Locating a critical facility such as a facility dealing with hazardous materials (HAZMAT) requires a thoughtful process. The most critical aspect of HAZMAT transportation is the risk to the surrounding population. However, ignoring network vulnerability and path resilience, can seriously hamper route planning and increase the propensity of happening of an incident which eventually converts to transportation risk. Avoiding vulnerable sections of the underlying network can save the shipment from a potential disaster, however, it may not always be possible to avoid network vulnerabilities, therefore the resilience of the route plays a vital role in risk mitigation and must be given equal consideration in HAZMAT route planning. Since the location of a HAZMAT facility and designated routes are interrelated, the facility sites should be selected based on a rigorous route appraisal considering at least these three key factors. A common approach to deal with the location and routing problem simultaneously is to solve a location-routing problem (LRP) which provides optimal sites from a pool of candidate sites. The limitation of the solutions obtained from LRP raises the question that how the optimal sites can be compared to each other or which site among the non-optimal solutions can be selected as an alternate. Since locating a HAZMAT facility is multidisciplinary and works with other disciplines such as geotechnical, geological, or land use, there is a need to have some information which could help to support the location decision in case of any constraint. This thesis presents a chronological study of a special case of LRP in the context of HAZMAT transportation. The thesis is important in two key aspects, first, it provides approaches to explicitly rank the critical HAZMAT facility sites, second, it is inclusive in providing route appraisal and ranking methodologies by specifically considering risk, vulnerability, and resilience. The thesis provides flexibility to incorporate decision-makers preferences in route appraisal and site ranking. Three pertinent research questions are answered in three successive research modules. First, how can the potential HAZMAT sites be ranked based on the transportation risk posed by the routes? Second, what is a valuable trade-off a decision-maker can consider for an alternate route if the designated or shortest paths are not considered feasible? Lastly, how resilient are the designated paths against unknown disruptions and how can the potential sites be ranked based on path resilience behaviour? The first research module incorporates transportation risk and presents an optimization and ranking framework for potential HAZMAT facilities. The procedure is demonstrated with improved risk functions to obtain ranking based on stochastic risk assessment on the preferred routes. The proposed stochastic model relaxes some assumptions made in the traditional deterministic risk estimation approaches and provides a relatively accurate risk estimation. The stochastic process allows determining the probability of optimality to rank the sites. The second research module develops a route evaluation procedure based on network topological vulnerability by using the concept of Random Walk with Restart and (Personalized) PageRank. This method evaluates the designated or shortest routes against potential hazards. This research module augments the routing part of the developed ranking methodology and answers whether an alternate path, rather a more resilient path, can be selected if the decision-maker is willing to trade off path length (or path-determining criteria) with vulnerability. Incorporating real traffic data in finding the critical network elements makes the process dynamic. Finally, the last module proposes a site ranking methodology by addressing how resilient the designated paths are against unknown disruptions and proposes a stochastic model based on the path resilience that offers flexibility to incorporate transportation preference, risk and or route evaluation attributes in the overall site ranking methodology. The path resilience states can be modelled as Discrete Markov Chain to rank the sites at an individual point of interest (POI) level, combined POI level, or multicriteria level. Although the sites change with the hazard circle radius and the actual routes between the POIs, for the scenario considered, the analysis based on transportation risk locates the consistent optimal sites around multimodal transfer terminals at the city of saskatoon and Regina, and at the west side of the province between Highway_1 and Highway_16 along Highway_7. The analysis based on path resilience gives optimal sites on the eastern side of the border along Highway_3 and Highway_9, although other optimal sites are a bit scattered and spread over the Southern part of the province. These sites may be investigated further by Government agencies to guide the facility site licensure processes.