key: cord-0335829-gggu5idl
authors: Fesen, Robert A.; Drechsler, Marcel; Weil, Kathryn E.; Strottner, Xavier; Raymond, John C.; Rupert, Justin; Milisavljevic, Dan; Subrayan, Bhagya M.; Cicco, Dennis di; Walker, Sean; Mittelman, David; Ludgate, Mathew
title: Far UV and Optical Emissions from Three Very Large Supernova Remnants Located at Unusually High Galactic Latitudes
date: 2021-02-24
journal: nan
DOI: 10.3847/1538-4357/ac0ada
sha: 79055efae21a84b432ca248bc9876f6bda4fe0bd
doc_id: 335829
cord_uid: gggu5idl

Galactic supernova remnants (SNRs) with angular dimensions greater than a few degrees are relatively rare, as are remnants located more than ten degrees off the Galactic plane. Here we report a UV and optical investigation of two previously suspected SNRs more than ten degrees in both angular diameter and Galactic latitude. One is a proposed remnant discovered in 2008 through 1420 MHz polarization maps near Galactic coordinates $l$ = 353, $b$ = $-$34. GALEX far UV (FUV) and H$alpha$ emission mosaics show the object's radio emission coincident with a 11 x 14 degree shell of UV filaments surrounding a diffuse H$alpha$ emission ring. Another proposed high latitude SNR is the 20 x 26 degree Antlia nebula (G275.5+18.4) discovered in 2002 through low-resolution all-sky H$alpha$ and ROSAT soft X-ray emissions. GALEX UV and H$alpha$ mosaics along with optical spectra indicate the presence of shocks throughout the Antlia nebula with estimated shock velocities of 70 to over 100 km s$^{-1}$. We also present evidence that it has collided with the northeast rim of the Gum Nebula. We find both of these nebulae are bona fide SNRs with ages less than 10$^{5}$ yr despite their unusually large angular dimensions. We also present FUV and optical images along with optical spectra of a new high-latitude SNR (G249.7+24.7) some 4.5 degrees in diameter which has also been independently discovered in X-rays and radio (Becker at al. 2021). We find this remnant's distance to be $leq$400 pc based on the detection of red and blue shifted Na I absorption features in the spectra of two background stars.

In spite of an increasing variety of supernovae (SNe) sub-types, the majority of SNe are generally believed to release 0.5 − 2 × 10 51 erg, although superluminous SNe may release substantially more (Howell 2017) . Only a small percentage of this comes out as visible light, with most of a SN's energy initially carried away in the form of kinetic energy. It is this enormous point deposition of energy that has a significant impact on the structure and energy content of a galaxy's interstellar medium through an expanding remnant that can last largely intact up to 10 5 yr (Cox & Smith 1974; McKee & Ostriker 1977; Shull et al. 1989; Blondin et al. 1998) .

Currently, there are some 300 confirmed Galactic supernova remnants (SNRs) catalogued with dozens of other suspected SNRs and more added every few years (Ferrand & Safi-Harb 2012; Green 2019) . Most Galactic SNRs are less than a degree in angular size, more than 1 kpc distant and well evolved, with typical estimated ages between 10 4 and 10 5 yr.

Out-sized attention relative to their population percentage has been directed to the handful of the Milky Way's young remnants with ages less than 5000 yr. These include the Crab Nebula, Cassiopeia A, the remnants of Tycho and Kepler, the Vela remnant, and Puppis A. This increased attention is due to their high expansion velocities, metal-rich ejecta, bright emission across the entire electromagnetic spectrum, and clearer connections to the various core-collapse and thermonuclear SN sub-types.

The majority of Galactic SNRs are first detected through radio observations due to their characteristic synchrotron nonthermal radio emission associated with shocked gas (Downes 1971; Chevalier 1977; Green 1984 Green , 2004 . Such nonthermal emission leads to a power law flux density, S, with S ∝ ν −α where α is the emission spectral index with typical SNR values between -0.3 and -0.7. Rarer pulsar wind-driven SNRs exhibit a much flatter radio spectrum, with spectral indices around zero. While multi-frequency radio surveys have been historically the dominant tool for finding SNRs, X-ray studies have also led to the discovery of several additional SNRs. These include RX J1713.7-3946 (G347.3-0.5; Pfeffermann & Aschenbach 1996) and the young "Vela Junior" SNR (G266.2-1.2; Aschenbach 1998) coincident with the much larger Vela SNR. Recent examples of X-ray confirmed or discovered remnants include several new SNRs in the LMC (Maggi et al. 2014 ) and X-ray confirmation of the suspected Galactic remnant G53.4+0.0 (Driessen et al. 2018) .

Despite the fact that less than 50% of Galactic SNRs exhibit any appreciable associated optical emission, a remnant's optical emission can be useful in confirming the presence of high-velocity shocks and in defining a remnant's overall size and morphology. Although discoveries of new SNRs in the optical is relatively rare, there have been several recently reported discoveries (Stupar et al. 2007; Boumis et al. 2009; Stupar & Parker 2012; Fesen et al. 2015; Neustadt et al. 2017; Stupar et al. 2018; How et al. 2018; Fesen et al. 2019 ) along with many proposed SNR candidates (Stupar et al. 2008; Stupar & Parker 2011; Boumis et al. 2009; Alikakos et al. 2012; Sabin et al. 2013) .

The main criterion for optical SNR identification is a line ratio of I([S II])/I(Hα) ≥ 0.4 which has proven useful in identifying the shocked emission of SNRs in both the Milky Way and nearby local group galaxies (Blair et al. 1981; Dopita et al. 1984; Fesen et al. 1985; Leonidaki et al. 2013; Long 2017) . Recently, the near infrared [Fe II] line emissions at 1.27 and 1.64 µm have also been used to detect dust obscured SNRs in nearby galaxies (see review by Long 2017) .

One wavelength regime that has not yet been fully exploited to search for new Galactic SNRs is the ultraviolet (UV). Here we present results of an initial study of SNRs located far away from the Galactic plane using wide field-of-view (FOV) UV images assembled from the Galaxy Evolution Explorer (GALEX) All-Sky survey (Bianchi 2009 ). We began this research by investigating two unusually high latitude suspected supernova remnants including the exceptionally large Antlia remnant (McCullough et al. 2002 ) (see Table 1 ). During this work, we also found a new apparent SNR. However, after our initial submission, W. Becker kindly sent us a preprint of their independent discovery of this new remnant based on SRG/eRosita X-ray and Parks All-Sky Survey observations. Since their paper has now been published (Becker et al. 2021) , we have cited their results in reference to our independent findings. UV and optical images plus some follow-up optical spectra are described in Section 2, with results presented in Section 3. We discuss the general properties of these SNRs in Section 4, with our conclusions and discussion of helpful follow-up observations given in Section 5.

The GALEX satellite was a NASA science mission led by the California Institute of Technology who operated it from July 2003 through February 2012. The main instrument was a 50 cm diameter modified Ritchey-Chrétien telescope, a dichroic beam splitter and astigmatism corrector, and two microchannel plate detectors to simultaneously cover two wavelength bands with a 1.25 degree field of view with a resolution of 1.5 pixel −1 . Direct images were obtained in two broad bandpasses: a far-UV (FUV) channel sensitive to light in the 1344 to 1786Å range, and a near-UV (NUV) channel covering 1771 to 2831Å (Morrissey et al. 2007) . Resulting images are circular in shape with an image FWHM resolution of ∼ 4.2 and ∼ 5.3 in the FUV and NUV bands, respectively.

