| Summary: | Marine populations are controlled by a series of drivers which act independently or synergistically. These drivers can be in the form of bottom-up processes such as climate cycles and food availability or top-down drivers such as predation. There is much debate over drivers of ecosystems and populations, especially when trying to understand the causes of events termed regime shifts. Calanus finmarchicus is a key copepod species in temperate and polar ecosystems. With the use of a novel methodology, Genetic Programming (GP), a collection of 87 biological, hydrographical and climatic time series are used to identify the most relevant drivers of C. finmarchicus abundance. Regime shift detection methods were also compared with the results being used to further understand the drivers of C. finmarchicus. GP identified 14 variables as relevant whilst 7 variables were identified as highly-relevant. Relevant variables are variables which occur more than by chance, whilst highly-relevant variables are variables which only appear with other relevant variables in the approximating functions created by GP. GP also produced 12 models which can capture the dynamics of the C. finmarchicus population using only a small set of variables. The 7 highly-relevant variables encompassed only 4 potential drivers: SST, circulation, food and predation. No large scale climate patterns were deemed highly-relevant but AMO, SCA and NAO were selected as relevant variables. Cumulative Sums analysis was preferred over STARS as the variability of the results was far less. From the highly-relevant variables and the CuSUM results a new model of the drivers was created and a potential explanation as to why C. finmarchicus abundance has declined in the North Sea over the past four decades. By increasing our understanding of what drives important marine populations we can improve our management of ecosystems and resources.
|
|---|