The Possible Responses of Polar Ozone to Solar Proton Events in March 2012 by FengYun‐3 Satellite Observations

In this work, we use observations by the Solar Backscatter Ultraviolet Sounder and Space Environment Monitor on FengYun‐3 to analyze the polar ozone depletion during the solar proton events (SPEs), which occurred in early March 2012. The ozone distributions changed evidently with the increasing energetic proton flux (the particle energy is over 100 MeV) at the approximate altitude of 30 km. From the ozone profile relative changes, the short‐term impacts of SPEs can be distinguished from the long‐term effects of ozone season variations after the SPEs take place and cause about 4–17% of the short‐term polar ozone decreases at the different levels in the upper stratosphere of both hemispheres. In the upper stratosphere, the SPE‐related polar ozone depletion is more significant and continuous in the Northern Hemisphere but shows the short‐term effects in the Southern Hemisphere during the March SPEs. The ozone depletion responses to the first SPE on 7 March are more pronounced in this altitude region than the second one on 13 March in both hemispheres due to the “harder” particle energy spectrum.


Introduction
It is known that severe solar activities, such as solar flares or coronal mass ejections (CMEs), will erupt an intense flux of charged energetic particles (protons, ions, and so on) (Reames, 1999) and cause solar proton events (SPEs) with particle energies over 10 MeV. When these charged particles are ejected toward the Earth, they are guided by the geomagnetic field and precipitate into the polar atmosphere (Patterson et al., 2001). Energetic particles generate odd hydrogen and odd nitrogen, such as HO x and NO x , which participate in the catalytic chemical process of polar ozone loss in the middle atmosphere. As shown in Figure 1, the particles penetrating into the atmosphere generate ions, such as N + 2 , O + 2 , N + , O + , and NO + , which react with O 2 and water clusters to form NO x and HO x that are well known to cause ozone loss by the NO x and HO x cycles [Swider & Keneshea, 1973;Crutzen et al., 1975;Frederick, 1976;Rusch et al., 1981;Solomon et al., 1981;Solomon et al., 1983;Jackman & McPeters, 1985;Jackman & McPeters, 2004]. Many investigations about the polar ozone depletion during SPEs show that severe SPEs may cause significant polar ozone changes in the middle atmosphere (Weeks et al., 1972;Heath et al., 1977;McPeters et al., 1981;Reid et al., 1991;Zadorozhny et al., 1992;Jackman et al., 2001;Seppälä et al., 2004;Jackman et al., 2005a;Jackman et al., 2005b;Degenstein et al., 2005;López-Puertas et al., 2005;Rohen et al., 2005;von Clarmann et al., 2013;Jackman et al., 2014).
The SPEs that occurred in early March 2012 are the twelfth largest SPEs in the satellite era since 1963 based on the study in (Jackman et al., 2008;von Clarmann et al., 2013) and caused significant impact on the polar middle atmosphere (von Clarmann et al., 2013;Jackman et al., 2014). In this study, for the first time, we apply the measurements of Solar Backscatter Ultraviolet Sounder (SBUS) and Space Environment Monitor (SEM) on board the FengYun-3 meteorological satellites to investigate how the March SPEs of 2012 affected the polar ozone density in the upper stratosphere of both hemispheres, and we provide the simultaneous observations of proton fluxes and stratospheric ozone depletion during the SPEs at different latitudes. This study focuses on the middle to upper stratosphere, while the previous publications were more focused on the mesosphere.

