Discussion about models,patterns and seasonal forecasts

[Altima2005]

i've read about the sal before...looks like it can throw off the inner-core development/longevity of the storms pretty badly.

yeah...and the monsoon. another feature that is somewhat difficult to predict in advance. i guess the azores high could show a little bit about the monsoon perhaps.

[MGC]

If I recall, didn't we have a simular argument last winter concerning SSTA? The Tropical Atlantic basin is always warm enough to support TC development year round. Strong UL winds are the limiting factor during the winter and spring months........MGC

[Anonymous]

there's no way on this rock mass that i will try and predict whether or not we will have a trough protecting the ec this summer. dr. gray has alluded that it is impossible to do this. thus, he does not say exactly what regions will get hit during a season. i will take my hat off and bow to anyone who can do that (with decent consistency, too). it's nearly impossible to predict daily changing weather patterns months in advance. perhaps in the late 90's we could predict (somewhat) that the ec would get hit. i do believe that year-to-year trends exist. however, many seasons feature pattenrs that differ from year-to-year.

You bring up some good points but I have to disagree...more detailed stuff such as where hurricanes will strike and form, as well as when, CAN be done. It can. Impossible? No. If Dr. Gray did indeed say that (I don't recall him), then it's time for seasonal hurricane forecasting to get a boost. TWW and I have been doing research over the past several months on the progged steering flow pattern, as well as atmospheric favorability...add in a little bit of climatology/stats and you have a reasaonable detailed idea of what the landfall pattern, formation hotspots, and month-by-month activity will be. This is type of stuff we are aiming to discuss in depth in our seasonal forecast come May 25.

for the reasons above, analogs aren't useful for doing much more than predicting the amount of activity in a season. look back at some of the more reliable records from 1950-present. many years don't feature hits in the same places. if they do, some may be prone to keep forecasting activity to strike the same area. what happens then? the pattern changes.

Of course many years don't feature hits in the same places. A year of that sort would be considered highly anomalous. Now, climatology isn't what TWW and I use solely (I'm sure you know why, it's like that famous saying from Twain)...BUT there ARE patterns that we've found by closely linking and categorizing each year into groups by the many factors that influenced them. Not only do we find total activity correlations, but also LANDFALL correlations. Obviously, if it's just based on 1 or 2 years...or seems subject to error based on what is expected in 2004...then the pattern will be dismissed. But if it seems to go well with the anticipated set up, and/or involves many years, then I'd consider it good to use to forecast.

[Anonymous]

If I recall, didn't we have a simular argument last winter concerning SSTA? The Tropical Atlantic basin is always warm enough to support TC development year round. Strong UL winds are the limiting factor during the winter and spring months........MGC

Yep, SSTs are, but SST ANOMALIES change all the time. Like Derek said if they're cooler or warmer than average during the season it doesn't have a huge influence, but it DOES help CV disturbance if the waters are above average, and vice-versa.

[Altima2005]

Supercane posted:
"You bring up some good points but I have to
disagree...more detailed stuff such as where hurricanes will strike and form, as well as when, CAN be done. It can. Impossible? No. If Dr. Gray did indeed say that (I don't recall him), then it's time for seasonal hurricane forecasting to get a boost. TWW and I have been doing research over the past several months on the progged steering flow pattern, as well as atmospheric favorability...add in a little bit of climatology/stats and you have a reasaonable detailed idea of what the landfall pattern, formation hotspots, and month-by-month activity will be. This is type of stuff we are aiming to discuss in depth in our seasonal forecast come May 25."

okay...there is some degree of predictability for atmospheric favorability. in addition, the pattern that sets up in late july and early august will sometimes hint at what might be in store for those peak months. for example, 2003 had conditions that were favorable for cv development by early august. the waves were juicy and the upper level winds were generally favorable. this past season, in general, had a much different look to it compared to 2001/2002.

and it is generally acceptable to say that la nina may enhance the cv season...as will neutral conditions with other favorable factors combined.

still though...i really don't think it is possible to predict months in advance what areas are more at risk during a specific hurricane season.

from gray's forecast (it has been on every single forecast he's ever written):
"Moreover, these forecasts do not specifically predict where within the Atlantic basin these storms will strike. The probability of landfall for any one location along the coast is very low and reflects the fact that, in any one season, most US coastal areas will not feel the effects of a hurricane no matter how active the individual season is. However, it must also be emphasized that a low landfall probability does not insure that hurricanes will not come ashore. Regardless of how active the 2004 (or any) hurricane season is, a finite probability always exists that one or more hurricanes may strike along the US coastline or the Caribbean Basin and do much damage."

