Search

The mechanistic basis for problem-specific, question-driven ecological simulation models often includes general physical phenomena. Although linking general physical environmental process models with custom-written spatially-explicit ecological models appears conceptually straightforward, practical implementation of such structures is often problematic for ecological modellers and may require specific software engineering knowledge. We describe a general coupling framework that introduces an intermediate layer between the two types of models. We used a popular programming language (NetLogo) to implement a spatially-explicit simulation model for insect population growth and migration, and a commonly-used complex atmospheric dispersion model (HYSPLIT) to project air particle (here insect pest) transport, dispersion and deposition. We developed an algorithm that facilitates an ongoing dialogue between these two models. The dialogue takes place along a temporal scale with frequency and duration constrained only by the study objective. Our approach should be readily adaptable to analogous problems.

Spatially-explicit individual-based simulation models provide a valuable tool for exploring complex ecological and evolutionary processes that are not easily empirically measured. Here, we present modifications of a spatially-explicit individual-based simulation model (CDMetaPOP) to accommodate a two-species system and simulations involving interspecific hybridization. We first describe how a hybrid (H) index is used to distinguish individuals of interspecific descent from those of either parental species. User-defined thresholds provide flexibility in the degree of admixture tolerated for classifying ‘pure’ individuals. We then detail relationships further informed by the H index, including individual growth, temperature-based fitness and selection, and mate preference behavior. Empirically derived species- and system-specific information can be incorporated into these relationships, for example, to produce differential growth among hybrids and parental species. Lastly, we demonstrate an application of this simulation framework by exploring the relative effects of temperature-based selection, mate preference behavior, and hybrid fitness on the rate and spatial extent of sympatric hybridization between two native riverine fish species, bull trout (Salvelinus confluentus) and Dolly Varden (Salvelinus malma), in the upper Skagit River system (United States and Canada). Results from this demonstration provide guidance for future empirical studies of bull trout, a federally threatened species. Understanding factors that contribute to the initiation and maintenance of hybridization, as well as the ecological and evolutionary consequences of this phenomenon, is of increasing importance given shifting species ranges due to large-scale landscape modification and a changing global climate. Our framework can be used to study a wide range of hybridization dynamics in any terrestrial or aquatic system, including comparisons of distinct environmental conditions or potential management responses.

We demonstrate the use of a surrogate-based optimization framework for large-scale and high-resolution landscape management optimization, using irrigated corn production systems in eastern Colorado, USA as a case study. An artificial neural network was employed to create a surrogate of the DayCent biogeochemical simulation model. Our optimization considered trade-offs among seven different objectives at different scales, including farm profits, irrigation water use, corn grain, corn stover, soil organic carbon (SOC), greenhouse gas (GHG) emissions, and nitrogen leaching. The results show that the surrogate captured greater than 99% of the variations in the DayCent’s simulated outputs and was 6.2 million times faster than the DayCent model for our analysis. Farm-level optimization increased farm profits by 83%–150%, SOC by 16%–53%, grain yield by 10.1–11.3%, and reduced GHG emissions by 19%–55% compared to the ‘business-as-usual’ scenario.

Evaluating stochastic abiotic factors and their combined effects on fish and wildlife populations have been challenging in environmental sciences. Contributing to this challenge is the paucity of data describing how observations made on individuals related to exposure to chemical compounds ultimately effect population vital rates, and how this exposure interacts with other abiotic drivers. Using three smallmouth bass populations in Pennsylvania as a case study, we explored both single-factor and combined effects of water temperature, flow velocity and chemical exposure on populations through a model simulation. Although there are many pathways through which chemicals can affect population vital rates, we focused on one where chemical exposure leads to reduced reproduction. Among the three populations considered, two (the Juniata and Susquehanna populations) have experienced adverse health, including the potential adverse effects of environmental stress and chemical contamination that may cause disease and mortality of young-of-year (YOY), various skin lesions and a high prevalence of intersex or testicular oocytes in adults. The third population (The Alleghany population) has not encountered mortality events of YOY and intersex prevalence is much lower. The simulation involved projecting populations using a length-based model under different environmental conditions. In the simulations, abiotic factors influenced population dynamics through their impacts on growth, survival and recruitment. Response to the same environmental stress varied by population and life-stage of the species. Factors affecting young adult and adult life-stages had great influence on proportional stock density (PSD) and the probability of having PSD within the suggested range (PSD probability). Increases in water temperature had a negative effect and dominant role in the combined effect on population size structure (e.g., PSD and PSD probability) – increases in flow velocity during the spring season also had a negative effect on abundance. Populations with high recruitment rates sustained relatively large population size, even under high water temperature and/or high flow velocity, which suggests that factors and management strategies that benefit recruitment (such as reduced chemical contaminants) may compensate for the negative effects of warming water temperatures and high spring flow velocity on population size.

We developed and applied a wildfire simulation package in the Envision agent-based landscape modelling system. The wildfire package combines statistical modelling of fire occurrence with a high-resolution, mechanistic wildfire spread model that can capture fine scale effects of fire feedbacks and fuel management, and replicate restoration strategies at scales that are meaningful to forest managers. We applied the model to a landscape covering 1.2 million ha of fire prone area in central Oregon, USA where wildland fires are increasingly impacting conservation, amenity values and developed areas. We conducted simulations to examine the effect of human versus natural ignitions on future fire regimes under current restoration programs, and whether contemporary fire regimes observed in the past 20 years are likely to change as result of fire feedbacks and management activities. The ignition prediction model revealed non-linear effects of location and time of year, and distinct spatiotemporal patterns for human versus natural ignitions. Fire rotation interval among replicate simulations varied from 78 to 170 years and changed little over the 50-yr simulation, suggesting a stable but highly variable and uncertain future fire regime. Interestingly, the potential for fire-on-fire feedbacks was higher for human versus natural ignitions due to human ignition hotspots within the study area. We compare the methods and findings with other forest landscape simulation model (FLSM) studies and discuss future application of FLSMs to address emerging wildfire management and policy issues on fire frequent forests in the western US.

