AI-supported early warning systems for natural hazards and critical infrastructure

Self-powered sensors measure where debris flows, floods and slope movements begin. GeoSenseAI® combines the measurements into an up-to-date situation overview – as a basis for decisions by municipalities, authorities and infrastructure operators.

Natural hazards announce themselves in the catchment – not only at the protective structure

Heavy rainfall, snowmelt, rising water levels and sediment movement often change a catchment hours before an event reaches settlements or transport routes. If you only measure on the valley floor, you see this development late.

That is why GMD® starts where hazard processes begin: with self-powered sensors in the channel, on the slope and on the rock face, which continuously transmit readings to GeoSenseAI®.

How monitoring and early warning work together

approx. 20%
of all buildings in Austria are located in areas potentially at risk from natural hazards. Source: BMLUK
approx. 12,500
torrent catchments are listed in the register of the Austrian Torrent and Avalanche Control service. Source
over €340 bn
in losses from weather and climate extremes were recorded in the Alpine countries Germany, Austria, Switzerland and Italy between 1980 and 2023 – more than 40% of losses across all 38 EEA countries in Europe. Source: EEA

Impact of early warning · Worked example

Early warning can prevent up to 50% of damage

Weather-related natural catastrophe losses worldwide, USD bn per year

100%no warningBaseline

80%GMD conservativeGMD planning assumption

70%24 h warningWMO / GCA

50%GMD complete systemGMD target value

Weather-related losses worldwide in USD bn per year
ScenarioRemaining lossesAvoided
No warning244.70
GMD conservative (–20%, planning assumption)195.848.9
24 h warning (–30%, WMO/GCA)171.373.4
GMD complete system (–50%, target value)122.4122.4

up to USD 122 bn

in losses avoidable per year if early warning works as in the GMD complete system (–50%, target value). According to WMO/GCA, just 24 hours’ warning can cut damage by up to 30% – around USD 73 bn.

  1. Detect Sensors measure in the catchment
  2. Alert Municipality and emergency services
  3. Act Drone verifies · closures, evacuation
  4. Prevent damage fewer casualties and losses

In the EU, only around a quarter of losses from extreme events are insured – in some countries less than 5%. Every euro avoided mainly relieves governments, municipalities and businesses.

Basis: inflation-adjusted 10-year average of natural catastrophe losses USD 266 bn p.a. × 92% weather-related (2025 share) = USD 244.7 bn (Munich Re, January 2026). Reduction of –30% with 24 h warning (Global Commission on Adaptation, cited by WMO/UN); –20% and –50% are GMD planning assumptions. Insurance share: EIOPA, ECB. Simplified example assuming area-wide early warning.

From reading to decision in three stages

Early warning is more than an alarm signal. The system supports those responsible from the ongoing overview through to documentation after an event.

  1. Continuous

    Nowcasting

    Water level, precipitation and movement data are condensed into a traffic-light status for each catchment. This allows a pre-warning level to be triggered early and emergency services to be prepared.

  2. When defined thresholds are exceeded

    Early warning

    GeoSenseAI® automatically notifies the designated recipients – via app, SMS or email. The responsible bodies decide which measures follow.

  3. After the event

    Situation overview

    Drone surveys and terrain comparison show what has changed and which buildings and transport routes are affected – without sending staff into the danger zone.

GMD® systems provide measurement data and decision support. They do not replace official warnings or the assessments and operational decisions of the competent authorities and emergency services.

Hazard processes we monitor

Each process needs its own measured variables. The colour identifies the hazard process – consistently across the website and in GeoSenseAI®.

  • Debris flow with boulders and driftwood in an alpine torrent with check dams

    Debris flows and torrents

    Water level, precipitation and vibrations in the channel to detect sediment transport and early stages of debris flows.

  • Flooding mountain stream at a wooden bridge in an Alpine village

    Heavy rainfall and floods

    Water levels, flow velocity and precipitation in tributaries and urban streams – including the detection of blockages.

  • Rockfall on a cliff above a gallery on an Alpine road

    Rockfall and rock avalanches

    Automated recording of vibrations and micro-movements on rock, masonry and protective structures.

  • Landslide with debris at a mountain road in conifer forest

    Slope movements

    Observe tilt, movement and soil parameters over long periods and make trends visible.

  • Powder avalanche on a steep snowy mountain slope

    Avalanches

    Additional measurement data for assessment by avalanche commissions and operators.

  • Forest fire in a mountain forest with smoke in front of an Alpine peak

    Wildfire

    Sensor-based early detection in forest areas at risk.

Illustrative images, AI-generated.

Everything from a single source: sensors, platform and operation

GMD® develops its own measurement technology, installs it in the field and operates the system on a rental basis – municipalities do not need their own specialist staff.

Self-powered edge IoT sensors
Energy-autonomous sensors for hydrology, meteorology and geotechnics – even where there is neither power nor cable.
GeoSenseAI®
The platform combines sensor, weather and geodata, shows developments and triggers warnings based on individually defined thresholds.
Drones and digital twin
Autonomous drones and digital terrain models make changes visible and provide reliable documentation.
Installation, maintenance, operation
Planning, mounting in the channel or on the steep face, calibration and ongoing maintenance by the GMD® team.

Who we work for

Municipalities and authorities
A situation overview for command staff and fire brigades, and a traceable basis for closure and reopening decisions.
Rail and road
Monitoring of torrents, embankments and rockfall areas along transport corridors.
Mountain railways and tourism
Slope and channel monitoring above settlements and facilities, plus data-based visitor flow management.
Water supply
Self-powered monitoring of spring discharge, water level and consumption – even in high mountains.
Insurers
Measurement data and event documentation as a basis for risk assessment and claims settlement.

Case studies

Projects with partners in Graubünden and Tyrol, from a torrent catchment to the walls of a historic fortress.

Rhaetian Railway train on the Hinterrhein bridge in Graubünden
Debris flow and torrent

Rhaetian Railway: early warning system for the Carrerabach

Graubünden, with the Rhaetian Railway and the Graubünden Office for Forests and Natural Hazards

Instead of measuring at the structure, GMD® monitors the catchment of around 14 km² in real time – off-grid, year-round and without existing infrastructure on site.

View project

Neue Prager Hütte in the high mountains with a monitoring station in the foreground
Water resources

German Alpine Club: water monitoring at Neue Prager Hütte

Tyrol, at 2,796 m

A self-powered sensor system measures spring discharge, water level and consumption and transmits the data without any existing power or data infrastructure.

View project

“Not least because of climate change, our sensor-based monitoring solutions are becoming increasingly important for ensuring safety in the Alpine region.”

Steve Weingarth, founder and CEO of GMD® GmbH

GMD® – founded in 2024 in Innsbruck

An interdisciplinary team of geoinformatics, hydrology, data science and embedded development specialists. With offices in Innsbruck and Landquart and projects beyond the Alpine region.

More about GMD®

Partners and network

We work with public authorities, research institutions, infrastructure operators and companies in Austria, Switzerland, Germany and beyond.

Logos of our partners: German Alpine Club, Infineon, State of Tyrol, HTB Baugesellschaft, University of Innsbruck, SLF, TIROLER VERSICHERUNG, Standortagentur Tirol, Impact Hub Tirol, Canton of Graubünden, Rhaetian Railway, Graubünden Office for Forests and Natural Hazards, Graubünden Office for Economy and Tourism, Aalen University of Applied Sciences, Fraunhofer and CSEM

Planning a monitoring or early warning project?

Tell us about your site and the hazard to be monitored. We will get back to you with an initial expert assessment.