The Interplay of the Drivers
Global Change Ecology has highlighted the joint effects of the multiple drivers of community structure and ecosystems functioning [31]. But the complexity of the interplay among the drivers challenges by itself our ability to address it. Another aspect of my research consists in participating facing this challenge.
Trajectories under multiple drivers
The community and ecosystem structures are shaped by external forces, including the different aspects of Global Environmental Changes, from climate and land-use changes to biological invasions. The investigations of community and ecosystem responses to these changes monopolize large part of recent ecological studies, which often focus on single driver. However, I brought back from my overwinter for the French Polar Institute on the sub-Antarctic Kerguelen island first evidence for the need to consider the potential joint effects. The modelling of the larvae cycle of the introduced Calliphora vicina under low temperature showed that the establishment of the species was possible only because of the climate warming recorded since late 70’s [32]. The spread of the invasive blowfly threatens now the native populations of the apterous species Anatalenta aptera and Calycopteryx moseleyi.
In less isolated systems, human land-use becomes a major player. In temperate mountain, climate and land-uses changes are the most probable drivers to jointly affect the ecosystem. Alpine and Arctic systems are expected to be more vulnerable to climatic changes while changes in agro-pastoral practices induced land-use redistribution. As regularly reported in the literature, we found a subalpine grassland rather resistant to simulated heatwave [33]. However, a focus on the carbon storage of this conservative species also showed that the stressed induced by the extreme summer heat and drought could alleviate the sensitivity F. paniculata to mowing jeopardizing the grassland management strategy [34].
The consequences of the interplay between multiple drivers of the ecosystem are hard to predict from single-focused studies [31]. On the other hand, long-term experiments authorize going back in time and investigating the consequences of joint multiple drivers. The transplant experiment monitored during 23 years in the Northern Fennoscandia associated the functional divergence of heath tundra communities toward alternative states to the interplay between environmental perturbation, grazing pressure and soil wetness [14](Fig. 10 ).
The multiple time scale of dynamic
Time is a crucial dimension of ecological processes and my research aims at integrating this aspect. Field experimentation implies questioning delays in the target response or treatment effect. Pulse events induced by manipulative experiment are also the opportunity to explore the diverse aspects of ecological dynamics. My works in communities dominated by long-lived species such as forest and low-productive systems integrate the concept of biological inertia that refers to time interval between the driver action and the ecological response.
Joint effect does not mean necessarily synchronicity. In the slow growing systems of mountain arctic tundra, we showed that neither the community responds to the multiple drivers homogeneously, nor the drivers operate simultaneously. The monitoring of a sowing experiment, in a design crossing grazing, soil wetness and neighbour removal treatments revealed a shift between the prevalence of sowing and neighbour removal during the first years and the increasing importance of grazing and environmental conditions for the sown species establishment [35] (Fig 11 ). Combining multiple drivers and multiple time scale we conclude that even a medium-term study is a poor predictor of the long-term response of community, in particular regarding the asynchrony of species responses [36 ] (Fig.12).
The inertia of the system, the remaining of effect long time after the end of the factor or tolerance of the community can accentuate such asynchrony. We demonstrated, for instance, the current indenture of subalpine grassland biogeochemistry to half century past agro-pastoral practices (Fig. 13).
Alluvial forests are often facing the joint effects of flood regulation and biotic invasions [16,37]. Along the bioclimatic gradient of the Rhone River, the more recent history of flood regulation management and associated biological inertia upstream could explain the increasing biotic containment with condition harshness that contradicted the Stress Gradient Hypothesis [38] (Fig.14).
This temporal inertia is one of the reasons why long-term monitoring programs are so important and must be preciously maintained. Initiatives as GLORIA are often based on simple designs to be robust and reproducible. Then they are also mainly cost-less and their simplicity is often a good argument to seduce a large enough panel of partners to build networks and databases at significant scales. Here, the numerical approach allows to draw the main patterns and strong messages (e.g. the concepts of ‘thermophilisation’ [39] or ‘homogenisation’ [40]). On the other hand, 20 years for a program as GLORIA focused on alpine vegetation is shorter than the life span of most of the monitored species. If the first results could be interpreted as an increase of species richness of the alpine summits with warming [41], such fine monitoring allows also to detect the shrinking of cryophyte species [42] and to suggest the extinction debt hypothesis [43]. Along with the development of ‘omic’ and numeric approaches, the support and maintain of such filed-based long-term monitoring is crucial to root data science approaches in the reality of the ecosystems. It is, here, one of the main motivations for my involvement in GLORIA. The complexity of threshold dynamics from one stable state to an alternative one cannot be captured by other methods but the empirical approaches.