Being mainly a mission to study the UV properties of galaxies in the local universe, GALEX survey images largely avoided the complex and external galaxy poor Galactic plane. However, even its All-Sky Survey program focused away from the plane of the Milky Way did not cover all possible areas, leading to numerous gaps (see Fig. 1 ). Nonetheless, some 26,000 square degrees were imaged to a depth of m AB = 20.5 (Bianchi 2009 ). Using the on-line GALEX images, large FUV and NUV image mosaics were examined of several regions typically more than 10 degrees off the Galactic plane.

Several different sources of narrow passband optical images were used in our investigation of the three SNR candidates. Table 2 lists details of these optical images. SALT Observatory 10 m · · · · · · · · · λ4050 -λ7120 5.0Å

G354-33 SHASSA Survey f/1.6 52 mm lens Hα, λ6440, λ6770

Wide-field, low-resolution Hα images of regions around the suspected SNRs G249+24 and Antlia obtained as part of the MDW Hydrogen-Alpha Survey 1 were used to provide general optical size and emission features. This survey uses a 130 mm f/2 camera at the New Mexico Skies Observatory, with a FLI Pro-Line 16803 CCD and 3 nm filter centered on Hα. This telescope-camera system produced field-of-view (FOV) of approximately 3.5 • × 3.5 • with a pixel size of 3. 17. Each field was observed 12 times each with an exposure of 1200 s.

We have also made large mosaics from Southern Hα Sky Survey Atlas (SHASSA; Gaustad et al. 2001 ). This wide-angle robotic survey covered δ = +15 • to −90 • with 13 • square images with an angular resolution of 0.8 . Narrow passband continuum filters on blue and red sides of Hα centered at 6440 and 6770Å allow for stellar and background subtraction. The survey had a native sensitive level of 1.2×10 −17 erg cm −2 s −1 arcsec −2 which could be significantly lowered with smoothing.

The western region of the Antlia remnant where it abuts the northeastern limb of the Gum Nebula was imaged using Hα + [N II], [O III] and [S II] filters using a 200 mm f/2 lens and a 9576 × 6388 pixel ZWO ASI6200mm Pro CCD. This system provided a FOV of 10.3 • × 6.9 • with a 3.87 pixel −1 image scale. Total exposure times for the three filters listed above were 6.9 hr, 13.3 hr, and 6.8 hr respectively.

Higher resolution images of the the northwestern limb of the suspected G354-33 remnant were obtained with a 0.5 m Newtonian telescope at the CHILESCOPE Observatory located 25 km south of the Gemini South telescope in the Chilean Andes. A 4096 × 4096 FLI PRO-LINE 16803 CCD camera was used that provided a FOV of 1. Images of optical filaments detected in MDW images associated with GALEX FUV filaments in G249+24 were obtained with a 2.4 m telescope at the MDM Observatory at Kitt Peak, Arizona using using the Ohio State Multi-Object Spectrograph (OSMOS; Martini et al. 2011) in direct imaging mode. With a 4096×4096 CCD, this telescope/camera system yielded a clear FOV of 18 × 18 . With on-chip 2 × 2 or 4 × 4 pixel binning, spatial resolution was 0.55 and 1.10 , respectively. Three exposures of 600 to 1200 s each using an Hα + [N II] filter were taken at four positions where long slit OSMOS spectra were also obtained (see below).

Low-dispersion optical spectra of filaments in the northern suspected SNR, G249+24, were obtained with the MDM 2.4m Hiltner telescope using Ohio State Multi-Object Spectrograph (OSMOS; Martini et al. 2011) . Employing a blue VPH grism (R = 1600) and a 1.2 arcsec wide slit, exposures of 2 × 900 s were taken covering 4000-6900Å with a spectral resolution of 1.2Å pixel −1 and a FWHM = 3.5Å. Spectra were extracted from regions clean of appreciable emission along each of the 15 long slits.

As part of a program looking for high-velocity interstellar Na I absorption features in background stars in the direction of Galactic SNRs, moderate-dispersion spectra of over a dozen early-type stars with projected locations toward G249+24 and with estimated Gaia distances between 300 and 2000 pc were also observed on with the MDM 2.4m Hiltner The data were taken using a Boller & Chivens Spectrograph (CCDS) which employs a Loral 1200 × 800 CCD detector. Using a 1800 grooves mm −1 grating blazed at 4700Å yielded a wavelength coverage of 330Å with a spectral scale of 0.275Å per pixel. Exposure times varied from 1 − 2 × 400 s to 3 × 1200 s depending on star brightness and sky conditions. A 1.3 wide slit was used which resulted in a measured FWHM of 2.9 pixels, corresponding to a spectral resolution of 0.80Å providing an effective R 7400 at the Na I λλ5890,5896 lines. Although this spectral resolution is low compared to conventional interstellar absorption studies where R values typically equal or exceed 30 000, this spectrograph provided sufficient velocity resolution at the Na I lines for detecting ≥ 40 km s −1 interstellar absorption features seen in SNRs (Jenkins et al. 1976; Danks & Sembach 1995; Welsh & Sallmen 2003; Sallmen & Welsh 2004; Fesen et al. 2018; Ritchey 2020) . Wavelength calibration was done using Hg and Ne comparison lamps and have been corrected to local standard of rest (LSR) values. Deblended absorption feature velocities are believed accurate to ±4 km s −1 .

For southern hemisphere Antlia nebula, spectra were taken of suspected SNR filaments using the Robert Stobie Spectrograph (RSS) on the 10 m SALT telescope in South Africa. Using a 900 lines per mm grating and a 1.5 wide slit, spectra were obtained covering 4050 to 7120Å region with a FWHM resolution of 5Å and a dispersion of 1.0Å pix −1 . Exposures ranged from 2 × 300s to 2 × 600s.

MDM spectra were reduced using using Pyraf software and OSMOS reduction pipelines 2 in Astropy. and PYRAF 3 . L.A. Cosmic (van Dokkum 2001) was used to remove cosmic rays and calibrated using a HgNe or Ar lamp and spectroscopic standard stars (Oke 1974; Massey & Gronwall 1990) . SALT spectra were reduced using specific SALT reduction software.

A few large and unusually high Galactic latitude nebulae have been recently proposed as previously unrecognized SNRs. These objects have angular diameters well in excess of the largest confirmed SNRs, such as the 8 • diameter Vela SNR. If these objects were to be found to be true SNRs, they would expand the notion of how large an identifiable Galactic remnant can be and how far off the Galactic plane should SNR researchers be looking.

The proposed new SNRs include the huge ∼24 • diameter Antlia nebula (G275. 5+18.4; McCullough et al. 2002) and the ∼10 • diameter radio remnant G354-33 (Testori et al. 2008) . Due to limited data on it, Antlia is listed only as a possible SNR in the most recent catalogue of Galactic supernova remnants (Green 2019) , whereas no mention is made of G354-33 in either this 2019 SNR catalogue or in online SNR lists 4 . Below we discuss GALEX far UV (FUV) and optical properties of these two suspected SNRs, plus a third object, G249.2+24.4, which appears to be a new SNR based both on our FUV and optical data, and X-ray and radio data presented by Becker et al. (2021) . The presentation order follows that of our research work.

