Network analysis shows the role of exogenous and endogenous processes in shaping spatial synchrony of intertidal macroalgal assemblages

Spatial synchrony, defined as correlated fluctuations of spatially disjunct assemblages, is a fundamental aspect influencing ecosystems dynamics. In fact, the degree of spatial synchronization may strongly affect the stability of a system. Spatial synchrony might be regulated by multiple drivers. For example, exogenous disturbances (i.e., abiotic events leading to biomass removal from a system) may variably shape spatial synchrony in natural assemblages. On the other hand, in some natural systems, endogenous local-scale processes can lead assemblages to self-organize in space and exhibit emergent spatial patterns at large scale. A well-known example of a self-organized system is provided by rocky intertidal mussel beds, where scale-invariant patterns emerge from local disturbance-recovery dynamics. Although exogenous and endogenous processes have been historically studied separately, they may jointly shape spatial synchrony. The aim of the thesis is to investigate the contribution of exogenous and endogenous processes to spatial synchronization in rocky intertidal communities dominated by the brown algae Cystoseira compressa (Esper) Gerloff & Nizamuddin. In 2017, 18 transects with an initially homogenous cover of C. compressa were selected along the coast of Calafuria (Livorno). Transects - made of 6 contiguous quadrats (30 x 30 cm) – were disturbed by removing the organisms from the substratum to simulate the impact of strong waves. Three treatments were applied to simulate different spatial patterns of disturbance: 1) a homogeneous pattern, where all the six quadrats in a transect received the same amount of disturbance (HOM), 2) a gradient-like pattern, where quadrats received a varying intensity of disturbance along the transect (GRAD), and 3) a pattern of disturbance in which the first event was distributed homogenously along a transect, whereas subsequent events were imposed at the margins of previously disturbed areas (EDGE). This treatment tested whether patterns at the scale of the transect could emerge from disturbance-recovery dynamics operating at smaller scales (i.e., at the edges of previous disturbed areas), which is a hallmark of self-organization, as also observed in mussel beds. To determine whether the effects of spatial patterns of disturbance varied in relation to the intensity of disturbance, each of the three treatments was crossed with two levels of intensity of disturbance in a factorial design. Three replicate transects were assigned to each combination of spatial pattern and intensity of disturbance. Disturbances were imposed seven times over the course of the experiment, which lasted four years. Transects were sampled four months after each disturbance treatment. Data from this experiment were analysed with network-based approaches. A network is an object made of elements (nodes) connected by links. The degree of spatial synchronization among the algae and invertebrates associated with C. compressa was determined for each pair of quadrats in the experimental transects, and each quadrat was considered a node. Therefore, for each treatment, a network was assembled by linking highly synchronized quadrats, and the properties of the resulting networks were examined. One of the main focuses of the analysis was the partition of networks into “communities”, groups of nodes tightly interconnected with each other, but less connected with nodes of other groups: in our case, groups of highly synchronised quadrats. Three alternative models for the emergence of spatial synchrony in our study system led to as many hypotheses. If exogenous disturbance was the dominant driver of spatial synchrony, then HOM and EDGE treatments would produce similar networks, both characterized by two communities (correspondent to the two intensity levels). In fact, if endogenous mechanisms do not influence spatial synchrony, whether disturbances were imposed homogeneously along the transect or at the edges of previously homogenously disturbed areas should not make any difference. The GRAD treatment, instead, would have more communities, reflecting differences in the overall intensity of disturbance received by the different quadrats along a transect (Hypothesis 1). In contrast, if endogenous dynamics were prevalent, networks produced by HOM and GRAD treatments should be similar among each other and both should be different from the EDGE treatment, as this was the only treatment mimicking endogenous forces (Hypothesis 2). Finally, if exogenous and endogenous processes were not mutually exclusive forces shaping synchrony, then different results from those expected under hypotheses 1 and 2 would emerge, with all networks possibly becoming different from each other (Hypothesis 3). Our analysis revealed similarities between the networks generated by treatments EDGE and HOM, while both differed from GRAD. These results supported the first hypothesis: exogenous processes dominate the synchronization of temporal fluctuations of algae and invertebrates on rocky shores. Yet, a small, desynchronizing effect of small-scale endogenous processes was detectable. A subsequent analysis has investigated the relationship between synchrony and stability, by simulating the consequences of habitat loss on the different networks. Understanding the processes underpinning synchrony and stability is key to foster ecosystem conservation in the face of global change and escalating anthropogenic disturbance