Scientific programme

The season's scientific programme.

An open weather probe in the Drake Passage, compared against the ship that carries it: verification, observation, drift

A school team is building three autonomous meteorological probes of its own open design — an ESP32 microcontroller, a research-grade barometer, an Iridium satellite modem, solar-assisted battery power, a printed Ø200 mm sealed sphere — and operates them in the Drake Passage during the 2026/27 season. Two drifting buoys of one build, SEA-1 and SEA-2, are released together from the expedition vessel mid-passage, north of 60 °S — SEA-1 towing a drogue, an underwater sail about four metres down, and SEA-2 towing none; the third, VESSEL, is the same electronics without the hull or the satellite modem, aboard the ship on its WiFi for the whole voyage.

The programme answers a metrological question first — how accurate is a low-cost open instrument in polar conditions, when continuously compared against the ship's bridge barometer? — and applies the answer where observations are scarcest: sea-level pressure series from the Drake Passage with a stated uncertainty, a geostrophic wind estimate across the passage, and a measured answer to how much of a small buoy's drift is the ocean's and how much the wind's, from two drift tracks across the fronts of the Antarctic Circumpolar Current. All data are open under CC-BY 4.0.

1. Who we are

SouthPing is a private school science project, not a funded research programme, and this page says so plainly. The team is sufficient to execute the programme as designed; the institutions around it add legitimacy and reach, they do not substitute for the work.

Ildar Gabdrakhmanov — project lead and principal investigator

Engineer. System architecture, electronics, enclosure, power, deployment, and all institutional correspondence. Travels with the 2026/27 expedition and personally performs the release and runs VESSEL aboard.

Timur, 10th grade — student lead, software

Owns the probe firmware, the server-side processing and the web dashboard — the software half of the project, not an assistant role. He attends Founders Classical Academy of Prosper, Texas, USA.

Jason K. Brogden — faculty mentor

Dean of Science Instruction at Founders Classical Academy and Director of the Lynx Science Academy, Prosper, Texas. Guides the project's scientific method and the student's research, and has offered to help prepare the results for a journal for high-school researchers after the season.

The school

Founders Classical Academy of Prosper, Texas, USA, a public charter school, is the project's school: its Dean of Science Instruction mentors the project. There is no second school.

2. Background and motivation

The observing gap. The Southern Ocean is the least observed ocean on Earth, and the Drake Passage is its narrowest, stormiest strait. Ship and buoy observations there are sparse; forecast-sensitivity studies rank drifting-buoy sea-level pressure at high southern latitudes as the most valuable single observation type per observation in the Global Observing System, precisely because there is so little of it. The Data Buoy Cooperation Panel (WMO/IOC) explicitly prioritises barometers on buoys in this region. This is the network's own stated gap, not a school project's rationalisation.

The instrument question. Professional drifting buoys cost tens of thousands of dollars each. Barometric sensors of research grade now cost a few dollars, and the rest of a probe — microcontroller, satellite modem, power, hull — can be built openly for under a thousand dollars in parts. What is not known is whether such an instrument, built by students and calibrated honestly, produces data of measurement quality in polar conditions. Answering that is a genuine metrological question, and the answer determines whether the low-cost approach can be repeated every season by other teams.

The sentence that defines the mission: an open-hardware probe costing under a thousand dollars, verified against a ship's reference barometer, measuring sea-level pressure where almost nothing else does.

3. Scientific questions

Each question has an external, checkable answer — a reference the result is compared against — which is what makes this science rather than a demonstration.

Q1

Metrological — the anchor, and load-bearing: how accurate is a low-cost meteorological probe in polar conditions, and what corrections does it require, when continuously compared against the ship's bridge barometer?

Method
All three boards sit on deck beside the bridge barometer before the release, and VESSEL stays there for the whole voyage; the bridge reading is logged by hand at the same hours. The residual pressure difference is fitted as a temperature-dependent correction, p_corr = p_raw + a + b·(T_die − T₀), with a drift term if the residuals demand one.
Output
A per-unit correction function and an honest uncertainty budget: reference uncertainty, fit residual, transfer uncertainty and drift. This question carries the other three — without it the drifters' data are unattributable — and it is the only one that does not depend on a drifter surviving.
Reference
The bridge barometer's hourly log, kept by hand in the deployment log.
Q2

Meteorological — the highest external value: what is the sea-level pressure in the Drake Passage, where there are no observations?

