Research Centres
Research projects
Explore current and recent research projects that the Centre for Palaeobiology and Biosphere Evolution has taken part in.
Our projects
Hydrogen Isotopes in Lignin: A New Proxy for the Amazon Hydrological Cycle?
NE/X012921/1 NERC Discovery Grant
The Amazon basin is the world's largest catchment and sustains the largest continuous rainforest on Earth. Global change is affecting the Amazon climate in uncertain ways, with recent decades seeing an increase in Amazon wet season rainfall, as well as, increasing drought frequency. To better understand the drivers behind these changes, long-term accurate records of the hydrological cycle are needed.
In this project we will test a new climate proxy derived from hydrogen isotopes in tree rings that can contribute towards such reconstructions. In contrast to existing oxygen isotope proxies in the Amazon, hydrogen isotopes in lignin methoxyl groups are thought to preserve the signal of hydrogen isotopes in precipitation. We will test to what degree hydrogen isotopes growing along an elevational transect from the Amazon to the Andes indeed preserve variation in hydrogen in precipitation, and apply this new method for the tropical tree species Cedrela odorata at a dry and wet site to reconstruct precipitation oxygen isotopes for the past 50 to 150 years at annual resolution.
If successful this new proxy can be used to generate insights in the response of the tropical hydrological to climate change. Reconstructions of hydrogen isotopes in the source water can further be useful to interpret long-term change in plant physiology to CO2 when used in combination with oxygen and carbon isotopes from tree rings.
CongoPeat: Past, Present and Future of the Peatlands of the Central Congo Basin
NE/R016860/1 NERC Large Research Grant
We recently discovered the world's largest tropical peatland complex, spanning an area larger than England, in the heart of Africa. This proposal brings together an interdisciplinary team of scientists to study this newly discovered ecosystem. Our goal is to understand how the peatland became established, how it functions today, and how it will respond to human-induced climate change and differing future development pathways. We will use the results to inform critical policy decisions about the region.
Peat is partially decomposed plant matter. Peatlands are some of the most carbon-dense ecosystems on Earth. Covering 3% of Earth's land surface, they store one-third of soil carbon. A recent NERC-funded PhD, led by CongoPeat PI Professor Lewis, showed for the first time that the largest wetland in Africa, in the central Congo Basin, contains extensive peat deposits. This research, published in 2017 in Nature, estimates that the peatland stores 30 billion tonnes of carbon (C). By comparison, in 2016, UK emissions were 0.1 billion tonnes of C. Our discovery increases global tropical peatland C stocks by 36%.
We know very little about this new globally important ecosystem. Our data show peat accumulation began about 10,600 years ago, when central Africa's climate became wetter. Accumulation has been slow - on average just 2m has accumulated over this period - but it is unknown whether this is due to a constant slow build-up of peat and C, or fast rates interspersed with losses in drier periods. Our evidence suggests that the peatlands are fed by rainfall, but such peatlands usually form domes ('raised bogs'), yet satellite data do not show this feature. Thus, we do not know how this peatland system developed, how it functions today, or how vulnerable it is to future climate and land use changes.
Tropical peatlands in SE Asia have been extensively damaged by drainage for industrial agriculture, particularly oil palm, with serious biodiversity, climate and human health implications. Oil palm is now rapidly expanding across Africa. Congolese peatlands could become a globally significant source of atmospheric CO2 if they are drained, leading to their decay. A prerequisite of following a different development pathway is a scientific understanding of the region.
The CongoPeat proposal therefore brings together leading experts from six UK universities, a science-policy communication specialist, and five Congolese partner organisations, to gain:
- An integrated understanding of the origin and development of the central Congo peatland complex over the last 10,000 years. We will analyse peat deposit sequences from across the region, extracting preserved pollen grains, charcoal, and chemical markers, to reconstruct the changing environment through time. We will use an unmanned aerial vehicle to map peatland surface topography, and develop a mathematical model of peatland development.
- A better estimate of the amount of C stored in the peat, its distribution, and the amounts of important greenhouse gases, CO2, methane, and nitrous oxide, being exchanged with the atmosphere. This will be achieved via extensive fieldwork to map peat distribution, and by installing intensive measurement stations to determine the flows of C into and out of the ecosystem.
