Andreas Huth

The simulation of the movement of fish schools

Many species of fish schools organize for short or longer periods of time without a leader. We searched for the behaviour patterns of the individual fish, which allow movement of such a school. On the basis of biological facts we constructed a number of behaviour models and tested them with computer simulations against biological reality. Basic assumptions of our simulations are: (1) The motion of a fish is only influenced by the position and orientation of its nearest neighbours. (2) The new velocity and the turning angle of each fish (after a time step) are calculated by probability distributions taking into account random influences. (3) The movement of each model fish is based upon the same behaviour model, i.e. the modelled fish group swims without a leader. The basic behaviour patterns are attraction, repulsion and parallel orientation. Our investigations show that it is very important how a fish mixes the influences of its neighbours. If a fish averages the influences of its neighbours, the model fish group shows the typical characteristics of a real fish school: strong cohesion and high degree of polarization. If a fish only responds to a single neighbour, the model creates a confused fish group.

Statistical inference for stochastic simulation models – theory and application

Statistical models are the traditional choice to test scientific theories when observations, processes or boundary conditions are subject to stochasticity. Many important systems in ecology and biology, however, are difficult to capture with statistical models. Stochastic simulation models offer an alternative, but they were hitherto associated with a major disadvantage: their likelihood functions can usually not be calculated explicitly, and thus it is difficult to couple them to well-established statistical theory such as maximum likelihood and Bayesian statistics. A number of new methods, among them Approximate Bayesian Computing and Pattern-Oriented Modelling, bypass this limitation. These methods share three main principles: aggregation of simulated and observed data via summary statistics, likelihood approximation based on the summary statistics, and efficient sampling. We discuss principles as well as advantages and caveats of these methods, and demonstrate their potential for integrating stochastic simulation models into a unified framework for statistical modelling.

The simulation of fish schools in comparison with experimental data

Abstract Fish do not need a leader or external stimuli for their school organization. The model presented shows that the group movement of a school can be maintained by interactions in which each individual controls its movement in relation to its neighbours. Our three-dimensional simulations reproduce the typical characteristics of real schools, if the behaviour of the single fish is based on four patterns: attraction, repulsion, parallel orientation and averaging the influences of at least four neighbours. The results of our simulations agree with experimental data in many points, as is demonstrated here for the polarization, nearest neighbour distance and internal dynamics.

