Movement-based ecological network design based on realistic simulation by individual-based model: an opportunity for process-based ecological network design with real conservation efficacy

Current ecological network (EN) design primarily relies on least-cost paths simulated by minimum-cumulative resistance (MCR) method. It oversimplifies animal behaviors, where resulting EN often fails to support actual inter-habitat transfer. We argue corridors emerge from the repeated spontaneous traverse of animal movement and advocate using individual-based model (IBM) to simulate realistic movement for EN design. We develop an IBM that deconstructs animal behavior by perception, decision and actions, incorporating 7 parameters related to animal traits and movement mechanisms to model movement in natal dispersal. An new corridor identification algorithm is proposed to identify asymmetric corridors for EN design from the simulated spatially-explicit movement paths. Taking Mustela sibirica as example, its movement is simulated by IBM. ENs designed from IBM-simulated movement (ENMEND) and MCR method (ENMCR) are compared. ENMEND comprises multiple corridors, landscape corridors and linear corridors, whereas ENMCR only contains linear corridors. Connectivity of ENMCR is nearly 5 times to that of ENMEND. We reduce cost-value in ENMEND and ENMCR to represent the effect of implementing respective EN, forming 2 scenarios. Movement of M. sibirica is simulated in the scenarios using IBM. ENMEND achieves 3% increase in successful transfer by reducing 0.04 mean cost-value over 4.3% study area, while ENMCR reduces 11.1% successful transfer by reducing 10 mean cost-value in 46.1% study area. The results reveal the inequivalence between “enlarging corridor extent” and “increasing successful transfer” in conventional methods, highlighting the need for EN design that better aligns with animal behaviors to achieve real increase in successful transfer.