- An understanding of the possible future scenarios for the Congo peatlands. A range of models will be used to simulate the possible impacts of future climate and land-use change on the peatland, at local to global scales.
Finally, we will effectively communicate these results to policy-makers in Africa and internationally via briefings and active media engagement.
The CongoPeat team will produce the first comprehensive assessment of the genesis, development, and future of the world's largest tropical peatland, enabling the UK to retain world-leading expertise in understanding how the Earth functions as an integrated system and how humans are changing it.
SIOPACT - Developing a 250 Year Climate Record for the Southeastern Indian Ocean Sector of the Indo-PACific Warm Pool and its Global Climate Teleconnections
Deutsche Forschungsgemeinschaft (DFG), German Science Foundation
February 2019 - June 2024
Hedwig Krawczyk (PhD Researcher), Jens Zinke, Arnoud Boom (Supervisors), Paul Wilson and Bastian Hambach (NOC), Neal Cantin (AIMS)
Our main aim in the SIOPACT project was to develop >200 years of paired high-resolution (bimonthly and monthly-resolved) coral Sr/Ca (a proxy for SST; Zinke et al., 2019; Pfeiffer et al., 2019) and coral δ18O (proxy for SST and/or precipitation-evaporation balance/ocean advection; Zinke et al., 2004, 2008), in order to capture for the first time the societally important seasonal aspects of climate connectivity in sea surface temperature and salinity between the Pacific and Indian Ocean within the main Timor passage ITF ocean pathway. Coupled measurements of coral Sr/Ca ratios and δ18O allow reconstruction of past changes in δ18O of seawater, related to ocean advection and/or local air-sea interaction influencing sea surface salinity, by subtracting the thermal component of δ18O from the Sr/Ca-SST estimates (Zinke et al., 2004; Nurhati et al., 2011). This will also help to eliminate uncertainties with quantifying absolute SST from coral proxies Sr/Ca and δ18O. At the same time, we will develop a multi-decadal record of past salinity (inferred from δ18O of seawater) fluctuations extending far beyond current observational capabilities which only reach back to 1984 or extrapolated based on model-reanalysis data to 1958 (Feng et al., 2018). The bi-monthly proxy data will enable the reconstruction of seasonally stratified climate variability at the two locations since heat wave and hydrological events are strongly phase-locked to the annual cycle. We proposed to develop at least two coral chronologies from nearby sites with common climate forcing at bi-monthly resolution (~8-10 samples/year) to ensure sufficient replication of and confidence in the climate signals (Pfeiffer et al, 2009; DeLong et al., 2011). Browse Island and the Rowley Shoals (Clerke Reef) were chosen as target sites to generate (for both sites) 1 x ~250 years (1500 samples) and 1 x 100 years (500 samples; for replication of calibration period with instrumental SST data) bi-monthly records of trace element/Ca ratios at CAU Kiel (Sr/Ca, Mg/Ca, U/Ca, Ba/Ca ratios) and coral δ18O (at University of Southampton) with a total sample volume of 4000 samples. The Rowley Shoals cores were planned to be analysed at lower resolution (annual) to provide an additional dataset to verify the long-term interannual to multidecadal variability in SST and SSS.
Earth Indices: Evidence and Experiment
Haus der Kulturen der Welt – Anthropocene Working Group
December 2019 – December 2022
Corals are unique in the suite of proposed Anthropocene Global Boundary Stratotype Section and Point (GSSP) archives, as living organisms that produce aragonite exoskeletons preserved in the geological record that contain highly accurate and precise (<±1 year) internal chronologies. The GSSP candidate site North Flinders Reef in the Coral Sea (Australia) is an offshore oceanic reef, and therefore less vulnerable to local human influences than those closer to the coast. Here, we present geochemical records from two Porites sp. corals sampled at an annual to pluri-annual (i.e., 3 to 5 years) resolution that shows clear global and regional human impacts. Atmospheric nuclear bomb testing by-products (14C, 239+240Pu) show a clear increase in the Flinders Reef corals coincident with well-dated nuclear testing operations. By contrast, the radionuclides 241Am and 137Cs are present at low or undetectable levels, as are spheroidal carbonaceous fly-ash particles. Coral δ13C shows centennial variability likely influenced by growth effects in the 18th century and with a progression to lower values starting in 1880 and accelerating post-1970. The latter may be related to the Suess Effect resulting from 13C-depleted fossil fuel burning. Coral δ15N decreased between 1710 and 1954 with a reversal post-1954. Coral temperature proxies indicate prominent centennial variability with equally warm conditions in the 18th and end of 20th century. However, the exact mechanisms responsible for the mid-20th century changes in these parameters need to be scrutinised in further detail.
