Documentation

Introduction

This document in intended to be a short but concise user's manual regarding the handling and use of the "Mark-I" database. Additional references are given for a proper scientific exploitation and even to develop different calibration pipelines by each user, in addition to the used till now. The simplicity (Sun-as-a-star measurement: 1 pixel) and reduced size of each daily file (below 150 Kb) allows an easy process of all database at once (few Gb).

The "Mark-I" solar spectrophotometer

This instrument, located and operated at Observatorio del Teide (Altitude: 2.390 m; Longitude 16° 30' 35" W and Latitude 28° 18' 00" N) since 1976, provides precise radial velocity observations of the Sun-as-a-star at the Potassium KI 7699 Å absorption solar line. For a complete description on the technical details of the instrument, see the reference paper by Brookes et al. 1978.

As this data base corresponds to daylight observations gathered at a single ground-based site, it will lack of continuity (day-night cycle) but it possess some interesting and unique particularities: a) Observations extent since 1976 till present (currently 2012) with only summer campaigns from 1976 till 1983 and since then on a daily base all year around. b) Up to 1983 it was the only solar spectrometer fully devoted to Helioseismology. c) Its crucial instrumental parts remained "unchanged" and has a unique long term stability and sensitivity in terms of radial velocity of the Sun. d) Its systematic errors are well known and parameterized.

Data span and distribution

The completed database of daily Mark-I files covers the period July 1976 till December 2012. Observations were performed basically over summer season till 1984 when they become systematically on a daily base. In total, there are 8549 daily files out of the possible 1337 ones (64%). For the latest period (1985 till 2012), data is available for 8549 days out of the possible 10227 ones (84%). Each data file (a day of observation) has been subjected to the following processes: cleaning of the data (bad weather, instrumental failures), dead time correction of the photon counting values, resampling of the data, computation of the mean values of solar radial velocity Sun-instrument (un-calibrated) and of the mean intensity and calibration of the instrumental clock. The overall data characteristics (distribution and length of useful hours of data for each existing file) are shown in Figure 1.

Figure 1. Temporal distribution of the Mark-I data files. The percentage of hours of good data relative to the possible hours of sunlight is shown in grey for each existing data file. The annual values are represented by the black dots.	Figure 2. Mean values of the monthly duty cycle averaged over the period 1976-2012. The error bars correspond to the standard deviation of the mean value.

As in any ground-based observatory, the amount of solar data collected at Observatorio del Teide along the year is not uniform. Summer season use to allow much more hours of observations with clear skies. The illustrate this effect, in Figure 2 is shown the mean seasonal behaviour averaged over the 38-year length of the database.

Daily files: structure and contents

Daily file are named "izyyyymmdd.dat", where yyyy is the integer value of the year, mm of the month and dd of the day (i.e. August 5, 2009 data corresponds to the iz20090805.dat file). They are written is Ascii-format containing 12 columns of equal length. The length of the columns could vary from day to day depending on the amount of good data. In Appendix 1 their structure is described and in Figure 3 a snapshot of the contents of a given file (iz19970612.dat) is shown.

Figure 3.- Screen dump of the contents of a daily file (corresponding to June 12th, 1997). Only the first 35 rows of the file are shown

As described in Appendix 1, the first column of any file contains all required information to: properly characterize the data (date in various units), perform the correct time stamp of the events (behaviour of the instrumental oscillator -clock-, calibration against UTC, leap seconds correction), atmospheric quality of the day, ephemerides values to calibrate the instrumental velocity (line-of sight orbital and spin velocity of the instrument relative to the Sun) and some complementary ones (solar declination, solar axis angles, etc). In Palle et al. 1986 and Palle et al., 1993, a detailed description on various calibration procedures can be found.

Finally, in Figure 4 a sample plot of the daily instrumental velocity and of the transmitted light along the instrument are shown. The dominant feature of the line-of-sight velocity is the sinusoidal variation. By comparing the variation of the spin velocity of the observatory (directly from the ephemerides) with the amplitude of the observed sinusoidal variation of the instrumental velocity, a value of K ≈ 3000 (m/s) can be estimated that it resulted to be correct within ±5%. By applying any high-passband filter, adjusting any high order (n>1) polynomial function of properly modelling the observed signal (Palle et al. 1986 and Palle et al. 1993), the residuals can be obtained, showing clearly the 5-minute oscillation signal (see Figure 5).

Figure 4. Sample plot directly obtained from data contained in file iz19970612.dat. Time stamp (x-axis) is obtained from the instrumental time (column 2) and the use of the time check (rows 6 and 7 of the first column); the instrumental velocity r corresponds to column 3 and the sum of the scattering light (s) to column 5. The calibration constant K is about 3000, so the line-of-sight velocity varies for that day between 420 and 1050 m/s.

Figure 5.- Residuals of the line-of-sight velocity derived from the data shown in Figure 4. The 5-min oscillation pattern is clearly seen.

Appendix 1

Structure of the files:

Column	#1:	Headers and infos
	L1	day
	L2	month
	L3	year
	L4	Modified Julian Day (MJD)
	L5	N. blocks (of 40s or 42s) == number or rows
	L6	Hours UT of the TimeCheck
	L7	Block. Reading of the instrument's clock (counts) at T.CH
	L8	Bouger's fit: extinction coef.(slope)
	L9	Bouger's fit: Mv (ordinate origin)
	L10	tcross. U.T time of the crossing (seconds).(0 if no crossing)
	L11	etcross. Error in tcross (seconds)
	L12	Vgrs crossing in cm/s
	L13	evgrs. Error in Vgrs crossing (cm/s)
	L14	Tnoon (hours UT)
	L15	Vspi at noon (m/s)
	L16	Vorb at noon (m/s)
	L17	slopvorb (m/s)/hour. To compute Vorb at any time during the day: Vorb(t) = Vorb(tnoon) + slopvorb*(t-tnoon)
	L18	B angle (deg)
	L19	P0 angle (deg)
	L20	slopeclock (dimensionless, 8 decimal precision). Correction (multiplicative factor) to "instrumental seconds". Perfect clock= 1, faster clock < 1, slower clock > 1. When converting "block number" to "UT hours" using the daily time check, perform:
	L21	Leaps. Leap seconds added since UTC-TAI difference was defined as 10 seconds from January 1972.
	L22	Solar declination (decimal degrees) at Tnoon
	L23	daily slope of the solar declination (degrees/hour).
Column	#2:	Central time of the block (instrumental clock=counts). From 1984 onwards the length of a block has been of 40 s. and before was 42 s.
Column	#3:	Mean value of the instrumental velocity (ratio) computed over the length of the block (20 or 21 pairs). Outliers removed (2 runs of 3σ threshold)
Column	#4:	Standard deviation (σ) of <r>
Column	#5:	Mean value of the sum: <s> = Red + Blue
Column	#6:	Standard deviation (σ) of <s>
Column	#7:	Mean value of the Transmitted light: <tr> = TRed + TBlue
Column	#8:	Standard deviation (σ) of <tr>
Column	#9:	ADC Channel 1
Column	#10:	ADC Channel 2
Column	#11:	ADC Channel 3
Column	#12:	ADC Channel 4

References:

• Brookes J.R., Isaak G.R., van der Raay H.B., (1978), "A resonant-scattering solar spectrometer", MNRAS, 185,1.
• Palle P.L. et al., (1986), "The global oscillation spectrum of the sun. I - Analysis of daily power spectra of velocity measurements", A&A, 169, 313.
• Palle P.L. et al., (1993), "Full-disk helioseismic IRIS raw data calibration", A&A, 280, 324.