Biology subjects

Yackulic, C. B.

Publications and source records attributed to Yackulic, C. B..

5 recordsLinked to original sources

bioRxiv2026

Bridging Ecological Inference and Decision Optimization for Conservation Using Artificial Intelligence

The ability to model the complex and uncertain population dynamics of endangered species has improved dramatically in recent decades. However, approaches to identify optimal decisions often require a simplified representation of population dynamics. This leads to a conundrum where managers may be unsure about the output of dynamic decision models because they rely on simplified assumptions of the underlying population dynamics. Here, by pairing integrated population models (IPM) that synthesize diverse ecological data with deep reinforcement learning (DRL) capable of optimizing decisions with high-dimensional uncertainty, we introduce a framework that delivers data-driven and ecologically detailed adaptive management strategies. We demonstrate its utility through application to the supplementation program for the endangered Rio Grande silvery minnow. Using our IPM-DRL framework, we developed an adaptive decision model that selects production and distribution decisions of the supplementation program in response to the observed demographic, hydrological, and genetic environment. The decision model outperformed all heuristic approaches in the simulation across management objectives that weighed persistence and effective population size-related genetic impact differently. For example, the currently deployed supplementation strategy performed 5.3% worse than the decision model under the persistence-focused objective scoring and 185% worse under the genetics-focused one. Analysis of the models decisions in relation to demographic and environmental covariates revealed that minimum sub-population size and total population size were primary drivers of the models decisions. The results demonstrate that the IPM-DRL framework offers a high-performing and interpretable decision-support tool for managing endangered species. SignificanceConservation problems, like imperiled species management, are often challenging because the system dynamics are complex and uncertain. We demonstrate how combining an integrated population model that infers key demographic processes from noisy ecological data with a deep reinforcement learning framework that optimizes management actions addresses these challenges by generating high-performing supplementation strategies for a conservation-dependent species. Our approach embeds two decades of monitoring data within a multi-objective decision-making environment that accounts for ecological uncertainty. The result is a generalizable framework that links ecological inference directly to actionable policy outcomes, enabling scientists and managers to move beyond describing system states and processes toward identifying optimal management actions.

ecology↗

bioRxiv2024

Declining reservoir elevations following a two-decade drought increase water temperatures and non-native fish passage facilitating a downstream invasion

River ecosystems are threatened by interactions among river regulation, non-native species, and climate change. Water use has exceeded supply for two decades in the USAs Colorado River basin draining its two largest storage reservoirs (Lake Powell and Lake Mead). In 2022, after more than two decades of declining reservoir levels and warming downstream river water temperatures, Lake Powell began releasing water from its lower epilimnion into the Grand Canyon segment of the Colorado River. While managers were concerned about the risk of non-native, predatory smallmouth bass (Micropterus dolomieu) entrainment and reproduction, they lacked a quantitative tool to understand entrainment rates and population growth potential under different reservoir management strategies. To fill this void, we developed models in spring 2022 that: 1) predicted propagule pressure at different reservoir elevations, and 2) linked reservoir storage/operations, water temperatures, and smallmouth bass population dynamics to forecast population growth rates under different hydrologic and management scenarios. In the summers of 2022 and 2023, smallmouth bass were documented reproducing in the lower Colorado River for the first time. Our models accurately forecasted adult catch of smallmouth bass in 2022 and 2023 and forecasted that reproduction would occur in both years for the first time in the history of this river segment. Above average runoff in 2023 increased reservoir elevations, however the potential for smallmouth bass establishment remains high because of long-term forecasts of reduced reservoir inflows and lake levels significantly below full pool. Maintaining Lake Powell elevations above 1,088 m (3,570 ft) would likely minimize propagule pressure from the reservoir and would likely create downstream conditions that minimize smallmouth bass population growth.

ecology↗

bioRxiv2023

Fast and flexible Bayesian Jolly Seber models and application to populations with transients

O_LIJolly Seber (JS) models are an appealing class of capture-recapture models for modeling open populations because they allow for inferences about an array of population parameters, including abundance, survival, and recruitment. Multiple formulations of JS models have been developed and include both maximum likelihood and Bayesian approaches. Bayesian approaches offer greater flexibility; however, they are often extremely slow when applied to moderate to large populations because the latent states of all observed individuals, as well as any potential unobserved individuals (parameter-expanded data augmentation), must be simulated at each time step. C_LIO_LIJS model implementations can be sped up dramatically through marginalization, whereby the conditional likelihoods of states are tracked for unique capture histories, rather than the simulated states of all individuals. Here we describe a marginalized implementation of a multistate JS model and compare its performance to that of an implementation using simulated discrete states. We fit models to data generated under two scenarios, one in which no data was missing and another in which 25% of data was randomly removed. To illustrate how marginalization can accommodate more complex models and datasets, we describe a modified version of the model for application to populations with transients and fit the model to simulated data and to data collected on Kentucky Warblers (Geothlypis formosa) as part of the Monitoring Avian Productivity and Survivorship (MAPS) program. C_LIO_LIBoth the discrete latent state and marginalized JS model implementations performed similarly with respect to bias and coverage; however, the marginalized implementation was roughly 1,000 times faster when no data were missing and 100 times faster when 25% of data was missing. The marginalized model accommodating transients also converged quickly and a more complex version of the model applied to the larger multi-site Kentucky Warbler data set yielded useful estimates of most parameters within a reasonable time frame (hours). C_LIO_LIGains in efficiency provided by marginalization shown here should stimulate additional study and development of this useful class of models and provide new opportunities for real-world applications to large complex data sets. C_LI

ecology↗

bioRxiv2023

Models of underlying autotrophic biomass dynamics fit to daily river ecosystem productivity estimates improve understanding of ecosystem disturbance and resilience

Directly observing autotrophic biomass at ecologically relevant frequencies is difficult in many ecosystems, hampering our ability to predict productivity through time. Since disturbances can impart distinct reductions in river productivity through time by modifying underlying standing stocks of biomass, mechanistic models fit to productivity time series can infer underlying biomass dynamics. We incorporated biomass dynamics into a river ecosystem productivity model for six rivers to identify disturbance flow thresholds and understand the resilience of primary producers. The magnitude of flood necessary to disturb biomass and thereby reduce ecosystem productivity was consistently lower than the more commonly used disturbance flow threshold of the flood magnitude necessary to mobilize river bed sediment. The estimated daily maximum percent increase in biomass (a proxy for resilience) ranged from 5% to 42% across rivers. Our latent biomass model improves understanding of disturbance thresholds and recovery patterns of autotrophic biomass within river ecosystems.

ecology↗

bioRxiv2020

Quantifying the demographic vulnerabilities of dry woodlands to climate and competition using range-wide monitoring data

Climate change is expected to alter the distribution and abundance of tree species, impacting ecosystem structure and function. Yet, anticipating where this will occur is often hampered by a lack of understanding of how demographic rates, most notably recruitment, vary in response to climate and competition across a species range. Using large-scale monitoring data on two dry woodland tree species (Pinus edulis and Juniperus osteosperma), we develop an approach to infer recruitment, survival, and growth of both species across their range. In doing so, we account for ecological and statistical dependencies inherent in large-scale monitoring data. We find that warming and drying conditions generally lead to declines in recruitment and survival, but there were some idiosyncrasy in the strength of responses across species. Climate conditions lead to vulnerable regions, such as Pinus edulis in N. Arizona, where both survival and recruitment are low. Our approach provides a path forward for leveraging emerging large-scale monitoring and remotely sensed data to anticipate the impacts of global change on species distributions.

ecology↗