- University of Leicester, UK
- Molly Agg
- Arnoud Boom
- Adam Cox
- Sue Sampson
- Genevieve Tyrrell
- Jens Zinke
- Australian Institute of Marine Science, Australia
- Neal E Cantin
- Grace Frank
- Louisiana State University, USA
- Kristine L DeLong
- Kylie Palmer
- ETH Zurich, Laboratory of Ion Beam Physics, Switzerland
- Irka Hajdas
- Max Planck Institute for Chemistry (Otto Hahn Institute), Germany
- Nicolas Duprey
- Alfredo Martínez-García
- University College London, London
- Neil L Rose
- Lucy R Roberts
- Sarah L Roberts
- Simon D Turner
- Handong Yang
- University of Southampton, UK
- Andrew B Cundy
- Pawel Gaca
- James Andy Milton
Earth System Dynamics at the Dawn of Animal Life
The Leverhulme Trust
2023-2026
Mark Williams and Tom Harvey (Leicester), Alex Liu (Cambridge) and Alexandre Pohl (Bourgogne)
Climate change is coupled to the evolution of life and there is strong geological evidence that the two have co-evolved to make our planet habitable for over 3 billion years. One of the big changes evident in the fossil record between 600 and 500 million years ago is the evolution of the first animal-rich ecosystems in the oceans. We know from the geological record that this interval was associated with significant changes in Earth’s geography and climate, with the break-up of a major supercontinent, and interchanges between globally cold and warm climate states. But we do not know how, or even if, these changes influenced the development of the animal-rich biosphere.
In this project we use state-of-the-art climate models for the time interval when animals were first evolving. Our models are underpinned by global geological datasets that identify the climate state at key intervals of time in animal evolution. Our aim is to answer a fundamental question: How important was climate in driving the early evolution of animals, and is there any evidence that animals themselves could, through their behaviour, shape aspects of the climate system themselves?
Image: Earth geography and climate half a billion years ago. Adapted from work by Tom Wong Hearing and Alexandre Pohl. The names of oceans and continents are from the geological literature. White hexagons represent the ancient sedimentary rocks that served as “anchor points” for the climate reconstruction.
Enduring Impacts of Climate and People on the Landscape of South West Madagascar - Harnessing the Power of Automated Archaeological Remote Sensing in Holocene SW Madagascar
National Geographic Enduring Frontiers
June 2020 - December 2023
Kristina Douglass (Columbia University, US), Dylan Davies (Columbia University, US), Jens Zinke (University of Leicester) and Tanambelo Rasolondrainy (CEDRATOM, Madagascar)
We propose to use satellite-based remote sensing, machine learning and oxygen isotope records from fossil corals to reconstruct changes in human settlement distribution, land and sea use and climate in southwest coastal Madagascar over the course of the Holocene. We will develop open-source AI tools and build local capacity for reconstructing and analyzing long-term landscape-level human-climate-environment dynamics, so that our approach can be scaled up beyond southwest Madagascar. By harnessing the power of remote sensing and machine learning, this project will generate and analyze large datasets on long-term landscape change and past human response that are vital for understanding and successfully tackling critical threats to local livelihoods, cultural heritage and biological diversity today. Specifically, this project will provide a model, accessible tools and a common language for researchers, local communities and policy makers to reconstruct and understand landscape-level changes over time and clarify co-evolutionary relationships between people, climate and environment. Second, this project lays a critical foundation for testing theories about the adaptive role and importance of human mobility for climate resiliency, within current policy frameworks that favor sedentarization of rural communities. The power of AI to see landscape-level phenomena is key to understanding the scale of human action and developing strategies for a sustainable and resilient future.
