A CATALOG OF ULTRA-COMPACT HIGH VELOCITY CLOUDS FROM THE ALFALFA SURVEY: LOCAL GROUP GALAXY CANDIDATES?

ALFALFA is an extragalactic spectral line survey making use of the Arecibo 305 m telescope. The survey maps 7000 deg² of sky in the H i 21 cm line, covering the spectral range between 1335 and 1435 MHz (roughly −2500 km s⁻¹ to 17500 km s⁻¹ for the H i line), with a spectral resolution of 25 kHz, or ∼5.5 km s⁻¹ (at z = 0). ALFALFA is designed to outperform previous blind H i surveys. With an angular resolution of ∼35, ALFALFA can resolve structures one-fourth the angular size possible with the HI Parkes All Sky Survey (HIPASS; Meyer et al. 2004) and one-ninth that possible with the Leiden Dwingeloo Survey (LDS; Hartmann & Burton 1997). Its flux density sensitivity is nearly one order of magnitude higher than that of HIPASS and more than two orders of magnitude better than that of the LDS. ALFALFA can detect a ∼5 × 10⁴ M_☉ cloud of 20 km s⁻¹ linewidth at a distance of 1 Mpc. A full description of the observational mode of ALFALFA is given in Giovanelli et al. (2007), while the definition and goals of the survey are described in Giovanelli et al. (2005). Only a summary of the observational details is given here.

ALFALFA surveys the sky using a seven-feed multi-beam receiver in "drift" mode: the telescope is normally parked along the local meridian and 14 tracks (7 feeds, 2 polarizations each) of spectral data of 4096 channels—each acquired continuously and recorded at a 1 Hz rate as the sky drifts by. All regions of the sky are visited twice with the two visits typically a few months apart in time. Upon completion of data taking of a region of the sky, data cubes of 24 × 24 in spatial coordinates are produced and sampled over a regular grid of 1' spacing in R.A. and decl. After Hanning smoothing to 11 km s⁻¹ resolution, the rms noise per channel of the data is typically 2–2.5 mJy per beam. In general, sources are extracted from the data cubes through a two-step process. An automated signal identification algorithm is first run over each data cube, producing a preliminary source catalog (Saintonge 2007). Then each source in the catalog is visually inspected and remeasured. The measurement tool fits ellipses to contours of constant flux density level and delivers a source position, given by the center of the ellipse encircling half of the total flux density of the source, source sizes (as the major and minor axes of said ellipse), flux density, velocity, and linewidth.

As mentioned above, the standard source identification and measurement in ALFALFA uses the algorithm developed by Saintonge (2007) to identify sources and is then followed by measurement of the source by hand. Briefly, the identification algorithm is a one-dimensional matched filtering scheme. The spectrum in each pixel of an ALFALFA grid is matched to a series of Hermite polynomial templates. The detection of a galaxy requires the detection of spectra of similar velocity widths with a high significance in 5 or more contiguous pixels.

