5.1 Introduction

5.2 Regression Methods Used for the Barometric Coefficient Calculation

5.3 Discussion

5.4 Conclusion

5.1 Introduction

Particle detectors of the Aragats Space Environmental Center (ASEC, Chilingarian et al., 2003, 2005,) are located on the slopes of mountain Aragats and in CRD headquarters in Yerevan, Armenia; geographic coordinates: 40°30'N, 44°10'E, altitudes - 3200m, 2000m and 1000m. a.s.l. Various ASEC detectors, measuring fluxes of diverse secondary cosmic rays, are sensitive to different energetic populations of primary cosmic rays. Two neutron monitors (18NM-64) operating at the Nor-Amberd and the Aragats research stations detect secondary neutrons. The Nor-Amberd muon multidirectional monitor (NAMMM) detects low energy charged particles and muons with energies above 350 MeV. The Aragats Multidirectional Muon Monitor (AMMM) registers high energy muon flux (threshold energy - 5 GeV). The Aragats Solar Neutron Telescope (ASNT) measures neutrons and charged particles. ASNT is a part of a world-wide network coordinated by the Solar-Terrestrial Laboratory of the Nagoya University. Another monitoring system, based on the scintillation detectors of the Extensive Air Shower (EAS) surface arrays, MAKET-ANI and GAMMA (3200 m a.s.l.), detects low energy charged particles. New world-wide particle detector networked, named SEVAN, operates now in Armenia, Bulgaria and Croatia (Chilingarian & Reymers 2008, Chilingarian et al., 2009). SEVAN detectors measure low energy charged particles, neutral particles (gammas and neutrons) and high energy muons. NAMMM and ASNT measuring channels are equipped with Amplitude-to-Digital (ADC) convertors and microcontroller based advanced electronics. Data Acquisition (DAQ) electronics and flexible software triggers allow not only to register the count rates of the detector channels, but also histograms of energy releases; correlations of the charged and neutral fluxes; and many other physical phenomena. Details of detector operation can be found in (Chilingarian et al., 2007 and Arakelyan et al., 2009).
ASEC detectors measure time series of secondary particles born in cascades originating in the atmosphere caused by primary protons and stripped nuclei. The networks of particle detectors can predict upcoming geomagnetic and radiation storms hours before the arrival of Interplanetary Coronal Mass Ejections (ICMEs) at the ACE and SOHO spacecraft. The less than one hour lead time (the time it takes for the ICME to travel from the spacecraft to the magnetosphere) provided by particle detectors located at ACE and SOHO at the libration point 1,5 million kilometers from the Earth is too brief to take effective mitigating actions to protect satellites and surface from the harm of major geomagnetic storms. For reliable and timely forecast we need adequate models of the major solar energetic events in progress. The information on the highest energy solar cosmic rays, available from surface based particle detectors can be used to test such models and to obtain overall knowledge on the particle acceleration in flares and by fast shock blasts; on transient modulation effects posed by sun activity on the Galactic Cosmic Ray (GCR) flux; on the interactions of solar wind with magnetosphere; on the dynamic of the magnetosphere and many others.

Cosmic Ray flux incident on the terrestrial atmosphere and measured elementary particles on the Earth surface comprise very different entities although genetically connected with each other. Primary particles interactions with atmospheric nuclei and different meteorological effects can hide genuine variations of the primary flux and prevent from understanding of dynamics of ongoing physical processes in solar-terrestrial chain.

For recovering the primary particles fluxes incident on the Earth’s atmosphere it is necessary to know the relationship between observed count rates of the detectors and the primary particles fluxes, as well as the influence of the meteorological effects on the flux of secondary particles reaching the Earth surface. Dorman’s theory of meteorological effects (Dorman, 2004) gives detailed classification of the effects; mentioned the barometric one as a major influencing particle fluxes (at least for highest energies – 10-100 GeV). Therefore, it is of greatest importance to accurately measure the barometric coefficients to “unfolding” the solar modulation effects. Besides this main goal there exists several independent research problems connected with barometric coefficient dynamics:

• rigidity dependence;
• solar cycle phase dependence;
• height dependence;
• detected particle type dependence

