Originating from outer space, cosmic rays are high-energy charged subatomic particles that travel close to the speed of light. Most of these cosmic rays come from outside our solar system where blackholes and massive stars that undergo supernovae are among some of the main violent sources that accelerate and shoot high-energy charged particles into space. Our Sun is also a source of cosmic rays. When powerful solar activity like solar flares or coronal mass ejections occur, massive clouds of hot gas and solar protons are blown away from the sun. The earth is constantly bombarded by cosmic rays from every direction. Upon striking Earth's atmosphere, the cosmic rays lose energy through interactions with air molecules and shower into secondary particles that rain down onto the surface of the planet. As a result, weather conditions and height-dependent state variables such as pressure and temperature affect the production of extensive air showers that can be detected by observatories on the ground.
The High-Altitude Water Cherenkov (HAWC) Gamma Ray Observatory is a ground-array detector currently under construction at 4,100 m above sea level on the Sierra Negra volcano near Puebla, Mexico. It is being built at a high elevation because as air showers cascade downward, the showers reach a maximum size and then decrease in density. At higher altitude, the observatory will be able to see more particles in air showers and increase its sensitivity to lower energies. Once completed, the facility will have 300 water Cherenkov tanks in total. It is able to observe high-energy gamma rays and cosmic rays in the 100 GeV to 100 TeV spectrum. The detector has a large field of view and at any given time is able to see 15% of the sky allowing it to observe half of the sky over a course of 24 hours. It uses the water Cherenkov technique to observe extensive air showers. Similar to how sonic booms are produced when objects move faster than the speed of sound, particles traveling faster than light will produce Cherenkov radiation. When air shower particles penetrate a tank and continue through the water, they are moving faster than the speed of light since the index of refraction of water causes photons to travel slower. Thus, the particles produce Cherenkov radiation. Cherenkov light has wavelengths in the ultraviolet spectrum and eminates in a distinct forward cone of 41 degrees from the direction of motion of the particle. There are four photomultipliers (PMTs) in each tank that are able to detect Cherenkov photons. The timing and charge in each PMT is then used to reconstruct the primary particle.
Scaler Data Acquisition System
HAWC uses two data acquisition systems (DAQs) to collect information on extensive air showers: the Trigger System and the Scaler System. This particular research is focused on data obtained by the Scaler DAQ. The scaler system operates through PMT pulse counting from the amount of charge each PMT has accumulated. Counts are due to air showers, naturally occurring radioactivity near the PMTs, and thermal noise inside the PMT electronics. The scaler system is sensitive to gamma ray and cosmic ray transient events. It is unable to provide the directional information of an incoming particle but the system is used in complement with the trigger DAQ and other detectors sensitive to gamma ray transients.
Sensitivity to Heliospheric Activities
HAWC has the ability to observe solar flares and coronal mass ejections (CMEs) that occur on the surface of the sun. When the sun experiences solar flares, a large amount of solar energetic particles (SEPs) are released into space that increase the quantity of cosmic rays bombarding the Earth. The raised radiation levels detected on Earth's surface are then called ground level enhancements (GLEs). In the event of a CME, however, massive amounts of plasma are blasted out into the solar system. The magnetic field from the resulting plasma solar wind sweep away some of the galactic cosmic rays entering Earth and the rapid decrease of cosmic ray levels detected is known as a Forbush decrease.
The main objective of this research is to identify decreases in scaler rate due to heliospheric events while removing normal variations due to weather. Two principle atmospheric factors that affect the counting rate are pressure and temperature. Air pressure is important to record because as the air column density where an air shower is traveling through increases, there are more air molecules that collide and block the incoming particles. Temperature data is also important since as the ambient temperature increases, the air density decreases which allows more particles to pass through the atmosphere unimpeded.
Fig. 1 . Forbush Decrease
Figure 1. is an illustration of how a detected Forbush decrease will look. The red curve depicts the daily normal fluctuations in counts caused by small changes in the atmosphere around HAWC while the blue curve depicts a Forbush decrease with a characteristic rapid fall in count rates over several hours and a slow gradual recovery over the course of a few days.
