Imagine being able to forecast, outsmart, and defend communities from wildfires. When Rocio Frej Vitalle and the Improving Aviation team developed WindTL, a tool powered by data and artificial intelligence, they had that goal in mind.
Building a service-model tool that could manage enormous data streams and intricate AI models in order to provide first responders with real-time wildfire predictions was necessary to make this vision a reality. Utilising Google Cloud’s scalable infrastructure for data processing and model deployment, the Improving Aviation team was able to construct the SkyTL core architecture to WindTL’s platform.
It took both technical know-how and a solid strategic alliance to scale the new tool. Google for Startups specialists offered leadership advice, and Google Cloud engineers offered advice on AI, streaming data, and runtime structures. Improving Aviation was able to quickly expand WindTL’s capabilities and greatly increase their influence in the fire and rescue sector because to this support.
End users now have science-based AI decision-making capabilities with Improving Aviation’s WindTL wildfire software, which gives them a far better understanding of the likelihood and seriousness of wildfire ignition threats both before and during a disaster. WindTL is more than just another fire model; it’s a lifeline, created using Google Cloud technology to fight ember spread, the silent killer of wildfire damage that causes an astounding 90% of structural losses.
Improving Aviation created WindTL, a decision support tool for predictive wildfire modelling, pre-fire risk assessment, and real-time situational awareness. With the use of AI-powered models, WindTL provides precise predictions for wildfire propagation, ember behaviour, and ignition danger, giving field workers the knowledge they need to take proactive, evidence-based decisions.
Scalable wildfire response
Scalable wildfire response: An explanation of WindTL’s Google Cloud foundation and architecture
Utilising a modular, cloud-based architecture that guarantees real-time situational awareness and predictive analytics, WindTL functions as a fully integrated product within SkyTL. These elements make up the foundation of WindTL’s architecture:
Pipeline for data ingestion
Gathering information from weather stations, Internet of Things sensors, satellites, and drones to analyse wildfire conditions in real time.
Gathering of historical information (Gemini)
By combining and organising many publically accessible internet data sources, a thorough historical wildfire dataset was created, which greatly enhanced the forecasting capabilities of the machine learning model.
Modelling powered by AI (Vertex AI)
Predicting fire behaviour and ember risk using physics-informed neural networks (PINNs) and unique machine learning models installed on Google Cloud’s Vertex AI.
Analytics of BigQuery data
Utilising BigQuery to store large datasets in a scalable manner and analyse them in real time, giving decision-makers useful information.
Interactive online user interface (integration with SkyTL)
An intuitive web-based interface for visualising fire spread models, high-risk areas, and assisting in the formulation of well-informed suppression strategies that is connected with Improving Aviation’s SkyTL backend.
Scalability in the cloud (GKE)
In order to ensure concurrent access for thousands of users and flexibility in the face of large-scale wildfire events, Google Kubernetes Engine (GKE) is used for quick scalability and low-latency replies.
WindTL’s smooth integration into emergency response processes is demonstrated in the following architecture diagram, which offers insurance companies, emergency responders, and firefighters dependable, real-time intelligence for crucial decision-making.
WindTL revolutionised wildfire forecast with the use of Google Cloud
In order to scale and improve the WindTL tool, the Improving Aviation team has been working closely with Google Cloud experts since 2024. Improving Aviation has accomplished the following by utilising Google’s cloud and artificial intelligence technologies:
Enhanced model performance
Improving Aviation decreased AI model run times over local execution for wildfire simulations by leveraging Google Kubernetes Engine (GKE) and optimised cloud computing.
Data processing that is scalable
Improving Aviation provided vital up-to-the-minute insight by enabling the ingestion and analysis of large, real-time wildfire datasets through the use of Google Cloud’s BigQuery and Vertex AI.
Enhanced accuracy of predictions
The accuracy of ember spread prediction was greatly improved by integrating Google’s physics-informed neural networks (PINNs), which resulted in more accurate risk assessments and resource allocation.
Highlight of WindTL: promoting innovation in a Google Cloud video
The brief film, Predicting wildfire behaviour with WindTL & Google AI, highlights Improving Aviation’s groundbreaking AI. The film, which deconstructs the science of ember propagation and shows how real-time decision support is preventing fatalities and catastrophic property losses. It delves deeply into how WindTL uses Google AI to forecast wildfire behaviour.
A Google Cloud team visited Improving Aviation’s Tampa headquarters to chronicle WindTL’s fundamental engineering processes and capture the magic that goes on behind the scenes. Read more about Google Cloud’s Innovation in Enhancing Aviation’s Experience in Transforming Wildfire Management.
Lessons learnt: partnering with Google Cloud to accelerate AI development
Beyond merely technical improvements, the collaboration between Improving Aviation and Google Cloud produced important information that developers creating AI-powered applications might use:
Scalability is a fundamental necessity
Real-time application design necessitates intrinsically scalable structures, especially in crucial fields like wildfire management. For effective data processing and dynamic resource allocation, cloud-native services must be used.
Model of collaboration
Improving Aviation engineers were able to apply iterative model optimisation through direct collaboration with Google’s AI and Outbound Strategy & Enable Cloud engineering teams, which greatly increased predictive accuracy. Its development cycles were hastened by sprints, shared expertise, and access to specialised tools.
Create AI-driven solutions with Google Cloud
Conventional wildfire response techniques are not keeping up with the increasing intensity of wildfires. Improving Aviation is using cutting-edge AI applications to innovate catastrophe management with WindTL and Vertex AI. Improving Aviation is developing WindTL to be a scalable and resilient system to handle the intricate data processing and prediction issues related to catastrophic wildfires, with the use of Google Cloud’s infrastructure and artificial intelligence tools.