Predicting the Conditions of the Next Season Using Current Conditions

Seasons are defined as periodic changes of temperature and precipitation for a given location. These changes occur due to variations in the solar radiation that is received as the Earth revolves around the sun. Due to the elliptic orbit and angle of the Earth's tilt, locations at mid-latitudes experience the greatest noticeable change from season to season (Hidore and Oliver, Page 37). While there are great differences between alternating seasons (winter vs. summer), the same seasons (summer vs. next summer) tend to display similar trends. Small fluctuations occur from year to year and are taken into account to produce average seasonal conditions for each location. There is great interest as to whether an upcoming season will be warmer or colder, or perhaps wetter or drier than normal. Slight changes that wouldn't even be noticed by the general public can have devastating effects on farming, ecosystems, health, and property. Temperature increases of only a few degrees have lead to the destruction of marine habitats, and the spread of disease through insects (Hidore and Oliver, Pages 298-299). While it is immensely complex to understand and predict future conditions, the demand is still there. What if there was a simple, almost elementary way of predicting the next season? It would be known if conditions would be close to normal, or if there were to be devastating deviations that could be prepared for. Forget analyzing atmospheric pressures, ocean temperatures, wind patterns, and the hundreds of other complex variables. What if an upcoming season's conditions could be determined solely be analyzing the preceding season? If there turns out to be a correlation, then there must be some underlying mechanism responsible. This will be discussed at a later time; for now the focus should be on potential correlations.

The Experiment

First, an area of study needs to be established. The cities that will be tested will be Binghamton, NY, Scranton, Pa, and Syracuse, NY. All of these locations are mid-latitude cities in the Northern Hemisphere within a few hundred miles of each other. This will limit variables pertaining to differences in geography, climate, and general wind flow. Next is the time period for study. Data from 1971-2000 will be used to evaluate a possible correlation in summer and winter conditions. For the summer months, June, July and August will be used, whereas December, January and February will be used for winter. For each location, an average temperature will be established for each "season". This will be done by taking the average of the mean temperature (1971-2000) of the three months in question. Then the actual average temperature for those three months for each year will be compared to the thirty year average, resulting in a departure from normal value. When this has been done for every year, a correlation will be made between departures from the summer months to the winter months. An index will be assigned to each set of data, representing the strength of any potential correlations. This index, named the Visco Index, will have a value that ranges from -30 to 30. Each season that exhibits similar conditions to the previous one will receive a value of +1, whereas if they are opposite, -1. The final index will reveal the strength and nature of the correlation, with a value of zero representing no predictable seasonal pattern. The same thing will be done for precipitation amounts. Additionally, a correlation will be made comparing winter temperatures and precipitation to the following summer, to see if there is any change in noticeable patterns. There are over 1,000 data points that need to be taken into account, so extreme care must be taken to ensure accurate results. Once a correlation, if any, is established the focus will turn to causes.

The Results

From all the calculated data, a few notable trends can be observed. Beginning with Scranton, PA, there is a minimal correlation for temperature. In terms of temperature, the index was recorded to be -2, which would lead to a slightly higher probability of the following winter being the opposite deviation. If looked at with winter occurring first and affecting the following summer, the index was 0, meaning there was absolutely no correlation. Precipitation however produced an index of -6 for summer vs. next winter. For winter vs. summer, the index was 0. This means that the winter season is 60% likely to oppose the precipitation patterns of the previous summer, while the winter has no dominant affect on the next summer. Next up is Binghamton, NY.

In terms of temperature, Binghamton showed a fair correlation supporting similar conditions from season to season. For summer vs. winter, the index was 4, while winter vs. summer had an index of 6. While not overwhelming, they are both 6 points higher than the corresponding data for Scranton, PA (For some perspective, that's 20% more likely). Precipitation for Binghamton was a real stand out, not in terms of a commanding pattern, but rather the drastic differences found from taking different approaches. If looking at a summer vs. winter correlation, the index was -4, supporting a fair reversal in precipitation amounts from season to season. However, if the data was studied in the opposite fashion, winter vs. summer, the index was 6. One direction involves a 56% chance of opposite patterns, while the other has a 60% of similar. This difference was the most astonishing for the entire experiment. Syracuse NY demonstrated fairly consistent results. For temperature, summer vs. winter produced an index of 6 (60% likely to repeat), and winter vs. summer produced an index of 8 (63% likely to repeat; highest magnitude index of all data). Precipitation data also yielded fair correlations in favor of repeating conditions. For summer vs. winter, and winter vs. summer the respective indices were 6 and 4.