3.1. The Suspected Radio SNR G354.0-33.5

Using a 1420 MHz linear polarization survey of the southern hemisphere sky, Testori et al. (2008) identified a large ∼10 • depolarized shell at α(J2000) = 20 h 25 m , δ(J2000) = −47 • , roughly corresponding to Galactic coordinates l = 353 • b = −34 • . This shell could also be seen in a 1.4 GHz polarization all-sky survey map (Reich Reich 2009 ). Due to the object's spherical morphology and radio properties, Testori et al. (2008) suggested a SNR identification with a size suggesting a distance between 300 and 500 pc, a physical diameter of 17.4 pc × d 100pc and a z distance of 57.4 pc ×d 100pc . Testori et al. (2008) noted that the object could be seen in 408 MHz data (Haslam et al. 1982) and weakly in the SHASSA Hα images (Gaustad et al. 2001 ). This object also shows up in gradients of linearly polarized synchrotron radio emission (Iacobelli et al. 2014 ). More recently, Bracco et al. (2020) briefly commented on the presence of some GALEX FUV filaments coincident with this radio shell.

However, to our knowledge, there is no work showing this suspected SNR's FUV emissions or any associated optical emissions. Below we present large mosaics of GALEX FUV images along with a SHASSA Hα image and compare these to the low resolution 1420 MHz radio images.

3.1.1. Far UV and Hα Images The image mosaic reveals a large, coherent set of sharp UV emitting filaments arranged in a fairly circular and complete shell-like structure which is positionally coincident with the 1420 MHz radio shell. 5 Multiple and overlapping filaments are common along the shell's northern and western limbs. Because these 5 Because of distortions due to the large size of this emission structure, coordinates shown are only accurate to a few arc minutes.

filaments appear brightest in the GALEX FUV images compared to NUV images, we will concentrate mainly on the shell's far UV emission structure. Although many of the FUV filaments are unresolved at GALEX's 4.6 FWHM resolution, some filaments appear partially resolved in places, especially along its northernmost edge. It is not clear if this appearance is due to closely spaced multiple filaments or resolution of a single emission filament.

The strong shock-like morphology of these filaments and their arrangement in a nearly continuous shell lends support to Testori et al. (2008) 's suggestion for it to be a previously unrecognized Galactic supernova remnant, albeit with unusually large angular dimensions and distance from the Galactic plane.

Based on its FUV filaments, we estimate somewhat larger angular dimensions than the 10 • cited by Testori et al. (2008) . Instead, we find dimensions of 11 • East-West and 14 • North-South, with a distinct indentation along its south-western limb. Its southern extent is poorly constrained in our GALEX FUV mosaic but FUV filaments extend south to at least a Declination of −53.5 • . While largely coincident in location with the 1420 MHZ radio emission map of Testori et al. (2008) , the ∼11 • East-West diameter seen in the FUV images is centered around 20 h 16 m , -45 • 50 (l = 354.0 • , b = −33.5 • ). Hence, we will use the shorten name G354-33 based on the FUV images (see Table 1 ) which is slightly different from the G353-34 name cited by Testori et al. (2008) .

Due to its large size, many of the filamentary details visible in the full resolution FUV mosaic are lost in Figure 2. Consequently, we show in Figure 3 blow-ups of four regions, along G354-33's northern, western, southeastern, and southwestern rims to give a better sense of the object's fine-scale UV emission structure and morphology. Note that some individual filaments are two to three degrees in length, equivalent to the diameters of some of the largest confirmed Galactic SNRs.

A continuum subtracted SHASSA Hα image mosaic of the G354-33 remnant is shown in Figure 4 . This image reveals a thick diffuse shell of emission located within the boundaries of the FUV filaments. The Hα emission structure is far more diffuse and broader than that seen in the GALEX FUV image with very few obvious thin filaments. We estimate an Hα flux of around 1−2×10 −17 erg cm −2 s −1 arcsec −2 for most of the diffuse emission and about 3 times times this for the area of brighter emission along the nebula's northwestern limb.

While its Hα emission's location matches that seen in the FUV image, some weak Hα emission is seen to extend farther to the east and south than is readily visible in the GALEX FUV mosaic image. In general, however, the FUV filaments appear to mark the outer edges of the diffuse Hα emission shell. This is shown in Figure 5 where we compare FUV vs. Hα images for two northeastern sections along its limb. The upper images show the presence of weak Hα emission coincident with the bright main section of the FUV filament, with diffuse Hα emission extending toward the east bordered by the long FUV filament. The lower panels show a section along the nebula's northeastern limb where again the UV filaments are seen to mark the extent of the diffuse Hα emission. The SHASSA image also gives a hint of filamentary structure that resembles that seen in the FUV image.

Higher resolution optical images give a better sense of the remnant's true emission structure. Figure 6 presents a higher resolution CHILISCOPE Hα + [N II] image of a small portion of G354-33's northwestern limb where Figure 4 revealed a particularly bright emission patch. The upper panel of this figure shows this emission patch to be diffuse and situated mainly west of a ∼ 45 long, thin Hα filament which likely marks the western extent of the remnant in this region.

The three lower panels of Figure 6 compare the emission appearance of this Hα filament in the GALEX FUV image and in an equally deep [O III] image. The similarity of the filament as seen in the FUV and Hα images suggests G354-33's optical structure would consist of numerous thin, unresolved filaments like that seen in the FUV mosaic image shown in Figure 2 .

Based on the filament's absence of [O III] emission, its thin morphology and its brightness in the UV, it is likely to represent a non-radiative or a 'Balmer dominated' emission filament. Given the sharpness of the other G354-33 filament's in the FUV, they are also probably Balmer dominated emission filaments marking the location of the object's shock front and outer boundary.

As already noted, Testori et al. (2008) used 1420 MHz Villa Elisa radio data (Testori et al. 2001 ) to identify the radio emission ring, G354-33, as a possible SNR.

The upper panel of Figure 7 shows a 60 • × 25 • wide section of the Villa Elisa 1420 MHz polarization southern sky survey data. The emission ring appears distinct in shape and separate from other emission features along the southern edge of the Galactic plane (i.e., the righthand side of the image).

The images in the lower panels of Figure 7 show the GALEX FUV image mosaic (left panel) and the SHASSA Hα image (right panel) overlayed with the 1420 MHz Villa Elisa radio image contours. Overall, there is good agreement in both size and location for the FUV and Hα emission shells and radio ring seen in these images. This includes the greater extent toward the southern limb area. We note that the brightest portion of the radio emission lies on the object's western limb, the side facing the Galactic plane.

Examination of ROSAT on-line data showed no obvious associated X-ray emission with this nebula. This is not surprising given its location 33.5 • below the Galactic plane. Moreover, in view of the lack of any bright FUV emission nebulae along its limbs, this suggests that it lies in a region with few or none large-scale interstellar clouds which might generate significant X-ray flux.

3.2. SNR G249.7+24.7

Using GALEX FUV mosaics, we discovered a region containing several long, thin, shock-like morphology filaments aligned mostly N-S and arranged in a roughly elliptical shape ∼ 2.5×4.0 • in size. These filaments appear centered at α(J2000) = 9 h 30.5 m δ(J2000) = −16 • 40 . However, as discussed below, we believe the object to be larger and centered α(J2000) = 9 h 33 m δ(J2000) = −17 • 00 which corresponds to Galactic coordinates l = 249.7 • , b = +24.7 • . We call this object G249+24 for short. Below, we present images of this object's UV and Hα emission structure along with optical spectra of four of its filamentary regions. Figure 8 presents a 5.2 • ×5.2 • mosaic of GALEX FUV images centered on G249+24's. The image shows several UV filaments along with a few smaller diffuse emission patches. Although GALEX imaging of this region is fairly incomplete, enough imaging exists to indicate G249+24's minimum size and extent 6 .

