7.4 GPS/GNSS and Field Data Collection

Key takeaways

• GPS is one of several GNSS (Global Navigation Satellite Systems) — GLONASS, Galileo, BeiDou are others.
• Accuracy ranges from 10 m (consumer) to 1 cm (RTK survey) — understand which tier your project needs.
• Field collection tools (Mergin, QField, ArcGIS Field Maps, Fulcrum) structure data entry for reliability.

Introduction

Satellite imagery and open data cover most GIS inputs, but sometimes you need to collect your own observations — a field biologist counting nests, an engineer locating utility poles, a disaster responder mapping damage. This lesson covers GNSS fundamentals and field data collection workflows.

GNSS basics

GNSS = Global Navigation Satellite System. Four operational constellations:

• GPS (USA, 1978–) — 31 satellites.
• GLONASS (Russia, 1982–).
• Galileo (EU, 2016–).
• BeiDou (China, 2018–).

A modern receiver uses multiple constellations simultaneously to improve accuracy.

How it works

A GNSS receiver:

1. Picks up signals from visible satellites.
2. Measures time-of-flight to each.
3. Triangulates its position from at least 4 satellites (3 for position, 1 for clock).
4. Reports latitude, longitude, and sometimes altitude.

Accuracy tiers

Tier	Typical accuracy	Hardware
Consumer GPS	3–10 m	Smartphone
Dual-frequency consumer	1–3 m	Phone with L1+L5 support
DGPS / SBAS	0.5–2 m	Augmented consumer
Mapping-grade	30–50 cm	Trimble R-series, Emlid Reach
Survey-grade RTK	1–3 cm	Survey receiver + base or NTRIP
Post-processed kinematic (PPK)	1–5 cm	Survey receiver + post-processing

Choose accuracy to match your question. For mapping individual trees in a park, 1 m is fine. For cadastral survey, centimetre accuracy is required.

Augmentation systems

• SBAS (Satellite-Based Augmentation — WAAS US, EGNOS EU, MSAS Japan) — improves consumer accuracy to ~1 m.
• RTK / RTCM — base station broadcasts corrections in real time via radio or NTRIP over cell.
• PPP — Precise Point Positioning; uses precise orbit/clock products, typically post-processed.

Factors degrading accuracy

• Multipath — reflections from buildings confuse time-of-flight measurements.
• Canopy / heavy cover — signal attenuation; forest-floor readings are erratic.
• Indoors / urban canyons — few satellites visible; dead reckoning takes over.
• Ionospheric activity — varies with solar cycle.
• Cold start — a receiver with no recent fix can take minutes to acquire.

Best practices: wait for a stable fix, record dilution of precision (DOP), use external antennas for precision work.

Field data collection tools

Mergin Maps + QField

• Open source, QGIS-aligned.
• Forms, validation, offline support.
• Sync to Mergin cloud or self-host.

ArcGIS Field Maps (formerly Collector)

• Esri ecosystem.
• Deep integration with ArcGIS Online/Enterprise.

Fulcrum / Survey123

• Form-first; well suited for asset inspections.

Custom apps

• Mapbox / MapLibre + React Native for bespoke flows.
• ol-mobile, leaflet.mobile for lightweight web captures.

Designing a field form

Good forms:

• Constrain input — dropdowns, not free text.
• Validate on the device — reject out-of-range values immediately.
• Capture photos + metadata — GPS tag, timestamp, compass bearing.
• Autofill — user, device, weather from API.
• Offline-capable — no cell service is a given in the field.
• Short — the more fields, the more errors.

A sample workflow

1. Define schema — what attributes do you need?
2. Build a form in QField / Field Maps.
3. Test it (in your kitchen, then in the field) — identify usability problems.
4. Train field team.
5. Collect.
6. Sync to central store (PostGIS, cloud service).
7. QA / QC — check for impossible coordinates, missing required fields, duplicates.
8. Integrate with analytic dataset.

Privacy and consent

When collecting locations of people (residences, behaviours) consider:

• Data minimisation — collect only what you need.
• Precision reduction — H3 cell, hex grid, or randomised jitter rather than exact address.
• Consent from participants.
• Compliance with local privacy law (GDPR in EU, HIPAA for US health data).

Self-check exercises

1. Your iPhone reports a location with 5 m accuracy. You need 10 cm. What hardware do you need?

A survey-grade RTK GNSS receiver (e.g., Emlid Reach RS2, Trimble R12) plus a base station or NTRIP subscription. The receiver must be dual- or multi-frequency (L1/L2/L5) and pair with corrections to achieve centimetre fix. Consumer smartphones top out around ~1 m even with dual-frequency hardware.

2. Why does GPS accuracy drop in a dense city?

Multipath: buildings reflect signals, so the receiver mixes direct and reflected arrivals and miscalculates time-of-flight. Urban canyons also hide satellites low on the horizon, reducing the geometry quality (high PDOP). Using multi-constellation receivers (GPS + Galileo + GLONASS + BeiDou) recovers some satellites; dead-reckoning (IMU-assisted) helps in gaps.

3. Your field team's data has strings like "Yea", "ye", "Yes" for the same intended value. What should have been done?

The form should have been built with a constrained input — a dropdown or boolean toggle — rather than a free-text field. On-device validation and constrained types prevent free-text variability. For existing data, a normalisation pass with fuzzy matching + rules is needed.

Summary

• GNSS = GPS + GLONASS + Galileo + BeiDou; modern receivers use them all.
• Accuracy tiers span 10 m (consumer) to 1 cm (RTK survey).
• Field forms should constrain input, validate on device, and sync reliably.
• Plan for multipath, canopy, privacy, and offline from the start.

Further reading

• GPS.gov — public GPS system documentation.
• Emlid blog — accessible RTK tutorials.
• QField documentation — field GIS on phones / tablets.
• El-Rabbany, A. — Introduction to GPS: The Global Positioning System.