Fast growth of the human population puts high pressure on pasture habitats due to increasingly high grazing intensities on shrinking grazing land. Adapted herbivore management is needed to maintain the long term productivity of the rangeland ecosystem, especially in dry climates where precipitation is highly erratic and where forage yield and quality of the pasture are highly dynamic. We used an agent-based approach to develop a spatially explicit model for a case study region in Madagascar. Our model “RaMDry” integrates the movements and feeding metabolism of domesticated ruminants in order to assess the potential of adaptive livestock production in a highly dynamic, heterogeneous, semi-arid rangeland system. It evaluates the additional metabolic energy costs due to pastoral herd movements in search of forage, incorporates seasonal dynamics in forage quality in terms of feed digestibility and relates forage availability and quality to climatic conditions. In the presented study, we focused on describing the processes simulated in RaMDry in detail, simplifying model conditions by implying free-ranging conditions. We verified the results of the model through global and local sensitivity analysis and pattern-oriented modeling and compared findings with observed patterns and existing data from literature. Our model provides a useful tool to assess strategies and effects of locally adapted herd and rangeland utilization for sustained food security and household economy of livestock keepers in the semi-arid tropics and sub-tropics. Subject for further analysis of the model is its capability to simulate different management strategies and climatic conditions as well as the development of the models’ capacity to predict accumulative effects of environmental degradation.

Effective decision support tools are required in order to provide agricultural practitioners with advice regarding appropriate and economic pest management strategies. Process-based simulation models could enhance farmer’s abilities to make knowledge-based decisions regarding both the timing and extent to which chemical management is relied upon in an agronomical context. This work describes a new simulation model that quantifies the size and timing of grain aphid populations in response to temperature. The simulation model is comprised of compartmentalised aspects of the aphid’s life cycle, which interact with one another to produce the population dynamics. The model is subjected to an independent evaluation in order to assess its descriptive capacity in comparison with previous simulation models, and is shown to constitute an improved tool to describe the seasonal population dynamics of S. avenae.

We present a modeling approach to assess the impact of different conservation management options on the population dynamics of the moor frog (Rana arvalis, Amphibia, Anura) in dynamic river floodplains. The model is targeted at supporting amphibian conservation actions within floodplain management efforts for human flood protection. FloMan-MF is an individual-based, spatially explicit simulation model that combines three aspects: the hydrological dynamics of the landscape influenced by floods, frog reproductive behavior modeled through migration and selection of suitable reproduction sites (ponds), and population structure and trajectory over time. We validated each of the three systems and applied the baseline scenario to a two-year case study conducted in the Middle Elbe River floodplain (Germany). The FloMan-MF model aims to identify suitable compositions of floodplain ponds and habitat characteristics to optimize the reproductive habitat for moor frogs in a degraded floodplain. Therefore, to show potential applications of the model, we simulated floodplain habitat degradation and five different management scenarios. The population responded differently to each of the simulated management actions, ranging from rapid extinction to recovery with low numbers of individuals. The partly unexpected results reveal complex interactions in species-habitat interactions over time. According to our results, the establishment of dynamic temporary ponds seems to be a basic condition for population persistence of this species and can be realized within floodplain management activities. Extending the rationale and basic design of FloMan-MF to other amphibian species living in floodplains is straightforward.

Projected climate change implies that site conditions can no longer be expected to remain constant over a tree’s lifetime. The fast and complex changes in site characteristics and growth patterns diminish the value of traditional knowledge and profoundly alter the conditions of forest management. One way to tackle the inherent uncertainties are simulation studies addressing these new dynamics and mechanisms. The aim of this study is to present such a simulation model system comprising various established and validated process-based and statistical models assessing the complex and dynamic response of a forest stand to climate change. For a given climate scenario, these coupled models estimate the potential growth and yield and various risks considering changing site and stand conditions. As an example, the model system is applied to managed forest stands of Norway spruce ( Picea abies (L.) H. Karst.) in a forest district located in central-western Germany. For the changing climate conditions according to SRES B1 and A1B, the model results suggest a positive effect on the site index and, by contrast, a negative impact on tree survival of increasing risks regarding drought stress mortality, wind damage, and bark beetle infestation given the climate change scenario. The annual contribution margin of timber production under consideration of damage risks by drought stress mortality, wind, and bark beetle infestation reveals that, in this case, the increased growth is able to compensate for the higher risks with few exceptions. Furthermore, we discuss the advantages and challenges of employing a dynamic complex simulation model system for climate change impact assessment based on high-resolution climate data.

Mosquitoes are vectors of numerous pathogens that cause infectious diseases, and they pose a significant global health burden as a result. As such, more reliable field-relevant models to study mosquito population dynamics and life history traits such as development time and survival of mosquito larva would be of great value. In Aedes mosquitoes, progression through early life stages is known to be density-dependent. Despite its importance, density dependence is largely ignored or oversimplified in many existing simulation models, leading to less accurate predictions of development and survival during the early life stages. Furthermore, density dependence is frequently assumed to impact only larval survival and not development time in models, despite empirical evidence for density-dependent development. Here, we develop a discrete-time model of mosquito larval population dynamics which accounts for density impacts on both survival and development time. We demonstrate the validity of our model using publicly available semi-field data of larval density and pupation time across a six-month experiment. Using our model, we found that incorporating density dependence during larval development is important to the accurate prediction of mosquito pupation. This is especially true when considering density-dependent development time for mosquito larva as opposed to density-dependent larval survival. We determined that the incorporation of simple functional forms to describe density dependence in simulation models gives improved prediction results over models that ignore density dependence entirely. Such simple functional forms can easily be incorporated into existing models, and thus help improve field-relevant models of mosquito population dynamics, particularly in Aedes and other container-inhabiting mosquitoes that are known to experience density dependence during larval development.