In comparison with the extragalactic sources identified in the α.40 catalog, the UCHVCs are typically spatially extended and have narrow velocity widths. This is illustrated in Figure 2 where the distribution of H i angular diameters and velocity widths are plotted for the α.40 extragalactic sources, α.40 HVCs and the UCHVCs of this work. The UCHVCs are spatially extended compared to the extragalactic sources but generally small compared to the full HVC population of the α.40 survey. The minimum velocity width used in the templates of the Saintonge (2007) identification algorithm is 30 km s⁻¹, the typical maximum width of the UCHVCs. For this reason, a special source identification algorithm was developed for the UCHVCs in addition to the standard ALFALFA pipeline. This method is based on the philosophy of Saintonge (2007), but with three main differences: a limited velocity range, three-dimensional matched filtering, and the use of Gaussian templates. Only a limited velocity range of the ALFALFA data set, −500 < v_☉ < 1000 km s⁻¹, is selected as this is the expected velocity range for objects within the Local Volume. Because only a limited velocity range is examined, it is reasonable to perform a full three-dimensional matched filtering, matching both the spectrum of the source and the spatial position and size simultaneously. Gaussian templates are used to describe both the spatial extent and the velocity profile of the UCHVCs. The templates range from a spatial FWHM size of 4' to 12' in steps of 2' and the spectral line FWHM ranges from 10 km s⁻¹ to 40 km s⁻¹ in steps of 6 km s⁻¹. The lower bound of the spatial templates is set by the beam size of Arecibo. The upper size bound is near the median size value of the UCHVCs and represents our emphasis on detecting ultra-compact clouds. UCHVCs can be larger in size than 12' and the matched filtering of the 12' template to a UCHVC with H i diameter greater than 12' is robust. A velocity FWHM of 10 km s⁻¹ represents the narrowest source that can be spectroscopically resolved in the ALFALFA data. The warm neutral medium is thought to be the dominant phase of the ISM in minihalos (e.g., Sternberg et al. 2002); for a reasonable range of temperatures (6000–10000 K) for the warm neutral medium in the UCHVCs, thermal broadening results in linewidths of ∼16–21 km s⁻¹. Thus for a cloud of 40 km s⁻¹ linewidth, we would expect the large scale motion to be ∼34 km s⁻¹ for the warmest clouds, after subtracting the thermal broadening contribution in quadrature. For a typical size of 10' at an indicative distance of 1 Mpc, the dynamical mass based on this unbroadened linewidth is ∼10⁸ M_☉. This is a reasonable upper limit to the dynamical mass we may expect to be traced out for a more massive dark matter halo of ≲ 10¹⁰ M_☉, and matches the dynamical mass traced by the baryon extent of the presumably more massive SHIELD galaxies (Cannon et al. 2011).

Zoom In Zoom Out Reset image size

Many of the UCHVCs are missed by the standard identification algorithm. Since the ALFALFA pipeline also involves visual inspection of the dataset, most of these sources are identified by eye and included in the α.40 catalog as HVC detections. The specialized UCHVC identification algorithm does find sources that are missed by the standard ALFALFA pipeline; of the 59 UCHVCs identified here (listed in Table 1), 5 sources are not included in the α.40 catalog. Three of these are in the spring sky and two in the fall sky. Figure 3 shows the measured properties of all the UCHVCs compared to the five sources not included in the α.40 catalog. The additional sources tend to have low integrated flux densities and narrow linewidths (W₅₀). While they have a range of H i diameters, they are not the most compact clouds. Most strikingly, the UCHVCs not included in the α.40 catalog are the sources with the lowest average column densities, suggesting that these sources are the tip of the iceberg for further clouds to be detected.

Zoom In Zoom Out Reset image size

Table 1. ALFALFA UCHVCs in the α40 Survey

Source	AGC	R.A.+decl.	cz☉ VlsrVgsr VLG	W50(w)	a × b	S21	S/N	N3	N10	Notes
		(J2000)	(km s−1)	(km s−1)	(')	(Jy km s−1)
HVC111.65-30.53-124a	103417	000554.3+312014	−128 −124 55 139	21(8)	27 × 15	2.31	12	0	9	g, O
HVC123.11-33.67-176	102992	005206.2+291204	−177 −176 −19 61	21(3)	24 × 10	1.28	9	0	5	g, O
HVC123.74-33.47-289c	102994	005431.6+292402	−290 −289 −133 −52	21(1)	6 × 5	0.67	15	0	13	g, O
HVC126.85-46.66-310	749141	010237.8+160752	−308 −310 −186 −112	23(6)	10 × 8	0.81	9	1	21	g, O
HVC131.90-46.50-276a,c	114574	011703.4+155548	−273 −276 −160 −88	27(4)	10 × 6	0.71	9	1	22	g, O
HVC137.90-31.73-327	114116	014952.1+292600	−325 −327 −199 −124	34(8)	29 × 16	3.93	13	1	9	g, O
HVC138.39-32.71-320	114117	015031.4+282259	−317 −320 −194 −119	22(2)	19 × 13	4.41	30	1	7	g, O, S12
HVC154.00-29.03-141	122836	025229.7+262630	−135 −141 −55 8	27(3)	29 × 15	6.90	31	0	15	g, O
HVC205.28+18.70+150★	174540	074559.9+145837	162 150 59 42	23(4)	10 × 6	2.06	28	0	2	g, S12, O
HVC196.50+24.42+146	174763	075527.1+244143	156 146 88 79	20(2)	16 × 11	2.80	20	3	11	g, S12, O
HVC196.09+24.74+166	174764	075614.8+250900	175 166 110 101	24(6)	10 × 5	0.66	9	3	10	p, O
HVC198.48+31.09+165	189054	082546.7+251128	173 165 104 90	26(1)	19 × 13	1.77	13	0	8	g, O
HVC204.88+44.86+147★	198511	093013.2+241217	152 147 80 53	15(1)	8 × 6	0.73	14	0	0	g, S12, O
HVC234.33+51.28+143	208315	102701.1+084708	148 143 29 −22	20(2)	15 × 10	4.96	35	0	16	g, S12
HVC250.16+57.45+139	219214	110929.8+052601	142 139 25 −32	20(5)	7 × 4	0.56	10	0	9	g, G10, S12
HVC252.98+60.17+142	219274	112119.6+062132	143 142 35 −22	27(5)	28 × 15	8.55	37	1	10	g, S12
HVC253.04+61.98+148	219276	112624.8+073915	149 148 47 −8	36(1)	14 × 12	2.06	14	1	11	g
HVC255.76+61.49+181	219278	112855.6+062529	182 181 77 19	18(2)	11 × 6	0.90	13	0	7	g, S12
HVC256.34+61.37+166c	219279	112928.6+060923	167 166 61 3	24(1)	12 × 11	1.49	14	2	11	g
HVC245.26+69.53+217★	215417	114008.1+150644	216 217 146 97	17(4)	10 × 9	0.70	9	0	1	g, G10
HVC277.25+65.14-140★	227977	120920.0+042330	−142 −140 −234 −294	23(1)	7 × 4	0.46	8	0	1	g, G10
HVC274.68+74.70-123★	226067	122154.7+132810	−128 −123 −182 −232	54(13)	5 × 4	0.92	11	0	0	p, G10
HVC290.19+70.86+204	226165	123440.2+082408	200 204 135 80	21(1)	10 × 6	0.90	11	1	15	g
HVC292.94+70.42+159a	229344	123758.5+074849	154 159 89 34	15(4)	17 × 14	1.67	13	0	18	g
HVC295.19+72.63+225	226170	124204.6+095405	220 225 164 112	28(7)	14 × 12	1.17	10	3	16	p, G10
HVC298.95+68.17+270★	227987	124529.8+052023	265 270 196 139	26(1)	16 × 9	5.58	44	0	4	g, G10
HVC324.03+75.51+135	233763	131242.3+133046	127 135 102 56	29(1)	7 × 5	0.94	18	1	12	g
HVC320.95+72.32+185	233830	131321.5+101257	177 185 141 92	23(9)	21 × 16	1.70	9	0	15	g, G10
HVC330.13+73.07+132	233831	132241.6+115231	124 132 100 53	16(1)	6 × 3	0.63	11	0	11	g, G10
HVC326.91+65.25+316★	238713	133043.8+041338	308 316 264 210	26(4)	12 × 10	1.25	11	0	0	p, G10
HVC 28.09+71.86-144★	249393	141058.1+241204	−157 −144 −111 −136	43(6)	15 × 9	1.12	8	0	0	g, O
HVC353.41+61.07+257★	249323	141948.6+071115	246 257 244 201	20(4)	13 × 9	1.34	13	3	4	g, G10
HVC351.17+58.56+214★, b	249282	142321.2+043437	203 214 196 151	40(8)	7 × 5	1.45	17	0	4	p, G10, S12
