SEARCH · Search USGS

Results for “Biodiversity”

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

SourceResearch connectionFrom yearThrough year

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

At least 811 records · Page 45Linked to original sources

USGS2022

Shifting precipitation regimes alter the phenology and population dynamics of low latitude ectotherms

Predicting how species respond to changes in climate is critical to conserving biodiversity. Modeling efforts to date have largely centered on predicting the effects of warming temperatures on temperate species phenology. In and near the tropics, the effects of a warming planet on species phenology are more likely to be driven by changes in the seasonal precipitation cycle rather than temperature. To demonstrate the importance of considering precipitation-driven phenology in ecological studies, we present a case study wherein we construct a mechanistic population model for a rare subtropical butterfly (Miami blue butterfly, Cyclargus thomasi bethunebakeri ) and use a suite of global climate models to project butterfly populations into the future. Across all iterations of the model, the trajectory of Miami blue populations is uncertain. We identify both biological uncertainty (unknown diapause survival rate) and climate uncertainty (ambiguity in the sign of precipitation change across climate models), and their interaction as key factors that determine persistence vs. extinction. Despite uncertainty, the most optimistic iteration of the model predicts that Miami blue butterfly populations will decline under the higher emissions scenario (RCP 8.5). The lack of climate model agreement across the projection ensemble suggests that investigations into the effect of climate change on precipitation-driven phenology require a higher level of rigor in the uncertainty analysis compared to analogous studies of temperature. For tropical species, a mechanistic approach that incorporates both biological and climate uncertainty is the best path forward to understand the effect shifting precipitation regimes have on phenology and population dynamics.

Florida↗

USGS2017

The practice of prediction: What can ecologists learn from applied, ecology-related fields?

The pervasive influence of human induced global environmental change affects biodiversity across the globe, and there is great uncertainty as to how the biosphere will react on short and longer time scales. To adapt to what the future holds and to manage the impacts of global change, scientists need to predict the expected effects with some confidence and communicate these predictions to policy makers. However, recent reviews found that we currently lack a clear understanding of how predictable ecology is, with views seeing it as mostly unpredictable to potentially predictable, at least over short time frames. However, in applied, ecology-related fields predictions are more commonly formulated and reported, as well as evaluated in hindsight, potentially allowing one to define baselines of predictive proficiency in these fields. We searched the literature for representative case studies in these fields and collected information about modeling approaches, target variables of prediction, predictive proficiency achieved, as well as the availability of data to parameterize predictive models. We find that some fields such as epidemiology achieve high predictive proficiency, but even in the more predictive fields proficiency is evaluated in different ways. Both phenomenological and mechanistic approaches are used in most fields, but differences are often small, with no clear superiority of one approach over the other. Data availability is limiting in most fields, with long-term studies being rare and detailed data for parameterizing mechanistic models being in short supply. We suggest that ecologists adopt a more rigorous approach to report and assess predictive proficiency, and embrace the challenges of real world decision making to strengthen the practice of prediction in ecology.

Ecological Complexity↗

USGS2026

Toxicity of anticoagulant rodenticides on Pacific salmon: Assessing lethal and sublethal effects

To restore native biodiversity on island ecosystems containing invasive rodents, partial- and whole-island eradications generally rely on broadcast baiting with anticoagulant rodenticides (ARs). This approach can result in bait pellets entering aquatic environments, raising concerns about effects to non-target fish. Salmonids are a dominant group of fishes on many temperate islands targeted for rodent eradication, and AR toxicity data for salmonids are limited. Our goal was to determine if coho salmon ( Oncorhynchus kisutch ) are susceptible to coagulopathy and death via exposure to commonly used ARs. We assessed risk of ARs to coho using dose-response curves generated through intraperitoneal injections after determining that coho would not directly ingest the AR baits. Median lethal doses (96-h LD 50 ) estimated using 100 % corn oil carrier were 85.7 µg/g for brodifacoum and 54.0 µg/g for diphacinone. Acetone (30–41 %), used to dissolve ARs in corn oil, reduced the toxicity of diphacinone (LD 50 = 102.3 µg/g, p < 0.001) but not brodifacoum (LD 50 = 73.3 µg/g, p = 0.126) indicating that solvent choice can influence toxicity outcomes. Behavioral changes and onset of mortality differed between the two ARs, with diphacinone acting more rapidly. Tissue analysis supported a difference in toxicokinetics between the two ARs, with significant decreases in liver and muscle residues for diphacinone but not brodifacoum. Sublethal brodifacoum exposure (53.9 µg/g; LD 13 ) impaired blood clotting at 72- and 96- h but returned to baseline by 120 h. No clotting impairment was observed up to 144 h after diphacinone exposure (45.5 µg/g; LD 4 ), suggesting a non-coagulopathy mode of action. These findings will inform risk assessments when considering use of these ARs for rodent management near streams and shorelines and clearly demonstrate that brodifacoum causes coagulopathy in coho.

