Nikolaos KamakiotisWildfire Risk Analyst

Every fireleaves traces.

06:30Pisak, 21 June 2025Replay the Pisak fire

I rebuild how real wildfires moved, minute by minute, from the evidence they leave behind. Then I feed what each fire teaches back into the risk assessment of the next one.

Nikolaos KamakiotisWildfire Risk Analyst, FlEvo Ltd · Larnaca, Cyprus

Two lines of work, one cycle

In industrial safety, an incident investigation does not stop at what happened: it traces every contributing factor, so that each one can be addressed before the next incident. I apply the same discipline to wildfire, on two parallel tracks.

Research

Forensic reconstruction of extreme wildfires, carried out independently and with scientific partners.

  • Evidence-based fire spread reconstructionTen-minute arrival-time maps built from citizen footage, webcams, satellite passes and aircraft tracks.
  • NERO NetworkMember of the EU COST Action on extreme wildfire behaviour since 2024.
  • Scientific missionsNational Observatory of Athens (2024) and the Croatian Meteorological and Hydrological Service, DHMZ (2026).
  • PresentationsFrench National Academy for Fire Officers (ENSOSP, EU FIRE-IN project), European University Cyprus Risk Forum, Cyprus Insurance Institute, Safe Thessaloniki.

Practice

Through FlEvo Ltd, the findings are turned into services for the people who carry wildfire risk.

  • Post-incident analysisWhen did the fire reach a property, from which direction and under what wind: arrival windows tied to dated evidence, for insurers, loss adjusters and authorities.
  • Wildfire risk assessment with PYRODOTUSLong-term property risk, daily fire weather alerts and community maps for underwriting and portfolios.
  • Root cause analysisContributing factors behind losses, from fuel and town planning to evacuation and incident management decisions.

Eight years of reconstructing extreme fires

Each year is coloured as an isochrone on the maps: the earliest in deep purple, the latest in green.

2018

Mati, Attica

An independent spatiotemporal reconstruction from citizen and media data, with an incident animation. The fire scar was redrawn around every building, plot and road, and the fuel that was there before the fire was mapped across the whole burned area. Presented at the European University Cyprus Risk Forum and, by invitation, at the French National Academy for Fire Officers (ENSOSP) under the EU FIRE-IN project.

Fire spread reconstruction, 23 July 2018.
Detailed fire scar at Mati with unburned buildings and roads cut out
Detailed fire scar: unburned buildings, plots and roads cut out one by one.
Pre-fire fuel distribution across the Mati burned area
Pre-fire fuel distribution.
2021

Arakapas, Cyprus

Four fatalities, and at the time the largest wildfire recorded in Cyprus. The perimeter progression was reconstructed and presented at the Cyprus Insurance Institute annual conference.

Fire perimeter progression, accelerated.
Presenting the Arakapas analysis at the Cyprus Insurance Institute conference
Cyprus Insurance Institute annual conference, November 2021.
2022

Mati and Arakapas compared

Rapid front propagation under extreme conditions, presented at Safe Thessaloniki 2022 with the Greek Ministry of Climate Crisis and Civil Protection.

Presentation at Safe Thessaloniki 2022
Safe Thessaloniki 2022, HELEXPO, Thessaloniki.
2023

Lahaina, Hawaii

A different landscape, the same pattern of impact. Independent monitoring and data collection on fire spread and evacuation.

Burned area and building losses mapped over Lahaina.
2024

Varnavas, East Attica

The two most violent fire heads of the first 24 hours, in ten-minute steps, as the fire reached the north-eastern suburbs of Athens. The flight tracks of every firefighting aircraft were analysed: drops in speed and height mark water drops and add time-stamped evidence of where the fronts were. Within the perimeter, two communities with very different layouts were compared side by side. Followed by a scientific mission at the National Observatory of Athens on wildland–urban interface fires.

Flight tracks of firefighting aircraft over the Varnavas fire
Firefighting aircraft flight analysis: every track during the incident.
The Varnavas fire perimeter with the two communities analysed
Two communities inside the perimeter: Varnavas and Dioni.
Street-level view of a house in Dioni in 2012 with dry grass and pines
Dioni, January 2012: dry grass and pines around the house.
The same house in 2014 with vegetation grown up to its walls
October 2014: natural vegetation in direct contact with the walls.
The same plot in 2026 after the fire
March 2026: the only total loss in Dioni. Its neighbours survived.

Street-level imagery © Google. The Varnavas work was sponsored by Altex S.A., E.N. Manos and Funky Buddha.

2025

Pisak, Croatia

A full-day reconstruction of the fire of 21 June 2025, with the synoptic analysis of the Croatian Meteorological and Hydrological Service (DHMZ) and records from the local fire service. In March 2026 a NERO scientific mission brought the method to DHMZ in Zagreb, with training and field work with the fire brigade in Split.