In this study we investigate the relationships of emergy storage or flow to biodiversity using three different models—a dynamic simulation model, a static scenario model, and a modified ecological network model. These models attempt to explain how disturbance regime, latitude, and trophic complexity are related to observed patterns of renewable emergy flows and storages and biodiversity. A prior hypothesis, which this work seeks to examine, suggests that as renewable emergy flow increases biodiversity will increase. In this regard, we simulate how H.T. Odum’s original CLIMAX model, which tracks forest biomass and diversity over 100 years of succession, responds to a periodic disturbance. The static scenario model compares emergy flow, storage and diversity in five forest eco-regions along the east coast of the United States. An emergy flow matrix ecological network model was used to simulate biodiversity in a mature forest ecosystem and in a typical suburban forest system to investigate how the complexity of a forest system will affect emergy throughput. Comparisons were made for the Shannon diversity index and transformity at the individual trophic level. These comparisons seek to further our understanding of the relationship of emergy and biodiversity and to validate the use of renewable emergy flow to explain ecological phenomena (e.g., biodiversity increasing as latitude decreases, biodiversity increasing through the stages of forest succession).

Freshwater systems provide iconic examples of ecological tipping points. In some lakes and ponds, changes to nutrient levels can drive sudden shifts between primary producer communities dominated by either submerged or floating plants. Several models, ranging in complexity, have been developed to understand the interaction between these primary producer groups. Previous field studies suggest that spatial (e.g., water body size) and temporal (e.g., seasonality) processes are important for the dynamics of this system in nature, but these processes cannot be included in most models without a significant increase in model complexity. Therefore, I developed a spatially- and temporally-explicit model of this system with moderate model complexity that extends a previous model, in which alternative states are known to occur. I found that under low (approximately ≤2 mg total nitrogen L −1 ) or high (approximately ≥6 mg total nitrogen L −1 ) nutrient levels, either submerged or floating plants dominated, respectively. At intermediate nutrient levels, simulations resulted in different final plant states, depending on the initial cover of floating and submerged plants, providing evidence for alternative states. Under most conditions, stable intermediate states were uncommon. Water body size had a large effect on the dynamics of the system, as observed in the field, but only if wind strengths increased with water body size and there was a prevailing wind direction. Surprisingly, species composition and trait diversity did not appear to have major effects on the final plant states. This model allows the integration of processes on multiple scales of biological organization, from species traits and composition, to climate and seasonality or ecosystem-level properties, and it complements the growing realization that spatial context has significant impacts on the dynamics of alternative states in nature.

Individual-based modeling is considered an important tool in ecology and other disciplines. A major challenge of individual-based modeling is that it addresses complex systems that include a large number of entities, hierarchical levels, and processes. To represent these, individual-based models (IBMs) usually comprise a large number of submodels. These submodels might be complex by themselves and interact with each other in many ways, which in turn can affect the overall system behavior in ways that are not always easy to understand. As a result, both the validity and credibility of IBMs can be limited. We here demonstrate how a cascaded design of simulation experiments (cDOE) may support the validity and efficiency of the analysis of IBMs and other ecological simulation models. We take a systematic approach that adopts a divide-and-conquer strategy. In a preparatory phase, submodels and their parameters are configured in “subexperiments”. Consequently, the “top-level experiments” of the simulation model can assess the research questions in a more valid and efficient way. Our strategy thus supports the structural realism of individual-based models because both the behavior of their main components and the relationships between these components are explicitly addressed.

Forest managers are faced with the difficulty of promoting multiple ecosystem services (ES) with an incomplete understanding of the complex ways these ES interact. We developed a quantitative model to help managers better understand the effect of different management options on eight ES over time. These ES were (1) Mexican spotted owl (Strix occidentalis lucida) habitat, (2) the harvest of merchantable timber, (3) the harvest of woody biomass for potential conversion to bioenergy, (4) northern goshawk (Accipiter gentilis) habitat, (5) scenic beauty, (6) fire hazard reduction, (7) carbon storage, and (8) restoration of pre-European settlement forest structure represented through larger quadratic mean diameters (QMD). We integrated production functions for these eight ES with a forest growth-and-yield model to simulate the effects of three management actions over a 45-year period in northern Arizona, USA. We scaled ES values based on their observed maximum and minimum values as a means to compare achievement levels between ES. These scaled ES values were input into a goal-programming model to identify the optimal management regimes given a corresponding set of five different management objectives. The combination of simulation modeling with goal-programming is a novel approach for evaluating ES responses to management and for optimizing ES attainment objectives. To illustrate the value of this novel approach, we designed management objectives to reflect the likely goals a forest manager may choose for each ES in this region. ES goals were distinguished at the forest- and stand-level spatial extents. Results demonstrate the flexibility of the ES optimization model to plan for a variety of situations and preferences. For instance, two ES management goals viewed as conflicting are the protection of Mexican Spotted Owl habitat and the reduction of fire hazard risk. One ES optimization scenario (Scenario 5) produces a plan that reduces area considered as having a high fire hazard from 56% to 7% of the study area in 2015 without any reduction in owl habitat quality throughout the time horizon. We recommend that our methodology be applied and developed further by those in both the research and applied fields of ecosystem sciences.

Geography investigates changes in physical structures and distributions of objects in spatiotemporal world, which are shaped by geographic process (geo-process). With extensive simulation models used to study geo-process, this paper examines the status of geo-process modeling (namely model-based simulation) for multidisciplinary geo-processes across scales in virtual geographic environments (VGEs). The conceptual framework of integrated modeling in VGEs is proposed with a review of specific issues, including model sharing and management, collaborative modeling and uncertainty analysis. The contribution of a model base in model reusability and modeling management, concerning input data, parameterization, and simulation output, is detailed. Finally, this paper concludes with a discussion of future research directions for holistic geo-process modeling.