The nebula's brightest UV filaments are concentrated in its southern region where there is a bright, long gently curved filament along its southwestern edge. This feature itself consists of several separate, closely aligned filaments. Although a smaller but a similarly bright filament caught on the western edge of a more northern GALEX image might give the impression that this filament's full length and extent is partially missing in this mosaic, this is not the case based on Hα images (see below).

Whereas the object's FUV emission filaments are readily apparent in the GALEX image mosaic shown in Figure 8 , its optical emissions are relatively faint. Although some of its brighter filaments are weakly visible on broad red passband Digital Sky Survey images, they are so faint and scattered over such a large area that it is not surprising that this nebula had not attracted prior attention.

The Hα images of the MDW All-Sky survey reveal more fully G249+24's optical emissions. In Figure 9 , we show a side-by-side comparison of G249+24's central emission structure in GALEX FUV and the MDW Hα images. In general, there is a good agreement between the nebula's UV and Hα emissions. The bright FUV and curved southwestern filament shows up strongly in Hα as does a shorter filament a degree to its east. Correlated UV -Hα emission is also seen along the upper left-hand portion of Figure 8 , where diffuse and filament type of emission is seen in both. (Note: Faint diffuse emission seen along the lower portion of the Hα image extends several degrees to both the east and west of the region and hence does not seem to be connected to G249+24's other optical features (see also Figure 10 below).

However, little in the way of Hα emission can be seen along the object's southeast limb, where in contrast one finds considerable FUV emission. A similar difference between FUV and Hα fluxes is seen for the south-central area around 9 h 28 m , −18 • 15 . Consequently, it is clear that this nebula's overall structure is more readily visible in the FUV than in even fairly deep Hα images.

Higher resolution MDM Hα images of selected regions of the nebula were also obtained and presented in Figure 11 . Although these images were taken mainly in preparation for follow-up spectral observations, they showed a surprisingly amount filamentary detail to G249+24's optical emission than initially indicated in Figure 9 . Numerous long and delicate filaments are visible in several on these images, indicating a rich and beautiful optical shock emission structure largely hidden due to its relative faintness.

Low-dispersion, exploratory slit spectra were taken at four positions (P1 through P4) in G249+24 as shown in Figures 10 and 11 in order to investigate the emission nature G249+24's optical filaments. Since only the spectra taken at P1 exhibited any [O III] emission, the four spectral plots shown in Figure 12 cover only the wavelength region 6200 to 6900Å.

The resulting spectra reveal clear evidence for its filaments being shock heated ISM like that commonly seen in SNRs. For example, spectra obtained at P2, P3 and P4 exhibited [S II]/Hα ratios of 1.26±0.20, 0.45±0.04, and 0.91±0.15, respectively, thus well above the standard criteria ratio of 0.40 indicative of shock emission. Added support for shocks is the detection of [O I] λ6300 emission in the spectra of both P3 and P4. The presence of [O I] is a secondary indicator for shock emission commonly observed in evolved SNRs (Fesen et al. 1985; Kopsacheili et al. 2020) .

The [S II] λ6716/λ6731 line ratio can be used to estimate the electron density in the S + recombination zone and is nearly independent of electron temperature (Osterbrock & Ferland 2006) . The λ6716/λ6731 ratio was found to be near the low density limit of 1.43 indicating n e <100 cm −3 ; 1.48 ± 0.04 for P2, and 1.42 ± 0.03 for P4. A bit surprisingly, this ratio for P3 was observed to be 1.15 indicating an n e density much higher than the other locations of around 500 cm −1 . Interestingly, the spectrum seen at P3 also showed much weaker [N II] λ6583 line emission compared to that seen at P2 and P4.

The spectrum taken at P1 deserves special attention due to significant line emission variations with respect to distance along the E-W aligned slit. This is shown in Figure 13 , where line emission can be seen to start at the location of the mainly N-S aligned filament extending many arcseconds to the east (left). The emission line ratio of Hα/[O III] can be seen to vary considerably down stream from the shock front. Behind the leading edge of the shock front marked by start of line emissions, Hα emission becomes strong with little or no emission in other lines. The intensity of Hα then briefly drops in strength within a distance of a few arcseconds, then increases for a short distance followed by a long, gradual decline. The [O III] emission is barely detectable at the shock front where the Hα starts but substantially increases some 10-20 arcseconds behind the shock front, to where it dominates the downstream spectrum. This is consistent with the picture of an extended post-shock cooling zone. The presence of such strong [O III] emission behind the shock front suggests shock velocities at least 100 km s −1 are present in certain filaments. In contrast, spectra at P2, P3, and P4 showed no appreciable [O III] emission indicating much lower velocity shocks, less than about 70 km s −1 .

Similar 2D emission structures have been observed in the high latitude Galactic halo SNRs G70.0-21.5 (Raymond et al. 2020) and G107.0+9.0 . They can be interpreted as emission from a ∼ 100 km s −1 shock in partially neutral gas. Neutral hydrogen atoms swept up by the shock are quickly ionized, but before that happens, some of them are excited to produce Table 3 .

Position 1 of G249+24 suggests a longer cooling length, as would occur in a slightly faster shock, perhaps ∼120 km s −1 . These structures are spatially resolved thanks to the low density of the Galactic halo and the relatively small distances to these SNRs.

The flux level of Hβ was either undetected or too faint at P2, P3, and P4 to accurately measure an Hβ/Hα ratio which could be used to estimate extinction. Only the spectrum P1 showed a weak but measurable level of Hβ flux. There the measured Hα/Hβ ratio of 2.85±0.20 suggests a very low amount of extinction. Adopting a theoretical Hα/Hβ ratio of 2.87 for 10 4 K we estimate a foreground E(B − V ) < 0.07. However, in view of the weakness of the Hβ detection, this estimate comes with a large uncertainty. Nonetheless, such a low interstellar extinction is not unexpected due to the remnant's high Galactic latitude of 24.4 • .

If this low extinction estimate is correct, then the lack of [O III] emission at P2, P3, and P4 suggests a fairly low shock velocity, around 70 km s −1 or less, in contrast to that indicated at P1. Consequently, it appears that a variety of shock speeds, from <70 to ≥120 km s −1 appear to be present throughout the nebula's structure. This conclusion is supported by Na I absorptions described below.

Moderate-dispersion optical spectra (R 7200) of several stars with projected positions within the estimated radio boundary of G249+24 (see Becker et al. (2021) and Section 3.2.5 below) were obtained in a search for high-velocity Na I interstellar absorption components, which could yield both a direct measurement of its expansion velocity plus maximum remnant distance information. Of the stars examined, two showed evidence for the presence of high-velocity (> 40 km s −1 ) Na I absorption features.

The two stars were: HD 82509 (B = 10.29, V = 10.4; A2 IV) with an estimated Gaia Early Release DR3 (EDR3) distance estimate of 825 ± 25 pc, and HD 83636 (B = 9.74, V = 9.57: A2 IV/V) with an EDR3 distance estimate of 386 ± 4 pc. Figure 10 shows the locations of these stars relative to the object's FUV and optical emissions.

Reduced and normalized spectra of both stars for the wavelength region around the Na I D lines at λ5889.95 and λ5895.92 indicating the presence of high-velocity Na I absorptions are shown in Figure 14 . Using IRAF deblending software and allowing initially chosen absorption component wavelengths to vary but keeping the Gaussian line profiles and FWHMs all the same, we found red and blue absorption features in the observed Na I profiles in both stars with LSR velocities around +55 to −73 km s −1 . In addition, to achieve a good fit to the Na absorption in the spectrum of HD 83636, an additional blue-shifted components with a radial velocity of around −123 km s −1 was needed (see Table 3 ). We note that the lower velocity components are consistent with the observed spectra seen for slit positions P2, P3, and P4 where no [O III] was seen indicating shock velocities below 70 km s −1 , while the presence of the higher velocity 123 km s −1 component is consistent with the strong [O III] seen at slit position P1.

