Profiles of Operational and Research Forecasting of Smoke and Air Quality Around the World

Biomass burning has shaped many of the ecosystems of the planet, and for millennia humans have used it as a tool to manage the environment. When widespread fires occur, the health and daily lives of millions of people can be affected by the smoke, often at unhealthy to hazardous levels leading to a range of short-term and long-term health consequences such as respiratory and cardiovascular illness, and mortality. It is critical to adequately represent and include smoke and its consequences in atmospheric modeling systems to meet needs such as addressing the global climate carbon budget and informing and protecting the public during smoke episodes.

Smoke impacts from biomass burning are highly episodic, with great variability from day to day and year to year, making it extremely difficult to simulate and predict. Smoke plumes can linger close to the ground where people breathe or be lofted high into the atmosphere and transported long distances. Further, fire interacts with the atmosphere, creating its own fire weather, such as smoke-induced density currents and pyrocumulonimbus clouds. On regional to global scales, smoke interacts with the atmosphere, changing the radiative budget of the atmosphere, modifying winds and temperature, and interacting with cloud processes. Within the smoke plume, the rich mixture of trace gases and aerosols transform continuously through a variety of chemical and physical processes.

Many scientific and technical challenges are associated with modeling the complex phenomenon of smoke. Variability in fire emissions estimates has an order of magnitude level of uncertainty, depending upon vegetation type, natural fuel heterogeneity across the landscape, and fuel combustion processes. Methods for quantifying fire emissions also vary from ground/vegetation-based methods to those based on remotely sensed fire radiative power data. These emission estimates are input into dispersion and air quality modeling systems, where their vertical allocation associated with plume rise, and temporal release parameterizations influence transport patterns, and in turn affect chemical transformation and interaction with other sources. These processes lend to another order of magnitude of variability on the downwind estimates of trace gases and aerosol concentrations. Unlike retrospective modeling studies, forecast modeling systems have a particularly challenging task because they cannot use observations to set fire locations, fire behavior, emissions composition, timing of emissions, or how high emissions are lofted and distributed vertically in the atmosphere. They need to make assumptions, such as persistence, where the fire information from yesterday is assumed to apply to every future day in the forecast period. Coupled atmosphere-fire behavior systems are promising, as they track the evolution fires, but as of yet these systems are too computationally intensive to implement on regional and global scales. Data assimilation techniques combine numerical model predictions with observational datasets to provide a powerful means of initializing model runs to address some of these forecasting challenges.

In this presentation we profile many of the global and regional smoke prediction systems currently operational or quasi-operational in real time or near-real time. It is not an exhaustive review of systems, but rather a profile of many of the systems in use to give examples of the creativity and complexity needed to simulate the phenomenon of smoke. The systems described reflect the needs of different agencies and regions, where various systems are tailored to the best available science to address challenges of a specific region. Smoke forecasting requirements range from warning and informing the public about potential smoke impacts to planning burn activities for hazard reduction or resource benefit. Different agencies also have different mandates, and the lines blur between the missions of quasi-operational organizations (e.g., research institutions) and agencies with operational mandates. The global smoke prediction systems are advanced, and many are self-organizing into a powerful ensemble, the International Cooperative for Aerosol Prediction (ICAP). Regional and national systems are being developed independently for example in Europe (11 systems), North America (7 systems), and Australia (3 systems). Finally, the World Meteorological Organization (WMO) is bringing together global and regional systems to form an ensemble and building Vegetation Fire and Smoke Pollution Warning Advisory and Assessment Systems (VFSP-WAS) to support countries with smoke issues and which lack resources. For each system we discuss how fire activity information is obtained, how fire emissions are calculated, and how atmospheric transport and chemical transformation of the smoke plume is treated. This presentation is based on a book chapter in the upcoming Wylie AGU publication Fire, Smoke and Health: tracking the modeling chain from flames to health and wellbeing.