THE PANCHROMATIC STARBURST IRREGULAR DWARF SURVEY (STARBIRDS): OBSERVATIONS AND DATA ARCHIVE^*

Table 1 lists the dwarf irregular galaxies that comprise the STARBIRDS sample and their basic properties. The sample includes known starburst galaxies, post-starburst galaxies, and galaxies newly identified as starbursts based on the significant populations of young massive stars found in their optical color–magnitude diagrams (CMDs; McQuinn et al. 2010a). The sample was originally selected based on observations available in the HST archive that were suitable for measuring the duration of starbursts in dwarf galaxies (McQuinn et al. 2009, 2010a). With this goal in mind, optical imaging was chosen to meet three conditions. First, both V and I band images were required for each system. Second, the I band observations were required to reach a minimum photometric depth of ∼2 mag below the tip of the red giant branch; a requirement for accurately constraining the recent star formation history of a galaxy (Dolphin 2002; McQuinn et al. 2010b). Third, the galaxies were close enough such that HST imaging instruments are able to resolve their stellar populations (D ≲ 6 Mpc). These criteria provided an appropriate dataset in which to study the temporal and spatial characteristics of starbursts.

The strength of this HST archival-defined sample is that the dwarf galaxies cover a range in luminosity ($-17.94\;\lt$ ${M}_{B}\;\lt -9.80$), morphology (SBdm, IAB(s)m, pec, IBm), and metallicity ($7.39\;\leqslant \;12\;+$ (O/H) $\leqslant \;8.38$). The majority of the galaxies have high Galactic latitudes and low foreground extinction (i.e., ${A}_{R}\lesssim 0.2$; Schlegel et al. 1998). NGC 1569 and NGC 6822 are two exceptions, with A_R extinction values of 1.9 and 0.6 mag, respectively. Note, however, that the sample is not a homogeneous, volume limited sample. Thus, care must be taken when interpreting the results from the sample.

Figure 2 presents three measured properties as a function of the absolute B-band magnitudes of the sample, including the stellar mass of the galaxies in the HST FOV (top panel), the SFR calculated from the GALEX FUV emission (middle panel), and the fraction of neutral hydrogen to total mass within the Holmberg radius (bottom panel; Karachentsev et al. 2004). Starburst galaxies are plotted with filled circles, while post-starburst galaxies are plotted with unfilled circles.

Zoom In Zoom Out Reset image size

The stellar mass estimates in the top panel of Figure 2 were derived from the lifetime-average SFRs from McQuinn et al. (2010b) with a 30% correction for the amount of mass returned to a galaxy over the lifetime of the stellar populations (Kennicutt et al. 1994). As expected, there is a general trend between stellar mass and luminosity. Note, however, that while the majority of the galaxies have a small enough angular size to have their stellar mass well-represented within in the HST FOV, the optical disk of three galaxies significantly extend beyond the observational footprints (see Table 1). These three galaxies, NGC 6822 (${M}_{B}=-15.22$), Ho ii (${M}_{B}=-16.72$), and NGC 4214 (${M}_{B}=-17.19$), show deviations from the general trend as the stellar mass is derived from a region that is smaller the region encompassed by the luminosity measurements.

The SFRs in the middle panel of Figure 2 were calculated from the extinction corrected FUV fluxes using the scaling relation from Hao et al. (2011). Because the FUV emission originates primarily from the photospheres of young stars of intermediate and high mass (M ≳ 3 ${M}_{\odot }$), it is a direct tracer of the recent star formation (t ∼ 100 Myr). The most widely used UV scaling relation, presented in Kennicutt (1998), is based on the theoretical results from Madau et al. (1998) over the wavelength range 1500–2800 Å. Recently, the theoretical relation was re-calibrated using updated stellar evolution libraries and spectral models (Hao et al. 2011; Murphy et al. 2011) with the Starburst99 code (Leitherer et al. 1999; Vázquez & Leitherer 2005). The resulting revision is ∼5% lower than the original value from Kennicutt (1998) for the same IMF. Uncertainties on the Hao et al. (2011) FUV–SFR relation are not well-quantified. Note that the theoretical scaling relation applied here may under-estimate the recent SFRs in these systems by 65% or more. We present a full comparison of the FUV-based scaling relation with CMD-based SFRs, including quantified uncertainties on the FUV–SFR calibration in our companion paper (McQuinn et al. 2015). No FUV data are available for one galaxy, IC 4662 (see Section 3.1). Similar to the top panel, there is a general trend between SFR and luminosity in the sample.