We integrated a widely used forest growth and management model, the Forest Vegetation Simulator, with the FSim large wildfire simulator to study how management policies affected future wildfire over 50 years on a 1.3 million ha study area comprised of a US national forest and adjacent lands. The model leverages decades of research and development on the respective forest growth and wildfire simulation models, and their integration creates a strategic forest landscape model that has a high degree of transparency in the existing user communities. The study area has been targeted for forest restoration investments in response to wildland fires that are increasingly impacting ecological conditions, conservation areas, amenity values, and surrounding communities. We simulated three alternative spatial investment priorities and three levels of management intensity (area treated) over a 50-year timespan and measured the response in terms of area burned, fire severity, wildland-urban interface exposure and timber production. We found that the backlog of areas in need of restoration on the national forest could be eliminated in 20 years when the treatment rate was elevated to a maximum of 3× the current level. However, higher rates of treatments early in the simulation created a future need to address the rapid buildup of fuels associated with understory shrub and tree regeneration. Restoration treatments over time had a large effect on fire severity, on average reducing potential flame length by up to 26% for the study area within the first 20 years, whereas reductions in area burned were relatively small. Although wildfire contributed to reducing flame length over time, area burned was only 16% of the total disturbed area (managed and burned with prescribed fire) under the 3× management intensity. Interactions among spatial treatment scenarios and treatment intensities were minimal, although inter-annual variability was extreme, with the coefficient of variation in burned area exceeding 200%. We also observed simulated fires that exceeded four times the historically recorded fire size. Fire regime variability has manifold significance since very large fires can homogenize fuels and eliminate clumpy stand structure that historically reduced fire size and maintained landscape resiliency. We discuss specific research needs to better understand future wildfire activity and the relative influence of climate, fuels, fire feedbacks, and management to achieve fire resiliency goals on western US fire frequent forests.

Process-based simulation models are promising tools to integrate biophysical process with soil and climate conditions and then simulate genetic and management impacts on forest productivity. The aim of this study was to adapt, calibrate, evaluate and improve the performance of the APSIM Next Generation Eucalyptus model for different major Brazilian Eucalyptus clones. To these ends, experimental stemwood production data from 2012 to 2017 from eight Eucalyptus clones distributed over 23 locations with contrasting environmental conditions in Brazil were used. The APSIM Next Generation Eucalyptus model, when properly adapted and calibrated, performed well in simulating stemwood biomass and volume, basal area and leaf area index in subtropical and tropical regions and for different genetic entries. For stemwood biomass, the R2 ranged from 0.76 to 0.93 and the Willmott Agreement Index ranged from 0.93 to 0.98, indicating satisfactory precision and accuracy, respectively. As the model performed well, it may be a valuable decision support tool to help foresters in matching suitable genotypes to their sites, to simulate the best management strategies and to assist in long-term forest planning.

Reduction of fire hazard is becoming increasingly important in managed landscapes globally. Fuels reduction prescribed burn treatments are the most common form of reducing fire hazard on landscapes around the world but often result in homogenized fuel age structures and habitats. Alternatively, the size of unplanned fires, and hence fire hazard, can be reduced by controlling the size and patterning of fuels treatments in a patch mosaic arrangement on landscapes. Patch mosaic burning is being implemented globally as a means to increase heterogeneity to mimic natural fire regime results. Funding for prescribed fire programs is often justified primarily on hazardous fuels reduction with secondary consideration given for ecological effectiveness, which can be increased by particular fire mosaic patterns in some systems. The question we address is: Which of two prescribed fire treatment regimes, fuels reduction or patch mosaic burning, reduces fire hazard most effectively? We address the question using computer simulation modeling on synthetic landscapes representing both fire regime treatments. Treatment scale was important. Among fuel reduction treatments, large blocks burned less area than small blocks. For the mosaic treatments, small blocks reduced fire size the most (out of all treatments) and had the least variance in area burned. It is possible to reduce fire hazard and to provide heterogeneous age fuels structure on the landscape, simultaneously benefiting humans and many native fire-dependent species requiring mosaic habitat patterns.

International audience; Ecology makes an increasing use of complex simulation models. As more processes and model parameters are added, a comprehensive model calibration with process-level data becomes costly and predictions of such complex models are therefore often restricted to local applications. In this context, inverse modelling techniques enable to calibrate models with data of the same type than model outputs (thereafter called population data for the sake of clarity, although other data types can be used according to model outputs), which are usually simpler to collect and more readily available. This study aims at demonstrating how such data can be used to improve ecological models, by recalibrating the most influential parameters of a complex model in a Bayesian framework, and at providing general guidelines for potential users of this approach. We used the individual-based and spatially explicit forest dynamics simulation model Samsara2 as a case study. Considering the results of an initial calibration and of a sensitivity analysis as prerequisites, we assessed whether we could use approximate Bayesian computation (ABC) to recalibrate a subset of parameters on historical management data collected in forests with various ecological conditions. We propose guidelines to answer three questions that potential users of the approach will encounter: (1) How many and which parameters are we able to recalibrate accurately with such low-informative data? (2) How many ABC simulations are required to obtain a reasonable convergence of the parameter posterior estimates? (3) What is the variability of model predictions following the recalibration? In our case study, we found that two parameters by species could be recalibrated with forest management data and that a relatively low number of simulations (20,000) was sufficient. We finally pointed out that the variability of model predictions was largely due to model stochasticity, and much less to ABC recalibration and initial calibration uncertainties. Combining direct process-level calibration to ABC recalibration of the most influential parameters opens the door to interesting modelling improvements, such as the calibration of forest dynamics along environmental gradients. This general approach should thus help improve both accuracy and generality of model-based ecological predictions.

