Search USGS⌕ Search

SEARCH · Search USGS

Results for “Ecological Modelling”

Search indexed USGS publications on groundwater, aquifers, geologic maps, mineral resources and earthquakes. Explore source records by subject and place.

Quote a phrase for an exact phrase match. Source license links do not imply unrestricted reuse.

At least 433 records · Page 24Linked to original sources
USGS2008

Integrated monitoring and information systems for managing aquatic invasive species in a changing climate

Changes in temperature, precipitation, and other climatic drivers and sea-level rise will affect populations of existing native and non-native aquatic species and the vulnerability of aquatic environments to new invasions. Monitoring surveys provide the foundation for assessing the combined effects of climate change and invasions by providing baseline biotic and environmental conditions, although the utility of a survey depends on whether the results are quantitative or qualitative, and other design considerations. The results from a variety of monitoring programs in the United States are available in integrated biological information systems, although many include only non-native species, not native species. Besides including natives, we suggest these systems could be improved through the development of standardized methods that capture habitat and physiological requirements and link regional and national biological databases into distributed Web portals that allow drawing information from multiple sources. Combining the outputs from these biological information systems with environmental data would allow the development of ecological-niche models that predict the potential distribution or abundance of native and non-native species on the basis of current environmental conditions. Environmental projections from climate models can be used in these niche models to project changes in species distributions or abundances under altered climatic conditions and to identify potential high-risk invaders. There are, however, a number of challenges, such as uncertainties associated with projections from climate and niche models and difficulty in integrating data with different temporal and spatial granularity. Even with these uncertainties, integration of biological and environmental information systems, niche models, and climate projections would improve management of aquatic ecosystems under the dual threats of biotic invasions and climate change

Conservation Biology↗
USGS2016

Estimated use of water in the Delaware River Basin in Delaware, New Jersey, New York, and Pennsylvania, 2010

The Delaware River Basin (DRB) was selected as a Focus Area Study in 2011 by the U.S. Geological Survey (USGS) as part of the USGS National Water Census. The National Water Census is a USGS research program that focuses on national water availability and use and then develops new water accounting tools and assesses water availability at both the regional and national scales. One of the water management needs that the DRB study addressed, and that was identified by stakeholder groups from the DRB, was to improve the integration of state water use and water-supply data and to provide the compiled water use information to basin users. This water use information was also used in the hydrologic modeling and ecological components of the study. Instream and offstream water use was calculated for 2010 for the DRB based on information received from Delaware, New Jersey, New York, and Pennsylvania. Water withdrawal, interbasin transfers, return flow, and hydroelectric power generation release data were compiled for 11 categories by hydrologic subregion, basin, subbasin, and subwatershed. Data availability varied by state. Site-specific data were used whenever possible to calculate public supply, irrigation (golf courses, nurseries, sod farms, and crops), aquaculture, self-supplied industrial, commercial, mining, thermoelectric, and hydroelectric power withdrawals. Where site-specific data were not available, primarily for crop irrigation, livestock, and domestic use, various techniques were used to estimate water withdrawals. Total water withdrawals in the Delaware River Basin were calculated to be about 7,130 million gallons per day (Mgal/d) in 2010. Calculations of withdrawals by source indicate that freshwater withdrawals were about 4,130 Mgal/d (58 percent of the total) and the remaining 3,000 Mgal/d (42 percent) were from saline water. Total surface-water withdrawals were calculated to be 6,590 Mgal/d, or 92 percent of the total; about 54 percent (3,590 Mgal/d) of surface water withdrawn was freshwater. Total groundwater withdrawals were calculated to be 545 Mgal/d (8 percent of the total), all of which was freshwater. During 2010, calculated withdrawals by category, in decreasing order, were: thermoelectric power, 4,910 Mgal/d; public supply, 1,490 Mgal/d; self-supplied industrial, 350 Mgal/d; irrigation, 175 Mgal/d; self-supplied domestic, 117 Mgal/d; mining, 41.3 Mgal/d; aquaculture, 19.3 Mgal/d; livestock, 6.72 Mgal/d, and commercial, 5.89 Mgal/d. The amount of instream use for hydroelectric power generation purposes in 2010 was reported to be 273 Mgal/d for the Wallenpaupack Plant and 127 Mgal/d for the Mongaup River system. Total return flows in the DRB were 2,960 Mgal/d in 2010. Although municipal wastewater-treatment plants accounted for 539 (97 percent) of the return-flow sites, they accounted for about 70 percent of the total return flows in the DRB. There was limited information on return flows from thermoelectric power.

