2016 Perseid Meteor Shower

A screen shot from the 16th of August 2016 - The Geminid Meteor shower putting on a display

The data from 2016 is incomplete and was captured during the calibration process undertaken when the LVST was set up. The data for the 2016 Perseid shower covers the period from the 4th to the 16th of August. Ideally a longer period from say the 25th July to the 25th of August would have better shown the increase and decrease in meteor activity to and from maximum.

Summary of findings

August 2016 LVST Data: Summary of Observations and Interpretation of data analysis

Seasoned visual meteor hunters are used to scanning the sky in the early hours of the morning to see a handful of meteors between midnight and dawn. Meteor storms, in which literally hundreds of bright ‘shooting stars’ per hour fall from the sky, are at best, a once in a lifetime experience.

Once the LVST went on line in 2016, all members of the Jodrell Plank Observatory Team and Comet the Observatory cat were amazed by the constant level of meteor activity being detected by the very basic equipment being employed.

Over a period from the 4th to the 16th of August 2016, the amount of data generated by the LVST was quite astonishing – 32,800 reflection observations were recorded.

In order to better investigate the ‘Perseid Shower’ meteor trail reflections, separating shower data from the constant background level of non-shower associated activity became a priority. As the 2016 Perseid shower data was the first to be analysed by the team, a number of iterative methods of enquiry were adopted to both test the efficacy of the procedure and the sensibility/validity of the results.

The Analytical Process Adopted

  • Once extreme data outliers were removed, that is observations having a duration > than 720 cycles (approx 6 cycles =1 sec), what data remained was used as the source information data set containing 32,729 observations.
  • With such a large data set, visual sorting of individual trails was determined impractical and statistical methods of filtering source data to create data subsets was adopted.
  • The accepted classification of meteor trails was adopted ie. Separation of observations into Over dense O, Under dense U and Transition T trail observations.
  • The source data set of all observations was then the U+O set (containing T)
  • Following investigation (see page Separating U, O and T trail reflection observations), it was decided to use ‘duration of meteor trail’ as a measure by which to filter the U +O set to create a subset O-T. This contains meteor trails detected that have a duration = or > 18 cycles. This subset is likely to exclude 99% of U reflection observations but will include; Perseid shower trail O reflection observations, other shower trail reflections, sporadic meteor trail reflections and non meteor related noise.
  • Filtering the O+U set by duration < 18 created a new subset U+T. This data subset is likely to contain all U and T reflection observations for all shower and non shower meteors and include some non meteor related noise.
  • Detailed analysis of the U+T subset, see the graph logmaxpowerxduration (a Energy) plotted against Radial Velocity V, indicates a maximum value for log maxpowerxDuration =3.15 below which the standard deviation from the mean radial velocity suddenly increases.
  • As T is the subset O of reflections that have short durations like U but similar velocity and power profiles to O, the Jodrell Plank Observatory Team decided to filter the U+O source data by log maxpowerxduration to leave only observations with a logmaxpowerxduration =>3.15. The new subset created is being interpreted as O+T.
  • The relationships between the day date and the time on which observations were made for the O+T subset, particularly with regard to radial velocity and MaxpowerxDuration, were regarded as very useful in interpreting the meteor shower profile. The use of hexagonal binned data plots was considered effective in presenting a big picture view of a large data set.

Observations and Interpretations

Observations:

  • General activity: (based on the largest data set O+U – 32,729 observations)
    • Overall activity reflection observations peaked at 4:00 UT on the 6th August 2016
    • Between the 4th and 16th August reflection observations were consistently higher between 0:00 and 17:00 UT than between 18:00 and 22:00 UT.
    • The mean radial velocity (Line of sight with the LVST) was +379 metres per sec with a standard deviation of 218.8.
  • Shower activity: ( based on subsets of data O-T - 1547 observations and O+ T - 16950 observations)
    • O-T reflection observations peaked at 5:00 UT on the 10th August 2016.
    • Between the 4th and 16th August O-T reflection observations were consistently higher between 3:00 and 8:00 UT than at any other time and were very low between 12:00 and 22:00 UT.
    • The mean radial velocity of O-T reflection observations was +385.7 metres per sec with a standard deviation of 105.8.
    • O+T reflection observations peaked at approximately 18% between 2:00 and 8:00 UT and possibly at 5:00 UT on the 6th August 2016. Between 11.62 and 12.6 % were recorded on the 10th of August. Between 10.4 and 11.34 % were recorded on the 7th of August. The minimum number of observations - between 2.92 and 3.44 % occurred on the 4th August.
    • The mean radial velocity for O+T reflection observations (Line of sight with the LVST) was +366.8 metres per sec with a standard deviation of 103.6.
    • For O+T set, on the 7th and 14th of August, there were observations clustered around radial velocities (+109 and +69 m/sec respectively) significantly lower than the mean radial velocity for the O+T set.

