6.5.1.4 What Are the Links Between Orbital Forcing and Mid-Holocene Monsoon Intensification?

Lake levels and vegetation changes reconstructed for the early to mid-Holocene indicate large precipitation increases in North Africa (Jolly et al., 1998). Simulating this intensification of the African monsoon is widely used as a benchmark for climate models within PMIP. When forced by mid-Holocene insolation resulting from changes in the Earth’s orbit (see Box 6.1), but fixed present-day vegetation and ocean temperatures, atmospheric models simulate NH summer continental warming and a limited enhancement of summer monsoons but underestimate the reconstructed precipitation increase and extent over the Sahara (Joussaume et al., 1999; Coe and Harrison., 2002; Braconnot et al., 2004). Differences among simulations appear related to atmospheric model characteristics together with the mean tropical temperature of the control simulation (Braconnot et al., 2002). As already noted in the TAR, the vegetation and surface albedo feedbacks play a major role in the enhancement of the African monsoon (e.g., Claussen and Gayler, 1997; de Noblet-Ducoudre et al., 2000; Levis et al., 2004). New coupled ocean-atmosphere simulations show that the ocean feedback strengthens the inland monsoon flow and the length of the monsoon season, due to robust changes in late summer dipole SST patterns and in mixed layer depth (Braconnot et al., 2004; Zhao et al., 2005). When combined, vegetation, soil characteristics and ocean feedbacks produce nonlinear interactions resulting in simulated precipitation in closer agreement with data (Braconnot et al., 2000; Levis et al., 2004). Transient simulations of Holocene climate performed with EMICs have further shown that land surface feedbacks are possibly involved in abrupt monsoon fluctuations (see Section 6.5.2). The mid-Holocene intensification of the North Australian, Indian and southwest American monsoons is captured by coupled ocean-atmosphere climate models in response to orbital forcing, again with amplifying ocean feedbacks (Harrison et al., 2003; Liu et al., 2004; Zhao et al., 2005).

6.5.1.5 What Are the Links Between Orbital Forcing and mid-Holocene Climate at Middle and High Latitudes?

Terrestrial records of the mid-Holocene indicate an expansion of forest at the expense of tundra at mid- to high latitudes of the NH (MacDonald et al., 2000; Prentice et al., 2000). Since the TAR, coupled atmosphere-ocean models, including the recent PMIP-2 simulations, have investigated the response of the climate system to orbital forcing at 6 ka during the mid-Holocene (Section 6.6.1, Box 6.1). Fully coupled atmosphere-ocean-vegetation models do produce the northward shift in the position of the northern limit of boreal forest, in response to simulated summer warming, and the northward expansion of temperate forest belts in North America, in response to simulated winter warming (Wohlfahrt et al., 2004). At high latitudes, the vegetation-snow albedo and ocean feedbacks enhance the warming in spring and autumn, respectively and transform the seasonal orbital forcing into an annual response (Crucifix et al., 2002; Wohlfahrt et al., 2004). Ocean changes simulated for this period are generally small and difficult to quantify from data due to uncertainties in the way proxy methods respond to the seasonality and stratification of the surface waters (Waelbroeck et al., 2005). Simulations with atmosphere and slab ocean models indicate that a change in the mean tropical Pacific SSTs in the mid-Holocene to conditions more like La Niña conditions can explain North American drought conditions at mid-Holocene (Shin et al., 2006). Based on proxies of SST in the North Atlantic, it has been suggested that trends from early to late Holocene are consistent with a shift from a more meridional regime over northern Europe to a positive North Atlantic Oscillation (NAO)-like mean state in the early to mid-Holocene (Rimbu et al., 2004). A PMIP2 intercomparison shows that three of nine models support a positive NAO-like atmospheric circulation in the mean state for the mid-Holocene as compared to the pre-industrial period, without significant changes in simulated NAO variability (Gladstone et al., 2005).