Delaware, New Jersey, New York, Pennsylvania↗
USGS2019

Riverscape correlates for distribution of threatened spotfin chub Erimonax monachus in the Tennessee River Basin, USA

Globally, aquatic biodiversity is imperiled at an increasing rate, especially in diversity hotspots such as the southeastern USA. The spotfin chub Erimonax monachus is a federally threatened minnow with a disjunct distribution resulting from numerous impoundments on the Tennessee River and its tributaries in the heart of the southeastern USA. Recovery actions required to remove federal protection for E. monachus are dependent on the establishment of additional populations within the historical range of the species, but little is known regarding macroscale habitat requirements that could guide conservation planning. We analyzed local- and network-scale watershed attributes to develop an ecological niche model (ENM) for E. monachus useful for directing conservation actions at sampled and unsampled sites across the Tennessee River Basin. We found E. monachus occurred most often in larger streams with large upstream catchment areas and minimal alteration to forested uplands, but all of these sites were in close proximity to high densities of downstream dams due to populations being restricted to large-stream habitat upstream of reservoirs. The ENM showed the highest probability of E. monachus occurrence among catchment locations with known extant populations; however, additional historical and previously unoccupied catchments showed potential for successful (re)introductions, provided that fine-scale habitats are appropriate. Our framework can be used to identify potential survey and (re)introduction sites for E. monachus as well as other rare riverine fishes and represents a method for identifying areas of high priority for conserving aquatic biodiversity.

Alabama, Georgia, Kentucky, Mississippi, North Car↗
USGS2023

Enhancing the predictability of ecology in a changing world: A call for an organism-based approach

Ecology is usually very good in making descriptive explanations of what is observed, but is often unable to make predictions of the response of ecosystems to change. This has implications in a human-dominated world where a suite of anthropogenic stresses are threatening the resilience and functioning of ecosystems that sustain mankind through a range of critical regulating and supporting services. In ecosystems, cause-and-effect relationships are difficult to elucidate because of complex networks of negative and positive feedbacks. Therefore, being able to effectively predict when and where ecosystems could pass into different (and potentially unstable) new states is vitally important under rapid global change. Here, we argue that such better predictions may be reached if we focus on organisms instead of species, because organisms are the principal biotic agents in ecosystems that react directly on changes in their environment. Several studies show that changes in ecosystems may be accurately described as the result of changes in organisms and their interactions. Organism-based theories are available that are simple and derived from first principles, but allow many predictions. Of these we discuss Trait-based Ecology, Agent Based Models, and Maximum Entropy Theory of Ecology and show that together they form a logical sequence of approaches that allow organism-based studies of ecological communities. Combining and extending them makes it possible to predict the spatiotemporal distribution of groups of organisms in terms of how metabolic energy is distributed over areas, time, and resources. We expect that this “Organism-based Ecology” (OE) ultimately will improve our ability to predict ecosystem dynamics.

Frontiers in Applied Mathematics and Statistics↗
USGS2019

LANDFIRE remap prototype mapping effort: Developing a new framework for mapping vegetation classification, change, and structure

LANDFIRE (LF) National (2001) was the original product suite of the LANDFIRE program, which included Existing Vegetation Cover (EVC), Height (EVH), and Type (EVT). Subsequent refinements after feedback from data users resulted in updated products, referred to as LF 2001, that now served as LANDFIRE’s baseline datasets and are the basis for all subsequent LANDFIRE updates. These updates account for disturbances and vegetation transitions changes that may not represent current vegetation conditions. Therefore, in 2016 LANDFIRE initiated the Remap prototype to determine how to undertake a national-scale remap of the LANDFIRE primary vegetation datasets. EVC, EVH, and EVT were produced (circa 2015) via modeling for ecologically variable prototyping areas in the Pacific Northwest (NW) and Grand Canyon (GC). An error analysis within the GC suggested an overall accuracy of 52% (N = 800) for EVT, and a goodness of fit of 51% (N = 38) for percent cover (continuous EVC) and 53% (N = 38) for height (continuous EVH). The prototyping effort included a new 81-class map using the National Vegetation Classification (NVC) within the NW. This paper presents a narrative of the innovative methodologies in image processing and mapping used to create the new LANDFIRE vegetation products.