Interpretations and Speculation:

  • Activity associated with the 2016 Perseid Shower substantially commenced on the 5th of August and reached its peak in the early hours of the 6th of August. Activity remained high on the 7th August but was falling to almost similar levels as had been experienced on the 4th. The 8th of August presented a rising level of activity which peaked on the 10th of August at lower levels than had been experienced on the 7th. Apart from a small rise in activity on the 14th Aug, levels of activity fell back towards those experienced on the 4th.
  • The increase in reflection activity observed on the 14th is not thought to be associated with the Perseid shower. This is conjecture based on observed differences in radial velocity ( and that meteors from showers take parallel paths) - By all accounts, Comet the cat finds this more difficult to swallow than Mr. Shrodinger's goldfish!
  • The fluctuations in activity (ie. number of observations) are possibly proportional to the fluctuations in shower dust density. The Earth, in its annual path around the Sun, probably passed through the densest part of the dust from Comet Swift-Tuttle on the 6th and 7th of August 2016.
  • The number of observations O-T having a duration =>3secs peaked on the 10th of August between 4:00 and 5:00 UT. Enhanced activity was observed on the 6th 9th and 11th although the number of observations was significantly lower than on the 10th.
  • The larger pieces of matter in the cometary stream stream, were probably encountered by the Earth on the 10th of August. This hypothesis is based on a relatively small data sample of observations (O-T sub set -1547 reflections and associating increased reflection duration with increased meteoroid mass).
  • The O+T data set contains reflection observations from at least one other meteor shower with increased activity being recorded after midday on the 7th and 14th of August. These observed reflections, sharing similar radial velocity and hours of the day, may be associated with the Capricornid, Alpha Capricornid or Delta Aquarid meteor showers which are active in July and August each year.

The Detailed Data Analysis and Data Plots

The graph of all reflection observations O+U plotted against the (24 hour) day date indicates :

  • Overall the period during which data was collected 32,729 reflections were recorded.
  • With 95% confidence:
  1. The maximum number of observations - between 18 and 19% - were recorded on the 6th August.
  2. Between 10.8 and 11.5% were recorded on the 7th August.
  3. Between 9.4 and 10% were recorded on the 10th of August.
  4. Between 6.25 and 7.37% were recorded on the 5th 8th 9th 11th and 14th of August.
  5. Between 5 and 6.37% were recorded on the 12th 13th 15th and 16th August.
  6. The minimum number of observations - between 3.3 and 3.7% was recorded on the 4th August.


The graph of all reflection observations O+U plotted against the hour of the day (UT) indicates with 95% confidence :

  • Two max peaks of observations:
  1. Between 2:00 and 7:00 UT
  2. Between 13:00 and 14:00 UT
  • A deep minimum between 18:00 and 22:00 UT


The graph of all reflection observations O+U plotted against the radial velocity V indicates:

  1. 50% of all observations are positive and have a radial (line of sight) velocity component moving towards the LVST receiver at between +286 and +446 metres per sec.
  2. The median velocity is + 374 metres per sec.
  3. The mean velocity is = +379 metres per sec. (SD=218.8)
  4. The maximum velocity away from the LVST receiver is -102 metres per sec.
  5. The maximum velocity towards the LVST receiver is +948 metres per sec.

The graph of O+U log duration plotted against radial velocity is interesting.

