Optimal foraging: Some simple stochastic models

Summary Some simple stochastic models of optimal foraging are considered. Firstly, mathematical renewal theory is used to make a general model of the combined processes of search, encounter, capture and handling. In the case where patches or prey items are encountered according to a Poisson process the limiting probability distribution of energy gain is found. This distribution is found to be normal and its mean and variance are specified. This result supports the use of Holling's disc equation to specify the rate of energy intake in foraging models. Secondly, a model based on minimization of the probability of death due to an energetic shortfall is presented. The model gives a graphical solution to the problem of optimal choices when mean and variance are related. Thirdly, a worked example using these results is presented. This example suggests that there may be natural relationships between mean and variance which make solutions to the problems of energy maximization and minimization of the probability of starvation similar. Finally, current trends in stochastic modeling of foraging behavior are critically discussed.     

Optimal harvesting of ecologically interdependent fish species

The optimal exploitation of a two-species predator-prey system is considered, using Lotka-Volterra-type equations. Due to the density-dependence of ecological efficiency, both species should be harvested simultaneously over a range of relative prices. Beyond the limits of this price range, either the prey species should be utilized indirectly by harvesting the predator, or the predator should be eliminated to maximize the prey yield. Neglecting harvesting costs, the simultaneous harvest of prey and predators requires that a unit of prey biomass increase in value by being “processed” by predators. Certain results from single-species fishery models are shown not to apply to multispecies models. These are as follows: (i) Optimal regulation of a free access fishery may call for subsidizing instead of taxing the harvest of predator species. (ii) Increasing the discount rate may, at “moderate” levels, imply that the optimal standing stock of biomass increases instead of decreasing. (iii) A rising price or a falling cost per unit fishing effort of a species may raise and not lower the optimal standing stock of that species.     

Optimal harvesting of a plant-hervibore system: lichen and reindeer in northern Finland

The optimal harvesting policy for a plant-herbivore system consisting of lichen and reindeer in Finnish Lapland is investigated. Using a discrete time model with no age structure it is shown that the optimal procedure to reach target levels for lichen and reindeer involves a possible initial harvest of reindeer and then a sequence of no-harvesting years until the lichen has recovered. After two adjusting harvests the system will settle to an equilibrium. The optimal solution is compared with the fastest possible approach to the target levels. The two solutions coincide if future yields are sufficiently discounted. With a discount factor near one there will be a heavier initial harvest in the optimal solution. It is seen that allowing some harvesting also in the no-harvesting years has no marked effect on the total yield. The target levels for lichen and reindeer are not unique but depend both on discounting and on the length of the planning period.     

Optimal exploitation of a fishery employing a non-linear harvesting function

A harvesting function is developed to described the rate of removal of fish from a fish population. The function incorporates the effects of both the handling or processing time of the catch and the competition, between boats in the fleet, for the fish.We will assume that the growth rate of the fish population can be modelled with a concave, dome shaped growth curve. With this assumption, it has been shown that if the rate of harvesting the fish is linearly related to both effort (which can be thought of as some measure of the number of boats in the fleet) and the population size, then the population will tend towards a single equilibrium level which is globally stable. This paper shows that the saturation effects due to the handling time may generate two equilibrium levels (one stable, one unstable) rather than a single globally stable equilibrium. The results of competition between boats are economically undesirable because of the decrease in efficiency. However, this competition may be beneficial to the exploited fish population.Using the harvesting model derived earlier, the steady state or long term optimal harvesting policies as well as the transition paths to these states are developed. The only constraint is on the maximum allowable effort which is effectively an upper limitation on the fleet size or number of man-hours of fishing.     

Survival analysis of stochastic single-species population models in polluted environments

This paper reports on the behaviors of single species models with and without pollution. We consider three basic models, one is deterministic, and others are stochastic. We first obtained the acute thresholds between local extinction and (stochastic) weakly persistent in the mean for population respectively. Then we study the attainability of population size 0 for the stochastic cases and show that a randomized non-autonomous logistic equation will be stochastic permanent under some conditions. Finally, we introduce some numerical simulink graphics to illustrate our main results.     

Integral projection models for finite populations in a stochastic environment

Continuous types of population structure occur when continuous variables such as body size or habitat quality affect the vital parameters of individuals. These structures can give rise to complex population dynamics and interact with environmental conditions. Here we present a model for continuously structured populations with finite size, including both demographic and environmental stochasticity in the dynamics. Using recent methods developed for discrete age-structured models we derive the demographic and environmental variance of the population growth as functions of a continuous state variable. These two parameters, together with the expected population growth rate, are used to define a one-dimensional diffusion approximation of the population dynamics. Thus, a substantial reduction in complexity is achieved as the dynamics of the complex structured model can be described by only three population parameters. We provide methods for numerical calculation of the model parameters and demonstrate the accuracy of the diffusion approximation by computer simulation of specific examples. The general modeling framework makes it possible to analyze and predict future dynamics and extinction risk of populations with various types of structure, and to explore consequences of changes in demography caused by, e.g., climate change or different management decisions. Our results are especially relevant for small populations that are often of conservation concern.     

