Mentoring Undergraduate Research in Mathematical Modeling.

In writing about undergraduate research in mathematical modeling, I draw on my 31 years as a mathematics professor at the University of Nebraska-Lincoln, where I mentored students in honors' theses, REU groups, and research done in a classroom setting, as well as my prior experience. I share my views on the differences between research at the undergraduate and professional levels, offer advice for undergraduate mentoring, provide suggestions for a variety of ways that students can disseminate their research, offer some thoughts on mathematical modeling and how to explain it to undergraduates, and discuss the challenges involved in broadening research participation to include early career students and mid-tier students and how to deliver a research experience in a classroom setting. While different situations pose different challenges, different problems require different approaches, and different experiences lead to different conclusions, it is my hope that my experiences will be of broad value to a wide audience.

Water transport through tall trees: A vertically explicit, analytical model of xylem hydraulic conductance in stems.

Trees grow by vertically extending their stems, so accurate stem hydraulic models are fundamental to understanding the hydraulic challenges faced by tall trees. Using a literature survey, we showed that many tree species exhibit continuous vertical variation in hydraulic traits. To examine the effects of this variation on hydraulic function, we developed a spatially explicit, analytical water transport model for stems. Our model allows Huber ratio, stem-saturated conductivity, pressure at 50% loss of conductivity, leaf area, and transpiration rate to vary continuously along the hydraulic path. Predictions from our model differ from a matric flux potential model parameterized with uniform traits. Analyses show that cavitation is a whole-stem emergent property resulting from non-linear pressure-conductivity feedbacks that, with gravity, cause impaired water transport to accumulate along the path. Because of the compounding effects of vertical trait variation on hydraulic function, growing proportionally more sapwood and building tapered xylem with height, as well as reducing xylem vulnerability only at branch tips while maintaining transport capacity at the stem base, can compensate for these effects. We therefore conclude that the adaptive significance of vertical variation in stem hydraulic traits is to allow trees to grow tall and tolerate operating near their hydraulic limits.

Scaling for Dynamical Systems in Biology.

Asymptotic methods can greatly simplify the analysis of all but the simplest mathematical models and should therefore be commonplace in such biological areas as ecology and epidemiology. One essential difficulty that limits their use is that they can only be applied to a suitably scaled dimensionless version of the original dimensional model. Many books discuss nondimensionalization, but with little attention given to the problem of choosing the right scales and dimensionless parameters. In this paper, we illustrate the value of using asymptotics on a properly scaled dimensionless model, develop a set of guidelines that can be used to make good scaling choices, and offer advice for teaching these topics in differential equations or mathematical biology courses.

Incorporating mass vaccination into compartment models for infectious diseases.

The standard way of incorporating mass vaccination into a compartment model for an infectious disease is as a spontaneous transition process that applies to the entire susceptible class. The large degree of COVID-19 vaccine refusal, hesitancy, and ineligibility, and initial limitations of supply and distribution require reconsideration of this standard treatment. In this paper, we address these issues for models on endemic and epidemic time scales. On an endemic time scale, we partition the susceptible class into prevaccinated and unprotected subclasses and show that vaccine refusal/hesitancy/ineligibility has a significant impact on endemic behavior, particularly for diseases where immunity is short-lived. On an epidemic time scale, we develop a supply-limited Holling type 3 vaccination model and show that it is an excellent fit to vaccination data. We then extend the Holling model to a COVID-19 scenario in which the population is divided into two risk classes, with the high-risk class being prioritized for vaccination. In both cases, with and without risk stratification, we see significant differences in epidemiological outcomes between the Holling vaccination model and naive models. Finally, we use the new model to explore implications for public health policies in future pandemics.

Dynamic Energy Budget models: fertile ground for understanding resource allocation in plants in a changing world.

Climate change is having dramatic effects on the diversity and distribution of species. Many of these effects are mediated by how an organism's physiological patterns of resource allocation translate into fitness through effects on growth, survival and reproduction. Empirically, resource allocation is challenging to measure directly and so has often been approached using mathematical models, such as Dynamic Energy Budget (DEB) models. The fact that all plants require a very similar set of exogenous resources, namely light, water and nutrients, integrates well with the DEB framework in which a small number of variables and processes linked through pathways represent an organism's state as it changes through time. Most DEB theory has been developed in reference to animals and microorganisms. However, terrestrial vascular plants differ from these organisms in fundamental ways that make resource allocation, and the trade-offs and feedbacks arising from it, particularly fundamental to their life histories, but also challenging to represent using existing DEB theory. Here, we describe key features of the anatomy, morphology, physiology, biochemistry, and ecology of terrestrial vascular plants that should be considered in the development of a generic DEB model for plants. We then describe possible approaches to doing so using existing DEB theory and point out features that may require significant development for DEB theory to accommodate them. We end by presenting a generic DEB model for plants that accounts for many of these key features and describing gaps that would need to be addressed for DEB theory to predict the responses of plants to climate change. DEB models offer a powerful and generalizable framework for modelling resource allocation in terrestrial vascular plants, and our review contributes a framework for expansion and development of DEB theory to address how plants respond to anthropogenic change.

A discrete/continuous time resource competition model and its implications.

