Via Pennsylvania State University's Earth System Science Center: North Atlantic Hurricane Season Prediction. Excerpt:
ESSC scientists Dr. Michael E. Mann and Daniel J. Brouillette and alumnus Dr. Michael Kozar have released their seasonal prediction for the 2020 North Atlantic hurricane season, which officially starts on 1 June and runs through 30 November.
The prediction is for 19.8 +/- 4.4 total named tropical cyclones, which corresponds to a range between 15 and 24 storms, with a best estimate of 20 named storms. This prediction was made using the statistical model of Kozar et al. (2012, see PDF here). This statistical model builds upon the past work of Sabbatelli and Mann (2007, see PDF here) by considering a larger number of climate predictors and including corrections for the historical undercount of events (see footnotes).
The assumptions behind this forecast are (a) the persistence of current North Atlantic Main Development Region (MDR) sea surface temperature (SST) anomalies (+1.1 °C in early to mid-April 2020 from NOAA’s Coral Reef Watch) throughout the 2020 hurricane season, (b) the development of mild El Niño-Southern Oscillation (ENSO)-negative conditions by boreal late summer and early fall 2020 (ENSO forecasts here; we used mid-April 2020), and (c) climatological mean conditions for the North Atlantic Oscillation in boreal fall/winter 2020-2021.
If no La Niña develops, then the prediction will be slightly lower: 18.3 +/- 4.3 storms (range of 14-23 storms, with a best guess of 19).
Using an alternative model that uses "relative" MDR SST (MDR SST with the average tropical mean SST subtracted) in place of MDR SST yields a considerably lower prediction (13.6 +/- 3.7 total named storms). This alternative model also includes mild ENSO-negative conditions.