SBUS and SEM
The SEM and the SBUS are two of the payloads on board the FengYun-3 satellites, which are Chinese Sun-synchronous polar orbit operational meteorological satellites at the altitude of 830 km. The FengYun-3 series of satellites are developed in three groups: FY-3A and FY-3B in the first group were launched on 7 May 2008 and 5 November 2010, respectively; the second group consists of FY-3C and FY-3D, which were launched on 23 September 2013 and 15 November 2017; the third group containing FY-3E (dawn-dusk orbit), FY3F, FY-3G, and FY-Rainfall (instruments focus on the origins of rainfall) will be launched in the future (after 2020). FY-3A has been stopped after March 2018, but FY-3B/C/D are still running. SEM has the capability to detect proton flux with an energy range from 3 to 300 MeV (six bands: 3-5, 5-10, 10-26, 26-40, 40-100, and 100-300 MeV) and electron flux with energy range from 0.15 to 5.7 MeV (five bands: 0.15-0.35, 0.35-0.65, 0.65-1.2, 1.2-2.00, and 2.00-5.70 MeV). From the 1960s, many missions launched to detect the radiation belts [Northrop & Teller, 1960;Mcllwain, 1961;Hess, 1968;Heckman & Nakano, 1969] and major satellite series, which are POES (Polar Orbiting Environmental Satellites) and FengYun-3, provide operational particle products presently. POES have been running for more than 40 years from 1979, and now, NOAA-15/18/19/20 satellites are still in operation. SBUS applies Solar Ultraviolet Backscattered technique, which has been developed for 40 years from the 1970s (Bhartia et al., 1996;Flynn et al., 2009;World Meteorological Organization [WMO], 2010) to derive ozone profile from the troposphere (several kilometers high) to the mesosphere (about 70 km). According to the in-orbit cross calibration with the ozone vertical profiles from National Oceanic and Atmospheric Administration Solar Backscatter Ultraviolet (NOAA SBUV), the uncertainty of the SBUS measurements is approximately 6-7% (F. X. . SBUS has 21 layer detections from 1,013 hPa to the top of atmosphere like SBUV on board POES, and each layer's value is the average value over this layer. It should be noted that the detections above 70 km are unreliable. Due to the reflection impacts of the aerosols in the troposphere and the rare ozone density in the mesosphere, we consider that the ozone profiles retrieved by SBUS are more reliable in the stratosphere (about 10-55 km) (F. X. . So in this work, we used the ozone profiles measured by SBUS from 30 to 50 km (Layers 11 to 16, from 10.13 to 1.013 hPa) to investigate the March SPEs' impacts on the ozone in the upper stratosphere. The relationships between the layers (11 to 16) and the altitudes in polar region are shown in Table 1. Previous studies about ozone profile measured by SBUS can be found in (Huang et al., 2010;Liu et al., 2011). The SBUS data are provided in Dobson units (DU). The number density or volume mixing ratio might be more suitable for this work; however, the SBUS on board FY-3 satellite is designed to monitor the ozone in the low to middle stratosphere. More details of SEM and SBUS instruments, data retrievals, calibrations, and validations can be found in (C.  and (F. X. . In this work, we applied the particle detections by SEM on board two satellites (FY-3A and FY-3B) and the ozone measurements by SBUS on board one satellite (FY-3B) because the SBUS on board FY-3A suffered a mechanical failure during the SPE days. It should be noted that SEM works both day and night but SBUS only works at dayside. So, to one satellite, the available particle detections are about twice as much as the ozone detections.
We also applied the proton measurements from the SEM on board FengYun-2D (FY-2D, launched on 8 December 2006 and located at 86.5 • E) geosynchronous meteorological satellite, which has the ability to detect the proton flux in three bands (P 1 : 10-30 MeV; P 2 : 30-100 MeV; and P 3 : 100-300 MeV) and the electron flux in two bands (≥ 350 keV and ≥ 2 MeV), to derive the situation of the proton flux during the March SPEs ( Figure 2, P 1 , P 2 , and P 3 are combined to derive the proton flux in the bands of 10-300 and 30-300 MeV).

