Environmental controls on the development of early life stages of European smelt in the Elbe estuary

• First Individual-based model for European smelt in a German estuary. • Model simulating egg to larvae stage development under changing conditions. • Sensitivity testing of in- and extrinsic factors on growth via random sampling technique. • Model validated using field data and catch monitoring across multiple years. The anadromous European smelt ( Osmerus eperlanus ) is a commonly observed fish in the German estuaries of Elbe, Ems and Weser. Their substantial contribution to local biomass and its ecological role as a “wasp-waist” species, show their importance as a key species in estuarine ecosystems. However, recent studies indicate a rapid decrease in smelt populations across all German estuaries. Potential extrinsic drivers of the decrease include climate change driven temperature increase, shifts in food availability and increasing anthropogenic pressures across the estuarine environment. To investigate possible drivers of mortality of early life stages we present an individual-based model (IBM) for the European smelt, applied in a 1D setup. The model includes development during endogenous life stages, the transition to first feeding and growth-related bioenergetic calculations of 0+ larvae. The sensitivity of the model to changes of intrinsic coefficients and extrinsic drivers was tested using ensemble simulations. Ensembles input data was randomly sampled using a Latin Hypercube Sampling routine (LHS). The model’s ability to reproduce interannual development variability is tested using a combined input dataset including field measurements and bioecological model output. The model allows us to determine important processes that impact the individual survival throughout its lifetime. We present the model validation along with a dedicated parameter sensitivity study on the individual survival. Using field data, the interannual variability of development rates in the model are compared to available monitoring datasets.