A tide gauge built from reflected satellite signals.

Africa's coast is a hole in the sea-level map

Knowing how fast the sea is rising depends on measuring it, and Africa is one of the largest gaps in the global sea-level observing system. Tide gauges are sparse, unevenly distributed and of varying quality, and the East African coast is especially poorly instrumented — despite low-lying, erosion- and flood-prone shorelines. In Kenya, the port of Mombasa hosts one of the region's few long records, while the remaining ~530 km of coast — Lamu, Malindi, Kilifi, Diani, Shimoni — is effectively unmonitored.

Ground-based GNSS Interferometric Reflectometry offers a low-cost way to close that gap. A single geodetic station at the coast listens for the navigation signals that reflect off the sea and reads sea level from the interference between the direct and reflected paths recorded in its signal-to-noise data. It performs like a tide gauge — with daily-mean accuracy below 1.5 cm and around 350 water-level estimates a day when four constellations are tracked — and, crucially, it simultaneously measures the antenna's own vertical land motion.

The method: counting usable sea-reflection arcs

This study is a side-project of GWO's spaceborne GNSS-R work and reuses that validated orbit and constellation code, adding the ground-based observing geometry. For six candidate stations along the Kenyan coast, the simulation propagates all four constellations and, at every epoch, computes each satellite's elevation and azimuth in the station's local frame.

A reflection is counted as usable only when two conditions hold together — and each continuous arc through that window yields one sea-level estimate:

• Condition one

  Low elevation band

  The satellite sits between 5° and 25° elevation, where GNSS-IR sea-level retrieval is performed and the reflection grazes the water.

  Footprint ≈ antenna height ÷ tan(elevation)
• Condition two

  Seaward azimuth sector

  The satellite's azimuth falls within the station's open-water sector — the range of directions actually facing the sea rather than land.

  ~180° at open sites, ~160° where the coast curves in

The analysis is deliberately geometric: it quantifies coverage and sampling rate, not retrieval precision, which is taken from the published literature (< 1.5 cm daily-mean). The question it answers is the first one any deployment must clear — is there enough usable geometry, often enough, facing the right way?

What the simulation shows

All six sites are strongly feasible. The headline numbers below carry the finding; multi-GNSS is what lifts every station to tide-gauge-equivalent performance.

Where the network would sit

The six candidate stations span the coast from Lamu in the north to Shimoni near the Tanzanian border — co-locating one with the existing Mombasa gauge for cross-validation, and placing the other five along the stretch that has no record at all.

The sky plot makes the geometry concrete: at Mombasa the open-water sector (≈110° ±90°) is densely filled with low-elevation arcs from every constellation, so the sea surface is sampled across all offshore directions rather than along a single track.

Multi-GNSS is the whole story

The single most decisive result is how much the other three constellations add. GPS alone yields only ~70–83 measurements a day per site — useful, but well short of tide-gauge cadence. Stacking Galileo, GLONASS and BeiDou on top roughly quadruples the rate to ~270–310/day and shortens the median gap between estimates from about sixteen minutes to three.

Table I — Per-site feasibility

Site	Lat, Lon	Multi-GNSS /day	GPS /day	Median gap	Seaward coverage
Lamu	−2.27, 40.90	304	83	3.2 min	full (180°)
Malindi	−3.22, 40.13	299	75	3.2 min	full (180°)
Kilifi	−3.63, 39.85	279	74	3.5 min	160°
Mombasa	−4.06, 39.67	307	82	3.2 min	full (180°)
Diani	−4.28, 39.59	306	82	3.2 min	full (180°)
Shimoni	−4.65, 39.38	267	70	3.4 min	160°

Sites with a narrower open-water sector (Kilifi, Shimoni) lose only modestly. Every station clears tide-gauge-class sampling.

Why it beats a tide gauge, not just matches it

The network's real value is greatest precisely where Kenya is blind today — and it does one thing a tide gauge fundamentally cannot. A gauge measures sea level relative to the land it stands on; if that land subsides — a real risk in deltaic and reclaimed coastal ground — it silently over-reports sea-level rise. A GNSS-IR station measures both at once.

• Conventional tide gauge

  Relative sea level

  Reads the water against its own structure. Land motion is invisible to it, so subsidence and true rise are entangled in a single number.

  One measurement, two confounded signals
• GNSS-IR station

  Absolute sea level + land motion

  Measures the reflecting sea surface and its own vertical position — yielding absolute sea-level change and an independent subsidence estimate from one instrument.

  The key advantage for attributing flood risk

That separation is the crux for a coast where both processes are in play. Add ~3-minute sampling — fast enough to resolve tides and storm surges, not just long-term trends — and the network supports flood early-warning as well as climate monitoring, on infrastructure that is cheap to install and maintain.

What it would take to deploy

The study is a feasibility screen, not a deployment plan. Three steps stand between this result and stations in the ground.