Observations and Speculations:

  • For log duration = or > 1.0 - as radial velocity increases beyond the mean value of +379 metres per sec the observations become bluer (captured between 8th and the 12th August).
  • Log duration against Radial velocity roughly follows a Gaussian distribution centred upon the mean radial velocity.
  • For log duration = or < 1.0 - as radial velocity decreases below the mean value of +379 metres per sec the observations become greener (captured between 4th and the 7th August).
  • There appears to be a spike of green and mauve coloured observations centred approximately on V = +100 metres per sec. Could this be an indication of a meteor stream other than Perseids?


The graph of reflection observations O-T plotted against the (24 hour) day date indicates :

  • Overall the period during which data was collected 1,547 reflections were recorded. This is only 4.7% of the total number of reflections observed and represents only those reflections having a duration =or>18 cycles (approx 3 secs).
  • With 95% confidence:
  1. The maximum number of observations - between 27 and 32% - were recorded on the 10th August.
  2. Between 10 and 15.5% were recorded on the 6th 9th and 11th of August.
  3. Between 4 and 9.5% were recorded on the 7th 8th and 12th of August.
  4. The minimum number of observations - between 2 and 5% occurred on the 4th 5th 13th 15th and 16th of August.


The graph of reflection observations O-T plotted against the hour of the day (UT) indicates with 95% confidence :

  • One max peak and one trough of observations:
  1. Peak between 3:00 and 8:00 UT
  2. A deep minimum between 12:00 and 22:00 UT


The graph of reflection observations O-T plotted against the radial velocity V indicates:

  1. 50% of all observations are positive and have a radial (line of sight) velocity component moving towards the LVST receiver at between +343 and +423 metres per sec.
  2. The median velocity is + 380 metres per sec.
  3. The mean velocity is = +385.7 metres per sec. (SD=105.8)


The graph of O-T log duration plotted against radial velocity.

Observations and Speculations:

  • For radial velocity < +300 or > +500 m/sec the density of observations decreases. There remains an increased density of green observations (captured on days 4 to 8) for radial velocities below mean radial velocity.
  • Log duration against Radial velocity follows a Gaussian distribution centred upon the mean radial velocity.
  • The spike seen in O+U observations centred on V = +100 m/sec has disappeared and the SD at 105.8 has more than halved.
  • there are a number of long duration trails with radial velocities > +600 m/sec and less than +200 m/sec that may be representative of Perseid meteor trails with trajectories that have been affected by high level winds in the upper atmosphere or are representative of 'sporadic meteors'

The graph of all reflection observations U+T plotted against the (24 hour) day date indicates :

  • Overall the period during which data was collected 31,182 reflections were recorded.
  • With 95% confidence:
  1. The maximum number of observations - between 18.6 and 19.5% - were recorded on the 6th August.
  2. Between 11 and 11.8% were recorded on the 7th August.
  3. Between 8.4 and 9% were recorded on the 10th of August.
  4. between 6.85 and 7.42% was recorded on the 14th of August
  5. between 6.38 and 7% were recorded on the 5th 8th 9th and 11th August .
  6. Between 5.2 and 6.56% were recorded on the 12th 13th 15th and 16th August.
  7. The minimum number of observations - between 3.4 and 3.8% was recorded on the 4th of August.


The graph of reflection observations U+T plotted against the hour of the day (UT) indicates with 95% confidence :

  • Two max peaks and two troughs of observations:
  1. A peak between 1:00 and 7:00 UT.
  2. A trough between 8:00 and 12:00 UT.
  3. A peak between 13:00 and 14:00
  4. A deep minimum between 18:00 and 22:00 UT.


The graph of reflection observations U+T plotted against the radial velocity V indicates:

  1. 50% of all observations are positive and have a radial (line of sight) velocity component moving towards the LVST receiver at between +279 and +448 metres per sec.
  2. The median velocity is + 374 metres per sec.
  3. The mean velocity is = +378.6 metres per sec. (SD=222.9)


Comparing the reflection observations against radial velocity graphs for U+O, O-T and U+T:

  • The O-T sub set of reflections is not only much smaller than the U+O set but also much smaller than the U+T sub set.
  • The O-T sub set has a much lower SD (standard deviation from the mean) than either the full O+U set or the sub set U+T
  • The sub sets were created by filtering the O+U set using an arbitrary value chosen for reflection duration.
  • The U+O and U+T graphs show a consistent base level of observations across the full range of radial velocities (from -102 to +948 m/sec) underlying the 'Gaussian bell curve' of data that is centred on approx +379 m/sec). This base level set of observations warrants further analysis.

