AT-AT · Calibration control and recovery pilot

ICA + HYCSA · 121 control vertices · 12 source groups · 15 October 2026

Executive summary

Provisional recovery shortlist

This is a proposed field-recovery plan, not a list of approved checkpoints or base stations. Begin with ICA vertices 3217, 3210, 3186 and 3185 around KM30–39. If those recoveries are useful, expand to four more ICA points and six HYCSA points in the Santa Clara and T3 areas, planned as separate visits. Investigate vertex 3155 separately because its two workbook UTM entries differ by 24.84 mm. Candidates were selected for geographic spread, shared calibration membership and proximity to the drawing—not confirmed access, monument stability, sky visibility or survey accuracy. Every point must be recovered, identified and independently surveyed before its intended role is accepted.

Before accepting a recovered point

For any role: confirm the exact vertex and marker with the survey crew, photograph and inspect it, establish safe access, obtain its survey record and independently verify XYZ, uncertainty, coordinate frame, epoch and height reference. Resolve the UTM / planas calibration. Workbook coordinates alone do not establish approved control.

As a checkpoint: identify the exact surveyed center in multiple images, or independently survey a visible nearby target. Withhold checkpoints from model adjustment and calibration; validate their reference coordinates independently of the flight-base solution. Distribute checks across the area and evaluate against the intended deliverable accuracy.

As a temporary RTK base: verify stability, centering, sky visibility and multipath; establish an accepted geodetic position including ellipsoidal height; document antenna height and reference point; check correction coverage and separate control, and retain raw observations. Do not enter planas as the geodetic base position.

The ICA workbook provides no point elevations. HYCSA source Z values have an unspecified vertical datum; zero Z values are unavailable. No permanent CORS installation site is approved by this shortlist.

Recover, observe, solve, verify

Establish a defensible reference position for each recovered control using the Emlid Reach RS2+ and NRCan CSRS-PPP. Recovering a physical mark and accepting its coordinates are separate decisions. This is the proposed project protocol; the durations below are planning targets, not service guarantees or proof of accuracy.

Emlid advises at least four hours for PPP and allows longer sessions according to the accuracy needed. The 6–8-hour initial and 4–6-hour repeat targets are our conservative pilot planning choices. See Emlid’s PPP logging guidance.

1 · Prepare the recovery visit

Take the point’s dataset, vertex ID, project label, source UTM/planas, approximate map location and discrepancy notes. Request the original monument record and photographs. Obtain access, assess construction activity and traffic, and agree the horizontal/vertical accuracy targets and decision criteria for its intended role before surveying. Check battery, storage, tripod/tribrach, centering optics, level calibration, tape and weather protection.

Keep the original coordinate files unchanged. Record the mark name exactly; a nearby nail, painted target or replacement monument is a different point until its relationship is surveyed. For 3155, retain both source entries and resolve the discrepancy through observations rather than selecting whichever matches best.

2 · Recover and document the physical mark

Confirm identity with the survey crew. Photograph the inscription and precise survey center, a close view with scale, the approach, four horizon directions and the final antenna setup. Describe material, cracking, looseness, settlement, construction risk and obstructions. Record an access sketch and measured ties to stable features. A BN/GPS/RT/PA label alone does not establish monument type or stability. If identity is uncertain or the mark is damaged, record unconfirmed / unsuitable and stop its acceptance process.

3 · Set up the RS2+ and verify antenna height

Use a stable tripod/tribrach or properly braced setup, precisely centered and leveled over the identified center. Avoid vehicles, fences, metal roofs, reflective surfaces and growing obstructions. Measure the vertical distance from the survey center to the receiver’s bottom antenna reference point, following the selected Emlid logging workflow. Record the actual value, units, reference point and measurement method. Check centering, level and height before and after the session; investigate any movement or height change instead of averaging it away.

Set the actual point name and measured height in the RINEX logging information. Preserve receiver model, serial number and firmware, plus antenna/radome identification. Verify the exported RINEX antenna fields and height against the field sheet. Confirm which physical point the final PPP coordinates refer to. Do not add or subtract 0.134 m by habit: the project’s CAL8 note is not a universal correction, and antenna-model offsets must not be applied twice. See Emlid’s antenna-height conventions.

