Reconstruction of asteroid spin states from Gaia DR2 photometry

Reconstruction of asteroid spin states from Gaia DR2 photometry

J. Ďurech Astronomical Institute, Faculty of Mathematics and Physics, Charles University, V Holešovičkách 2, 180 00 Prague 8, Czech Republic
durech@sirrah.troja.mff.cuni.cz
   J. Hanuš Astronomical Institute, Faculty of Mathematics and Physics, Charles University, V Holešovičkách 2, 180 00 Prague 8, Czech Republic
durech@sirrah.troja.mff.cuni.cz
Received ?; accepted ?
Key Words.:
Minor planets, asteroids: general, Methods: data analysis, Techniques: photometric
Abstract

Context:In addition to stellar data, Gaia Data Release 2 (DR2) also contains accurate astrometry and photometry of about 14,000 asteroids covering 22 months of observations.

Aims:We used Gaia asteroid photometry to reconstruct rotation periods, spin axis directions, and the coarse shapes of a subset of asteroids with enough observations. One of our aims was to test the reliability of the models with respect to the number of data points and to check the consistency of these models with independent data. Another aim was to produce new asteroid models to enlarge the sample of asteroids with known spin and shape.

Methods:We used the lightcurve inversion method to scan the period and pole parameter space to create final shape models that best reproduce the observed data. To search for the sidereal rotation period, we also used a simpler model of a geometrically scattering triaxial ellipsoid.

Results:By processing about 5400 asteroids with at least ten observations in DR2, we derived models for 173 asteroids, 129 of which are new. Models of the remaining asteroids were already known from the inversion of independent data, and we used them for verification and error estimation. We also compared the formally best rotation periods based on Gaia data with those derived from dense lightcurves.

Conclusions:We show that a correct rotation period can be determined even when the number of observations is less than 20, but the rate of false solutions is high. For , the solution of the inverse problem is often successful and the parameters are likely to be correct in most cases. These results are very promising because the final Gaia catalogue should contain photometry for hundreds of thousands of asteroids, typically with several tens of data points per object, which should be sufficient for reliable spin reconstruction.

1 Introduction

The ESA Gaia mission (Gaia Collaboration et al., 2016b) has been in the science operations phase since August 2014. So far, there have been two Data Releases, the first in September 2016 (DR1, Gaia Collaboration et al., 2016a) and the second in April 2018 (DR2, Gaia Collaboration et al., 2018a). The main output of DR2 is accurate astrometric data for more than a billion stars. However, unlike DR1, DR2 also contains astrometric and photometric data for about 14,000 asteroids (Gaia Collaboration et al., 2018b).

Time-resolved photometry of asteroids, i.e. lightcurves, can be used for the reconstruction of the rotation period, spin axis orientation, and shape (Kaasalainen et al., 2002; Ďurech et al., 2007a; Hanuš et al., 2016; Marciniak et al., 2007, 2018; Michałowski et al., 2004; Kryszczyńska, 2013, for example). Also, photometry that is sparse in time with respect to the rotation period can be successfully used with the same lightcurve inversion method (Kaasalainen, 2004; Ďurech et al., 2007b, 2009, 2016; Hanuš et al., 2011, 2013). Gaia provides this type of sparse-in-time photometry with unprecedented accuracy. After the end of mission, these data will be used to determine periods, spins, and triaxial shape models (Cellino et al., 2006, 2007; Cellino & Dell’Oro, 2012; Santana-Ros et al., 2015). As shown by Santana-Ros et al. (2015), the probability of deriving a correct spin model is related to the shape (spherical asteroids have small lightcurve amplitudes), spin axis latitude (low-latitude asteroids are sometimes seen pole-on with small lightcurve amplitude), and the number of data points. Until now, real Gaia asteroid photometry was not available and the performance of inversion techniques was tested on simulated data. DR2 has changed this situation and we can now use real high-quality Gaia photometry and test whether the expectations were met. Here we use asteroid photometry released in DR2 with the aim of testing the limits of lightcurve inversion and the information content of the data. We also derive new asteroid models.

