New Methanol Maser Transitions and Maser Variability Identified from an Accretion Burst Source G358.93-0.03

The high-mass young stellar object G358.93-0.03 underwent an accretion burst during the period from 2019 January to June. Given its extraordinary conditions, a number of new maser transitions may have been naturally excited during the burst stage. Searching for new maser lines and monitoring maser variability associated with the accretion burst event are important for understanding the complex conditions of the massive star formation toward G358.93-0.03. In this work, using the Shanghai 65 m Tianma Radio Telescope, we continuously monitored the multiple maser (including methanol and water) transitions toward G358.93-0.03 during the burst in the period from 2019 March 14 to May 20. There were 23 CH3OH maser transitions and one H2O maser transition detected from the monitoring. Nearly all the detected maser transitions toward this source have dramatic variations in their intensities within a short period of ∼2 months. Eight new methanol transitions from G358.93-0.03 were identified to be masering in our observations based on their spectral profile, line width, intensity, and the rotation diagram. During the monitoring, the gas temperature of the clouds in the case of saturated masers can show a significant decline, indicating that the maser clouds were going through a cooling process, possibly associated with the propagation of a heat wave induced by the accretion burst. Some of the maser transitions were even detected with the second flares in 2019 April, which may be associated with the process of the heat-wave propagation induced by the same accretion burst acting on different maser positions.


Introduction
Episodic accretion comprises prolonged quiescent and transient periods punctuated by intense bursts of accretion. Generally, the quiet stage has a low accretion rate and luminosity along with a cooling of the disk. However, most of the protostellar mass is accumulated in the burst stage with luminosity outbursts and the heating and stabilization of the surrounding disk (Vorobyov & Basu 2006;Stamatellos et al. 2011). There is observational and theoretical evidence that accretion of material is often episodic in early evolution for low-mass star formation (Herbig 1977;Peneva et al. 2010). Further evidence for episodic accretion comes from the periodically spaced knots seen in bipolar jets (Reipurth 1989). Apart from that, the luminosity problem provides indirect observational support for episodic accretion (Kenyon et al. 1990;Evans et al. 2009), as the luminosity outbursts are intermittent and most protostars are observed between bursts. Recent studies have shown that massive star formation can also experience phenomena similar to episodic accretion and accretion bursts that occur in low-mass star formation. The discoveries of accretion bursts in three massive young stellar objects (MYSOs), NGC6334I−MM1 (Hunter et al. 2017), S255IR−NIRS3 (Caratti o Garatti et al. 2017), and G358.93 −0.03 (Chen et al. 2020a), provide vital evidence for episodic accretion in massive star formation (Cesaroni et al. 2018;Brogan et al. 2018). The accretion burst source studied in this paper (G358.93−0.03) has a central protostar mass of ∼12 M e in its MM1 region (Chen et al. 2020a), a bolometric luminosity in a range of 5700−22,000 L e , and an accretion rate of 3.2 × 10 −3 M e yr −1 (Stecklum et al. 2021). Observational studies of the Orion molecular clouds show that episodic accretion accounts for >25% of a star's mass (Fischer et al. 2019). And even theoretical considerations show that massive stars can gain 40%-60% of their mass during accretion bursts (Meyer et al. 2021), suggesting that they are an essential rather than serendipitous process for massive star formation (Cesaroni et al. 2018;Brogan et al. 2018;Chibueze et al. 2021). If we are to gain a clear understanding of whether episodic accretion is a common phenomenon in the formation of all young stars, the study of episodic accretion bursts in massive star-forming regions (MSFRs) is crucial. So far, due to the lack of sufficient observational evidence, this process of high-mass star formation is still poorly understood.