4 · Log an uninterrupted static session

Use Emlid Flow’s appropriate CSRS-PPP logging preset/settings and a RINEX version supported by the service; Emlid recommends RINEX 3.03 when unsure, subject to service support. Enable UBX backup. Use a 1 Hz raw archive as the project default where supported; retain all available tracked constellations and dual-frequency observations. CSRS-PPP uses its supported signals; do not strip the original archive to match an older processing recipe.

Confirm that recording has actually started and file size/elapsed time are increasing. Note start/end times with time system explicitly identified, horizon conditions, interruptions, receiver resets, weather and nearby activity. Use a stable external-power arrangement if needed and secure the site. End logging normally, wait for the file to close, download it, verify its duration and make a second copy before clearing receiver storage. Do not treat the displayed real-time position, FIX badge or a short SINGLE average as the accepted PPP reference. Emlid logging and export instructions.

5 · Process with the Canadian CSRS-PPP service

Inspect the RINEX before submission: actual epochs, data gaps, observation types, approximate location, antenna model/radome and antenna height. Submit each occupation separately in static mode to NRCan CSRS-PPP, using its international ITRF output for the Mexico project. Confirm the returned realization, coordinate epoch and height convention in the report; do not relabel the output as generic WGS84 or as planas.

For project acceptance, review a Rapid-product result and reprocess with Final products when available. An Ultra-rapid result can be an early diagnostic. NRCan currently reports Rapid availability 17–18 hours after the observation day ends and Final availability 12–15 days after the week ends; product availability can change. Save the report, summary, position files, submission settings and exact RINEX submitted. Record the product type actually used. NRCan product, reference-frame and output documentation.

6 · Review quality and compare independent sessions

Review reported uncertainties and their confidence level, convergence, data completeness, residuals, rejected observations and antenna warnings. Short, noisy or poorly converged sessions remain on hold. A processing-success email or small formal sigma is not an independent accuracy test. A split of one long file can expose instability, but is not a replacement for a separate setup and occupation.

Compare the independent solutions as east, north, horizontal magnitude and ellipsoidal-height differences in a common reference realization and epoch. Compare Rapid and Final results, and investigate disagreements larger than the pre-agreed project tolerance or inconsistent with the reported uncertainties. Do not average solutions across unresolved shifts. Where feasible, check against independently established control or a verified CORS solution; separate occupations reduce repeatability risk but can still share antenna-model, frame and environmental biases.

Keep a geodetic master record with XYZ or latitude/longitude/ellipsoidal height, realization, epoch, antenna convention and uncertainty. Separately document the transformation to the project’s coordinate system. An inverse/forward UTM projection alone does not perform a reference-frame or epoch transformation. Do not convert PPP ellipsoidal heights into construction elevations without an accepted vertical reference/geoid relationship. The workbook’s assumed WGS84 and HYCSA source Z values remain unverified until reconciled.

7 · Accept a specific role, then use the point

Checkpoint: require a stable, unambiguous, image-identifiable center or a separately surveyed visible target. Keep it out of the photogrammetric adjustment. Its coordinate uncertainty must support the intended validation accuracy, with an independent reference check so a shared flight-base error cannot be hidden.

Temporary RTK base: require repeatable centering, adequate sky and correction coverage, accepted geodetic coordinates and the documented antenna-height convention. Recheck height on every setup; use the correct reference point in Emlid Flow and the drone workflow, log raw data throughout the flight and verify a separate check mark. With one RS2+, complete the checkpoint occupations beforehand or use a separate independently referenced survey arrangement; the receiver cannot be both local base and roaming checkpoint rover at the same time.

Permanent CORS: a recovered point may support local ties and checks, but permanent-site stability, sky, power, security, communications and long-term monitoring require a separate design and commissioning exercise. PPP acceptance of one occupation does not commission a CORS.

An optional differential-static cross-check can use RS2+ and verified reference-station observations in Emlid Studio: local observations in Static receiver, reference observations in Base, with overlapping data, navigation, authoritative reference coordinates and correct antenna heights. Review session behavior rather than only the single best epoch. Emlid static-processing workflow.

Field record and acceptance gates

Methodology prepared 18 October 2026. Session durations are planning allowances, not certified accuracy classes or a substitute for the project survey specification. Retain raw observations and processing history; revise the plan when field conditions or results require it.

Data and field files

Downloads are generated from data embedded in this block. Keep the QGIS project and its GeoJSON layers in the same folder. Basemaps require an internet connection.