The only viable means of acquiring long term measurements of ocean current and ice drift are autonomous sub-sea instruments. During expeditions on several icebreakers in August 2003, such instruments were moored for the first time within all three principal gateways for flow between the Arctic Ocean and Baffin Bay – Nares Strait, Cardigan Strait and Lancaster Sound. The instruments were programmed to measure current, temperature, salinity, ice thickness, ice drift and sea level. All arrays were maintained to the end of the IPY in August 2009. Since that time, observations have been curtailed but not discontinued. The CAT observational capability is depicted on the adjacent map. In addition to conventional temperature-salinity recorders and the sonar suite for measuring current, ice velocity and thickness, the moorings in Lancaster support an ICYCLER – a winched sensor package to collect data in the dangerous top 40 m – an autonomous water sampler to capture water samples at monthly intervals, a sediment trap and a mammal vocalization recorder. Surveys of seawater temperature, salinity and chemical properties were completed when ships were used to service moorings. Seawater samples have been analyzed for trace chemicals that reveal the fractions originating from melted ice, river run-off and inflow from the Pacific and the amount of carbon dioxide exchanged with the atmosphere. The CAT study also includes a theoretical component, based on numerical ocean models configured to represent the Canadian Archipelago in great detail. One is a finite-element model that simulates the tides (heights and depth-average current); it is available via the Internet for user-directed tidal predictions. The other is the NEMO model adapted for Canadian waters 1º and ¼º resolutions. The NEMO sub-grid that covers the Arctic has 10 km resolution over the Canadian Archipelago (see illustration at the right). The NEMO ocean has been linked to the atmosphere as simulated by the MSC’s GEM weather forecast model. A low resolution model of the entire Arctic to needed to deliver acceptable data to the boundary of a high resolution Archipelago model. These nested models are being used to develop an understanding of CAT and its impacts on Arctic climate and marine ecosystems.