Canopy structure in Late Cretaceous and Paleocene forests as reconstructed from carbon isotope analyses of fossil leaves

Abstract

While modern forests have their origin in the diversification and expansion of angiosperms in the Late Cretaceous and early Cenozoic, it is unclear whether the rise of closed-canopy tropical rainforests preceded or followed the end-Cretaceous extinction. The “canopy effect” is a strong vertical gradient in the carbon isotope (δ13C) composition of leaves in modern closed-canopy forests that could serve as a proxy signature for canopy structure in ancient forests. To test this, we report measurements of the carbon isotope composition of nearly 200 fossil angiosperm leaves from two localities in the Paleocene Cerrejón Formation and one locality in the Maastrichtian Guaduas Formation of Colombia. Leaves from one Cerrejón fossil assemblage deposited in a small fluvial channel exhibited a 6.3‰ range in δ13C, consistent with a closed-canopy forest. Carbon isotope values from lacustrine sediments in the Cerrejón Formation had a range of 3.3‰, consistent with vegetation along a lake edge. An even-narrower range of δ13C values (2.7‰) was observed for a leaf assemblage recovered from the Cretaceous Guaduas Formation, and suggests vegetation with an open canopy structure. Carbon isotope fractionation by Late Cretaceous and early Paleogene leaves was in all cases similar to that by modern relatives, consistent with estimates of low atmospheric CO2 during this time period. This study confirms other lines of evidence suggesting that closed-canopy forests in tropical South America existed by the late Paleocene, and fails to find isotopic evidence for a closed-canopy forest in the Cretaceous. INTRODUCTION Closed-canopy tropical forests are the most diverse modern biome and can drive water, carbon, and climate dynamics at continental and global scales (Burnham and Johnson, 2004). Although tropical rainforests comprise only ∼12% of Earth’s surface, they account for ∼45% of the carbon in terrestrial biomass (Watson et al., 2000; Malhi et al., 2002). These forests help maintain consistent temperatures and the wet conditions (mean annual precipitation ≥2000 mm/yr) to which they are adapted via their low albedo and massive movement of transpired water across continents, both of which influence large-scale atmospheric circulation and temperatures (Bastable et al., 1993; Betts, 1999; Bonan, 2008; Boyce et al., 2010). It is not well understood when angiospermdominated closed-canopy tropical forests first developed, and estimates of their origin range from the mid-Cretaceous to the early Paleogene (Burnham and Johnson, 2004). Time-calibrated molecular phylogenetic trees constructed for extant angiosperms place the modern tropical rainforest lineages as far back as 100 Ma and could indicate that angiosperm-dominated, closed-canopy forests have been present since the mid-Cretaceous (Soltis and Soltis, 2004; Davis et al., 2005), except that fossils documenting the morphological and ecological traits common to canopy-forming angiosperms are rare until the Paleocene (Bruun and Ten Brink, 2008; Herrera et al., 2014). Further, leaf features that indicate dense canopy can reflect multiple drivers, leaving few empirical tools that can be used to assess ancient forest structure (Beerling and Royer, 2002; Feild et al., 2011; Carins Murphy et al., 2014). In modern forests, it has been observed that the stable carbon isotope composition of leaves (δCleaf) declines strongly downward from upper canopy to understory (Vogel, 1978). This “canopy effect” provides a promising approach that could be applied to relatively common leaf compression fossils. If this isotope gradient is preserved in fossils, it would allow canopy placement to be estimated for fossil leaves and leaf fragments. Three major mechanisms contribute to the canopy effect. (1) High rates of respiration by soil biota combined with restricted atmospheric mixing create elevated CO2 concentrations and 13C-depleted CO2 (δCatm) in the understory (Brooks et al., 1997; Medina and Minchin, 1980). (2) Higher humidity lower in the understory permits stomata to remain open without loss of leaf water, resulting in a fuller expression of 13C fractionation during enzymatic carbon fixation (Δleaf; Ehleringer et al., 1986; Madhavan et al., 1991). (3) High light in the upper canopy increases the rate of photosynthesis up to four times that of leaves in the understory, and leads to less 13C discrimination (Zimmerman and Ehleringer, 1990; Hanba et al., 1997). As a result of these pronounced gradients in CO2, water, and light, closed-canopy forest δCleaf values range as much as 10‰ from the sun-lit canopy top to the dark and humid understory. A Monte Carlo–style leaf resampling model from closed-canopy forest litter has shown