Improving the forecast for biodiversity under climate change

BACKGROUND As global climate change accelerates, one of the most urgent tasks for the coming decades is to develop accurate predictions about biological responses to guide the effective protection of biodiversity. Predictive models in biology provide a means for scientists to project changes to species and ecosystems in response to disturbances such as climate change. Most current predictive models, however, exclude important biological mechanisms such as demography, dispersal, evolution, and species interactions. These biological mechanisms have been shown to be important in mediating past and present responses to climate change. Thus, current modeling efforts do not provide sufficiently accurate predictions. Despite the many complexities involved, biologists are rapidly developing tools that include the key biological processes needed to improve predictive accuracy. The biggest obstacle to applying these more realistic models is that the data needed to inform them are almost always missing. We suggest ways to fill this growing gap between model sophistication and information to predict and prevent the most damaging aspects of climate change for life on Earth. ADVANCES On the basis of empirical and theoretical evidence, we identify six biological mechanisms that commonly shape responses to climate change yet are too often missing from current predictive models: physiology; demography, life history, and phenology; species interactions; evolutionary potential and population differentiation; dispersal, colonization, and range dynamics; and responses to environmental variation. We prioritize the types of information needed to inform each of these mechanisms and suggest proxies for data that are missing or difficult to collect. We show that even for well-studied species, we often lack critical information that would be necessary to apply more realistic, mechanistic models. Consequently, data limitations likely override the potential gains in accuracy of more realistic models. Given the enormous challenge of collecting this detailed information on millions of species around the world, we highlight practical methods that promote the greatest gains in predictive accuracy. Trait-based approaches leverage sparse data to make more general inferences about unstudied species. Targeting species with high climate sensitivity and disproportionate ecological impact can yield important insights about future ecosystem change. Adaptive modeling schemes provide a means to target the most important data while simultaneously improving predictive accuracy. OUTLOOK Strategic collections of essential biological information will allow us to build generalizable insights that inform our broader ability to anticipate species’ responses to climate change and other human-caused disturbances. By increasing accuracy and making uncertainties explicit, scientists can deliver improved projections for biodiversity under climate change together with characterizations of uncertainty to support more informed decisions by policymakers and land managers. Toward this end, a globally coordinated effort to fill data gaps in advance of the growing climate-fueled biodiversity crisis offers substantial advantages in efficiency, coverage, and accuracy. Biologists can take advantage of the lessons learned from the Intergovernmental Panel on Climate Change’s development, coordination, and integration of climate change projections. Climate and weather projections were greatly improved by incorporating important mechanisms and testing predictions against global weather station data. Biology can do the same. We need to adopt this meteorological approach to predicting biological responses to climate change to enhance our ability to mitigate future changes to global biodiversity and the services it provides to humans. Emerging models are beginning to incorporate six key biological mechanisms that can improve predictions of biological responses to climate change. Models that include biological mechanisms have been used to project (clockwise from top) the evolution of disease-harboring mosquitoes, future environments and land use, physiological responses of invasive species such as cane toads, demographic responses of penguins to future climates, climate-dependent dispersal behavior in butterflies, and mismatched interactions between butterflies and their host plants. Despite these modeling advances, we seldom have the detailed data needed to build these models, necessitating new efforts to collect the relevant data to parameterize more biologically realistic predictive models. New biological models are incorporating the realistic processes underlying biological responses to climate change and other human-caused disturbances. However, these more realistic models require detailed information, which is lacking for most species on Earth. Current monitoring efforts mainly document changes in biodiversity, rather than collecting the mechanistic data needed to predict future changes. We describe and prioritize the biological information needed to inform more realistic projections of species’ responses to climate change. We also highlight how trait-based approaches and adaptive modeling can leverage sparse data to make broader predictions. We outline a global effort to collect the data necessary to better understand, anticipate, and reduce the damaging effects of climate change on biodiversity.

Tree Mortality in Gap Models: Application to Climate Change

Gap models are perhaps the most widely used class of individual-based tree models used in ecology and climate change research. However, most gap model emphasize, in terms of process detail, computer code, and validation effort, tree growth with little attention to the simulation of plant death or mortality. Mortality algorithms have been mostly limited to general relationships because of sparse data on the causal mechanisms of mortality. If gap models are to be used to explore community dynamics under changing climates, the limitations and shortcomings of these mortality algorithms must be identified and the simulation of mortality must be improved. In this paper, we review the treatment of mortality in gap models, evaluate the relationships used to represent mortality in the current generation of gap models, and then assess the prospects for making improvements, especially for applications involving global climate change. Three needs are identified to improve mortality simulations in gap models: (1) process-based empirical analyses are needed to create more climate-sensitive stochastic mortality functions, (2) fundamental research is required to quantify the biophysical relationships between mortality and plant dynamics, and (3) extensive field data are needed to quantify, parameterize, and validate existing and future gap model mortality functions.

Statistical inference for stochastic simulations models : theory and application

Statistical models are the traditional choice to test scientific theories when observations, processes or boundary conditions are subject to stochasticity. Many important systems in ecology and biology, however, are difficult to capture with statistical models. Stochastic simulation models offer an alternative, but they were hitherto associated with a major disadvantage: their likelihood functions can usually not be calculated explicitly, and thus it is difficult to couple them to well-established statistical theory such as maximum likelihood and Bayesian statistics. A number of new methods, among them Approximate Bayesian Computing and Pattern-Oriented Modelling, bypass this limitation. These methods share three main principles: aggregation of simulated and observed data via summary statistics, likelihood approximation based on the summary statistics, and efficient sampling. We discuss principles as well as advantages and caveats of these methods, and demonstrate their potential for integrating stochastic simulation models into a unified framework for statistical modelling.

