The DIARAD/SOVIM instrument performance
The SOVIM package is composed of three PMO6 absolute radiometers, two sun- photometers, one pointing sensor (TASS) and one DIARAD radiometer. SOVIM is mounted in the SOLAR payload of the International Space Station (ISS) together with the SOLSPEC (Thuillier et al., 2008) and SOLACE (Brunner et al., 2008) instruments. The DIARAD instrument is developed at the Royal Meteorological Institute of Belgium.
The extended open sequence is used to increase the number of acquisition points.
The active channel has its shutter closed for 90 seconds. It is maintained afterwards opened during 270 seconds instead of 90 seconds. This cor responds to a succession of one closed state and three open states. This sequence is illustrated in Figure 1. During the ﬁrst open state following a closed state, the servo system reduces the regulated power to reach the equilibrium as during the classical operating mode. During the following two open states, the servo system starts from an equilibrium state. As a consequence all the power acquisitions made 56 at seconds 20, 50, 60, 80 and 90 of the second and third open state can be exploited as equilibrium measurement points. In total the succession of extended sequences during Solar measurements gives 11 acquisition of the TSI every six minutes (the last acquisition of the ﬁrst open state and each of the ﬁve acquisitions of the following two open states). Compared to the classical operating mode, this multiplies by a factor of 5.5 the number of acquisitions during an orbit.
Figure 1: Top: the extended open sequence is composed of one closed state and three open states. The localisations of the acquired temperatures are also displayed. Bottom: the ’x’ are the acquired dissipated power in the regulated cavity (left) during the diﬀerent states. The data are linearly interpolated for a better visualization. The ’+’ are the weighted average of the regulated power of the two closed states surrounding the three open states.
The TSI is calculated using two closed states occurring at time t=0 s and t=360 s (state 1 and state 2) surrounding three open states . For each of the open states starting at time t=90 s, t=180 s and t=270 s, the diﬀerence between the reference (right) and regulated (left) powers is computed according to:
∆Popen (t) = Pref open (t) − Preg open (t) (1)
This diﬀerence is compared to the linear interpolation ∆P closed (t) of ∆Pclosed 1 (t) and
∆Pclosed 2 (t) where:
∆Pclosed 1 (t) = Pref closed 1 (t) − Preg closed 1 (t) (2)
∆Pclosed 2 (t) = Pref closed 2 (t) − Preg closed 2 (t) (3)
The ﬁnal TSI equation (Equation 5) uses the Solar Irradiance (SI) equation:
The diﬀerent parameters of Equation 4are described in what follows. Their uncertainties are summerised in Table 1. The diﬀerents parameters are :
- The absorption coeﬃcient of the cavity αabs : this coeﬃcient correspond to the amount of light trapped by the radiometric cavity. For DIARAD/SOVIM we have αabs = 0.999748 for both the left and right cavity with 45 ppm uncertainty. This coeﬃcient is calculated on the basis of the geometric dimension of the cavity and the measured hemispheric absorptivity of the black paint.This paint absorptivity is measured in ground using a synchronous detection setup. It is of the order of 0.972 ± 0.003 (Mekaoui, 2008)
- The eﬀect of backscattered radiation within the view limiting volume (Σ): we have Σ = 0.75 e − 5.
- The scattering of radiation by the front aperture (Σ ): we have Σ = 4.5 e − 5
- The diﬀraction of radiation (δ=0.0005)
- The thermal eﬃciency or eﬀective absoptivity of the cavity (αth ): from the part of the absorbed radiation in the cavity only a proportion is eﬀectively detected by the heat ﬂux sensor. The other part is dissipated by the external wall of the cavity. The eﬃciency of the cavity sensor association is determined in air and vacuum. We have for DIARAD/SOVIM: αth leftth = 0.99849089 and αth right = 0.99893007 in vaccum with 120 ppm uncertainty. In air αth left = 0.99586386 and αth right = 0.99584252
- The shutter correction (∆shutter ): this correction takes into account the contribution of the internal thermal emission of the active shutter. This contribution is removed when the shutter is open. For DIARAD/SOVIM the shutter correction is determined during deep space pointing. During this phase Equation 4 becomes.
It is therefore possible to derive ∆shutter . The contribution is of the order of 1.24 W.m−2 .
- The parasitic wire heating.
- The value of the resistor used to measure the heating currents.
