DEMAND MODEL FOR CRUDE OIL RAIL AND PIPELINE SHIPMENTS IN CANADA

Abstract

Global energy demand is expected to increase significantly over the next 10-20 years, and in the absence of an increase in pipeline capacity, Canadian crude oil shipments are likely to be diverted to rail. The current rail loading capacity originating in western Canada is 754,000 barrels/day. This rail capacity is expected to meet the 1 million b/d deficit between western Canadian production forecasted for 2025 (5 million b/d) and existing pipeline capacity (4 million b/d). Western Canadian production is further forecast to reach 5.4 million b/d by 2030, representing a 39% increase from 2016.

The primary objective of this research is to obtain accurate current estimates of crude oil shipments by pipeline and rail in Canada (tonnages and volumes), and to use these data to calibrate and evaluate an empirical model of crude oil shipments by shippers’ mode choice (pipeline and rail) and route selection. Modelling crude oil shipments allows for an assessment of the impacts of future changes in pipeline/rail network connectivity, modal attributes, and shipment protocols, on the expected pattern of crude oil shipments.

Origin Destination (OD) demands were based on empirical trade data. In particular, crude oil shipments beginning and ending in Canada are available from CANSIM, Statistics Canada’s key socioeconomic database. These data were supplemented with other data from the National Energy Board (NEB) and the U.S. Energy Information Administration. Due to data limitations and the need for more disaggregated zones to characterize crude oil shipment patterns more precisely, shipments originating from (destined for) British Columbia or Alberta were split based on total production (attraction) data from the Canadian Association of Petroleum Producers, Alberta Energy, and the Canadian Fuels Association.

Numerous data sources were compiled to estimate cost functions for shipping crude by pipeline and rail in Canada. For pipelines, cost performance functions (shipper tolls) were found to depend significantly on shipping distance, route pipe diameters and shipment destination (domestic vs international). Moreover, medium and heavy crude were found to be more expensive to ship compared to condensate and light crude due to their lower viscosities. For rail, the distance shipped, a terminal fee and an international tariff surcharge were found to be statistically significant in explaining shipper costs along a route. These pipeline and rail cost models yielded R2 values of 0.85 and 0.83, respectively.

Conventional Random Utility Models (RUM) fail to capture the complex interactions of pipeline shipments to determine mode and route choice shipper decisions, resulting in the development of a rule-based approach for mode choice and route assignment. This research found that the mode split, and route assignment of crude oil shipments are jointly determined by shipper types, destination types, prioritization rules, and allocation rules. This approach was validated by comparing predicted throughputs with those reported for Canada’s Group 1 and Group 2 export pipelines, and applied to determine the impact of a 39% increase in western Canadian crude oil production, as forecasted by CAPP for 2030. Assuming no new pipeline infrastructure is constructed before 2030, several rail lines would carry increased amounts of crude oil.