Disturbance and recovery of litter fauna: a contribution to environmental conservation

Disturbances play a great role in ecosystem functioning and, with the increasing anthropogenic activities, they have more and more influence on ecosystems. They have been studied for several decades but recovery, the ecological phenomenon following a disturbance, has seldom been the focus of research. In this thesis, I studied the impact of disturbances on the structure of soil and litter fauna communities and their ensuing recovery in varying environmental conditions, combined with the effect of productivity, life-history traits and community structure. I combined all the results to draw some conclusions on the main factors involved in recovery, how to improve recovery of ecosystems and how to make better predictions on recovery. In the second chapter of this thesis, I studied how the structure of soil and litter fauna communities from two climatically contrasting biomes was affected by a similar disturbance and how these communities recovered. I sampled litter macrofauna in a temperate and a boreal forest and, to be able to determine whether the communities had recovered, I created a “recovery index” that took into account the pre- and post-disturbance conditions of the disturbed and the control communities, taking into account natural variations. I hypothesised that the temperate communities would recover more rapidly due to the warmer temperatures and to higher species richness and abundance. Recovery was as fast in both biomes, which also had similar species richness. Contrary to my assumption, higher pre-disturbance species abundance did not favour the resistance of communities to the disturbance; on the contrary, high-abundance communities suffered a proportionally greater loss than other communities. Analyses based on life-history traits revealed that dispersal capabilities were the most relevant traits for species facing a disturbance and also for re-establishing. An unexpected factor that influenced the outcome of the disturbance was the litter layer, which, thick in the temperate forest and almost inexistent in the boreal one, protected the fauna of this former biome. The aim of my third chapter was to consider the main two theories of species assembly, the niche and neutral theory, in the context of a recovery. I considered these two theories not as mutually exclusive but as if they were at opposite ends of a stochasticity gradient. The neutral theory predicts recovered communities in a similar environment to be dissimilar from one another and the niche theory predicts the opposite, because, in this case, species assembly is driven by deterministic factors inherent to communities and to the environment. I used the same experiments as in the previous chapters and hypothesised that the more constraining environmental conditions of the boreal forest would lead to a species assembly rather driven by deterministic factors, with recovered communities more similar to each other than the temperate ones. These latter ones, from a less constraining environment, would be more dissimilar to each other. I found that the structure of each community before and one year after the disturbance was indeed more similar in the boreal forest. This would mean that, in low-productivity environments, the response of communities being less variable, it could be more easily predictable. In Chapter 4, I studied the structure of communities from a different perspective, using the density – body mass (DBM) relationship to detect changes in the structure of communities after a disturbance. I hypothesised that the slope of the relationship would be less steep if smaller organisms were mainly impacted or that it would be steeper if larger organisms were mainly impacted. By collecting the soil and litter fauna before, just after and again one and two months after a disturbance, I could establish that the DBM relationship reflected the changes of the structure of communities responding to modifications of the environment. In disturbed conditions, the slope of the DBM relationship of a community was less steep, because mainly the small organisms were impacted by the disturbance. I also showed that, at the very early stage of the recovery, the slope was even less steep, because of the large body mass of the first colonisers. This study confirmed the necessity to sample a broad spectrum of body masses and it was the first time that the DBM relationship was shown to be able to reflect changes of the structure of communities. I concluded by suggesting that it could be used for environmental biomonitoring. After the satisfying results of Chapter 4, I decided to test the ability of the DBM relationship to reflect different structures of communities living in environments varying by their productivity and subjected or not to a disturbance. I hypothesised that communities from low-productivity areas would have a less steep slope than high-productivity area communities and that disturbed communities would also have a less steep slope. To test this, soil and litter fauna were collected from a salt marsh at four elevation levels (hence subjected to varying sea inundation frequencies, from daily to annually), half of which were subjected to cattle grazing (i.e., the disturbance). I assumed that the least inundated sites were more productive and used the quantity of plant litter to confirm this. The only significant result was between the daily and annually inundated ungrazed areas, confirming that communities from high-productivity areas have a steeper DBM relationship slope. High productivity does not seem to equally affect all the trophic levels, certainly due to inefficient transfers of energy from one level to the other. In the synthesis, I suggested that recovery should first be properly defined to establish when a community has reached that stage and I advise to use pre- and post-disturbance states of control communities for that purpose. Besides, several environmental factors have to be taken into account instead of only focusing on one species or one ecosystem service, as I have showed that the species richness and abundance of communities, and the productivity and heterogeneity of the environment can influence the resistance and recovery of ecosystems. I also propose, in a first time, to study species assembly in constraining environments, where stochastic factors are limited, in order to obtain a better mechanistic understanding of the processes involved. As there is yet not such understanding, I suggest that managers in charge of environmental conservation rather use a phenomenological approach to quickly estimate outcomes of recovery.