Long-term impacts of logging in a tropical rain forest — a simulation study

Abstract Simulation models for growth of tropical forest can be a useful tool to investigate the long-term impacts of logging. In this paper the rain forest growth model FORMIX3 is used for such a study. As main processes, the FORMIX3 model includes tree growth, mortality, regeneration and competition. The calculation of tree growth is based on a carbon balance approach. Trees compete for light and space; dying large trees fall down and create gaps in the forest. Different logging scenarios for an initially undisturbed forest stand at Deramakot (Malaysia) were simulated. Two different logging methods (conventional and low impact logging) in combination with different cutting cycles (from 20 to 100 years) were investigated for their long-term impact on the forest. We characterize the impacts with four indicators: total yield, yield per cut, species composition, and canopy opening. Our simulation results indicate that too short logging cycles (

The effects of tree species grouping in tropical rainforest modelling : Simulations with the individual-based model FORMIND

Abstract Due to high biodiversity in tropical rainforests, tree species are aggregated into functional groups for modelling purposes. In this article the influences of two different classifications of tropical tree species into functional groups on the output of a rainforest model are analysed. The F ormind model is documented. F ormind simulates the tree growth of tropical rainforests. The model is individual-based and developed from the F ormix3 model. In the model, trees compete for light and space in plots of 20×20 m in size. A carbon balance is calculated based on the processes of photosynthesis and respiration. Using a tree geometry submodel, typical tree variables (e.g. diameter, height, crown length) are calculated. The mortality process is mainly driven by falling trees and the canopy gaps they create. Trees of the same functional group and diameter class are represented in one cohort. Simulation results for a primary lowland dipterocarp rainforest in Sabah, Malaysia, are discussed. Detailed structural characteristics of the rainforest stands can be analysed, e.g. the simulation results support the hypothesis that rainforests grow in a layer structure. A comparison of results for the aggregation of 436 tree species into five or 22 functional groups respectively, shows that typical results, such as total stem volume or crown closure, achieved with five groups are as accurate as simulations with a huge number of groups. However, some features such as interspecific competition can only be analysed with a huge number (e.g. 22) of functional groups. The model has the potential to analyse extinction processes and spatial structure of gap formations in rainforests. The analysis of logging scenarios could estimate the effects of human impacts in tropical rainforests.

Simulation of the growth of a lowland Dipterocarp rain forest with FORMIX3

Abstract In this paper a new model for simulation of the growth of tropical rain forest is presented (FORMIX3). The model describes growth, mortality, recruitment of trees and competition between trees. The calculation of tree growth is based on a carbon balance. The carbon gain of a tree depends on the photo production of its leaves, respiration and other losses. Trees compete for light and space. Dying large trees fall down and create gaps in the forest. Based on an extensive field data review, a parametrisation was developed for the simulation of lowland Dipterocarp rain forest at Deramakot, Malaysia. A total of 28 variables describing different aspects of forest structure and growth were compared with field data. The model reproduces most parts of the forest dynamics well. A new concept for sensitivity analysis is presented: 46 model parameter were varied and analysed in respect to their influence on 26 variables describing the forest state. The influence of the different processes on forest structure is complex. Some general trends can be observed: The growth characteristics of the two Dipterocarp species groups strongly influences species composition in the forest, but not general forest structure (biomass, leaf area, production, leaf area, tree size distribution)

Concepts for the aggregation of tropical tree species into functional types and the application to Sabah's lowland rain forests.

For analysing field data as well as for modelling purposes it is useful to classify tree species into a few functional types. In this paper a new aggregation of tree species of the dipterocarp rain forests in Sabah (Borneo), Malaysia, is developed. The aggregation is based on the two criteria successional status and potential maximum height. Three classes of successional status (early, mid and late successional species) and five classes of potential maximum heights ( 36 m) lead to a combination of 15 functional types. The criteria of the developed classification are chosen to suit for applications with process-based models, such as FORMIx3 and FORMIND, which are based on photo- synthesis production as the main process determining tree growth. The concept is universal and can easily be applied to other areas. With this new method of group- ing a more realistic parametrization of process-based rain forest growth models seems to be possible.