2 Inversion of Gaia asteroid photometry

The DR2 contains G-band brightness measurements with uncertainties for about 14,000 asteroids. The observations cover 22 months and the number of data points per object varies from a few to 50. As described by Gaia Collaboration et al. (2018b), the reported brightness values are constant for a single transit and they were computed as average values over the transit. Gaia Collaboration et al. (2018b) also tested the accuracy of the asteroid photometry and reached the conclusion that it is probably better than 1–2%. This is much better than the accuracy of sparse photometry from ground-based surveys, which is hardly better than 0.1 mag (Ďurech et al., 2009). A unique reconstruction of the shape/spin model is possible only if there are enough photometric data points with good accuracy covering a sufficiently wide interval of geometries. With ground-based surveys, the poor photometric quality is compensated with the number of data points, typically several hundred, observed over many apparitions. Even so, unique solutions are rare; the success rate of deriving a robust and reliable model is less than one percent (Ďurech et al., 2016). In the case of Gaia, the final catalogue will contain data that fulfil all three requirements: they will be very accurate, there will be several tens of measurements per object, and they will cover several apparitions for a typical main-belt asteroid. According to simulations, several tens of accurate measurements should be sufficient to derive a unique spin solution and an approximate shape (Santana-Ros et al., 2015).

Figure 1: Comparison between the best period derived from Gaia photometry with convex models (left) and ellipsoids (right). Each point represents an asteroid for which the best-fitting period from Gaia was determined and with a period in the LCDB with the uncertainty code . The number of Gaia observations is expressed by the colour. The left panel contains fewer points because convex models were used only when .

DR2 photometry allowed us to test how successful the inversion of real Gaia data is. We selected all 5413 asteroids for which the number of brightness measurements was . We computed the geometry with respect to the Sun and the Gaia spacecraft for each observation and processed the data in the same way as in Hanuš et al. (2011) and Ďurech et al. (2016, 2018) by using the lightcurve inversion method of Kaasalainen et al. (2001). Our approach is similar to that of Torppa et al. (2018) with the main difference that we do not deal with any error analysis. For each processed asteroid, we searched for a shape/spin model that gives the best fit to the data, measured by the lowest between the observed and modelled brightness. All data points were given the same weight; we did not take into account errors of individual measurements. The reason was that the relative formal errors are mostly below 2% (90% of all data points), and in this range the difference between, for example, 1% and 0.1% accuracy plays no role because the errors introduced by the model are larger (simplified shape approximation and scattering model assuming uniform albedo).

2.1 Rotation periods

As the first step, we computed periodograms using convex shapes and ellipsoids and tested the reliability of the formally best-fit period. In Fig. 1, we show the comparison between the best period derived from DR2 using either convex shapes or ellipsoids and the values compiled in the Lightcurve Database (LCDB) of Warner et al. (2009); we used the version from November 12, 2017. We used only reliable LCDB periods with the uncertainty code . The colour-coding correlates with the number of data points . Convex models do not provide any periods when (see the discussion below). The points concentrating on the diagonal line represent the correctly determined periods (Gaia and LCDB periods are the same). The points off the diagonal are likely incorrect Gaia periods because LCDB records with should be reliable. The minor diagonal in the left panel are false solutions with the derived period being half of the real one; this can happen with convex shapes as they produce lightcurves with only one minimum/maximum per rotation. Ellipsoidal models do not have this disadvantage of producing false half periods, but the periods based on a small number of points are often wrong. In general, when there are more data points, it is more likely that the derived period will be correct.

The dependence of the number of false periods on the number of data points is shown in Fig. 2. The fraction of correctly determined periods (defined as those that agree with LCDB values within ) increases above 0.5 when . For fewer points, the formally best periods are not reliable. The clustering of points around shorter periods is likely a consequence of the way the model is constructed; it is easier to formally fit the sparse points with an incorrect period that is shorter than the true period. For , the success rate seems to be high, but the sample size of asteroids in this range is very small.

We also tested if there is any difference in the photometric errors of Gaia data between the asteroids with correctly and incorrectly determined rotation periods. For the two bins with between 21–25 and 26–30 (where the fraction of correctly determined periods is about 50% and the number of periods is large), we compared the photometric errors of points belonging to asteroids with correctly determined periods with those that belong to asteroids with incorrect Gaia-based periods. The t-test did not reveal any significant difference in the means of these two groups. Also, the distribution of observations in time was very similar for the two groups. We did not reveal any statistical difference in, for example, the number of observations separated by minutes, which is the spacing related to the scanning pattern of Gaia corresponding to two field-of-view transits.

2.2 Spins and shapes

The best-fit periods discussed above are often just random global minima in the periodograms. To distinguish between random and real periods, we have to define some level of significance measured by the fit. We defined the uniqueness of the best solution by the depth of the minimum with respect to other local minima. The formula we used for the threshold is a modification of the formula we used in Ďurech et al. (2018); now there is no factor of 1/2. This is an arbitrary borderline that is based on a trade-off between the total number of new models and their contamination with incorrect models. Here is the number of degrees of freedom, which is formally the difference between the number of points and the number of parameters . For ellipsoids, and the parameters are the sidereal rotation period , the spin axis direction in ecliptic coordinates and , one parameter for the linear slope of the phase curve, and two parameters ( and ) for the ellipsoid axes ratios. With Gaia observations the phase angle is almost always , which means that the phase function can be reduced to only a linear part with one parameter (Kaasalainen et al., 2001). With convex models, we used the spherical harmonics representation of the order and degree of three (Kaasalainen et al., 2001), which corresponds to 16 shape parameters, so the total number of parameters is . For spin/shape reconstruction, we only used asteroids with .

When the number of points was small, in many cases we obtained RMS residuals of almost zero for many different periods. Such periodograms were excluded from the analysis. To avoid fitting noise, we only selected periodograms with all RMS values mag. Another requirement was that there should be only one period with RMS below 0.01. If there were more, we considered it a non-unique solution even if the threshold limit was satisfied. The verification procedure was the same as in Ďurech et al. (2018), see Fig. 3 there, with the only difference that we did not use for the degree and order of the spherical harmonics series. The visual inspection of the periodograms was crucial because in many cases we obtained false solutions for  h or  h. Even with a correct rotation period and pole, the corresponding shapes were often unrealistic with sharp edges and triangular pole-on silhouettes. This is a consequence of the order and degree of the spherical harmonics series () being too low, but with a small number of photometric points there is not enough information to reconstruct higher resolution models. In this sense, convex shape models derived from DR2 data should not be taken as real shapes; they are just models that fit the data best with the given resolution, and they are likely to change significantly when more data points are available and a higher degree of resolution is possible. On the other hand, the rotation periods and pole directions are not that sensitive to the resolution and they are more reliable (Hanuš & Ďurech, 2012).

2.3 Comparison with independent models

By processing all asteroids with and rejecting unreliable solutions, we derived models of 173 asteroids. Of these, 44 were already in the Database of Asteroid Models from Inversion Techniques (DAMIT, Ďurech et al., 2010) and we used them for an independent test of the accuracy of our solutions based on DR2. Most of our models agreed with those in DAMIT: their periods were the same within the errorbars and the mean difference between their pole directions was . However, there was a group of clear outliers with differences between pole directions . We looked in detail into these cases (seven in total). In one case – asteroid (2802) Weisell – the DAMIT model based only on sparse data (Hanuš et al., 2016) was clearly incorrect because the period search was done in the wrong local minimum. We removed this model from DAMIT. In five cases, the periods were very similar but differed more than their uncertainty, so the DR2-based solution was apparently a different local minimum leading to a different pole. In one case, the periods were completely different.

Figure 2: Success rate of deriving correct rotation periods (compared with LCDB) for ellipsoidal and convex models. The number above each histogram bar is the number of asteroids with periods in the LCDB with in that interval of data points.

2.4 New models

In Table 1, we list 129 new models and their spin axis directions (sometimes there are two possible solutions), the sidereal rotation period, and the period reported in the LCDB. The LCDB period agrees with our value in most cases. The asteroids for which the periods do not agree and the LCDB period is reliable (higher uncertainty code U) are marked with an asterisk.

To further check the reliability of these new models, we repeated the period search using reduced data sets. For each asteroid from Table 1, we randomly selected and removed 10% of the data points (i.e. 2–5 points) from the original data set. In most cases, this reduction had no effect and the new periodogram showed the same unique period. In 44 cases, the reduced data set still provided the same best period, but it did not pass the threshold limit and thus was not considered a unique solution. In one case the new period was different from the original one. These asteroids are marked with an exclamation mark. This test shows that with our definition of , we are at the critical limit of the number of data points in many cases. Removing just a few points may lead to formal rejection of the model even if it is correct (as is independently confirmed by the agreement between and ).

Because the geometry of observations is restricted mostly to the ecliptic plane, the lightcurve inversion usually produces two mirror shape solutions with about the same pole latitude and the difference in longitude of (Kaasalainen & Lamberg, 2006). However, there are a surprising number of solutions with just one pole direction in Table 1 (compared with the results of Ďurech et al., 2018, for example). This is likely caused by loosely constrained shapes that often are too elongated along the rotation axis. These shapes are removed by the pipeline. It means that the mirror pole solutions cannot be rejected even if they are not listed in Table 1.

3 Discussion

Our analysis of Gaia DR2 asteroid photometry shows that the excellent photometric accuracy enables us to derive reliable spin directions and rotation periods from the first 22 months of Gaia observations. Although the number of unique models derived from DR2 is small compared to the number of all asteroids with photometry, this is mainly because the number of asteroids with measurements is still limited. However, the prospect for the next data releases is very high. With more than 50 points covering several years, the inversion should provide unique results in most cases (apart from very spherical asteroids or those with extreme rotation) and the shape models will be more robust. Moreover, the photometric calibration in the future data releases should be even more accurate due to further improvement of the reduction pipeline that in the case of DR2 did not include correction for flux loss due to moving objects or a more sophisticated filtering of outliers, for example (Gaia Collaboration et al., 2018b).

Before DR3 (likely in the first half of 2021), the full potential of DR2 can be exploited when Gaia photometry is combined with archived lightcurves or sparse photometry from ground-based surveys. In this way, even a small number of accurate Gaia photometric measurements with a higher statistical weight can help to reconstruct uniquely the shape and spin state of many asteroids with the lightcurve inversion method.

Acknowledgements.
The authors were supported by the grant no. 18-04514J of the Czech Science Foundation. This work has made use of data from the European Space Agency (ESA) mission Gaia (https://www.cosmos.esa.int/gaia), processed by the Gaia Data Processing and Analysis Consortium (DPAC, https://www.cosmos.esa.int/web/gaia/dpac/consortium). Funding for the DPAC has been provided by national institutions, in particular the institutions participating in the Gaia Multilateral Agreement.

References

  • Cellino et al. (2006) Cellino, A., Delbò, M., Zappalà, V., Dell’Oro, A., & Tanga, P. 2006, Advances in Space Research, 38, 2000
  • Cellino & Dell’Oro (2012) Cellino, A. & Dell’Oro, A. 2012, Planet. Space Sci., 73, 52
  • Cellino et al. (2007) Cellino, A., Tanga, P., Dell’Oro, A., & Hestroffer, D. 2007, Advances in Space Research, 40, 202
  • Ďurech et al. (2018) Ďurech, J., Hanuš, J., & Alí-Lagoa, V. 2018, A&A, 617, A57
  • Ďurech et al. (2016) Ďurech, J., Hanuš, J., Oszkiewicz, D., & Vančo, R. 2016, A&A, 587, A48
  • Ďurech et al. (2007a) Ďurech, J., Kaasalainen, M., Marciniak, A., et al. 2007a, A&A, 465, 331
  • Ďurech et al. (2009) Ďurech, J., Kaasalainen, M., Warner, B. D., et al. 2009, A&A, 493, 291
  • Ďurech et al. (2007b) Ďurech, J., Scheirich, P., Kaasalainen, M., et al. 2007b, in Near Earth Objects, our Celestial Neighbors: Opportunity and Risk, ed. A. Milani, G. B. Valsecchi, & D. Vokrouhlický (Cambridge: Cambridge University Press), 191
  • Ďurech et al. (2010) Ďurech, J., Sidorin, V., & Kaasalainen, M. 2010, A&A, 513, A46
  • Gaia Collaboration et al. (2018a) Gaia Collaboration, Brown, A. G. A., Vallenari, A., et al. 2018a, A&A, 616, A1
  • Gaia Collaboration et al. (2016a) Gaia Collaboration, Brown, A. G. A., Vallenari, A., et al. 2016a, A&A, 595, A2
  • Gaia Collaboration et al. (2016b) Gaia Collaboration, Prusti, T., de Bruijne, J. H. J., et al. 2016b, A&A, 595, A1
  • Gaia Collaboration et al. (2018b) Gaia Collaboration, Spoto, F., Tanga, P., et al. 2018b, A&A, 616, A13
  • Hanuš & Ďurech (2012) Hanuš, J. & Ďurech, J. 2012, Planet. Space Sci., 73, 75
  • Hanuš et al. (2013) Hanuš, J., Ďurech, J., Brož, M., et al. 2013, A&A, 551, A67
  • Hanuš et al. (2011) Hanuš, J., Ďurech, J., Brož, M., et al. 2011, A&A, 530, A134
  • Hanuš et al. (2016) Hanuš, J., Ďurech, J., Oszkiewicz, D. A., et al. 2016, A&A, 586, A108
  • Kaasalainen (2004) Kaasalainen, M. 2004, A&A, 422, L39
  • Kaasalainen & Lamberg (2006) Kaasalainen, M. & Lamberg, L. 2006, Inverse Problems, 22, 749
  • Kaasalainen et al. (2001) Kaasalainen, M., Torppa, J., & Muinonen, K. 2001, Icarus, 153, 37
  • Kaasalainen et al. (2002) Kaasalainen, M., Torppa, J., & Piironen, J. 2002, Icarus, 159, 369
  • Kryszczyńska (2013) Kryszczyńska, A. 2013, A&A, 551, A102
  • Marciniak et al. (2018) Marciniak, A., Bartczak, P., Müller, T., et al. 2018, A&A, 610, A7
  • Marciniak et al. (2007) Marciniak, A., Michałowski, T., Kaasalainen, M., et al. 2007, A&A, 473, 633
  • Michałowski et al. (2004) Michałowski, T., Kwiatkowski, T., Kaasalainen, M., et al. 2004, A&A, 416, 353
  • Santana-Ros et al. (2015) Santana-Ros, T., Bartczak, P., Michałowski, T., Tanga, P., & Cellino, A. 2015, MNRAS, 450, 333
  • Torppa et al. (2018) Torppa, J., Granvik, M., Penttilä, A., et al. 2018, Advances in Space Research, 62, 464
  • Warner et al. (2009) Warner, B. D., Harris, A. W., & Pravec, P. 2009, Icarus, 202, 134

