African biomes are most sensitive to changes in CO<sub>2</sub> under recent and near-future CO<sub>2</sub> conditions

Abstract

Abstract. Current rates of climate and atmospheric change are likely higher than during the last millions of years. Even higher rates of change are projected in CMIP5 climate model ensemble runs for some Representative Concentration Pathway (RCP) scenarios. The speed of ecological processes such as leaf physiology, demography or migration can differ from the speed of changes in environmental conditions. Such mismatches imply lags between the actual vegetation state and the vegetation state expected under prevailing environmental conditions. Here, we used a dynamic vegetation model, the adaptive Dynamic Global Vegetation Model (aDGVM), to study lags between actual and expected vegetation in Africa under a changing atmospheric CO2 mixing ratio. We hypothesized that lag size increases with a more rapidly changing CO2 mixing ratio as opposed to slower changes in CO2 and that disturbance by fire further increases lag size. Our model results confirm these hypotheses, revealing lags between vegetation state and environmental conditions and enhanced lags in fire-driven systems. Biome states, carbon stored in vegetation and tree cover in Africa are most sensitive to changes in CO2 under recent and near-future levels. When averaged across all biomes and simulations with and without fire, times to reach an equilibrium vegetation state increase from approximately 242 years for 200 ppm to 898 years for 1000 ppm. These results have important implications for vegetation modellers and for policy making. Lag effects imply that vegetation will undergo substantial changes in distribution patterns, structure and carbon sequestration even if emissions of fossils fuels and other greenhouse gasses are reduced and the climate system stabilizes. We conclude that modelers need to account for lag effects in models and in data used for model testing. Policy makers need to consider lagged responses and committed changes in the biosphere when developing adaptation and mitigation strategies.

Climate change promotes transitions to tall evergreen vegetation in tropical Asia

“…In these regions, environmental conditions showed higher interannual variation and stochastic processes in aDGVM2, such as fire occurrence or demographic processes, had a greater effect on vegetation dynamics. This model behaviour confirms a previous aDGVM result indicating that fire‐driven and open ecosystems are more variable and take longer to reach an equilibrium state with prevailing environmental conditions than forest (Scheiter et al., 2020).…”

“…Modelled transition rates of biomes and dominant phenology types might be overestimated compared to real transition rates. While it has been shown that species and biomes might not be able to keep pace with climate change (Loarie et al., 2009; Scheiter, Moncrieff, Pfeiffer, & Higgins, 2020), most DGVMs ignore seed dispersal and associated migration lags (Corlett & Westcott, 2013). Models assume that local seed pools contain all functional types or ecological strategies that persisted during model spin‐up, and that climate change or other disturbances influence the relative abundance of these types.…”

“…This modelling approach represents an establishment bottleneck that prevents or delays invasion of inferior ecological strategies or functional types into local communities, but at the same time it represents a full dispersal scenario (Thuiller, Lavorel, Araujo, Sykes, & Prentice, 2005). While this caveat is well known (Blanco et al., 2014; Corlett, 2009; Corlett & Westcott, 2013; Scheiter et al., 2020), it has rarely been addressed in DGVMs (but see Sato & Ise, 2012). Reasons include the mismatch between the coarse resolution of DGVM studies (typically 0.5° or coarser) and shorter dispersal distances, and knowledge gaps about dispersal distances, probabilities and pathways (Corlett, 2009).…”

“…While it has been shown that species and biomes might not be able to keep pace with climate change (Loarie et al, 2009;Scheiter, Moncrieff, Pfeiffer, & Higgins, 2020), most DGVMs ignore seed dispersal and associated migration lags (Corlett & Westcott, 2013).…”

Climate change will cause non-analog vegetation states in Africa and commit vegetation to long-term change

“…In addition, delayed responses to environmental drivers may unexpectedly push vegetation beyond tipping points towards alternative stable states long after the change in forcing has occurred. Particularly in connection with African savanna ecosystems, such multi-stable ecosystem states have been proposed and studied by a variety of authors (e.g., Staal et al, 2016;Li et al, 2019;Pausas and Bond, 2020). Conservation management needs to be aware that the vegetation state at any given time may not be the vegetation state expected under prevailing environmental conditions, and managers need to decide whether to preserve the status quo or allow vegetation development towards its anticipated equilibrium state.…”