Accessibility models calculate the cost distance from any point on the surface to any target pixel. The target pixels could be a market, city, port or any other point features that are the destinations of interest. For this project we are using settlements of over 1000 people as the target. The settlement data is taken from the Gridded Population of the World Version 3 (CIESIN et al., 2004). Due to errors in spatial representation, many towns from the settlement data set do not fall exactly on transportation networks. We used the POINTNODE command in ArcInfo to automatically shift those towns to the nearest transportation link. We did not put this step in the AML or the ModelBuilder model because there are some interactive steps that need to be taken to make sure that the POINTNODE command works correctly.
The threat surface actually used in the threat assessment combines accessibility with population. An occurrence is deemed to be under greatest threat if it is both highly accessible (low travel time to the nearest populated place) and in the vicinity of high population. This logic attempts to capture the fact that there is likely higher threat to an occurance found 15 minutes travel time from Sao Paolo in Brazil (population > 10m) compared to an occurrence 15 minutes travel time from the town of Villa de Leyva (population approx. 50,000) in Colombia.