Wendy C. Gentleman

A chronology of plankton dynamics in silico: how computer models have been used to study marine ecosystems

Research on plankton ecology in the oceans has traditionally been conducted via two scientific approaches: in situ (in the field) and in vitro (in the laboratory). There is, however, a third approach: exploring plankton dynamics in silico, or using computer models as tools to study marine ecosystems. Models have been used for this purpose for over 60 years, and the innovations and implementations of historical studies provide a context for how future model applications can continue to advance our understanding. To that end, this paper presents a chronology of the in silico approach to plankton dynamics, beginning with modeling pioneers who worked in the days before computers. During the first 30 years of automated computation, plankton modeling focused on formulations for biological processes and investigations of community structure. The changing technological context and conceptual paradigms of the late-1970s and 1980s resulted in simulations becoming more widespread research tools for biological oceanographers. This period saw rising use of models as hypothesis-testing tools, and means of exploring the effects of circulation on spatial distributions of organisms. Continued computer advances and increased availability of data in the 1990s allowed old approaches to be applied to old and new problems, and led to developments of new approaches. Much of the modeling in the new millennium so far has incorporated these sophistications, and many cutting-edge applications have come from a new generation of plankton scientists who were trained by modeling gurus of previous eras. The future directions for modeling plankton dynamics are rooted in the historical studies.

Evaluating the synopticity of the US GLOBEC Georges Bank broad-scale sampling pattern with observational system simulation experiments ☆

Abstract A set of observational system simulation experiments (OSSEs) have been designed to assess quantitatively the synopticity of the broad-scale surveys of Georges Bank carried out as part of the US GLOBEC program. The approach uses model simulations that contain realistic spatial and temporal fluctuations of adult females of the planktonic copepod Calanus finmarchicus observed during February and March 1995. Simulations are constructed with two types of assimilation procedures (nudging and the adjoint method), which are used to dynamically interpolate between the two broad-scale surveys taken one month apart. Using these simulations as representations of the real ocean, the model fields are subsampled in space and time along a typical cruise track. These simulated data are then objectively analyzed and the resulting maps compared with “reality” as represented in the original simulation. Results indicate a total error of approximately 50%, which is comprised mostly of simple mapping error (incomplete spatial sampling) and a smaller contribution from space/time smearing. Adjustment of the station positions for displacement by the mean flow reduces the latter error by about half.

The legacy of Gordon Arthur Riley (1911–1985) and the development of mathematical models in biological oceanography

Gordon Arthur Riley (1911-1985) is remembered for his pioneering work in the development of marine ecosystem models during the mid-20th century. Using models that were necessarily simple because of the limited understanding of plankton physiology at the time, as well as the fact that calculations had to be done by hand, Riley studied the processes that control plankton stocks, production, and nutrient cycling, notably at Georges Bank. His great achievement lay not so much in the simulation of plankton dynamics per se, but rather in bringing to the fore the concept of using modeling as a means of explaining and interpreting the dynamics of marine ecosystems. In this article, we examine Rileys approach and philosophy to ecosystem modeling, which we discuss in context of modern day approaches. In particular, we focus on his landmark paper describing a model study of the dynamics of phytoplankton production on Georges Bank (Riley, 1946: J. Mar. Res., 6, 54-73). After reconstructing the model, we show how Riley created new mathematical characterizations of the environmental dependencies of each process in the phytoplankton equation, and how these relate to modern day formulations. We then reproduce Rileys results and conduct further analyses and sensitivity tests which serve to illustrate Rileys conviction that mathematical models can provide clear, rational explanations for the observed temporal changes in ecosystems. Rileys methods and outlook are discussed in context of the ongoing debate about the merits of complex versus simple marine ecosystem models. Based on our analyses of Rileys model, as well as his own critiques, we argue that although recent decades have seen a proliferation of complex ecosystem models that are intended to reflect our expanded understanding, the doctrines proposed by Riley are no less relevant today. In particular, Riley noted that while increasing model complexity is generally desirable, it can only be done within the confines afforded by observational data and knowledge of the physiology and ecology of key species and their interactions.

Modelling rates of random search over the transition from diffusive to ballistic movement of plankton

The rate of search for food (i.e. maximum clearance rate), F, of a plankter is essential to the prediction of encounter rates, and is dependent on movement. Classic encounter rate models assume diffusive or ballistic movements, which represent opposing extremes of directional persistence. From the perspective of the predator, the directional persistence of prey is determined by the ratio of the persistence length (i.e. “run length” of a random walker), λ, and the radius of prey detection, r. We developed an individual-based model to (i) describe variation in F due to λ/r and time, and (ii) evaluate the utility of published corrections (that take into account the effect of λ/r on F ) to the classic models. Our results illustrate that classic models overestimate F when their assumptions of movement are invalid, and indicate that the effect of time variation in F on food consumption is most substantial near the middle of the diffusive to ballistic transition (i.e. λ/r ≈ 1). At λ/r ≪ 1, predators may exploit high clearance rates by “jumping”, provided that the far-field concentration of prey is sufficiently high. We recommend a published Michaelis–Menten type correction to the classic models, and discuss the assumptions and applications of our model system.

Considering non-predatory death in the estimation of copepod early life stage mortality and survivorship

Estimation of early stage mortality is essential for predicting copepod population dynamics and ecological linkages. Standard methods do not distinguish among types of mortality, nor do they consider that samples may include dead individuals, which can lead to misinterpretation and bias. Here, we develop theory to explain how non-predatory death, or “expiration”, influences in situ abundances. We present an amended population dynamics model that accounts for the production of non-viable eggs, expiration of live individuals and losses of dead individuals. This model is used to derive generalizations of four vertical mortality estimation methods, including the widely used Vertical Life Table approach. These new formulae are applied to data for Calanus finmarchicus in the Labrador Sea to illustrate the potential effects of reduced viability on estimated early stage loss rates and survivorship. Results show that even slight reductions in viability can impart significant changes, with the nature of the effect varying among methods, consistent with previous studies. We explain the reasons for these differences and how the common practice of aggregating stages masks the ecological significance of egg viability. Our analysis reinforces previous recommendations for scientists to consider expiration in their estimates of mortality, and in designing their empirical studies.

Explanatory Power of Human and Environmental Pressures on the Fish Community of the Grand Bank before and after the Biomass Collapse

Ecosystem based fisheries management will benefit from assessment of how various pressures affect the fish community, including delayed responses. The objective of this study was to identify which pressures are most directly related to changes in the fish community of the Grand Bank, Northwest Atlantic. These changes are characterized by a collapse and partial recovery of fish biomass and shifting trophic structure over the past three decades. All possible subsets of nine fishing and environmental pressure indicators were evaluated as predictors of the fish community structure (represented by the biomasses of six fish functional-feeding groups), for periods Before (1985 – 1995) and After (1996 – 2013) the collapse, and the Full time series. We modelled these relationships using redundancy analysis, an extension of multiple linear regression that simultaneously evaluates the effect of one or more predictors on several response variables. The analysis was repeated with different lengths (0 to 5 years) and types (moving average vs. lags) of time delays imposed on the predictors. Both fishing and environmental indicators were included in the best models for all types and length of time delays, reinforcing that there is no single type of pressure impacting the fish community in this region. Results show notable differences in the most influential pressures Before and After the collapse, which reflects the changes in harvester behavior in response to the groundfish moratoria in the mid-1990s. The best models for Before the collapse had strikingly high explanatory power when compared to the other periods, which we speculate is because of changes in the relationships among and within the pressures and responses. Moving average predictor sets generally had higher explanatory power than lagged sets, implying that trends in pressures are important for predicting changes in the fish community. Assigning a carefully chosen delay to each predictor further improved the explanatory power, which is indicative of the complexity of interactions between pressures and responses. Here we add to the current understanding of this ecosystem while demonstrating a method for selecting pressures that could be useful to scientists and managers in other ecosystems.