Ecotoxciology and Environmental Safety↗

USGS2023

Invaders at the doorstep: Using species distribution modeling to enhance invasive plant watch lists

Watch lists of invasive species that threaten a particular land management unit are useful tools because they can draw attention to invasive species at the very early stages of invasion when early detection and rapid response efforts are often most successful. However, watch lists typically rely on the subjective selection of invasive species by experts or on the use of spotty occurrence records. Further, incomplete records of invasive plant occurrences bias these watch lists towards the inclusion of invasive plant species that may already be present in a land management unit, because the occurrences have not been formally integrated into publicly accessible biodiversity databases. However, these problems may be overcome by an iterative approach that guides more complete detection and compilation of invasive plant species records within land management units. To address issues from unobserved or unrecorded occurrences, we combined predicted suitable habitat from species distribution models and aggregated invasive plant occurrence records to develop ranked watch lists of 146 priority invasive plant species on >4000 land management units from five different administrative types within the United States. Based on this analysis, we determined that on average 84% of priority invasive plants with suitable habitat within a given land management unit were as yet unobserved, and that 41% of those were ‘doorstep species’ – found within 50 miles of the unit boundary yet not detected within the unit. Two case studies, developed in collaboration with staff at U.S. Fish and Wildlife Service Refuges, showed that by combining both habitat suitability models and invasive plant occurrence records, we could identify additional problematic invasive plants that had been previously overlooked. Model-based watch lists of ‘doorstep species’ are useful tools because they can objectively alert land managers to threats from invasive plants with high likelihood of establishment.

contiguous United States↗

USGS2024

A spatiotemporal optimization engine for prescribed burning in the Southeast US

Many ecosystems in the Southeast US are dependent upon frequent low-intensity surface fires to sustain native biodiversity, ecosystem services, and endangered species populations. Today, landscape-scale prescribed fire is required to manage these systems for conservation objectives and to mitigate wildland fire risk. Successful application of prescribed fire in this region requires careful planning and assessment of the risks and tradeoffs involved when deciding whether or not to conduct a burn. Many of these risks are closely tied to ambient environmental conditions and are reflected in sets of ‘prescription’ parameters that define safe and effective operating conditions to meet objectives or regulatory requirements. To facilitate effective decision making and acknowledge growing uncertainties related to climate change effects on wildland fire operations, we developed a spatiotemporal optimization engine to identify near-term optimal burning opportunities for prescribed fire implementation. By mining historical 3-day numerical weather forecasts and observation-based weather data for 2015–2021, we have developed a Bayesian hierarchical model for forecast verification that provides calibrated daily weather forecasts and joint uncertainty estimates on meteorological variables of interest, with the latter serving as a measure of risk associated with prescribed fire activities. Burn allocation decisions are then optimized by considering this risk jointly with the utility of burning a particular habitat parcel. The initial iteration of the optimization engine is demonstrated through a case study of short-term meteorological conditions for the Eglin Air Force Base, located in Florida, USA. Results indicate agreement between the optimization engine and the observed past decision-making, with the largest divergences likely arising primarily from differences between utility functions presumed important and used to develop the optimization engine versus the true utility functions driving management behavior in practice.