Fire↗
USGS2024

Evaluating spatially explicit management alternatives for an invasive species in a riverine network

Invasive species have substantial ecological and economic costs and removing them can require large investments by management agencies. Optimal spatial allocation of removal effort is critical for efficient and effective management of invasive species. Using a series of ecologically informed model simulations, we evaluated and compared different spatially explicit removal strategies for invasive rusty crayfish ( Faxonius rusticus ) in the John Day River, USA. We assessed strategies in terms of their performance on three likely management objectives: suppression (minimise overall population abundance), containment (minimise the spatial extent of invasion) and prevention (minimise spread into a specific area). We developed five spatial removal strategies to achieve those objectives, denoted as: Target Abundance (removal at locations with the highest population abundance), Target Growth (removal at locations with the highest population growth), Target Edges (removal at the most distant locations in the river), Target Downstream (removal at the most downstream invaded segments on the Mainstem), and Target Random (removal at randomly selected locations). Each strategy was assessed at various effort levels, referring to the number of spatial segments in the river in which removals were conducted, after seven years of management. We identified the alternative that best achieved each objective, based on decision criteria for risk-neutral and risk-averse decision-makers and further evaluated strategies based on Pareto efficiency, which identifies the set of alternatives for which an improvement on one objective cannot be had without a decline in performance on another. We found that Target Abundance and Target Growth strategies best achieved the suppression objective, for risk neutral and risk averse decision-makers, respectively and Target Downstream was always best in achieving the prevention objective across both types of decision-makers. No single strategy consistently performed best in terms of the containment objective. In terms of all three objectives, Target Downstream was consistently Pareto efficient across all levels of management effort and both decision criteria. The modelling framework we provided is adaptable to a variety of riverine invasive species to help assess and compare spatial management strategies.

Oregon↗
USGS2005

Sexual differentiation in the distribution potential of northern jaguars ( Panthera onca )

We estimated the potential geographic distribution of jaguars in the southwestern United States and northwestern Mexico by modeling the jaguar ecological niche from occurrence records. We modeled separately the distribution of males and females, assuming records of females probably represented established home ranges while male records likely included dispersal movements. The predicted distribution for males was larger than that for females. Eastern Sonora appeared capable for supporting male and female jaguars with potential range expansion into southeastern Arizona. New Mexico and Chihuahua contained environmental characteristics primarily limited to the male niche and thus may be areas into which males occasionally disperse.

Arizona;Chihuahua;New Mexico;Sonora;Texas↗
USGS2019

Implications of climate scenarios for Badlands National Park resource management

Badlands National Park (BADL) hosts a myriad of natural and cultural resources, including bison and black-footed ferrets, the mixed grass prairie they live in, 37-75 million-year-old fossils, and historic buildings, trails, and roads. All are sensitive to climate, but anticipating precisely how each will be affected by climate change is difficult. In the face of this challenge, park resource managers must nevertheless make forward-looking decisions and take action to meet resource management goals. Fortunately, tools exist to identify strategies and actions likely to succeed under a range of potential future climate conditions. Two such tools—qualitative scenario planning and quantitative ecological simulation modeling—were used to anticipate management challenges and identify solutions for BADL and adjacent federal and tribal lands in the coming decades (through 2050). This brief summarizes and synthesizes results of this work. Although the brief focuses on BADL, it also includes several key insights gained from examining management approaches on adjacent lands.

South Dakota↗
USGS2023

BatTool: Projecting bat populations facing multiple stressors using a demographic model

Bats provide ecologically and agriculturally important ecosystem services but are currently experiencing population declines caused by multiple environmental stressors, including mortality from white-nose syndrome and wind energy development. Analyses of the current and future health and viability of these species may support conservation management decision making. Demographic modeling provides a quantitative tool for decision makers and conservation managers to make more informed decisions, but widespread adoption of these tools can be limited because of the complexity of the mathematical, statistical, and computational components involved in implementing these models. In this work, we provide an exposition of the BatTool R package, detailing the primary components of the matrix projection model, a publicly accessible graphical user interface ( https://rconnect.usgs.gov/battool ) facilitating user-defined scenario analyses, and its intended uses and limitations (Wiens et al., US Geol Surv Data Release 2022; Wiens et al., US Geol Surv Softw Release 2022). We present a case study involving wind energy permitting, weighing the effects of potential mortality caused by a hypothetical wind energy facility on the projected abundance of four imperiled bat species in the Midwestern United States.

