Hydrological controls of fisheries production in a major Amazonian tributary

Amazonian rivers are characterized by a strongly seasonal flood pulse, which is being altered by ongoing and planned hydropower development across the basin. Changes in hydrology have a cascade of physical, ecological, and social effects, and some ecohydrological changes in the Amazon are expected to be irreversible. To better understand these linkages, we investigated shared trends and causal factors driving fish catch in a major Amazonian tributary before dam construction to derive relationships between catch and natural hydrologic dynamics that can be used to assess likely post‐dam fisheries impacts. We applied a time series‐based dimension reduction technique (dynamic factor analysis) to investigate dynamics in fish catch across 10 commercially important species using daily fish landings and hydrological data. We found a 4‐trend dynamic factor model to best fit the observed data, with fitted trends exhibiting variation representative of seasonal and longer‐term hydrologic variation. We next considered 11 candidate explanatory time series and found the best dynamic factor model used 4 explanatory variables: maximum water level, flooding duration, previous year's flow, and rate of change in flow; however, each species showed a unique response to this set of hydrological variables. Species‐specific responses suggested that future dam operating rules need to closely mimic the natural hydrologic regime in order to maintain the biological dynamics of this and similarly diverse systems. In particular, high flows associated with increased catch the following year are important to maintain, but may be difficult to achieve in the context of reservoir regulation and energy production goals.