- Joined
- Jun 3, 2007
- Messages
- 1,304
- Reaction score
- 2,307
- Golden Thread
- 0
- Location
- Sebastian, Florida
- Detector(s) used
- A sharp eye, an AquaPulse and a finely tuned shrimp fork.
- Primary Interest:
- Shipwrecks
- #1
Thread Owner
Hi folks.
I was doing some analysis of 'where they said it is' and 'where it really is' when you have positions that were taken by the many and varied types of recording a location at sea.
On the 1715 Fleet, prior to the advent of GPS, position sightings were done with beach markers and a sideways sextant. I was not around at that time and have always thought that those positions were a lot less than accurate. As always, accuracy and precision is very dependent on the skill of the person(s) doing the measuring, but this discussion makes me believe that those fixes are a lot more accurate that I previously believed.
GPS was brought to life in 1993 but was a degraded signal until 2000 and has improved with every new generation of equipment, both satellites and recievers. What I did NOT realize was just how much the accuracy was degraded for those 7 years. Then in 2003 we got WAAS and things got better again. NOW I have to rethink some things...
After i got to the point that I thought I knew what questions to ask Chat GPT, I asked it to answer a series of questions. I believe that it is enlightening to all of us who attempt "to go where no one else has gone." or at least to not dig right on top of other excavations...
and yes...it IS kind of long.
Annnnnnnnd discuss!!
"Good morning, Bill. Here’s a detailed rundown of GPS positional accuracy since its operational start in 1993 and how it’s evolved:
👉 In short:
Here’s the chart showing how civilian GPS accuracy has improved from 1993 to 2025 — from ~100 meters with Selective Availability to meter-level accuracy today, and down to centimeters with RTK/PPP.
That’s a really sharp observation, Bill — and you’re right to question it.
Here’s the way it really shakes out when you compare pre-GPS fixes (sextant & shore ranges) versus early GPS (1993–2000) and beyond:
Do you want me to sketch you a timeline overlay that shows “reliability of wreck fixes” — sextant vs early GPS vs modern — so you’ve got a quick visual to judge old chart dots by?
Here’s the overlay:
Want me to also mock up a decision guide (flow-style graphic) for “if a wreck mark came from year X, treat it like Y”? That could make it really quick to apply in your workflows.
Sure...let's take a look
Here’s a decision guide flow you can drop straight into your workflows:
I was doing some analysis of 'where they said it is' and 'where it really is' when you have positions that were taken by the many and varied types of recording a location at sea.
On the 1715 Fleet, prior to the advent of GPS, position sightings were done with beach markers and a sideways sextant. I was not around at that time and have always thought that those positions were a lot less than accurate. As always, accuracy and precision is very dependent on the skill of the person(s) doing the measuring, but this discussion makes me believe that those fixes are a lot more accurate that I previously believed.
GPS was brought to life in 1993 but was a degraded signal until 2000 and has improved with every new generation of equipment, both satellites and recievers. What I did NOT realize was just how much the accuracy was degraded for those 7 years. Then in 2003 we got WAAS and things got better again. NOW I have to rethink some things...
After i got to the point that I thought I knew what questions to ask Chat GPT, I asked it to answer a series of questions. I believe that it is enlightening to all of us who attempt "to go where no one else has gone." or at least to not dig right on top of other excavations...
and yes...it IS kind of long.
Annnnnnnnd discuss!!
"Good morning, Bill. Here’s a detailed rundown of GPS positional accuracy since its operational start in 1993 and how it’s evolved:
Early Operational Phase (1993–2000)
- Initial Capability: When the U.S. declared the Global Positioning System fully operational in 1993, it had 24 satellites in orbit.
- Accuracy (Civilian, Standard Positioning Service – SPS):
- Purposefully degraded by Selective Availability (SA).
- Civilian users typically got 100 meters horizontal accuracy (95% of the time) and worse vertically.
- Accuracy (Military, Precise Positioning Service – PPS):
- Using encrypted P(Y)-code and dual-frequency corrections, military users saw 10–20 meters or better.
- Reality: Even with SA, many surveyors and advanced users could achieve better than 10 m by averaging positions, differential GPS (DGPS), or post-processing.
Selective Availability Turned Off (2000)
- May 2, 2000: U.S. discontinued SA by presidential order (Clinton).
- Civilian Accuracy: Jumped from ~100 m to 10–20 m instantly.
- Military Accuracy: Remained ~3–5 m with PPS.
Differential GPS (1990s–present)
- DGPS:Ground reference stations broadcast correction signals.
- Accuracy improved to 1–3 m.
- Widely adopted by maritime, aviation, and surveying users.
- By the late ’90s, U.S. Coast Guard had DGPS beacons covering most coastlines.
WAAS and Global SBAS Systems (2003–present)
- WAAS (Wide Area Augmentation System)deployed for aviation in the U.S.
- Accuracy: 1–2 m horizontally and vertically.
- Integrity monitoring ensures high reliability for aircraft navigation.
- Similar systems followed: EGNOS (Europe), MSAS (Japan), GAGAN (India), etc.
Modernized GPS Signals (2005–present)
- New Civil Signals:
- L2C (2005), L5 (2010), L1C (2020s deployment ongoing).
- Allow dual-frequency civilian corrections, removing ionospheric error directly.