It is relatively difficult to observe bursts of accretion in massive protostars due to their rapid evolution (much shorter timescales than low-mass stars), and the fact that accretion bursts tend to be relatively short compared to the more common quiescent periods (Stamatellos et al. 2011). Moreover, these MYSOs are usually buried in very dense clouds of dust and gas. Therefore, accretion bursts in MSFRs are very difficult to capture from a temporal and environmental point of view. Fortunately, masers are powerful tracers of several astronomical events, as they are commonly believed to be extremely sensitive to changes in the physical conditions of their natal clouds, especially those caused by the enhancement of radiation fields and collisions of matter. The increased local radiation field due to the stellar luminosity burst induced by an accretion burst will result in the increase of incident photons, thus enhancing the maser emission. Class II CH 3 OH (methanol) masers are pumped by infrared radiation and are thought to be closely associated with massive protostellar luminosity outbursts. Additionally, class II methanol masers are well established as tracers of the early stage of massive star formation (Minier et al. 2001;Ellingsen 2006) and exclusively observed near MYSOs (Minier et al. 2002;Xu et al. 2008;Paulson & Pandian 2020). It is worth mentioning that a direct link between 6.7 GHz class II CH 3 OH maser flaring and an accretion burst has recently been established for the three known MYSOs (NGC6334I-MM1, S255IR-NIRS3, and G358.93-0.03) with accretion burst events (Moscadelli et al. 2017;Hunter et al. 2018;Sugiyama et al. 2019).
The target source of this paper, G358.93-0.03 (hereafter G358), was identified as undergoing an accretion burst from variability monitoring of the 6.7 GHz methanol maser by the Maser Monitoring Organization (M2O, which is a global cooperative of maser monitoring programmers). 9 The 6.7 GHz maser burst started in 2019 January (10 Jy; Sugiyama et al. 2019), reached its peak emission (1156 Jy) in a short period of ∼2 months (MacLeod et al. 2019), and subsequently decayed rapidly returning to a normal accretion state. The burst thus lasted only about 5 months, such a rapid timescale that no current theory can adequately explain. Therefore, further methanol maser monitoring is needed to inform a theoretical explanation of episodic accretion process in MSFRs. Monitoring maser variability can also yield valuable information on changing conditions around the maser regions and the potential kinematics of the maser clouds.
In addition to the 6.7 GHz methanol maser, multiple new maser transitions have been detected in G358, such as the new class II methanol maser transitions, some of which are in torsionally excited states (v t = 1 and 2), at both centimeter MacLeod et al. 2019;Volvach et al. 2020) and millimeter wavelengths , and new molecular maser species 13 CH 3 OH, HDO, and HNCO (Chen et al. 2020a(Chen et al. , 2020b. The latter three new species of masers accurately depict spiral-arm accretion flow structures tracing fragmentations caused by the instability of large-mass disks (Chen et al. 2020a). The discoveries of these new maser transitions suggest that the episodic accretion process of G358 has a special physical environment to effectively excite new and rare masers from methanol and other molecule species. Nearly all these new maser transitions have dramatic changes within a short period. The rapid variability of the maser emission suggests that it is a transient phenomenon, probably associated with rapid changes in the thermal radiation field due to an accretion outburst (Chen et al. 2020a;Burns et al. 2020). Moreover, the accretion burst in G358 was decisively confirmed by multiepoch SOFIA observations (Stecklum et al. 2021). The event is found to be the first near infrared (NIR)/(sub)millimeterdark and far infrared (FIR)-loud MYSO accretion burst, showing an increase in the flux of the source only in the FIR, and not in the NIR or (sub)millimeter (Stecklum et al. 2021). The dense monitoring of methanol masers at multiple transitions will help us to further investigate more details of the accretion burst phenomenon in this source.
In this paper, we reported the monitoring results for the multiple methanol maser lines accompanied with accretion burst phase using the Shanghai 65 m Tianma Radio Telescope (TMRT) toward G358 during the period of 2019 March 14 to May 20. We detected eight new methanol maser transitions that have not previously been known to show maser emission.