All these dependences can be investigated at ASEC and by SEVAN network due to different altitudes, various detected particle fluxes and planned long-term operation.
The main drivers of these dependences are changing according to solar cycle phase primary flux, type of secondary flux, and location of the detector. At minimum of solar activity, the GCR flux is enriched by abundant low energy (below 10 GeV) particles, blown out from solar system by intense solar wind at maximum of solar activity. Particle detectors located at high latitudes also are sensitive to lower primary energies compared with detectors located at middle low latitudes, because of lower rigidity cutoff. Detectors located at high altitudes are registering more cascade particles than sea level detectors due to attenuation of cascade in the atmosphere. Therefore, because pressure effects should be more pronounced for cascades initiated by particles of lower energies and at cascades containing more particles, following relation can be expected:

• Barometric coefficient absolute value for same secondary particle flux is greater for detectors located at high latitudes compared with low latitudes;
• Barometric coefficient absolute value for same secondary particle flux should be greater at minimum of solar activity compared with maximum;
• Barometric coefficient absolute value for same secondary particle flux should be greater for high altitudes compared with sea level location;
• Barometric coefficient absolute value should be greater for neutrons than for muons;
• Barometric coefficient absolute value should be inverse proportional to muon energy;
• Barometric coefficient absolute value should be inverse proportional to zenith angle of incident particle flux.
• Barometric coefficient absolute value should be lower for high multiplicities detected in Neutron Monitors and for greater dead times of DAQ electronics.

All mentioned dependences were investigated and discovered during the last 50 years by the networks of neutron monitors and muon detectors (see details in Dorman, 2004). Nonetheless, because of peculiarities of detection techniques, scarce statistics, highly different local meteorological conditions, cycle-to cycle variations of solar activities dependencies yet are more qualitative and additional investigations of dynamic and interrelations of barometric coefficients are needed. ASEC provides ideal platform for such research.

During more than 50-years operation Neutron Monitors (NM) network prove to be extremely effective in observing solar modulation effects. Several attempts were made to enlarge NM information contain: put additional channels without lead coverage, measure so called multiplicity (number of multiple counts), etc. The monitors are equipped with new electronics providing time integration of counts by three dead times. The first dead time equals to 400ns for collecting almost all secondary neutrons generated in the lead of NM. The second dead time is equal to the 0.25ms and the third one equal 1.25ms (as most of NM from world-wide network).
Physical analysis of the 3 time series from one and the same monitor and comparison of data from 2 monitors located at 2 altitudes will be presented in the report. Barometric coefficients of all 6 time series will be calculated and compared.

The paper is organized in the following way: the second chapter will explain the statistical techniques used for the barometric coefficient calculations; the third section will present the main results obtained for ASEC monitors at beginning of solar cycle 24; in discussion section we’ll compare our results with previously obtained data and will check consistency of obtained results with expectations.

5.2 Regression Methods Used for the

Barometric Coefficient Calculation

Experimentally, the intensity I of any secondary cosmic ray component varies with a small change in the atmospheric pressure P (Dorman ,1974) as

where µ is the absorption coefficient for the secondary component under consideration. For
µ = constant, the equation (1) gives

Where P is pressure and P₀is reference pressure, usually the average pressure at station. I and I₀ are counting rates at these pressures, β is barometric coefficient.
After simple transformation we readily get equation of linear regression:

Empirically value of the barometric coefficient can be found by means of liner correlation between intensity of cosmic-rays I_i and data of atmospheric pressure P_i.

Where r correlation coefficient:

The relative error of estimation β can be calculated as follows:

Data for barometric coefficient calculation is selected at time periods when there were no disturbances of the Interplanetary Magnetic Field (IMF) and magnetosphere; and in addition there were significant changes in the atmospheric pressure. The least square method was used to obtain the regression coefficients. Large values of correlation coefficient prove correct selection of the reference data.
In Tables 1we summaries the calculated barometric coefficients of ASEC monitors. In the columns accordingly are posted the altitude; cut-off rigidity; barometric coefficient; goodness of fit – the correlation coefficient; minute count rate; relative error of count rate; “Poisson” estimate of relative error. Values posted in the last two columns should be very close to each other if the particle arrival can be described by the Poisson process. Any small deviation manifested the correlation between detector channels; any large correlation – failures in electronics or data acquisition software.

