Throughout Earth’s history, the abundance of oxygen in our atmosphere has varied, but by how much remains debated. Previously, an upper limit for atmospheric oxygen has been bounded by assumptions made regarding the fire window: atmospheric oxygen concentrations higher than 30–40% would threaten the regeneration of forests in the present world. Here we have tested these assumptions by adapting a Dynamic Global Vegetation Model to run over high atmospheric oxygen concentrations. Our results show that whilst global tree cover is significantly reduced under high O₂ concentrations, forests persist in the wettest parts of the low and high latitudes and fire is more dependent on fuel moisture than O₂ levels. This implies that the effect of fire on suppressing global vegetation under high O₂ may be lower than previously assumed and questions our understanding of the mechanisms involved in regulating the abundance of oxygen in our atmosphere, with moisture as a potentially important factor.

Fire regimes are changing across the globe, with new wildfire behaviour phenomena and increasing impacts felt, especially in ecosystems without clear adaptations to wildfire. These trends pose significant challenges to the scientific community in understanding and communicating these changes and their implications, particularly where we lack underlying scientific evidence to inform decision-making. Here, we present a perspective on priority directions for wildfire science research—through the lens of academic and government wildfire scientists from a historically wildfire-prone (USA) and emerging wildfire-prone (UK) country. Key topic areas outlined during a series of workshops in 2023 were as follows: (A) understanding and predicting fire occurrence, fire behaviour and fire impacts; (B) increasing human and ecosystem resilience to fire; and (C) understanding the atmospheric and climate impacts of fire. Participants agreed on focused research questions that were seen as priority scientific research gaps. Fire behaviour was identified as a central connecting theme that would allow critical advances to be made across all topic areas. These findings provide one group of perspectives to feed into a more transdisciplinary outline of wildfire research priorities across the diversity of knowledge bases and perspectives that are critical in addressing wildfire research challenges under changing fire regimes. This article is part of the theme issue ‘Novel fire regimes under climate changes and human influences: impacts, ecosystem responses and feedbacks’.

Several experimental studies have shown that climate-warming-induced permafrost thaw releases previously unavailable nitrogen which can lower nitrogen limitation, increase plant productivity, and counteract some of the carbon released from thawing permafrost. The net effect of this belowground fertilisation effect remains debated and is yet to be included in Earth System models. Here, we included the impact of thaw-related nitrogen fertilisation on vegetation in the Joint UK Land Environment Simulator (JULES) land surface model for the first time. We evaluated its ability to replicate a three-year belowground fertilisation experiment in which JULES was generally able to simulate belowground fertilisation in accordance with the observations. We also ran simulations under future climate to investigate how belowground nitrogen fertilisation affects the carbon cycle. These simulations indicate an increase in plant-available inorganic nitrogen at the thaw front by the end of the century with only the productivity of deep-rooting plants increasing in response. This suggests that deep-rooting species will have a competitive advantage under future climate warming. Our results also illustrate the capacity to simulate belowground nitrogen fertilisation at the thaw front in a global land surface model, leading towards a more complete representation of coupled carbon and nitrogen dynamics in the northern high latitudes.

Atmospheric oxygen concentrations have remained remarkably stable over the past ~400 million years (Myr), suggesting the presence of robust regulatory mechanisms. Because of its sensitivity to oxygen, wildfire was traditionally assumed to control oxygen levels by limiting terrestrial vegetation; however, this feedback is nullified by high moisture levels in tropical ecosystems. Using vegetation modeling, we show that where oxygen-fire effects are dampened by high moisture, photorespiration becomes more effective through increased temperatures. Together, these processes interact to drive an 86% reduction in modeled global biomass when oxygen levels reach 35%. This coregulation imposes substantially tighter control of atmospheric oxygen than wildfire alone, providing previously unknown insights into the spatial and interactive feedbacks that may explain the remarkable stability of oxygen levels since the evolution of forests.

Oxygen has varied in its abundance in the atmosphere throughout Earth’s long –term evolutionary history. Laboratory experiments have shown that fire responds rapidly to oxygen changes. Therefore it has been suggested that increases and decreases in atmospheric oxygen levels have influenced fire frequency, fire behaviour and plant evolutionary adaptations to fire.Recent research has indicated that periods with high atmospheric oxygen, such as the Cretaceous period, which also saw the evolution of new plant groups had the coupled effect of altering fire behaviour. Such modelled fire behaviour has been able to estimate that fires during this period would have been more intense and spread more rapidly which likely fed back to changes in ecosystem dominance. However, we are lacking understanding of how oxygen driven changes in fire might feedback to influence the dominance and distribution of land-surface vegetation cover across Earth’s surface throughout Earth history.Here I will present, a series of oxygen-fire-land cover simulations using the LPJ-LMfire Dynamic Global Vegetation Model that considers how oxygen-mediated changes in fire frequency and behaviour lead to changes in dominance of selected plant functional types within Earth’s biomes and influence the total land area covered by forest.Our aim being to explore the coupled influence of oxygen and a climate on vegetation distributions mediated by fire throughout Earth’s past such that we can work towards understanding the balance of natural fire feedbacks to the Earth system versus human interrupted fire feedbacks in our modern day.

Wildfire risk is increasing in temperate regions like the UK and NW Europe, but we lack operational tools to support wildfire management decision-making needs. We developed FireInSite to address the need for a user-oriented system for predicting fire behaviour. FireInSite is a fire behaviour prediction system in the form of a web-based application that forecasts the probability of ignition, surface fire rate of spread, flame length and fireline intensity for a user selected location for a set of core UK fire prone fuels. By seamlessly integrating geolocated weather forecasts up to 5 days ahead, topographic data, and in-built UK specific fuel models, FireInSite creates an accessible system that removes barriers like the need to gather data from multiple sources and is designed to minimise the number of inputs and decisions users must make before being able to predict fire behaviour. FireInSite can be used to assess the risk of fire in a particular area, plan for fire prevention and suppression, assess the potential effects of fuel load reduction, and educate the public about fire behaviour. We envision FireInSite being useful as a land management planning tool to assess the potential impacts of proposed landscape changes on potential fire behaviour.FireInSite is built on over four years of intensive data collection of fuel moisture, fuel flammability, and energy contents measured across the UK for key fire prone vegetation types, which have been used to develop fuel models that describe the fire prone fuel types of the UK landscape for the first time. No other fire behaviour prediction system contains fuel models that have been specifically designed and tailored to UK vegetation and are ready inbuilt for use in the system. It also allows the user to select custom developed fuel moisture models, explore past fire behaviour using historical weather records back to 1970, and compare weather and fuel moisture forecasts to conditions in previous years. As FireInSite fuel models capture seasonal variability in fuel flammability and moisture for a range of temperate, humid fuels, we anticipate that FireInSite will also be transferable and of interest for wildfire management in other temperate regions like north western Europe.