Fires are a fantastic phenomenon to study. The complexity in something that starts with a spark is astounding. Often, fire is envisioned as a destructive process, but this viewpoint strongly depends on which part of the world you are talking about.For example, the total number of fires that occur every year is between about 1.5-3.5 millions fires, according to satellite-based measurements from the NASA Moderate Resolution Imaging Spectroradiometer (MODIS), which are shown in the figure. In the USA, fire is something no one wants in their backyard, or really anywhere close to them. Fire, however, is a vital part of many ecosystems, and it is an economical way to manage cultivated landscapes such as cropland or pasture. The USA is experiencing an increasing number of large fires, and because of the impact of human life and livelihood that arises from the overlapping boundaries between urban and wildland environments, we hear a lot about these fires. For perspective, however, only about 4-5% of fires that occur on Earth every year occur in the USA. Looking again at the figure, you can see that the tropics are where fires occur. In fact, nearly 75% of all fires occur in the tropics, and about 50% occur on the continent of Africa alone. Those numbers suggest that humans in Africa have a much different relationship with fire than humans in the USA. There is a lot of evidence supporting this hypothesis, but quantifying the specific human-fire relationship is not simple. On top of human-fire interaction, there is the climate-fire relationship, which is much more apparent in the data. Separating the two is crucial to understanding the role of fire in the Earth system, but as you might guess, that’s also not very simple. Humans have played a major role in the environment for perhaps 1000s of years. Climate is changing and has changed in the past. What can we do to parse out the effects? That’s a question I’m trying to answer in collaboration with colleagues who are interested in similar questions.
Fire measurementsI started this component of my research when I was in southern Africa as a part of the the large, multi-agency field campaign called SAFARI-2000. This was an effort spearheaded by NASA to characterize fire emissions and their impacts on the southern African region. I studied this from the perspective of aerosol physics and chemistry, and even now have not fully explored the data we collected. As a postdoc at NOAA GFDL, I became increasingly interested in how aerosol and trace gas emissions from fires are determined and started exploring the simulation of fires from the ignition of vegetation to emissions into the atmosphere.
PaleofireAlong with colleagues at Yale University (Jenn Marlon, twitter) and University of Oregon (Patrick Bartlein), and my PhD student Alex Schaefer, I am exploring the idea that a fire model can be used to test hypotheses about the roles that humans and climate play in determining global fire patterns. The collaboration is built on the foundation of measurements of charcoal in lake sediment that are a proxy for regional fire activity. These measurements have been made at over 800 different lakes in the world, so when combined, fire activity from charcoal records provides a proxy for global fire activity since the Last Glacial Maximum over 20,000 years ago! This project is funded by NSF.
Fire modelingI collaborate with researchers at Princeton University in the development of a global interactive fire model. My long term goal is to simulate global fires as a process that is fully coupled to the human and physical dimensions of the Earth system. In the near-term, I am test functions that accurately predict fire patterns. The global fire model operates within the framework of the GFDL Land Model and Earth System Model. The lead on this project is a Princeton University PhD student named Sam Rabin, who I have worked with for many years.
MethodsI use many different approaches to address the research questions described above. At the heart of any good model are observationally-based constraints. Complete knowledge of fire processes at all spatial and temporal scales is probably an unrealistic goal. Where the physics are poorly understood, I use measurements to constrain the representation of that missing physics. Measurements include NASA remote sensing data products, ground-based fire records, analysis of the culture of fire in different regions of the world, and paleofire records. There are plenty of other data that could be used as well (such as tree-ring records of forest disturbances). The computer simulations are then tested against these observations and the scientific literature is scoured for information that can be used to help me address any problems. By necessity, the work requires a high degree of comfort with not being comfortable. I read ecology, remote sensing, programming, climate, and dedicated fire science. Most of those are not part of my disciplinary expertise in atmospheric sciences.
Marlon et al. (2015): Fire trends over the past 22,000 years suggest we are presently in a strong upward trend, and charcoal influxes are higher now than since glaciers covered Wisconsin.
Rabin et al. (2015): We show the pasture fires account for over 40% of global burned area.
Magi et al. (2012): We show that regional fire seasonality depends on whether fires are used for agricultural management or occur as a result of non-agricultural burning, such as lightning ignited fire.
Magi (2009): I show that over 80% of aerosol mass in a fire-laden atmosphere is from carbonaceous species.
Magi et al. (2009): Using an atmospheric model, we show that carbon aerosol mass from a fire-prone region is underestimated, and that increasing emissions by 2-4 times is needed to have the model more closely match the data.
Magi et al., 2008: A report from AIMES Young Scholars Network meeting on the cultural practices and physical phenomenon of fire, and the allure of interdisciplinary fire research