Aragats and Nor Amberd neutron monitors operate with 3 different dead times. The shortest dead time collected all secondary neutrons generated in lead by primary hadrons. As it was demonstrated in (Cilingarian & Oganissyan, 2009), secondary neutrons can be registered in neighboring channels of monitor. Therefore, due to this embedded correlation “Poisson” and measured relative errors for shortest dead time deviated from each other. When enlarging the dead time, the one-to-one relation between high energy hadron entering detector and detector count is established, inter-channel correlation vanished and Poisson and measured relative errors get equal.
The Aragats Multichannel Muon Monitor (AMMM) after changing data acquisition electronics, demonstrates large deviations of measured and Poisson relative errors. Therefore, electronics of AMMM is moved and under repair now.
In table 2 we present barometric coefficients for the SEVAN detectors located at Aragats and in Yerevan (Chilingarian and Reymers, 2008, Chilingarian et al., 2009). SEVAN detectors have 3 layers inter-layered with lead filters. Middle thick layer is sensitive to the neutral particles. Analyzing the outputs from each layer we can outline different species of the incident on detector particles. For instance combination (010 – signal only in middle scintillator) “selects” neutral particles. Probability that neutral particle give signal in upper 5 cm. thick scintillator less than 5%; and – the signal probability that neutron will give signal in middle 25 cm. thick scintillator is ~ 25%. The combination (111 signals in all scintillators) “selects: muons with energies greater than 250 MeV – the energy necessary to cross 10 cm. of lead.
In Table 3 we compare barometric coefficients of neutron monitors sending data to the Neutron Monitor Data Base (NMDB), a new European project to collect and present minute data from Eurasian detectors. The cutoff rigidities of selected monitors ranging from 0.81 to 7.1 GeV gave good representation of the network and, in addition, the Table provides some hints to compare monitor sensitivity to transient solar events and check of chamber failures. Different data reliability checks are of upmost importance when you are collected and compared data obtained from different detectors using various electronics and data acquisition software.

5.3 Discussion

Large diapason of the barometric coefficient values, covering approximately one order of magnitude, from 0.08% for the >5 GeV muon flux till 0.73% for the neutron flux demonstrates unique sensitivity of ASEC detectors to primary rigidities from 7 to 50 GV.

ASEC neutron monitors simultaneously measure count rates corresponding to the 3 preselected dead times: 0.4 us, 250 us and 1250 us. This additional information will provide possibilities to access different primary energies. Indeed, from Figure 1 we can see that for both ANM and NANM larger dead times are correspondent to smaller barometric coefficients, i.e. to higher primary energies. As it was expected the absolute value of barometric coefficients increase with decreasing dead time, because of increasing sensitivity to lower energy primaries more influenced by pressure changes. In Figure 1 in addition have depicted barometric coefficient obtained from data of two proportional counters located in Nor Amberd without lead filters. As it was expected, these chambers are most influenced by atmospheric pressure, due to their sensitivity to the lowest energy atmospheric neutrons

Figure 1. Comparison of barometric coefficients different dead times ASEC neutron monitors.

In Table 3 and 4 you can also see barometric coefficients for neutron monitors of Izmiran (Moscow) and Oulu stations. Data were taken from Neutron Monitor Data Base (NMDB) in Kiel, Germany.

Table 1. Barometric coefficients, count rates and relative errors of ASEC monitors

Table 2. Simulated and experimental count rate relative increases of the Aragats and Nor-Amberd neutron monitors at 7:15UT on 20 January 2005

Table 3. Barometric coefficients, count rates and relative errors of Aragats (us=0.4),  Izmiran(Moscow) and Oulu (Finland)  neutron monitors, data from NMDB

Figure 2. Comparison of Aragats (ARNM), Nor Amberd (NANM), Izmiran (Moscow) and Oulu
neutron monitors with barometric coefficients for neutron component calculated by L.Dorman.

In Figure 2 we compare barometric coefficients of Aragats (ARNM, 18 NM 64), Nor Amberd (NANM, 18 NM 64), Izmiran (Moscow 24 NM 64) and Oulu (9 NM 64) neutron monitors with barometric coefficients for neutron monitors calculated by (Dorman et al., 1968) during minimum of solar activity in 1964-1965.

