Jump to content

  • Member Statistics

    17,506
    Total Members
    7,904
    Most Online
    SnowHabit
    Newest Member
    SnowHabit
    Joined

2018 North Atlantic tropical storm / hurricane season forecast contest

Roger Smith

By Roger Smith,
in Tropical Headquarters

 Share

Recommended Posts

Roger Smith

Roger Smith

Posted
Posted

What kind of North Atlantic tropical cyclone season do you foresee? 

Our annual contest asks you to predict the total number of storms, and break it down by months. As you know, the "count" is total number of named storms, then total number of those that become hurricanes, and finally total number of hurricanes that become major (cat-3 or stronger). 

Last year the final count was 17/10/6 and our contest winner was Rockchalk83. 

The rules are fairly simple. You need to post a seasonal total by the deadline which I am setting as June 4th as we often seem to be getting entries past the season's June 1st starting point. I will try to expand the field by publicizing the contest in regional forums this year. 

Your seasonal total should include any developments in May, but if there is an outcome (GFS keeps showing a late May hurricane in the eastern Gulf) earlier forecasts that are backed up by monthly predictions will be adjusted if it appears that the May storm is not anticipated. Seasonals that have no complete monthly component will be left unadjusted as it could be assumed that the May storm is included. 

Last year, all entrants gave monthly predictions and almost all left them unedited, but the contest rules allow you to submit these up to 06z of the first of each month, or to revise them by that deadline. Note, there is no requirement for your monthly numbers to add up to your seasonal numbers and you can use decimal points to express uncertainty. Since most of the seasonal forecasts were the sum of the monthly predictions, I assumed that any May activity could be factored in. Some past years, we have had to deal with even earlier storms that happened before the contest announcement (January, April). 

I am going to post my forecast mostly as a guide to how your forecast should appear (not the numbers but the format) ...

Roger Smith ____ 21 16 7 ____ (Jun) 2 1 0 _ (Jul) 1 1 0 __ (Aug) 4 3 2 __ (Sep) 7 5 3 __ (Oct) 5 4 2 __ (Nov-Dec) 1 1 0

(this assumes a 1 1 0 May, would be adjusted to 20 15 7 if 0 0 0 or 21 15 7 if 1 0 0)

Scoring for the contest is as follows:

50% of the score is based on the seasonal. You start with 50 points and in each category, you lose half of (1 point per error plus that number squared). Example, you predict 16 storms, the actual is 19, your error is (3 + 3 squared)/2 = 12/2 = 6. If you had similar errors for hurricanes and majors, your total score would then be 32/50. 

The other 50% of the score is based on your monthly forecasts starting with June and ending with Nov-Dec. These are worth 4, 6, 12, 16, 10 and 2 points in order from June to Nov-Dec. Then the errors reduce your possible score in the same way as the seasonal formula, except that June, July and Nov-Dec go with half the reductions (in other words, you lose points at half the rate of the other more active months). 

In the past few years, highest scores in the contest have been close to 90 and seasonal scores of 48 to 50 have been achieved. A good monthly set will earn 41-43 points. 

Good luck if you enter. The deadline will be made more precise around June 1st and all entrants can edit up to the eventual deadline without notice, you can assume that I won't be copying down or storing any forecasts until a firm deadline is posted in the thread and up to that firm deadline, so no need to post new numbers, just edit the old ones. Once a table of entries appears, your numbers are set for the seasonal and June portions but you have the option of posting revised monthly numbers at any point during the contest before monthly deadlines. Late monthly revisions will be penalized at a rate of 10% per day but will not be valid after any named storm is declared in the month. 

As you see, I am expecting a very active season. There seems to be no particular reason why the outcome from 2017 would change very much this year, hopefully the majors will stay out to sea more often and avoid the sorts of landfalls that caused so much devastation in 2017. 

 

  • Like 1
Link to comment
Share on other sites
Roger Smith

Roger Smith

Posted
  • Author
Posted

Thanks for entering the contest ... further entries will be accepted at least to June 4th then we will make up a table of entries.

We use a "contest normal" which is basically derived from the average of 1989 to 2017, the active season last year has bumped up the values slightly from last year's contest. Also we add in the out-of-season May 1,0,0 to that contest normal as we have done the past two years for January to May activity. This makes the contest normal 

 

_________________________ SEASON __ MAY _ JUN _ JUL _ AUG_ SEP _ OCT_ N-D

 

Adjusted Normal 1989-2017____ 16_8_3 __ 100 _ 100__100__421__641__321__000

(note that N-D has an actual average of something like 0.4/0.1/0.01 so it rounds down to 000)

::::::::::::::::::::::::::::::::::::::::::::::

We also take the NHC "mid-range" from their published seasonal forecast. It is assumed that the May activity is part of their forecast. 

The mid-range of their 2018 forecast is 13 7 2.5, which is almost identical to the Weather Channel at 13 7 2. (source: Wikipedia article)

To make their entry complete, I have scaled their forecast to the contest normal as follows:

NHC mid-range ___________ 13 7 2.5 __ 100 _ 000 _ 100 _ 321 _ 531 _ 3 2 0.5 _ 000

If you have entered only a seasonal prediction, I will scale a set of monthlies for you and you can always adjust them going forward.

If you have entered monthlies and they do not include a 100 count for May but add up to your seasonal total, I will add the May 100 to your seasonal forecast. If you have that in your mix then I won't adjust your total. A reminder -- you can adjust monthly forecasts up to the start of each month if you spot new trends. Your seasonal is carved in stone once we set the table of entries though. 

 

 

Link to comment
Share on other sites
ncforecaster89

ncforecaster89

Posted
Posted
11 hours ago, AfewUniversesBelowNormal said:

I mean, the most named storms ever in El Nino is 15 (2004) and second is 12 in 11 years

(Since 1978. The 1960s went higher)

Let me preface my comments by stating that I appreciate your viewpoint, and also don't expect some of the higher predictions to materialize. 

That aside, it's important to remember that the ENSO cycle isn't the only factor influencing seasonal activity, as well as the fact that not all El Nino phases are created equal, so to speak.  Moreover, it's still possible an El Nino could be delayed or even fail to transpire.  

The other significant factor that could lead to suppressed activity (anomalously cold MDR) could very well moderate as we get to the peak months of the season.           

The El Nino year of 1969 (you alluded to) was pretty hyper-active...so a similar active season wouldn't even be unprecedented.      

To reiterate, I too, don't anticipate an unusually active season.  But, there's very little discernable skill in seasonal forecasts at this range.  Considering that, and the recognition that weather forecasting in general is an inexact science...not too much surprises me anymore; relative to the atmospheric sciences.  My own totals are slightly enhanced based on the additional May TS, and the tendency for the NHC to name a system that might otherwise not have been classified back when I interned there in the late 80's.  

Regardless,  I'm looking forward to observing whatever this particular season has in store. 

Link to comment
Share on other sites

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!

Register a new account

Sign in

Already have an account? Sign in here.

Sign In Now
 Share
Go to topic listing

  • Recently Browsing   0 members

    • No registered users viewing this page.
×
×
  • Create New...