Integrating approaches for a revised theoretical framework on metacommunity dynamics using a distance-decay approach.
A decrease of the community similarity between sites as the distance between them increases is usually observed. This decrease can be observed both with geographic and environmental distances, i.e. with an increase in the separation between sites and an increase in the differences between the environment of two sites, respectively. It is called ‘distance-decay’ relationship and it can be modelled with a regression. There are various factors that can modulate the intensity of the decay. Firstly, it has been observed that there is a differential response of the organisms to distance according to their dispersal mode: active or passive. Secondly, the spatial extent of the region studied also affects since the dispersal of the organisms can be limited by the distance. In fact, we have observed that the majority of the studies on this matter have been conducted at large spatial extents or even continental or inter-continental extents, but there are few studies conducted at small spatial extents, where dispersal is supposed to be less limited. Thus, taking into account all this, we have adapted an existing proposal that integrated the type of response that community similarity would have for both geographic and environmental distances according to the dominant metacommunity mechanism (i.e. neutral model, mass effects or species sorting). Additionally, we have differentiated the possible response between regions with small vs. large spatial extents. Furthermore, we used empirical data to try to elucidate which was the prevailing mechanism in two sites (two pond networks of the same habitat type, temporary ponds) with different spatial extents and taking into account organisms with different dispersal modes (active and passive dispersers). We found a clear predominance of the species sorting mechanism in the metacommunities studied independently of the organisms’ dispersal mode in the site with a larger extent. In contrast, in the site with a small extent, dispersal mode seems to be more decisive since each group is likely to be mainly driven by a different mechanism.
Burning the landscape: pond metacommunity resilience against simulated wildfire impacts.
Disturbances such as wildfires are main drivers of biodiversity changes and can determine ecosystems functioning. Surprisingly, iconic systems in terms of its biodiversity value have still received little attention regarding wildfire impacts. This is the case of Mediterranean temporary ponds. These systems are usually found along areas of landscape forming pond networks, whose communities are connected throughout dispersal of individuals (i.e. metacommunities). Empirical approaches to analyse wildfire impacts under a metacommunity perspective are complicated due to the large-scale study needs, both in time and space. Thus, in these situations, modelling approaches can become an interesting tool to better understand the possible impacts on communities and how wildfire can compromise their recovery. In 2012, a large wildfire that burned more than 10000 ha affected a Mediterranean pond network in NE Iberian Peninsula. The wildfire spread fast due to strong winds and left a patchy burned landscape. Our aim is to analyse the resilience of the metacommunity against wildfires using a modelling approach to better understand how this disturbance can compromise the system. The simulation process followed 4 steps: 1) we represented the affected pond network as a graph; 2) simulated the metacommunity with a neutral-based model; 3) generated 400 simulated wildfires having different characteristics: a) burned area, and b) the percentage of ponds burned within this area (i.e. patchiness); and 4) simulated the recolonization process of the burned ponds, again under a neutral-based model dynamic. We then repeated this simulation process but with different network connection distances, to obtain the response of organisms with different landscape perceptions (ranging from 250 to 5000 meters). Our results indicated that the empirical network is strongly connected configuring a highly resilient metacommunity, in some cases in the less connected scenarios. Moreover, simulated wildfires also allowed us to detect that the degree of patchiness was more important than total burned area for the system resilience. In fact, results with a lower degree of patchiness, so the more homogenously affected landscapes, were the ones that showed lower resilience. Finally, in all the simulated wildfires burned ponds rapidly recovered their pre-fire values and only in the most isolated ponds of the network wildfire impact really compromised recovery. Our study is a perfect example, of how simulation and metacommunity models could be useful tools to asses impacts at large scales, providing interesting information to face disturbance consequences.