Projets de recherche réalisés
1. Mid-Late Miocene and Early Pliocene grass communities of the Central Andean Plateau (CAP)
|
The Central Andean Plateau was about ~1700 m.a.s.l. in the Middle Miocene and had reached near-modern elevations (~4000 m.a.s.l.) by the Early Pliocene. The paleobotanical record of the Descanso formation, in the Peruvian Altiplano (Cuzco region) was recently investigated by Martinez et al. (2020) who documented a shift from of a montane ecosystem in the Miocene to Puna-like vegetation in the Early Pliocene, this latter representing the earliest evidence of a Puna-like ecosystem. Climate inferences from fossil evidence suggest wetter conditions in the Miocene and modern precipitation levels in the Pliocene. The processes that lead to the evolution of modern high-elevation Andean ecosystems (Paramo, Puna, austral alpine vegetation) are debated, and wide range of climatic (e.g., hydric gradients), geographic and ecological factors have been hypothesized to have led to floristic differences between these ecosystems. I used phytolith assemblages to better characterize the Neogene plant communities of the Descanso formation, and shed further light on the origin of the modern day Puna in the CAP. I also extracted phytoliths from modern soil surface samples and plant specimens collected in different high elevation Andean ecosystem (e.g. Puna, Paramo) in order to compare them with the Descanso formation phytolith assemblages and increase the accuracy of vegetation reconstructions. Preliminary results comfirm the establishment of modern-like grassland ecosystems since the Pliocene. Furthermore, grass phytoliths form the Descanso formation (from both Miocene and Pliocene members) are unique in their morphology, suggesting the occurrence of either early-diverging (propably extinct or currently rare) grass species, or the occurrence of taxa for which the phytolith morphology of modern representatives is currently unknown or differs from that of their ancestors. This article is in press in the journal Paleobiology |
|
2. Using C3:C4 grasses to build a paleotherometer for the Northern Andes
|
|
I worked on this project at the ISEM (in Monptellier) in collaboration with Laurent Bremond, Juan Carlos Berrio, Arnoud Boom (University of Leicester), and Diego Cañas (Universidad Nacional de Colombia).
I used the data collected by my collaborators to build a transfer function in that predicts temperature from grass phytolith assemblage composition. Grass phytoliths were extracted from soil samples collected along an altitudinal and temperature gradient in the Northern Andes (Colombia). The proportion of grass C3/C4 phytoliths shows a strong negative correlation with temperaure. I used logistic regression to model this relationship. The results of this work have been submitted for publication and are currently under review. |
1. Sampling phytoliths in modern analogue studies
|
To improve paleoenvironmental reconstruction based on fossil phytoliths, it is vital that modern modern calibration studies rely on methods applicable to the deep time fossil record. However, soil phytolith inventories from extant vegetation types, representing modern analogues suffer from several methodological shortcomings limiting inter-study comparisons, and the development of a single, and repeatable protocol for soil sample collection.
In the first chapter of my dissertation I addressed two main methodological questions, regarding the number of phytolith samples required to capture the vegetation signal in a small area, and the soil level to be sampled in a dry forest and a rainforest in Costa Rica. I compared soil phytolith assemblages from several samples collected in a small quadrat (10 x 10 meters) and from at two different soil depths (in the upper and lower soil A-horizon). Results suggest that the use of a single point samples from the lower A-horizon of paleosols, an approach typically used in deep-time paleoecology is valid for the two studied vegetation types. The use of this approach is strongly justified because a) the upper section of the A-horizon is often eroded in paleosols, and b) because point samples can be time-averaged, the combination of multiple point samples might amplify time-averaging rather than providing more accurate reconstructions of the vegetation at a given time. |
Soil sampling
Phytoliths from La Selva Rainforest
|
2. Spatial resolution of the phytolith record
Vegetation structure plays a key role for ecosystems (e.g., it contributes to the water cycle through evapotranspiration, and to the carbon carbon cycle as a carbon sink; it influence the albedo; it controls fire regimes and erosion; it is a source of soil nutrients, and controls moisture and temperature in soils; it is a source of food and habitat for animals; it influences plant and animal species composition and diversity...)
In the fossil record, direct (i.e. paleobotanical) evidence for habitat structure is rare.
Phytolith analysis has emerged in the last two decades as one of the most promising tools for vegetation reconstruction because phytoliths are thought to represent a local vegetation signal. My PhD research focused on understanding which aspect of the vegetation (structure or openness, composition, diversity, heterogeneity in this characteristics) can be reconstructed using phytoliths.
I analyzed phytolith assemblages form modern soils collected along vegetation transects through Palo dry forest and La Selva rainforest in Costa Rica. I compared soil phytolith assemblages with the standing vegetation, its Leaf Area Index (LAI) -a measure of canopy openness-, and tree species composition.
In the fossil record, direct (i.e. paleobotanical) evidence for habitat structure is rare.
