Quaternary fire activity in Northern Africa reconstructed from charcoal records in marine sediments.

Abstract

Fire in the Earth’s ecosystems is a significant driver of ecological change, fires remove dead biomass and prevent the reestablishment of woody plants acting to maintain and/or expand the high light environments. Frequent fires, create positive feedbacks that affect the environment by driving it towards increasingly arid conditions, by promoting the fire and often arid adapted grasses. The increased fire activity during the Late Miocene is hypothesized to have led to the large scale, global near synchronous expansion of C4 tropical grasses and consequently, the evolution of the savanna biomes. Little, however, is known about the role of fire activity in the geologic past. Existing terrestrial and lacustrine records are often discontinuous due to periods of erosion or non-deposition, cover limited time intervals, may have poor age control and can be strongly biased by very local events. Marine sediments provide an underexploited resource to circumvent these issues. In this thesis, I present a number of new reconstructions of regional fire activity spanning several Quaternaryglacial-interglacial cycles, reconstructed from deep sea sedimentary cores located on the Northwest African Margin. These new records allow me to shed light on long-lived patterns of palaeo-fire activity in response to intervals of climate change documented across northern Africa at both orbital and millennial time-scales. Chapter 4 explores the factors that determine the validity of fire activity reconstructions from marine sediments by examining inter-site discrepancies in charcoal recovery and comparing the results of three distinct techniques to reconstruct fire activity, including both geochemical and optical approaches. I report major differences in charcoal concentration between different sites along the northwest African margin, and suggest that oxidation could be a potential control on charcoal preservation in certain locations. In addition, I show that there can be large discrepancies in reconstructions of fire activity using optical and geochemical methods, with the optical method of charcoal extraction revealing more charcoal flux variability than geochemical techniques. Chapter 5 examines the evolution of spatial variability in fire activity along a transect (9°N-21°N)along the Northwest African Margin over the last 50 kyr. This interval covers events driven by both high- and low- latitude climatic forcing which occur over millennial and orbital timescales respectively. Each of the three sites in the transect shows a different relationship between hydroclimate and fire activity. The northern site exhibits a pattern of increased fire activity in response to increased humidity. In direct contrast, the southern site displays reduced fire activity when humidity increases. The central site (15°N) displays peak fire activity during intermediate humidity, with a reduction in fire activity during both arid and humid climate intervals. I attribute this to the proximity of this site to the highly rainfall sensitive grassy-woody savanna boundary, which experiences major changes in floral assemblage and therefore fuel loads associated with even small changes in rainfall. Greater aridity leads to a contraction of vegetation cover and a reduction of fuel loads, whilst increased humidity produces a floral assemblage that generates more moisture rich fuel loads, consequently in both cases resulting in decreased fire activity. Chapter 6 presents a long record of fire activity response to climatic changes through the Last Glacial Cycle (140-0 kyr before present) at Site Ocean Drilling Program (ODP) 658 (21°N). In general, a clear relationship between fire activity and precipitation is clearly observed, higher fire activity is observed when humidity increases due to the greater availability of biomass, supported by the higher precipitation levels facilitated by the northward displacement of the Intertropical Convergence Zone (ITCZ). However, during the times of highest humidity (African Humid Periods 5and 1), fire activity is dampened. I discuss three possible hypotheses to explain this suppression of burning: 1) High-latitude climate forcing exerted a secondary control on vegetation through remnant effects of the glacial, 2) Changes in atmospheric carbon dioxide exerts a secondary control on fire activity by promoting forest growth at the expense of grasslands, 3) The most extreme insolation peaks triggers conditions that were sufficiently humid to encourage the expansion of wetter-adapted less flammable ecosystems with higher fuel moisture contents. Chapter 7 compares and contrasts charcoal flux records from two glacial-interglacial cycles experiencing similar orbital geometry: the Last Glacial Cycle (MIS 1-5, 0-140 ka) and the Mid-Brunhes (MIS 9-11, 300-450 ka). The same relationship between precipitation and fire activity as observed in the Last Glacial Cycle is present through the Mid-Brunhes, with fire activity increasing with increasing humidity, except during peak interglacials. The absence of strong charcoal flux peak associated with MIS 11, a major interglacial associated with only weak insolation forcing, implies that it is the interglacial state rather than insolation forcing that is responsible for the suppression of burning during interglacials. This result is enigmatic, however, my favoured hypothesis to explain the low fire activity observed, is the development of a wetter less flammable vegetation assemblage leading to a smaller moister fuel source that experiences less climatic curing during interglacials. Also due to the similarity in the deglaciation from MIS 2 where there is the potential for a human impact and MIS 11 without a human impact in fire activity in the charcoal flux trends this study allows me to suggest that there is no or minimal anthropogenic forcing of fire activity during the Holocene.

More information

Identifiers

ORCID for Harriet Ruth Moore: orcid.org/0000-0002-1292-1953

ORCID for Paul Wilson: orcid.org/0000-0001-6425-8906

Catalogue record

Export record