The detection of these high-velocity Na I absorptions in these two stars can be used to estimate a maximum distance to G249+24. HD 82509's Gaia EDR3 parallax derived distance sets a distance limit of less than 825 pc, while HD 83636 sets a maximum distance of less than 390 pc. Assuming an angular diameter of 4.5 • (see below), this shorter maximum 390 pc distance implies a maximum physical size for G249+24 of ∼30 pc, similar to that estimated for the Cygnus Loop SNR ).

Examination of several on-line radio maps including the NRAO VLA Sky Survey (NVSS; Condon et al. 1998) did not show obvious radio emission in the G249+24's region. However, the 1420 MHz Bonn survey (Reich 1982) did show some emission along G249+24's FUV and optical filaments near α(J2000) = 9 h 27.5 m , δ(J2000) = −16 • 10 .

The 408 MHz all-sky survey (Haslam et al. 1981 (Haslam et al. , 1982 ) also showed enhanced emission in a roughly spherical shape approximately 4.5 • in diameter centered at 9 h 32.7 m , −16 • 45 , a position that is nearly the same indicated by FUV and optical images (see Fig. 15 ). A seemingly unrelated broad, vertical emission feature borders the emission shell on its eastern side and complicates the assessment of this emission. Assuming this emission shell is associated with the nebula, it would suggest a far more spherical SNR than indicated by either the GALEX FUV or MDW Hα images. Taking the object's largest N-S angular dimensions based of its FUV filaments (see Figs. 8 and 9) and requiring that both stars, HD 82509 and HD 83636, must lie inside the remnant's shell, we estimate G249+24 is 4.5 • in diameter assuming a circular morphology. This size estimate is in excellent agreement with that found by Becker et al. (2021) using eRosita X-ray images and archival Parks All-Sky Survey radio data. Becker et al. (2021) report the detection of diffuse Xray emission filling almost the entire remnant with a gas temperture of order kT = 0.1 keV, along with radio data indicating a spectral index of −0.69 fully consistent with its SNR nature.

McCullough et al. (2002) reported a discovery of a very large ∼24 • diameter Hα emission shell with interior ROSAT 0.25 keV X-ray emission located in the southern hemisphere and at a high Galactic latitude of +19 • . They proposed it to be a previously unrecognized SNR and named it Antlia for the constellation it is found in. McCullough et al. (2002) suggested it was an extremely old remnant with an estimated age of ∼1 Myr and located relatively nearby with an uncertain distance of between 60 pc and 340 pc.

The Antlia remnant does appears in the catalogue of high energy SNRs (Ferrand & Safi-Harb 2012) but not in the Green (2019) catalogue of confirmed Galactic SNRs, which cited the need for further observations to confirm its nature and parameters. Except for a 2007 far UV study finding weak coincident C III λ977 and C IV λλ1548,1551 line emissions from the object's interior (Shinn et al. 2007) , and a brief AAS abstract by Orchard et al. (2015) who reported Wisconsin H-alpha Mapper data indicating [S II] emission from the Antlia Nebula consistent with a SNR interpretation, there has been no detailed follow-up investigation of this exceptionally large suspected SNR. This situation plus its location at an unusually high Galactic latitude led us to include it in our high Galactic latitude GALEX FUV investigation of SNRs.

In the top panel of Figure 16 , we present a mosaic of SHASSA Hα images of the Antlia remnant at a higher resolution than the VTSS Hα image (Finkbeiner 2003) which led McCullough et al. (2002) to discover it. It shows a ∼ 20 • × 26 • Hα emission shell with a well determined boundary roughly centered at α(J2000) = 10 h 38 m , δ(J2000) = −37 • 18 corresponding to Galactic coordinates l = 275.5 • b = +18.4 • . (Note: These Galactic coordinates differ slightly from those given by McCullough et al. 2002.) A mosaic of FUV GALEX images covering just the remnant's northern limb is shown in the lower panel of Figure 13 . This image reveals a long and nearly continuous set of thin, bright FUV emission filaments and filament clusters extending approximately 20 degrees along the remnant's northeastern limb. The line of FUV filaments starts at RA = 10 h 10 m and Dec = −24 • and extends southward to the bottom left of the figure at RA = 11 h 40 m and Dec = −32 • .

Because the morphology of Antlia's FUV filaments is similar to that seen in the proposed SNRs G354-33 and G249+24, we examined MDW Hα images of the FUV filaments and adjacent regions above the −32 • Declination limit of the MDW survey. Overall, we found excellent positional agreement between FUV and Hα filaments.

To illustrate this agreement Figure 17 shows a comparison of GALEX FUV and MDW survey Hα images for three sections along the Antlia's northeastern boundary. While we find there is a close correlation of the outer FUV filaments with many Hα filaments, there are also considerable differences in terms of the nebula's internal emissions. That is, far more Hα emission is seen 'behind', i.e., to the west, of the sharp, outlying filaments than seen in the FUV images. This is most striking in the lower panel images of Figure 16 for the remnant's southeastern limb region. While the SHASSA image shows considerable internal diffuse and filamentary emission, the brightest Hα emissions are found along Antlia's northwestern, western and southeast limbs.

If McCullough et al. (2002) is correct in their assessment that the Antlia Hα nebula is a SNR, as seemingly supported by the GALEX FUV and our Hα images described above, the remnant would then be both many times larger than any confirmed SNR and be located at an unusually high Galactic latitude. On the other hand, considering its size and location immediately adjacent to the enormous Gum Nebula with its extended emission, outer 'filamentary wisps', and numerous H I wind blown bubbles (Brandt et al. 1976; Chanot & Sivan 1983; Purcell et al. 2015) , there is some uncertainty about its true origin.

Optical spectra are a powerful tool for confirming the presence of shock emission and hence can be used to in- vestigate the SNR origin of optical nebulosities. Consequently, we obtained long-slit, low-dispersion spectra of seven filaments and emission clumps distributed across the whole of the Antlia nebula to explore the nature of its optical emissions. Figure 18 shows the location of the seven filaments observed in the Antlia remnant, along with the resulting spectra. Table 2 lists the relative emission line strengths uncorrected for extinction. Listed relative line strengths are believed accurate to 10%. Based on the observed Hα/Hβ ratios of 3.0 to 4.30, we find a variable degree of extinction across the Antlia nebula, with E(B − V ) values ranging from 0.0 to 0.35 assuming an intrinsic ratio of 3.0 and an R value of 3.1. Here we have chosen an Hα/Hβ value greater than the theoretical value of 2.87 for 10 4 K due to the likelihood of significant collisional excitation of the n = 3 level at postshock temperatures seen in SNRs. Such a range of extinction values is not unexpected in view of the nebula's large dimensions.

