NOAA/WDS Paleoclimatology - East African 675 Year MCEOF Hydroclimatic Lake Proxy Data Synthesis

The recent decades-long decline in East African rainfall suggests that multidecadal variability is an important component of the climate of this vulnerable region. Prior work based on analysing the instrumental record implicates both Indian and Pacific ocean sea surface temperatures (SSTs) as possible drivers of East African multidecadal climate variability, but the short length of the instrumental record precludes a full elucidation of the underlying physical mechanisms. Here we show that on timescales beyond the decadal, the Indian Ocean drives East African rainfall variability by altering the local Walker circulation, whereas the influence of the Pacific Ocean is minimal. Our results, based on proxy indicators of relative moisture balance for the past millennium paired with long control simulations from coupled climate models, reveal that moist conditions in coastal East Africa are associated with cool SSTs (and related descending circulation) in the eastern Indian Ocean and ascending circulation over East Africa. The most prominent event identified in the proxy record - a coastal pluvial from 1680 to 1765 - occurred when Indo-Pacific warm pool SSTs reached their minimum values of the past millennium. Taken together, the proxy and model evidence suggests that Indian Ocean SSTs are the primary influence on East African rainfall over multidecadal and perhaps longer timescales. Synthesis of lacustrine hydroclimatic proxy records from East Africa using a Monte Carlo empirical orthogonal function (MCEOF) approach. MCEOF1 timeseries and spatial loadings plus 68% and 95% uncertainty bounds are reported. Site information for Lake proxy records utilized Lake Latitude Longitude Proxy Average DT References Challa -3.32 37.70 "BIT, dDwax, varve thickness" 33 "Verschuren et al. (2009), Nature; Tierney et al., (2011) QSR; Wolff et al. (2011) Science" Naivasha -0.77 36.35 Lake level reconstruction 3 "Verschuren et al. (2000), Nature" Victoria -1.00 33.00 % shallow water diatoms 5 "Stager et al. (2005), J. Paleolimnol." Edward -0.25 29.50 % Mg in authigenic calcite 4 "Russell and Johnson (2007), Geology" Tanganyika -6.70 30.00 Charcoal 10 "Tierney et al. (2010), Nature Geosci." Masoko -9.33 33.76 Magnetic susceptibility 10 "Garcin et al. (2006), Palaeo3; Garcin et al. (2007), J. Paleolimnol." Malawi -10.00 34.22 Terrigenous mass accumulation rate 6 "Brown and Johnson (2005), G3; Johnson and McCave (2008), Limnol. Oceanogr." Excel file of input proxy data also provided. It includes raw and interpolated proxy data from the seven sites used in the MCEOF synthesis. Note that provided ages correspond to the previously published age models. Original proxy data for all seven East African lakes are also archived at NOAA/NCDC/WDC Paleoclimatology: Lake Challa varve thickness: ncdc.noaa.gov Lake Challa dDwax: ncdc.noaa.gov Lake Naivasha lake levels: ncdc.noaa.gov Lake Victoria SWD: ncdc.noaa.gov Lake Edward %Mg: ncdc.noaa.gov Lake Tanganyika Charcoal: ncdc.noaa.gov Lake Masoko Mag Sus06: ncdc.noaa.gov Lake Masoko Mag Sus07: ncdc.noaa.gov MCEOF1 site loadings are as follows: Lake Median Loading Lower 95% Lower 68% Upper 68% Upper 95% Challa 0.54 0.29 0.48 0.58 0.62 Naivasha 0.17 -0.41 -0.18 0.38 0.48 Victoria -0.19 -0.51 -0.41 0.15 0.39 Edward -0.43 -0.55 -0.51 -0.26 0.07 Tanganyika -0.20 -0.46 -0.37 0.01 0.35 Masoko -0.24 -0.45 -0.34 -0.10 0.11 Malawi -0.50 -0.29 -0.44 -0.54 -0.59 For Lake Challa, the first principal component of three different hydroclimate proxies was used to represent the site (PC1). All three raw proxy time series are provided in addition to the interpolated series and PC1. The age model used for this site is the varve-based age model of Wolff et al., 2011. Please refer to the Supplementary Material in Tierney et al. (2013) for a description of how PC1 was calculated. For Lake Masoko, magnetic susceptibility from the longer core (Garcin et al. 2006) was used incorporating the additional chronological controls from Garcin et al. 2007. See Supplementary Information and Anchukaitis and Tierney (2012) Clim. Dyn. for more information. No depth data are available from the Lake Malawi core. As this is a varve sequence that does not depend on depth-age translation, depth is not needed to iterate time uncertainty. See Supplementary Information and Anchukaitis and Tierney (2012) for information on how varve counting uncertainty was modeled.