Geology Reports⌕ Search

SEARCH · Geology Reports

Results for “Ecological Modeling”

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 649 records · Page 36Linked to original sources
USGS2014

The effects of harvest on waterfowl populations

Change in the size of populations over space and time is, arguably, the motivation for much of pure and applied ecological research. The fundamental model for the dynamics of any population is straightforward: the net change in the abundance is the simple difference between the number of individuals entering the population and the number leaving the population, either or both of which may change in response to factors intrinsic and extrinsic to the population. While harvest of individuals from a population constitutes a clear extrinsic source of removal of individuals, the response of populations to harvest is frequently complex, reflecting an interaction of harvest with one or more population processes. Here we consider the role of these interactions, and factors influencing them, on the effective harvest management of waterfowl populations. We review historical ideas concerning harvest and discuss the relationship(s) between waterfowl life histories and the development and application of population models to inform harvest management. The influence of population structure (age, spatial) on derivation of optimal harvest strategies (with and without explicit consideration of various sources of uncertainty) is considered. In addition to population structure, we discuss how the optimal harvest strategy may be influenced by: 1) patterns of density-dependence in one or more vital rates, and 2) heterogeneity in vital rates among individuals within an age-sex-size class. Although derivation of the optimal harvest strategy for simple population models (with or without structure) is generally straightforward, there are several potential difficulties in application. In particular, uncertainty concerning the population structure at the time of harvest, and the ability to regulate the structure of the harvest itself, are significant complications. We therefore review the evidence of effects of harvest on waterfowl populations. Some of this evidence has focussed on correspondence of data with more phenomenological models and other evidence relates to specific mechanisms, including densitydependence and heterogeneity. An important part of this evidence is found in the evolution of model weights under various adaptive harvest management programmes of the U.S. Fish and Wildlife Service for North American waterfowl. Overall, there is substantial uncertainty about system dynamics, about the impacts of potential management and conservation decisions on those dynamics, and how to optimise management decisions in the presence of such uncertainties. Such relationships are unlikely to be stationary over space or time, and selective harvest of some individuals can potentially alter life history allocation of resources over time – both of which will potentially influence optimal harvest strategies. These sources of variation and uncertainty argue for the use of adaptive approaches to waterfowl harvest management.

Wildfowl↗
USGS2020

Understanding and documenting the scientific basis of selenium ecological protection in support of site-specific guidelines development for Lake Koocanusa, Montana, U.S.A., and British Columbia, Canada

Modeling of ecosystems is a part of the U.S. Environmental Protection Agency’s protocol for developing site-specific selenium guidelines for protection of aquatic life. Selenium as an environmental contaminant is known to bioaccumulate and cause reproductive effects in fish and wildlife. Here we apply a modeling methodology—ecosystem-scale selenium modeling—to understand and document the scientific basis for predicting and validating ecological protection for Lake Koocanusa, a transboundary reservoir between Montana and British Columbia. A comprehensive set of site-specific data compiled from public databases (Federal, State, and Provincial) and reports by Teck Coal Ltd., is available in a companion U.S. Geological Survey data release. The tissue guideline used within modeling here to assess protection is the U.S. Environmental Protection Agency’s national selenium guideline for whole-body fish (dry weight); however, other numeric values for a whole-body guideline or other tissue types may be assumed if applicable tissue-to-tissue conversion factors are available. We consider the report assembled here as a working document that presents a model that can effectively address and structure the needs of (1) scientific understanding in representing the lake’s ecosystem and selenium biodynamics and (2) policy and management development during a decision-making process, but it is open to modification and updating as more ecologically detailed data become available. The approach brings together the main concerns involved in selenium toxicity: likelihood of high exposure, inherent species sensitivity, and close connectivity of ecosystem characteristics and behavioral ecology of predators. Detailed site-specific modeling equations are provided to document the linked factors that determine the responses of ecosystems to selenium. A series of scenarios quantifies the implications of choices of site-specific variables including food-web species, bioavailability of particulate material, and partitioning between the dissolved and particulate phases at the base of food webs. A gradient mapping tool applied to Lake Koocanusa provides a precedent for ecosystem-scale modeling of lakes by recognizing the importance of lake strata and hydrodynamics as components of modeling. Data requirements for ecosystem modeling, including ecological and hydrological process information fundamental to the dietary biodynamics of selenium in site-specific food webs, were assessed as a precursor to model validation for Lake Koocanusa. Understanding these relationships is necessary to connect modeling outcomes to reproductive effects and establish boundaries, in the case of Lake Koocanusa, for the influences of dam operation, fish-community viability, and its Clean Water Act impaired 303(d)-listing status on ecosystem function. We find that an assemblage of conditions affects the representation of Lake Koocanusa’s ecosystem within modeling scenarios but that the constructed gradient maps, mechanistic model, and associated bioaccumulation potentials portray and quantify the variables that are determinative to protection of predator species. Ecological and hydrological sorting of compiled individual data points on a site- and species-specific basis helps identify and address model uncertainties. Sources of uncertainty include (1) the scarcity of data for some environmental media compartments across time and locations, (2) the complexity of hydrodynamic conditions that can lead to seasonal ecological disconnects such as in selenium partitioning from water into particulates, and (3) the functional status of Lake Koocanusa’s ecosystem because of cumulative effects of various environmental stresses (for example, fish-community changes, flow regime changes, parasites, gonadal dysfunction, and increasing mining input-selenium concentrations since 1984). To this last point, it is important to determine where Lake Koocanusa is in an impairment-restoration cycle so as not to base protection on survivor bias, the maintenance of a currently degraded ecosystem, or normalized toxicity. In a broader context, one of the overall consequences of revised selenium regulations is that their derivation is now dependent on being able to define and understand the status of the ecosystem on which protection is based.