The fraction of neutral hydrogen to "total" mass in the bottom panel of Figure 2 is from Karachentsev et al. (1999). In that study, the hydrogen mass was calculated based on H i flux measurements; the total mass inside the Holmberg radius was estimated based on the galaxy rotation amplitude, corrected for inclination and turbulence. There is no apparent trend between the fraction of hydrogen to total mass and overall galactic luminosity.

Figure 3 compares the SFR density based on SFRs derived from the GALEX FUV emission and the area containing the majority of the FUV emission, with the SFRs derived from the GALEX FUV emission using the calibration from Hao et al. (2011). The dashed lines denote star-forming regions of constant radius from 0.1 to 10 kpc from top to bottom. Starburst galaxies are represented as filled circles; post-starburst galaxies are unfilled circles. The SFR density spans a wide range, reflecting the diversity of the star formation properties in the sample. All galaxies show elevated levels of recent or ongoing star formation levels based on their historical averages, thus representing the low-mass end of the starburst galaxies.

Zoom In Zoom Out Reset image size

Figure 4 compares the SFRs derived from the GALEX FUV emission with the stellar masses of the galaxies. Post-starburst galaxies are shown as unfilled circles and predominate at the lower SFRs. A sample of dwarf Irregular galaxies spanning an overlapping range in FUV based SFRs and stellar mass range, highlighted in shaded orange, was studied by Lee et al. (2011). Their sample was based on a sensitivity limit which probes higher SFRs in these lower mass galaxies. Also shown are best fit lines from a number of other surveys with samples of more massive galaxies at z = 0 from the SDSS (long dashed green line; Brinchmann et al. 2004), at z = 1 from the GOODS survey (dotted red line; Elbaz et al. 2007), and at z = 2 from the GOODS-S (short dashed blue line; Daddi et al. 2007). Note that the axes of the plot were chosen to cover the range in SFR and stellar mass in our data, as well as the higher rates of star formation in the more massive systems of the high redshift surveys.

Zoom In Zoom Out Reset image size

The GALEX instrument provides simultaneous FUV (1350–1750 Å) and NUV (1750–2750 Å) imaging using a dichroic beam splitter. The native pixel resolution for each detector is 1  5 pixel⁻¹. The GALEX instrument is described in detail in Martin et al. (2005), and its in-orbit performance and calibration by Morrissey et al. (2005, 2007).

New, deep GALEX observations (PID 60026) were obtained with the goal of reaching a uniform exposure time of 10.5 ks (i.e., seven orbits) in each of the GALEX UV bands when combined with existing archival data. This integration time is equivalent to reaching an FUV photometric depth of 22.5 mag, corresponding to the surface brightness expected from a constant SFR per unit area of 8 × 10⁻⁴ ${M}_{\odot }$ yr⁻¹ kpc⁻² at a distance of 3 Mpc. This depth was targeted to facilitate an exploration of the outer regions in starburst galaxies where lower SFRs predominate. Preceding the execution of the new observations, the FUV detector aboard GALEX failed. Thus, the data for the FUV band is comprised solely of archival data with integration times ranging from ∼1.5 ks (1 orbit) to ∼21 ks (14 orbits). One system, IC 4662, did not have archival observations and therefore lacks FUV observations; the NUV observations were executed as planned. We include IC 4662 NUV data for completeness. Further, due to a calibration change that occurred as a result of Coarse Sun Point event on 2010 May 4th,⁴ some of the affected observations failed the quality assurance tests and are not included. Table 2 lists the final GALEX exposure time achieved per system.

Table 2.  Summary of GALEX Observations

	Arch.	Arch.		Total	Total
	FUV	NUV	Archival	FUV	NUV	Final	Bkgd. Sub.	Masked	FluxFUV × 10−14	FluxNUV × 10−14
Galaxy	(ks)	(ks)	Tile name	(ks)	(ks)	Tile name	Method	Image	(erg s−1 ${\mathrm{cm}}^{-2}\;\mathring{\rm{A}}$)	(erg s−1 ${\mathrm{cm}}^{-2}\;\mathring{\rm{A}}$)
(1)	(2)	(3)	(4)	(5)	(6)	(7)	(8)	(9)	(10)	(11)
Starburst Galaxies
UGC 4483	1.5	1.4	$\mathrm{NGA}\_$ UGC4483	1.5	11.2	GI6_026202_UGC4483	pipeline	Y	1.7 ± 0.5	0.87 ± 0.13
UGC 6456	1.6	1.6	$\mathrm{NGA}\_$ VIIZw403	2.1	11.6	GI6_026203_UGC6456	pipeline	Y	3.0 ± 0.6	1.6 ± 0.2
	0.4	0.4	$\mathrm{AIS}\_117$
DDO 165	1.7	1.7	$\mathrm{NGA}\_$ DDO165	1.7	11.3	GI6_026208_DDO165	pipeline	Y	6.4 ± 0.9	3.8 ± 0.3
IC 4662	⋯	⋯		⋯	9.9	GI6_026007_IC4662	pipeline	Y	⋯	16 ± 1
NGC 6822	6.3	7.8	$\mathrm{NGA}\_$ NGC6822	6.3	10.2	GI6_026214_NGC6822	custom	Y	3.7 ± 0.7	1.7 ± 0.2
						GI6_026214_NGC6822	custom	Y	4.4 ± 0.8	2.3 ± 0.2
						GI6_026214_NGC6822	custom	Y	5.6 ± 0.9	2.3 ± 0.2
NGC 4068	2.6	3.5	GI4_095022_NGC4068	2.6	11.3	GI6_026206_NGC4068	pipeline	Y	7.5 ± 1.0	4.1 ± 0.3
NGC 2366	2.9	2.9	$\mathrm{NGA}\_$ NGC2366	2.9	7.2	GI6_026210_NGC2366	pipeline	Y	31 ± 2	17 ± 1
ESO 154-023	1.7	1.7	GI4_095006_ESO154_G023	1.7	9.9	GI6_026209_ESO154m023	pipeline	Y	9.1 ± 1.1	5.1 ± 0.3
NGC 784	2.2	2.2	GI4_095004_NGC0784	2.2	7.6	GI6_026212_NGC784	pipeline	Y	6.8 ± 0.9	4.7 ± 0.3
Ho II	21.1	21.1	GI3_050003_HolmbergII	22.8	22.8	GI6_026218_HolmbergII	custom	Y	32 ± 2	17 ± 1
	1.7	1.7	$\mathrm{NGA}\_$ HolmbergII
NGC 4214	2.0	2.0	GI4_095023_NGC4214	2.0	10.2	GI6_026215_NGC4214	custom	N	54 ± 3	33 ± 1
NGC 5253	1.7	1.7	GI4_095049_NGC5253	2.3	4.5	GI6_026213_NGC5253	pipeline	Y	33 ± 2	24 ± 1
	0.6	0.6	$\mathrm{NGA}\_$ NGC5253
IC 2574	13.0	13.0	GI3_050005_IC2574	15.0	16.5	GI6_026219_IC2574	custom	Y	34 ± 2	18 ± 1
	1.9	3.5	$\mathrm{NGA}\_$ IC2574
NGC 1569	7.1	7.1	$\mathrm{NGA}\_$ NGC1569	7.1	7.1	NGC_NGC1569	pipeline	Y	3.2 ± 0.7	2.8 ± 0.2
NGC 4449	3.3	1.7	GI4_095051_NGC4460	4.2	15.5	GI6_026216_NGC4449	custom	Y	160 ± 5	85 ± 1
	0.9	0.9	$\mathrm{NGA}\_$ NGC4449
Post-starburst Galaxies
Antlia Dwarf	13.8	17.2	NGA_Antlia_Dw	13.8	17.2	NGA_Antlia_Dw	custom	Y	0.33 ± 0.22	0.39 ± 0.09
UGC 9128	2.6	5.8	GI4_016007_DDO187	2.6	7.6	GI6_026201_UGC9128	pipeline	Y	1.2 ± 0.4	0.82 ± 0.13
NGC 4163	1.7	1.7	GI3_061001_NGC4163	12.1	15.6	GI6_026217_NGC4163	pipeline	Y	2.5 ± 0.5	1.8 ± 0.2
	10.3	13.4	GI4_015004_NGC4163
NGC 6789	1.5	1.5	$\mathrm{NGA}\_$ NGC6789	1.5	2.9	GI6_026204_NGC6789	pipeline	Y	0.8 ± 0.3	0.64 ± 0.11
NGC 625	4.5	4.5	GI1_047007_NGC0625	4.5	11.3	GI6_026211_NGC625	custom	N	9.7 ± 1.2	6.2 ± 0.4