Florida↗

USGS2022

Grassland conservation supports migratory birds and produces economic benefits for the commercial beekeeping industry in the U.S. Great Plains

Although declines in grassland birds have been documented, national initiatives to conserve grasslands and their biota have fallen short in part because the non-market values of natural ecosystems and species are often not recognized in political decision making. Identifying shared, anthropogenic threats faced by market-valued and non-market-valued species may generate additional support for grassland conservation. We quantify the relationship between the market value of grasslands to commercial beekeepers and the importance of grasslands for birds of conservation concern in North and South Dakota. Our models estimated beekeeping annual revenue increased by $7525 USD and grassland bird abundances increased 2 to 7% per 10-km 2 increase in grassland area. We estimated grassland conversion from 2006 to 2012 resulted in a $2.0 to $2.8 M USD decrease in annual revenue for beekeepers in the Dakotas. Through this study we demonstrate both the market value of grasslands to commercial beekeepers and the non-market benefits of grasslands in supporting migratory birds and discuss the implications of future land-use change. As grassland conversion and subsequent biodiversity loss continue, understanding the co-benefits of grassland conservation may be necessary to illuminate their contributions to society.

North Dakota, South Dakota↗

USGS2014

Remote sensing analysis of riparian vegetation response to desert marsh restoration in the Mexican Highlands

Desert marshes, or cienegas, are extremely biodiverse habitats imperiled by anthropogenic demands for water and changing climates. Given their widespread loss and increased recognition, remarkably little is known about restoration techniques. In this study, we examine the effects of gabions (wire baskets filled with rocks used as dams) on vegetation in the Cienega San Bernardino, in the Arizona, Sonora portion of the US-Mexico border, using a remote-sensing analysis coupled with field data. The Normalized Difference Vegetation Index (NDVI), used here as a proxy for plant biomass, is compared at gabion and control sites over a 27-year period during the driest months (May/June). Over this period, green-up occurred at most sites where there were gabions and at a few of the control sites where gabions had not been constructed. When we statistically controlled for differences among sites in source area, stream order, elevation, and interannual winter rainfall, as well as comparisons of before and after the initiation of gabion construction, vegetation increased around gabions yet did not change (or decreased) where there were no gabions. We found that NDVI does not vary with precipitation inputs prior to construction of gabions but demonstrates a strong response to precipitation after the gabions are built. Field data describing plant cover, species richness, and species composition document increases from 2000 to 2012 and corroborate reestablished biomass at gabions. Our findings validate that gabions can be used to restore riparian vegetation and potentially ameliorate drought conditions in a desert cienega.

Ecological Engineering↗

USGS2017

Evapotranspiration by remote sensing: An analysis of the Colorado River Delta before and after the Minute 319 pulse flow to Mexico

The unique hydrologic conditions characterizing riparian ecosystems in dryland (arid and semi-arid) areas help maintain high biodiversity and support high levels of primary productivity compared to associated uplands. In western North America, many riparian ecosystems have been damaged by altered flow regimes (e.g., impoundments and diversions) and over utilization of water resources (e.g., groundwater pumping for agriculture and human consumption). This has led some state and national governments to provide occasional environmental flows to address the declining condition of such riparian systems. In a historic agreement between the United States and Mexico, 130 million cubic meters (mcm) of water was released to the lower Colorado River Delta in Mexico, with the intent to evaluate the hydrological and biological response of the ecosystem. We used the Moderate Resolution Imaging Spectroradiometer (MODIS) Enhanced Vegetation Index (EVI) to estimate long term (2000–2014) and short term (pre- and post-pulse; 2013 and 2014) evapotranspiration (ET; used herein as an indicator of plant health) of the delta’s riparian corridor. We found the pulse flow helped reverse a decline in ET from 2011 to 2013, with a small, but statistically significant increase in 2014 (P < 0.05). ET was greater than 100 mcm in all years analyzed (even in years without surface flows) and exceeded surface flows in all years except 2000 (result of excess flows following an El Niño cycle in 1997) and 2014 (year of the pulse flow). Based on groundwater salinities and MODIS ET estimates, we estimated groundwater flow into the delta to be ∼103 mcm. Shallow groundwater salinities in the riparian zone increased from 1.30 g L −1 in the most upstream reach to 2.77 g L −1 in the most downstream reach we measured, partly due to uptake of water by riparian vegetation and partly to intrusion of saline agricultural return flows. The disparity between surface flows and ET can likely be explained by the predominantly phreatophytic plants characterizing the area, which draw water from the aquifer. These results also suggest that the deteriorated condition of vegetation within the riparian zone might not be reversed by a single pulse event and could instead require subsequent pulse flows as a long term strategy to restore vegetation in this riparian ecosystem.

