Limits...
A Novel, Unbiased Analysis Approach for Investigating Population Dynamics: A Case Study on Calanus finmarchicus and Its Decline in the North Sea.

Papworth DJ, Marini S, Conversi A - PLoS ONE (2016)

Bottom Line: This methodology benefits of having no a priori assumptions either on the ecological parameters used or on the underlying mathematical relationships among them.The results produced by the GP based procedure are consistent with the literature published to date, and validate the use of GP for interpreting species dynamics.We propose that this methodology holds promises for the highly non-linear field of ecology.

View Article: PubMed Central - PubMed

Affiliation: Faculty of Science and Technology, School of Marine Science and Engineering, Plymouth University, Plymouth, Devon, PL4 8AA, United Kingdom.

ABSTRACT
Marine populations are controlled by a series of drivers, pertaining to both the physical environment and the biological environment (trophic predator-prey interactions). There is heated debate over drivers, especially when trying to understand the causes of major ecosystem events termed regime shifts. In this work, we have researched and developed a novel methodology based on Genetic Programming (GP) for distinguishing which drivers can influence species abundance. This methodology benefits of having no a priori assumptions either on the ecological parameters used or on the underlying mathematical relationships among them. We have validated this methodology applying it to the North Sea pelagic ecosystem. We use the target species Calanus finmarchicus, a key copepod in temperate and subarctic ecosystems, along with 86 biological, hydrographical and climatic time series, ranging from local water nutrients and fish predation, to large scale climate pressure patterns. The chosen study area is the central North Sea, from 1972 to 2011, during which period there was an ecological regime shift. The GP based analysis identified 3 likely drivers of C. finmarchicus abundance, which highlights the importance of considering both physical and trophic drivers: temperature, North Sea circulation (net flow into the North Atlantic), and predation (herring). No large scale climate patterns were selected, suggesting that when there is availability of both data types, local drivers are more important. The results produced by the GP based procedure are consistent with the literature published to date, and validate the use of GP for interpreting species dynamics. We propose that this methodology holds promises for the highly non-linear field of ecology.

No MeSH data available.


Time series of Calanus finmarchicus and the 9 relevant variables identified by the GP procedure.The time series are ordered from left to right of most frequently occurring. All time series were normalised by dividing each value by the time series maximum value before the GP process.
© Copyright Policy
Related In: Results  -  Collection

License
getmorefigures.php?uid=PMC4930201&req=5

pone.0158230.g003: Time series of Calanus finmarchicus and the 9 relevant variables identified by the GP procedure.The time series are ordered from left to right of most frequently occurring. All time series were normalised by dividing each value by the time series maximum value before the GP process.

Mentions: The relevance analysis of the population pool identified the variables that occurred with a frequency superior to chance among the approximating functions of the population-pool. According to the statistic test based on the Bernoulli trial (see S1 File for details), the variables whose occurrence frequency was larger than 7 were deemed relevant. Out of the original 86 potential drivers, 9 variables, statistically relevant for the approximation of C. finmarchicus, where thus identified. These 9 relevant variables (6, once seasons and multiple datasets are combined, e.g., herring biomass and abundance represent 1 potential driver, herring) are, in order of frequency of occurrence in the final population-pool: herring (total stock biomass and total abundance estimates), cod (spawning stock biomass and abundance at age 1), phytoplankton (as represented by the PCI), SST (spring and winter), and two circulation variables, N. Atlantic net flow in winter (wNAtlNET) and English Channel eastward flow in summer (smEnglChanE) (Table 2, Fig 3). It is worth noticing that a) the relevant variables encompass both top-down/ bottom-up (predators, prey proxies) biological and physical (climate, temperature, circulation,) potential drivers, and b) they do not encompass large-scale climate indices.


A Novel, Unbiased Analysis Approach for Investigating Population Dynamics: A Case Study on Calanus finmarchicus and Its Decline in the North Sea.

Papworth DJ, Marini S, Conversi A - PLoS ONE (2016)

Time series of Calanus finmarchicus and the 9 relevant variables identified by the GP procedure.The time series are ordered from left to right of most frequently occurring. All time series were normalised by dividing each value by the time series maximum value before the GP process.
© Copyright Policy
Related In: Results  -  Collection

License
Show All Figures
getmorefigures.php?uid=PMC4930201&req=5

pone.0158230.g003: Time series of Calanus finmarchicus and the 9 relevant variables identified by the GP procedure.The time series are ordered from left to right of most frequently occurring. All time series were normalised by dividing each value by the time series maximum value before the GP process.
Mentions: The relevance analysis of the population pool identified the variables that occurred with a frequency superior to chance among the approximating functions of the population-pool. According to the statistic test based on the Bernoulli trial (see S1 File for details), the variables whose occurrence frequency was larger than 7 were deemed relevant. Out of the original 86 potential drivers, 9 variables, statistically relevant for the approximation of C. finmarchicus, where thus identified. These 9 relevant variables (6, once seasons and multiple datasets are combined, e.g., herring biomass and abundance represent 1 potential driver, herring) are, in order of frequency of occurrence in the final population-pool: herring (total stock biomass and total abundance estimates), cod (spawning stock biomass and abundance at age 1), phytoplankton (as represented by the PCI), SST (spring and winter), and two circulation variables, N. Atlantic net flow in winter (wNAtlNET) and English Channel eastward flow in summer (smEnglChanE) (Table 2, Fig 3). It is worth noticing that a) the relevant variables encompass both top-down/ bottom-up (predators, prey proxies) biological and physical (climate, temperature, circulation,) potential drivers, and b) they do not encompass large-scale climate indices.

Bottom Line: This methodology benefits of having no a priori assumptions either on the ecological parameters used or on the underlying mathematical relationships among them.The results produced by the GP based procedure are consistent with the literature published to date, and validate the use of GP for interpreting species dynamics.We propose that this methodology holds promises for the highly non-linear field of ecology.

View Article: PubMed Central - PubMed

Affiliation: Faculty of Science and Technology, School of Marine Science and Engineering, Plymouth University, Plymouth, Devon, PL4 8AA, United Kingdom.

ABSTRACT
Marine populations are controlled by a series of drivers, pertaining to both the physical environment and the biological environment (trophic predator-prey interactions). There is heated debate over drivers, especially when trying to understand the causes of major ecosystem events termed regime shifts. In this work, we have researched and developed a novel methodology based on Genetic Programming (GP) for distinguishing which drivers can influence species abundance. This methodology benefits of having no a priori assumptions either on the ecological parameters used or on the underlying mathematical relationships among them. We have validated this methodology applying it to the North Sea pelagic ecosystem. We use the target species Calanus finmarchicus, a key copepod in temperate and subarctic ecosystems, along with 86 biological, hydrographical and climatic time series, ranging from local water nutrients and fish predation, to large scale climate pressure patterns. The chosen study area is the central North Sea, from 1972 to 2011, during which period there was an ecological regime shift. The GP based analysis identified 3 likely drivers of C. finmarchicus abundance, which highlights the importance of considering both physical and trophic drivers: temperature, North Sea circulation (net flow into the North Atlantic), and predation (herring). No large scale climate patterns were selected, suggesting that when there is availability of both data types, local drivers are more important. The results produced by the GP based procedure are consistent with the literature published to date, and validate the use of GP for interpreting species dynamics. We propose that this methodology holds promises for the highly non-linear field of ecology.

No MeSH data available.