The March SPEs of 2012
SPEs caused by intense solar flares or CMEs are always accompanied by high-energy protons that appear at high latitudes. The SPEs that occurred on 7 (Day 67) and 13 (Day 73) March were caused by the solar active region of AR1429. In Figure 2, the proton flux measured by FengYun-2D Chinese geosynchronous meteorological satellite shows that the proton flux with particle energy over 100 MeV reached over 100 pfus (proton flux units) during the SPE on 7 March when the proton (10-100 MeV) fluxes increased 3-4 orders of magnitude. From Figure 2, it is found that the first SPE on 7 March (Day 67) has much higher flux (about an order of magnitude) and "harder" particle energy spectrum with the flux of energetic proton (particle energy is over 100 MeV) than the second SPE on 13 March (Day 73). More "harder" particles may bring more significant impacts on lower atmosphere at about 20 hPa (von Clarmann et al., 2013) because higher energy particles have the ability to penetrate deeper into the atmosphere and produce the ozone destruction materials at lower altitudes. The proton flux measured by FY-2D is slightly different from GOES-13 measurements in Figure 2 of Jackman et al. [2014] due to the different locations of the two geostationary satellites. Figure 3 shows the ozone distribution at 10 hPa (Layer 11, about 30 km high) measured by SBUS and the proton flux with particle energy over 100 MeV detected by SEM in the March SPEs. The ozone distribution plot is made by the bilinear interpolation method. SBUS data were gridded onto a 5 • grid map, and we used bilinear interpolation to derive every latitudinal gridding's value. Then, we generated the ozone distribution   map. The particles with energy over 100 MeV have the ability to reach the level of 10 hPa and induce ozone variations there (Hargreaves, 1992). In early March 2012, the north polar ozone distribution at 10 hPa shows very significant depletion responses and distribution changes to the SPEs on 7/8/9 March (the area of very low ozone actually agrees well with the area of highest proton fluxes) and no obvious feedback to the SPE on 13 March. Figure 4 shows the proton flux in 40-100 MeV (the protons in this energy band have the ability to reach approximately 40 km) and the ozone distribution at 2.51 hPa level (about 41 km). The protons in this energy band have the ability to reach approximately 40 km. The ozone distribution at 2.51 hPa changed significantly from 7 to 10 March with the high proton flux that the ozone depleted in the regions of high particle flux, but there still exists the ozone increasing somewhere. It should be noted that the significant ozone depletion at 2.51 hPa occurred on 10 March, which is different from that at 10 hPa. The direct impact of the particles lasts longer (about 5 days) at 2.51 hPa than that at 10 hPa (about 3 days), which might be due to the longer duration of the energetic particle flux. The top panel of Figure 5 shows In general, the north polar ozone depletion responses to the first SPE on 7 March (Day 67) are more significant than the second SPE on 13 March (Day 73) because the first March SPE has more "harder" particle energy spectrum, which may affect the lower level of stratosphere. Figure 6 presents the ozone distribution at 10 hPa and the proton flux over 100 MeV from FY-3 observations in the Southern Hemisphere. From Figure 6, the significant ozone destruction in the region of high proton flux at 10 hPa level are observed after the SPE of 7 March, but the SPE of 13 March does not cause any ozone depletion at 10 hPa level. Similarly to Figure 4, Figure 7 shows the ozone distribution responses to the  Figure 2 in (von Clarmann et al., 2013). According to this figure, ozone levels are increasing from January to March in the upper stratosphere in the Southern Hemisphere, which may be a seasonal effect. In addition, Figure  2c in (von Clarmann et al., 2013) and Figure 6 in (Jackman et al., 2014) show that there is a peak of ozone just before the time of 7 March SPE at 42-43 km in the Southern Hemisphere (Figures 2c and 6, top right). So we tend to consider that this positive ozone anomaly might be part of the natural ozone variation. Now considering the negative anomaly, we compare it with Figure 2 in (von Clarmann et al., 2013). This shows ozone levels decreasing at low latitudes from February to March in the Southern Hemisphere. So we think the negative anomaly around 30 km may also be a seasonal effect. Similar to the Northern Hemisphere, the SPE-related ozone depletion on 7 March (Day 67) is significant, lasting more than a week but much shorter than the depletion in the Northern Hemisphere.

Conclusion and Discussion
The SPEs which occurred in early March 2012 brought significant impacts on both hemispheres' polar middle atmosphere. From FY-3 satellite observations, there exist short-term but obvious ozone changes in the upper stratosphere, and the direct impacts of the energetic particles last longer at higher altitude than that at lower altitude in both hemispheres. Due to the "harder" particle energy spectrum, the responses of SPE-related ozone depletion at 10 hPa are more pronounced on 7 March (Day 67) than that on 13 March (Day 73). Seppälä et al. (2008) reported the impacts of hard-spectra SPEs, the 2005 January SPEs, on the middle atmosphere, and the model results showed that the 2005 January SPEs caused insignificant impacts on the ozone in the stratosphere even if the SPEs lasted 24 hr. In our work, the hard-spectra SPE that occurred on 7 March brought significant impacts on the ozone in the upper and middle stratosphere. We think the cause of the differences between the March 2012 SPEs and the January 2005 SPEs may the long duration of the high flux of energetic protons (about 48 hr), the more "harder" particle spectrum in the March 2012 SPEs. In our work, the short-term ozone depletion after the March SPEs is more evident in the northern polar upper stratosphere, and there exists continuous ozone reduction in the north polar upper stratosphere, which may be due to the downward motion of odd nitrogen caused by the Arctic vortex from the upper level . The NO x downward transport inside the polar vortex is common (Funke et al., 2011;Funke, López-Puertas, Holt, et al., 2014;Fytterer et al., 2015). In the work of von Clarmann et al. (2013), they reported that the March SPEs made the continuous ozone depletion in March (see Figure 1l in their paper) around the altitude of 40 km in the Northern Hemisphere, but only a short-term negative response of ozone was found after the March SPEs in the Southern Hemisphere (see Figure 2c Figure 6 in their paper). According to the results in Jackman et al. (2014), the peaks of the ionization rate and the ozone depletion during the March SPEs occurred on 8 March. In our work, we also find that the significant peak of ozone destruction is on 8 March in the Southern Hemisphere (see it in Figure 6), which may be because the ionization rate reaches a maximum on 8 March and coincides with the MLS observations. The process of energetic particle precipitation effects on ozone deserves to be studied so that we can better understand how space weather events influence middle atmosphere and also the mechanism between solar activities and climate changes of the Earth.