Montana, British Columbia↗
USGS1994

An integrated model of human-wildlife interdependence

This paper attempts to integrate wildlife-related ecologic and economic variables into an econometric model. The model reveals empirical evidence of the presumed interdependence of human-wildlife and the holistic nature of humanity's relationship to the ecosystem. Human use of biologic resources varies not only with income, education, and population, but also with sustainability of humankind's action relative to the quality and quantity of the supporting ecological base.

Ecological Economics↗
USGS2006

Classification tree and minimum-volume ellipsoid analyses of the distribution of ponderosa pine in the western USA

Aim? Ponderosa pine (Pinus ponderosa Douglas ex Lawson & C. Lawson) is an economically and ecologically important conifer that has a wide geographic range in the western USA, but is mostly absent from the geographic centre of its distribution - the Great Basin and adjoining mountain ranges. Much of its modern range was achieved by migration of geographically distinct Sierra Nevada (P. ponderosa var. ponderosa) and Rocky Mountain (P. ponderosa var. scopulorum) varieties in the last 10,000 years. Previous research has confirmed genetic differences between the two varieties, and measurable genetic exchange occurs where their ranges now overlap in western Montana. A variety of approaches in bioclimatic modelling is required to explore the ecological differences between these varieties and their implications for historical biogeography and impending changes in western landscapes. Location? Western USA. Methods? We used a classification tree analysis and a minimum-volume ellipsoid as models to explain the broad patterns of distribution of ponderosa pine in modern environments using climatic and edaphic variables. Most biogeographical modelling assumes that the target group represents a single, ecologically uniform taxonomic population. Classification tree analysis does not require this assumption because it allows the creation of pathways that predict multiple positive and negative outcomes. Thus, classification tree analysis can be used to test the ecological uniformity of the species. In addition, a multidimensional ellipsoid was constructed to describe the niche of each variety of ponderosa pine, and distances from the niche were calculated and mapped on a 4-km grid for each ecological variable. Results? The resulting classification tree identified three dominant pathways predicting ponderosa pine presence. Two of these three pathways correspond roughly to the distribution of var. ponderosa, and the third pathway generally corresponds to the distribution of var. scopulorum. The classification tree and minimum-volume ellipsoid model show that both varieties have very similar temperature limitations, although var. ponderosa is more limited by the temperature extremes of the continental interior. The precipitation limitations of the two varieties are seasonally different, with var. ponderosa requiring significant winter moisture and var. scopulorum requiring significant summer moisture. Great Basin mountain ranges are too cold at higher elevations to support either variety of ponderosa pine, and at lower elevations are too dry in summer for var. scopulorum and too dry in winter for var. ponderosa. Main conclusions? The classification tree analysis indicates that var. ponderosa is ecologically as well as genetically distinct from var. scopulorum. Ecological differences may maintain genetic separation in spite of a limited zone of introgression between the two varieties in western Montana. Two hypotheses about past and future movements of ponderosa pine emerge from our analyses. The first hypothesis is that, during the last glacial period, colder and/or drier summers truncated most of the range of var. scopulorum in the central Rockies, but had less dramatic effects on the more maritime and winter-wet distribution of var. ponderosa. The second hypothesis is that, all other factors held constant, increasing summer temperatures in the future should produce changes in the distribution of var. scopulorum that are likely to involve range expansions in the central Rockies with the warming of mountain ranges currently too cold but sufficiently wet in summer for var. scopulorum. Finally, our results underscore the growing need to focus on genotypes in biogeographical modelling and ecological forecasting.