Observations
The TMRT was used to conduct monitoring observations of a series of molecular lines, including masers, toward the flaring 6.7 GHz maser, G358 (J2000 position: 17 h 43 m 10 1014, 29°51′ 45 693; ). These observations began on 2019 March 14, and concluded on 2019 May 20, with a number of epochs in order to sample the different phases of the bursting source. We used the cryogenically cooled C-, Ku-, K-, Ka-and Q-band receivers covering a frequency range of 4 −50 GHz and the Digital Backend System (DIBAS) to receive and record signals. DIBAS is an FPGA-based spectrometer designed on the basis of the Versatile GBT Astronomical Spectrometer (Bussa 2012). Observations were first made in the wideband mode with bandwidths of 187.5 MHz in C band, 500 MHz in Ku band, and 1500 MHz in the K, Ka and Q bands, and detected emission was monitored using zoom-band mode with a high spectral resolution. In the zoom-band mode, each narrowband window has a bandwidth of 23.4 MHz. Using the active surface correction system, the achieved aperture efficiency of the TMRT ranges from 53% to 65%, corresponding to a sensitivity ranging from 1.28 to 1.59 Jy K −1 . The uncertainty in the absolute flux density for both wideband and zoom-band spectra is less than 10% by checking the flux density of nearby continuum calibrators. More details of our TMRT observations are listed in Table 1.
All observations were performed in position-switching mode, as a series of 1 or 1.5 minutes ON/OFF cycles. For each epoch, the total on-source time ranges from 12 to 56 minutes, depending on the signal-to-noise ratio of each detected line.
The GILDAS/CLASS package was used to perform the spectral line data reduction. The linear baseline of the spectrum was first fitted and then removed from the averaged spectrum of all scans. The rms noise levels achieved for each line are listed in Tables 2 and 3.

Eight New Methanol Transitions
In total, eight new class II methanol transitions at rest frequencies of 26.12, 27.28, 28.97, 31.98, 34.24, 41.11, 46.56, and 48.71 GHz were detected toward G358 in our observations. In addition to the 28.97 GHz transitions, all of the others are also discovered for the first time in interstellar space. Spectra of the eight new CH 3 OH transitions are shown in Figure 1 and their line properties are summarized in Table 2. The parameters and profiles of the Gaussian fits for the new transitions detected with the zoom-band mode are given in Appendices A and B, respectively. The new transitions have E u /k ranging from 121.3 to 950.7 K and the majority are from the torsional ground state v t = 0, with two from the first torsionally excited state v t = 1. As seen in Figure 1, the flux density of these methanol transitions changed significantly within ∼1 month, but the velocity extent was always contained within the range of −18.9 to −14.3 km s −1 .
26.12 GHz (10 1 -11 2 A − v t = 1): This transition was monitored over five epochs from March 17 to May 7. According to the Gaussian fit, the three velocity components of this emission are detected near −17.5, −16.4, and −15.7 km s −1 . All three velocity components have shown significant variability in flux density during the monitoring (see Figure 4). The peak-flux density reached 1218 Jy on April 12 at −17.5 km s −1 , and on May 3 the velocity component at −15.8 km s −1 reached 1265 Jy,   (5): inherent parameters in different backend modes including channel number, bandwidth, and velocity resolution. Columns (6)-(9): system temperature, sensitivity, aperture efficiency, and beam size at the corresponding receiver band. Column (10): total integration time.  Note. Columns (1)-(2): methanol maser transitions and adopted rest frequency with uncertainties given in parentheses from CDMS https://cdms.astro.uni-koeln.de/ and JPL (denoted with * ) https://spec.jpl.nasa.gov/. Column (3): observation epoch (YY/MM/DD). Columns (4): velocity range of the emission, which is determined by those velocity channels with emission above 3 σ rms . Columns (5)- (6): peak velocity and peak-flux density, which are directly measured for the brightest component of the emission at the given transition. Column (7): integrated flux density, which is the area enclosed by spectral profile within the velocity range in Column (4). Column (8): observational rms noise. Column (9): velocity resolution of each spectral line.     (5): velocity range of the emission which is determined by those velocity channels with emission above 3 σ rms . Columns (6)- (7): peak velocity and peak-flux density, which are directly measured for the brightest component of the emission at the given transition. Column (8): integrated flux density, which is the area enclosed by spectral profile within the velocity range in Column (4). Column (9): observational rms noise. Cragg et al. (2005). The observations made on nearly same date around March 17-18 show that the spectral profile of the 26.12 GHz is similar, but with weaker emission compared to that of the 20.97 GHz presented in Volvach et al. (2020), in accordance with the model predictions of Cragg et al. (2005). 28.97 GHz (8 2 -9 1 A − ): This is the first detection of this transition in G358, observed twice on April 1 with the wideband mode and on May 3 with the zoom-band mode. The spectrum of this transition shows a complex profile with at least five velocity components detected with the zoom-band mode (see Appendices A and B). In order to determine if this source varied, we smoothed the zoom-mode data (0.015 km s −1 ) to match the spectral resolution of the wideband data (0.95 km s −1 ). At the same resolution, we found that the flux density of the ∼−16 km s −1 component slightly decreased from 14.24 Jy to 10.68 Jy on April 1 and May 3, respectively. Notably, this transition has been detected in other sources. It was emitted in the quasi-thermal from the methanol emission center, about 10″ south from the hot core region in Orion KL, and with a peakflux density of about 1.1 Jy (Wilson et al. 1993). But toward W3(OH), the 28.97 GHz emission is masering and the peakflux density reached 15 Jy (Wilson et al. 1993). Shuvo et al. On April 6 with the wideband mode, the peak-flux density reached 9.7 Jy at −15.9 km s −1 , then on May 2, the peak flux decreased to 6.3 Jy at −17.0 km s −1 with the zoom-band mode. This is the line from the same series as the 229.589 GHz (15 4 -16 3 E) line detected by . Comparisons of spectra of these lines is impossible because the 229.589 GHz line is weak and its spectrum is not shown. Anyhow, detection of the maser line from the same series can be considered as a support for the maser nature of the detected line.
46.56 GHz (20 7 -21 6 A + ): This transition was observed twice on April 6 with the wideband mode and on May 2 with the zoom-band mode. The spectral profile of the wideband data shows a clear double-peaked structure (see Figure 1) with a peak-flux density of 2.6 Jy at −15.5 km s −1 . Zoom-band mode observations less than a month later with an rms noise of 0.09 Jy revealed no emission.

Previously Known Maser Transitions in G358
In this section, we report a series of previously known maser transitions including 15 methanol masers and 1 H 2 O maser detected in these observations toward G358. Figure 2 shows the spectra of these maser transitions and the line properties of each maser transition are summarized in Table 3. The parameters and profiles of the Gaussian fits for these transitions detected with the zoom-band mode are also given in Appendices A and B, respectively. It can be clearly seen from Figure 2 that the intensities of all maser emissions varied significantly with their peak variations in the range of 30% to 100% during the monitoring. Their velocity ranges, however, (2) decreasing first, then increasing followed by decreasing; (3) gradually increasing. Most of the class II methanol maser emissions showed their brightest intensity at the beginning of the observations in March and then gradually decreased (i.e., case 1). But there are some others such as at 19.97, 20.35, 20.97, and 23.12 GHz where their flux densities decreased first, then increased, and then decreased during the monitoring period (i.e., case 2), peaking around April 12 (see Section 5.2.1). In addition, for class I methanol maser transitions at 36.17 and 44.07 GHz, and a H 2 O maser transition at 22.24 GHz, their flux densities gradually increase from April to May (i.e., case 3).