A new synergetic paradigm in environmental numerical modeling: Hybrid models combining deterministic and machine learning components

A new type of environmental numerical models, hybrid environmental numerical models (HEMs) based on combining deterministic modeling and machine learning components, is introduced and formulated. Conceptual and practical possibilities of developing HEM, as an optimal synergetic combination of the traditional deterministic/first principles modeling (like that used for solving PDEs on the sphere representing model dynamics of global climate models) and machine learning components (like accurate and fast neural network emulations of model physics or chemistry processes), are discussed. Examples of developed HEMs (hybrid climate models and a hybrid wind–wave ocean model) illustrate the feasibility and efficiency of the new approach for modeling extremely complex multidimensional systems.     

Some stochastic differential equation models of an aquatic ecosystem

A set of stochastic differential equations has been used to model an aquatic ecosystem. The randomness in the system has been introduced through initial conditions of the state variables, parameters, and input variables (light and temperature). These models were analysed using Monte Carlo simulation procedures and the results were similar to those observed in the experimental and field data. They were different, however, from the results of a deterministic simulation. This approach allows us to incorporate the maximum degree of information in the model and to study the behavior of the system without arbitrarily manipulating the values of the parameters. Some possible refinements and generalizations of this approach are also discussed.     

Space-time estimation of grid-cell hourly ozone levels for assessment of a deterministic model

We present an approach to estimate hourly grid-cell surface ozone concentrations based on observations from point monitoring sites in space, for comparison with grid-based results from the SARMAP photochemical air-quality model for a region of northern California. Statistical estimation is carried out on a transformed (square root) scale, followed by back-transforming to the original scale of ozone in parts per billion, adjusting for bias and variance. We estimate a spatially-varying diurnal mean structure and a non-separable space-time correlation structure on the transformed scale. Temporal pre-whitening is followed by modelling of a spatially non-stationary, diurnally-varying spatial correlation structure using a spatial deformation approach. Comparisons of SARMAP model results with the estimated grid-cell ozone levels are presented.     

Using stochastic volatility models to analyse weekly ozone averages in Mexico City

In this paper we make use of some stochastic volatility models to analyse the behaviour of a weekly ozone average measurements series. The models considered here have been used previously in problems related to financial time series. Two models are considered and their parameters are estimated using a Bayesian approach based on Markov chain Monte Carlo (MCMC) methods. Both models are applied to the data provided by the monitoring network of the Metropolitan Area of Mexico City. The selection of the best model for that specific data set is performed using the Deviance Information Criterion and the Conditional Predictive Ordinate method.     

Unveiling human-assisted dispersal mechanisms in invasive alien insects: Integration of spatial stochastic simulation and phenology models

Capturing the spread of biological invasions in heterogeneous landscapes is a complex modelling task where information on both dispersal and population dynamics needs to be integrated. Spatial stochastic simulation and phenology models have rarely been combined to assist in the study of human-assisted long-distance dispersal events.Here we develop a process-based spatially explicit landscape-extent simulation model that considers the spread and detection of invasive insects. Natural and human-assisted dispersal mechanisms are modelled with an individual-based approach using negative exponential and negative power law dispersal kernels and gravity models. The model incorporates a phenology sub-model that uses daily temperature grids for the prediction and timing of the population dynamics in each habitat patch. The model was applied to the study of the invasion by the important maize pest western corn rootworm (WCR) Diabrotica virgifera ssp. virgifera in Europe. We parameterized and validated the model using maximum likelihood and simulation methods from the historical invasion of WCR in Austria.WCR was found to follow stratified dispersal where international transport networks in the Danube basin played a key role in the occurrence of long-distance dispersal events. Detection measures were found to be effective and altitude had a significant effect on limiting the spread of WCR. Spatial stochastic simulation combined with phenology models, maximum likelihood methods and predicted versus observed regression showed a high degree of flexibility that captured the salient features of WCR spread in Austria. This modelling approach is useful because it allows to fully exploit and the often limited and heterogeneous information available regarding the population dynamics and dispersal of alien invasive insects.     