Methods in Ecology and Evolution↗
USGS2019

A General Lake Model (GLM 3.0) for linking with high-frequency sensor data from the Global Lake Ecological Observatory Network (GLEON)

The General Lake Model (GLM) is a one-dimensional open-source code designed to simulate the hydrodynamics of lakes, reservoirs, and wetlands. GLM was developed to support the science needs of the Global Lake Ecological Observatory Network (GLEON), a network of researchers using sensors to understand lake functioning and address questions about how lakes around the world respond to climate and land use change. The scale and diversity of lake types, locations, and sizes, and the expanding observational datasets created the need for a robust community model of lake dynamics with sufficient flexibility to accommodate a range of scientific and management questions relevant to the GLEON community. This paper summarizes the scientific basis and numerical implementation of the model algorithms, including details of sub-models that simulate surface heat exchange and ice cover dynamics, vertical mixing, and inflow–outflow dynamics. We demonstrate the suitability of the model for different lake types that vary substantially in their morphology, hydrology, and climatic conditions. GLM supports a dynamic coupling with biogeochemical and ecological modelling libraries for integrated simulations of water quality and ecosystem health, and options for integration with other environmental models are outlined. Finally, we discuss utilities for the analysis of model outputs and uncertainty assessments, model operation within a distributed cloud-computing environment, and as a tool to support the learning of network participants.

Geoscientific Model Development↗
USGS1995

Population modeling and its role in toxicological studies

A model could be defined as any abstraction from reality that is used to provide some insight into the real system. In this discussion, we will use a more specific definition that a model is a set of rules or assumptions, expressed as mathematical equations, that describe how animals survive and reproduce, including the external factors that affect these characteristics. A model simplifies a system, retaining essential components while eliminating parts that are not of interest. ecology has a rich history of using models to gain insight into populations, often borrowing both model structures and analysis methods from demographers and engineers. Much of the development of the models has been a consequence of mathematicians and physicists seeing simple analogies between their models and patterns in natural systems. Consequently, one major application of ecological modeling has been to emphasize the analysis of dynamics of often complex models to provide insight into theoretical aspects of ecology. 1

Book chapter↗
USGS2026

A Bayesian hierarchical modeling approach for species diversity in ecology

Species diversity is the foundation of many ecological disciplines. This metric is often approximated using species richness and evenness, even though actual richness likely exceeds observations due to imperfect sampling methods. Estimating the “true” species richness, which includes identifying the number of missing species, has intrigued ecologists for decades. We adopted a parametric model that appeared in Fisher et al. (1943), which models the numbers of individuals from different species as random samples from a negative binomial distribution, and developed a Bayesian computational approach to directly estimate the distribution model parameters. The model parameters represent species abundance and evenness, and can be used to derive species richness. We evaluated our parametric approach using (1) a simulation study and (2) three historical data sets. Furthermore, we illustrated the hierarchical modeling approach to combine data from multiple parallel studies using a biannual fishery survey data set. Our parametric model formulation is computationally efficient, and the hierarchical structure facilitates embedding diversity estimation into broader application, such as assessing spatial and temporal trends in species diversity associated with environmental stressors. Additionally, because the two parameters of the negative binomial distribution model represent species abundance and evenness of a community, this parametric approach facilitates a deeper understanding of the ecological systems under study. The negative binomial distribution model works with a wide range of species frequency distribution types. As a result, our emphasis on a parametric model can help us characterize the structure of an ecosystem and provide a greater depth of ecologically meaningful information.

Ecological Informatics↗
USGS1995

Hydrologic modeling as a predictive basis for ecological restoration of salt marshes

Roads, bridges, causeways, impoundments, and dikes in the coastal zone often restrict tidal flow to salt marsh ecosystems. A dike with tide control structures, located at the mouth of the Herring River salt marsh estuarine system (Wellfleet, Massachusetts) since 1908, has effectively restricted tidal exchange, causing changes in marsh vegetation composition, degraded water quality, and reduced abundance of fish and macroinvertebrate communities. Restoration of this estuary by reintroduction of tidal exchange is a feasible management alternative. However, restoration efforts must proceed with caution as residential dwellings and a golf course are located immediately adjacent to and in places within the tidal wetland. A numerical model was developed to predict tide height levels for numerous alternative openings through the Herring River dike. Given these model predictions and knowledge of elevations of flood-prone areas, it becomes possible to make responsible decisions regarding restoration. Moreover, tidal flooding elevations relative to the wetland surface must be known to predict optimum conditions for ecological recovery. The tide height model has a universal role, as demonstrated by successful application at a nearby salt marsh restoration site in Provincetown, Massachusetts. Salt marsh restoration is a valuable management tool toward maintaining and enhancing coastal zone habitat diversity. The tide height model presented in this paper will enable both scientists and resource professionals to assign a degree of predictability when designing salt marsh restoration programs.