There were three first torsionally excited methanol maser transitions at 6.18, 20.97, and 44.96 GHz that were consistently detected from March to May. The 6.18 and 20.97 GHz emissions had higher peak intensities than a month ago Volvach et al. 2020). This also suggests that G358 has an specific pumping environment during the outburst. May, the components with a velocity larger than −16.5 km s −1 had completely disappeared.

Remarks on Individual Transitions
36.17 GHz and 44.07 GHz: These two class I methanol maser transitions have been only monitored at two epochs, with the wideband mode in April and the zoom-band mode in May. The 36.17 GHz emission has three components close together covering a velocity range of −21.0 to −16.2 km s −1 , which is wider than the class II transitions. On March 7, the peak-flux density of the 36.17 GHz transition is 0.5 Jy at −19.5 km s −1 as detected by  and it gets to 0.74 Jy at −18.2 km s −1 on May 2 as detected in our observations. Oddly enough, this transition was not detected on April 1 with the wideband mode, suggesting its flux was less than the threshold of 0.36 Jy (3σ rms ). When the zoom-band spectrum on May 2 was smoothed to the same velocity resolution as the wideband mode, the peak-flux density was ∼0.60 Jy, which is still larger than the above threshold taken on April 1. It is suggested that after excluding the effect using different observing spectral modes, the flux of this transition really has little increment from April to May. The 44.07 GHz transition also has three components sharing a very similar velocity range to the 36.17 GHz transition. The −21.1 km s −1 component was detected with a peak of 4.5 Jy on March 5 by , and 2.7 Jy on May 2 from our observations. Notably, the spectral and peak velocity, as well as the velocity range of these two class I methanol maser transitions, are different from that of the class II transitions detected toward this source. Both observational surveys and theoretical considerations show that the maser emission in 36.17 GHz and 44.07 GHz transitions comes from generally the same regions but their spectra do not coincide in detail, due to the different sensitivity to the pumping conditions (Sobolev et al. 2007;Voronkov et al. 2014;Leurini et al. 2016;Sobolev & Parfenov 2018). The locations of these two transitions ) are close to the position of the water maser cluster components II −3−4 which occurred in 2019 May−June (Bayandina & Burns 2022b). Velocity of the 44.07 GHz line emission is close to the velocity of the water masers detected in May, suggesting that these masers might also have close locations. Occurrence of new water maser clusters is likely associated with the accretion burst in G358 (Bayandina & Burns 2022b). Both water masers and class I methanol masers are pumped by collisions, so the variability of the 36.17 GHz and 44.07 GHz is also likely associated with the accretion burst phenomenon. In particular, it can explain the occurrence of the 36.17 GHz maser only on May 2. 37.70 GHz: This methanol transition has been monitored with two epochs with the wideband mode on April 1 and with the zoom-band mode on May 2. It has two velocity components at −16.2 and −17.6 km s −1 . Its peak-flux density was 250 Jy detected on March 7 at −17.3 km s −1 , 111 Jy on April 1 at −17.6 km s −1 , and 64 Jy on May 2 at −16.2 km s −1 detected in our observations. It seems that the emission of this transition gradually decreases from March to May.
44.96 GHz: This transition has been monitored with two epochs on April 6 and May 2 with the wideband mode. This transition appears to have a distinct double-peaked spectral structure with two velocity components peaked at −15.4 and −17.1 km s −1 during the period from April to May.  also detected this transition with a clear double-peaked structure on March 5, with the peak-flux density of 508 Jy at −15.4 km s −1 . The emission at this same velocity component gradually decreased to 61 Jy on April 6 and then to ∼15 Jy on May 2. 45.84 GHz: This transition has been monitored with two epochs with the wideband mode on April 6 and with the zoomband mode on May 2. This transition has two velocity components peaked at −16.2 and −17.3 km s −1 seen from the zoom-band mode spectrum. On March 5, the peak-flux density of this transition reached 414 Jy at −15.4 km s −1 ). On April 6, the peak-flux density reached 54 Jy at −17.2 km s −1 under a velocity resolution of 0.6 km s −1 from our observation. On May 2, at the same velocity component, the peak-flux density reached 86 Jy under a velocity resolution of 0.01 km s −1 . When the zoom-band spectrum on May 2 was smoothed to a velocity resolution of 0.6 km s −1 similar to the wideband spectrum, the peak-flux density is 46 Jy. Therefore, the peak flux of this emission gradually decreased after March 5.