Method
The two drifters report pressure eight times a day, every three hours, the standard synoptic hours (00/06/12/18 UTC) among them, corrected through the chain established by Q1.
Output
Sea-level pressure and sea-surface temperature time series from the passage with a stated uncertainty — valuable standalone, and directly assimilable if admitted to the WMO Global Telecommunication System.
Reference
ERA5 reanalysis, and any co-located ship or buoy reports.
Q3

Dynamical: how strong are the westerlies over the passage, and does the reanalysis reproduce them?

Method
The pressure difference between SEA-1 and SEA-2 yields the geostrophic wind — a wind speed measured by two barometers and no anemometer. The two are released together, so their baseline starts at zero and grows as the wind carries SEA-2 ahead; each drifter against VESSEL carries the first days, and the hours the pair spends side by side measure the noise of the pressure difference itself.
Output
A geostrophic wind estimate across the passage over the drifters' lifetime, compared against ERA5 10 m and geostrophic winds. Stated limitation: a two-point gradient over hundreds of kilometres is a coarse estimate, and the comparison is the result, not a claim of precision.
Reference
ERA5 reanalysis.
Q4

Oceanographic: how much of a small buoy's drift is the ocean, and how much is the wind — and where do the fronts of the Antarctic Circumpolar Current run?

Method
Each buoy's GPS track, three positions per message, gives its drift. SEA-1 tows a drogue — two crossed panels centred about four metres down — and follows the water at that depth; SEA-2 is the same hull with none, so the wind and the waves push it as well. Released together, the two part at the rate the bare hull outruns the drogued one, which is thereby measured directly, with the current common to both cancelled; over the season each buoy's drift, less the satellite-derived current, is compared with the reanalysis wind. The expected values are written down and dated before the release, so the result is a test and not a description. Hull-mounted temperature sensors on both buoys give sea-surface temperature; on SEA-1, crossing the Subantarctic and Polar Fronts should appear at once as a change in drift speed and a step in water temperature — two independent signals that cross-check each other.
Output
Each buoy's response to the wind, as a percentage of the wind speed and an angle to it; the current at SEA-1's depth; front positions along both tracks; and, if SEA-1 loses its drogue, the date it happened.
Reference
Satellite-derived currents (CMEMS), ERA5 wind, the published wind slip of undrogued drifters, and climatological front positions.

4. Instruments and methods

Three probes of one design, all built from the same open-hardware set. SEA-1 and SEA-2 are expendable drifting buoys of one build, released together from the expedition vessel mid-Drake Passage, north of 60 °S; VESSEL is the same electronics without a hull or a satellite modem, aboard the ship on its WiFi.

The three probes

SEA-1 and SEA-2 are released together, at one stop of the ship about six hours of its track before the 60th parallel. They differ in one thing: SEA-1 tows a drogue and SEA-2 does not. Both answer Q2 to Q4 and transmit until their power is exhausted. VESSEL is the reference leg for Q1: it sits beside the ship's bridge barometer for the whole voyage, and the drifters sit there with it on deck before they go over the side. Neither drifter comes back; VESSEL comes home with the team.

Hull

A Ø200 mm sphere printed in ASA, 3 mm walls, two shells bonded and bolt-clamped at an equatorial flange, with a hydrophobic membrane vent for pressure equalisation. All-up mass is about 1.9 kg with ballast, floating about half submerged. The design is verified by a 48 h submersion test before any hull is deployed.

The drogue

SEA-1's drogue is two crossed tarp panels on a plastic-pipe frame, 158 cm tip to tip, hung on a 3 m elastic line so that its centre sits about 4 m below the hull — modelled on the drogue of Pacific Gyre's Microstar drifter and scaled up. By calculation its drag area, taken face-on, is 40 to 48 times that of the hull and line together; drifter practice asks for more than 40, so it sits at the rule, with nothing to spare at the low end of the estimate. It is not yet built or tested: five tests gate it, and if it is not ready both buoys are released bare. It follows the water at 4–5 m, not at the 15 m of a standard drifter, and that offset is a stated limitation. SEA-2 carries no drogue and no weight in its place, so it floats about a centimetre higher.