- Current Civilian Accuracy (Standalone GPS, 2025):
- Typically 5 m horizontal, 10–15 m vertical (95% confidence).
- Under ideal conditions: <3 m horizontal.
- Military (M-Code, under rollout):
- Significantly better, exact specs classified, but thought to be <1 m routinely.
Augmentation with Real-Time Kinematic (RTK) & Precise Point Positioning (PPP)
- RTK GPS:Uses carrier-phase corrections from a base station.
- Accuracy: 1–2 cm.
- PPP (Precise Point Positioning):
- Global corrections via satellites/internet.
- Accuracy: 5–20 cm without local base stations.
- These are the state-of-the-art methods used in surveying, agriculture, and autonomous vehicles.
Summary of Accuracy Over Time
- 1993–2000: Civilian ~100 m, Military ~10 m.
- 2000–2003: Civilian ~10–20 m (post-SA).
- 2003–present (WAAS/DGPS): Civilian ~1–3 m.
- 2010–present (dual-frequency, modern signals): Civilian ~3–5 m, often better.
- Today with RTK/PPP: Centimeter to decimeter accuracy available to anyone with proper equipment.
👉 In short:
- 1993 civilians: 100 m (on purpose).
- Today civilians: 1–3 m routinely, cm-level with augmentation.
- Military: Always had better than civilians, now pushing sub-meter with M-code.
Here’s the chart showing how civilian GPS accuracy has improved from 1993 to 2025 — from ~100 meters with Selective Availability to meter-level accuracy today, and down to centimeters with RTK/PPP.
That’s a really sharp observation, Bill — and you’re right to question it.
Here’s the way it really shakes out when you compare pre-GPS fixes (sextant & shore ranges) versus early GPS (1993–2000) and beyond:
Pre-GPS Methods (Beach Ranges & Sextant on its Side)
- Technique: Salvors (and fishermen) would drive poles or use fixed landmarks on the beach. With a sextant laid sideways, they’d measure horizontal angles between markers.
- Accuracy:
- In skilled hands with clear sightlines, this could yield 5–20 meters accuracy in position — especially when cross-fixing with multiple markers.
- Errors came from human factors: wave motion, sighting precision, and shoreline marker stability (erosion, storms, or replacement by “eyeball” rather than surveyor).
- Reliability: Consistent in a local context — the same crew using the same landmarks could repeatably return to the same patch of bottom within a boat length.
- But not absolute in the sense of global coordinates (lat/long), since it wasn’t tied to a geodetic datum like WGS-84.
Early GPS (1993–2000, with Selective Availability)
Civilian accuracy: ~100 meters (intentionally degraded).- Reality:
- Some fixes were better (30–50 m) depending on geometry and satellite constellation.
- Differential GPS (DGPS) was not widely available to salvors yet — mostly surveyors and government users.
- Result: Early GPS fixes on charts can actually be worse than a good sextant/range fix, especially if they weren’t averaged over time.
- A charted wreck mark from 1995 might be off by the length of a football field or more.
After May 2000 (Selective Availability Off)
- Civilian GPS instantly improved to ~15 m accuracy.
- This finally outperformed sextant/range methods in both repeatability and absolute accuracy.
- WAAS/DGPS (early 2000s) dropped it to 1–3 m, blowing the old methods out of the water.
Practical Takeaway for Salvage Records
- Pre-1993 fixes: Sextant/marker based, often surprisingly tight locally but not in global terms. If you know the landmarks are gone, you lose repeatability.
- 1993–2000 GPS fixes: Can be worse than sextant fixes, with random errors up to 100 m. Don’t blindly trust chart dots from this period.
- 2000+ GPS fixes: Much more reliable, especially post-2003 with WAAS/DGPS.
- Modern fixes: Within 1–3 m, and even sub-meter with survey-grade gear — far beyond what salvors had in the 80s/90s.
- Sextant/marker fixes were often more repeatable than early GPS (pre-2000).
- Modern GPS beats everything before it hands-down.
- When you see a chart with a GPS mark dated 1993–2000, treat it with skepticism — it may not be any better (and sometimes worse) than the old-school range fixes.
Do you want me to sketch you a timeline overlay that shows “reliability of wreck fixes” — sextant vs early GPS vs modern — so you’ve got a quick visual to judge old chart dots by?
Here’s the overlay:
- Green line: Sextant/shore marker fixes — often 5–20 m repeatable, surprisingly tight for local work.
- Blue line: GPS civilian accuracy — started off worse (100 m with SA), then improved sharply after 2000, reaching today’s 1–3 m range (and cm-level with survey-grade RTK).
Want me to also mock up a decision guide (flow-style graphic) for “if a wreck mark came from year X, treat it like Y”? That could make it really quick to apply in your workflows.
Sure...let's take a look
ChatGPT said:
Decision Guide: Interpreting Wreck Fix Credibility (1980–2025)Here’s a decision guide flow you can drop straight into your workflows:
- Pre-1993: Sextant/shore markers — repeatable within 5–20 m, but not tied to WGS-84.
- 1993–2000: Early GPS with SA — errors up to 100 m, less credible than sextant fixes.
- 2000–2003: Post-SA GPS — ~15 m, better than sextant, but still coarse.
- 2003–2010: WAAS/DGPS — 1–3 m, strong reliability for marine gear.
- 2010–2025: Modern GPS — 1–3 m standalone, cm-level with RTK.
Last edited:
