| Some of what you suggest is being done by fire departments and companies but there are complications that make the problem more difficult than it may seem: * We don't currently have the satellites to monitor fires at both small scale and real time. Geostationary (real time) satellites have a resolution of ~1km, and higher resolution satellites of ~10m to ~1m may only make a pass once per day. * The FAA imposes flight restrictions around wildfires and doesn't make exceptions for research or scrappy startups, so it's tough to experiment or go to market with drone-based monitoring or firefighting. * A lot of research has been done into fire spread prediction, including some using AI, but it's a difficult problem. Like weather forecasting it's pretty accurate on short time horizons, but over many hours and days errors compound. * Data. High resolution fuel data is needed, which changes every season as vegetation changes, and changes every day with the weather. Quality elevation data is needed. We don't have much ground-truth data of what's going on inside a wildfire. Finally, we have a combination of: forest fires are natural, suppressing them increases future likelihood of fire, people are building houses in fire-prone areas, climate change is increasing the frequency of fires. We could probably do more to reduce the loss of life due to fire, but the biggest problems are social rather than technical: we need to reduce our impact on the climate, stop building houses in high-risk areas, and not get NIMBY about prescribed burns and forest management. |