Ecological Engineering↗

USGS2015

Phenology and climate relationships in aspen (Populus tremuloides Michx.) forest and woodland communities of southwestern Colorado

Trembling aspen ( Populus tremuloides Michx.) occurs over wide geographical, latitudinal, elevational, and environmental gradients, making it a favorable candidate for a study of phenology and climate relationships. Aspen forests and woodlands provide numerous ecosystem services, such as high primary productivity and biodiversity, retention and storage of environmental variables (precipitation, temperature, snow–water equivalent) that affect the spring and fall phenology of the aspen woodland communities of southwestern Colorado. We assessed the land surface phenology of aspen woodlands using two phenology indices, start of season time (SOST) and end of season time (EOST), from the U.S. Geological Survey (USGS) database of conterminous U.S. phenological indicators over an 11-year time period (2001–2011). These indicators were developed with 250 m resolution remotely sensed data from the Moderate Resolution Imaging Spectroradiometer processed to highlight vegetation response. We compiled data on SOST, EOST, elevation, precipitation, air temperature, and snow water equivalent (SWE) for selected sites having more than 80% cover by aspen woodland communities. In the 11-year time frame of our study, EOST had significant positive correlation with minimum fall temperature and significant negative correlation with fall precipitation. SOST had a significant positive correlation with spring SWE and spring maximum temperature.

Colorado↗

USGS2016

Development and assessment of indices to determine stream fish vulnerability to climate change and habitat alteration

Understanding the future impacts of climate and land use change are critical for long-term biodiversity conservation. We developed and compared two indices to assess the vulnerability of stream fish in Missouri, USA based on species environmental tolerances, rarity, range size, dispersal ability and on the average connectivity of the streams occupied by each species. These two indices differed in how environmental tolerance was classified (i.e., vulnerability to habitat alteration, changes in stream temperature, and changes to flow regimes). Environmental tolerance was classified based on measured species responses to habitat alteration, and extremes in stream temperatures and flow conditions for one index, while environmental tolerance for the second index was based on species’ traits. The indices were compared to determine if vulnerability scores differed by index or state listing status. We also evaluated the spatial distribution of species classified as vulnerable to habitat alteration, changes in stream temperature, and change in flow regimes. Vulnerability scores were calculated for all 133 species with the trait association index, while only 101 species were evaluated using the species response index, because 32 species lacked data to analyze for a response. Scores from the trait association index were greater than the species response index. This is likely due to the species response index's inability to evaluate many rare species, which generally had high vulnerability scores for the trait association index. The indices were consistent in classifying vulnerability to habitat alteration, but varied in their classification of vulnerability due to increases in stream temperature and alterations to flow regimes, likely because extremes in current climate may not fully capture future conditions and their influence on stream fish communities. Both indices showed higher mean vulnerability scores for listed species than unlisted species, which provided a coarse measure of validation. Our indices classified species identified as being in need of conservation by the state of Missouri as highly vulnerable. The distribution of vulnerable species in Missouri showed consistent patterns between indices, with the more forest-dominated, groundwater fed streams in the Ozark subregion generally having higher numbers and proportions of vulnerable species per site than subregions that were agriculturally dominated with more overland flow. These results suggest that both indices will identify similar habitats as conservation action targets despite discrepancies in the classification of vulnerable species. Our vulnerability assessment provides a framework that can be refined and used in other regions.

