Session: Benchmarking & cross-cutting Group 1 (Isotope modelling, COMPARE)
Author: Stephan Woodborne / email@example.com / iThemba LABS, Johannesburg, South Africa; Stable Isotope Laboratory, Mammal Research Institute, University of Pretoria, Pretoria, South Africa
Co-author: Qiong Zhang, Department of Physical Geography and Bert Bolin Centre for Climate Research, Stockholm University, Stockholm, Sweden;
Grant Hall, Stable Isotope Laboratory, Mammal Research Institute, University of Pretoria, Pretoria, South Africa;
Tamryn Hamilton, Department of Animal, Plant, and Environmental Sciences, University of the Witwatersrand, Johannesburg, South Africa;
Estelle Razanatsoa, Plant Conservation Unit, University of Cape Town, Cape Town, South Africa;
Adrian Patrut, Babeş-Bolyai University, Faculty of Chemistry and Chemical Engineering, Cluj-Napoca, Romania;
Roxanna Patrut, Babeş-Bolyai University, Faculty of Biology and Geology, Cluj-Napoca, Romania;
Christiaan Winterbach, Tau Consultants, Maun, Botswana;
Stephan Winkler, Tau Consultants, Maun, Botswana;
Edaphic soil moisture potential is the main determinant of leaf-level carbon isotope discrimination in savanna trees of southern Africa, and a proxy rainfall record can be obtained from radial (time series) analysis of carbon isotope ratios in trees in the region. The approach has been tested in baobab trees (Adansonia digitata, A. za, and A. grandidieri), Black Monkey Thorn (Acacia burkei) and Camelthorn (Acacia erioloba) trees. This diversity of species allows the proxy to be applied across a range of xeric conditions from the arid Namib Desert in the west, through the Kalahari Desert to Madagascar. In the mesic regions a record has been generated from Yellowwood (Afrocarpus falcatus) trees. The lack of annual rings in the stems of most of these tree species makes it necessary to generate age models using radiocarbon dates which introduces a degree of error in the age assigned to each sample. In general the sampling resolution is sub-annual, but the age error only allows decadal to centennial trends to be inferred. The outcome is a time/space matrix of rainfall variability in the region over the last 1000 years. For each sample site a composite record is generated from multiple trees. Comparison between the carbon isotope ratio proxy and the short, patchy coverage with instrumental records provides strong support for the authenticity of the proxy record.
The tree record indicates synoptic scale variability in response to climate forcing. In some instances the rainfall anomalies have the same sign across the entire region, and in others there is a clear dipole response with opposite sign anomalies in different regions. The underlying forcing is attributed to north/south and east/west displacement of the main rainfall systems. These dynamics provide a tangible basis for testing model climate simulations. The EC-Earth last millennium simulation of these displacements in response to the inferred climate forcing well matches the pattern observed in the tree records. The result suggests that future climate change scenarios for southern Africa are accurately captured in the climate simulation models.