Quantitative Measurement of Wildfire Behavior in the Field: Leveraging Remote Sensing for Reproducible Observation and Improved Understanding

野火在世界多个地区（包括美国西部）日益增加的社会，生态和经济影响促使人们研究火灾行为原则，以更好地了解，预测和减轻野火的不良影响。这项研究导致了一些火灾模型的开发，旨在支持全球对野火的复原力。此外，还进行了广泛的实验活动，以获得可用于模型开发和验证的观测结果。虽然实验室研究有助于了解火灾行为的基本原理，但火灾动力学所涉及的物理和化学现象的复杂性进一步需要在现场进行广泛的观察，以证实和推断实验室研究结果。然而，测量野外野火行为涉及重大挑战，野外获得的有关野火行为的定量数据仍然很少。在收集这些数据时，火行为的特点通常是广泛的空间和时间平均值，这阻碍了对植被，地形和天气的火灾反应的详细研究。这种数据空白阻碍了模型和模拟器的开发，这些模型和模拟器可以预测火灾的演变及其对环境的影响。为了克服这一限制，本项目将利用现代传感器技术和最先进的数据处理算法来自动获取详细的火灾行为信息。本项目的主要目标是开发遥感工具和技术，以促进野外野火行为的自动化和定量测量。首先，将使用商用现货（COTS）传感器构建一个便携式和负担得起的系统。该遥感系统将设计成多功能的，以便可以从各种平台上操作，包括无人驾驶飞行器。其次，将开发科学软件，从原始传感器数据中导出时间和空间上明确的火灾行为指标，如蔓延率（ROS），火灾辐射功率（FRP），火灾停留时间（RT）以及火焰和前沿几何形状。将特别注意图像的直接地理参照以及不同时间和不同观测点获得的数据的融合。第三，将设计一个基于地理信息系统（GIS）的数据可视化平台，以支持结合羽流动力学、植被、地形和天气等信息对火灾行为数据进行分析。最后，将监测系统、数据分析软件和数据可视化平台部署在两个现场实验中。该项目的成果将填补数据可用性方面的关键空白，并必将支持火灾行为研究和火灾模型开发。该奖项反映了NSF的法定使命，并通过使用基金会的知识价值和更广泛的影响审查标准进行评估，被认为值得支持。

The increasing social, ecological and economic impact of wildfires in multiple regions of the world, including the western United States, has spurred research into fire behavior principles in order to better understand, predict and mitigate the undesired effects of wildland fire. This research has resulted in the development of a number of fire models intended to support global resilience to wildfire. Additionally, extensive experimental campaigns have been conducted to acquire observations that can be used for model development and validation. While laboratory studies are useful to understand the fundamentals of fire behavior, the complexity of physical and chemical phenomena involved in fire dynamics further requires extensive observation in the field to corroborate and extrapolate laboratory findings. However, measuring wildfire behavior in the field involves significant challenges and the amount of quantitative data about wildfire behavior acquired in the field is still scarce. When such data is collected, fire behavior is usually characterized by broad spatial and temporal average values, which prevents the detailed study of fire response to vegetation, terrain and weather. This data void hinders the development of models and simulators that would allow predicting the evolution of an active fire as well as its effects on the environment. To overcome this limitation, this project will leverage modern sensor technology and state-of-the-art data processing algorithms to automate the acquisition of detailed fire behavior information.The main goal of this project is to develop remote sensing tools and techniques to facilitate the automated and quantitative measurement of wildfire behavior in the field. Firstly, a portable and affordable system will be built using commercial off-the-shelf (COTS) sensors. The remote sensing system will be designed for versatility so that it can be operated from a variety of platforms, including unmanned aerial vehicles (UAVs). Secondly, scientific software will be developed to derive temporally and spatially explicit fire behavior metrics, such as Rate of Spread (ROS), Fire Radiative Power (FRP), fire Residence Time (RT) and flame and front geometry, from the raw sensor data. Special attention will be paid to image direct georeferencing and the fusion of data acquired at different times and from different observation points. Thirdly, a data visualization platform will be designed based on a Geographic Information System (GIS) to support the analysis of fire behavior data in conjunction with information about plume dynamics, vegetation, terrain, and weather. Finally, the monitoring system, the data analysis software and the data visualization platform will be deployed in two field experiments. The outcomes of this project will fill a critical gap in data availability and are bound to support fire behavior studies and fire model development.This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.