Cenozoic grassland expansion

Today, grasslands cover more than 20% of the land on the planet. But grass-dominated ecosystems are geologically young, thought to first emerge some 30-40 million years ago. Why grasslands took so long to expand, and the conditions that drove their expansion, are central questions in the geosciences. The answer to these questions will inform topics such as grassland biogeography, paleo-ecological resilience, and mammalian evolution.

Fig. 3 of Edwards et al. (2010) shows the timing of C3 and C4 grassland expansion in different parts of the world. Two take-aways are (1) grass-dominated ecosystems are (geologically) recent; (2) grassland expansion is consistently unidirectional.

Between the late Oligocene-early Miocene closed forest habitats were largely replaced by open C3 woodlands, grasslands, and scrublands (e.g. Strömberg, 2005). This is one of the earliest confirmed open habitat expansions of the Cenozoic and its cause(s) remains debated. We are building large oxygen isotope and phytolith datasets to test possible climate-vegetation links across the western U.S. Work is ongoing.

Today, biomass in the western US (here between 40-47°N) has different sensitivities to precipitation seasonality. Biomass generally decreases with the winter rainfall fraction but correlates positively with summer rainfall when the winter fraction is already high (see work [not by me] of Berkelhammer et al., 2017; 2021). We are studying how the winter fraction controlled western U.S. open habitat expansion in the late Oligocene-early Miocene.

Whether C4 biomass increases or decreases with precipitation remains a topic of debate and may carry implications for the Neogene expansion of C4 grasslands. Today, most C4 grasslands thrive when potential evapotranspiration (PET) exceeds precipitation (P), but not by too much, suggesting precipitation can increase or decrease C4 biomass depending on initial conditions. The scale(s) of space and time where this water balance-C4 biomass link is relevant, however, is unclear.

What caused late Miocene wave(s) of C4 grassland expansion? Usually, it’s attributed to declining atmospheric CO2 levels or declining (or increasing) precipitation. The debate is central to our understanding of grassland dynamics. We are addressing these questions with monsoon energy balance modeling (e.g. Boos & Korty, 2016) and comparison to ESMs. Work is ongoing. Check out some exciting recent work (not by me) from Karp et al., 2018, Karp et al., 2021, and Polissar et al., 2021. (Above data from: Quade et al., ‘89; Herbert et al., ‘16; Polissar et al., ‘19).