Note. Column (1): galaxy name. Columns (2) and (3): FUV and NUV integration time for the GALEX archival observations. Column (4): tile name for archival observations. Columns (5) and (6): total FUV and NUV integration time for the GALEX observations including the archival observations plus new legacy observations (PID 60026). Column (7): final tile name for the GALEX co-added observations. Column (8): background subtraction was performed as part of the GALEX v7.1 pipeline for all but six galaxies; the remaining six galaxies (including three fields of view for NGC 6822) showed emission structure from the galaxy or cirrus contamination in the sky background images and thus the background subtracted intensity images were over- or under-subtracted. We measured and subtracted the background separately for these systems as described fully in Section 4.1. Column (9): foreground and background contamination were masked from the GALEX images as noted. Columns (10) and (11): integrated FUV and NUV flux measurements from the GALEX images in the HST field of view with appropriate foreground and background contamination masked. Uncertainties assume a poisson distribution and are based on the square root of the number of counts at each waveband.

Download table as:  ASCIITypeset image

This study used observations from two instruments aboard the HST, the Advanced Camera for Survey instrument (ACS; Ford et al. 1998) and the WFPC2 instrument (Holtzman et al. 1995). Details on the HST observations, including exposure times are presented in Table 3. The ACS observations were taken with the wide-field camera's V filter (F555W or F606W) and I filter (F814W). The ACS instrument has a native pixel scale of 0  05 pixel⁻¹ and a FOV covering 202' × 202'. The WFPC2 observations were taken with the V filter (F555W) and the I filter (F814W). The WFPC2 instrument has three wide field CCDs, each providing an 800 × 800 pixel format with a pixel scale of 0  1 pixel⁻¹, and one planetary camera CCD providing the same pixel format but with a smaller pixel scale of 0  05 pixel⁻¹ and a correspondingly smaller FOV.

We used archival images from the Spitzer obtained in the mid-IR regime from the MIPS instrument (Rieke et al. 2004) at 24, 70, and 160 μm. The MIPS instrument produces images with FOV for the 24, 70, and 160 μm bands of 5' × 5', 2&farcm;5 × 5', and 0&farcm;5 × 5', respectively, resampled and mosaicked with corresponding pixel scales of 1.5, 4.5, and 9  0 pixel⁻¹.