Red spruce ( Picea rubens Sargent) has exhibited widespread growth decline and high mortality for the last half century in the eastern United States. Good prediction of this species’ distribution in relation to environmental conditions is critical for effective management. This study projects red spruce distribution in response to multiple causal mechanisms in the Great Smoky Mountains National Park (GSMNP) of the Southern Appalachian Mountains by coupling a temporal simulation model of tree growth (ARIM.SIM) to a species distribution model (ARIM.HAB). ARIM.HAB computed habitat suitability, estimated from ARIM.SIM-generated red spruce growth, for every spatial 30 m grid cell in GSMNP. ARIM.SIM showed that different factors were responsible for habitat suitability and growth at higher vs. lower elevations. The air pollution variables (acid rain and cloud immersion frequency) caused low habitat suitability at higher elevations (1800–2028 m). Reduced air pollution but greater stress from climatic variables (high temperatures, reduced precipitation) caused medium suitability at lower elevations (1400–1600 m). And less stress from air pollution and climate variables combined with ample water to produce highest suitability at intermediate elevations (1600–1800 m). The projected range was verified with an existing geospatial database for red spruce and showed excellent correspondence with present-day distribution (AUC = 0.99, kappa = 0.87 and TSS = 0.88). This research shows that species distribution models coupled with a process-based temporal simulation models can improve the precision and accuracy of, respectively, habitat suitability and range projections for species at local scales.

Ecosystems dominated by big sagebrush, Artemisia tridentata Nuttall (Asteraceae), which are the most widespread ecosystems in semiarid western North America, have been affected by land use practices and invasive species. Loss of big sagebrush and the decline of associated species, such as greater sage-grouse, are a concern to land managers and conservationists. However, big sagebrush regeneration remains difficult to achieve by restoration and reclamation efforts and there is no regeneration simulation model available. We present here the first process-based, daily time-step, simulation model to predict yearly big sagebrush regeneration including relevant germination and seedling responses to abiotic factors. We estimated values, uncertainty, and importance of 27 model parameters using a total of 1435 site-years of observation. Our model explained 74% of variability of number of years with successful regeneration at 46 sites. It also achieved 60% overall accuracy predicting yearly regeneration success/failure. Our results identify specific future research needed to improve our understanding of big sagebrush regeneration, including data at the subspecies level and improved parameter estimates for start of seed dispersal, modified wet thermal-time model of germination, and soil water potential influences. We found that relationships between big sagebrush regeneration and climate conditions were site specific, varying across the distribution of big sagebrush. This indicates that statistical models based on climate are unsuitable for understanding range-wide regeneration patterns or for assessing the potential consequences of changing climate on sagebrush regeneration and underscores the value of this process-based model. We used our model to predict potential regeneration across the range of sagebrush ecosystems in the western United States, which confirmed that seedling survival is a limiting factor, whereas germination is not. Our results also suggested that modeled regeneration suitability is necessary but not sufficient to explain sagebrush presence. We conclude that future assessment of big sagebrush responses to climate change will need to account for responses of regenerative stages using a process-based understanding, such as provided by our model.

Introgressive hybridization with introduced species is threatening the genetic integrity of many native fish populations worldwide. To date, several studies have indicated a direct relationship between rates of hybridization (H) within a sub-population of native fish and the stream distance (D) separating that sub-population from the closest source of introduced non-native genes. However, the relationship between D and H is rarely quantified, and the effect of the spatial arrangement of stream bifurcations and sub-populations (i.e., the network topology) is typically ignored. For this study, we developed and applied a novel individual-based simulation model to compare the spread of non-native genes across three stream networks with differing network topologies: linear, trellis, and dendritic. The model we used simulates mating, survival, and the movements of each fish in the network, where the likelihood of a fish to move between two sub-populations is a function of both D and the number of bifurcations between the two sub-populations. To monitor the spatiotemporal spread of hybridization, the model keeps track of the genetic composition and the breeding location of each fish present in the network over time. Simulation results agree with past field studies that suggest that D is a strong predictor of H. However, our simulations further suggest that the nature and the strength of the H vs. D relationship are likely to be a function of the stream network topology. Specifically, in our simulations, network bifurcations act as permeable barriers to the spread of non-native genes across a network while the strength of the H vs. D relationship is inversely proportional to the complexity of the spatial arrangement of bifurcations and sub-populations. Our study shows that considering network topology could yield a better understanding of observed patterns of hybridization across stream systems, and thus may help inform effective strategies for the management of native fish populations threatened by introgression.

a b s t r a c t Information on tree canopy architecture is crucial in forestry practice because the quality and price of standing trees and final logs directly depend on it. Simultaneously, accurate empirical or functional-structural models require information based on field observations. Pinus radiata is a polycyclic species that follows an acrotony law when forming a new branch cluster, showing smaller branches in its base and larger ones at the top of a forming cluster. The objective of this study was to describe the acrotony of the branches in a branch cluster as a Markov chain. Markov chains represent stochastic pro- cesses in discrete time that undergo a transition from one state to another among a finite number of possible states. The probability of transition from state i to a state j depends only on the current state, i. For modelling acrotony, the relative vigour (expressed as relative branch diameter in relation to the largest one) of each branch was selected as the stochastic variable and the states corresponded to five possible relative size ranges. The branches observed within a P. radiata cluster were ordered following their relative sizes (from largest to smallest), and Markovian transition matrixes were calculated for each branch cluster (whorls of 3-12 branches). The transition matrixes were defined as the probabilities of one branch being fol- lowed by an equal-sized or smaller branch when observing the cluster from the top down. The obtained Markov chain matrixes were used in a stochastic data simulation, which was validated with an inde- pendent dataset. The presented matrixes can be incorporated into traditional simulation models or functional-structural models. The validation results show that the proposed methodology accurately reflects the variability of the branch sizes in a cluster, and we suggest its application to other species that display a clustered organisation of branches.