Journal of Biogeography↗
USGS2024

Molecular sexing of birds using quantitative PCR (qPCR) of sex-linked genes and logistic regression models

The ability to sex individuals is an important component of many behavioural and ecological investigations and provides information for demographic models used in conservation and species management. However, many birds are difficult to sex using morphological characters or traditional molecular sexing methods. In this study, we developed probabilistic models for sexing birds using quantitative PCR (qPCR) data. First, we quantified distributions of gene copy numbers at a set of six sex-linked genes, including the sex-determining gene DMRT1 , for individuals across 17 species and seven orders of birds ( n = 150). Using these data, we built predictive logistic models for sex identification and tested their performance with independent samples from 51 species and 13 orders ( n = 209). Models using the two loci most highly correlated with sex had greater accuracy than models using the full set of sex-linked loci, across all taxonomic levels of analysis. Sex identification was highly accurate when individuals to be assigned were of species used in model building. Our analytical approach was widely applicable across diverse neognath bird lineages spanning millions of years of evolutionary divergence. Unlike previous methods, our probabilistic framework incorporates uncertainty around qPCR measurements as well as biological variation within species into decision-making rules. We anticipate that this method will be useful for sexing birds, including those of high conservation concern and/or subsistence value, that have proven difficult to sex using traditional approaches. Additionally, the general analytical framework presented in this paper may also be applicable to other organisms with sex chromosomes.

Molecular Ecology Resources↗
USGS2016

Maxent modeling for predicting potential distribution of goitered gazelle in central Iran: the effect of extent and grain size on performance of the model

The spatial scale of environmental layers is an important factor to consider in developing an understanding of ecological processes. This study employed Maxent modeling to investigate the geographic distribution of goitered gazelle, Gazella subgutturosa (Güldenstädt, 1780), in central Iran using uncorrelated variables at a spatial resolution of 250 m. We used spatial downscaling to downscale WorldClim data to 250-m resolution. We evaluated the sensitivity of the model to different grain and extent sizes from 250 m to 3 km. We compared the performance of the model at different scales using suitability indexes (AUC) and predicted habitat areas. Two models performed with AUC values higher than random (AUCun = 0.957, AUCpu = 0.953). The distribution of potential habitats at 250- m grid size was strongly influenced by bioclimatic data, vegetation type and density, and elevation. There were few spatial divergences between uncorrelated and pruned models. The mean AUC across eight different spatial scales ranged from 0.936 to 0.959. There was a significant negative correlation between grain size and AUC (R2 = 0.57). An increase in grain size increased the predicted habitat area. The extent size and AUC showed a positive correlation (R2 = 0.18). Predicted suitability habitat also decreased as extent size increased (R2 = 0.49). Spatial congruence AUC fluctuated within a small range and the maximum difference occurred between models of 1 × 1 and 2.5 × 2.5 km. These results showed that an increase in extent size is more accurate than an increase in grain size, and the maximum accuracy for predicting distribution of goitered gazelle in Iran was obtained if the grain size and extent size were 750 m.