Appendix A List of new models

Asteroid U method
number name/designation [deg] [deg] [deg] [deg] [h] [h]
205 Martha 358 3 46 C
! 217 Eudora 131 317 11 3 29 C
333 Badenia 3 179 3 25 CE
561 Ingwelde 89 68 3 22 C
580 Selene 62 66 269 50 3 34 C
581 Tauntonia 203 2 32 C
659 Nestor 6 3 42 C
! 723 Hammonia 146 22 330 21 3 31 CE
! 838 Seraphina 18 4 192 32 2 28 C
! 842 Kerstin 18 78 23 CE
876 Scott 105 262 2 27 E
906 Repsolda 108 3 23 E
961 Gunnie 37 24 220 7 23 CE
976 Benjamina 354 80 3 23 E
1029 La Plata 119 30 274 44 3 21 C
! 1107 Lictoria 82 57 303 57 3 31 CE
1118 Hanskya 224 2 39 C
! 1165 Imprinetta 39 3 25 C
! 1168 Brandia 310 65 3 21 E
! 1220 Crocus 26 48 165 80 3 23 E
1431 Luanda 79 54 3 23 E
1437 Diomedes 320 1 3 26 E
1533 Saimaa 332 3 25 C
! 1540 Kevola 96 3 32 E
1542 Schalen 92 41 268 52 3 23 E
! 1604 Tombaugh 166 34 323 36 2+ 35 CE
! 1647 Menelaus 157 26 330 23 3 28 E
! 1762 Russell 8 89 3 22 C
! 1767 Lampland 186 340 29 CE
! 1786 Raahe 88 45 3 25 C
! 1799 Koussevitzky 58 3 22 E
1849 Kresak 140 61 2 27 E
1873 Agenor 117 56 2 26 E
! 1939 Loretta 21 201 1 33 CE
! 1975 Pikelner 74 38 32 C
2090 Mizuho 231 2+ 22 E
2104 Toronto 306 3 24 E
2111 Tselina 102 19 282 53 3 46 C
! 2127 Tanya 64 61 2 35 CE
2147 Kharadze 180 347 2 26 CE
2192 Pyatigoriya 139 44 338 81 27 C
2203 van Rhijn 63 2 21 E
2230 Yunnan 2 64 154 73 21 E
! 2386 Nikonov 52 51 242 33 21 E
2397 Lappajarvi 77 251 2 28 C
2429 Schurer 235 3 42 CE
! 2587 Gardner 351 49 2 30 E
2627 Churyumov 141 307 3 29 CE
2634 James Bradley 120 21 E
2683 Brian 112 294 33 CE
2686 Linda Susan 51 268 31 C
2760 Kacha 101 3 29 E
2884 Reddish 201 26 CE
3131 Mason-Dixon 294 57 2 41 C
3134 Kostinsky 276 88 2 41 CE
3210 Lupishko 42 55 2 35 E
3325 TARDIS 142 338 48 CE
3374 Namur 217 34 27 C
! 3420 Standish 355 76 25 CE
3451 Mentor 81 18 3 29 E
! 3525 Paul 10 214 27 2 21 C
3565 Ojima 99 66 294 68 32 C
! 3776 Vartiovuori 190 2 24 E
3788 Steyaert 67 29 CE
! 4075 Sviridov 17 89 329 62 22 C
4131 Stasik 31 203 25 CE
4271 Novosibirsk 105 58 320 60 3 27 CE
! 4352 Kyoto 155 42 341 28 3 32 CE
! 4366 Venikagan 14 38 188 37 32 C
4369 Seifert 15 170 3 29 C
! 4451 Grieve 177 341 3 28 E
! 4575 Broman 181 303 26 CE
4613 Mamoru 280 3 41 CE
4732 Froeschle 105 289 26 E
! 4930 Rephiltim 264 3 30 CE
! 5059 Saroma 33 215 3 22 E
5130 Ilioneus 128 303 3 22 E
5138 Gyoda 157 23 E
! 5285 Krethon 182 2 24 E
5344 Ryabov 9 90 286 71 26 E
! 5385 Kamenka 168 23 2 25 E
5594 Jimmiller 260 2 26 C
5755 1992 OP7 51 78 190 88 29 E
5883 Josephblack 38 7 228 21 27 CE
6173 Jimwestphal 93 3 27 C
6338 Isaosato 146 84 30 CE
6665 Kagawa 40 24 E
! 6794 Masuisakura 154 345 3 23 E
! 7022 1992 JN4 4 84 337 56 2 28 E
7238 Kobori 42 238 37 C
7457 Veselov 16 88 25 E
7458 1984 DE1 48 39 232 17 2 25 C
! 7616 Sadako 185 339 24 C
! 7650 Kaname 58 2 24 CE
! 8066 Poldimeri 205 342 28 E
8292 1992 SU14 32 20 1 25 E
8443 Svecica 261 48 3 31 C
8770 Totanus 118 294 41 C
9299 Vinceteri 228 22 E
10406 1997 WZ29 14 184 22 E
10763 Hlawka 128 30 21 E
10790 1991 XS 171 52 357 55 33 C
! 11429 Demodokus 31 2 25 E
! 11682 Shiwaku 25 68 206 74 21 CE
12003 Hideosugai 32 27 E
12291 Gohnaumann 252 52 22 C
13446 Almarkim 48 54 230 53 25 E
13809 1998 XJ40 66 225 2 22 E
14268 2000 AK156 119 3 26 E
! 14362 1988 MH 188 6 3 21 C
14376 1989 ST10 139 58 340 68 3 28 E
! 14410 1991 RR1 342 32 E
15105 2000 BJ4 108 71 2 29 CE
15496 1999 DQ3 23 89 27 E
15955 Johannesgmunden 327 22 E
16029 1999 DQ6 94 85 3 25 E
! 16771 1996 UQ3 49 63 253 41 28 CE
17567 Hoshinoyakata 325 2 22 E
18156 Kamisaibara 23 221 24 C
18666 1998 FT53 316 49 36 CE
20721 1999 XA105 74 259 11 34 C
! 21904 1999 VV12 225 73 22 E
22972 1999 VR12 90 57 27 CE
! 24324 2000 AT51 132 53 26 E
25846 2000 EF93 123 27 300 55 29 E
32497 2000 XF18 214 15 22 E
40165 1998 QP102 331 86 2 28 E
47678 2000 CT75 159 54 2 23 E
! 51857 2001 OA105 78 59 257 29 2 23 CE
Table 1: continued.
Comments 0
Request Comment
You are adding the first comment!
How to quickly get a good reply:
  • Give credit where it’s due by listing out the positive aspects of a paper before getting into which changes should be made.
  • Be specific in your critique, and provide supporting evidence with appropriate references to substantiate general statements.
  • Your comment should inspire ideas to flow and help the author improves the paper.

The better we are at sharing our knowledge with each other, the faster we move forward.
""
The feedback must be of minimum 40 characters and the title a minimum of 5 characters
   
Add comment
Cancel
Loading ...
307553
This is a comment super asjknd jkasnjk adsnkj
Upvote
Downvote
""
The feedback must be of minumum 40 characters
The feedback must be of minumum 40 characters
Submit
Cancel

You are asking your first question!
How to quickly get a good answer:
  • Keep your question short and to the point
  • Check for grammar or spelling errors.
  • Phrase it like a question
Test
Test description