Ecological Indicators↗

USGS2022

The role of monitoring and research in the Greater Yellowstone Ecosystem in framing our understanding of the effects of disease on amphibians

Emerging infectious disease threatens amphibian biodiversity worldwide, including in landscapes that are protected from many anthropogenic stressors. We summarized data from studies in the Greater Yellowstone Ecosystem (GYE), one of the largest and most complete temperate-zone ecosystems on Earth, to assess the current state of knowledge about ranaviruses and the novel amphibian chytrid fungus (Bd) in this landscape, and to provide insight into future threats and conservation strategies. Our comprehension of these amphibian diseases in the GYE is based on >20 years of monitoring, surveys, population studies, and opportunistic observations of mortality events. Research indicates that local species are affected differently, depending on temperature, community structure, and location in the GYE. Bd has not been linked to die-offs in the GYE but evidence for ongoing reductions in survival contributes to foundational data about the effects of this pathogen in North America. Localized mortality events attributed to, or consistent with, disease from ranaviruses, are widespread in the GYE, but there is less information on how ranaviruses affect amphibian vital rates. The significance of disease in the long-term persistence of amphibians in the GYE is linked to anticipated changes in climate, especially drought. Additionally, expected increases in visitor use, and its associated impacts, have the potential to exacerbate the effects of disease. Long-term information from this large, intact landscape helps to frame our understanding of the effects of disease on amphibians and provides data that can contribute to management decisions, mitigation strategies, and forecasting efforts.

Wyoming↗

USGS2022

Trends in vegetation and height of the topographic surface in a tidal freshwater swamp experiencing rooting zone saltwater intrusion

A decrease in the ground surface height of coastal wetlands is of worldwide concern because of its relationship to peat loss, coastal carbon, and biodiversity in freshwater wetlands. We asked if it is possible to determine indicators of impending transitions of freshwater swamps to other coastal types by examining long-term changes in the environment and vegetation. In a tidal Taxodium distichum swamp in Hickory Point State Forest, Maryland, the topographic surface height (ground surface height) decreased by as much as 25.6 ± 2.2 to 50.8 ± 3.8 cm at two Surface Elevation Tables from 2015 to 2021 following salinity intrusion events related to hurricanes and offshore storms (e.g., Hurricane Melissa). In 2019, rooting zone salinity exceeded 5 ppt for >24.9 % of the time, with a maximum salinity level of 12.5 ppt. Tree growth of T. distichum trees declined and 60 % of these trees died along a 4 m wide × 125 m transect in 2014–2016. Root biomass and ground surface height decreased roughly in conjunction with a salinity pulse in the rooting zone during Hurricane Melissa in 2019. Saplings survived but T. distichum seedlings were uncommon and did not survive in the study area. Typha × glauca increased in cover (0.2 to 5.6 % cover plot −1 ) from 2014 to 2016 so a vegetation shift toward T. × glauca was apparent by 2021. This work captures a multi-year trend of decreasing ground surface height, tree growth and health, and freshwater status in the rooting zone that may be an indicator of impending vegetation transition.

Maryland↗

USGS2023

Intra-specific variation in responses to habitat restoration: Could artificial reefs increase spatiotemporal segregation between migratory phenotypes of lake sturgeon?

Habitat restoration is an important tool used to conserve biodiversity and restore species, but its effects are notoriously difficult to predict. Although outcomes of restoration projects are usually assessed using indices of species abundance and diversity, phenotypic differences among individuals within species are likely associated with differing responses to restored habitats. Here, we use lake sturgeon ( Acipenser fulvescens ) as a case study to illustrate how responses to habitat restoration can differ between phenotypes and potentially lead to unanticipated effects on populations. North America’s St. Clair River supports one of the largest remaining populations of lake sturgeon but has lost much spawning habitat due to its role as a major industrial corridor between the Laurentian Great Lakes Erie and Huron. Two artificial reefs were recently built in the lower and middle segments of the river to increase the available sturgeon spawning habitat. Interestingly, lake sturgeon in the St. Clair River express different migratory phenotypes that may be associated with different likelihoods of colonizing artificial reefs. Acoustic telemetry revealed that artificial reefs were more likely to be used by sturgeon that migrated downstream to overwinter in Lake St. Clair than those that migrated upstream to overwinter in Lake Huron. Furthermore, increasing time spent at the artificial reefs by Lake St. Clair migrants was associated with later arrival to and shorter occupancy of the river’s only natural spawning site, the primary location where the two phenotypes have opportunity to interbreed. Additional research is necessary to determine the ultimate impacts of the artificial reefs on lake sturgeon populations; nevertheless, our study showed phenotype-specific opportunity to colonize restored habitat and a mechanism through which this could lead to changes in gene flow. Our results illustrate the importance of considering intra-specific diversity when planning restoration projects and assessing the effects on populations.

