Impacts of fire, climate and land-use change on terrestrial ecosystems

Abstract

Fire is an important component of the Earth system, affecting the land surface, releasing gases to the atmosphere, and altering the water cycle. Yet many Earth System Models lack full representation of this process, giving rise to uncertainty about its contribution to the development and stability of ecosystems now and in the future. In this PhD I investigate the impact of fire on the land surface today, and how this might change with drought events and with climate change in the future by developing the land surface model JULES to represent fire-vegetation interactions for the first time. I introduce a new fire disturbance term based on burnt area from the INFERNO fire model, and analyse the results of the coupling, together with changes in land-use, against observations of present day vegetation cover. I find that the simulation of vegetation cover is improved when disturbance is included, and that fire is important in the development of savanna regions. I apply the new modelling capability to assess the impact of the 2015/16 El Niño event on fire, where projections show that burned area and fire emissions were higher due to the El Niño. The largest impact was across South America, where carbon uptake was reduced due to increases in fire, inducing a shift from a net sink of carbon to a net source. Fire danger may be further exacerbated in years of higher temperatures and drought in the future as a result of climate change. I apply the capability to model different aspects of the fire regime with future scenarios of climate and land-use change across a range of emission scenarios. Using Representative Concentration Pathway scenarios, I show that burned area is projected to increase in the future, with hotter, drier conditions increasing with higher emission scenarios and greater changes in land-use, especially across South America but not homogeneously. Using a theoretical scenario of Solar Radiation Management to limit temperature rise to 1.5°C above pre-industrial, I show that meteorological fire danger is generally reduced compared to 2.0°C, although there are regional variations and some regions show an increase including USA and Asia. This work furthers our current modelling capability around fire vegetation interactions, and enhances our understanding of the response of ecosystems to changes in fire, climate and land-use.