Table 4 lists the program ID of the MIPS observations and the total exposure times for each galaxy. Sixteen of the twenty galaxies in the sample were originally observed as part of either the Local Volume Legacy Survey (LVL; Dale et al. 2009) or the Spitzer Infrared Nearby Galaxy Survey (SINGS; Kennicutt et al. 2003; Regan & The Sings Team 2004) projects. These MIPS observations were taken using a scan-mapping mode in two separate visits allowing for the removal of detector artifacts and transient objects (e.g., asteroids) in the final images. Observing in the scan-map mode causes integration times to vary per pixel in each image. The exact observational depth is provided in a weight map for each galaxy with an average exposure time in the center pixels listed in the image headers. For the remaining four galaxies, the MIPS observations available from the Spitzer Heritage Archive (SHA) were incomplete. The observations were short (t $\lt \;10$ s) and, in the case of IC 4662, did not cover the HST FOV. In addition, the archival data were available at only one or two of the three MIPS bandpasses.

Table 4.  Summary of Spitzer MIPS Observations

	Spitzer		24 μm	70 μm	160 μm	Flux24 μm	Flux70 μm	Flux160 μm	24 μm	70 μm	160 μm
Galaxy	Program ID	PI	(s)	(s)	(s)	(Jy)	(Jy)	(Jy)	Masked	Masked	Masked
(1)	(2)	(3)	(4)	(5)	(6)	(7)	(8)	(9)	(10)	(11)	(12)
Starburst Galaxies
UGC 4483	LVL	Kennicutt	160	80	16	0.07	0.1	ND	Y	N	N
UGC 6456	59	Rieke	3	⋯	10	0.03	⋯	ND	Y	⋯	N
DDO 165	LVL	Kennicutt	160	80	16	ND	0.06	ND	Y	Y	N
IC 4662	3567	Vacca	3	10	⋯	>5.0	$\gt 9.0$	⋯	Y	N	⋯
NGC 6822	SINGS	Kennicutt	160	80	16	0.2	3.2	6.0	Y	N	N
						0.01	3.2	6.9	N	N	N
						0.2	3.2	4.0	N	N	N
NGC 4068	LVL	Kennicutt	160	80	16	0.02	0.7	1.0	Y	N	N
NGC 2366	LVL	Kennicutt	160	80	16	0.7	5.0	3.7	Y	Y	Y
ESO 154-023	LVL	Kennicutt	160	80	16	46	0.7	1.1	N	N	N
NGC 784	LVL	Kennicutt	160	80	16	0.03	1.1	1.4	Y	N	N
Ho II	LVL	Kennicutt	160	80	16	0.2	2.7	3.0	Y	N	N
NGC 4214	LVL	Kennicutt	160	80	16	1.7	17	19	N	N	N
NGC 5253	LVL	Kennicutt	160	80	16	8.8	26	20.	Y	N	N
IC 2574	LVL	Kennicutt	160	80	16	0.2	3.6	6.9	Y	N	N
NGC 1569	69	Fazio	3	⋯	3	7.4	⋯	⋯	N	⋯	N
NGC 4449	LVL	Kennicutt	160	80	16	3.1	41	71	Y	N	N
Post-starburst Galaxies
Antlia Dwarf	LVL	Skillman	107	42	8.4	ND	ND	ND	N	N	N
UGC 9128	LVL	Kennicutt	160	80	16	ND	ND	ND	Y	N	N
NGC 4163	LVL	Kennicutt	160	80	16	0.003	0.07	0.2	Y	N	N
NGC 6789	40301	Fazio	10	⋯	⋯	0.008	⋯	⋯	Y	⋯	⋯
NGC 625	LVL	Kennicutt	160	80	16	0.8	5.6	5.8	N	N	N

Note. Column (1): galaxy name. Columns (2) and (3): Spitzer program ID and PI. Columns (4)–(6): exposure times for the LVL and SINGS data are the approximate times for the central pixel in each map. Columns (7)–(9): flux measurements in the HST field of view after any evident foreground and background contamination has been masked. ND is a non-detection at that wavelength. Lower limits on the flux are noted for IC 4662 where the MIPS observations did not cover the HST field of view. Columns (10)–(12): foreground and background contamination were identified and appropriately masked from the MIPS images as noted.