Turkish Journal of Zoology↗
USGS2004

Combining band recovery data and Pollock's robust design to model temporary and permanent emigration

Capture-recapture models are widely used to estimate demographic parameters of marked populations. Recently, this statistical theory has been extended to modeling dispersal of open populations. Multistate models can be used to estimate movement probabilities among subdivided populations if multiple sites are sampled. Frequently, however, sampling is limited to a single site, Models described by Burnham (1993, in Marked Individuals in the Study of Bird Populations , 199–213), which combined open population capture-recapture and band-recovery models, can be used to estimate permanent emigration when sampling is limited to a single population. Similarly, Kendall, Nichols, and Hines (1997, Ecology 51 , 563–578) developed models to estimate temporary emigration under Pollock's (1982, Journal of Wildlife Management 46 , 757–760) robust design. We describe a likelihood-based approach to simultaneously estimate temporary and permanent emigration when sampling is limited to a single population. We use a sampling design that combines the robust design and recoveries of individuals obtained immediately following each sampling period. We present a general form for our model where temporary emigration is a first-order Markov process, and we discuss more restrictive models. We illustrate these models with analysis of data on marked Canvasback ducks. Our analysis indicates that probability of permanent emigration for adult female Canvasbacks was 0.193 ( ) and that birds that were present at the study area in year i — 1 had a higher probability of presence in year i than birds that were not present in year i — 1.

Biometrics↗
USGS2011

Development of a bioenergetics model for age-0 American Shad

Bioenergetics modeling can be used as a tool to investigate the impact of non-native age-0 American shad (Alosa sapidissima) on reservoir and estuary food webs. The model can increase our understanding of how these fish influence lower trophic levels as well as predatory fish populations that feed on juvenile salmonids. Bioenergetics modeling can be used to investigate ecological processes, evaluate alternative research hypotheses, provide decision support, and quantitative prediction. Bioenergetics modeling has proven to be extremely useful in fisheries research (Ney et al. 1993,Chips and Wahl 2008, Petersen et al. 2008). If growth and diet parameters are known, the bioenergetics model can be used to quantify the relative amount of zooplankton or insects consumed by age-0 American shad. When linked with spatial and temporal information on fish abundance, model output can guide inferential hypothesis development to demonstrate where the greatest impacts of age-0 American shad might occur.

Report↗
USGS2025

Assessing survey design for long-term population trend detection in piping plovers

Determining appropriate spatio-temporal scales for monitoring migratory shorebirds is challenging. Effective surveys must detect population trends without excessive or insufficient sampling, yet many programs lack formal evaluations of survey effectiveness. Using data from 2012 to 2019 on Louisiana’s barrier islands (Whiskey, west Raccoon, east Raccoon, and Trinity), we assessed how spatial and temporal scales influence population trend inference for piping plovers ( Charadrius melodus ). Point count data were aggregated to grid sizes from 50 to 200 m and analyzed using Bayesian dynamic occupancy models. We found occupancy and colonization estimates varied by spatial resolution, with space–time autocorrelation common across scales. Smaller islands (east and west Raccoon) yielded higher trend detection power due to better detectability, while larger islands (Trinity and Whiskey) showed lower power. Detectability, more than sampling frequency, drove trend inference. Models incorporating spatial autocorrelation outperformed traditional Frequentist approaches but showed poorer fit at coarser scales. These findings underscore how matching analytical scale to ecological processes and selecting appropriate models can influence predictions. Power analysis revealed that increasing survey frequency may improve inference, especially in low-detectability areas. Overall, our study highlights how careful scale selection, model diagnostics, and survey design can enhance monitoring efficiency and support long-term conservation of migratory shorebirds.

Louisiana↗
USGS2015

Data Collection and Simulation of Ecological Habitat and Recreational Habitat in the Shenandoah River, Virginia