The graph of O+U log maximum power x duration plotted against radial velocity. ( log max power limited to =< 3.6 in order to better see what is going on at lower energy levels)

Observations and Speculations:

  • Maximum log power x duration (within the limits of what can be recorded by the LVST) is proportional to the energy of the reflected wave.
  • T reflections are more likely to have greater energy than U reflections.
  • Three columns of reflection observations may be seen detached from a continuous horizontal band of observations that begins to thin out at a logmaxpowerxdur value = +2.853. The most dense column is centred on the mean radial velocity +379 m/sec. A less dense column is centred on +86 m/sec. - obs made on 14th Aug. and an even less dense column is centred on +141 m/sec -obs made on the 6th,7th and 8th of Aug.
  • Filtering the O+U data set by energy level (log max power x duration) using = or >3.15 might enable the U reflections to be removed from the O+U reflections to create a O+T subset.


Filtering the full set of reflection observations O+U by removing all data where logmaxpower x duration (energy) = or>3.15 to create a new subset O+T

The graph of reflection observations O+T plotted against the (24 hour) day date indicates :

  • Overall the period during which data was collected 16,950 reflections were recorded. This is 51.79% of the total number of reflections observed and represents only those reflections having a logmaxpower x duration =or>3.15 .
  • With 95% confidence:
  1. The maximum number of observations - between 17.45 and 18.61 % - were recorded on the 6th August.
  2. Between 11.62 and 12.60 % were recorded on the 10th of August.
  3. Between 10.4 and 11.34 % were recorded on the 7th of August.
  4. The minimum number of observations - between 2.92 and 3.44 % occurred on the 4th August.



The graph of reflection observations O+T plotted against the hour of the day (UT) indicates with 95% confidence :

  • One peak (approximately 5.8% of all observations were recorded in the hour commencing 5:00 UT), an approximately even base line and one trough in observations:
  1. Max peak between 2:00 and 7:00 UT.
  2. A base line for hours 00:00 to 1:00 and 8:00 to 17:00
  3. A deep minimum between 18:00 and 22:00 UT

The plot of days against radial velocity for the O+T subset of observations is interesting in that it shows outliers of observations, particularly on 7th and 14th of August, grouped around radial velocities lower than the mean radial velocity (+366.8 m per sec with SD = 103.6) for the O+T subset.

  • For the Observations O+T the plot of Observations per day against Observations per hour coloured to show differences in radial velocity is interesting. To aid clarity and understanding the data has been 'binned' on a hexagonal basis. The size of the hexagon is proportional to the number of observations.
  • The observations made between 14:00 and 1900 (UT) on the 7th of August indicate a large number of reflections with radial velocities well below the mean value for V. The timing of these observations in the afternoon is also at odds with other days where observation density clustering is clearly skewed to the morning. Do these factors reinforce the hypothesis that a stream other than the Perseids is being detected? Could this be an outlier from the Capricornids or the Alpha Capricornids streams which are active into August?


  • The same observations coloured to show energy (maxpower x duration) indicate the reflections in the afternoon of the 7th of August have a greater proportion of lower energy reflections. The Capricornids are slow moving meteors and have lower velocities (lower energies?) as well as different velocity vectors than the Perseids.


O+T Reflection Observations on the 7th of August 2016

The plot of logmaxpower versus radial velocity for the 7th of August shows a secondary spike of observations at radial velocity = approx +109 metres/sec. Note that these observations are primarily coloured blue and mauve indicating that they were recorded in the afternoon and early evening.


O+T Reflection Observations on the 14th of August 2016

The plot of logmaxpower versus radial velocity for the 14th of August shows a secondary spike of observations at radial velocity = approx +69 metres/sec. Note that these observations are primarily coloured blue indicating that they were recorded in the afternoon .