Environmental Management↗
USGS2025

Temporal and spatial equivalence in demographic responses of emperor penguins (Aptenodytes forsteri) to environmental change

1. Population ecology and biogeography applications often necessitate the transfer of models across spatial and/or temporal dimensions to make predictions outside the bounds of the data used for model fitting. However, ecological data are often spatiotemporally unbalanced such that the spatial or the temporal dimension tends to contain more data than the other. This unbalance frequently leads model transfers to become substitutions, which are predictions to a different dimension than the predictive model was built on. Despite the prevalence of substitutions in ecology, studies validating their performance and their underlying assumptions are scarce. 2. Here, we present a successful case study demonstrating both space-for-time and time-for-space substitutions using emperor penguins (Aptenodytes forsteri) as the focal species. Using abundance-based species distribution models (aSDM) of adult emperor penguins in attendance during spring across 50 colonies, we predict long-term annual fluctuations in fledgling abundance and breeding success at a single colony, Pointe Géologie. Subsequently, we construct statistical models from time series of extended counts on Pointe Géologie to predict average fledgling abundance across 50 colonies. 3. Our analysis reveals that distance to nearest open water (NOW) exhibits the strongest association with both temporal and spatial data. aSDM’s space-for-time substitution performance, as measured by Pearson correlation coefficient was 0.63 and 0.56 when predicting breeding success and fledgling abundance time series, respectively. Linear regression of fledgling abundance on NOW yields similar time-for-space substitution performance when predicting abundance distribution of emperor penguin colonies with a correlation coefficient of 0.58. 4. We posit that such space-time equivalence arises because: 1) emperor penguins colonies conform to their existing fundamental niche; 2) there is not yet any environmental novelty when comparing the spatial vs temporal variation of distance to nearest open water; and 3) models of more specific components of life histories, such as fledgling abundance, rather than occurrence or total population abundance, are more transferable. Identifying these conditions empirically can enhance the qualitative validation of substitutions in cases where direct validation data are lacking.

Journal of Animal Ecology↗
USGS2019

Nonlinear reaction–diffusion process models improve inference for population dynamics

Partial differential equations (PDEs) are a useful tool for modeling spatiotemporal dynamics of ecological processes. However, as an ecological process evolves, we need statistical models that can adapt to changing dynamics as new data are collected. We developed a model that combines an ecological diffusion equation and logistic growth to characterize colonization processes of a population that establishes long-term equilibrium over a heterogeneous environment. We also developed a homogenization strategy to statistically upscale the PDE for faster computation and adopted a hierarchical framework to accommodate multiple data sources collected at different spatial scales. We highlighted the advantages of using a logistic reaction component instead of a Malthusian component when population growth demonstrates asymptotic behavior. As a case study, we demonstrated that our model improves spatiotemporal abundance forecasts of sea otters in Glacier Bay, Alaska. Furthermore, we predicted spatially varying local equilibrium abundances as a result of environmentally driven diffusion and density-regulated growth. Integrating equilibrium abundances over the study area in our application enabled us to infer the overall carrying capacity of sea otters in Glacier Bay, Alaska.

Alaska↗
USGS2014

Modelling non-Euclidean movement and landscape connectivity in highly structured ecological networks

Movement is influenced by landscape structure, configuration and geometry, but measuring distance as perceived by animals poses technical and logistical challenges. Instead, movement is typically measured using Euclidean distance, irrespective of location or landscape structure, or is based on arbitrary cost surfaces. A recently proposed extension of spatial capture-recapture (SCR) models resolves this issue using spatial encounter histories of individuals to calculate least-cost paths (ecological distance: Ecology , 94 , 2013, 287) thereby relaxing the Euclidean assumption. We evaluate the consequences of not accounting for movement heterogeneity when estimating abundance in highly structured landscapes, and demonstrate the value of this approach for estimating biologically realistic space-use patterns and landscape connectivity. We simulated SCR data in a riparian habitat network, using the ecological distance model under a range of scenarios where space-use in and around the landscape was increasingly associated with water (i.e. increasingly less Euclidean). To assess the influence of miscalculating distance on estimates of population size, we compared the results from the ecological and Euclidean distance based models. We then demonstrate that the ecological distance model can be used to estimate home range geometry when space use is not symmetrical. Finally, we provide a method for calculating landscape connectivity based on modelled species-landscape interactions generated from capture-recapture data. Using ecological distance always produced unbiased estimates of abundance. Explicitly modelling the strength of the species-landscape interaction provided a direct measure of landscape connectivity and better characterised true home range geometry. Abundance under the Euclidean distance model was increasingly (negatively) biased as space use was more strongly associated with water and, because home ranges are assumed to be symmetrical, produced poor characterisations of home range geometry and no information about landscape connectivity. The ecological distance SCR model uses spatially indexed capture-recapture data to estimate how activity patterns are influenced by landscape structure. As well as reducing bias in estimates of abundance, this approach provides biologically realistic representations of home range geometry, and direct information about species-landscape interactions. The incorporation of both structural (landscape) and functional (movement) components of connectivity provides a direct measure of species-specific landscape connectivity.