This report presents updates to methods, describes additional data collected, documents modeling results, and discusses implications from an updated habitat-flow model that can be used to predict ecological habitat for fish and recreational habitat for canoeing on the main stem Shenandoah River in Virginia. Given a 76-percent increase in population predictions for 2040 over 1995 records, increased water-withdrawal scenarios were evaluated to determine the effects on habitat and recreation in the Shenandoah River. Projected water demands for 2040 vary by watershed: the North Fork Shenandoah River shows a 55.9-percent increase, the South Fork Shenandoah River shows a 46.5-percent increase, and the main stem Shenandoah River shows a 52-percent increase; most localities are projected to approach the total permitted surface-water and groundwater withdrawals values by 2040, and a few localities are projected to exceed these values. The habitat model used for this study evaluates the suitability of ecological habitat, represented by fish, and recreational habitat, represented by canoeing, based on depth, velocity, and substrate conditions, which are weighted for the physical habitat types (riffles, runs, or pools) present within a stretch of river. Weighted usable-habitat area in the Lockes Mill reach was maximized for adult smallmouth bass and sub-adult smallmouth bass ( Micropterus dolomieu ) and river chub ( Nocomis micropogon ) when streamflows were equal to median flow (900 cubic feet per second) for summer months. Ecological maximum weighted usable-habitat areas for smaller fish, such as spotfin or satinfin shiner ( Cyprinella spp.), margined madtom ( Noturus insignis ), and juvenile redbreast sunfish ( Lepomis auritus ) occurred with 10 th percentile flows (482 cubic feet per second) and lower. Recreational weighted usable-habitat areas for canoeing were maximized when streamflows were above the 75 th percentile (1,410 cubic feet per second). During historic droughts, streamflows were less than the 10 th percentile, and adult smallmouth bass and sub-adult smallmouth bass habitat was below normal for the majority of days during at least 2 months of the summer. When streamflows were less than the lowest 7-day average in a 10-year period, or 7Q10 flow (357 cubic feet per second), margined madtom, river chub, and sub-adult redbreast sunfish habitat areas were below normal as well. Streamflows that limit most fish species habit availability range from 300 to 500 cubic feet per second. For the drought years simulated, flows that were equal to or less than the 10 th percentile for summer months did not provide adequate depth for canoe passage through riffle habitats. A modeling limitation for higher flows than those studied during development of the habitat-suitability criteria is that modeled habitat availability will decrease as flows increase. Time-series analyses were used to investigate changes in habitat availability with increased water withdrawals of 10, 20, and almost 50 percent (48.6 percent) up to the 2040 amounts projected by local water supply plans. Adult and sub-adult smallmouth bass frequently had habitat availability outside the normal range for habitat conditions during drought years, yet 10- or 20-percent increases in withdrawals did not contribute to a large reduction in habitat. When withdrawals were increased by 50 percent, there was an additional decrease in habitat. During 2002 drought scenarios, reduced habitat availability for sub-adult redbreast sunfish or river chub was only slightly evident with 50-percent increased withdrawal scenarios. Recreational habitat represented by canoeing decreased lower than normal during the 2002 drought. For a recent normal year, like 2012, increased water-withdrawal scenarios did not affect habitat availability for fish such as adult and sub-adult smallmouth bass, sub-adult redbreast sunfish, or river chub. Canoeing habitat availability was within the normal range most of 2012, and increased water-withdrawal scenarios showed almost no affect. For both ecological fish habitat and recreational canoeing habitat, the antecedent conditions (habitat within normal range of habitat or below normal) appear to govern whether additional water withdrawals will affect habitat availability. As human populations and water demands increase, many of the ecological or recreational stresses may be lessened by managing the timing of water withdrawals from the system.

Virginia↗
USGS2013

Relaxing the closure assumption in single-season occupancy models: staggered arrival and departure times

Occupancy statistical models that account for imperfect detection have proved very useful in several areas of ecology, including species distribution and spatial dynamics, disease ecology, and ecological responses to climate change. These models are based on the collection of multiple samples at each of a number of sites within a given season, during which it is assumed the species is either absent or present and available for detection while each sample is taken. However, for some species, individuals are only present or available for detection seasonally. We present a statistical model that relaxes the closure assumption within a season by permitting staggered entry and exit times for the species of interest at each site. Based on simulation, our open model eliminates bias in occupancy estimators and in some cases increases precision. The power to detect the violation of closure is high if detection probability is reasonably high. In addition to providing more robust estimation of occupancy, this model permits comparison of phenology across sites, species, or years, by modeling variation in arrival or departure probabilities. In a comparison of four species of amphibians in Maryland we found that two toad species arrived at breeding sites later in the season than a salamander and frog species, and departed from sites earlier.

