INRA-CEFS
CV
Research
Publications
CEFS lab
Ecosystem pics
SOM stands for 'soil organic matter', IN for' inorganic nutrient'.
with a dense understorey (bramble, privet, box holly...).
A perfect habitat for roe deer .
_________________________________________________________
1- Looking for patterns
Understory density
The density of the understory up to 1.2 meters (the upper limit accessible to roe deer) was assessed inside and outside the enclosure, in a homogeneous area (in terms of soil characteristics, history of forest management and tree stand age) with a standard contact-point method.
Foliage sampling
In the vein of Carline et al. (2005), we are currently assessing the impact of deer on nutrient (N and P) availability in the soil. We use foliar nutrient content as a proxy to nutrient availability in soil. We sampled 7 species, covering a wide range of phenotypic traits, in the understory and the tree layer: Arum italicum, Euphorbia amygdaloides, Ruscus aculeatus, Ligustrum vulgare, Rubus fructicosus, Viburnum lantana and Quercus pubescens. Three of these species (A. italicum, E. amygdaloides, Q. pubescens) are never browsed by roe deer, while two (R. fructicosus and L. vulgare) are intensely browsed. We purposely picked browsed and unbrowsed species in order to identify potential confounding factors due to chemical reactions (synthesis of secondary metabolites, rejuvenation of foliar tissues) to browsing that could affect nutrient concentrations in leaves. For each species, 20 random individuals were sampled on each side of the fence, each individual accounting for one replicate.
The density of the understory up to 1.2 meters (the upper limit accessible to roe deer) was assessed inside and outside the enclosure, in a homogeneous area (in terms of soil characteristics, history of forest management and tree stand age) with a standard contact-point method.
Foliage sampling
In the vein of Carline et al. (2005), we are currently assessing the impact of deer on nutrient (N and P) availability in the soil. We use foliar nutrient content as a proxy to nutrient availability in soil. We sampled 7 species, covering a wide range of phenotypic traits, in the understory and the tree layer: Arum italicum, Euphorbia amygdaloides, Ruscus aculeatus, Ligustrum vulgare, Rubus fructicosus, Viburnum lantana and Quercus pubescens. Three of these species (A. italicum, E. amygdaloides, Q. pubescens) are never browsed by roe deer, while two (R. fructicosus and L. vulgare) are intensely browsed. We purposely picked browsed and unbrowsed species in order to identify potential confounding factors due to chemical reactions (synthesis of secondary metabolites, rejuvenation of foliar tissues) to browsing that could affect nutrient concentrations in leaves. For each species, 20 random individuals were sampled on each side of the fence, each individual accounting for one replicate.
Chemical analysis
N and P contents in leaves are measured on ground milled (0.5 mm) oven dried material. Total N is determined with a CN gas analyser. Total P is measured after digestion, with the ceruleomolydic blue method. All the measurements are performed in the AGIR laboratory, in INRA Toulouse.
Some preliminary results
We observed a significant decrease of vegetation density below 1 m at high deer density (i. e., inside the enclosure), suggesting a control of deer over accesssible vegetation (see a graph depicting the decrease of vegetation density).
We observed as well an increase of foliar nutrients and a decrease of the foliar N:P ratio in presence of deer, for all the species sampled, consistently with the results obtained by Carline et al (2005) in a scottish birch regeneration intensively browsed by red deer. More results are still to come (see preliminary results for E. amygdaloides).
N and P contents in leaves are measured on ground milled (0.5 mm) oven dried material. Total N is determined with a CN gas analyser. Total P is measured after digestion, with the ceruleomolydic blue method. All the measurements are performed in the AGIR laboratory, in INRA Toulouse.
Some preliminary results
We observed a significant decrease of vegetation density below 1 m at high deer density (i. e., inside the enclosure), suggesting a control of deer over accesssible vegetation (see a graph depicting the decrease of vegetation density).
