Dronų žemėlapis
Your areas
Alerts come only while this page is open in a browser: not continuous monitoring.Map
Legend
Now
Protected zones
Airports, air bases and military training areas (outlines from OpenStreetMap). A zone counts as
reached when a model balloon passes within km of its outline (Passes over) or comes down
there (Comes down). Same levels as the places. Dashed outlines on the map; they turn amber or red when reached.
Reports and official statements
Balloon and drone reports from the news (LRT, 15min, LSM) and official statements (VSAT, Latvian border
guard, Lithuania's defence ministry, Estonia's Defence Forces, Latvia's defence ministry, Polish police); last 14 days,
checked every 15 min, placed at the places they name. Status comes from the item's own words: reported,
official statement, investigating, confirmed or corrected (denied or withdrawn). The date is
the publication's; when the text names a date in the 14 days before, it is shown as the incident's.
Red diamonds: balloon reports; orange triangles: drones. Smuggling balloons carry no transponder, so they are never in
the aircraft layer.
Balloon reports by area and quarter
Counts of reported balloon events (the incident list from Sept 2024 plus the live news reports), not of
balloons: one event can be many balloons, and reporting changes with attention. A drop in one area is worth
checking against a rise in the next: smuggling may have moved, not stopped.
Hour by hour
Each source over the last 7 days (grey bands: night). Shares are the aircraft with poor GPS among
those passing; Kaliningrad's reach is how far from it aircraft at 8 km or higher lose GPS. Hover for the numbers.
Kaliningrad's schedule
Kaliningrad has two modes: a normal reach (dashed yellow ring on the map) and a long, mostly
night-time reach (solid when on, dotted when off). The long reach follows a fixed nightly schedule: on at about
22:00 Kaliningrad time on most days checked (16 days, March–November 2026), strongest around 01:00–03:00, gone by
about 08:00. A GNSS base station near the Neman, about 23 km from Kaliningrad, lost signal on every band from 21:54
to 07:57 on 3–4 Nov: it reaches the ground there too. Raid nights were not stronger: it is a standing schedule.
The Gulf of Finland and Estonia have their own source. Aircraft do not fly over Russia, so it is placed from one
side: at night, three quarters of the GPS losses over Estonia fit one ring around the Ust-Luga port area on the
Gulf of Finland coast, about 20 km from the Estonian border at Narva-Jõesuu (ringed marker; the port is also named
by Windward and Finland's Coast Guard). Pskov fits clearly worse. The hatched cells over Estonia are where aircraft
lose GPS at the edge of its reach, not a jammer in Estonia. Over the sea west of Saaremaa at night it is
Kaliningrad's long reach. By day the losses do not fit a single place: more than one source is likely on.
Spoofing of aircraft
Jamming makes an aircraft report poor accuracy; spoofing makes it report a false position with good
accuracy, so the jamming counts miss it. Aircraft are flagged when their own ADS-B position jumps faster than
1,500 km/h, or stays frozen or circles within 2.5 km while they report cruise speed and height (OPSGROUP's 2024
report: over eastern Latvia and Lithuania aircraft were spoofed to Smolensk, as a fixed point or small circles).
The area east of the map's circle is watched too. Red rings: where a spoofed aircraft really was just before;
red crosses: the false position it reported (dashed line between, last 6 h). The transmitter is usually not at
the false position. Multilateration positions, computed by ground receivers, are not used: they cannot be spoofed.
Weather balloons
Radiosondes launched by the weather services (Kaunas daily about 13:30 local; Legionowo, Łeba,
Tallinn, Riga and others), as received by amateur stations and shared on SondeHub. They are real latex balloons on
known flights: a free daily check of the winds this forecast uses. Dashed lines: flights of the last 3 days.
Filled dots: in the air now.
All places
Unusual = how this night ranks among all nights since
. Normal = nights a year at Prepare and at High in this view.
Second opinion
The top district of each oblast for the first night, flown by the SondeHub predictor (Tawhiri,
its own copy of GFS at 0.5°), launch 21:00 local, position after 4 h. Large gaps mean the winds are uncertain.
Balloon forecast
- Two views. Passes over: model balloons float for up to 8 h; a place counts when one passes within 25 km. Use it for sensors and radar. Comes down: the same districts flown with the flight profiles that land as far inside the border as the reported finds do (median 26 km): slow latex balloons that burst at 7–10 km, or floats that sink after 1–2 h. A place counts when one lands within 25 km. Use it for patrols and pick-up spots.
- Levels. Passes over: Prepare ≥ 10 %, High ≥ 30 % of a district's model balloons. Comes down: Prepare ≥ 5 %, High ≥ 20 %. Two levels on purpose: "prepare" is for patrol planning and attention, not for closing an airport.