Use of stochastic patch occupancy models in the California red-legged frog for Bayesian inference regarding past events and future persistence

Assessing causes of population decline is critically important to management of threatened species. Stochastic patch occupancy models (SPOMs) are popular tools for examining spatial and temporal dynamics of populations when presence–absence data in multiple habitat patches are available. We developed a Bayesian Markov chain method that extends existing SPOMs by focusing on past environmental changes that may have altered occupancy patterns prior to the beginning of data collection. Using occupancy data from 3 creeks, we applied the method to assess 2 hypothesized causes of population decline—in situ die-off and residual impact of past source population loss—in the California red-legged frog. Despite having no data for the 20–30 years between the hypothetical event leading to population decline and the first data collected, we were able to discriminate among hypotheses, finding evidence that in situ die-off increased in 2 of the creeks. Although the creeks had comparable numbers of occupied segments, owing to different extinction–colonization dynamics, our model predicted an 8-fold difference in persistence probabilities of their populations to 2030. Adding a source population led to a greater predicted persistence probability than did decreasing the in situ die-off, emphasizing that reversing the deleterious impacts of a disturbance may not be the most efficient management strategy. We expect our method will be useful for studying dynamics and evaluating management strategies of many species.     

Reducing over-reporting of deterministic co-occurrence patterns in biotic communities

Null models of species co-occurrence are widely used to infer the existence of various ecological processes. Here we investigate the susceptibility of the most commonly used of these models (the C-score in conjunction with the sequential swap algorithm) to type 1 and type 2 errors. To do this we use simulated datasets with a range of numbers of sites, species and coefficients of variation (CV) in species abundance. We find that this model is particularly susceptible to type 1 errors when applied to large matrices and those with low CV in species abundance. As expected, type 2 error rates decrease with increasing numbers of sites and species, although they increase with increasing CV in species abundance. Despite this, power remains acceptable over a wide range of parameter combinations. The susceptibility of this analytical method to type 1 errors indicates that many previous studies may have incorrectly reported the existence of deterministic patterns of species co-occurrence. We demonstrate that in order to overcome the problem of high type 1 error rates, the number of swaps used to generate null distributions for smaller matrices needs to be increased to over 50,000 swaps (well beyond the 5000 commonly used in published analyses and the 30,000 suggested by Lehsten and Harmand, 2006). We also show that this approach reduces type 1 error rates in real datasets. However, even using this solution, larger datasets still suffer from high type 1 error rates. Such datasets therefore require the use of very large numbers of swaps, which calls for improvements in the most commonly used software. In general, users of this powerful analytical method must be aware that they need surprisingly large numbers of swaps to obtain unbiased estimates of structuring in biotic communities.     

Predicting tree species presence and basal area in Utah: A comparison of stochastic gradient boosting,generalized additive models,and tree-based methods

Many efforts are underway to produce broad-scale forest attribute maps by modelling forest class and structure variables collected in forest inventories as functions of satellite-based and biophysical information. Typically, variants of classification and regression trees implemented in Rulequest's^© See5 and Cubist (for binary and continuous responses, respectively) are the tools of choice in many of these applications. These tools are widely used in large remote sensing applications, but are not easily interpretable, do not have ties with survey estimation methods, and use proprietary unpublished algorithms. Consequently, three alternative modelling techniques were compared for mapping presence and basal area of 13 species located in the mountain ranges of Utah, USA. The modelling techniques compared included the widely used See5/Cubist, generalized additive models (GAMs), and stochastic gradient boosting (SGB). Model performance was evaluated using independent test data sets. Evaluation criteria for mapping species presence included specificity, sensitivity, Kappa, and area under the curve (AUC). Evaluation criteria for the continuous basal area variables included correlation and relative mean squared error. For predicting species presence (setting thresholds to maximize Kappa), SGB had higher values for the majority of the species for specificity and Kappa, while GAMs had higher values for the majority of the species for sensitivity. In evaluating resultant AUC values, GAM and/or SGB models had significantly better results than the See5 models where significant differences could be detected between models. For nine out of 13 species, basal area prediction results for all modelling techniques were poor (correlations less than 0.5 and relative mean squared errors greater than 0.8), but SGB provided the most stable predictions in these instances. SGB and Cubist performed equally well for modelling basal area for three species with moderate prediction success, while all three modelling tools produced comparably good predictions (correlation of 0.68 and relative mean squared error of 0.56) for one species.     