Methods in Ecology and Evolution↗
USGS2020

Ecological risk assessment of environmental stress and bioactive chemicals to riverine fish populations: An individual-based model of smallmouth bass Micropterus dolomieu

Ecological risk assessments play an important role in environmental management and decision-making. Although empirical measurements of the effects of habitat changes and chemical exposure are often made at molecular and individual levels, environmental decision-making often requires the quantification of management-relevant, population-level outcomes. In this study, we generalized a modeling framework to evaluate population-level ecological risk of environmental stress and bioactive chemicals. The modeling framework includes (1) a biological model module that incorporates complex and interacting biological and ecological processes, and environmental stochasticity, (2) an effect module that links the impacts of environmental changes and chemical exposure to individual characteristics, and (3) a population module that makes decisions on the choice of population-level properties to best capture the effects and thus to track in the model based on the target species and the research and management interest. This framework is a 3-module procedure that provides an alternative way for researchers to organize, present and communicate the risk assessment modeling studies. To demonstrate this framework, we used a socioeconomically important riverine fish species, smallmouth bass Micropterus dolomieu , as the model species. We developed an individual-based model as the biological model module. We evaluated the impacts of changing water temperature and flow regimes, and the impacts of exposure to estrogenic endocrine disrupting compounds (EEDC) on smallmouth bass populations in the Chesapeake Bay Watershed, USA. Warm summer water temperatures and year-round high flows had the most severe impacts on the smallmouth bass population. An increase in exposure level to EEDC, both year-round and in summer months, substantially reduced population size, spawner and recruit abundance, and the proportion of quality-length individuals. Acute exposure to EEDC was more detrimental to the population than chronic exposure. Acute exposure during spawning season had the most severe impacts. This modeling framework can be extended to other species, environmental factors and chemicals, and can be used to inform management and conservation decisions.

Delaware, Maryland, New York, Pennsylvania, Virgin↗
USGS2015

Model averaging and muddled multimodel inferences

Three flawed practices associated with model averaging coefficients for predictor variables in regression models commonly occur when making multimodel inferences in analyses of ecological data. Model-averaged regression coefficients based on Akaike information criterion (AIC) weights have been recommended for addressing model uncertainty but they are not valid, interpretable estimates of partial effects for individual predictors when there is multicollinearity among the predictor variables. Multicollinearity implies that the scaling of units in the denominators of the regression coefficients may change across models such that neither the parameters nor their estimates have common scales, therefore averaging them makes no sense. The associated sums of AIC model weights recommended to assess relative importance of individual predictors are really a measure of relative importance of models, with little information about contributions by individual predictors compared to other measures of relative importance based on effects size or variance reduction. Sometimes the model-averaged regression coefficients for predictor variables are incorrectly used to make model-averaged predictions of the response variable when the models are not linear in the parameters. I demonstrate the issues with the first two practices using the college grade point average example extensively analyzed by Burnham and Anderson. I show how partial standard deviations of the predictor variables can be used to detect changing scales of their estimates with multicollinearity. Standardizing estimates based on partial standard deviations for their variables can be used to make the scaling of the estimates commensurate across models, a necessary but not sufficient condition for model averaging of the estimates to be sensible. A unimodal distribution of estimates and valid interpretation of individual parameters are additional requisite conditions. The standardized estimates or equivalently the t statistics on unstandardized estimates also can be used to provide more informative measures of relative importance than sums of AIC weights. Finally, I illustrate how seriously compromised statistical interpretations and predictions can be for all three of these flawed practices by critiquing their use in a recent species distribution modeling technique developed for predicting Greater Sage-Grouse ( Centrocercus urophasianus ) distribution in Colorado, USA. These model averaging issues are common in other ecological literature and ought to be discontinued if we are to make effective scientific contributions to ecological knowledge and conservation of natural resources.

Ecology↗