that the wide diagnostic range of δCleaf values unique to the closed-canopy forest can be found by carbon isotope measurements from as few as 50 leaves (Graham et al., 2014). Here we use δCleaf to estimate canopy structure in leaf fossil assemblages from the Maastrichtian Guaduas Formation and Paleocene Cerrejón Formation of Colombia. We also use fossil δCleaf data in Downloaded from https://pubs.geoscienceworld.org/gsa/geology/article-pdf/4825714/g46152.pdf by Northwestern University user on 09 September 2019 2 www.gsapubs.org | Volume 47 | Number XX | GEOLOGY | Geological Society of America combination with the predicted δ13C values of paleoatmospheric CO2 to determine if photosynthetic fractionation (Δleaf) differs greatly between these leaves and their modern descendants. δCleaf values reflect the source CO2 composition as well as the carbon isotope discrimination occurring during photosynthesis. Fractionation is subject to environmental influences and genetic factors that affect isotopic expression trends (Hubick et al., 1990). Comparison of the Δleaf values from modern plants with those of their fossil ancestors would indicate how conserved these fractionation trends are within plant families. MATERIALS AND METHODS This study compares δCleaf from three fossil assemblages with that of leaves from modern forests to determine if the δCleaf range preserved in the fossil cuticles is consistent with a closed canopy (see representative specimens in Fig. 1). All three fossil assemblages were collected in Colombia (Fig. 2): two from the Paleocene Cerrejón Formation and one from the Late Cretaceous (Maastrichtian) Guaduas Formation. Both Cerrejón assemblages include many of the families dominant in modern closed-canopy forests of the Neotropics (e.g., Fabaceae, Arecaceae, Lauraceae), and physiognomic leaf features—size, entire margins, vein density—that indicate a closed-canopy, multi-layered rainforest (Wing et al., 2009; Herrera et al., 2011). In contrast, the Guaduas Formation paleoflora neither is physiognomically similar to contemporary closed-canopy communities nor includes taxa assigned to extant families dominant in Neotropical rainforests (Guierrez and Jaramillo, 2007). These assemblages were selected in order to compare reconstructed canopy isotope gradients before and after the events of the CretaceousPaleogene mass extinction. The Cerrejón Formation is a coal-bearing fluvial unit widely exposed in terraces of the Cerrejón Mine on the La Guajira Peninsula. Palynofloral assemblages indicate a middle to late Paleocene age, ca. 58–60 Ma (Jaramillo et al., 2007). The formation consists of a variety of lithologies (sandstones, mudstones, and coals) deposited in a mosaic of fluvial and lacustrine settings typical of an estuarine coastal plain. Leaf margin and size analyses indicate a mean annual precipitation (MAP) of 2.3–4.6 m/yr and mean annual temperature (MAT) of 24–31°C (Wing et al., 2009). The two Cerrejón localities (Site 0315 at 11°8′6′′N, 72°34′12′′W; Site 0318 at 11°7′41′′N, 72°33′18′′W) were separated by <2 km laterally and 100 m stratigraphically. Fossils from both localities were collected from small areas (4–6 m2) that represent distinct terrestrial settings. Leaves from site 0315 were deposited in heterolithic sediments suggestive of a lowto medium-energy channel. Sampled leaves were associated with ten morphotypes from nine families, as well as a selection of taxonomically indeterminate non-monocot (magnoliid or eudicot) angiosperm leaves. At site 0318, leaves were collected from a laterally extensive, thinly bedded, flat-laminated siltstone interpreted as a shallow lake deposit. Sampled leaves included 19 morphotypes from 10 families as well as a selection of indeterminate non-monocot leaves. Herrera et al. (2008) and Wing et al. (2009) described family identification and morphotype assignment. Late Cretaceous fossils were collected from the middle Guaduas Formation of Boyacá Department (5°55′45′′N, 72°47′43′′W). Palynoflora indicate an age of ca. 68–66 Ma (Muller et al., 1987). Leaf margin and size analyses indicate an estimated MAT of 22.1 ± 3.4°C and MAP of ∼2.4 m/yr (Gutierrez and Jaramillo, 2007). At the time of deposition, this location was a coastal plain similar to that of the Cerrejón Formation (Gutierrez and Jaramillo, 2007). Fossil leaves from the Guaduas Formation were taken from laminated and massive mudstones with sandstone interbeds above fine-grained sand beds intercalated with coal seams (Guerrero, 2002). Most angiosperm leaves from the Guaduas flora can only be described as indeterminate dicots coexisting with abundant gymnosperms in a community that has no modern analog. One leaf for this study could be assigned to a family, and nine others could be assigned to one of five morphotypes (Gutierrez and Jaramillo, 2007). A B C