Seasonal density dependence, timing of mortality, and sustainable harvesting

Birth-pulse populations are often characterized with discrete-time models, that use a single function to relate the post-breeding population size to the post-breeding size of the previous year. Recently, models of seasonal density dependence have been constructed that emphasize interactions during shorter time periods also. Here, we study two very simple forms of density-dependent mortality, that lead to Ricker and Beverton-Holt type population dynamics when viewed over the whole year. We explore the consequences of harvest timing to equilibrium population sizes under such density dependence. Whether or not individual mortality compensates for the harvested quota, the timing of harvesting has a strong impact on the sustainability of a harvesting quota. Further, we show that careless discretization of a continuous mortality scheme may seriously underestimate the reduction in population size caused by hunting and overestimate the sustainable yield. We also introduce the concept of the demographic value of an individual, which reflects the expected contribution to population size over time in the presence of density dependence. Finally, we discuss the possibility of calculating demographic values as means of optimizing harvest strategies. Here, a Pareto optimal harvest strategy will minimize the loss of demographic value from the population for a given yield.     

Forest production model for upland black spruce stands—Optimal site occupancy levels for maximizing net production

The objective of this study was to develop a forest production model for determining optimal density management regimes for upland black spruce (Picea mariana (Mill.) B.S.P.) stands based on the maximization of net production. This objective was attained via the development of an allometrically extended stand density management diagram (SDMD), which was used to describe the mass dynamics of biotic and abiotic tree components by initial density regime, site quality and fine root turnover rate. Specifically, periderm, stem, branch, foliage and abiotic crown masses were estimated employing multivariate allometric regression functions based on data derived from 125 destructively sampled trees. Below-ground mass estimates were obtained using generalized allometric relationships derived from the literature. Abiotic masses included three basic components: (1) allometrically estimated retained woody debris consisting of abiotic crown structures that remained attached to the main stem; (2) fine woody debris arising from needle loss, root turnover, and abscission of modular components; (3) coarse woody debris arising from trees which incurred mortality through self-thinning. The algorithmic version of the model (1) simultaneously calculates periodic annual net production estimates (Mg/ha/year) by 10-year intervals over 100-year rotation lengths for eight initial density conditions, (2) given (1), determines the occupancy level for which net production is maximized for each stage of development (decade interval), and (3) given (2), determines the optimal size–density trajectory within the context of a SDMD. Additionally, results derived from multiple model simulations employing a range of initial densities (1500, 1650,…, 16,350 stems/ha), site indices (9, 10,…, 15 m) and fine root turnover rates (0.2, 0.3,…,0.8 proportion/year), indicated that black spruce productivity was maximized when site occupancies were maintained slightly below the zone of imminent competition mortality. Instructions for acquiring an executable version of the model through the Internet are also included.     

Selective harvesting and habitat loss produce long-term life history changes in a mouflon population.

We examined the long-term effects (28 years) of habitat loss and phenotype-based selective harvest on body mass, horn size, and horn shape of mouflon (Ovis gmelini musimon) in southern France. This population has experienced habitat deterioration (loss of 50.8% of open area) since its introduction in 1956 and unrestricted selective hunting of the largest horned males since 1973. Both processes are predicted to lead to a decrease in phenotype quality by decreasing habitat quality and by reducing the reproductive contribution of individuals carrying traits that are targeted by hunters. Body mass and body size of both sexes and horn measurements of males markedly decreased (by 3.4-38.3%) in all age classes from the 1970s. Lamb body mass varied in relation to the spatiotemporal variation of habitat closure within the hunting-free reserve, suggesting that habitat closure explains part of these changes. However, the fact that there was no significant spatial variation in body mass in the early part of the study, when a decline in phenotypic quality already had occurred, provided support for the influence of selective harvesting. We also found that the allometric relationship between horn breadth and horn length changed over the study period. For a given horn length, horn breadth was lower during the second part of the study. This result, as well as changes in horn curve diameter, supports the interpretation that selective harvesting of males based on their horn configuration had evolutionary consequences for horn shape, since this phenotypic trait is less likely to be affected by changes in habitat characteristics. Moreover, males required more time (approximately four years) to develop a desirable trophy, suggesting that trophy hunting favors the reproductive contribution of animals with slow-growing horns. Managers should exploit hunters' desire for trophy males to finance management strategies which ensure a balance between the population and its environment. However, for long-term sustainable exploitation, harvest strategy should also ensure that selectively targeted males are allowed to contribute genetically to the next generations.     

On equilibrium in resource markets with scale economies and stochastic prices

In this paper, I show the existence and the characteristics of equilibrium in a non-renewable resource market where extraction costs are non-convex and market price is subject to stochastic shocks, an empirically relevant setting. In my model firms may be motivated to hold inventories to facilitate production smoothing, which allows them to continue producing at a smooth pace at any instant when extraction ceases, e.g. when reserves are exhausted. This aspect of the model then supports a competitive equilibrium in the presence of non-convex costs. Casual empirical evidence is provided that supports the central features of my model for a variety of non-renewable resources, lending credence to the explanation for equilibrium I propose.