22.24 GHz: This H 2 O maser transition has been monitored with seven epochs from March 17 to May 7. We can see that the H 2 O maser emission has two components peaked at −17.3 and −19.2 km s −1 with a velocity resolution of 0.15 km s −1 in April. The intensity of the H 2 O maser emission is very weak (with a peak of 0.5 Jy) in March, until April 12, when it has a significant increment with a peak of 4.3 Jy, and then reaches 8.6 Jy on April 15 at −17.3 km s −1 . On April 20, at the same velocity component, the peak-flux density reaches 24 Jy (MacLeod et al. 2019). Furthermore, on May 7, some new maser features were detected in the velocity range of −22 to −17 km s −1 (with peak at ∼-20 km s −1 ) and −16 to −13 km s −1 (with a peak at ∼-14.5 km s −1 ), which are close to that detected with the Very Large Array (VLA) in June by Bayandina & Burns (2022b) with the exception of a new maser component peaked at -21.5 km s −1 that appeared in June from the VLA detection. These new water components have the similar velocity ranges to those detected in both the 36.17 and 44.07 GHz methanol transitions, supporting the idea that they have the same pumping origination and are associated with accretion burst actions on the shocked regions where they are excited (see above). Overall, the water maser emission is barely visible in March until the flux density suddenly increases on April 12 and reaches a much higher intensity on April 20, with new velocity components detected in May-June, which is quite different from the class II methanol maser transitions. This means that the water maser emission burst occurred later than the methanol masers, likely because the water maser is distributed in regions far away from the burst source G358-MM1 compared to the methanol masers. This statement is supported by the VLA images of methanol and water masers (Chen et al. 2020a;Bayandina & Burns 2022b), thus there is an expected time delay between the heat-wave propagation to the methanol and to water maser regions.

The Maser Characteristics of the Eight New Methanol Lines
G358.93-0.03 harbors two molecular hot cores, MM1 and MM3, and MM1 shows significantly richer molecular spectra and the brightest millimeter continuum emission . So far, accretion bursts and all the discovered masers are detected toward the peak of MM1, implying this region is suitable for exciting the maser emission Chen et al. 2020b). The majority of these new CH 3 OH transitions show a velocity range from −18.9 to −14.3 km s −1 , which is similar to the 6.67 GHz spectrum in Figure 2. At the same time, some transitions (e.g., 26.12,27.28,34.24,46.56,and 48.71 GHz) show similar velocity components to the 6.7 GHz transition at one of −15.5, −16.0, and −17.3 km s −1 at least. It suggests that the region where the eight new lines possibly originate from the same as or close to the 6.67 GHz transition. Using a Gaussian fit, we obtain the line width of the strongest feature in each of the new transitions ranging from 0.27 to 0.55 km s −1 , except for the transitions at 31.98 and 46.56 GHz detected only with the wideband mode (see Table A1). For the two brightest spectral peaks at −17.4 and −15.6 km s −1 , the average line widths of the new transitions detected in the zoom band are 0.33 and 0.40 km s −1 , respectively. The line widths of the two main spectral features of the 6.67 GHz transition are 0.44 and 0.48 km s −1 , respectively. Therefore for the two main features, the average line widths of these new transitions are even narrower than the 6.67 GHz transition. In addition, like the other maser lines, the new transitions have complex spectral compositions and rapid variations.