Some additional trends:

There were a few interesting trends to be found within the data. Beginning with temperature, the correlation of similar seasons increased as we moved from Scranton, to Binghamton, to Syracuse, moving north and closer to Great Lakes (connection?). Additionally, all three stations had an index for winter vs. summer that was 2 points higher than summer vs. winter. This could lead to the conclusion that winter seasonal temperatures have a greater affect on summer than visa versa. Another aspect that was examined was the nature of the similar correlations. For Scranton, 43% of the years that recorded winters similar to the previous summers had conditions warmer than normal. For Binghamton, that same figure was just over 70%. Even more, Syracuse had a figure of 72%, again displaying a trait of more distinguished patterns the further north and closer to the Great Lakes a location is.

The precipitation results were more scattered. Binghamton and Scranton both had discrepancies in the correlation depending on what season was considered the starting point. Syracuse was more consistent, but it isn't enough to declare a pattern.

Conclusions

With this basic data it is apparent there are higher probabilities of having temperature conditions repeat themselves from season to season, especially closer to the Great Lakes. Syracuse is within 50 miles of Lake Ontario, while Binghamton is about 100 miles, and Scranton 150. But is that the reason? Could the proximity to the water have such a profound affect? "The great thermal inertia of the lakes also generally acts to reduce the diurnal and annual temperature ranges over and in the vicinity of the lakes" (Bates, et al. 1993). It takes longer for water to lose or gain heat in comparison to land due to the high specific heat of water. This creates a temperature lag for the water in relation to the surrounding areas. For example, as winter approaches and the temperature decreases rapidly, the water stays considerably warmer for longer. The opposite is true for summer, where water remains cooler than its surroundings. Perhaps if there was a warmer summer, the waters would be that much warmer reaching into the next winter. This could have an affect on the surrounding area's temperatures. The same is true if the water was colder during the summer. During the winter the water would become colder faster. While this could affect the area, the warmer water scenario would appear to have a higher impact, which the data shows (70%+ of same conditions seasons were warmer than normal for both locations closer to Great Lakes). Another effect warmer waters can have is increased precipitation as a result of lake-effect snow (Burnett, et al. 2003). With the lakes remaining unfrozen longer, more condensation and precipitation can occur. This increase in the latent heat process could potentially result in a warmer atmosphere.

There have been numerous studies that have found that "northern lakes can significantly influence the local water and energy cycles" (Long et al. 2007). It certainly appears the city with the closest proximity to the Lakes has the greatest instance of seasonal correlation. Whether the lakes are in fact the cause of seasonal patterns cannot be determined from this experiment. Additional research needs to be completed in order to fully understand the mechanisms and effects of sizeable water bodies. Other mechanisms that should be explored related to this topic could include patterns in solar radiation, sea surface temperatures and their respective oscillations, global events such as volcanic eruptions, global warming, topographical differences, surface albedo, and general wind patterns. More cities should be tested from a wider area to see if any correlations exist. For the cities that were tested for this experiment, there was a fair correlation of seasonal temperatures mimicking the previous season. For precipitation, the results were too scattered to make an interpretation. One city had conditions that repeated, while another had opposite precipitation amounts, and the third had both.

While this experiment provided basic seasonal patterns, the correlations are not set in stone. If a final conclusion had to be made, the upcoming season will more likely exhibit similar conditions to the present one, especially in terms of temperature.

References

Bates, G.T., F. Giorgi, and S.W. Hostetler, 1993: Toward the Simulation of the Effects of the Great Lakes on Regional Climate. Mon. Wea. Rev., 121, 1373-1387.

Burnett, A.W., M.E. Kirby, H.T. Mullins, and W.P. Patterson, 2003: Increasing Great Lake-Effect Snowfall during the Twentieth Century: A Regional Response to Global Warming? J. Climate, 16, 3535-3542.

Hidore, J., and J. Oliver, 2002. The Seasons. Climatology. Page 37

Hidore, J., and J. Oliver, 2002. Evidence for Global Warming. Climatology. Pages 298-299

Long, Z., W. Perrie, J. Gyakum, D. Caya, and R. Laprise, 2007: Northern Lake Impacts on Local Seasonal Climate. J. Hydrometeor., 8, 881-896.

National Weather Service. Updated March 18, 2008. Binghamton Weather Forecast Office. Local Climate Data. http://www.erh.noaa.gov/bgm/climate/bgm.shtml

Visco, T. 2008. Seasonal Predictions. Temperature and Precipitation. Binghamton, NY. Scranton PA. Syracuse, NY.