Feedbacks of soil properties on vegetation during the Green Sahara period

Abstract

During the early to middle Holocene, the Sahara received enhanced precipitation and was covered by steppe-like vegetation with a large-scale hydrographic network of lakes, wetlands and fans, which is known as the Green Sahara (GS). However, most coupled land-atmosphere models underestimate the precipitation and vegetation cover, suggesting that critical atmospheric or land surface processes are lacking in those models. Climate-induced vegetation cover change can modify soil texture and physical properties over the long term, which in turn have feedbacks on vegetation. In this study, we examine five plausible soil-vegetation processes in a land surface model, which are expected to increase soil moisture for plants and possibly sustain equilibrium vegetation for a lower rainfall level. The annual precipitation required during the GS epoch to match the modelled vegetation distribution with paleorecords is inferred. Results demonstrate that these soil-vegetation processes have strong positive impacts on vegetation and soil moisture, especially the increase of soil evaporative resistance. After including all soil feedbacks on vegetation, the model requires only a mean precipitation of~400 mm/yr to reproduce the pollen-inferred GS vegetation, instead of~600 mm/yr when no soil feedback is included. From the mid-Holocene to pre-industrial period, we infer that terrestrial carbon stocks decrease by~58 PgC due to the removal of carbon in vegetation, soil and litter pools of the GS. This work highlights the importance of soil-vegetation interactions for simulating dry-region vegetation coverage in models, and the impacts of natural land cover change on carbon budgets in the geological past.

Abrupt change in North African hydroclimate and landscape evolution 3.2 million years ago

“…Major biome changes in tropical Africa have been simulated without a total annual precipitation change, simply by altering rainfall seasonality 45 . Moreover, inclusion of soil feedbacks in GSP simulations can reproduce pollen-inferred vegetation shifts at around 400mm/yr mean precipitation, relative to ~600 mm/yr in the absence of soil feedbacks 46 .…”

“…African pollen evidence attests to a more arid Pleistocene compared to the Pliocene 32,71 , and a significantly more erodible landscape must exist before a stepwise amplitude increase in fluvial suspended sediment can occur. Furthermore, North African desert-soil albedo and vegetation feedbacks strongly amplify rainfall variability 46,72,73 Figure 2 ODP967 composite core image. Northern and Southern Hemisphere ice-volume changes (NH, red; SH, blue), in equivalent sea-level fall, and June 21 insolation at 65N (yellow) illustrate timing relationships between climate forcing and lithology.…”

Wetlands of North Africa During the Mid‐Holocene Were at Least Five Times the Area Today

“…The parameters of this TOPMODEL-based approach were calibrated for each 0.5° grid cell to reproduce the present-day wetland fractions and then the ORCHIDEE-PEAT model was forced by mid-Holocene climate fields to simulate mid-Holocene wetland areas. The boundary conditions of the simulation (i.e., soil texture), other than climate forcing and TOPMODEL parameters, are the same as Chen et al (2020).…”

Using the mid-Holocene 'greening' of the Sahara to narrow acceptable ranges on climate model parameters

“…In contrast, most GCMs have a resolution of around 150-200 km. Processes such as convection must therefore be parametrized, but all parametrizations are approximations, and the structure of many parametrizations is unavoidably far simpler than reality, leaving no way to choose a parameter value a priori or directly from observations (e.g., Stensrud, 2007) Paleoclimate changes can provide valuable "out-of-sample" tests for modeling the climate system (Harrison et al, 2015;Schmidt et al, 2014;Tierney et al, 2020;Valdes, 2011;Zhu et al, 2020). This is because past climate states provide examples of both fast and gradual changes that are larger in amplitude than historical climate changes and are therefore more comparable in amplitude to expected future climate change.…”

“…Moreover, the reduction in dust loading would have altered cloud formation through dustcloud interactions, and this has been shown to reduce stratiform precipitation (Thompson et al, 2019). Land-surface feedbacks can efficiently drive the monsoon northwards (Levis et al, 2004;Skinner & Poulsen, 2016;Texier et al, 2000) but there is little agreement on how the "greening" of the Sahara should be configured in models (Chen et al, 2020;Hopcroft et al, 2017;Lu et al, 2018;Street-Perrott et al, 1990;Texier et al, 2000). It is thus not trivial to judge whether or not a sufficient precipitation enhancement is achieved for the right reasons in model simulations of the mid-Holocene.…”