- The correction for the time constant of the system (servocst ): during the ﬁrst open sequence following a closed sequence, the open powers acquired at seconds : 20, 50, 60 and 80 do not correspond to an equilibrium state. It is possible to introduce a factor (servocst ) to estimate the assymptotic value of each acquisition. For DIARAD/SOVIM the triple open sequence guarantee that equilibrium is reached during the third open state. This assymptotic value is used to compute the correction factor for each of the previous state.We have chosen to apply a factor equal to 1 after 90 s and not to use the measurements up to 90 s.
- Precision aperture at temperature T=20 Celcius: the precision apertures have been measured at NIST and NPL; we use the average of the NIST and NPL measurements. We have Sleft = 0.0000794094 m2 and Sright = 0.0000794533 m2 with 150 ppm uncertainty. The expansion coeﬃcient of the precision aperture is considered to be negligeable.
- •Distance, velocity and pointing corrections: the distance r between the radiometer and the center of the Sun, the radial velocity dr/dt of the satellite relativly to the Sun and the angle between the direction of the Sun and the radiometer (θ) are used to ﬁnally compute the TSI according to Equation 5.
The absolute accuracy of each channel is determined by the TSI equation (Equation 4 and 5). These equations are based on power and instruments parameters determination. The power absolute accuracy is of the order of 220 ppm in open state . Instruments parameters are determined in ground during the characterisation phase with a total uncertainty of 450 ppm (Table 1). As a consequence the uncertainty on the absolute measurement is of the order of 670 ppm which corresponds to ± 0.91 W.m−2. Figure 2 shows the SI measurements during the entire mission. These measurements are then corrected for the distance, velocity and pointing eﬀects in Figure 3. Both the left and right channels measurements are presented. DIARAD/SOVIM is nominally operating with its left cavity. The right cavity is used for ageing monitoring. The dispersion of DIARAD/SOVIM is of the order of 0.13 W.m−2 RMS. The diﬀerence between the channels is of the order of 0.25 W.m−2 . This diﬀerence is within the individual channel absolute accuracies.
For comparison DIARAD/VIRGO measurements on SOHO (Mekaoui & Dewitte,2009a) are displayed in the same graph with an oﬀset of - 1.2 W.m−2 .
Figure 4 shows the left and right channels measurements for three days of the mission (13, 14 and 15 of June 2008). All the measurements of an orbit are representd.
For this period of measurements, the level of the TSI from DIARAD/SOVIM was1364.50 ± 0.91 W.m−2 for DIARAD/SOVIM left channel and 1364.75 ± 0.91 W.m−2 for the right channel. Compared to TIM/SORCE measurements, DIARAD/SOVIM is 4 W.m−2 higher (see Figure 5). This adds new independant measurements of the TSI absolute level that do not validate the TIM/SORCE ﬁndings.
Figure 2: SI measurements (Equation 5) during the 6 months of mission for DIARAD/SOVIM left channel. The SI decreases and then increases as the Earth moves around the Sun. The eﬀect of the distance is corrected in Figure 3 The outlyers are caused by some reﬂections on the ISS.
Figure 3: TSI orbital means from DIARAD/SOVIM left (red dots) and right(blues crosses) channels. The dispersion of measurements is of the order of 0.13 W.m−2. The measurements during each orbit are presented for three days of observations in Figure 4. For comparison, DIARAD/VIRGO 3-minute measurements are also displayed. DIARAD/VIRGO data are shifted by -1.2 W.m−2.
Figure 4: TSI measurements from DIARAD/SOVIM left (black) and right (red)channels during each orbit for three days ( 13, 14 and 15) of June 2008 .
Figure 5: TSI measurements on their native scale, DIARAD/SOVIM measurements (dark blue +) diﬀer from TIM/SORCE by + 4 W.m−2.
In this paper we have given rigourous description of the DIARAD/SOVIM TSI data processing and the related uncertainty analysis. A new extended open sequence has been introduced to increase the TSI sampling rate. Compared to previous versions of the DIARAD instruments eﬀorts have been done to reduce the absolute uncertainty of the electrical power measurements and the precision area determination.
The uncertainty of the electrical power measurements is 220 ppm. The diﬀerence beteween the independent left and right channel measurements is as low as 0.25 W.m−2 and is within the absolute uncertainty limit of ± 0.91 W.m−2. Although DIARAD/SOVIM TSI is 1.2 W.m−2 lower than DIARAD/VIRGO TSI, it is still 4 W.m−2 higher than TIM/SORCE TSI.