J-SCOPE forecast results for the simulation beginning in April of 2016 are shown in the first figure below. Each panel represents a two-month average for the region averaged over three model runs of the ensemble. Each panel represents ensemble averaged anomalies of two month averages for the region. The panels directly under the climatology depict the relative uncertainty from the ensemble for the same time period.

One bias is known to exist in the model that impacts these forecasted SSTs. The short wave bias is known to be too high (see 2013, Year in Review), which can contribute to forecast temperatures being too warm in summer.

For waters off the Washington coast, the forecast projects that SST will be warmer than climatology until September (May - June, July - August) and then return to near climatological temperatures for the rest of 2016. The relative uncertainty is highest offshore in November - December (10%), and lower earlier in the forecast.

For the waters off the Oregon coast, the forecast projects that SST will be warmer than the climatology until September, and then return to near climatological temperatures for the rest of 2016, much like Washington. Relative uncertainty is highest offshore in November - December (10%), and is lower for the rest of the forecast.

The SST was forecast using all three model runs of the ensemble for two point locations where buoys exist, Ćháʔba· off La Push, WA, and NH-10 off Newport, OR, are shown at the bottom of the page.

Finally, climatological cross-sections from the Newport Line in Oregon, as well as the Grays Harbor Line in Washington, are compared to the forecasted average of the ensemble members. The forecast projects that the temperatures for the upwelling season of 2016 will be higher than the climatology and extend down to over 200 meters depth below the surface covering much of the shelf in both Washington and Oregon.

We note that the NANOOS Climatology App shows that the NDBC buoys along the WA and OR shelf are recording sea surface temperatures approximately 2 standard deviations above long-term averages since the start of 2015. The ability of the model to forecast this anomaly will be evaluated in our Year in Review activities.