Transforming our Understanding of the Indonesian Throughflow Variability and its Climatological Influence, Long Coral Records from Strategic Sites
NE/W001314/1 NERC Discovery Grant
April 2023-March 2026
Padmasini Behera (Postdoc, Leicester), Jens Zinke and Arnoud Boom (Supervisors, Leicester), Paul Wilson and Bastian Hambach (NOC), Neal Cantin (AIMS), Mathew England (UNSW)
The Indonesian Throughflow (ITF) current is the world’s only major oceanic current situated in the tropics. It propagates heat and salt-content anomalies from the Earth’s largest heat engine, the vast mass of warm and relatively low-salinity ocean surface waters of the Western Pacific Warm Pool, into the Indian Ocean and beyond. ITF variability is, therefore, deeply implicated in modulating sea-level-, climate- and weather-anomalies seasonally over southeast Asia, Australia and Africa, affecting biodiverse ecosystems and the lives of millions of people. However, our understanding of ITF variability is limited to instrumented observational records that date back to only 1984, and these records are often discontinuous. These data sets provide a basic understanding of the linkages between very recent ITF variability and extreme climate events such as the El Niño-Southern Oscillation (ENSO) and the Indian Ocean Dipole (IOD). However, because the record is so short, these analyses struggle to isolate the anthropogenically-forced component of change from natural variability. This problem severely limits our ability to make confident predictions of the response of the ITF and the Asian, Australian and African monsoons to 21st century human-induced warming. Longer records of ITF-variability are desperately needed.
This project will transform our understanding of ITF variability. It will reach back into the past to pre-industrial times by extending the records of ITF temperature and salinity from 1984 to 1770. Our analysis will be based on the chemistry of giant coral (Porites sp.) skeletons. These corals live in a “bulls-eye” location for ITF outlet waters, on the Northwest Australian margin where ITF-temperature and ITF-salinity anomalies are greatest and ITF-control is unequivocal (CfS). The coral cores were already collected and dated by annual growth band counting (by Project Partners Lough and Cantin). Our pilot study data from the youngest layers of the coral skeletons from our study site demonstrate that these corals faithfully record ITF-temperature and -salinity anomalies (CfS). Our new 240 year-long record of ITF temperature and salinity will be combined with numerical ocean model outputs (involvement of Project Partners England and Feng) and published monsoon records (e.g., tree-ring data sets from Australia, South East Asia and Africa) to revolutionise understanding of ITF-variability and its climatological influence.
Our project will make a step-change contribution to our knowledge of the relationship between south-eastern Indian Ocean and onshore-offshore Austral-Asian climate, including rainfall, floods, drought and bushfires. This is vital for preparing societies for a warming climate and an ever-increasing human footprint in the Austral-Asian-African monsoon regions.
Timing and Impact of Variable Arctic Outflow on the Atlantic Meridional Overturning Circulation
NE/X011690/1 NERC Exploring the Frontiers
November 2022 – April 2024
The North Atlantic Deep Water (NADW) constitutes the major deep oceanic current of the Atlantic Overturning Meridional Circulation (AMOC), distributing heat throughout the globe and affecting nutrients distribution sustaining marine ecosystems and human fishing grounds, whilst variations in its intensity have been linked to changes in productivity, current strength, regional and global climate throughout the Cenozoic.
The NADW is a mixture of deep currents originating from the Labrador Sea and the Arctic Ocean. Contribution from the latter component to NADW is largely influenced by the depth of the two gateways connecting the Nordic Sea to the North Atlantic basins on either side of Iceland. These gateways are shallower than 1500m and the volume of Arctic overflow will vary directly in response to changes in the relative depth of the passages. Lower sea level during glaciations and periodic increased mantle activity under Iceland, effectively raising land and seafloor, will directly constrain the influx of cold and dense Arctic Water, weakening the NADW and overall AMOC on geological time scales.
This project will exploit the capacity of ferromanganese crusts to record the signature of oceanic currents in their stratigraphy to track past changes in the strength, source, and provenance of the NADW components. Using samples from two key localities in the North Atlantic, high-resolution isotope records will be produced to investigate how the NADW varied over the last 20-30 Ma in response to glaciations and mantle pulses. These records will be dated using a new method for time calibration based on astronomical parameters and their influence on the chemical signature of these samples. The information gained will be used to inform future climate models predicting the response of this major oceanic current to disturbances.
