The hunt is now on more fiercely than ever: with meteorology growing and advancing at exponential rates in the last few decades to the present day, researchers are working harder than ever to find patterns in the weather to accurately predict long range conditions, months in advance. One aspect of this hunt is being able to predict the upcoming winter using conditions observed in the fall, or even summer. We'll take the time today to discuss methods that may or may not be useful in such long range predictions.
1. Nature Signs
This is a personal favorite of mine, due to the mystery surrounding it. Many weather enthusiasts, and people in general, believe that nature can point us to the intensity (or lack thereof) of the upcoming winter. I've heard mentions of trees producing anomalous fruit in the summer, animals acting in a suspicious manner, and even the color of certain shrubs and plants in the preceding summer/fall. I've even taken it upon myself to monitor roadkill anomalies in the fall, wondering if increased (decreased) animal activity in the fall may indicate a more (less) intense winter ahead, hence the above normal (below normal) roadkill sightings.
The unfortunate truth is that we don't know if these methods work, and likely will never know. The concept is very similar to us as a human race. We monitor certain trees and animals for signs of oddity, but that's like monitoring a certain person for an oddity. It doesn't work, because each person is different than the person next to them. As a race, humanity is quite similar across the board. The same goes with certain species of trees, animals, etc. But we can't observe a few trees and/or animals and make a qualified deduction from it. In order to determine any possible connection between nature anomalies and winter, a large-scale (covering a plethora of states) operation would likely be needed to monitor great masses of the same type of animal/tree, and then attempt to make a reasonable deduction on any winter prediction abilities.
2. Six-Month Separation
This method involves monitoring conditions in June-July-August to see if conditions six months later (December-January-February) are correlated. The method likely stems from the question of cyclical patterns repeating on a regular, monthly scale; in this case, half a year. One might believe that Arctic 500mb heights in the summer may be related to Arctic 500mb heights that following winter, or the same scenario, except over North America, etc.
The truth behind this method is that it does not work a heavy majority of the time. I did some research last winter, where I compared monthly Arctic Oscillation (AO), North Atlantic Oscillation (NAO), and Pacific-North American (PNA) index values from June-July-August to values for the following December-January-February to determine any correlation. Although I no longer have the exact values with me, I can recall there being little to no correlation between any of the indices' summer and winter values. Looking back, I believe the PNA actually ended up seeing a near-perfect split of 50% of years recording a positive summer PNA and negative winter PNA, and 50% of years recording a positive summer PNA and positive winter PNA, or negative rather than positive for both timeframes. The Arctic Oscillation may have had the "highest" anomaly, with the split coming in at 40%-60% when comparing number of years with positive summer/winter value correlations and negative summer/winter value correlations. To make a long story short, this method is not one worth using.
So far, we've discussed items that either don't work in long range forecasting, or their effectiveness has gone undetermined. Let's now go over some items that do have merit in the long range forecasting field.
3. Sea Surface Temperature Anomalies
Long range forecasters often look to the oceans for their outlooks. Multiple oscillations reside in oceans all around the world, with many of them spreading their impacts on a global scale. Forecasters tend to believe that these long-term oscillations, which may remain in their same state for months or years at a time, can predict winter conditions as far out as summer.
There is definitely some truth to this claim. Persistent SST anomalies in a certain part of the globe can allow forecasters to get a glimpse at the weather for multiple months ahead. Long-term oscillations, like the Pacific Decadal Oscillation (PDO) or Atlantic-Multidecadal Oscillation (AMO), can remain in their same positive or negative phase for years, or even decades, at a time. Some shorter-term oscillations, like the El Nino-Southern Oscillation (ENSO) phenomenon, can still allow for long range outlooks, valid for months in advance.
4. Soil Moisture & Drought
This theory stems from the concept of feedback loops, which entail that something already in a bad situation gets continually worse, because that something is harming the things trying to improve the situation. In our case, we can apply this to drought. If a drought forms, the soil is anomalously below-normal in soil moisture, making the earth dry. As a result, clouds cannot form, meaning rain cannot fall. This makes the drought worse, and makes for a vicious cycle that continues the drought. Storm systems that move over the drought-affected region cannot produce as much rain as in other regions, because the storm can't draw water from the dry soil. Similarly, areas with above normal soil moisture may see rain continue, as more moisture is available in the soil than is necessary. Thus, the air becomes unusually humid, and any storms that form over this region can produce heavy rainfall, keeping the soil on a wet level.
This method of forecasting is more of a hit-or-miss method. This is because, while the theory of these feedback loops is sound (to an extent), a plethora of external forces, made up of oscillations in the oceans and upper atmosphere, can shift and change to bring rain to drought areas, or deprive precipitation to rain-soaked land. The atmosphere is one big, constantly-changing chess board, if you will, and some pieces happen to be weaker than others. While this drought/soil moisture method of long range forecasting can work, other factors must be taken into account as well.
We've only reviewed a few of many long range patterns known in the world of weather, but hopefully this gives you an understanding of some aspects of long range forecasting. While some methods don't tend to work well, others have experienced a decent amount of success. The hunt for a long range predictor with considerable accuracy will continue into the foreseeable future.
