Between 2024 and 2026 the wildfire-tech segment raised $350-400 million in fresh investment. A European thermal-satellite leader closed EUR 37M Series B extension, a Fraunhofer spinoff closed EUR 37M Series A, a Luxembourg player closed $60M Series B, FireSat-affiliated contractor and FireSat operated on $13M from Google, a dedicated camera AI vendor closed $44M Series B, and the market-leading fire-spread SaaS vendor raised a round with General Atlantic at a valuation of approximately $84M.
Scientific publications are scarce. Most companies on that list have no peer-reviewed papers describing their methodology in a form that an external group could reproduce. A peer-reviewed paper in a Q1 journal with full code and data is rare.
We build YourAirTest differently. Peer review is a discipline, not PR. One paper every twelve months. The first has already appeared: Canadian Journal of Forest Research, 2026. Four more will follow by end of 2028, across five different journals. The WildFiresUA system is our complete stack for Ukraine, and every subsequent paper documents one of its layers.
Why peer review is defensibility
When a startup submits to Horizon Europe, EIC Pathfinder, or an EU Mission, evaluators look for two signals in the track record: that a real-world deployment exists, and that the methodology has survived external scrutiny. Peer-reviewed publication is the most transparent signal of the second.
A Horizon Europe evaluator is often an academic from CSIC, CNR, or FMI, reading 15 applications per week. “Peer-reviewed publication in CJFR / ACP / ESSD” reads as “the text survived anonymous peers”. “Validation performed internally” forces the evaluator to verify alone, which lowers the credibility score.
Peer review does not guarantee a grant. Its absence makes a decision more cautious. When we talk with DSNS they want to see that the model is not a black box. A peer-reviewed paper is a contract with the reader: this is how we compute, these are the metrics, these are the limits.
CJFR 2026: what we published
In January 2026, Canadian Journal of Forest Research published our flagship paper: “Mapping and development of 30-meter landscape fuel data for Ukraine”. The first peer-reviewed 30 m fuel map for Ukraine, prepared together with Oles Honchar Dnipro National University, the Marzieiev Institute and UHMI.
We adapted the Fuel Characteristic Classification System (FCCS) to Ukrainian landscapes. For Kyiv and Dnipropetrovsk oblasts we built 13 classes under the Canadian Fire Behaviour Prediction System (FBP). The pipeline: FCCS map -> CONSUME computes consumed fuel -> FEPS generates hourly emissions of PM2.5, PM10, and CO -> CALPUFF v5 transports the plume through ERA5-driven WRF v4.3 meteorology -> validation against three KMDA stations.
The test case was a fire on 23 March 2022 in Kyiv oblast. PM2.5: BIAS +2.77 ug/m^3, RMSE 48.12 ug/m^3, Pearson r = 0.40. PM10: BIAS -6.29, RMSE 54.14, r = 0.40. CO: r = 0.48. Reviewers’ formulation: “Most modeled pollutant concentrations differ from the corresponding measurements by no more than a factor of 2.”
A global 300 m fuel map at the same stations yielded BIAS -9.66 and Pearson r = 0.28. Fuel resolution thus delivers a 10x accuracy improvement. This is the scientific novelty: quantitatively demonstrating why Ukraine requires a native 30 m map.
The manuscript was submitted in Q4 2024. The first round of review returned 31 comments. The second, 16. Acceptance in Q3 2025. Publication in January 2026. Total: 14 months, 47 substantive comments, a written response to each. Three were rejected with justification. The final version is stronger than the first.
The base that receives less attention: three earlier publications
CJFR 2026 is not our first publication. Three peer-review cycles preceded it.
ISPRS Archives 2023. Our team published a CNN architecture for land-cover classification on satellite imagery. Validation accuracy: 97%. The model became part of the fuel-map preparation pipeline.
DGPF Munich 2023. The same CNN, extended validation over European landscapes. First academic connections in the German community.
Springer LNDECT 2022. Our first peer-reviewed publication: multispectral data processing. It taught us the timeline of a cycle and how to structure a response to reviewers. Without these three papers CJFR 2026 would not have materialised.
Plan 2026-2028: one paper per 12 months
Each paper has a specific journal, dataset, and hypothesis, and each plays into the next grant application.
Q4 2026: ESSD (Earth System Science Data). A data paper. We publish the Ukrainian wartime fire corpus 2022-2025: VIIRS/MODIS thermal detections, OSINT verification of key events, metadata for sources (industrial sites, fuel depots, peatlands). ESSD is an open Q1 Copernicus Publications journal specialising in datasets.
Q2 2027: Atmospheric Chemistry and Physics (ACP). Coupled FLEXPART-WRF validation on the 2022 Kremenchuk refinery fire. We shift the focus from CALPUFF to the FLEXPART Lagrangian-particle approach. ACP is one of the top atmospheric-chemistry journals.
Q4 2027: Environmental Research Letters (ERL). Population-exposure attribution: a FLEXPART+CALPUFF hybrid against EcoCity ground truth for key events 2022-2025 (Kremenchuk, Lysychansk, Kakhovka, Chornobyl Exclusion Zone). Output: years of life lost to smoke exposure, with confidence intervals.