Accurate predictions of weed seed dispersal are important for spatial modelling of weed population dynamics. However there are few multi-year data sets available, and they are seldom analysed well. Here we gathered together and re-examined population growth and seed spread data from three grass weed species over 3 years. Using a new seed spread function in the simulation model SOMER it was possible to accurately parametrise population growth and seed spread to replicate the field data of Apera spica-venti (APESV) and Bromus sterilis (BROST). In contrast, the greatest increases in Alopecurus myosuroides (ALOMY) numbers occurred in the direction of machinery movement, which could not be predicted using this function. A probability-sum, exponential-type function gave an excellent fit to the field data for APESV and BROST, with changes in seed spread distances due to species differences, wind, and ‘shading’ from the hedge, accommodated by alterations in the scale, shape, and wind dispersal parameters. Here we describe a method of parametrising the probability function from within a stochastic simulation. Working within the simulation allowed empirical population size data sets to be successfully parameterised, without mathematical integration. This function is fully explained with stochastic simulated examples, and fitted to various two-dimensional longitudinal data sets. Enough detail to enable its parametrisation (including simplification) within a majority of spatial stochastic population models is included.

Soil water movement due to root water uptake is a key process for plant growth and transport of water in the soil plant system. The accuracy of prediction extraction rate by plants depends on selection of proper mathematical models. The water uptake by plant roots has been simulated both at the microscopic and macroscopic levels. The microscopic approach requires detailed information about the dynamic geometry of the plant root system that is practically not available. In the macroscopic approach, a sink term, representing water extraction by plant roots is included in the dynamic water flow equation and the spatial and temporal uptake is controlled by the soil moisture and the plant demand. Different pattern of moisture uptake including constant, linear and exponential are available in various models. Relevant literature points out that a non-linear, exponential and logarithmic macroscopic root water uptake model is popular due to their improved prediction efficiency. In this paper most commonly used models which are suitable for the estimation of moisture uptake by plants in different agro-climatic regions are reviewed and a field study of macroscopic model has been given. Model predicted soil-moisture parameters i.e. , moisture depletion, moisture status at various depths and soil moisture profile in root zone are compared with experimental results.

An assessment of the relative importance of vegetation change and disturbance as agents of landscape change under current and future climates would (1) provide insight into the controls of landscape dynamics, (2) help inform the design and development of coarse scale spatially explicit ecosystem models such as Dynamic Global Vegetation Models (DGVMs), and (3) guide future land management and planning. However, quantification of landscape change from vegetation development and disturbance effects is difficult because of the large space and long time scales involved. Comparative simulation modeling experiments, using a suite of models to simulate a set of scenarios, can provide a platform for investigating landscape change over more ecologically appropriate time and space scales that control vegetation and disturbance. We implemented a multifactorial simulation experiment using five landscape fire succession models to explore the role of fire and vegetation development under various climates on a neutral landscape. The simulation experiment had four factors with two or three treatments each: (1) fire (fire and no fire), (2) succession (dynamic and static succession), (3) climate (historical, warm-wet, warm-dry), and (4) weather (constant, variable). We found that, under historical climates, succession changed more area annually than fire by factors of 1.2 to 34, but one model simulated more landscape change from fire (factor of 0.1). However, we also found that fire becomes more important in warmer future climates with factors decreasing to below zero for most models. We also found that there were few differences in simulation results between weather scenarios with low or high variability. Results from this study show that there will be a shift from vegetation processes that control today's landscape dynamics to fire processes under future warmer and drier climates, and this shift means that detailed representations of both succession and fire should be incorporated into models to realistically simulate interactions between disturbance and vegetation.

While they can be useful tools to support decision-making in ecosystem management, robust simulation models of ecosystems with alternative states are challenging to build and validate. Because of the possibility of alternative states in model dynamics, no trivial criteria can provide reliable and useful metrics to assess the goodness-of-fit of such models. This paper outlines the development of the model TRITON, and presents simulation-based validation and analysis of model sensitivity to input parameters. TRITON captures the local dynamics of seaweed-based rocky reef communities in eastern Tasmania, which now occur in two alternative persistent states: (1) either as dense and productive seaweed beds, (2) or as sea urchin ‘barrens’ habitat, i.e. bare rock largely denuded of macroalgae and benthic invertebrates due to destructive grazing by sea urchins. Pattern-oriented-modelling, i.e. comparing patterns in model dynamics across Monte–Carlo simulations with direct observations of Tasmanian reef communities over large scales, provides a valuable approach to calibrate the dynamics of TRITON. Using the computationally efficient, model-independent extended Fourier amplitude sensitivity test, we identify fishing down of predatory lobsters, sea urchin recruitment rate, as well as seaweed growth rate as key parameters of influence on overall model behaviour. Through a set of independent sensitivity tests, we isolate different sets of drivers facilitating the ‘forward’ shift from the seaweed bed to the urchin-dominated state, and the reverse or ‘backward’ shift from denuded sea urchin barren to recovery of seaweed cover. The model suggests that rebuilding populations of large rock lobsters, which predate the urchins, will be effective in limiting ongoing formation of sea urchins barrens habitat, but that the chances of restoring seaweed beds from extensive barrens are relatively low if management relies solely on rebuilding stocks of large rock lobsters. Moreover, even when it does occur, seaweed bed restoration takes up to three decades in the simulations and so is arguably unrealistic to implement under short-term fishery management plans. The process of model validation provided both a better understanding of the key drivers of community dynamics (e.g. fishing of predatory lobsters), and an assessment of priority areas for future research.

Swedish Forest Soil Inventory (SFSI) estimates of SOC stocks and SOC changes for forest on mineral soils under Scots pine ( Pinus sylvestris )/lodgepole pine ( Pinus contorta ) or Norway spruce ( Picea abies ) were compared with estimates, including uncertainties due to parameter, input and climate variability, from two process-based models (Yasso07 and Q) for the period 1994–2000. We found that the stocks, changes, inter-annual variations and uncertainties were of the same magnitude among the different methods. The mean Swedish national stocks in 2000 were estimated to be 73 (±10) (95% CL) ton ha −1 C (SFSI); 69 (±9) (95% CL) ton ha −1 C (Yasso07); and, 67 (+10; −9) (5th and 95th percentiles) ton ha −1 C (Q). Between 1994 and 2000, the mean estimated SOC change were 6.6 (±7) Tg C yr −1 (SFSI), 1.7 (±8.8) Tg C yr −1 (Yasso07), and −3.2 (+10.5; −16.9) Tg C yr −1 (Q). Spatial variability was the main source of uncertainty for the SOC stocks and changes estimated with the SFSI. The uncertainties in the stock estimates originated from litter input for Yasso07 and from the model parameters for the Q model. In both models, litter input uncertainty was the major source of uncertainty for the estimated SOC changes, followed by climate variability and parameters. We concluded that the level of uncertainty for both methods was similar but the sources of uncertainties varied between models and measurements. Thus, comparing uncertainty between methods is difficult and further studies on SOC change estimates with related uncertainties are warranted.

