A Sea-Surface Temperature Homogenization Blend for the Northwest Atlantic

Abstract

As part of the Atlantic Zone Monitoring Program two different methods were applied to merge three Level-3 AVHRR SST products into a homogenized blend covering the time period from 1982 to present over the Northwest Atlantic, providing a regionally-tuned climatological base on which to assess whether current observations are below, near or above normal. Weekly and monthly SST composites were constructed by averaging daily anomalies within each time period and adding the result to the climatological mean for the period. This approach reduces biases introduced from missing data during a strong warming/cooling seasonal period. Since AVHRR SST data have many spatial and temporal gaps, a common difficulty is establishing how much data are sufficient to yield useful estimations of temperature anomalies. A statistical Monte Carlo method showed that monthly and weekly regional averages composed respectively of as little as 7% and 10% of possible data still yield useful results. A test case shows the increased usefulness of the blend for State of the Ocean reporting. Application of the data set confirmed the use of coastal air temperature as a useful proxy for SST allowing hindcasting past changes or forecasting future changes associated with global warming over Eastern Canadian waters.