Wednesday, December 8, 2010

A 5-Year 10-Meter Es Propagation Study Using PropNET - Part 3

The Spring/Summer “Es” Season:
When the first individual season study was first begun in 2005, my opinion based on experience was that for 10-Meters, the Spring/Summer Es season occurred from about May 1 till August 11. Once I compiled the data for 2005, I determined that I was slightly incorrect. The season begins a few days earlier and ends many more days later. I extended the data gathering efforts from April 21 till August 15 for the study. I should have extended the end date of each seasonal study much later (2 weeks at least), but other obligations, such as pursuing a college degree kept me from doing so efficiently for the first four years. I still was active many of those days and did retain my data. To make up for the lack of activity, I used data primarily from John Ainsworth, N5XYO in DM90 to complete 2 more weeks of probability analysis to help better define the end of the Es season.

The Purpose of the Study and What it Shows:
I do not pretend to have any solid answers to what causes Es. Many others, most notably Emil Peacock –W3EP, Patrick Dyer – WA5IYX, J.A. Pierce – W1JFO, and Melvin Wilson – W1DEI/W2BOC investigated and documented the phenomena. Many others continue to analyze it today. But after my own operating experience and documentation, I began to believe that most were concentrating on the answers to its creation and that maybe there was some bias towards the normality and predictability of what Es will display.

The purpose of the study was to show that Es was a very natural and predictable propagation phenomenon, and that there was nothing too magical about it. I researched this on a mathematical, not a scientific basis. Finally, it was incorrect for it to be called “Sporadic”. I believe that this study does accomplish it. Es propagation of this nature should have a totally different description. Along the way I made some interesting observations that should create forums for future discussions. I believe that the intensity of Es can influenced from outside elements. But, in no way should these outside elements have any influence on the central point of the study in that seasonal Es are predictable.

The study is broken down into the following parts:
1. Numerical activity, such as statistics daily, weekly, and hourly during the season.
2. Distance analysis hourly by segments.
3. Directional analysis hourly throughout the season.
4. Probability analysis by weekly segments, by week, by hourly segments, and hourly.
5. Spring/Summer Es Seasonal Calendar.

I want to thank everyone involved with the PropNET organization. Without the founders, programmers, forward-thinkers, and its participants, a study of this magnitude could not have been accomplished.

Solar Conditions During the 5-Year Study:
I have experienced on many occasions that solar conditions will have both an adverse and positive effect on Es conditions. Most notably, I have observed enhanced Es thanks to the beginning of a minor magnetic storm from a coronal mass ejection (CME). Some of my best late season 2-Meter Es events were due to these events. On the other hand, many excellent Es events on 10 and 6 Meters were abruptly ended by solar flares.

The 5 years I studied could not have been more perfect as we have experienced one of the most prolonged bottom period of a solar cycle. Solar flux was never high enough to support F2 propagation. MUF never was higher than 21 MHz. These low levels of solar activity kept disturbances to a minimum, therefore having only a minor impact. Only once could I determine a prolonged effect, and it was very early in the season (late April-early May 2008).

The following charts show average Solar Flux and the Ap indices for the 5 year period of the spring/summer Es season:

From 2005 to 2008, solar flux steadily declined, 2009 increased slightly. The Ap indices declined in succession for all five years. Solar flux was for monthly averages were between 90 and 100 for only 4 months in the 25 months measured. Overall it averaged at 80 or below for 4 of the 5 years.

The Spring/Summer Es Season, From its Beginning to the End:

The chart represents the actual number of 10-Meter PropNET PSK31 captures at my location (North Central Texas) for each day from April 20 till August 15, 2005–2009. At first it was difficult to see any clear trends using daily numbers, but after 5 years it finally called attention to several factors related to Es propagation. I was startled to see specific days being so active and other days close to those peaks were fairly quiet. Each annual season measured in this study was 131 days (concentrated data on 114), in which nearly 20 days were extremely active. The other days measure show a steady increase, a peak, and slow decline.

To more clearly demonstrate trends and determine peaks, I used a daily averaging approach. It was begun in the first year of analysis (2005) in order to smooth out the daily trends and focus activity to specific dates of the season. The following chart shows the average number of daily captures 6 days prior to and including the day charted. Rather than using actual dates, the chart indicates the number of days prior to and after the Summer Solstice (6/21). The chart better displays concentrations of Es propagation. The chart shows that from 22 to 12 days prior to solstice (5/31 to 6/09, 6/06 peak) was the most active period for Es. In the latter years of the study, a peak developed just prior to the Summer Solstice. During the season there are seven apparent concentrations of Es activity.

The events of a typical season show:
1. A quick rise in activity seven weeks prior to the summer solstice.
2. The absolute peak about 2 weeks prior to solstice.
3. A distinguishing lull approaching the solstice.
4. A second peak near the solstice.
5. A gradual decline throughout the remainder of the season along with occasional bursts.

Rather than a perfect bell shaped curve, activity appears to be skewed to the right and is right-tailed. The median (midpoint) of the total captures in this study occurs early in the day of 6/22, just one day after the Summer Solstice. The “median” capture point also proves that Es propagation is a seasonal phenomenon.

Trend Line:
Trends appeared to be prevalent although captures varied greatly for day to day. A polynomial regression analysis was applied to the 5-year daily average captures. Although not perfect, a good trend line was established for a two degree (x²) polynomial. Raising the degree did little improvement to the coefficient of determination at .558 (measured scale from 0 to1).

The trend line indicates that the beginning of the season was on April 26. The peak of approximately 167 captures occurs on June 23. The season trends towards an August 22 end. The trend line shows activity is at least 50% of peak from May 14 till August 5. Activity is 90% of peak from June 5 to July 13. Again, the trend line indicates a steady early rise and late steady decline, with a skewing towards the latter half of the season.

Actual, 7-day Average, Polynomial Trend

Another method to view Es activity was to chart the captures it into one week segments. The following chart shows total captures in one week periods from 8 weeks prior to and 8 weeks past the Summer Solstice. In the first years, the 6th week of the season was the best. There was right skewing of the data during that time. In the later years of the study, the 8th through the 11th weeks rapidly caught up in total. The Summer Solstice occurs the 2nd day of Week 9. A lull of total captures in weeks 7 and 9 was noticeable in the early years. It was not until 2008 that the 8th week became the most active of the season. One notable trend each year was an increase in activity beginning with Week 14 and Week 16. For this reason, I should have extended the study at least two more weeks, but will account for it in the probability analysis.

No comments: