********************************************************************* Response to referee's comments, suggestions and questions. ********************************************************************* Aki Takeda, Loren Acton and Nicole Albanese for the manuscript, "Solar Cycle Variation of Coronal Temperature, Emission Measure, and Soft X-ray Irradiance Observed with Yohkoh Soft X-ray Telescope" [Summary] We appreciate referee's careful reading of our manuscript and many useful suggestions. As most major revision, we added the Figure 14 to show 'pseudo' DEM of the corona on typical days in sample years in response to referee's comment No.6. We performed data refinement (bad data removal) in relation to referee's comment No.3 and updated Fig.4, 5 and 7 (but does not affect the results). The data refinement was also performed on the Northern hemispheric data used in the Discussion section, and Fig. 11 and 13 were also updated. We added large amount of description to the Discussion, so the paragraphs were separated into 2 subsections (4.1 "Disk-integrated v.s. spatially- resolved analysis" and 4.2 "Three temperature component analysis"). ---------------------------------------------------------------------- [1: I consider the results from the alternative method of an analysis at the image pixel level as being the more exciting results in this paper, so I suggest you expand the abstract a bit to reflect more of those results. For example, you can mention the cooler temperature emission measure trend that appears out of phase with solar cycle activity (that is, higher emission measure at cycle minimum instead of lower emission measure). You can also mention that the soft X-ray irradiance with both methods are very similar, thus the alternative method provides some validation for the simpler full-sun method.] ---------------------------------------------------------------------- In the previous manuscript, We could not fully include the results from the spatially-resolved analysis because of the 250 words limitation of the ApJ abstract. In the new manuscript, we decided to omit the description of our irradiance compared with GOES, and added the description on the irradiance from the spatially- resolved method, and results from the 3 temperature component analysis. We appreciate referee's suggestion to include more results from the spatially-resolved analysis. However, our 3 temperature component analysis is photometrically challenging, especially for the period of Solar Minimum around 1996. It is because the quality of stray light subtraction on the thinner filter(Al.1) critically affect the resulting filter-ratio temperature analysis, even though only using signals from the Northern hemisphere. In response to referee's question No.7(below), we conducted an additional analysis on the temperature and emission measure of 1996 data in attempt to derive their minimum values observed by SXT. What is found, however, is that the enhancement of the cool component (<1.5MK) early 1996 is most probably due to the under-correction of the stray light component on the Al.1 images. We will therefore withhold the discussion about the minimum state of the solar corona, i.e., the statement, "the cool component is enhanced in early 1996.", by frankly stating that the current analysis method do not optimized for the low signal period and do not have enough reliability. Regarding the other outcome from the 3 component analysis, i.e., different ratio of high temperature component between the descending(1992-1995) and ascending (mid 1998-2001) phases, on the contrary, it was based on the data with enough signal level and thus no reason to reject. We will have this as the main result from the 3 component analysis. ---------------------------------------------------------------------- [2. Your uncertainty estimate does not seem to include uncertainties for pre-flight instrument calibration and instrument degradation trends. With irradiance uncertainties of 20-30%, perhaps you have included them already but didn't discuss them in your paper. Anyway, I suggest that you also discuss (and include) additional uncertainties for instrument pre-flight and in-flight calibrations in your calculation for the total uncertainties for the irradiance, temperature, and emission measures.] ---------------------------------------------------------------------- Our pre-flight calibration data are mainly used to finalize the SXT effective area functions and unavailable for estimating other source of uncertainties of Yohkoh/SXT observation. Instead, a series of on-orbit calibration were performed and the results are reviewed in the following paper: Acton, L.W.,(2016) Solar Physics, vol.291, p.643. We added the description about the uncertainties by giving a brief summary of the above paper. It was added in subsection 3.1, as the fourth paragraph. "There are certainly other sources of uncertainties, as described in detail by Acton (2016). They are however difficult to quantify and not included in the uncertainties calculated by SXT's standard routine for data reduction (sxt_prep.pro): Scattered X-rays and telescope vignetting are corrected in the standard routine, but the errors from these process are difficult to quantify. The uncertainty in CCD amplifier gain will be as large as 10%, and this directly propagates to the calculation of the emission measure and irradiance as the same amount of systematic error. There is an orbital drift in CCD dark current, caused by the UV flood operation performed at the beginning of every daylight pass to relieve the damage from ionizing radiation. This drift is corrected in the standard reduction routine with a few percent of uncertainty, which may be considerable when the signal level of data is extremely low." Additionally, we made a minor update of the description on the uncertainty estimation due to stray light correction. ---------------------------------------------------------------------- [3. Figure 4: Why is the emission measure smoother looking over solar cycle than the irradiance or temperature trends? Also for this figure, the temperature results seem to have a larger spread from their smoothed trend than that compared to irradiance and emission measure. Does this larger spread in temperature signify larger uncertainty for the temperature result or could it be more related to solar intrinsic variability (e.g. flare influences)?] ---------------------------------------------------------------------- While emission measure and irradiance vary nearly 2 orders of magnitude through the solar cycle, temperature varies just a factor of 4 (0.6 in log scale). The plotting range applied for temperature in the initial Fig.4 ([5.8,7.0]MK in log scale) was too narrow and the scattered data points were emphasized. The spread of temperature data points appears to be larger, but actually is the same level as that of emission measure. The width of data spread (distance from the smoothed line) is within 0.5 in log scale for both cases. The eccentric data (data points far from the smoothed line) are in most cases due to the change of solar activity between the exposure times of image pairs (growing or decay phase of flares or transient brightenings). Saturation of small brightest area, cosmic ray hit, and low S/N data also result in eccentric temperature values. Many of these eccentric data (with log_Te over 7.4 or below 5.8) had been eliminated in the previous study before taking daily average, but we could identify 20+ new inadequate pairs in the course of this investigation. We therefore additionally removed them from the data set to be daily averaged, although the spread of daily data points was not improved significantly. The plots in Figure 4 were updated using the refined data set, and the plotting range for the temperature was expanded to [5.5,7.5]. Figure 5 (2D histogram) and 7 (scatter plots) were also updated accordingly. Also, we added the following description on the Fig.4 at the end of the first paragraph of subsection 3.1. "Over the whole mission period, coronal temperature smoothed over 14 days varies within a factor of 4 (most data points of logTe range from 6.1 to 6.7). On the other hand, emission measures smoothed over the same period vary with a factor of ~30, that is equivalent to the logEM range of ~1.5. The emission measure is calculated from the filter-ratio temperature using the coronal signal from either image of the pair (the Al.1 image in this work) and the corresponding temperature response of the filter (cf., Fig.2). When the derived temperatures (logTe) approaches to the value 7.0 at which the Al.1 filter has the highest response, the corresponding emission measures are calculated low to explain the observed signal level. Such extraordinary low emission measures are observed in 1992, 1996, and 2001, associated with flare activities. The variation of irradiance extends over 2 orders of magnitude, showing similar variation with that of original signals. The data points that have outstandingly higher(lower) temperature(emission measure) due to flare activities are canceled each other (although not completely) in the irradiance plot." ---------------------------------------------------------------------- [4. Figure 10 Caption: Change the "SXT irradiance for 0.1 to 0.1 nm" to "SXT irradiance for 0.1 to 1.0 nm".] ---------------------------------------------------------------------- Thanks for catching this error. We corrected it. ---------------------------------------------------------------------- [5. I suspect that the trend of the cool component in Figure 13 is related to how the coronal hole area evolves over the solar cycle. You could consider that option and if so, include references about coronal hole evolution in your paper. For example, Lowder, Qiu, and Leamon (Solar Physics, 292, ID 18, 2017) paper's Figure 3 shows coronal hole area trend that overlaps with your data for 1996-2002, and their coronal hole area trend looks similar to your cool corona emission measure trend.] ---------------------------------------------------------------------- It is reasonable to relate the variation of the cool component with the evolution of coronal holes, because the coronal holes are the major source of cool component plasma. We appreciate the referee to introduce the work on the long-term evolution of coronal holes by Lowder et al. (2017). The result of our additional investigation is as follows: In 1996, polar coronal holes are strong and takes large area (according to Lowder et al. (2017). However, it does not contribute much to our disk observation of 1996 corona. Since there is no AR or low-latitude coronal hole on the disk, the corona that SXT observed in 1996 should be categorized as the quiet Sun. It is therefore the quiet Sun that contributes to the emission measures of the cool component during the solar minimum. There are several periods after 1999 in which a large disk (or low latitude) coronal hole takes significant potion of the Northern hemispheric corona (e.g., early 1999 and early 2000). http://ylstone.physics.montana.edu/ylegacy/qlook/2000/dailyim_20000128.html In such periods, the EM of the cool component is enhanced, but does NOT reach (~30% of) the values as of 1996. Following the the findings in the above, we added the following description about the Figure 13 in the first paragraph of the subsection 4.2. "While the medium and hot components show the similar curve as the irradiance (bottom plot in Fig.11), the cool component appears to increase roughly by a factor of 10 in early 1996 when the solar activity is around minimum. However, this enhancement is likely to be caused by the under-correction of the stray light component on the Al.1 filter images, because the subtraction of the stray light component is difficult in this period due to the coronal signal level of the corona. Since the current method of stray light calibration is not optimized for this period, we need further analysis to discuss the minimum state of the corona. In the same vein, the 1996 plot of Figure 14 should be interpreted with caution. The distribution of the cool component may be corrected by the improvement of the stray light calibration. Outside this tricky period, the emission measures of each temperature component is more reliable and worth discussing. The cool component dominates other components until mid 1997, then drops rapidly during the year of 1998. This is the similar trend to that of observed coronal hole area reported by Lowder et al.(2017) with SOHO/EIT observations. According to their Figure 3, the strong polar magnetic field of the Solar Cycle 23 disappears rapidly in 1998. It is reasonable to associate the variation of cool component with the evolution of coronal hole area, as a source region of cool plasma. However, it is the quiet Sun corona that contributes substantially to the cool component observed by the SXT. It is typically shown in the 1996 plot of Figure 14, when the strong polar coronal hole does not affect much to the emission measures observed by SXT. The coronal holes appeared at low latitude do contribute to the cool component (see the January, 2000 plot in Figure 14), but their effect is mild due to their low emission measure. They will rather contribute through the depletion of hot plasma in the observed corona. We do not recognize the immediate relation between the temporal variation of the cool component observed by SXT and that of low-latitude coronal hole area reported by Lowder et al.(2017). Our data in Figure 13 is only for the Northern hemisphere, but the result differ little when we use the whole Sun area." ---------------------------------------------------------------------- [6. With your alternative method of analysis for temperature and emission measure at the image pixel level, you could compose a DEM for each image (or average image for a day) by co-adding all of the pixel temperature-emission measure results from a single image into discrete bins of log(Temperature). It would be interesting to compare this (pseudo) DEM result over the solar cycle; for example, you could show those results for a "typical" image in 1992, 1996, and 2000 to match the time frames as shown in Figures 7 and 12. This additional analysis has potential to elucidate the key components of soft X-ray variability (namely, quiet sun, coronal hole, active region).] ---------------------------------------------------------------------- We added a new Figure 14 to demonstrate the plots of the "pseudo" DEM for the 4 typical phases in the solar cycle: 12 February, 1992 (descending phase), 15 February, 1996 (near solar minimum), 23 January, 2000 (ascending phase with a large disk coronal hole), and 24 April, 2000 (ascending phase with active regions). We used these plots to discuss the mild effect of coronal holes to the cool component plasma, and to explain the larger ratio of hot component plasma in the ascending phase than in the descending phase. Please take a look at the descriptions added to the subsection 4.2, which is dedicated to the results from the 3-temperature component analysis. ---------------------------------------------------------------------- [7. You could discuss the state of the corona at solar cycle minimum more, as one way to put your results about the sun into context better for the stellar scientists. In particular, the lowest energy state (with log(Temp) of about 6.1) of the corona is still a "hot" corona compared to the photosphere temperature. Although not measured at the same time as SXT, full-sun spectral X-ray measurements from the recent MinXSS CubeSat also show similar lower limit for the corona at about 1.7 MK [Woods et al., Ap J, 835, 122, 2017]. This Woods et al. [2017] paper also provides comparisons to GOES XRS and that comparison also confirms the lower irradiance than GOES XRS during solar cycle minimum as you so clearly point out in your paper.] ---------------------------------------------------------------------- We conducted an additional analysis on the temperature and emission measure of 1996 data in attempt to derive their minimum values observed by SXT. What is found, however, is that the enhancement of the cool component (<1.5MK) early 1996 is most probably due to the under-correction of the stray light component on the Al.1 images. We selected several Al.1 images with good condition (longer exposure time, free from SAA or cosmic ray features, and less affected by other artificial patterns than the stray light). We then tried manually subtracting the stray light component until the residual stray light pattern is disappeared (by eye), and derived the maps of temperature and emission measure. The new result shows that the emission measure of the cool component is comparable or slightly lower than that of the medium component (1.5