Silver carp larva abundance in response to river flow rate revealed by cross-wavelet modelling

The complexity and non-linearity of increasingly available fisheries data, requires modern analytical approaches and statistical tools that can accommodate the temporal and scale structures of natural systems. Traditional analyses often assume uncorrelated or weakly correlated temporal structure in fisheries data time series. We propose that frequency and scale domain methods, in the form of wavelet transforms, can serve as useful probes in early investigations of fisheries data, finding out the mechanisms underlying the changes in the diversity and abundance of fishes. We used cross-wavelet analysis with phase angle and coherency parameters to study the effect of discharge on the occurrence of silver carp larvae in a large subtropical river. Simulation studies show that the spectral signature given by the cross-wavelet method provided a useful approach for statistically detecting and characterizing temporal dependency in the occurrence of fish larvae. In addition, our simulations provide an objective, quantitative basis for understanding how river flow rate acts as a trigger for the occurrence of larvae. And our results can be used to ecological operation silver carp larvae were found when discharge was greater than 104 m3 s−1. The flood peak persists for some days (at least 2 days) and can induce a peak in larval abundance. Our analyses also show that flood duration and the flood total volume decreased from 2006 to 2013. Moreover the flood occurred earlier each year. This may be why the period of most intense spawning (the day of maximum larvae collected) was earlier each year. We conclude with a discussion about the role wavelet analysis may have in appropriately guiding flexible probabilistic models connecting fisheries with environmental covariates. The wavelet approach is an especially promising tool for analyzing ecological data, such as the relationship between hydrological conditions and fish larva occurrence.