -0.003 -0.0028 1.3932
CORVA 0.0037 -0.0073 -0.0033 1.4946
D728       1.4782
MAG       1.2049
PEAV N/A N/A N/A 1.3193
Z714       1.2652
Q388RESET N/A N/A N/A 1.3712
PEAK N/A N/A N/A 1.4577

• Use only GPS data
• Partially constrained the HARN stations according
  to their reported NGS network accuracies
• Average 95 confidence on ellipsoidal height
  1.23 cm
• Only 1 mark exceeded 2 cm from the published NGS
  ellipsoidal height (N99RESET)
• J54 appears to have been disturbed

8
Next Steps

• Determine which benchmarks have "valid"
  orthometric heights
• Identify any needs to conduct geodetic leveling
  as a redundancy check
• Repeat study using other techniques
• GPSGLONASS (compare with GPS-only results)
• Rapid ephemerides instead of precise ephemerides
• Use of OPUS-S, OPUS-RS, and OPUS Projects
• Use of Oregon Real Time Network (ORGN)
• Single base versus real-time networks
• GPS only versus GLONASS
• Use of StarNET versus ADJUST versus
  OPUS-Projects
• Give recommendations for optimizing the
  determination of accurate heights