Andrew
1. Nature Signs
This is a personal favorite of mine, due to the mystery surrounding it. Many weather enthusiasts, and people in general, believe that nature can point us to the intensity (or lack thereof) of the upcoming winter. I've heard mentions of trees producing anomalous fruit in the summer, animals acting in a suspicious manner, and even the color of certain shrubs and plants in the preceding summer/fall. I've even taken it upon myself to monitor roadkill anomalies in the fall, wondering if increased (decreased) animal activity in the fall may indicate a more (less) intense winter ahead, hence the above normal (below normal) roadkill sightings.
The unfortunate truth is that we don't know if these methods work, and likely will never know. The concept is very similar to us as a human race. We monitor certain trees and animals for signs of oddity, but that's like monitoring a certain person for an oddity. It doesn't work, because each person is different than the person next to them. As a race, humanity is quite similar across the board. The same goes with certain species of trees, animals, etc. But we can't observe a few trees and/or animals and make a qualified deduction from it. In order to determine any possible connection between nature anomalies and winter, a large-scale (covering a plethora of states) operation would likely be needed to monitor great masses of the same type of animal/tree, and then attempt to make a reasonable deduction on any winter prediction abilities.
 |
| Squirrels are commonly looked to for their ability to predict the upcoming winter, with observers monitoring a squirrel's effort to collect nuts earlier (or later) on in the year. |
2. Six-Month Separation
This method involves monitoring conditions in June-July-August to see if conditions six months later (December-January-February) are correlated. The method likely stems from the question of cyclical patterns repeating on a regular, monthly scale; in this case, half a year. One might believe that Arctic 500mb heights in the summer may be related to Arctic 500mb heights that following winter, or the same scenario, except over North America, etc.
The truth behind this method is that it does not work a heavy majority of the time. I did some research last winter, where I compared monthly Arctic Oscillation (AO), North Atlantic Oscillation (NAO), and Pacific-North American (PNA) index values from June-July-August to values for the following December-January-February to determine any correlation. Although I no longer have the exact values with me, I can recall there being little to no correlation between any of the indices' summer and winter values. Looking back, I believe the PNA actually ended up seeing a near-perfect split of 50% of years recording a positive summer PNA and negative winter PNA, and 50% of years recording a positive summer PNA and positive winter PNA, or negative rather than positive for both timeframes. The Arctic Oscillation may have had the "highest" anomaly, with the split coming in at 40%-60% when comparing number of years with positive summer/winter value correlations and negative summer/winter value correlations. To make a long story short, this method is not one worth using.
So far, we've discussed items that either don't work in long range forecasting, or their effectiveness has gone undetermined. Let's now go over some items that do have merit in the long range forecasting field.
3. Sea Surface Temperature Anomalies
Long range forecasters often look to the oceans for their outlooks. Multiple oscillations reside in oceans all around the world, with many of them spreading their impacts on a global scale. Forecasters tend to believe that these long-term oscillations, which may remain in their same state for months or years at a time, can predict winter conditions as far out as summer.
There is definitely some truth to this claim. Persistent SST anomalies in a certain part of the globe can allow forecasters to get a glimpse at the weather for multiple months ahead. Long-term oscillations, like the Pacific Decadal Oscillation (PDO) or Atlantic-Multidecadal Oscillation (AMO), can remain in their same positive or negative phase for years, or even decades, at a time. Some shorter-term oscillations, like the El Nino-Southern Oscillation (ENSO) phenomenon, can still allow for long range outlooks, valid for months in advance.
 |
| SST anomalies are often believed to be one of the best & most reliable predictors of long range weather. |
This theory stems from the concept of feedback loops, which entail that something already in a bad situation gets continually worse, because that something is harming the things trying to improve the situation. In our case, we can apply this to drought. If a drought forms, the soil is anomalously below-normal in soil moisture, making the earth dry. As a result, clouds cannot form, meaning rain cannot fall. This makes the drought worse, and makes for a vicious cycle that continues the drought. Storm systems that move over the drought-affected region cannot produce as much rain as in other regions, because the storm can't draw water from the dry soil. Similarly, areas with above normal soil moisture may see rain continue, as more moisture is available in the soil than is necessary. Thus, the air becomes unusually humid, and any storms that form over this region can produce heavy rainfall, keeping the soil on a wet level.
This method of forecasting is more of a hit-or-miss method. This is because, while the theory of these feedback loops is sound (to an extent), a plethora of external forces, made up of oscillations in the oceans and upper atmosphere, can shift and change to bring rain to drought areas, or deprive precipitation to rain-soaked land. The atmosphere is one big, constantly-changing chess board, if you will, and some pieces happen to be weaker than others. While this drought/soil moisture method of long range forecasting can work, other factors must be taken into account as well.
We've only reviewed a few of many long range patterns known in the world of weather, but hopefully this gives you an understanding of some aspects of long range forecasting. While some methods don't tend to work well, others have experienced a decent amount of success. The hunt for a long range predictor with considerable accuracy will continue into the foreseeable future.
Andrew
1 comment:
Someone should just create a gigantic Bayesian model, with a huge dataset of weather station records for model training stretching back 100+ years. I would be curious to know of the success of this type of model in predicting the patterns for the next 6 months of weather.
Post a Comment