The maser emission in all detected transitions resides in a region of 0 2 around the bursting source (Bayandina et al. 2022a). Assuming that the new methanol emission originates in this region, we infer a lower limit of brightness temperature in the range from 3 × 10 4 to 8 × 10 7 K for these new transitions through the equation: T B = S υ λ 2 /2kΩ, where S υ is flux density, λ is wavelength, k is the Boltzmann constant, and Ω is solid angle. They are much higher than the maximum kinetic temperatures (100-300 K) of the typical thermal emission associated with hot molecular cores in MYSOs. In addition, the typical thermal line width expected under such a high-temperature condition (3 × 10 4 K) is ∼6 km s −1 , through the relation: Δv FWHM = 0.2(T/A) 1/2 km s −1 , where T is the temperature in kelvin and A is the atomic mass number of the methanol molecule. It is much bigger than the average line width of our detections. These all suggest that the new emissions are unlikely to be thermal in nature. Beyond this, we performed the rotation diagram analysis to obtain additional evidence that the excitation of these transitions shows significant deviation from the local thermodynamic equilibrium. The rotation diagram of the newly detected CH 3 OH transitions is shown in Figure 3 and the detailed method is described in Blake et al. (1987) and Chen et al. (2013). In general, the eight new transitions do not distinctly show a linear correlation between ln(3 kW/(8π 3 υSμ 2 )) and E u /k expected from thermal conditions in the rotation diagram. In particular, when E u /k < 500 K, the five transitions (observed at the same epoch on 2019 May 3) even show an inverse trend in the rotation diagram, inconsistent with thermal conditions (van Dishoeck et al. 1995;Ginsburg et al. 2017;Chen et al. 2019). Notably, the three transitions in the lower-right corner observed in April do not show an opposite trend to the above five transitions, probably because of their variabilities. Combining all these properties discussed above, we conclude that all the detected eight new methanol transitions from G358 deviate from the thermal condition.
There are two methanol maser transitions at 26.12 and 48.71 GHz from the first torsionally excited state (v t = 1) detected in our observations. They were not included in the published lists of class II methanol maser candidates predicted from theoretical calculations (Cragg et al. 2005). The detection of such two torsionally excited transitions with bright intensities suggests that we were observing a special pumping regime for class II methanol masers and therefore the current observations offer substantial new information for methanol maser pumping models.

Maser Variability
Dynamic spectra of the 11 CH 3 OH maser transitions and 1 H 2 O maser transition with more than 5 epoch observations over 2 months from 2019 March to May associated with G358 are presented in Figure 4. It is clearly seen that these masers show temporal variations with different velocity components (see panels (a) and (b) of each transition in Figure 4). 5.2.1. The New Flare Event of the 19.97, 20.35, 20.97, 23.12, and 26.12 GHz Maser Transitions on April 12 From Figure 4, we can see that during the monitoring from 2019 March to May, the flux density shows a declining trend in general for most class II CH 3 OH transitions with exceptions at the 19.97, 20.35, 20.97, 23.12, and 26.12 GHz transitions. These five transitions show a sharp increment in their flux densities in the April 12 observations. After this flare, the emission continues to gradually decay again. From the variability shown in Figure 4, we can derive that this new flare can be considered to start around April 8. In just 4 days, the maser emissions have increased by a factor of 5, 3.6, 3.25, 3.26, and 6.22 for the above five transitions, respectively.
To further confirm whether this new short-period flare event that occurred in these five maser transitions is real or not, we compared data observed with other telescopes at dates close in  Therefore, these two close-date observations also show the similarity in the flux density within the uncertainty of the flux scale at the 23.12 GHz transition. By comparing the peak-flux densities for the above three transitions detected in close-date observations using different telescopes, their agreement supports the increment of the 19. 97, 20.35, 20.97, 23.12, and 26.12 GHz maser transitions on April 12 being reliable.