O+T Reflection Observations Aug 4_16 2016 (Detailed Data)

Total number of observations: 16950

Summary of the distribution of day.cat:

Table of Counts:

4 5 6 7 8 9 10 11 12 13 14 15 16 Row Total

539 1101 3056 1843 1212 1322 2053 1293 980 894 931 876 850 16950

Table of Percentages:

4 5 6 7 8 9 10 11 12 13 14 15 16 Total Row N

3.18% 6.50% 18.03% 10.87% 7.15% 7.80% 12.11% 7.63% 5.78% 5.27% 5.49% 5.17% 5.01% 100% 16950

Inference of the distribution of day.cat:

Estimated Proportion with 95% Confidence Interval

Lower Estimate Upper

0.0292 0.0318 0.0344

0.0612 0.0650 0.0687

0.1745 0.1803 0.1861

0.1040 0.1087 0.1134

0.0676 0.0715 0.0754

0.0740 0.0780 0.0820

0.1162 0.1211 0.1260

0.0723 0.0763 0.0803

0.0543 0.0578 0.0613

0.0494 0.0527 0.0561

0.0515 0.0549 0.0584

0.0483 0.0517 0.0550

0.0469 0.0501 0.0534

Total number of observations: 16950

Summary of the distribution of integhour.cat:

Table of Counts:

0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 Row Total

708 796 906 909 983 989 925 928 797 691 702 741 726 665 707 704 569 649 448 362 365 422 537 721 16950

Table of Percentages:

0 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 Total Row N

4.18% 4.70% 5.35% 5.36% 5.80% 5.83% 5.46% 5.47% 4.70% 4.08% 4.14% 4.37% 4.28% 3.92% 4.17% 4.15% 3.36% 3.83% 2.64% 2.14% 2.15% 2.49% 3.17% 4.25% 100% 16950

Inference of the distribution of integhour.cat:

Estimated Proportion with 95% Confidence Interval

Lower Estimate Upper

0.0388 0.0418 0.0448

0.0438 0.0470 0.0501

0.0501 0.0535 0.0568

0.0502 0.0536 0.0570

0.0545 0.0580 0.0615

0.0548 0.0583 0.0619

0.0512 0.0546 0.0580

0.0513 0.0547 0.0582

0.0438 0.0470 0.0502

0.0378 0.0408 0.0437

0.0384 0.0414 0.0444

0.0406 0.0437 0.0468

0.0398 0.0428 0.0459

0.0363 0.0392 0.0422

0.0387 0.0417 0.0447

0.0385 0.0415 0.0445

0.0309 0.0336 0.0363

0.0354 0.0383 0.0412

0.0240 0.0264 0.0288

0.0192 0.0214 0.0235

0.0193 0.0215 0.0237

0.0226 0.0249 0.0272

0.0290 0.0317 0.0343

0.0395 0.0425 0.0456

Total number of observations: 16950

Summary of radial.velocity:

Min 25% Median 75% Max Mean SD Sample Size

-103 342 373 405 948 366.8 103.6 16950

Total number of observations: 16950

Summary of radial.velocity by day.cat:

Min 25% Median 75% Max Mean SD Sample Size

4 -31 312 338 363.0 921 342.1 95.69 539

5 -103 331 348 367.0 933 349.3 77.45 1101

6 -103 350 380 405.0 948 379.6 95.64 3056

7 -103 151 342 386.0 948 297.0 164.29 1843

8 -103 344 369 392.0 933 367.6 76.35 1212

9 -90 342 369 402.5 948 371.9 86.83 1322

10 -103 354 382 415.0 931 383.1 83.32 2053

11 -103 365 392 417.0 948 393.9 80.59 1293

12 -103 361 386 411.0 948 387.2 74.16 980

13 -103 352 375 398.0 948 377.2 92.41 894

14 -103 331 373 394.0 927 341.1 125.75 931

15 -92 363 386 405.0 891 385.5 80.43 876

16 -103 373 390 407.0 948 394.5 84.86 850