Our variable star (part one): Getting trends in solar observations right

Variations in the Sun's output affects our climate. How much the Sun change's by can only be determined by observing it continuously over a long period of time..

Variations in the Sun’s output affects our climate. How much the Sun changes by can only be determined by observing it continuously over a long period of time.

In order to quantify the role that the Sun has played in recent climate change, we need to know how the Sun changes over time. The only way to do that is to observe the Sun using satellites. But, the instruments on board do not last long enough to tell us how the Sun really varies on the multi-decadal time scales important for climate change. And, no two instruments tell quite the same story. We need to know that the observations are right before we link them to longer datasets and estimate how the Sun has varied in the past. Deciding how to put those instrument stories together can ultimately lead us to attributing a greater or lesser role in climate change to the Sun.

(Top plot) Total solar irradiance observations from different instruments between 1978 and 2013 with the published absolute values. (Bottom three plots) The three different composites of the observations. Colours in the composites match the colours of the instruments in the top plot. The thick black trends are 81-day running averages while the think black line is for visual guidance. This plot is reproduced from PMOD/WRC [1].

Figure 1: (Top plot) Total solar irradiance observations from different instruments between 1978 and 2013 with the published absolute values. (Bottom three plots) The three different composites of the observations. Colours in the composites match the colours of the instruments in the top plot. The thick black trends are 81-day running averages while the thin horizontal line is for visual guidance. This plot is reproduced from PMOD/WRC [1].

In principle, detecting changes in the Sun should be easy. Put a satellite in space and point your instrument at the Sun. There is nothing between the two to get in the way, so that should be it, right? Sadly not. Take a look at the top plot of Figure 1 [1]. Here we have all the separate observations of the total energy per second received from the Sun, through one square metre at the top of our atmosphere, since the 1970s: the total solar irradiance. It is immediately obvious that, even at the same time, the Sun appears to have a different brightness. That cannot be the case. While this is a problem when trying to account for all the incoming and outgoing radiation on Earth, that is really not the big issue. In fact, we know now that TIM/SORCE in that plot is at the correct value [2], but that’s another snack.

The serious problem is the trend. Let’s pick one of those datasets as our ‘true’ answer and then line all the other datasets up so they have the same value when they overlap. But wait, sometimes we have two or more datasets showing slightly different yearly or longer trends. Which do we choose to be correct?

Well, we need to know why the data might be different. The main problem when observing the Sun is that the very thing we are measuring, the light, is actually damaging our instrument and altering the way it responds: the same amount of change in the Sun in the first year might be seen, by degraded instrument eyes, as larger or smaller the next year. This affects the long-term trend. Since we can’t visit our satellite to do an eye test, and with the addition of breakdowns and solar `explosions’ (coronal mass ejections that hurl energetic particles into the solar system and sometimes towards Earth), it is difficult to be sure exactly how the instrument has been affected. We aren’t helpless though. We have some other indicators of what the Sun might be doing. We also know some physics that can guide us on how the Sun might have damaged the instruments and what the limits are on how much the Sun may have changed by over time. So, we have a way of correcting the data back to what it should be.

This has been done by different groups and there are three contenders for the crown of ‘best composite’ of the data (Figure 1, lower three plots). These composites show how the Sun may have varied over the last three ‘solar cycles’, approximately 11-year periods over which the Sun’s total irradiance increases by 0.1% (or about 1 Wm-2) before dropping back down in to a quiet, ‘minimum’ state. On time scales less than a few years, the three datasets show similar behaviour. The important thing for climate is the change on multi-decadal time scales – for the Sun this time scale is captured in the underlying, minimum state that it returns to at the end of each cycle. But, we only have 35 years of data here. This gives us (essentially) three data points. However, the composites tell very different stories: PMOD shows a decrease each minimum, ACRIM goes up and then down, while IRMB shows an upward trend. These different trends are the result of choosing different datasets to build the composites and how those data were adjusted to account for the degradation or any jumps in the data.

The key point here is that this observational record is used as the basis upon which to extend the record of the Sun back to the 18th century. From this, the Sun’s contribution to global temperature change can be calculated. How much we estimate the Sun to have contributed can depend partly on by how much, and in what way, those minima vary. We will discuss that quantity in part two of this snack.

Let’s get back to the problem of those pesky composites. How do we assess which one is best? A good way is to use models based on our knowledge of the surface features that produce changes in the Sun’s output. In addition, we want the model to be completely independent of the composites. We have a model that can do that [3]. In Figure 2 we have smoothed the three composites and put the model result [4] on top. The uncertainty range is independent of the composites and the nominal values match most closely with PMOD. The PMOD composite has had more consideration of instrument degradation than the others and, together with the agreement from the independent model, this suggests that PMOD is probably our best bet so far for a total solar irradiance record.

Figure 3: The three composites from Figure 1 in this plot are smoothed here. The SATIRE-S model [4,5] and its uncertainty range are also plotted. All data are normalised to the value of the SORCE/TIM measurements as in the top plot of Figure 1.

Figure 2: The three composites from Figure 1 in this plot are smoothed here. The SATIRE-S model [3,4] and its uncertainty range are also plotted. All data are normalised to the value of the SORCE/TIM measurements as in the top plot of Figure 1.

In my next snack we will look at the ways we can extend the record of total solar irradiance into the past. We can then gauge how much the Sun may have changed in the last three centuries.

Will’s final thought: We are not quite at the end of the story yet regarding the observations. Work is now underway, involving all the instrument scientists available (and more), to fold in all the uncertainties and produce a single, ‘master’ composite [5]. After all, there is only one Sun. Also, the total solar irradiance is the combined total of the full spectrum of light. So, even when we have the best record of total solar irradiance, we must remember that it is not the total solar irradiance, but the different parts of the spectrum, that interacts with the Earth. Over the solar cycle, the light in some regions of this spectrum can vary by 1000 times more, in relative terms, than the combined total does. How this highly variable energy physically interacts with our planet is something we will investigate in a future snack.

References:
[1] Physikalisch-Meteorologisches Observatorium Davos World Radiation Center
[2] Kopp, G., Lean, J.L., A new, lower value of total solar irradiance: Evidence and climate significance, GRL, 38, L01706, 2011
[3] Krivova N. A., Solanki S. K., Fligge M., Unruh Y. C., Reconstruction of solar irradiance variations in cycle 23: Is solar surface magnetism the cause?, A&A, 399, L1, 2003
[4] Ball, W.T., Unruh, Y.C., Krivova, N.A., Solanki, S., Wenzler, T., Mortlock, D.J., Jaffe, A.H., Reconstruction of total solar irradiance 1974–2009, A&A, 541, A27, 2012
[5] ISSI TSI Team: An Assessment of the Accuracies and Uncertainties in the Total Solar Irradiance Climate Data Record

DISCLAIMER

Advertisements

About Will Ball

I am a PostDoc in the Physics Department at Imperial College London and I am co-director of ClimateSnack. My research interests lie in two main areas: total and spectral solar irradiance variability, on all timescales from the solar rotation to millenia; and the interaction of the Sun with the Earth's atmosphere, in particular with stratospheric ozone.
This entry was posted in Climate, Solar and tagged , , , , . Bookmark the permalink.