Q2 2028: Fire (MDPI). Wartime-specific fuel models: adaptation of FCCS to destroyed urban infrastructure, peat with reduced moisture from Kakhovka drainage, and stubble. Applied models for DSNS.
Q4 2028: Nature Communications or ACP. If our ML-physics hybrid (WRF <-> FLEXPART with uncertainty quantification) produces a breakthrough result, Nature Communications. If it is strong but not a breakthrough, ACP, still Q1.
Five journals: CJFR (already), ESSD, ACP, ERL, Fire. Four of five are open access. All with open data and open code.
What reviewers look for
Methodological transparency. A reviewer must be able to reproduce the work from scratch. “Parameters tuned experimentally” without specific values will not pass.
Validation metrics. BIAS, RMSE, correlation coefficients, sensitivity, comparison with existing approaches. Our BIAS +2.77 / RMSE 48 / r = 0.40 is not perfect, but honest.
Open data and reproducibility. We build on open sources: ERA5, WRF, FLEXPART, CALPUFF, FCCS, VIIRS, MODIS. Code on GitHub or Zenodo with DOI. A Docker run executes the pipeline end-to-end. The DSNS deployment and operational use are an advantage over purely academic publications.
How publication strengthens a grant application
The publication plan is a vertical that feeds grants.
CJFR 2026 -> April 2026: EIC Pathfinder (EUR 3-4M), physics-ML hybrid. ESSD end of 2026 -> Q1 2027: Horizon Europe Cluster 5 Climate Services (EUR 8-15M), the wartime corpus under FAIR. ACP Q2 2027 -> EU Mission Adaptation regional demonstrator. ERL end of 2027 -> EU4Health / DG SANTE with the Marzieiev Institute, wildfire-smoke mortality at EU scale. Fire 2028 -> DG ECHO UCPM procurement with DSNS as reference customer. Nature / ACP end of 2028 -> the next EIC Accelerator round.
Five publications, six grant windows.
Dual-use: academic respect and operational trust
Risk: the publication track pulls the team into a pure academic mode. Each paper carries an operational component: data live to DSNS, models in production. When a reviewer asks whether it is real, we answer with screenshots of operational systems. In reverse, each deployment becomes material for the next paper. A pure academic startup would not survive the market. A pure product startup without science would not pass Horizon Europe.
Why the other players do not do this
The European thermal-satellite leader (EUR 37M), the Fraunhofer spinoff (EUR 37M), and the Luxembourg player ($60M) are satellite-first with hardware capital. Their defensibility sits on orbit. The market-leading fire-spread SaaS vendor ($84M round with General Atlantic) is a modelling SaaS with a 20-year history; a closed proprietary stack is their defensibility. The dedicated camera AI vendor ($44M) runs computer vision on camera networks. Muon/FireSat is a non-profit on Google sponsorship and does not produce papers at regular cadence.
When Horizon Europe reviewers sit in a room with the thermal-satellite leader, the fire-spread SaaS vendor, and YourAirTest, we hold one advantage no one can close in six months: external academic validation of the full pipeline. They hold capital; we hold time for scientific proof. Converting science into grants is our asymmetric bet.
FAQ
Is “one paper per 12 months” realistic for a small team?
Yes, if each grows from real operational work. ESSD 2026 formalises data we already collect. ACP 2027 is validation for DSNS. Half of the volume is reformatting internal documentation.
What happens when a paper is rejected?
CJFR accepted on the second round. Had it been rejected, we would have resubmitted to Atmospheric Environment. For each target journal we hold a fallback.
Is open code not a risk?
Real-time ingestion and the dashboard remain proprietary. A competitor would spend 18-24 months reproducing the stack and would still lack access to the wartime corpus.
Our tools
- Air-quality map: partner.yourairtest.com/map
- Fire forecast: partner.yourairtest.com/forecast
- Monitoring system: air-quality-monitoring-system
- Wildfires: wild-fires
- About the project: pro-proekt
Ukrainian startup ecosystem: follow TechUkraine and AIN.ua — the two leading outlets covering Ukrainian deep tech, climate tech, and environmental startups.
What to do today
- Check the YourAirTest air quality map for your city — recent PM2.5 readings.
- If the topic resonates, share with colleagues; we monitor referrals via Google Search Console.
- To contribute data (sensor measurements, regional models) reach out via the contact form.
References
- Our peer-review paper CJFR 2026: Mapping and development of 30-meter landscape fuel data for Ukraine. Canadian Journal of Forest Research, cjfr-2025-0035.
- Our CNN paper on land-cover classification (ISPRS Archives, 2023).
- Our CNN paper on European land-cover mapping (DGPF Munich, 2023).
- Our paper on multispectral data processing (Springer LNDECT, 2022).
- Sofiev et al. (2025). Wildfire smoke and mortality in Europe. Lancet Planetary Health.
- WHO Global Air Quality Guidelines (2021). EU AAQ Directive 2008/50/EC and draft 2024/2881.