We use dead time equal to 1250 microseconds, value commonly used in the world-wide network of neutron monitors in 60-ths. All coefficients relate to solar activity minimum years (1965 and 2008) and are in good agreement with each other. Also it is apparent increase of the absolute value of barometric coefficients with decreasing of cutoff rigidity.
From ASEC muon channels we can see that absolute value of barometric coefficients is inversely proportional to the muon energy.
In addition, by SEVAN barometric coefficients we can illustrate that indeed measured fluxes “selected” by detector electronics are enriched by different species of cosmic rays. Of course we cannot measure “pure” flux of neutrons, due to contamination of gamma-quanta and muons. However, as we see from Table 2, events selected as “neutrons” demonstrate barometric coefficients approximately twice as events selected as muons. Just the same behavior we expect from the neutron and muon fluxes. SEVAN detector measure in addition different combinations of signals in detector layers; therefore we can pose problem of finding barometric coefficients of the “pure” fluxes, as it was described in (Quang et all., Proc. ASA 2 (5) September 1974).

The summary of ASEC barometric coefficients we present in Figures 3 and 4. From simulations (Zazyan & Chilingarian, 2008) we estimate the most probable energy for each detector.

Figure 3. The dependence of barometric coefficient on the primary energy for Nor Amberd station detectors

Figure 4. The dependence of barometric coefficient on the primary energy for Aragats station detectors.

As we can see in Figure 3 and 4, energy of primary particle to which detector is sensitive and barometric coefficient correlate rather well.

5.4 Conclusion

1. Large diapason of the barometric coefficient values, covering approximately one order of magnitude, from 0.08% for the >5 GeV muon flux till 0.73% for the neutron flux demonstrates unique sensitivity of ASEC detectors to primary particles with rigidities from 7 to 50 GV.
2. Barometric coefficients of monitors belonging to the new particle detector network SEVAN demonstrate that 3 layers of monitors are sensitive to different species of the secondary cosmic rays, namely: low energy charged neutrons and high energy muons. It is independent check of the SEVAN, proving results obtained by simulations
3. Preliminary analysis of the barometric coefficients calculated for ASEC monitors proves expectations about its energy and altitude dependence. Obtained coefficients are used for correcting ANM and NANM data to appear in the Neutron Monitor Data Base (NMDB) in Kiel, Germany, a European project to collect on-line data of neutron monitors. Data also is transferred to mirror site in USA and will be transferred to new mirror sites of CRD site in Russia and Europe.

References

• Chilingarian A., Avakyan K., Babayan V. et al., Aragats space-environmental centre: status and SEP forecasting possibilities. J. Phys. G: Nucl. Part. Phys. 29, 9–951, 2003.
• Chilingarian A., Arakelyan K., Avakyan K. et al. Correlated measurements of secondary cosmic ray fluxes by the Aragats Space-Environmental Center monitors. NIM, A543, 483- 492, 2005.
• Chilingarian A., Melkumyan L., Hovsepyan G., Reymers A., The response function of the Aragats Solar Neutron Telescope, Nuclear Instruments and Methods in Physics Research A 574, 255-263, 2007.
• Chilingarian A. and Reymers A., Investigations of the response of hybrid particle detectors for the Space Environmental Viewing and Analysis Network (SEVAN), Ann. Geophys., 26, 249-257, 2008.
• Chilingarian A., Hovsepyan G., Arakelyan K., et al. Space environmental viewing and analysis network (SEVAN). Earth, Moon and Planets, 104, 195, 2009.
• Chilingarian A., Hovhannisyan A. Calculation of Neutron Mulitplication Coefficients, private communication, 2009.
• Dorman L.I. Cosmic Rays in the earth’s atmosphere and underground, Kluver Academic Publishers, Dordrecht, Boston, London, 2004.
• Dorman L.I., Rogava O.G., Shotashvili L.Kh. Planetary distribution of cosmic ray neutron component barometric coefficient during IQSY, Geomagnetizm and Aeronomy 8, 166-167, in Russian,1968.
• Dorman L.I., Rogava O.G. and Shatashvili L. Kh. Planetary distribution of cosmic ray neutron component barometric coefficients during IQSY, Geomagnetism and Aeronomy, 8, No. 1, 166-167, 1968.
• Dorman L.I., Cosmic Rays, Variations and Space Exploration. North-Holland, Amsterdam. 1974.
• Rogelio Caballero López and José Fco. Valdés-Galicia. Variations in cosmic radiation intensity associated with the barometric effect. Geofisica Internacional Vol.39, Num. 1, pp. 135-137, 2000.
• Quang T.T., Fenton A.G. and Fenton K.B.. Effect of the Muon Contribution on the Barometric Coefficient of Neutron Monitors, Proc. ASA 2 (5) September 1974