my point is, once again, the atlantic is big, the us coastline is big, and weather patterns change rapidly and in many cases unexpectedly.

interesting philosophy that you have there. i look forward to discussing and debating with everyone here this season, and, in particular, you two.

edit: i should also add that i have been reading wx message boards for a while. i find the topic of the "ec trough is finally weakening" to be rediculous. Troughs move through the eastern us on a daily, weekly, and monthly basis. most times, if a hurricane gets affected by a major trough, it's getting the slingshot, regardless of what the pattern over the atlantic may have been like in earlier times.

[cycloneye]

Great discussion going on here from all and that is what we encourage the members to do to make this quiet period before the season starts an active forum in which the members can expose all of their anaylisis and discussions about the upcomming season.

[Anonymous]

still though...i really don't think it is possible to predict months in advance what areas are more at risk during a specific hurricane season.

from gray's forecast (it has been on every single forecast he's ever written):
"Moreover, these forecasts do not specifically predict where within the Atlantic basin these storms will strike. The probability of landfall for any one location along the coast is very low and reflects the fact that, in any one season, most US coastal areas will not feel the effects of a hurricane no matter how active the individual season is. However, it must also be emphasized that a low landfall probability does not insure that hurricanes will not come ashore. Regardless of how active the 2004 (or any) hurricane season is, a finite probability always exists that one or more hurricanes may strike along the US coastline or the Caribbean Basin and do much damage."

my point is, once again, the atlantic is big, the us coastline is big, and weather patterns change rapidly and in many cases unexpectedly.

Weather patterns change rapidly...a very true statement (what fun would it be if the weather stayed constant...yuck). However, my point is, is it possible to FORECAST such rapid changes months in advance??? My answer has been, is, and always will be, YES. The seasonal forecast we made last year had some good wins but also quite a few looses...If our forecast we make on May 25 busts horrendously, it doesn't mean it can't be done...it only means we didn't research hard enough, or the results were misinterpreted. Doesn't mean it's impossible. And climatology will get better and better imo with every year. The more years available to research means a higher probability of good results, and maybe one day climatology will be the best predictor. My point is, if we can't figure out how to forecast the pattern months in advance during this time period (hopefully one of today's researchers can), then as technology and amount of data...somebody will.

interesting philosophy that you have there. i look forward to discussing and debating with everyone here this season, and, in particular, you two.

I look forward as well.

[MGC]

I agree with ya SC....One day we will have the ability to forecast the weather out months in advance. However, I doubt we will see such ability in the short term. Until a perfect computer model is produced, coupled along with more plentiful observation of the oceans and upper air, this dream is years and years away. Just think of the benefits of long range forecasting, a major hurricane could be predicted to impact a major population center months in advance allowing sufficient time to allow its modification.......MGC

[Anonymous]

I agree with ya SC....One day we will have the ability to forecast the weather out months in advance. However, I doubt we will see such ability in the short term. Until a perfect computer model is produced, coupled along with more plentiful observation of the oceans and upper air, this dream is years and years away. Just think of the benefits of long range forecasting, a major hurricane could be predicted to impact a major population center months in advance allowing sufficient time to allow its modification.......MGC

Yep...now predicting the exact city where a major hurricane will impact months in advance is probably a ways off, but I think general landfall area predictions months in advanced can be achieved much sooner, and with the available data. It isn't that more data is needed, just more research.

[mf_dolphin]

I'll have to disagree. Man's ability to forecast weather months away, especially down to a landfalling hurricane's location is a pipe dream in the next 30 lifetimes. They are simply too many variables for that accurate of a prediction. I've been in computers for the last 20 years and while we've made some great strides the hardware is the least of the issues. In order for a model to be written, we first have to uderstand the factors involved as well as the dynamics between all of those factors. Not going to happen for a loooooooooong time

[FLguy]

I agree with ya SC....One day we will have the ability to forecast the weather out months in advance. However, I doubt we will see such ability in the short term. Until a perfect computer model is produced, coupled along with more plentiful observation of the oceans and upper air, this dream is years and years away. Just think of the benefits of long range forecasting, a major hurricane could be predicted to impact a major population center months in advance allowing sufficient time to allow its modification.......MGC