Taking the Pulse of the Tropical Oceans Through the Coral’s Lens
RSWF\FT\180000, Royal Society Wolfson Fellowship
January 2019 - December 2023
Jens Zinke (University of Leicester), Arnoud Boom (University of Leicester), Walid Naciri (PhD Researcher, University of Leicester), Nicola Brown, Jennifer McIlvain, Bradley J. McDonald, Noreen Evans, Kai Rankenburg, (Curtin University (Perth, Australia), Ramasamy Nagarajan (Curtin University Malaysia)
My research focuses on understanding how the tropical oceans regulate our climate and to which extent global warming caused by humans has interrupted or modified the natural cycles. Additionally, I would like to decipher how tropical marine ecosystems, in particular coral reefs, respond to ever increasing pressures from unabated climate change and local human action. Such knowledge is crucial since tropical oceans cover 76% of the Tropics and its shallow coastal marine waters are of fundamental importance to human societies providing valuable services, especially for island nations and coastal populations that comprise 40% of the world’s population.
I have chosen to work with coral reefs because it is the tropical marine ecosystem that is most sensitive to current global warming and human impacts from coastal urbanization which is threatening their survival. The coral's sensitivity is due to their high adaptation to a specific narrow range of environmental conditions which includes temperature, nutrients and light levels. I research the impact of global warming, natural climate variability and human impacts with the help of massive, long-lived stony tropical corals. Corals build a robust skeleton that records climatic and environmental change throughout their multi-century life span in their chemical fingerprint. As such I develop monthly reconstructions of past and present tropical climate by far exceeding short instrumental observations. Such data fill an important knowledge gap in our understanding of the tropical ocean’s role in the climate system affecting billions of people far away from the Tropics and provide future references.
Climate Eparses
EP19R001-01 TAAF France
April 2019 – ongoing
Our proposed research program aims to provide solid environmental and climatic baseline data for the Eparses Islands on seasonal, interannual and decadal time scales from natural biogenic archives (massive corals). The aim is to fill an important data gap in meteorological and oceanographic processes through retrospective monitoring of climate and sea-level variability/long-term change from modern and fossil coral geochemistry and calcification records.
These data will serve as a guide for managers to identify key threats for future maintenance of biophysical conditions across the Eparses Islands. Our research will also address the affect of climate change and variability and oceanographic processes on the maintenance of coral growth and calcification in a low human footprint context of the Eparses.
We will target Tromelin, Glorieuse, Juan de Nova, Bassas da India and Europa Island where our French team members have GPS identified sites with highest potential for drilling living massive Porites sp. and Diploastrea sp. and Last Millenium/Holocene boulders from storm deposits. All sites are located along the climatologically important South Equatorial Current and Mozambique Channel ocean gateway, one of the crucial global surface ocean thermohaline circulation pathways.
We aim to provide a first comprehensive dataset on the marine climate history, recent sea-level change and coral calcification across the Eparses Islands, an ocean warming hotspot, for the Common Era and the Holocene.
MEMBRA: Understanding Memory of UK Treescapes for Better Resilience and Adaptation
NE/V021346/1 NERC Directed Grant (Treescapes)
Memory is the acquisition, retention and transmission of information guiding future action. It is used habitually by humans (e.g. in reading this Summary), is easily detectable in the behaviour of animals, and is the basis of machine-learning computer codes. Is it meaningful to talk of the 'memory of trees'? Science has established that plants can write and access a record of prior stress, making the phrase as meaningful to discuss as the selfishness of genes or the (artificial) intelligence of a machine. Writers and artists have responded creatively to how this new knowledge is altering our relationship with trees - perhaps most notably Richard Powers in the Pulitzer Prize winning novel 'The Overstory'. This co-ownership of the idea of tree memory across science and the Arts makes it particularly suited to the Future of UK Treescapes programme. Processes related to plant memory may be fundamental in allowing treescapes to swiftly adapt and therefore survive and thrive under the rapid environmental shifts of the Anthropocene.
The imprinting of memory in plants mostly happens by altering its epigenetic signature: i.e., changes that occur to the DNA that alter the activity of some genes, but that do not involve changes of the DNA sequence itself. The specific epigenetic marks known to be responsible for long-term memory, including transgenerational resistance, arise from the replacement of hydrogen atoms by methyl groups in the DNA base, cytosine. Very recently, a study in poplar has demonstrated transgenerational maintenance of epigenetic marks, pointing to a potential transmission of memory in long-lived perennial plants. However, whether long-lived plants can acquire, retain, and transmit memory from stress remains unknown. MEMBRA will study epigenetic changes and transgenerational memory as a result of stress in key tree species present in UK treescapes. Different abiotic (i.e. drought, frost, elevated CO2) and biotic stresses (i.e. insect infestation and disease) have been selected on the basis of our preliminary data showing that they alter important traits, e.g. growth, within UK treescapes. This project will also take into consideration past experiences. Trees have marks of past interactions with the environment recorded into their wood as they grow. We will analyse tree rings and isotope markers of drought stress to visualise a complete lifetime picture of the responses of trees to their past environment. The combination of ecology, tree rings and molecular techniques will allow us to assess the ability of previous stresses in improving the resilience of UK treescapes.