Sensors

A Bosch BMP581 barometer (±0.3 hPa) is the mission pressure sensor, with its die temperature feeding the Q1 correction; a BME280 measures air temperature and humidity inside the hull; a waterproof DS18B20 (±0.5 °C) measures sea-surface temperature from a sea-flooded pocket in the lower shell; a u-blox SAM-M8Q receiver supplies position and UTC.

Link and power

On the drifters a RockBLOCK 9603 Iridium Short Burst Data modem carries the data. Readings are taken hourly and held in a five-day buffer against failed transmissions, then sent as compact binary messages: each drifter eight times a day, every three hours with the synoptic hours among them; VESSEL over the ship's WiFi at a cadence its own design sets. The transmission schedule is designed to downshift when the battery runs low, so a poor week degrades the data rate rather than the mission. Power is a six-cell 18650 lithium pack, solar-assisted by a small panel on the crown; the mission is sized to succeed without solar yield, because a Southern Ocean deployment cannot count on the sun, and the charger itself stops charging below freezing to protect the cells.

Data handling and engineering basis

Raw messages are never rewritten. Every derived product — corrected pressure, sea-level reduction, geostrophic wind, drift velocity — is computed from the raw archive with versioned code, so any result can be recomputed from first principles, and quality-control flags are stored beside the data rather than applied to it. The design, the firmware, the message contract and the server-side decoder are in one open repository, with a test suite that keeps the encoder and decoder byte-identical against shared fixtures.

5. The traceability chain

Ship's bridge barometer (logged by hand, hourly)
        │   the whole voyage, side by side
        ▼
     VESSEL ──── on deck together before the release ─────►  SEA-1, SEA-2
        ▲
        └──── a fourth BMP581, the bench witness: never travels
A drifter at sea has nothing to compare itself to, so confidence in its data has to be carried to it. Stage 0, before departure, is the bench co-run: the three boards side by side with a fourth barometer that stays home as the bench witness, through a temperature sweep from −18 to +25 °C — the two-week co-run planned for October did not fit the build, and what remains is a co-run of at least 48 h with one temperature cycle. Aboard, the three boards sit on deck beside the ship's bridge barometer before the release, so SEA-1 and SEA-2 carry their offsets against it into the water. Stage 1 is VESSEL beside the bridge barometer for the whole voyage, its reading logged by hand at the same hours, giving the absolute anchor and the correction function.

6. Deployment plan and timeline

The schedule is set backward from the vessel's departure, which is not under the project's control.

WhenWhat
September – October 2026Done by 3 November: the carrier boards built and tested; the firmware and the server working end to end, with probes transmitting from the yard into the backend; the hull design finished; the charger's behaviour in the cold and the power budget measured. Not done: the enquiry to OceanOPS.
3–14 November 2026Hulls printed and sealed; the submersion test; Calibration Stage 0 at the minimum the traceability chain states; the outdoor end-to-end run with real satellite transmissions; VESSEL's firmware and ingestion. SEA-1's drogue built and put through its tests. Packing and spares. The OceanOPS enquiry goes out.
16 November 2026Vessel departs Ushuaia.
The crossing, from 16 NovemberThe three boards on deck beside the bridge barometer, sending. The drift prediction written down and dated on 16 November, before the release. SEA-1 and SEA-2 released together, about six hours of ship's track before the ship reaches the 60th parallel. Deployment log: UTC time, position, sea state, the bridge's wind reading then and hourly through the first day, photographs with each buoy's waterline, serial numbers; whether SEA-1's drogue went into the water, what it was, and that it hung on its line, uncut and attached to SEA-1, with a photograph of it going; the minute each buoy went in, and how far the ship moved between the two.
November 2026 – February 2027Season. SEA-1 and SEA-2 report eight times a day; VESSEL from the ship for the voyage. Live dashboard.
End of seasonThe drifters transmit until their power is exhausted; VESSEL comes home with the team.
March 2027 onwardCalibration fit, derived products, quality control, published dataset. WMO identifier and GTS applications continue.