Yep...now predicting the exact city where a major hurricane will impact months in advance is probably a ways off, but I think general landfall area predictions months in advanced can be achieved much sooner, and with the available data. It isn't that more data is needed, just more research.

seasonal forecasting is by NO means exact, however has come a long way over the past 20 years. but to expect to be able to predict exactly where an intense hurricane will make landfall months in advance, or for that matter, before it has even developed, im not convinced is even possible with numerical modeling. and IF it is, i dont think any of us will see it in our life time.

numerical models are great for making an assessment of where a tropical cyclone may make landfall, and how it may track, but they are by no means perfect.

we can use analog methods and climate factors to try to determine what parts of the east coast or the gulf coast might be under the gun based on systems in those years, which is the best method we have going currently. at times it works out, other times is does not.

[Stormsfury]

Although, model guidance has come a long way and especially in pattern recognition, the fact of the matter is that pinpointing certain events to certain locations aren't an exact science ...

Take a look at GSP's first paragraph in an area forecast discussion just last night ...

AREA FORECAST DISCUSSION
NATIONAL WEATHER SERVICE GREENVILLE-SPARTANBURG SC
405 PM EST THU FEB 5 2004

.DISCUSSION...
TODAY IS A GOOD EXAMPLE OF WHY HUMANS ARE STILL NEEDED IN THE
FORECAST PROCESS...AND HOW EASY IT IS TO BECOME RELIANT ON
MODELS AND FORGET THAT THEY CANNOT HANDLE EVERY SCENARIO BY
THEMSELVES. WINTER WEATHER CONTINUES TO DEVELOP ACROSS THE NORTH
CAROLINA MOUNTAINS THIS AFTERNOON...AND NOW FREEZING RAIN/SLEET
COMING IN FROM NORTHERN TIER OF UPSTATE SOUTH CAROLINA. FOR NOW...
WILL KEEP THE WINTER WEATHER ADVISORY IN PLACE FOR THE NORTH
CAROLINA AREAS THAT HAVE THEM. WILL HAVE TO MAKE LAST MINUTE GRID
CHANGES FOR PARTS OF THE UPSTATE THAT HAVE SLID TO 32 FAHRENHEIT OR
BELOW.

[Anonymous]

The debate really isn't about being able to forecast landfalls months in advance. I think we all know that we're still a long ways away from making such forecasts. In fact, it may never happen...but you have to be careful when you say things like that. What Rob and I are saying, is that you can predict landfall probabilities with at least some skill. Dr. Gray wouldn't begin posting probabilities in December if that weren't the case. Some people even question seasonal hurricane predictions and they often criticize Gray. but if you look at his stats over the years you will find that there is forecast skill. The only disagreement with I have is landfall %s. I rather go through all of the climatological stats and provide actual numbers. Rob and I are looking forward to seeing our results after a few years.

[Anonymous]

seasonal forecasting is by NO means exact, however has come a long way over the past 20 years. but to expect to be able to predict exactly where an intense hurricane will make landfall months in advance, or for that matter, before it has even developed, im not convinced is even possible with numerical modeling. and IF it is, i dont think any of us will see it in our life time.

numerical models are great for making an assessment of where a tropical cyclone may make landfall, and how it may track, but they are by no means perfect.

we can use analog methods and climate factors to try to determine what parts of the east coast or the gulf coast might be under the gun based on systems in those years, which is the best method we have going currently. at times it works out, other times is does not.

I agree with pretty much all of your comments. IMO, climatology will solve most of our forecasting issues. But we probably need much more time for climo stats to evolve before seasonal forecasting becomes an EXACT science.

[FLguy]

seasonal forecasting is by NO means exact, however has come a long way over the past 20 years. but to expect to be able to predict exactly where an intense hurricane will make landfall months in advance, or for that matter, before it has even developed, im not convinced is even possible with numerical modeling. and IF it is, i dont think any of us will see it in our life time.

numerical models are great for making an assessment of where a tropical cyclone may make landfall, and how it may track, but they are by no means perfect.

we can use analog methods and climate factors to try to determine what parts of the east coast or the gulf coast might be under the gun based on systems in those years, which is the best method we have going currently. at times it works out, other times is does not.

