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…
Recent advancements in modeling population dynamics of endangered species have led to improved understanding, yet decision optimization approaches frequently simplify these dynamics. This results in uncertainty regarding dynamic decision models' outputs due to reliance on assumptions about the underlying population dynamics. By integrating population models with deep reinforcement learning, a framework has been developed that delivers adaptive management strategies informed by data and ecologically detailed.
This framework was applied to the supplementation program for the endangered Rio Grande silvery minnow. Our IPM-DRL framework created an adaptive decision model that makes production and distribution decisions in response to observed demographic, hydrological, and genetic environments. This model outperformed heuristic approaches in simulations, excelling under management objectives that prioritize persistence and genetic impact.
The supplementation strategy currently in use performed 5.3% worse under the persistence-focused objective, and 185% worse under the genetics-focused one. The model's decisions regarding minimum sub-population size and total population size were strongly influenced. The results indicate that the IPM-DRL framework presents a high-performing and interpretable decision-support tool for managing endangered species.
Written by urgent.news from bioRxiv's reporting — not their text. Machine-written — may contain errors; check the original before relying on it.