We observed as well an increase of foliar nutrients and a decrease of the foliar N:P ratio in presence of deer, for all the species sampled, consistently with the results obtained by Carline et al (2005) in a scottish birch regeneration intensively browsed by red deer. More results are still to come (see preliminary results for E. amygdaloides).
_________________________________________________________
2-Modelling mechanisms
2-Modelling mechanisms
The model
In
order to identify mechanisms
that may lead to the observed patterns, we are currently studying a
soil-plant-herbivore model, in the vein of previous
soil-plant-herbivore models (e. g., De Mazancourt et al. 1999).
However, like the stoichiometric models developped in aquatic
ecosystems (e. g., Daufresne and Loreau 2001), our model takes
into account two nutrients (namely, N and P) instead of one. The main
peculiarity of the N and P cycles are considered, in particular, the
fact that the N cycle is more open that the P cycle (see an outline
of the model).
The model is a set of ordinary differential equations. The variables are the stocks of nitrogen and phosphorus in the different compartments of the system. As a preliminary approach, we are currently studying simplified versions of the model (see equations of a simplified version).
Model analysis: generating functional hypothesis
We analyzed simplified versions of the model, such as the one presented above. Among other analysis, we calculated the equilibrium values for all the variables and we addressed their partial derivative with regard to the herbivory rate h, and the recycling efficiency rc. The models predict an increase of nutrient content in plant biomass due to a decrease of competitive pressure when herbivory increases. Indeed, as browsing reduces biomass density in the understory, the remaining plants get more nutrients than they would get in the absence of deer. The model also predicts that given some assumptions, as herbivory increases and plant biomass decreases, the uptake of soil nutrient eventually decreases, leading to larger stocks of inorganic N and P in the soil. Since inorganic nitrogen is highly soluble, more nitrogen get leached out of the system when deer density is high. As a result, the total stock of nitrogen in the system decreases, which buffers the increase of plant nitrogen due to competitive relief. Hence, the nitrogen to phosphorus ratio in plants decreases at high deer density (see some preliminary results).
The model is a set of ordinary differential equations. The variables are the stocks of nitrogen and phosphorus in the different compartments of the system. As a preliminary approach, we are currently studying simplified versions of the model (see equations of a simplified version).
Model analysis: generating functional hypothesis
We analyzed simplified versions of the model, such as the one presented above. Among other analysis, we calculated the equilibrium values for all the variables and we addressed their partial derivative with regard to the herbivory rate h, and the recycling efficiency rc. The models predict an increase of nutrient content in plant biomass due to a decrease of competitive pressure when herbivory increases. Indeed, as browsing reduces biomass density in the understory, the remaining plants get more nutrients than they would get in the absence of deer. The model also predicts that given some assumptions, as herbivory increases and plant biomass decreases, the uptake of soil nutrient eventually decreases, leading to larger stocks of inorganic N and P in the soil. Since inorganic nitrogen is highly soluble, more nitrogen get leached out of the system when deer density is high. As a result, the total stock of nitrogen in the system decreases, which buffers the increase of plant nitrogen due to competitive relief. Hence, the nitrogen to phosphorus ratio in plants decreases at high deer density (see some preliminary results).
3-Back to the fields: testing functional hypotheses
If the functional hypothesis exposed above turns out to be true, one should observe an increase of nitrogen leaching from the soil when deer density is increased. To test this trend, we are currently collecting gravitation water on each side of the fence. We buried 20 tubes (10 on each side) designed to collect water at a depth of 60 cm. Repeated experiments ar being ran: 20 liters of pure water are poored on a one square meter surface around each tube. After 48 hours, the water accumulated in the tubes is collected, and the concentrations of disolved nitrogen (nitrate and ammonium) are measured. Our first prelimnary results are consistent with the hypothesis: we observe significantly more leaching of both nitrate and ammonium inside (high deer density) than outside (low deer density) the enclosure (see some preliminary results).