These optical spectra show clear evidence that the seven filaments observed distributed across the whole of the Antlia nebula exhibit line ratios indicative of shock emissions. In fact, emissions at Positions 4 and 6, located along Antlia's southeastern limb display nonradiative, pure Balmer line emission. 7 Such shock spectra are usually seen in situations where the shock velocities are quite high (> 500 km s −1 ) as seen in young Type Ia SNRs like Tycho and SN 1006. However, similar Balmer dominated spectra has also been seen in much older remnants (G107.+9.0; Fesen et al. 2020) . Relatively slow shocks can remain non-radiative if the ambient density is sufficiently low, like that expected in the Galactic halo. This requires that the cooling time of the shocked gas should be long enough to prevent the forbidden lines from becoming bright. Scaling the cooling times of Hartigan et al. (1987) with n −1 0 , the non-radiative shocks at the edge of a 100,000 year old remnant in a density of 0.1 cm −3 would be at least 170 km s −1 .

The other five filaments show optical spectra commonly seen in evolved SNRs like the Cygnus Loop and IC 443. In all these cases, the I([S II])/I(Hα) ratio exceeds the 0.4 criterion for shock emission. Indeed, filaments 7 and 8 display ratios above 2, which are among the highest values observed in SNRs. Interestingly, these five filaments lack appreciable [O III] spectra along with pure Balmer emission spectra along the remnant's leading edges, there must be a considerable range of shock velocities present in the nebula. This is not too surprising given the large size of Antlia and a nearly 18 • range in Galactic latitude.

4. DISCUSSION Based on the data presented above, we find that two suspected, high latitude SNRs, namely G354-33 and Antlia, are likely bona fide SNRs. Moreover, their UV emission filaments appear best at marking these objects' forward shock front locations. Below we discuss supporting evidence for our conclusions, followed by estimates regarding their physical dimensions and general properties. We then briefly comment of the usefulness of far UV imaging for SNR detection particularly in areas far off the Galactic plane. 4.1. G354.0-33.5

There are several observations that support the SNR identification of this large FUV, Hα, and radio emission nebula. At its Galactic latitude of −33.5 • , there are no bright and nearby early type stars near the center of this UV shell that might have generated its UV emission shell through stellar winds. There is also no known nearby recurrent nova projected inside, and the emission shell is at least two orders of magnitude larger and does not have a similar appearance to known nova shells.

Then there is the large 1420 MHz radio emission shell roughly centered on and lying entirely within the borders of the FUV emission filaments. The FUV filaments also exhibit a shock-like morphology such as commonly seen in Galactic SNRs. Furthermore, its nearly continuous ring of thin filaments surrounds a broad, diffuse Hα emission shell. In summary, its shock-like filamentary appearance and the positional agreement between UV, Hα and radio emissions for such a large object located far off the Galactic plane leaves few viable options other than a SNR origin.

While we have not yet obtained optical spectra of its Hα and UV filaments, based on the morphology of the Hα emission filament along its northwestern limb and the lack of [O III] emission, we suspect this remnant may resemble that of nonradiative filaments seen in a few remnants such as those that surround the northern and eastern limb of the Cygnus Loop Fesen et al. (1982) ; Raymond et al. (1983); Hester et al. (1994) ; Sankrit et al. (2000) ; Medina et al. (2014) . A nonradiative or 'Balmer dominated' shock is one that heats plasma to a relatively high temperature and has encountered the plasma either so recently or is of such a low density that the gas has had insufficient time to radiatively cool and present optical line emissions. Such filaments are UV bright (Shull & McKee 1979; Raymond et al. 2003 ) a property consistent with GALEX FUV images for this object. Although the confirming test for the presence of nonradiative fila-ments in this remnant is spectra, the deep Hα and [O III] images presented in Figure 6 supports this conclusion.

Centered at a Galactic latitude around −33.5 • , G354-33 would rank as the highest Galactic latitude SNR found yet. In addition, if not for the Antlia remnant, this object would also be among the largest angular dimensions among the roughly 300 SNRs in the Green (2019) catalogue. Because of its ∼ 11 • × 14 • size and the fact that the largest known SNRs have physical dimensions ∼100 pc, we can make some crude estimates of its distance and properties.

If its diameter is taken to be 100 pc, then its distance is around 400 pc. Based on the presence of its stronger FUV than NUV emission, its shock velocity is likely greater than 100 km s −1 but probably less than 150 km s −1 , according to Figure 4 of Bracco et al. (2020) . That shock speed is too high for this remnant to be in the snowplow phase of SNR evolution, but the presence of radiative shock waves around much of the rim indicates that it is no longer in the Sedov phase, at least in those places. That suggests that it is in the pressuredriven shell phase, when the recently shocked gas has cooled, but there is still relatively hot, high-pressure gas in the interior.

From the calculations of Cioffi et al. (1988) , this velocity range and a diameter of around 90 pc would be consistent with an age 10 5 years and a preshock density of 0.1 cm −3 (see Fig. 14 of Fesen et al. 2020) . One might expect detectable X-ray emission during the pressuredriven shell phase, but Shelton (1998) has discussed halo SNRs in which the interior gas is too cool to produce Xrays, but still rich in high ionization states such as O VI.

GALEX FUV, SHASSA, and MDM images show it displays a highly filamentary morphology like that commonly seen in SNRs. Moreover, optical spectra of its filaments reveal emission-line ratios consistent with the presence of shocks. The discovery of its X-ray emission along with associated radio emission and archival Parks radio data (Becker et al. 2021) leave little doubt that this object is a new SNR. With an angular dimension of 4.5 • , the G249+24 remnant is among the largest of known SNRs. Situated at Galactic latitude of over 24 degrees, G249+24 also lies farther from the Milky Way's plane than any other confirmed SNR -that is, if it were not for G354-33.

The detection of high-velocity Na I absorptions in the spectrum of the star HD 83636 which has a Gaia EDR3 estimated distance of 386 pc, sets a maximum limit on G249+24's distance of less than 400 pc. This is a bit less than the lower distance range of 400-500 pc estimated by Becker et al. (2021) . If high-velocity Na I absorption lines are confimed by subsequent higher resolution echelle spectra, then the G249+24 remnant would then also rank among the closest SNRs known and with an especially robust distance limit.

The 400 pc upper limit on the distance implies an upper limit on its diameter of about 30 pc. It also provides an estimate of the ambient density. Fesen et al. (2020) used the separation between the sharp Hα filaments of G107.0+9.0 and the diffuse [O III] emission to determine that remnant's preshock density. The angular separation between the Hα and [O III] seen in Figure 13 is similar to that seen in G107.0+9.0, but here the distance upper limit is 2.5 times smaller than the 1 kpc distance adopted for G107.0+9.0. Therefore, the preshock density is at least 0.25 cm −3 .

Because the emission from G249+24 is so faint, the density cannot be much higher than 0.25 cm −3 , indicating that the distance is close to the upper limit. This is in accord with its morphology, which suggests that G247+24 is significantly above the Galactic disk and interacting with it on its southern side, indicating that it is more than 100 pc from the plane, and therefore more than 250 pc distant.

The pure Balmer line emission at P1 indicates the that remnant is still in the Sedov phase there, while the emission and absorption spectra indicate shock speeds ∼60 -100 km s −1 in most of the rest of the remnant. The faint X-ray emission and the 0.1 keV temperature found by Becker et al. (2021) also indicate that the remnant has left the Sedov phase and entered the pressuredriven shell phase. Assuming that the cooling occurred recently, when the shock speed fell below 400 km s −1 , we can combine the Sedov equations for SNR radius and shock speed to estimate and age around 15,000 years for a 10 51 erg explosion. Thus the parameters of G247+24 are similar to those of the Cygnus Loop ) but is much fainter because it lacks the dense clouds that make parts of the Cygnus Loop bright at optical and X-ray wavelengths. McCullough et al. (2002) claimed this extraordinarily large emission nebula was a likely SNR based on its appearance on the deep Hα image of the VTSS survey and on the presence of diffuse soft X-ray emission in its interior. However, this conclusion does not appear to be widely accepted, as measured by the remnant having attracted little subsequent attention.