I agree with pretty much all of your comments. IMO, climatology will solve most of our forecasting issues. But we probably need much more time for climo stats to evolve before seasonal forecasting becomes an EXACT science.

your basically right, but even still i dont think seasonal forecasting, be it for the tropical season or northern hemisphere winter will ever become an "exact" science. why is that? because we still dont completely understand how all climate factors relate to and effect one another, not to mention factors we still have yet to discover.

its mostly the indirect relationships between climate factors which are seemingly distant and non-connected that we have the hardest time making an assessment of. so until we are able to un-cover all the different relationships both direct and indirect between climate factors we WILL NEVER see a "perfect" seasonal forecast.

[FLguy]

what we must understand however is that all climate factors here on earth are in some way connected to one another, and either directly or indirectly effect one another. the same goes for the solar cycles, which IMO are a much larger influence on global climate patterns than any amounts of fossil fuel we burn, or for that matter any known climate factor here on earth.

the principal behind this drawn back to the route of the reason why our climate is the way it is here on earth. that of course is because of the sun. therefore, it is safe for one to assume that it is the most important feature impacting global climate patterns at all times.

[cycloneye]

Nothing is perfect in this world and making a seasonal forecast is not the exception but there has been progress on those forecasts as new tools haved comed to help the forecasters to make more precision seasonal forecasts like what Dr Gray does and he has for the most part correct on his outlooks in terms of the patterns that are going to be in the atlantic and in the pacific.About landfall % that part is more difficult to forecast and to pinpoint an area that will have a hurricane landfall but as Dr Gray points out he generalizes not says a citie nor an island in particular as he looks at the pattern shaping up at the time of the outlook comes out.Yes the numbers are important but the patterns every year are more important and that is what we have to be more focused on.In other words giving numbers for a seasonal forecast is great but it is not the most important part of an outlook.

[Derek Ortt]

Weather will NEVER be able to be predicted more than 2 weeks in advance as the equations of the atmosphere cannnot be solved precisely by the models, they are approximated. By the end of 14 days, the errors have simply grown too large

[KatDaddy]

Even predicting weather across a region with current data can be a bust. Once the current data is applied to upcoming days the forecast error becomes extremely large. Just look at all the MRF discrepansies that occur on daily runs. There are just way to many variables to be defined.

Perhaps if we have surface sensors placed 1 mile apart across 100s of miles and layered upward every 1000ft to 30,000ft, the models may have enough data to process a forecast correctly. At this point a regional forecast for that day may be close to 100% Again I say may be close to 100%. Forecast contain many predictions for a single day such as low temp, high temp, precentage of rain, sky conditions, and wind speed and direction. These weather conditions can be affected by many micro-scale affects.

Additionally the senors must record data in realtime for model input. The model runs need to issued at least 6 times a day based on the realtime data. The output of a group of model runs will then need to be input in another model equation and run to get even more precise model run outputs. Even this will likely not be enough as new models with more variance would need to be developed to handle all the data input.


Making any forecast for a season is just a guess and unreliable. Notice how vague Dr. Gray is with his precentages (X% for East Coast). Very general statement for a large coastal area.

If someone is correct for landfall precentages for a specific coastal area they are lucky.

I agree with Derek that currently we cannot predict weather 2 weeks today and will likely never be able to make such forecasts. The data being input from sensors can change dramatically from day to day. This would apply to data at different levels of the atmosphere changing as well as.

Pattern recognition does hold some importance and may yield some overall insight to an upcoming season but has way to many local variables which would be applied to seasonal variables and pattern variables. Again another meteorological equation nightmare.

A very tough job forecasting the weather. What is sad is the fact the majority of the Worlds population have no idea what weather forecasting really consists of and how far we have come.

[FLguy]

Weather will NEVER be able to be predicted more than 2 weeks in advance as the equations of the atmosphere cannnot be solved precisely by the models, they are approximated. By the end of 14 days, the errors have simply grown too large

which is why pattern recognition is ALMOST ALWAYS more effective in meadium range forecasting, than just taking numerical model output at face value. and analoging is much more effective in seasonal forecasting as compared to staring at a long-range climate model. case in point, the AGCM's preformance prior to the 2001-02 winter. which was laughable at best.

the GFS (AVN/MRF) is a better example of the inconsistencies in meadium range model preformance. beyond 240hrs, the output is for the most part useless.

And the long range model data is way useless because for the same reason you said FLguy.