Fig 1.
(A) Map of the Trentino-Alto Adige region. (B) The Tagliamento river. (C) Detailed map of the study floodplain reach and the nine sampling sites (as in September 1999). Description of sites and their connectivity level (in %) is as follows: MC = main channel (100%); SC = side channel (100%); EC = ephemeral channel (25%); BW = backwater (75%); EP = ephemeral pond (0%); 30 = pond (17%); 24 = pond (8%); 33 = pond (0%); 53 = pond (0%).
Fig 2.
Hydrograph of the study period.
Plot of the hydrograph showing stage height (cm) from April 1999 to March 2000, measured at the S. Pietro gauging station (3 km downstream of the study reach). Vertical bars indicate the time of invertebrate sampling with exact date listed in text boxes above. The red horizontal dashed line indicates the level where 80% of the study floodplain was inundated. Four main flooding events occurred during the sampling period.
Fig 3.
Biodiversity–connectivity relationships.
Relationship between mean (±SE) taxonomic richness (A), mean diversity of order q = 2 (B), mean density (C) and mean CA (Correspondence Analysis) axis 1(D) with lateral hydrologic connectivity across nine waterbodies in the Tagliamento floodplain (NE Italy). Grey area = 0.95 confidence interval of regression.
Fig 4.
Temporal species accumulation.
Relationship (±0.95 CI) between the rate of temporal species accumulation and lateral hydrologic connectivity across nine waterbodies in the Tagliamento floodplain (NE Italy).
Fig 5.
Change in community metrics after floods.
Relationship between hydrologic connectivity and percentage changes in density and richness after each flooding event. Each panel shows changes in either density or richness between consecutive months during which flooding occurred (e.g. April-May). Zero change is indicated by a solid grey horizontal line. No significant correlations were observed. For visualisation purposes, a dashed grey line indicates the direction of the relationship.
Fig 6.
Temporal changes in community composition.
Correspondence Analysis (CA) axis scores for each floodplain waterbody over the hydrologic year. The colour code follows the gradient of lateral hydrologic connectivity (%) indicated in the text boxes. Black and blue lines represent the main and side channel, respectively. The brown line indicates backwater, while the others, mostly isolated, waterbodies are shown in orange. The dashed line indicates the ephemeral pond. The ephemeral channel, inundated for only three months, is not shown. Vertical blue lines indicate the occurrence of bankful floods. Temporal patterns show that the composition of the floodplain waterbodies tended to converge (towards a ‘main channel’ condition) after the floods, but to diverge during months of stable hydrologic conditions, reflecting differential species sorting.
Fig 7.
Inferring community assembly mechanisms.
Temporal patterns in mean pairwise beta-deviation across waterbodies (A), species co-occurrence z-scores (B) and the proportion of variation in composition explained by local environmental parameters (C). Vertical blue lines indicate the occurrence of bankful flood events. The three indicators consistently showed that flooding weakened the signal of deterministic assembly processes. In (A), mean beta-deviation declined after each flood. In (B), months with stable hydrologic conditions were associated with non-random species co-occurrence patterns (above the red line at z-score = 1.96), which shifted to random patterns after flooding. In (C), flooding tended to decrease the match between local abiotic factors and species composition. Abiotic data for the first and last months were not available.