The understanding of stress memory in trees opens new paths to consider, e.g. the re-conceptualisation of environmental ethics and even tree consciousness. MEMBRA will collate information on how memory has been represented as a characteristic of trees. The project will study how this alters our understanding of the evolution and resilience of treescapes, how a consideration and appreciation of memory in trees can foster moral understanding of treescapes, and how the resulting ethical valuing of trees challenges a utilitarian and monetised 'ecosystem services' valuation of treescapes.
MEMBRA will provide the tools to identify the species and populations that will result in better resilience and adaptation and that will therefore be used in conservation and planting strategies. With the knowledge on how a consideration and appreciation of memory in trees can foster moral understanding of trees, we will identify and use new language to incorporate the concept of tree memory into policy-based initiatives. The final convergence of the results of MEMBRA will feed into policy development and the flourishing of the first stages of a memory-inspired intentional forest: the MEMBRA Treescape.
Unlocking the Full Potential of Tree Ring Oxygen Isotope Archives
Leverhulme REF 106609 Standard grant
Tree rings contain invaluable information on past climates locked away in the form of oxygen isotopes (oxygen atoms with different mass) in the wood. However, tree ring oxygen isotopes are a mixture of climate signals from rain and the plant. We will develop novel chemical procedures to separate these signals through analysis of isotope ratios at different positions in cellulose molecules and in other compounds from tree rings. This will fully unlock the potential of tree ring isotopes to provide better climate histories and reveal insights on plant physiological responses to climate and CO2.
PhD projects
Unlocking Western Tropical Indian Ocean temperature and hydroclimate back to the Little Ice Age and the Holocene, reconstructed from coral geochemistry
University of Leicester, CSE PhD stipend Manlin Zhang
September 2020 - September 2024
Indian Ocean sea surface temperatures (SST) play an important role in determining the location of precipitation over the tropics and the Indian Monsoon regions. Future global warming on land and ocean is thought to increase the land-ocean temperature gradient driving Indian Ocean circulation, leading to droughts or floods in adjacent biodiversity hotspots in Africa. Indian Ocean SST, precipitation and circulation anomalies also show a remote response to El Niño-Southern Oscillation (ENSO) events originating in the tropical Pacific. Atmospheric and oceanic teleconnections between the Indian Ocean and the Pacific are thought to vary on interannual to multidecadal time scales. A better knowledge of past SST and hydroclimate variability over the Indian Ocean and their interaction with the tropical Pacific over long time scales is needed to characterize natural climate variability. Such information is crucial in addressing the effects of anthropogenic climate change in light of such natural forcing factors.
Therefore, the aim of this project is to produce several centuries long climate records from and nearby Madagascar with bimonthly temporal resolution in order to unravel natural changes in the teleconnectivity with the tropical Pacific (abundant ENSO archives). Coral records dating back several centuries will be used for the reconstruction of past surface ocean variability under different boundary conditions, such as during the Little Ice Age (1600-1850). We therefore propose to build a 300-400 year modern climate and a Holocene (snapshots from fossil corals) data base for the western Indian Ocean biodiversity hotspot Madagascar and nearby islands through a seasonally resolved multi-proxy geochemical study of corals complimented by instrumental climate data analysis.
The coral data will be compared to terrestrial climate archives and instrumental climate data from Sub-Saharan Africa and Madagascar to assess past and present land-ocean climate teleconnections. Here, we collaborate with colleagues in Germany, South Africa and the USA who simultaneously work on tree ring, speleothem and lake level records. Furthermore, the paleoclimate data will be assessed in the context of newly discovered archaeological sites in Madagascar to shed light on the influence of climate variability and early human impact on megafaunal extinctions in Madagascar since the Little Ice Age.