Michigan, Ontario↗

USGS2024

Passive acoustic monitoring and convolutional neural networks facilitate high-resolution and broadscale monitoring of a threatened species

Population monitoring is an essential component of biodiversity conservation and management, but low detection probabilities for rare and/or cryptic species makes estimating abundance and occupancy challenging. Passive acoustic monitoring combined with machine learning algorithms represents a potential path forward to effectively and efficiently monitor the occurrence of rare vocalizing species across entire forest landscapes. Our objectives were to develop and implement a convolutional neural network (PNW-Cnet) to identify vocalizations of a rare and threatened forest nesting bird species – the marbled murrelet ( Brachyramphus marmoratus ) – in the Pacific Northwest, U.S.A., 2018–2021. We used PNW-Cnet predictions from broadscale passive acoustic monitoring data to examine spatiotemporal patterns in the distribution of murrelets. PNW-Cnet showed sufficiently high prediction accuracy (overall precision > 0.9) to enable broadscale population monitoring. Spatiotemporal analysis showed that annual peak murrelet call abundance occurs in ordinal weeks 28–32 (late July–Mid August) but this varied by study area. The greatest number of detections typically occurred in the Olympic Peninsula and Oregon Coast Range where late-successional forest dominates and nearer to ocean habitats. We demonstrate that passive acoustic monitoring can be used to understand intensity of use across broad scales for a rare and cryptic species in addition to the typical detection/non-detection data that are often collected. Passive acoustic monitoring combined with PNW-Cnet offers considerable promise for species distribution modeling and long-term population monitoring for rare species.

Oregon, Washington↗

USGS2010

Accuracy of gap analysis habitat models in predicting physical features for wildlife-habitat associations in the southwest U.S.

Despite widespread and long-standing efforts to model wildlife-habitat associations using remotely sensed and other spatially explicit data, there are relatively few evaluations of the performance of variables included in predictive models relative to actual features on the landscape. As part of the National Gap Analysis Program, we specifically examined physical site features at randomly selected sample locations in the Southwestern U.S. to assess degree of concordance with predicted features used in modeling vertebrate habitat distribution. Our analysis considered hypotheses about relative accuracy with respect to 30 vertebrate species selected to represent the spectrum of habitat generalist to specialist and categorization of site by relative degree of conservation emphasis accorded to the site. Overall comparison of 19 variables observed at 382 sample sites indicated ???60% concordance for 12 variables. Directly measured or observed variables (slope, soil composition, rock outcrop) generally displayed high concordance, while variables that required judgments regarding descriptive categories (aspect, ecological system, landform) were less concordant. There were no differences detected in concordance among taxa groups, degree of specialization or generalization of selected taxa, or land conservation categorization of sample sites with respect to all sites. We found no support for the hypothesis that accuracy of habitat models is inversely related to degree of taxa specialization when model features for a habitat specialist could be more difficult to represent spatially. Likewise, we did not find support for the hypothesis that physical features will be predicted with higher accuracy on lands with greater dedication to biodiversity conservation than on other lands because of relative differences regarding available information. Accuracy generally was similar (>60%) to that observed for land cover mapping at the ecological system level. These patterns demonstrate resilience of gap analysis deductive model processes to the type of remotely sensed or interpreted data used in habitat feature predictions. ?? 2010 Elsevier B.V.