Ecology↗
USGS2010

An Adaptive Management Approach for Summer Water Level Reductions on the Upper Mississippi River System

The primary purpose of this report is to provide an adaptive management approach for learning more about summer water level reductions (drawdowns) as a management tool, including where and how drawdowns can be applied most effectively within the Upper Mississippi River System. The report reviews previous drawdowns conducted within the system and provides specific recommendations for learning more about the lesser known effects of drawdowns and how the outcomes can be influenced by different implementation strategies and local conditions. The knowledge gained can be used by managers to determine how best to implement drawdowns in different parts of the UMRS to help achieve management goals. The information and recommendations contained in the report are derived from results of previous drawdown projects, insights from regional disciplinary experts, and the experience of the authors in experimental design, modeling, and monitoring. Modeling is a critical part of adaptive management and can involve conceptual models, simulation models, and empirical models. In this report we present conceptual models that express current understanding regarding functioning of the UMRS as related to drawdowns and highlight interactions among key ecological components of the system. The models were developed within the constraints of drawdown timing, magnitude (depth), and spatial differences in effects (longitudinal and lateral) with attention to ecological processes affected by drawdowns. With input from regional experts we focused on the responses of vegetation, fish, mussels, other invertebrates, and birds. The conceptual models reflect current understanding about relations and interactions among system components, the expected strength of those interactions, potential responses of system components to drawdowns, likelihood of the response occurring, and key uncertainties that limit our ability to make accurate predictions of effects (Table 1, Fig. 4-10). Based on this current understanding, the main questions still associated with drawdowns include (1) the effects of frequency of drawdowns (from once every few years to multiple years in succession); (2) timing of the beginning of drawdowns (follow the descending arm of the flood pulse versus always beginning in early summer); (3) long-term benefits (greater than 5-6 years), especially as compared to known short-term loses (e.g., mortality of mussels in exposed areas, loss of submersed vegetation in exposed areas, cost of advanced dredging); and (4) the effects in northern (above pool 14) versus southern pools (pool 14 and below, and the Illinois River). An adaptive management design should address these questions to reduce uncertainty in predictions of drawdown effects and help determine if different implementation strategies are needed in different parts of the system. Given that drawdowns will continue to be used as a management tool on the UMRS, we suggest that some drawdowns be conducted in an adaptive management context that helps meet management objectives, but also provides efficient learning about the questions listed above. We propose two different, but interrelated, experimental designs to address these questions. Both designs call for conducting multiple drawdowns in multiple pools (2-4 pools) to allow direct comparison of results and produce rapid learning. However, the report does not provide a detailed scope of work for carrying out the designs. If managers choose to implement one of the experimental designs, specifics of choosing appropriate pools and developing a monitoring plan will need to be determined through collaboration among managers, researchers, and statisticians. We suggest characteristics to consider in selecting treatment and reference pools (study sites) and also provide guidance for developing a monitoring plan. Some aspects of these two designs could be implemented individually, but by implementing individual elements, direct comparisons of some design features

Illinois;Iowa;Minnesota;Missouri;Wisconsin↗
USGS2012

Assessing the vulnerability of human and biological communities to changing ecosystem services using a GIS-based multi-criteria decision support tool

In this paper we describe an application of a GIS-based multi-criteria decision support web tool that models and evaluates relative changes in ecosystem services to policy and land management decisions. The Santa Cruz Watershed Ecosystem Portfolio (SCWEPM) was designed to provide credible forecasts of responses to ecosystem drivers and stressors and to illustrate the role of land use decisions on spatial and temporal distributions of ecosystem services within a binational (U.S. and Mexico) watershed. We present two SCWEPM sub-models that when analyzed together address bidirectional relationships between social and ecological vulnerability and ecosystem services. The first model employs the Modified Socio-Environmental Vulnerability Index (M-SEVI), which assesses community vulnerability using information from U.S. and Mexico censuses on education, access to resources, migratory status, housing situation, and number of dependents. The second, relating land cover change to biodiversity (provisioning services), models changes in the distribution of terrestrial vertebrate habitat based on multitemporal vegetation and land cover maps, wildlife habitat relationships, and changes in land use/land cover patterns. When assessed concurrently, the models exposed some unexpected relationships between vulnerable communities and ecosystem services provisioning. For instance, the most species-rich habitat type in the watershed, Desert Riparian Forest, increased over time in areas occupied by the most vulnerable populations and declined in areas with less vulnerable populations. This type of information can be used to identify ecological conservation and restoration targets that enhance the livelihoods of people in vulnerable communities and promote biodiversity and ecosystem health.