With the Split fire brigade, March 2026
Split, March 2026: with the local fire brigade.
Field work on the Pisak slope
Field work on the Pisak slope.
Open the interactive reconstruction
2026

Oraiokastro, Thessaloniki

A night-time fire on 4–5 July 2026 that burned into an industrial zone and residential streets on the edge of the city. The reconstruction maps the fire against every factory, warehouse and home in its path, including combustibles stored outdoors and illegal dump sites that standard vegetation maps do not show.

Layered baseline map of Oraiokastro with vegetation, buildings and industry
Layered baseline: vegetation, farmland, homes and industry.

It looks like a simulation. Every line is evidence.

A simulation starts from equations and predicts where a fire could go. This method starts from records of where the fire actually was, and at what time. Physics is used only to connect those records, never to replace them.

  1. Collect. Videos and photos posted by residents, news footage, interviews with responders and witnesses, webcam recordings, satellite images and firefighting aircraft tracks.
  2. Locate and time every item. Where it was filmed from and where the front is in the frame, matched against pre-fire imagery; when it was filmed, from the post itself, other footage, the sun's position or shadow lengths.
  3. Fix the milestones. Each located and timed observation becomes a fire-front line, starting with the ignition point, which sets the clock for everything else.
  4. Cross-check them. A milestone that the evidence, wind, slope and fuel cannot support is re-examined or rejected.
  5. Fill the gaps. Ten-minute curves are drawn between milestones, inside the burned area, spaced by recorded wind, slope and gullies, vegetation, roads and man-made fuel.
  6. Animate. The curves are played in sequence on satellite imagery or 3D terrain.
slower: road, orchards faster: gully aligned with wind 06:5007:0007:1007:2007:3007:4007:50
Thick lines: milestones fixed by evidence, here a resident's video (06:50) and a webcam frame (07:50). Thin lines: ten-minute curves drawn between them, spaced by wind, slope and fuel.
Arrival-time mapsIsochrones at ten-minute intervals, showing when the front reached each part of the burned area.
Fire-front behaviourHow rate and direction of spread changed through the incident, which hourly reconstructions cannot capture.
AnimationThe spread replayed on satellite imagery or 3D terrain with the clock running.
Evidence registerEvery photo, video, webcam frame and satellite detection used, with its location, time and source.
Fire behaviour analysisThe weather and terrain mechanisms behind each change in speed and direction.
Technical reportMethod, findings and limits, written to be reviewed and challenged.

From incident analysis to risk assessment

Reconstruction and risk assessment are two stages of the same cycle, as in industrial safety management. Every incident investigated shows how losses actually happen; that knowledge is fed back to assess the next risk more accurately and reduce it before the next fire.

  1. IncidentFire reconstructed from evidence
  2. Loss mechanismsWhat made buildings burn and fronts accelerate
  3. Risk factorsWeighted at country, district, community and property level
  4. Risk reducedUnderwriting conditions, mitigation and in-season alerts

PYRODOTUS is where that cycle runs before the incident. It scores the long-term wildfire risk of a property and turns it into an underwriting decision, carrying the lessons of reconstructed fires. Built for Cyprus first and endorsed by the Cyprus Department of Forests.

PYRODOTUS, Wildfire Risk Intelligence Platform
  • Long-term riskA composite score from 74 property factors, from site and surroundings to construction and maintenance, with a confidence index and a printable risk action report.
  • Daily alert systemA seven-day fire weather outlook for each insured location, computed with the Canadian Fire Weather Index chain.
  • Community mapSettlement-level risk for portfolio accumulation.

Engineer, investigator, wildfire risk analyst

Fifteen years of investigating industrial incidents taught me to ask what turned an event into a loss. Public attention after a wildfire usually stops at the ignition. My work starts there: how the fuel was distributed, how properties were protected, how the fire moved and how the incident was managed.

Engineering

  • Mechanical Engineering, Aristotle University of Thessaloniki
  • Postgraduate studies in Water Resources, National Technical University of Athens
  • Member of the Technical Chamber of Greece (TEE) and the Cyprus Scientific and Technical Chamber (ETEK)

Safety and investigation

  • 15+ years as SHEQ manager at port, oil storage and airport operators, including DP World and VTTI B.V.
  • Lecturer on accident investigation (CERIDES / European University Cyprus)
  • ISO 45001 Lead Auditor; NEBOSH Oil & Gas; industrial firefighting and incident command training

Wildfire science

  • Member of the NERO Network (EU COST Action on extreme wildfires) since 2024
  • Scientific missions with the National Observatory of Athens (2024) and DHMZ, Croatia (2026)
  • Wildfire Risk Analyst, FlEvo Ltd, since 2024
NERO
NERO Network
EU COST Action
DHMZ
Croatian Meteorological and
Hydrological Service
National Observatory of Athens
National Observatory
of Athens

When did the fire reach the property?

For post-incident analysis, wildfire risk assessment, research collaboration or a talk, write or call.