Ecological Modelling↗

USGS2020

Spatial conservation planning under uncertainty using modern portfolio theory and nash bargaining solution

In recent years, researchers from interdisciplinary teams involving ecologists, economists and operations re- searchers collaborated to provide decision support tools to address the challenges of preserving biodiversity by optimizing the design of reserves. The goal of this paper is to further advance this area of research and provide new solutions to solve complex Spatial Conservation Planning (SCP) problems under uncertainty that consider risk preferences of decision makers. Our approach employs modern portfolio theory to address uncertainties in SCP problems, and involves two conflicting objectives: maximizing return and minimizing risk. We apply concepts from game theory such as the Nash bargaining solution to directly compute a desirable Pareto-optimal solution for the proposed bi-objective optimization formulation in natural resource management problems. We demonstrate with numerical examples that by directly computing a Nash bar- gaining solution, a Binary Quadratically Constrained Quadratic Program (BQCQP) can be solved. We show that our approach (implementable with commercial solvers such as CPLEX) can effectively solve the proposed BQCQP for much larger problems than previous approaches published in the ecological literature. Optimal solutions for problems with less than 400 parcels can be computed within a minute. Near optimal solutions (within at most 0.2% gap from an optimal solution) for high-dimensional problems involving up to 800 parcels can be computed within 8 hours on a standard computer. We have presented a new approach to solve SCP optimization problems while considering uncertainty and risk tolerance of decision makers. Our new approach expands considerably the applicability of such SCP optimization methods to address real conservation problems.

Ecological Modelling↗

USGS2022

Modeling the impact of invasive species litter on conditions affecting its spread and potential regime shift

Many introduced plants pose invasion risks globally and threaten the biodiversity of native ecosystems. Such non-native plants can become invasive when they have advantages over native plants, such as having fewer natural enemies. Invasive plants often have the ability to alter ecosystem properties after they have become established, which can make it difficult to eliminate the invasive. In principle, this can cause a regime shift that may not be reversed through intense control efforts that increase mortality and reduce growth of the invasive species. Here we use spatially explicit agent-based modeling to simulate the invasion of an introduced tree species into a habitat occupied by a native species. The model describes an invasive tree with fast growth and high seed production and, in addition, produces litter that has a suppressive effect on native seedlings. These are properties, for example, shared by the invasive Melaleuca quinquenervia in southern Florida habitats. We use simulation modeling to test the following logical hypotheses: Partial suppression of native tree seedlings by the invasive tree's litter (1) will accelerate the spread of the invasive tree into native vegetation, (2) will impede efforts to control invasive spread through biocontrol, and (3) can cause a regime shift that is not reversed even if the biocontrol lowers invasive growth and reproduction to levels substantially lower than those of the native species. Additionally, (4) the earlier in the invasion biocontrol is introduced, the more effective it will be in reversing the invasion. The simulations support all four hypotheses. While these results highlight the potential for biocontrol of invasive tree species, our findings also suggest that successful elimination of positive litter feedbacks and invasive spread may critically depend on the timing of control efforts within the invasion process.

Ecological Modelling↗

USGS2018

Machine learning for ecosystem services

Recent developments in machine learning have expanded data-driven modelling (DDM) capabilities, allowing artificial intelligence to infer the behaviour of a system by computing and exploiting correlations between observed variables within it. Machine learning algorithms may enable the use of increasingly available ‘big data’ and assist applying ecosystem service models across scales, analysing and predicting the flows of these services to disaggregated beneficiaries. We use the Weka and ARIES software to produce two examples of DDM: firewood use in South Africa and biodiversity value in Sicily, respectively. Our South African example demonstrates that DDM (64–91% accuracy) can identify the areas where firewood use is within the top quartile with comparable accuracy as conventional modelling techniques (54–77% accuracy). The Sicilian example highlights how DDM can be made more accessible to decision makers, who show both capacity and willingness to engage with uncertainty information. Uncertainty estimates, produced as part of the DDM process, allow decision makers to determine what level of uncertainty is acceptable to them and to use their own expertise for potentially contentious decisions. We conclude that DDM has a clear role to play when modelling ecosystem services, helping produce interdisciplinary models and holistic solutions to complex socio-ecological issues.

Ecosystem Services↗