Santa Cruz watershed↗
USGS2022

Machine learning for understanding inland water quantity, quality, and ecology

This chapter provides an overview of machine learning models and their applications to the science of inland waters. Such models serve a wide range of purposes for science and management: predicting water quality, quantity, or ecological dynamics across space, time, or hypothetical scenarios; vetting and distilling raw data for further modeling or analysis; generating and exploring hypotheses; estimating physically or biologically meaningful parameters for use in further modeling; and revealing patterns in complex, multidimensional data or model outputs. An important research frontier is the injection of limnological knowledge into machine-learning models, which has shown great promise for increasing such models’ accuracy, trustworthiness, and interpretability. Here we describe a few of the most powerful machine learning tools, describe best practices for employing these tools and injecting knowledge guidance, and give examples of their applications to advance understanding of inland waters.

Book chapter↗
USGS2014

Projecting climate effects on birds and reptiles of the Southwestern United States

We modeled the current and future breeding ranges of seven bird and five reptile species in the Southwestern United States with sets of landscape, biotic (plant), and climatic global circulation model (GCM) variables. For modeling purposes, we used PRISM data to characterize the climate of the Western United States between 1980 and 2009 (baseline for birds) and between 1940 and 2009 (baseline for reptiles). In contrast, we used a pre-selected set of GCMs that are known to be good predictors of southwestern climate (five individual and one ensemble GCM), for the A1B emission scenario, to characterize future climatic conditions in three time periods (2010–39; 2040–69; and, 2070–99). Our modeling approach relied on conceptual models for each target species to inform selection of candidate explanatory variables and to interpret the ecological meaning of developed probabilistic distribution models. We employed logistic regression and maximum entropy modeling techniques to create a set of probabilistic models for each target species. We considered climatic, landscape, and plant variables when developing and testing our probabilistic models. Climatic variables included the maximum and minimum mean monthly and seasonal temperature and precipitation for three time periods. Landscape features included terrain ruggedness and insolation. We also considered plant species distributions as candidate explanatory variables where prior ecological knowledge implicated a strong association between a plant and animal species. Projected changes in range varied widely among species, from major losses to major gains. Breeding bird ranges exhibited greater expansions and contractions than did reptile species. We project range losses for Williamson’s sapsucker and pygmy nuthatch of a magnitude that could move these two species close to extinction within the next century. Although both species currently have a relatively limited distribution, they can be locally common, and neither are presently considered candidates for prospective endangerment. We project range losses of over 40 percent, from its current extent of occurrence, for the plateau striped whiptail, Arizona black rattlesnake, and common lesser earless lizard. Currently, these reptile species are thought to be common or at least locally abundant throughout their ranges. The total contribution of plants in each distribution model was very small, but models that contained at least one plant always outperformed models with only physical variables (climatic or landscape). The magnitude of change in projected range increased further into the future, especially when a plant was in the model. Among bird species, those that had the strongest association with a landscape feature during the breeding season, such as terrain ruggedness and insolation, exhibited the smallest contractions in projected breeding range in the future. In contrast, bird species that had weak associations with landscape features, but strong climatic associations, suffered the greatest breeding range contractions. Thus, landscape effects appeared to buffer some of the negative effects of climate change for some species. Among bird species, magnitude of change in projected breeding range was positively related to the annual average temperature of their baseline distribution, thus species with the warmest breeding ranges exhibited the greatest changes in future breeding ranges. This pattern was not evident for reptiles, but might exist if additional species were included in the model. Our results provide managers with a series of projected range maps that will enable scientists, concerned citizens, and wildlife managers to identify what the potential effects of climate change will be on bird and reptile distributions in the Western United States. We hope that our results can be used in proactive ways to mitigate some of the potential effects of climate change on selected species.

