Evacuation is often a necessary volcanic risk mitigation measure . However, the focus is typically on primary evacuation from immediate life safety from the eruption, compared to secondary evacuation from loss of critical services and networked infrastructure. Evacuation, and the associated decision -making process, can be streamlined if probable evacuation can be forecast, and hence prepared for. This paper provides a framework using geospatial data, probabilistic modelling and cost -benefit analysis to produce a geospatial evacuation decision support tool. In this paper , we present a case study for the tephra fall hazard to dairy livestock for Mt Taranak i, New Zealand, applying Bayesian Event Tree for Volcanic Hazard (BET_VH) modelling in conjunction with tephra dispersal modelling using TEPHRA 2 . We model two eruption types: 1) smaller Merapian; and 2) larger sub -Plinian , estimating tephra fall coverage with associated probabilities of occurrence. A comprehensive geospatial inventory is compiled, associating dairy farms with essential services: modelled electricity and water supply sources, and road access. Fault tree analysis and fragility functions are used to assess impact to: 1) feed supply; 2) water supply; and 3) milking ability for both on farm and external services e.g. electricity and water supply for each farm in the region. Without any of these, animal welfare will be threatened, even within a short timeframe e.g. two weeks. Cost -benefit analysis establishes if and where it will probably be beneficial to evacuate. This paper provides the first quantitative geospatial inter -dependency study for the Taranaki dairy industry from tephra fall, but also provides a framework that can be used to assess the impact to any geophysical hazard for any community or sector.