The role of fungi in soil ecosystem sustainability is poorly understood, as is the extent to which it is affected by the microscale heterogeneity of soils with respect to structure, chemistry and biology. This is due to the complexity of soil ecosystems, presenting significant challenges to their study in situ. Many theoretical and simulation models have been developed to link nutrient levels to colony dynamics. Unfortunately, there is currently no model that can take both structural and nutritional microscale heterogeneity into account, and is parameterised for the soil environment. In this context, the objective of this article is to develop such a 3D spatially explicit model of fungal dynamics, and to calibrate it for a soil system using data from the literature. A sensitivity analysis is carried out to better understand the uncertainties in the input parameters and their effect on colony dynamics in terms of biomass yield and respiration rates. The results highlight simulation outcomes that are most suited to validation by experimentation. The results also indicate that predictions in biomass yield are sensitive to uncertainties in model parameters relative to the soil–fungal complex that at this point are insufficiently understood experimentally and still have to be estimated by model fitting. The latter parameters, which influence biomass yield and respiration, are associated with biomass recycling processes such as adsorption (,αni) desorption (βni, βi), insulation (ζni) and biomass yield efficiency (ɛ1), and translocation (Dv). The model now opens up great opportunities for hypothesis-driven research, combining theoretical models and novel types of experimentation, especially given the recently acquired ability to generate artificial, replicable soil-like microcosms on which to test model predictions.

Recent increases in computation power have prompted enormous growth in the use of simulation models in ecological research. These models are valued for their ability to account for much of the ecological complexity found in field studies, but this ability usually comes at the cost of losing transparency into how the models work. In order to foster greater understanding of the functioning of computer simulation models, we develop an analytical approximation of the Landscape Age-class Demographics Simulator (LADS; Wimberly, 2002 ), a representative example of broad group of models that simulate landscape-scale forest dynamics in response to a series of recurring disturbances that interact spatially with existing landscape structure. Much of the model output was produced mathematically, without generating a series of disturbances (in this case, fire) or simulating the forest response to each disturbance. The approximation provides a detailed understanding of the modeled fire regime. Also, it provides equations that directly specify the roles of key input parameters rather than having to infer these roles indirectly from model output in a sensitivity analysis. The application of analytical methods typically has been limited to simple scenarios that lack feedbacks or spatial interactions, but in this exercise, analytical methods address much of the complexity more commonly addressed by simulation: the modeled landscape is composed of two provinces, each with a unique fire frequency and fire-size distribution; stochastic variation in the number of fires per year and the size of each fire; and two levels of fire severity that each have different effects on forest structure. Analytical approximation is not suggested as an alternative to simulation models, but rather, as a complementary approach aimed at improving insight into model function.

a b s t r a c t Because of human pressures, the need to understand and predict the long-term dynamics and develop- ment of subtropical dry forests is urgent. Through modifications to the ZELIG simulation model, including the development of species- and site-specific parameters and internal modifications, the capability to model and predict forest change within the 4500-ha Guanica State Forest in Puerto Rico can now be accomplished. Published datasets and additional data from the U.S. Forest Service Forest Inventory Anal- ysis were used to parameterize the new gap model, ZELIG-TROP. We used data from permanent plots

a b s t r a c t Different methodologies for evaluating aspects of model performance going beyond the pure agreement between measured and simulated data have been recently proposed. These indicators and criteria for the evaluation of, e.g., complexity and robustness can be used in conjunction with well-known metrics for the evaluation of model accuracy - such as root mean square error and modelling efficiency - to get a deeper knowledge about models structure and behaviour. The aim of this paper is to propose an indi- cator of model plasticity, defined as the aptitude of a model to change the sensitivity to its parameters while changing the conditions of application. Sensitivity was here analyzed using the Sobol' method for sensitivity analysis (SA). Concordance among parameters relevance (total order effect) estimated under different conditions allowed to quantify changes in the way models react to different environments. The concordance among the different SA results was related to the variability of a normalized agrometeoro- logical indicator used to characterize the explored conditions. The plasticity indicator was tested using three different crop models (WARM, CropSyst, WOFOST; rice was simulated), 10 European locations, and 10 years for each location, for a total of 5,939,200 simulations and 300 SA experiments. Results indicated WOFOST as the most plastic, both within location, year, and using all the combinations location × year, whereas WARM showed to be the less plastic across the conditions explored. Previous studies carried out on the same models in northern Italy seem to suggest a direct relationship between model complexity and plasticity, whereas model accuracy seems to be unrelated to these features. This consideration under- lines that, in case of availability of different models with a similar degree of accuracy, different choices should be performed for different modelling studies, characterized by different aims and conditions of application.

Previous ecological models for disturbance from energetic perspectives have focused on destructive pulses by which storages in a system are quickly drained during disturbance events and recovered thereafter. However, considering the wide range of disturbance intensities, frequencies, and durations in nature, disturbance effects on ecosystem energetics would be better understood by diversifying the disturbance effects on specific system configurations or energy pathways. Based on two hypotheses, we built simulation models of the variable disturbance–productivity relationships observed in a freshwater aquatic microcosm study. First, we hypothesized that disturbances will differentially alter the intrinsic rates of energy pathways in a system. Second, we hypothesized that there is a disturbance threshold where response of the intrinsic rates changes abruptly. Simulation results showed variable patterns of gross primary productivity (GPP) during the disturbance and post-disturbance periods under the diverse scenarios of disturbance effects on the intrinsic rates. Simulation results confirmed that the second hypothesis (i.e., disturbance threshold) was essential to achieve a U-shaped or peaked disturbance–productivity relationship. We evaluated the models by comparing them with the results of the microcosm tests, and suggested possible mechanisms of the variable disturbance–productivity relationships by varying parameters related to the disturbance effects on the intrinsic rates and the disturbance thresholds.