Download table as:  ASCIITypeset image

The FUV and NUV data were processed through the GALEX pipeline (v7.1) which converts the GALEX satellite telemetry data (i.e., photon lists) into calibrated intensity maps in counts per second. The v7.1 pipeline accounts for the change in the NUV calibration due to the Coarse Sun Point event on 2010 May 4th. The intensity maps include corrections from flat-fielding and account for a drift in the photometric calibration (0.25% and 1.5% fainter per year for the FUV and NUV bands, respectively). The GALEX pipeline also co-adds the archival and new observations into one FUV and one NUV image per target. As the original observations used to create the deeper, co-added tiles can have slightly varying FOV, the GALEX pipeline provides high resolution relative response maps which measure the depth of the observations per pixel. Table 2 provides the tile names for the co-added intensity maps.

In addition to a co-added intensity map, the GALEX data pipeline also provided a background image and a background subtracted intensity map (in counts s⁻¹; Morrissey et al. 2007, see Section 3.3). For most observations, the pipeline background images provide a robust measurement of the sky flux and thus, the pipeline background subtracted images were used in our analysis and in the archive. However, for higher surface brightness galaxies, the background images included clearly visible structure from the galaxies. In addition, observations for one galaxy (Antlia Dwarf) showed cirrus contamination in the subtracted image. For these systems, we replaced the background flux in a 3 × D₂₅ region centered on the target galaxies with the median flux value from a region of at least 1000 pixels within a 3 × D${}_{25}-4\;\times$ D₂₅ mag arcsec⁻² annulus of the galaxy, where D₂₅ is the major diameter measured to B-band 25 mag arcsec⁻² isophote reported in Karachentsev et al. (2004). The galaxies for which we performed this custom background subtraction process are identified in Table 2.

The HST data include single and multiple pointing observations. For all single pointing observations, we used the co-added images available through the Hubble Legacy Archive (HLA). For the five galaxies with multiple pointings (NGC 4449, NGC 5253, NGC 2366, IC 2574, and Ho II), co-added mosaicked images of only one system (NGC 4449) were available through the HLA. For the remaining four galaxies, we created the co-added mosaicked images using MultiDrizzle software on the STScI python 2.7 platform. No additional processing was performed on the HST FITS mosaic images. These HST images provide the reference FOV of the star formation and are used for comparison at other wavelengths (McQuinn et al. 2015). Note that photometry is best performed on the raw images available through MAST, which are processed using the standard HST pipeline.

For the sixteen galaxies observed as part of the LVL or SINGS Spitzer legacy programs, we used the enhanced data products provided by these programs. A detailed description of the data processing can be found in Dale et al. (2009) and in the data delivery documents of the surveys. Here, we list the main data reduction steps used by the LVL and SINGS team.

The MIPS observations were processed using the MIPS Data Analysis Tool (MIPS DAT Gordon et al. 2005) and calibrated in MJy sr⁻¹. Data processing for the MIPS 24 μm data included a droop correction, a nonlinearity correction in the ramps, and a dark current correction. The images were flat-fielded, latent images were masked, and jailbar patterns corrected. The background was estimated using a fit with a third order polynomial and subtracted from the images. Scattered light was also subtracted.

Data processing for the MIPS 70 and 160 μm data included a linear fit to the ramps, which also removes cosmic rays and readout jumps, and applies an electronic nonlinearity correction. Stimflash frames, or on-board calibration stimulator frames used to track the detector response carefully, were used for responsivity corrections. The images had dark current subtractions, illumination corrections, and drift corrections. Short-term variations caused by drift were subtracted and cosmic rays were masked.

For the four galaxies that were not observed as part of these larger Spitzer legacy programs, we created the MIPS mosaics from the data available through the SHA. Using the Spitzer MOPEX software, we deleted the first frames (which were taken in the high dynamic mode), flat-fielded the observations, and created mosaic images from the archival data. The median background was estimated in a region outside of the galaxy using the IRAF task imstat and subtracted from the mosaic. The 70 μm observations for NGC 1569 and NGC 6789 were unusable due to the presence of stimflash in all frames.