- How well they fit the past (balloons from Belarus, Sep 2024 – Oct 2026, launches spread over each area). Passes over: the reported place was at Prepare or higher on 89 % of reported place-nights, a level reached on about half of all nights; High on 74 %, about a third of nights. Comes down: Prepare on 75 % (31 % of nights), High on 50 % (17 % of nights). With town launches it was 90/78 % and 73/49 %: the same skill. Comes down is sharper but misses more; it also says short flights rarely land far inland (Panevėžys about 4 nights a year).
- This is the wind, not a detection. It says where balloons can arrive tonight if someone launches them, not that anyone will.
- Reported balloons so far came from Belarus: Hrodna oblast to Vilnius and the south, Vitebsk oblast (Braslaŭ, Miory, Verkhnyadzvinsk) to Latvia and north-east Lithuania. No reported balloon fits a Kaliningrad wind; that column is a watch, not a pattern.
- On the past nights with reports, this model gave the reported place a reach in the top fifth of all nights in most cases. A night without reports is not a night without balloons.
- Weather balloons come from SondeHub (amateur receivers). The page asks it for sondes within 450 km of central Lithuania once a minute while it is open; the recent flights are the ones our tracker saved. Forecast wind error = how far the forecast wind at 5–11 km was from the wind the sonde drifted in.
- Tested on 293 real weather balloons (Sep–Oct 2026): with the right flight profile, these winds a day ahead put the landing a median 5 km from the real one (90 % within 11 km) after a 2-hour flight. A 20 % wrong burst height doubles that, so for smuggling balloons the unknown fill matters more than the forecast.
- Model: GFS 1° winds; launches from anywhere in each border district (a 15 km grid of points over its area, up to 30 km from its town, not from the town itself: smugglers launch from fields and forests), every two hours 17:00–01:00 local. Passes over: ascent 4 m/s, float at 1.2–11 km. Comes down: four profiles fitted on the finds, not on any known balloon fill, so read landing spots as an area tens of kilometres wide; the night and the direction are what is solid.
GPS jamming
- Aircraft from open ADS-B (adsb.lol), modern (version 2) transponders only. Poor GPS = position accuracy (NACp) below 8, or no accuracy at all after a good fix in the last 2 hours. Aircraft that never report good accuracy are left out as equipment, not jamming. The live server and the back-filled history use the same rule.
- Yellow dots: where aircraft lost or regained GPS (last 24 h). Hatched cells: where aircraft lost GPS in the last hour (their last good position: a degraded aircraft carries the fault along its route, so where it lost GPS says more than where it is now).
- Ringed yellow marker: an estimated jammer, with its likely area and, dashed, how far it reaches aircraft at 10 km. Where aircraft fly all around a source it appears once enough points agree (at least 15 that surround it, likely area under 80 km). The source east of Estonia is fitted from one side (no aircraft over Russia): from night-time losses over Estonia and the Gulf of Finland, trying every place to the east and keeping the one where most losses fit one radio-horizon ring (live once there are 80 night-time losses in 14 days, else the 16-day result: Ust-Luga area, 76 % of 687 losses).
- Yellow squares: jammer sources named in official and research reports (Latvia's regulator, Gdynia Maritime University, Finland's Coast Guard); a dashed square and dotted ring mean only the area or oblast is known.
- Last night / Latest day: the share of aircraft time with poor GPS per cell, night (20–06 UTC) and day (08–17 UTC) kept apart. A daily map, as published elsewhere, is mostly a daytime map: on 3 Nov only 16 % of aircraft time over this area was at night. Empty means too few aircraft: Belarus and western Russia have almost no civil traffic, so the coloured band stops at the border even where the jamming does not.
- Ground checks: EUREF and Centipede GNSS stations' signal strength. Detection only: nothing here jams, spoofs or decodes anything.
How sure are the signals
Every signal on this page has a test record. A new rule appears only after a test; until then it is
logged for research or marked "learning". Inferred signals interpret measurements, they do not measure a cause.
Sources and coverage now
Each source's state and age, where it can see, the least data it needs to say anything, and what it
means when it is missing. "Nothing seen" from a missing source is not "nothing happening".
Sources
- Winds: NOAA GFS. Second opinion: SondeHub predictor (Tawhiri). Weather balloons: SondeHub.
- Aircraft: adsb.lol (open data, ODbL). Weather radar: RainViewer. Ships: Fintraffic Digitraffic AIS.
- Air-raid alerts (logged only): Ukrainian public alert map, ubilling.net.ua.
- News: VSAT, LRT, 15min, LSM, Latvian State Border Guard.
- Drone zones: ANS UTM (utm.ans.lt), a copy saved by hand; zones change by notice, so check its date.