Provenance: the deployment log above, agreed with the vessel's captain aboard. Without it the dataset has no provenance.

7. Data and publication

All data are open. The dataset is published openly and in full under CC-BY 4.0, in CSV and CF-compliant NetCDF, together with the calibration record and the uncertainty budget.

Anyone can follow the probes in real time on the Monitor page; the raw archive and the dataset are open.

GTS. The Q2 pressure series is intended for the WMO Global Telecommunication System, by one route: the US route for the two drifting buoys through NOAA/NDBC as the US representative to the Data Buoy Cooperation Panel. Admission to the GTS is amplification, not a precondition: the series is published regardless.

Instrument design, firmware and analysis code are in an open repository — code under Apache-2.0, and documents, drawings and data under CC-BY-4.0.

8. Expected results

A quantified accuracy characterisation

An open, low-cost meteorological probe measured in polar conditions over a whole voyage and traceable to a ship's reference barometer: a published correction function and uncertainty budget.

Two pressure series from the Drake Passage

Sea-level pressure and sea-surface temperature from the passage, one series per drifter, with a stated uncertainty.

A geostrophic wind estimate

Across the passage, from the pressure difference between the two drifters, and its comparison with ERA5.

Two drift tracks, one drogued and one bare

The wind's share of a small buoy's drift, measured against a prediction written before the release; the current at the drogue's depth, compared with satellite-derived currents; and the positions of the ACC fronts from drift speed and water temperature.

An open, citable dataset

Published under CC-BY 4.0, with the instrument design and code in an open repository.

A student who has taken a measurement from a design to a published dataset

With a professional reference standing behind it.

9. Risks and honest limitations

Two properties make the plan robust. Q1 does not depend on a single buoy: VESSEL rides with the team, and SEA-1 and SEA-2 are two independent instruments, so a drifter lost on day three still leaves a reference-anchored result. And Q2 does not depend on admission to the GTS — the pressure series is valuable and publishable on its own.

RiskWhat we do about it
One drifter fails earlyThe other carries Q2 alone, and Q3 falls back to the surviving drifter against VESSEL. Q4 loses its comparison: what remains is one buoy's response to the wind, checked against published figures rather than against the other buoy.
The hull leaksA bonded and bolt-clamped flange, a leak sensor, a submersion test article, and a 24 h confirmation dip before departure.
SEA-1's drogue is lost, or never worksShown by the data: SEA-1 begins to answer the wind the way SEA-2 does, and a pair that does not part in the first days has no working drogue. The bare hull's wind slip and the date of the loss remain as results; Q1 to Q3 are unaffected.
The drogue is not ready by departureBoth buoys are released bare. Q4 falls back to the bare hull's response to the wind and the temperature fronts; Q1 to Q3 are unaffected.
Calibration Stage 0 is shortThe two-week co-run planned for October has become a bench co-run of the three boards of at least 48 h plus one temperature cycle before departure, followed by the deck co-run beside the bridge barometer before the release; VESSEL's voyage beside the bridge barometer supplies the temperature coefficient over a wider range. Below that minimum, the drifters' data are unanchored, and this page says so.
PermitsThe release lies outside the Antarctic Treaty area.

Limitations carried into the results rather than hidden: the two-point gradient of Q3 is coarse; Q4 rests on one pair of buoys, with no replication; SEA-1's drogue sits at 4–5 m, not at the 15 m of a standard drifter; a partial loss of the drogue — a torn panel, an arm out of its hub — would probably not show in the data; wind and waves push a bare hull together and cannot be told apart here; the wind in that comparison is a reanalysis's, not a measurement at the buoy; the sea-surface temperature sensor's absolute accuracy serves front detection rather than absolute temperature; a Ø200 mm hull in Southern Ocean swell spends much of its time submerged, so per-attempt transmission success will be well below bench figures — which is what the retry logic and the buffer exist for; and solar yield at sea is expected to be far below bench values.

10. Long-term perspective

The programme is designed to be repeatable. An open, low-cost instrument can be built and released every season, extending the series and bringing in new school teams each year, and each season's probes are anchored the same way, against the ship that carries them.

Season 2026/27 is the proof of concept; the team is ready to design a multi-season programme with any partner the result interests.