Wildfire modelling and impacts on carbon cycle, climate, and vegetation distribution induced by climate change

Abstract

Wildfire is a natural and unavoidable feature of the environment in many terrestrial ecosystems and has a strong influence on global ecosystem patterns and processes, including vegetation distribution and structure, the carbon cycle, and climate. However, as illustrated in recent studies, fire is a risk to human societies, emits atmospheric pollutants, and causes death and damage to homes and businesses. Fire regimes have been altered by complex interactions among climate change, anthropogenic activities, and biome community changes and are predicted to change further with global warming, population growth, and urbanization. Therefore, the frequency and spatial distribution of wildfires, along with their drivers of recent trends, requires better understanding to enable robust future predictions. Many studies have had a central focus on the interactions between the carbon cycle and climate. However, as an important component in modulating future climate change, the climate-carbon feedback derived from changing fires is poorly understood. Here, we develop a fully coupled climate-vegetation-carbon-fire framework and perform simulations to investigate global Burnt Area (BA) trends and their drivers, climate-carbon feedbacks derived from changing fires, and the role of fire on the global carbon cycle. Meanwhile, the individual contribution of fire and climate on regional vegetation distribution is assessed using models, remote sensing technology and statistics. The main results and conclusions are provided below: (1) Global BA has decreased, and this is reproduced well in our simulation framework, capturing recent climate and human limits of change. Global BA will increase under climate change, rapid population growth and urbanization, and this is especially through high latitude warming and increases in human-induced fires over the tropics and subtropics. However, urbanization will also offset the potential future large increases in BA by enhancing fire suppression capacities and management. (2) Over the period 2005-2014, the carbon-cycle effect (i.e., caused via release of extra carbon into the atmosphere) of fire has increased the additional mean annual global land temperature by ~0.16 °C yr-1 relative to a world without fires. Future mean annual global fire carbon emissions will be sensitive to anthropogenic CO2 emissions and human behaviour (2.2 to 3.0 PgC yr-1), providing an approximate increasing mean annual atmospheric CO2 concentration of 15-22 ppm yr-1 and global land-mean warming of 0.13 to 0.18 °C yr-1 over the period 2081-2100, with warming being greatest at high latitudes. However, although the largest impact of fire on atmospheric CO2 concentrations is associated with the more severe anthropogenic emissions scenarios, the associated fire-climate feedbacks on enhanced warming are the lowest under these scenarios. (3) The vegetation distribution for the present day is mainly determined directly by climate (35%) rather than fire (1%-10.9%). However, global warming will change the balance between fire and climate in driving variations of boreal forest distribution under four global warming scenarios. With a future global warming of 1.5 °C, the fire control on local vegetation composition will grow to exceed that of climate (36.3% > 29.3%). Above a 1.5 °C warming, temperature will be more important than fires in regulating vegetation distribution, although other factors like precipitation will also contribute. Our analysis will help to develop wildfire management strategies in conditions of different possible future socioeconomic development pathways, to assess future ecological changes, and to help inform adaptation and mitigation options to global climate change.