Vegetation cover and hydroclimate are of great importance for the world's water supply, agriculture, and biological diversity. Accurate projections of their response to future emission scenarios are required to design effective mitigation and adaptation strategies. However, the response of hydroclimate and vegetation to radiative forcing variations is far less understood and more uncertain than temperature changes (Stocker et al., 2013). Paleoclimate proxies from natural archives, which extend far beyond instrumental records, facilitate the evaluation of Earth System Models under climate conditions different from present-day (Braconnot et al., 2012). Thereby, they provide guidance for improved representations of the hydrosphere and biosphere in complex numerical models and contribute to tighter constraints on future projections of moisture availability and land cover (Harrison et al., 2015).

<p>At the current decarbonization rate, we are set on a path towards re-shaping a substantial share of land for carbon dioxide removal (CDR) over the following decades. However, existing Earth system models which could help to quantify the character of resulting CDR side effects and their consequences for the cumulative CO<sub>2</sub> removal do not yet resolve dynamic CDR cover in space. Here, we embark on shedding light on this CDR uncertainty space, scrutinizing CDR impacts in spatial simulations with a comprehensive Earth system model. Assuming CDR to be driven by solar irradiation in the style of photovoltaics, our model is the first to simulate an idealized approach to land-based CDR with its physical, biospheric, and land use couplings on a grid box scale. We analyze dynamic CDR simulations for spatial deployment scenarios according to the country-wise burden of past CO<sub>2</sub> emissions, to livelihood constraints, and to optimal irradiation conditions. Shared socio-economic pathways drive the overall global CDR use for a range of potential future emission scenarios. Aside from these spatio-temporal scenarios, the simulations also cover different ways of releasing excess energy from the solar-to-carbon conversion, permitting either local cooling through carbon storage, heat dissipation resulting from system losses or co-benefits for energy production. Based on simulation ensembles for the different scenarios, we quantify Earth system impacts of CDR and their consequences for CO<sub>2</sub> removal if grid-scale feedbacks are properly resolved. With new spatially resolved CDR representations in Earth system models we will be able to test CDR-induced Earth system dynamics and CDR promises in greater detail than with existing globally forced projections. This spatially explicit modeling strategy could also open a way toward more comprehensive modeling strategies which include consequences for land use decisions on CDR.</p>

<p>Vegetation responds to local climate and carbon dioxide changes with response times ranging from decades to millennia, depending on location, spatial scale, and vegetation characteristic. Here, we focus on orbital timescales, for which all available estimates suggest an equilibrium of vegetation and climate. Over the course of the last glacial period, global mean temperature varied between minima during Marine Isotope Stage (MIS) 4 and MIS2, and a maximum in MIS3. If orbital-scale climate changes followed this global trend across most of the globe, we would expect vegetation changes to feature a similar temporal evolution.</p> <p>Leveraging a global compilation of pollen records, we quantify the synchronicity of orbital-scale vegetation changes within and across regions during the last glacial period. We use the arboreal pollen fraction, statistical mode decompositions, and key taxa as indicators for forest cover changes. Our results suggest that a globally coherent forest cover minimum occurred during MIS2. However, we do not find evidence for other periods of coherent forest cover trends across the globe or either hemisphere. This indicates that vegetation changes were more regionally confined during earlier parts of the last glacial. As chronologies become more uncertain further back in time, we examine the likelihood of dating errors to explain the absence of globally coherent vegetation changes during MIS4 and MIS3. Finally, we compare our results with simulations of climate and vegetation to assess if models capture the diagnosed forest cover trends found in the pollen records. Moreover, this comparison allows us to infer the influence of temperature, moisture availability, and carbon dioxide on vegetation variations during the last glacial period.</p>