Colorado Plateau;Sonoran Desert↗
USGS2015

Empirical evaluation of the conceptual model underpinning a regional aquatic long-term monitoring program using causal modelling

Conceptual models are an integral facet of long-term monitoring programs. Proposed linkages between drivers, stressors, and ecological indicators are identified within the conceptual model of most mandated programs. We empirically evaluate a conceptual model developed for a regional aquatic and riparian monitoring program using causal models (i.e., Bayesian path analysis). We assess whether data gathered for regional status and trend estimation can also provide insights on why a stream may deviate from reference conditions. We target the hypothesized causal pathways for how anthropogenic drivers of road density, percent grazing, and percent forest within a catchment affect instream biological condition. We found instream temperature and fine sediments in arid sites and only fine sediments in mesic sites accounted for a significant portion of the maximum possible variation explainable in biological condition among managed sites. However, the biological significance of the direct effects of anthropogenic drivers on instream temperature and fine sediments were minimal or not detected. Consequently, there was weak to no biological support for causal pathways related to anthropogenic drivers’ impact on biological condition. With weak biological and statistical effect sizes, ignoring environmental contextual variables and covariates that explain natural heterogeneity would have resulted in no evidence of human impacts on biological integrity in some instances. For programs targeting the effects of anthropogenic activities, it is imperative to identify both land use practices and mechanisms that have led to degraded conditions (i.e., moving beyond simple status and trend estimation). Our empirical evaluation of the conceptual model underpinning the long-term monitoring program provided an opportunity for learning and, consequently, we discuss survey design elements that require modification to achieve question driven monitoring, a necessary step in the practice of adaptive monitoring. We suspect our situation is not unique and many programs may suffer from the same inferential disconnect. Commonly, the survey design is optimized for robust estimates of regional status and trend detection and not necessarily to provide statistical inferences on the causal mechanisms outlined in the conceptual model, even though these relationships are typically used to justify and promote the long-term monitoring of a chosen ecological indicator. Our application demonstrates a process for empirical evaluation of conceptual models and exemplifies the need for such interim assessments in order for programs to evolve and persist.

California, Colorado, Idaho, Montana, Nevada, Oreg↗
USGS2017

The value of information for woodland management: Updating a state–transition model

Value of information (VOI) analyses reveal the expected benefit of reducing uncertainty to a decision maker. Most ecological VOI analyses have focused on population models rarely addressing more complex community models. We performed a VOI analysis for a complex state–transition model of Box-Ironbark Forest and Woodland management. With three management alternatives (limited harvest/firewood removal (HF), ecological thinning (ET), and no management), managing the system optimally (for 150 yr) with the original information would, on average, increase the amount of forest in a desirable state from 19% to 35% (a 16-percentage point increase). Resolving all uncertainty would, on average, increase the final percentage to 42% (a 19-percentage point increase). However, only resolving the uncertainty for a single parameter was worth almost two-thirds the value of resolving all uncertainty. We found the VOI to depend on the number of management options, increasing as the management flexibility increased. Our analyses show it is more cost-effective to monitor low-density regrowth forest than other states and more cost-effective to experiment with the no-management alternative than the other management alternatives. Importantly, the most cost-effective strategies did not include either the most desired forest states or the least understood management strategy, ET. This implies that managers cannot just rely on intuition to tell them where the most VOI will lie, as critical uncertainties in a complex system are sometimes cryptic.

Victoria↗
USGS2022

Bayesian applications in environmental and ecological studies with R and Stan

Modern ecological and environmental sciences are dominated by observational data. As a result, traditional statistical training often leaves scientists ill-prepared for the data analysis tasks they encounter in their work. Bayesian methods provide a more robust and flexible tool for data analysis, as they enable information from different sources to be brought into the modelling process. Bayesian Applications in Evnironmental and Ecological Studies with R and Stan provides a Bayesian framework for model formulation, parameter estimation, and model evaluation in the context of analyzing environmental and ecological data. Features: An accessible overview of Bayesian methods in environmental and ecological studies Emphasizes the hypothetical deductive process, particularly model formulation Necessary background material on Bayesian inference and Monte Carlo simulation Detailed case studies, covering water quality monitoring and assessment, ecosystem response to urbanization, fisheries ecology, and more Advanced chapter on Bayesian applications, including Bayesian networks and a change point model Complete code for all examples, along with the data used in the book, are available via GitHub The book is primarily aimed at graduate students and researchers in the environmental and ecological sciences, as well as environmental management professionals. This is a group of people representing diverse subject matter fields, who could benefit from the potential power and flexibility of Bayesian methods.

Book↗