2016 Predictions

Named Atlantic Storms in 2016

Seasonal predictions of the number of named storms issued by NOAA, CSU, and by the OLR-based algorithm developed in Karnauskas and Li (2016) are shown below. Finalized OLR-based predictions will be posted on or around June 2 (pre-season), July 2 (one month in), and August 2 (two months in) of each year, as there is a ~2 day lag in the retrieval and processing of satellite OLR measurements. The OLR-based predictions are calculated using two different techniques: linear regression (LR) and random forest (RF; a machine learning technique). Predictions made via both techniques are provided.

A note about uncertainties: NOAA issues ranges only, and CSU issues a single number only. The uncertainties on the OLR-based predictions published here will match those of NOAA each year. In other words, if NOAA's prediction is N ± 3, then the number of storms predicted by the OLR-based algorithm ± 3 will be displayed here. If the NOAA range (maximum minus minimum) is an odd number, the extra storm in the range will go into the direction from which the OLR-based prediction was rounded to the nearest whole number.

2016 OLR Visualizations

The maps below show the optimal OLR anomaly patterns, and the actual OLR anomaly patterns for the current year (as the data become available). The optimal maps reveal the patterns of OLR anomaly most condusive to an active Atlantic hurricane season. The extent to which the current year's OLR patterns match the optimal OLR patterns provides insight into why the OLR-based predictions for the current year may be relatively low or high.

Latest Short Term Outlooks

The maps below show the 2-day and 5-day tropical weather outlooks issued by the NOAA National Hurricane Center (NHC). Click on the images for the NHC page with more information.