We use a mixed time model to study the dynamics of a system consisting of two consumers that reproduce only in annual birth pulses, possibly at different times, with interaction limited to competition for a resource that reproduces continuously. Ecological theory predicts competitive exclusion; this expectation is met under most circumstances, the winner being the species with the greater 'power', defined as the time average consumer level at the fixed point. Instability of that fixed point for the stronger competitor slightly weakens its domination, so that a resident species with an unstable fixed point can sometimes be invaded by a slightly weaker species, leading ultimately to coexistence. Differences in birth pulse times can lead to qualitatively different long-term coexistence behaviour, including cycles of different lengths or chaos. We also determine conditions under which the timing of an annual pulse of a toxin can change the balance of power.

The interspecific growth-mortality trade-off is not a general framework for tropical forest community structure.

Resource allocation within trees is a zero-sum game. Unavoidable trade-offs dictate that allocation to growth-promoting functions curtails other functions, generating a gradient of investment in growth versus survival along which tree species align, known as the interspecific growth-mortality trade-off. This paradigm is widely accepted but not well established. Using demographic data for 1,111 tree species across ten tropical forests, we tested the generality of the growth-mortality trade-off and evaluated its underlying drivers using two species-specific parameters describing resource allocation strategies: tolerance of resource limitation and responsiveness of allocation to resource access. Globally, a canonical growth-mortality trade-off emerged, but the trade-off was strongly observed only in less disturbance-prone forests, which contained diverse resource allocation strategies. Only half of disturbance-prone forests, which lacked tolerant species, exhibited the trade-off. Supported by a theoretical model, our findings raise questions about whether the growth-mortality trade-off is a universally applicable organizing framework for understanding tropical forest community structure.

Type II functional response for continuous, physiologically structured models.

The goal of this work is to formulate a general Holling-type functional, or behavioral, response for continuous physiologically structured populations, where both the predator and the prey have physiological densities and certain rules apply to their interactions. The physiological variable can be, for example, a development stage, weight, age, or a characteristic length. The model leads to a Fredholm integral equation for the functional response, and, when inserted into population balance laws, it produces a coupled system of partial differential-integral equations for the two species, with a nonlocal integral term that arises from rules of interaction in the functional response. The general model is, typically, analytically intractable, but specialization to a structured prey-unstructured predator model leads to some analytic results that reveal interesting and unexpected dynamics caused by the presence of size-dependent handling times in the functional response. In this case, steady-states are shown to exist over long times, similar to the stable age-structure solutions for the McKendick-von Foerster model with exponential growth rates determined by the Euler-Lotka equation. But, for type II responses, there are early transient oscillations in the number of births that bifurcate in a few generations into either the decaying or growing steady-state. The bifurcation parameter is the initial level of prey. This special case is applied to a problem of the biological control of a structured pest population (e.g., aphids) by a predator (e.g., lady beetles).

Continuous and pulsed epidemiological models for onchocerciasis with implications for eradication strategy.

Onchocerciasis is an endemic disease in parts of sub-Saharan Africa. Complex mathematical models are being used to assess the likely efficacy of efforts to eradicate the disease; however, their predictions have not always been borne out in practice. In this paper, we represent the immunological aspects of the disease with a single empirical parameter in order to reduce the model complexity. Asymptotic approximation allows us to reduce the vector-borne epidemiological model to a model of an infectious disease with nonlinear incidence. We then consider two versions, one with continuous treatment and a more realistic one where treatment occurs only at intervals. Thorough mathematical analysis of these models yields equilibrium solutions for the continuous case, periodic solutions for the pulsed case, and conditions for the existence of endemic disease equilibria in both cases, thereby leading to simple model criteria for eradication. The analytical results and numerical experiments show that the continuous treatment version is an excellent approximation for the pulsed version and that the current onchocerciasis eradication strategy is inadequate for regions where the incidence is highest and unacceptably slow even when the long-term behavior is the disease-free state.

Biological control does not imply paradox.

Is the classical predator-prey theory inherently pathological? Defenders of the theory are losing ground in the debate. We will demonstrate that detractors' main argument is based on a faulty model, and that the conceptual and predictive bases of the theory are fundamentally sound.

Forest defoliation scenarios.

We consider the mathematical model originally created by Ludwig, Jones, and Holling to model the infestation of spruce forests in New Brunswick by the spruce budworm. With biologically plausible parameter values, the dimensionless version of the model contains small parameters derived from the time scales of the state variables and smaller parameters derived from the relative importance of different population change mechanisms. The small time-scale parameters introduce a singular perturbation structure to solutions, with one variable changing on a slow time scale and two changing on a fast time scale. The smaller process-scale parameters allow for the existence of equilibria at vastly different orders of magnitude. These changes in scale of the state variables result in fast dynamics not associated with the time scales. For any given set of parameters, the observed dynamics is a mixture of time scale effects with process-scale effects. We identify and analyze the different scenarios that can occur and indicate the relevant regions in the parameter space corresponding to each.

Dynamic energy budget models with size-dependent hazard rates.

We formulate and analyze two dynamic energy budget models, a net assimilation model with constant allocation strategy and a net production model with a 2-stage allocation strategy, with the objective of determining strategies that maximize the expected lifetime reproductive energy. The per capita death rate depends on the organism's size, as for example when the main cause of death is predation. In the analysis of the net production model, the size at maturity is calculated along with the probability of reaching that size. We show that a small probability of survival to maturity is incompatible with the simple assumption of an exponential survival probability. We demonstrate that when the hazard rate is significantly greater for small individuals than large ones, it is possible for the optimum strategy to be for an individual to grow to a large size in spite of an arbitrarily small probability of survival to maturity. Numerical simulations indicate how the optimal allocation strategies depend on the parameter values.