This situation might in part be due to a reluctance by SNR researchers to accept its huge 20 • × 26 • angular size, more than 3-5 times larger than the largest known confirmed SNRs, plus its location so close to the even larger 36 • diameter Gum Nebula with its complex of large emission shells and wind-blown bubbles plus the embedded Vela SNR. Consequently, except for a far UV study by Shinn et al. (2007) and a brief AAS abstract by Orchard et al. (2015) , the Antlia remnant has not be studied in any detail, leaving open its properties and nature.

Our GALEX FUV mosaics show a well-defined shell in Hα with many individual and overlapping filaments that strongly resemble ISM shocks. In addition, the locations of sharp UV filaments along the boundary of the object's Hα emission are consistent with a SNR where such UV filaments mark the location of the remnant's shock front.

In addition, results of our seven optical spectra of Antlia's filaments clearlu indicate shock emissions throughout the nebula and hence a SNR. These spectra include two textbook cases of non-radiative Balmer dominated spectra (Positions 4 & 6) , plus several other filaments exhibiting high [S II]/Hα line ratios well above the 0.4 value distinguishing shocked from photoionized nebulae. We conclude that if the Antlia remnant was not so large, it would be an easy case for supernova remnant classification.

If the FUV and optical images and spectra presented here were not enough to confirm the Antlia nebula as a SNR, the recently released eRosita All-Sky soft X-ray 0.3-2.3 keV image reveals a large emission shell at the site of the Antlia remnant (Predehl et al. 2020; Becker et al. 2021) . Both the size and morphology of this X-ray bright limb shell matches that of the remnant's Hα as shown in Figure 16 . Although a full analysis of Antlia's X-ray emission properties is needed to assess its properties, the positional and morphological similaritiy of this X-ray shell to its FUV and Hα emissions are strong additional indicators for a SNR classification.

The physical size of the Antlia SNR is uncertain due to its unknown distance. McCullough et al. (2002) estimated a wide range of possible distances, from 60 pc to 320 pc, and believed the remnant to be extremely old, at least 1 Myr, and hence in the final snowplow phase of SNR evolution. However, our optical spectra do not support such an old object or a distance less than 200 pc.

A 1 Myr old SNR is expected to have a a very low expansion velocity of around 10 to 20 km s −1 (Chevalier 1977) . In contrast, our optical spectra show line emissions indicating shock velocities between 50 and 150 km s −1 , making it far younger (< 10 5 yr) remnant in the pressure-driven shell phase. Moreover, there are other data supporting a much younger remnant with relatively high-velocity gas inside.

Before the Antlia nebula was discovered, Bajaja et al. (1989) reported high-velocity clouds in the remnant's direction, with Penprase & Blades (1992) reporting highvelocity Ca II absorption lines in the spectrum of the B9/A0 III/IV star HD 93721 which lies in the direction of the Antlia SNR (see Fig. 15 ). This star has a estimated Gaia Early Data Release 3 (EDR3) distance of 512 ± 7 pc. Penprase & Blades (1992) found at least 10 Ca II absorption components with v LSR ranging from −65 km s −1 to +75 km s −1 . Because a second star, HD 94724 (d = 210 pc) lying in the same general direction did not show any high-velocity components, they concluded that the high-velocity absorbing cloud's distance was between 200 and 500 pc. However, because not all stars behind a SNR display high-velocity absorp-tion lines, a lack of high-velocity components does not provide a robust minimum distance estimate.

An alternative means of estimating the distance to the Antlia remnant is its apparent collision with the Gum Nebula (Gum 1952 (Gum , 1955 Brandt et al. 1971; Sivan 1974; Chanot & Sivan 1983) . Although the positional coincidence or abutment of the remnant's southwestern rim with the Gum Nebula's bright northeastern emission cloud has been noted in some studies of the Gum Nebula (Shinn et al. 2007; Iacobelli et al. 2014; Purcell et al. 2015) , no proposal has been made about there being evidence for a physical interaction between the two nebulae.

Nonetheless, the SHASSA imaging of this region strongly indicates such a physical collision has occurred between the Antlia remnant and the Gum Nebula. The left panel of Figure 19 presents a section of the VTSS Hα image of the Galactic plane showing the Antlia emission shell, similar to the VTSS image presented in the McCullough et al. (2002) Antlia discovery paper. This shows Antlia's position relative to the Gum Nebula. The remnant's bright southern limb coincides with the northern rim of the Gum Nebula.

The right panel of Figure 19 shows a low contrast version our SHASSA image of Antlia. An unusual line of bright filaments nearly 10 • in length can be seen in the projected overlap region of both shells (see also top panel of Fig. 18 ). The simplest explanation for such bright filaments situated only in the exact Antlia/Gum overlap region is that the Antlia remnant has collided with the outer northeastern rim of the Gum Nebula. Support for an Antlia-Gum collision is shown in Figure Table 4 ), the observed strong [S II] emission in the proposed collision filaments leads support for a collision scenario. This conclusion is further indicated by the right panel which shows the observed ratio of F([S II])/F(Hα + [N II]) with the stars removed. The sharp dark filaments in this panel image represent ratios between 1 and 1.3, values consistent with the measured spectral line strengths of Antlia's shock filaments (see Table 4 ).

If the Antlia remnant's expansion has led to a collision of it with an outer portion of the Gum Nebula, then the Antlia's distance must be roughly the same as for the Gum Nebula. Unfortunately, the distance to the Gum Nebula is poorly known, with values ranging from 200 − 500 pc (Brandt et al. 1971; Zealey et al. 1983 ; If we adopt recent distances estimate to the Gum of 300 -450 pc, which include estimates for some of the Gum Nebula's ionizing stars (ζ Pup; Howarth & van Leeuwen 2019), then the Antlia remnant's physical diameter is ∼ 120−180 pc. However, given the Gum Nebula's 36 • diameter, its outer parts extend over a distance of some 200 to 300 pc along the line of sight, this greatly complicates using a collision scenario to constrain the Antlia remnant's distance.

A few large Galactic SNRs and superbubbles have been recently studied in terms of their far UV emissions. For example, studies of far UV emissions have been reported for the Vela SNR (Nishikida et al. 2006) , the Cygnus Loop (Seon et al. 2006) , the Lupus Loop (Shinn et al. 2006) , and the Orion-Eridanus Superbubble (Kregenow et al. 2006) . Many of these made use of the SPEAR imaging spectrograph .

However, there have been few papers reporting discoveries of large interstellar emission structures using FUV emissions. One such paper is that of Bracco et al. (2020) who reported finding a 30 • long UV arc in Ursa Major using GALEX images. That work was a follow-up to an earlier detection of a much shorter 2.5 • filament by McCullough & Benjamin (2001) using deep Hα imaging. However, only through the GALEX's FUV images was the full extent of this faint, long interstellar filament finally revealed.

The FUV emission of moderate velocity shocks is dominated by the hydrogen 2-photon continuum, resonance lines of C IV and Mg II, and intercombination lines such as C III] and Si III]. Bracco et al. (2020) computed the GALEX FUV and NUV count rates for shocks from the Sutherland & Dopita (2017) MAPPINGS models. Shocks slower than about 100 km s −1 are dominated by the 2-photon continuum, while faster shocks have strong contributions from C IV and other lines. The predicted ratios range from about 0.1 to 0.9. Based on our own model calculations, a reddening E(B − V ) ∼ 0.1 does not affect the ratios very much, though it reduces both the NUV and FUV count rates by a factor of 2. A major caveat, however, is that scattering in edge-on sheets of emitting gas, in particular SNR filaments, can strongly reduce the intensities of resonance lines such as C IV (Cornett et al. 1992) , reducing the FUV/NUV ratio.