Phytolith analysis has emerged in the last two decades as one of the most promising tools for vegetation reconstruction because phytoliths are thought to represent a local vegetation signal. My PhD research focused on understanding which aspect of the vegetation (structure or openness, composition, diversity, heterogeneity in this characteristics) can be reconstructed using phytoliths.
I analyzed phytolith assemblages form modern soils collected along vegetation transects through Palo dry forest and La Selva rainforest in Costa Rica. I compared soil phytolith assemblages with the standing vegetation, its Leaf Area Index (LAI) -a measure of canopy openness-, and tree species composition.
|
|
3. Change in vegetation and habitat structure during the Middle Miocene Climatic optimum of Southern Patagonia
Cañodón de las Vacas, Santa Cruz Formation
The Santa Cruz Formation (SCF) of coastal Patagonia (Argentina) is the southernmost (51˚) site spanning the Middle Miocene Climatic Optimum (MMCO) - the last warmest period in the second half of the Cenozoic, characterized by atmospheric CO2 content comparable to the levels predicted for the next century.
SCF fossiliferous horizons can be traced laterally for tens of kilometers and have yielded one of the Earth's most diverse and well-preserved terrestrial vertebrate assemblages. This unique set of geologic sections allow us to, for the first time, combine detailed records of faunal diversity and ecology, vegetation composition and structure, and local climate, and to test hypotheses on the relationship between habitat structure and species diversity.
While we use phytoliths to reconstruct changes in habitat structure (heterogeneity) through time, our collaborators are studying tooth stable isotope data, faunal community structure, and tooth micro- and meso-wear. By integrating these different proxies we will test climate and ecological hypothesis, including the relationship between habitat structure and diversity.
This article is in preparation with my PhD advisor Caroline Strömberg and colleagues.
SCF fossiliferous horizons can be traced laterally for tens of kilometers and have yielded one of the Earth's most diverse and well-preserved terrestrial vertebrate assemblages. This unique set of geologic sections allow us to, for the first time, combine detailed records of faunal diversity and ecology, vegetation composition and structure, and local climate, and to test hypotheses on the relationship between habitat structure and species diversity.
While we use phytoliths to reconstruct changes in habitat structure (heterogeneity) through time, our collaborators are studying tooth stable isotope data, faunal community structure, and tooth micro- and meso-wear. By integrating these different proxies we will test climate and ecological hypothesis, including the relationship between habitat structure and diversity.
This article is in preparation with my PhD advisor Caroline Strömberg and colleagues.
Using leaf vein density to trace the emergence of angiosperms in the canopy in the fossil record
What is leaf vein density?
Vein density (Dv) is the total vein length per leaf area. Increasing Dv reduces the overall hydraulic resistance to water flow through the plant (in the mesophyll) by delivering water closer to its site of evaporation (stomata). Thereby, high Dv values favor high transpiration rates and gas exchange capabilities.
Like other leaf traits associated with plant physiology and ecology, Dv varies vertically within a forest.
Vein density (Dv) is the total vein length per leaf area. Increasing Dv reduces the overall hydraulic resistance to water flow through the plant (in the mesophyll) by delivering water closer to its site of evaporation (stomata). Thereby, high Dv values favor high transpiration rates and gas exchange capabilities.
Like other leaf traits associated with plant physiology and ecology, Dv varies vertically within a forest.
A trait preserved in the fossil record
Although high vein density is a unique characteristic of flowering plants, the first angiosperms were characterized by low vein density values. A four fold increase in vein density during angiosperm evolution lead to transpiration rates comparable to the modern ones at least since the end of the Paleocene. Our study suggests that Angiosperms dominate the forest canopy since at least the Paleocene, with high Dv values corresponding to canopy leaves and low density values corresponding to understorey leaves, and matching the Dv distribution in modern stratified forests (figure 2).
Although high vein density is a unique characteristic of flowering plants, the first angiosperms were characterized by low vein density values. A four fold increase in vein density during angiosperm evolution lead to transpiration rates comparable to the modern ones at least since the end of the Paleocene. Our study suggests that Angiosperms dominate the forest canopy since at least the Paleocene, with high Dv values corresponding to canopy leaves and low density values corresponding to understorey leaves, and matching the Dv distribution in modern stratified forests (figure 2).
Vein density (Dv) probability curves
showing high vein density in the canopy (C) leaves (black) and low vein density in the understory (U) leaves (gray) in two tropical forests in Panama (A,C) and a temperate forest in Maryland (B). Black dotted line in C represents vein density of the forest litter spamming from low, understorey-like Dv values to high, canopy-like Dv values. |
Vein density (Dv) probability curves of eight leaf macrofossil assemblages (gray) from the early Cretaceous to the Paleocene compared to Dv density curve of a tropical forest litter (black) (same as dotted curve in figure above). (Fossil Dv data from Feild et al., 2011).
|