In order to remove noise caused by Earth's atmosphere, the atmospheric conditions surrounding HAWC must be known. The Global Data Assimilation System (GDAS) and HAWC's weather station are two systems that are used to analyze the weather. GDAS is an atmospheric model developed by National Centers for Environmental Prediction (NCEP). The system provides analysis four times a day at 0, 6, 12, and 18 UTC on meteorological state variables that are altitude-dependent including temperature, pressure, and humidity. It also offers 3-, 6-, 9-hour forecast. Weather data used in the research, acquired from GDAS, are based on the 3-hour forecasting. HAWC also has its own weather station that is built on one of its tanks. The weather station provides local temperature, humidity, wind speed, pressure, and rainfall at the detector, 4100 m above sea level.
The graphs below are compilations of the daily variations in pressure and temperature at ground-level, throughout April 2013, at the HAWC site.
Plot 1 . Ground-level GDAS data for April 2013.
Although both the daily and overall changes in the pressure and temperature are small, the two state variables are still able to affect the count rates by adding unwanted noisy fluctuations recorded by the observatory.
The following graphs are a comparison between GDAS data and HAWC's weather station data. Each graph is segmented into day and night cycles, with the white segment beginning with sunrise and the dark segment beginning with sunset. The pressure data from GDAS and the weather station were similar with only a difference of approximately 0.1 hPa to 2 hPa.
Plot 2 . Pressure data comparison between GDAS and weather station. Second graph shows small difference of the two systems with respect to pressure.
However, when comparing the GDAS temperature with the weather station temperature, the difference was quite significant. It was observed that the weather station reading of the local temperature drastically spiked during sunrise and plunged over the course of a day. These large swings in temperature are not yet understood but are thought to be due to the reflection of sunlight from the surrounding tanks back on the weather station during the day.
Plot 3 . Temperature data comparison between GDAS and weather station. The large peaks in temperature from the weather station may suggest problems due to local environmental factors (i.e. reflective surfaces of detector)
Thus, the comparisons between GDAS data and weather station data suggests that the pressure read by both systems is suitable for further use in studying atmospheric effects on extensive air showers but there are some problems that will need to be addressed with the weather station recording temperature data.
April 2013 Coronal Mass Ejection
The reason for assessing the environmental conditions during April is because on April 11, 2013, the sun experienced a medium-sized flare, ranked M6.5, close to the central meridian of the solar disk. A CME associated with the flare was also discharged close to Earth at roughly 960 km/s. The solar wind spent three days to propagate from the sun to the earth and HAWC detected the Forbush decrease connected with the CME on April 14. A constant Forbush decrease of approximately 3% was observed throughout the whole detector.
Fig. 2 . Constant Forbush decrease sweeping through 30 PMT channels. Count rate for each channel is slightly different since each PMT is hand-crafted. Large spikes in counts near the beginning and end of April are due to thunderstorms dumping excess charge into the detector.
Correlating Weather with Scaler Count Rate
The acquired April 2013 scaler rate data is analyzed with the April 2013 pressure and temperature. As expected, there is a very distinct anti-correlation between the pressure and channel rate. When the pressure increases the scaler counts decrease. Plot 4. is a graph of Pressure vs. Channel Rate from channel 38. Included below is also a list of Pressure vs. Channel Rate from twenty-nine other channels.
Plot 4 . Clear anti-correlation between pressure vs. channel rate.