International audience; Carbon is an important determinant of the sustainability of West African farming systems and of the atmospheric greenhouse effect. Given the complexity of C dynamics, various simulation models have been developed. Few include socioeconomic factors or handle system heterogeneity. This study pro- poses a generic, multi-agent model for the analysis of C dynamics at village level. It assumes that a better analysis of carbon dynamics at village level requires account to be taken of social, economic, physical and biological factors as well as of the actions of individuals and their interdependence. The Carbon of Territory Multi-Agent Simulator (CaTMAS) model is based on the Organization-Role-Entity-Aspect (OREA) meta-model and the Multi-Agent Systems (MAS) approach. OREA enables C dynamics to be studied from various points of view through the roles played by entities within organizations and also allows various entities to play the same role in various ways through the notion of aspects. The model was coupled with the Century model and a geographical information system to provide a realistic representation of C dynamics. CaTMAS provides not only a framework for the explicit description of the carbon dynamics of farming systems but can also be used to assess the viability of farming systems using various socioeconomic and biophysical scenarios. The model includes interactions between human activities and the environment. Simple simulations involving two cropping systems and focusing on the impact of population growth and different climate regimes on the C dynamics indicate that CaTMAS accounts realistically for the relationships between population, agriculture, climate and SOC dynamics. In simulation, population growth, which drives food demand, leads to agricultural expansion, land scarcity and decrease in fallow duration. These effects are accentuated by increasing temperature and decreasing rainfall which affect the SOC dynamics controlling soil fertility and thus crop production. Improvements to the model should make it possible to extend the scale of the simulation of C dynamics and include refinements such as the inclusion of the trading of carbon credits.

The impact of anthropogenic disturbance on wildlife is increasing becoming a source of concern as the popularity of outdoor recreation rises. There is now more pressure on site managers to simultaneously ensure the continued persistence of wildlife and provide recreational opportunities. Using ‘Simulation of Disturbance Activities’, a model designed to investigate the impact of recreational disturbance on wildlife, we demonstrate how a simulation modelling approach can effectively inform such management decisions. As an example, we explored the implications of various design and management options for a proposed recreational area containing a historic breeding bird colony. By manipulating the proximity, orientation and intensity of recreation, we were able to evaluate the impact of recreational activities on the behaviour of black-crowned night-heron nestlings (Nycticorax nycticorax). Using a classification and regression tree (CART) procedure to analyse simulation output, we explored the dynamics of multiple strategies in concert. Our analysis revealed that there are inherent advantages in implementing multiple strategies as opposed to any single strategy. Nestlings were not disturbed by recreation when bird-watching facility placement (proximity and orientation) and type were considered in combination. In comparison, proximity alone only led to a

The spread of invasive species is a major ecological and economic problem. Dynamic spread modelling is a potentially valuable tool to assist regional and central government authorities to monitor and control invasive species. To date a lack of suitable data has meant that most broad scale dispersal models have not been validated with independent datasets, and so their predictive ability and reliability has remained unscrutinised. A dynamic, stochastic dispersal model of the widely invasive plant Buddleja davidii was calibrated on European spread data and then used to project the temporal progression of B. davidii's distribution in New Zealand, starting from several different historical distributions. To assess the model's performance, we constructed an occupancy map based on the average number of simulation realisations that have a population present. The application of Receiver Operating Characteristic (ROC) curves to occupancy maps is introduced, but with specificity substituted by the proportion of available area used in a realisation. A derivative measure, the partial area under these curves when assessed through time (pAUC), is introduced and used to assess overall performance of the spread model. The model was able to attain a high level of model sensitivity, encompassing all of the known locations within the occupancy envelope. However, attempting to simulate the spread of this invasive species beyond a decade had very low model specificity. This is due to several factors, including the exponential process of spread (the further a population spreads the more sites exist from which it can spread stochastically), and the Markovian chain property of the stochastic system whereby differences between realisations compound through time. These features are seen in many reports of spread models, without being explicitly acknowledged. Our measure of pAUC through time allows a model's temporal performance and its specificity to be simultaneously assessed. While the rapid deterioration in model performance limits the utility of this type of modelling for forecasting long-term broad-scale strategic management of biological invasions, it does not necessarily limit its attractiveness for informing smaller scale and shorter term invasion management activities such as surveillance, containment and local eradication.

The HFire fire regime model was used to simulate the natural fire regime (prior to European settlement) on Kennedy Space Center, Merritt Island National Wildlife Refuge, Canaveral National Seashore, and Cape Canaveral Air Force Station, Florida. Model simulations were run for 500 years and the model was parameterized using information generated from previously published empirical studies on these properties (e.g., lightning fire ignition frequencies and ignition seasonality). A mosaic pattern of frequent small fires dominated this fire regime with rare but extremely large fires occurring during dry La Nina periods. This simulated fire size distribution very closely matched the previously published fire size distribution for lightning ignitions on these properties. A sensitivity analysis was performed to establish which parameters were most influential and the range of variation surrounding empirically parameterized model output. Dead fuel moisture and wind speed had the largest influence on model outcome. A wide range of variance was observed surrounding the composite simulation with the least being 6% in total burn frequency and the greatest being 49% in total area burned. Because simulation modeling is the best option for fire regime reconstruction in many rapidly growing shrub dominated systems, these results will be of interest to scientists and fire managers for delineating the natural fire regime on these properties, the southeastern United States and other fire adapted shrub systems worldwide.