Above we have presented large GALEX FUV mosaic images of two large nebulae which appear to be true Galactic SNRs, but were either only suspected or missed in most radio studies which have concentrated their searches near the Galactic plane. The object G249+24 also appears to have been missed in wide FOV optical surveys due to its relatively weak Hα emissions.

Both our investigations and that of Bracco et al. (2020) indicate that broad, far UV imaging can be an especially useful means for detecting and distinguishing interstellar shocks and, in some cases, is more sensitive compared to Hα imaging. As noted by Bracco et al. (2020) , both line emissions and two photon continuum emissions contribute to FUV emission in shocks with velocities above 50 km s −1 .

To illustrate the usefulness of far UV emission imaging for detecting SNRs, Bracco et al. (2020) noted the presence of networks of thin FUV filaments in both the Antlia SNR and the recently discovered Galactic remnant G70.0-21.5 (Boumis et al. 2002; Fesen et al. 2015; Raymond et al. 2020) . In Figure 21 we show a comparison of MDW's Hα image and the GALEX FUV image of G70.0-21.5. Until the study of the SNRs discussed here, this remnant at b = −21.5 • had been the remnant with the highest Galactic latitude. This figure nicely demonstrates how that the FUV image makes the remnant easy to detect and helps to define its full dimensions despite the many missing individual GALEX images. Knowing about the existence of these GALEX FUV images could have helped Fesen et al. (2015) and Raymond et al. (2020) in their analyses of this remnant in regard to the remnant's true physical size. In summary, therefore, far UV images appear to be an especially useful tool for identifying interstellar shocks like those found in SNRs, albeit though best suited for high Galactic latitude searches.

We have investigated the nature of two large and suspected supernova remnants located at unusually high Galactic latitudes through GALEX far UV emission mosaics and optical images and spectra. This research also has uncovered one new Galactic SNR. Our findings include:

1) The large radio emission shell, G354.0-33.5, seen in 1420 MHz and 1.4 GHz radio polarization maps is very likely a SNR. The remnant exhibits numerous sharp FUV emission filaments in a thin, unbroken shell with angular dimensions of 11 • × 14.0 • . It also exhibits a coincident Hα emission shell.

2) A group of bright, sharp FUV emission filaments coincident with numerous but faint Hα filaments appears to be a previously unrecognized SNR, G249.7+24.7. Optical spectra of several filaments show evidence for the presence of 50 -150 km s −1 shocks. Deep Hα images reveal a highly filamentary morphology like that seen in evolved SNRs. Becker et al. (2021) independently discovered this remnant as a circular 4.4 • shell of diffuse, soft X-ray emission (0.1 keV) with an estimated radio spectral index of −0.69 ± 0.08. Moderate dispersion optical spectra of two stars lying in the direction of this remnant have been found to exhibit high-velocity, red and blue shifted Na I absorption lines in the range of 55 − 73 km s −1 , with one star showing an added blue absorption feature at −125 km s −1 . The closest of these two stars lies at a Gaia estimated distance of 386 pc which, if confirmed by follow-up higher dispersion data, sets a strict maximum distance to the G249.7+24.7 SNR at less than 400 pc making it one of the closest Galactic SNRs known.

3) Despite its enormous angular dimensions (20 • × 26 • ), our GALEX FUV mosaic images, plus wide-angle Hα images and optical spectra strongly support a SNR origin for the Antlia nebula. This conclusion is in line with a reported ±70 km s −1 Ca II absorption in one background star. We estimate an age ∼10 5 yr, which is an order of magnitude less than the earlier estimate ∼1 Myr. We also find the remnant is in likely in direct physical contact along its southwestern rim with the Gum Nebula. 4) Our investigation of suspected SNR located at unusually high Galactic latitudes (> 15 • ) highlights the value of UV images to detect interstellar shocks.

Follow-up work on these three objects could include optical spectra of the FUV and Hα filaments of the G354-33 remnant, wide-field [O III] λ5007 imaging of G249+24 exploring its [O III] emission, and highdispersion echelle spectra of the two apparent background stars toward G249+24 to investigate our reported high-velocity Na I absorption lines. Higher resolution Hα images and optical spectra of the series of the long and very bright filaments seen in southwestern limb of the Antlia remnant could also provide an additional test of our conclusion regarding Antlia's collision with the Gum Nebula.

Although we have found that all three nebula in our study are SNRs, only the G249.7+24.7 remnant has a distance or upper limit estimate, leaving us with only approximate physical parameters and evolutionary status. This problem can be addressed especially well for the other two objects, G354.0-33.5 and Antlia, given their enormous angular dimensions and hence offering numerous potential background stars to explore highvelocity ISM absorptions. Given the era of accurate Gaia parallaxes, obtaining high-dispersion spectra looking for high-velocity Na I λλ5890,5896 and Ca II λ3934 absorptions in background stars like that done for several SNRs could be quite fruitful (e.g., Vela: Jenkins et al. 1976; Danks & Sembach 1995; Cha & Sembach 2000; IC 443: Welsh & Sallmen 2003 ; S147: Sallmen & Welsh 2004; Cygnus Loop: Fesen et al. 2018; W28: Ritchey 2020) . Such observations might also help gauge the range of shock velocities across the very large and expansive remnants of G354.0-33.5 and Antlia.

Our findings that both the G354.0-33.5 and Antlia nebula are real SNRs was a bit of a surprise. Out of the 300 or so confirmed SNRs, there are only about a dozen SNRs larger than two degrees and less than half that number located more than ten degrees off the Galactic plane (Green 2015 (Green , 2019 . Now with the realization that some Galactic remnants can reach angular sizes several times larger than even Vela, it encourages the investigation of other large emission shells suspected as possible SNRs or previously viewed as unlikely SNR candidates. It also raises the value of SNR searches in the | b | > 10 • range, generally viewed before as unproductive.

Finally, we note that although far UV imaging appears to be a sensitive new tool for uncovering the presence of interstellar shocks, it is most useful in searching uncomplicated regions, like the objects discussed here located in the Galactic halo. However, the number of similar but unrecognized high latitude Galactic remnants is probably pretty small. Nonetheless, given the dozens of unconfirmed but seemingly likely or suspected SNRs in the literature (see list in Green 2019's catalogue), many new SNR discoveries may be aided by the use of UV imaging.

The Observatory

Superluminous Supernovae

Galactic and Extragalactic Samples of Supernova Remnants: How They Are Identified and What They Tell Us

American Astronomical Society Meeting Abstracts #225

Astrophysics of gaseous nebulae and active galactic nuclei

Roentgenstrahlung from the Universe

Cosmic Magnetic Fields: From Planets, to Stars and Galaxies

We thank Eric Galayda and the entire MDM staff for making the optical observations possible, and the SALT Observatory and Resident Astronomer staff for obtaining the excellent RSS spectra despite disruptions due to COVID-19 restrictions. We also thank R. Benjamin for helpful discussions. This work made use of the Simbad database, NASA's Skyview online data archives, and the Max Planck Institute for Radio Astronomy Survey Sampler. This work is part of R.A.F's Archangel III Research Program at Dartmouth. D.M. acknowledges support from the National Science Foundation from grants PHY-1914448 and AST-2037297.