Channel 39 Rate w/ Pressure
Channel 77 Rate w/ Pressure
Channel 82 Rate w/ Pressure
Channel 113 Rate w/ Pressure
Channel 114 Rate w/ Pressure
Channel 119 Rate w/ Pressure
Channel 120 Rate w/ Pressure
Channel 124 Rate w/ Pressure
Channel 126 Rate w/ Pressure
Channel 130 Rate w/ Pressure
Channel 178 Rate w/ Pressure
Channel 185 Rate w/ Pressure
Channel 187 Rate w/ Pressure
Channel 189 Rate w/ Pressure
Channel 190 Rate w/ Pressure
Channel 192 Rate w/ Pressure
Channel 236 Rate w/ Pressure
Channel 242 Rate w/ Pressure
Channel 246 Rate w/ Pressure
Channel 248 Rate w/ Pressure
Channel 297 Rate w/ Pressure
Channel 298 Rate w/ Pressure
Channel 306 Rate w/ Pressure
Channel 308 Rate w/ Pressure
Channel 364 Rate w/ Pressure
Channel 365 Rate w/ Pressure
Channel 366 Rate w/ Pressure
Channel 368 Rate w/ Pressure
Channel 432 Rate w/ Pressure
The correlation between the pressure and the channel rate, however, seems to be very weak. Since the pressure and rate are anti-correlated, the resulting graph of pressure vs. rate should yield a clear negative sloped trend. It is currently not known why the correlation is so weak. There may be other significant factors that have not been examined yet. From analyzing the graphs below, there is a very minor negative trend in the data but it is definitely not enough to suggest a well-defined anti-correlation. For the temperature vs. rate, there should be a clear positive slope trend but there appears to be no correlation at all. Below is a list of channel rates correlated with pressure and temperature from the previous twenty-nine channels.
Plot 5 . Very weak negative trend for pressure vs. channel rate. Temperature vs. channel rate has no correlation.
Channel 39 Rate w/ PT
Channel 77 Rate w/ PT
Channel 82 Rate w/ PT
Channel 113 Rate w/ PT
Channel 114 Rate w/ PT
Channel 119 Rate w/ PT
Channel 120 Rate w/ PT
Channel 124 Rate w/ PT
Channel 126 Rate w/ PT
Channel 130 Rate w/ PT
Channel 178 Rate w/ PT
Channel 185 Rate w/ PT
Channel 187 Rate w/ PT
Channel 189 Rate w/ PT
Channel 190 Rate w/ PT
Channel 192 Rate w/ PT
Channel 236 Rate w/ PT
Channel 242 Rate w/ PT
Channel 246 Rate w/ PT
Channel 248 Rate w/ PT
Channel 297 Rate w/ PT
Channel 298 Rate w/ PT
Channel 306 Rate w/ PT
Channel 308 Rate w/ PT
Channel 364 Rate w/ PT
Channel 365 Rate w/ PT
Channel 366 Rate w/ PT
Channel 368 Rate w/ PT
Channel 432 Rate w/ PT
One method that was tried to produce a better pressure-channel rate correlation was to calculate a running average of the scaler rate and the pressure separately and from there subtract the average from the original data. This generates pressure and rate data that oscillates around 0 hPa and 0 Hz respectively. The pressure and rate were then graphed together to obtain a correlation trend. The pressure data and the scaler rate data on PMT channel 38 were used to test this process.
Plot 6 . A running average was computed for the scaler rate and pressure.
Plot 7 . Running average subtracted rate and pressure.
Plot 8 . Better negative correlation for pressure vs. channel rate but still not well-defined.
Once again, the correlation is still rather weak, however, it does slightly show a clearer negative trend.
Another possible solution that will be done to achieve a well-defined correlation between the pressure and channel rate is to perform a fast Fourier transform (FFT) on the pressure data. Undesired noise in the frequency domain will be filtered out and then the data will be converted back to the time domain with reduced interferences. Attempts will be made to apply pressure corrections to the data even if they are weak. Other important factors will also be taken into account in order to minimize their noise contribution within the data such as the temperature of the internal circuitries of devices used in gathering and evaluating the scaler counts.
I would like to say thank you to the following who have made my summer 2013 REU program possible.
Dr. Segev BenZvi for advising and teaching me throughout the summer program
Eric Hooper for being a fun program coordinator and keeping me updated with activities
The UW-Madison HAWC Group
The WIPAC Group
Fellow REU Students
The National Science Foundation
Everyone was so kind, patient, and willing to help me when I had difficulty understanding what to do. It was a fantastic learning experience for me! Finally, I wish all my fellow 2013 astrophysics REU members the best of luck in everything they do.
For more information regarding HAWC, particle astrophysics, and astronomy, please visit:
Wisconsin IceCube Particle Astrophysics Center
SIMBAD (Stellar/Galactic database)
NED (Extragalactic database)
NASA Astrophysics Data Service
Research projects of other REU students