The first GALEX FOV covers the HST observational footprint. The GALEX images were cropped using the WCSTOOLS task getfits based on the center coordinates, full axes size, and orientation from the HST astrometry solution in the FITS headers. While these cropped, square images match the size of the HST FITS files, they do not match the actual HST FOV as the ACS instrument is affected by distortion and the WFPC2 instrument has an irregularly shaped footprint. Thus, the cropped, square GALEX images were matched pixel-to-pixel to the HST observations; flux values in the UV images were set to zero if the corresponding locations in the HST images were not observed. For non-dithered, split orbit, HST ACS observations, the chip gap was also excluded in these cropped images. Note that the UV images are up to 1 pixel larger in length and width than the HST data due to the different pixel scales of the GALEX and HST data. Figure 6 presents a representative image of the GALEX data for the same galaxy shown in Figure 5, NGC 4068.

Zoom In Zoom Out Reset image size

The MIPS images were cropped to the FOV of the HST observations following the same two step process employed for the GALEX data. First, a square cropped image was created using the WCSTOOLS task getfits based on the center coordinates, full axes sizes, and orientation from the HST astrometry solution in the FITS headers. Second, these cropped images were matched pixel-to-pixel to the HST observations; flux values in the MIPS images were set to zero if the corresponding location around the edges in the HST images or in the HST ACS chip gap were unobserved. Some of the MIPS FOV did not encompass the entire HST FOV. In these cases, the MIPS total flux measurements are lower limits for the HST FOV. Figure 7 presents representative images of the MIPS data for the same galaxy shown in Figures 5 and 6, NGC 4068.

Zoom In Zoom Out Reset image size

The second GALEX FOV covers 4 × D₂₅ for each galaxy. Whereas the FOV defined by the HST instruments does not cover the complete stellar disk of the more extended galaxies, this larger FOV provides complete areal coverage beyond the stellar disks on each system. The native 125 circular FOV of the GALEX images was cropped using the WCSTOOLS task getfits centered on the RA and Dec coordinates of each system with a length and width equal to 4 × D₂₅. The coordinates and D₂₅ values are provided in Table 1.

The HST V and I band images have a pixel scale of 0  05 for the ACS instrument and 0  1 for the WFPC2 instrument after processing by the HLA. The FWHM of the PSF is ∼2 pixels for the HST images. Following the procedure employed by the NRAO Astronomical Image Processing System software, we convolved the smaller HST PSFs with the larger ∼5'' PSF of the GALEX detectors. Once convolved with the larger PSF, we re-binned the images using the IDL task frebin and mapped the images to same coordinate system using the world coordinate systems header information and the IDL tasks triangulate and griddata. The resulting images enable a comparison of the optical morphological features with the morphological features at UV and MIR wavelengths.

Comparing the reprocessed HST images with the GALEX images, there are numerous morphological features that are traced at both wavelengths, as expected. There are also notable differences between some of the features seen in one and not the other image. In contrast, while the reprocessed HST images and the MIPS 24 μm images often trace the same star-forming regions, the details of the morphological features are generally different. These similarities and differences are primarily due to the different sources of the emission at UV (massive, young stars), optical (stars of varying masses and ages) and MIR (dust heated by stars) wavelengths. Detailed analysis of the star formation at these three wavelength regimes are reported in a companion paper (McQuinn et al. 2015).

Grayscale JPEG files were created using the Montage software routine mJPEG for all of the fits files. The JPEG files were made using a Gaussian stretch of the full range of the original image up to a maximum flux level of 99.999% of all pixel values. Two color JPEG images were created from the HST V and I band optical FITS files using the same software and stretch parameters. The V band image was used for both the blue and green colors, and the I band image was used for the red color. The JPEG images showing the HST footprint on a larger DSS blue optical image were created using Aladin Sky Atlas (Bonnarel et al. 2000).

All JPEG images that covered the HST FOV were re-sized to be $\sim \displaystyle \frac{1}{2}$ page using the software KStudio. The larger GALEX images extending to the 4 times D₂₅ distance were resized to be approximately a full page using the same software. Note, therefore, that these larger GALEX JPEG images do not have the same angular scale.