Occupancy dynamics in small‐stream habitats: niches define the responses to floods by two neotropical fishes

Abstract

We modeled the occupancy dynamics of the benthic chum‐chum (Pimelodella cf. gracilis) and the midwater threespot leporinus (Leporinus friderici) to determine how these fish species respond to disturbances caused by stream flooding. Fieldwork was conducted in two neotropical streams in central Brazil. Both species occupied fewer sites during floods than during periods with stable hydrology. The percentage of sites occupied by threespot leporinus increased from 57 to 75 %, and the percentage occupied by chum‐chum increased from 27 to 87 %. Threespot leporinus moved upstream after floods, first recolonizing the sites farthest downstream and then recolonizing those farthest upstream. In contrast, chum‐chum remained in several scattered sites along the stream during floods and recolonized nearby sites from these sources (i.e., colonization was not spatially concentrated). During wet hydroperiods, chum‐chum used sites with water velocities ranging from 0.35 to 0.75 m/s. During the dry period, this range increased to 0.20–0.90 m/s (in sites that were colonized after the floods ended). Our results also showed that both species were more easily detected by snorkeling during the dry hydroperiod and that characteristics such as the substrate type and water velocity had different effects on the detectability of the two species. Overall, our study indicated that the annual persistence of threespot leporinus depends largely on seasonal upstream‐biased movement, whereas the chum‐chum population depends on the seasonal scattered colonization of temporary sites.

Scale‐dependent effects of isolation on seasonal patch colonisation by two Neotropical freshwater fishes

“…Moreover, it has been successfully applied to improve species conservation and management, especially given the increasing rates of landscape fragmentation (Akçakaya, Mills, & Doncaster, 2007). These results confirm previous findings (e.g., Hakamada & Penha, 2014), but in a slightly different way. Here, we analysed the occupancy dynamics of TA B L E 2 Estimates of the β parameters (±95% confidence interval) and their standard errors from best-ranked models of third model set (final) for multi-season dynamic occupancy models of Pimelodella cf.…”

“…to 06:00 p.m.) to determine species occurrences in each survey k (see more details in Hakamada & Penha, 2014). The choice of the stream section length reflected a compromise between (a) an area that could be thoroughly surveyed in five minutes (i.e., high probability of species detection given its presence and precise estimation of demographic parameters) and (b) large enough to encompass the home range of a mid-sized fish living in streams.…”

“…The river bottom habitat was described using its substrate composition, which was characterised by sixty 15 × 15 cm samples distributed along six perpendicular transects one meter apart from each other (Hakamada & Penha, 2014). The predominant substrate of each sample was classified and ordered from 1 (smooth surfaces: sand, silt, loam, clay and sludge) to 6 (irregular bedrock: upper sides of large cobbles and blocks, bedrock; blocks >40 cm of width) (Parasiewicz, 2007).…”

“…To estimate detectability, we used substrate structure, hydroperiod (raining-t1 plus t2, or drying-t3 plus t4), or season (t1, t2, t3, t4) because these variables were identified as important in a former study (Hakamada & Penha, 2014). We set to two the maximum number of covariates included simultaneously to parametrise γ i,t and ɛ i,t , and implemented our modelling exercise in three phases (detailed in the following paragraph).…”

“…Graceful pimelodella (Pimelodella cf. Both are mid-sized species with seasonal variation in habitat selection that in turn affects their occupancy (i.e., occupancy is higher during the dry than during the rainy season; Hakamada & Penha, 2014). Both are mid-sized species with seasonal variation in habitat selection that in turn affects their occupancy (i.e., occupancy is higher during the dry than during the rainy season; Hakamada & Penha, 2014).…”