Apart from the new line at 26.12 GHz, we have collected the peak-flux densities for the remaining four lines during the flare preceding the one on April 12. For the 19.97 GHz transition at −15 km s −1 , the peak-flux density was 104 Jy detected on A direct question related to the two maser flares concerns whether these flares are associated with two different accretion burst events, or are they the result of the influence of the same accretion burst in different maser locations? As suggested by Volvach et al. (2020), due to the changes in line width of the 20.97 GHz transition experienced at different flares, it is possible that the same accretion burst leads to the maser being excited in different regions along the path of the heat-wave passage. If the two flares of masers are the result of two different accretion bursts acting on the same maser locations, then it suggests that the maser spatial locations would not vary with different flares. However, the current interferometric observations have revealed that the spatial distributions of multiple methanol maser transitions are actually changed at different burst phases (Burns et al. 2020;Bayandina et al. 2022a). Therefore, we argue that the new flare detected in April in our monitoring is mainly contributed to by the same burst acting on different maser clouds.  Figure 4(d)). The standard theory of masers predicts line narrowing during unsaturated amplification and rebroadening to the full Doppler width during saturated amplification (Goldreich & Kwan 1974;Hirota et al. 2014). It means that these maser spots may be saturated. Under the saturation case, the standard theory of masers predicts that the profile in the Doppler velocity becomes the same as the Gaussian velocity distribution of the masing molecules (Litvak 1970;Nedoluha & Watson 1991;Watson et al. 2002). Therefore, for these saturated maser spots, we can derive the gas temperature from the Doppler velocity extent through the relation Δv FWHM = 0.2(T/A) 1/2 km s −1 (see Section 5.1). The relationship of gas temperature with observing time is presented in Figure 4(c) by a dashed line. It can be seen that in general the gas temperature of most maser spots gradually decreased during the monitoring from 2019 March to May. This might suggest that the maser components were going through a cooling process during the monitoring, possibly associated with the propagation of a heat wave induced by the accretion burst event in G358. It is also noted that the 12.18 GHz maser emission is generally stronger than the 6.67 GHz on the same observing dates (see Figure 4). Usually, the 6.67 GHz maser emission is stronger than the 12.18 GHz in almost all of the sources with both maser transition detections (Breen et al. 2012;Song et al. 2022). But our observations present a reversal detection in G358, suggesting that the physical conditions at the time of the accretion burst event in G358 might significantly change with respect to those expected in MSFRs without the accretion bursts.

Summary
Monitoring for multiple maser (including methanol and water) transitions was made with the TMRT during the period from 2019 March to May toward an accretion burst MYSO G358.93-0.03. This period mainly corresponds to the decay phase of this accretion burst. The main results and scientific insights obtained from this monitoring are summarized as follows: 1. The special conditions associated with the accretion burst are currently known to be able to excite new methanol maser transitions. Eight new methanol maser lines at 26.12, 27.28, 28.97, 31.98, 34.24, 41.11, 46.56, and 48.71 GHz were first detected during the accretion burst of MYSO G358.93-0.03. This large number of new methanol maser transitions are not reported for the other two accretion bursts in S255IR and NGC 6334I. So, the G358 burst has a kind of special condition. Indeed, it has a much shorter duration, suggesting that the accreted mass is smaller than in the other detected events. Sophisticated estimates in Stecklum et al. (2021) show that the accretion burst event was produced by accreting a clump with the planetary mass. In this case, the star does not experience substantial bloating and its emission during the burst can be considerably harder. This will result in the detection of unknown highly excited maser transitions. 2. Nearly all the maser lines have obvious and dramatic changes within a short period (∼order of a month). Some of the maser transitions showed repeated flares in 2019 April. This may be related to the passage of the heat wave induced by the same accretion burst event acting on different maser positions. 3. During the monitoring, the gas temperatures of clouds under the hypothesis of saturated masers generally showed a significant decay. This indirectly proves that the maser flares are not due to a kinetic process in such a short period, but to the propagation of thermal radiation from the MYSO's luminosity outburst, with a cooling process detected in our monitoring.