HAWC as a Ground-Based Space-Weather Observatory

Abstract

The High Altitude Water Cherenkov (HAWC) gamma-ray observatory is located close to the equator (latitude $18^{\circ }$ N), at an altitude of 4100 m above sea level. HAWC has 295 water Cherenkov detectors (WCD), each containing four photomultiplier tubes (PMT). The main purpose of HAWC is the determination of the energy and arrival direction of very high energy gamma rays produced by energetic processes in the universe, HAWC also has a scaler system which counts the arrival of secondary particles to the detector. In this work we show that the scaler system of HAWC is an ideal instrument for solar modulation and space-weather studies due to its large area and high sensitivity. In order to prepare the scaler system for low energy heliospheric studies, we model and correct the efficiency variation of each PMT of the array, which result in a capability to measure variations $> 0.01\%$ with high accuracy. Using the singular value decomposition method, we correct the rate deviations of all PMTs of the array, due to changes in efficiency, gain and operational voltage. We isolate and remove the atmospheric modulations of the PMTs count rates measured by the TDC-scaler data acquisition system. In particular, the atmospheric pressure at the HAWC site exhibits an oscillating behavior with a period of ∼12 hours and we make use of this periodic property to estimate the pressure coefficients for the HAWC TDC-scaler system. These corrections performed on the TDC-scaler system make the HAWC TDC-scaler system an ideal instrument for solar modulation and space-weather studies. As examples of this capability, we present the preliminary analysis of the solar modulation of cosmic rays at three time scales observed by HAWC, with an unprecedented accuracy.

Appendices

Appendix A: Comparison of FFT Method with Conventional Method

The application of the FFT filter on the TDC-scaler rate has effectively isolated the pressure modulation from the coupled solar modulation in the TDC-scaler rate, allowing a more accurate measurement of $\beta _{P}$. In this appendix, we demonstrate the effectiveness of this method over the conventional method. Figure 13 shows both the conventional method and the FFT filter method, the fitting of a linear function is most suitable, and the fit quality is better for the FFT method. The estimated $\beta_{P}$ during three months of observation using both methods are given in Table 4, from these values we can see how consistent and effective is the FFT method for estimating $\beta _{P}$. Using the conventional method, the estimated $\beta _{P}$ values are not consistent, the deviation is much higher than the required accuracy of HAWC TDC-scaler system, whereas the FFT method estimated the $\beta _{P}$ with an accuracy within the required limits. The variation of $\beta _{P}$ using the conventional method is due to the presence of solar modulation, this method was not able to decouple the solar modulation from the atmospheric one.

Figure 13

The rows from top to bottom show data of months of September, October, and November in the corresponding order. The first column shows the dependency of TDC-scaler rate on pressure difference using the conventional method. The second column shows the same with the FFT filter method.

Table 4 $\beta _{P}$ estimated using different methods.

Appendix B: Checking the Health of PMTs Using $\beta _{P}$

Once the pressure modulations were isolated and the pressure coefficient $\beta _{P}$ was estimated for all 1180 PMTs, we take advantage of the fact that this modulation is a purely physical phenomenon (due to the atmospheric properties) and therefore, the effects of pressure modulation are independent of the detector. This means that the obtained $\beta _{P}$ should be similar for the 1180 PMT and the deviation of $\beta _{P}$ from its mean value of a particular PMT/WCD is due to malfunction of that particular PMT or its associated components.

The distribution of $\beta _{P}$ ($R1$) of all 1180 PMTs for the months of September, October, and November of 2016 is shown in Figure 14, which is Gaussian in nature. A Gaussian fit to this distribution gives us a standard deviation as shown in Table 1. We chose a cut-off limit of 2 and $3\sigma$ levels. Using this criterion we have classified the PMTs and WCD s, which are shown in Table 5.

Figure 14

The distributions of the pressure coefficient $\beta _{P}$ of all PMTs ($R1$) for the months of September (top), October (middle), and November (bottom), 2016 are shown. The first column shows the distribution obtained using the exponential method and the second column shows that by linear method. The red lines show the Gaussian fit for these distributions.

Table 5 Classification of PMTs and WCDs. Third row gives the number of PMTs/WCDs that are good, fourth row gives the number of PMTs/WCDs with in 3 – $5\sigma$ range, fifth one shows that above $5\sigma$, and the last one shows that with no data. Column 2,3 and 4 correspond to single PMT rates R1, column 5, 6 and 7 to Multiplicity M2, column 8, 9, 10 for Multiplicity M3, and column 11, 12 and 13 to Multiplicity M4.

The value of $\beta _{P}$ and its deviation from the mean value are the quantitative measure of the health of a PMT. These give us a measure of how well a PMT is functioning within its normal gain mode. Considering this criterion we isolated the PMTs and WCDs outside this 3$\sigma$ range, and included PMTs and WCDs within 3$\sigma$ range for our further analysis of solar modulations. This selection process improved the accuracy of the measurement and reduced the systematic errors. This process has made the data more suitable for accurate solar modulation studies.