HVC352.45+59.06+263★	249283	142357.7+052340	252 263 248 203	32(9)	16 × 11	1.11	8	3	4	g, G10
HVC356.81+58.51+148★	249326	143158.8+063520	136 148 141 100	38(11)	6 × 5	0.70	10	0	1	p
HVC 5.58+52.07+163★	258459	150441.3+061259	149 163 176 141	24(8)	11 × 10	1.33	13	0	4	g
HVC 13.59+54.52+169★	258237	150723.0+113256	155 169 200 170	23(3)	10 × 5	1.34	17	1	3	g
HVC 13.60+54.23+179★	258241	150824.4+112422	164 179 210 180	17(1)	15 × 7	0.99	11	1	4	g
HVC 13.63+53.78+222★	258242	151000.6+111127	207 222 253 224	21(2)	9 × 6	0.71	9	0	1	g, G10
HVC 26.11+45.88+163	257994	155354.0+144148	146 163 232 217	23(3)	12 × 7	2.04	22	2	8	g
HVC 26.01+45.52+161	257956	155507.5+142929	144 161 230 215	25(6)	8 × 6	1.54	14	2	8	g
HVC 29.55+43.88+175	268067	160529.4+160912	158 175 255 244	37(11)	10 × 6	1.91	20	2	6	g, G10
HVC 28.07+43.42+150	268069	160532.6+145920	132 150 227 214	29(4)	10 × 5	1.15	11	0	10	g, G10
HVC 28.47+43.13+177	268070	160707.0+150831	160 177 255 243	20(3)	17 × 9	1.48	11	2	6	g, G10
HVC 28.03+41.54+127	268071	161236.8+141226	109 127 206 194	62(15)	12 × 7	2.67	18	1	8	g
HVC 28.66+40.38+125	268072	161745.3+141036	108 125 208 197	42(5)	16 × 9	3.17	21	3	7	g
HVC 19.13+35.24-123	268213	162235.7+050848	−139 −123 −63 −81	17(1)	12 × 10	2.83	22	0	7	g, G10, S12
HVC 27.86+38.25+124★	268074	162443.4+124412	107 124 207 197	23(4)	11 × 9	1.28	13	2	4	g
HVC 84.01-17.95-311	310851	215406.2+311249	−324 −311 −98 −21	21(4)	26 × 14	2.60	17	0	5	g
HVC 82.91-20.46-426	310865	215802.9+283735	−439 −426 −217 −140	22(1)	12 × 6	0.99	10	0	17	g, S12
HVC 80.69-23.84-334	321318	220100.7+244404	−345 −334 −131 −55	23(1)	18 × 9	1.47	13	0	5	g
HVC 86.18-21.32-277	321455	221121.8+295402	−288 −277 −68 10	17(1)	13 × 7	1.76	15	0	5	g, O
HVC 82.91-25.55-291	321320	221238.6+244311	−302 −291 −90 −13	24(2)	15 × 6	1.31	13	0	7	g, O
HVC 84.61-26.89-330	321351	222134.4+243638	−341 −330 −130 −53	21(4)	13 × 11	1.03	9	0	8	g, O
HVC 92.53-23.02-311	321457	223823.4+315257	−321 −311 −104 −23	28(2)	19 × 9	1.68	12	0	5	g, O
HVC 87.35-39.78-454a	334256	230056.4+152014	−461 −454 −282 −206	26(4)	11 × 8	1.57	16	0	1	g, O
HVC 88.15-39.37-445a	334257	230211.3+160048	−452 −445 −271 −195	22(11)	12 × 4	0.68	10	0	4	g, O
HVC108.98-31.85-328	333613	235658.8+293235	−333 −328 −147 −64	19(2)	13 × 5	0.55	8	1	19	g, O
HVC109.07-31.59-324	333494	235702.1+294846	−329 −324 −143 −60	17(5)	12 × 7	1.80	23	1	19	g, O

Notes. ^★Part of the extremely isolated MIS subsample. ^aNot included in the α.40 catalog. ^bAlso included in the compact cloud catalog of Saul et al. (2012). ^cPossible kinematic association with larger structure.

Download table as:  ASCIITypeset image

To be included as a UCHVC in the catalog, a source must have |v_LSR| > 120 km s⁻¹, have an H i major axis less than 30' in size, and have an S/N ⩾ 8 to ensure reliability. The v_LSR limit is imposed to focus on a class of clouds that are well separated from Galactic emission and that could trace dark matter halos within the LG. Some dark matter halos would be expected to have |v_LSR| < 120 km s⁻¹ (Leo T, for example) but disentangling their emission from Galactic hydrogen is challenging and left to future work. The 30' size limit corresponds to a physical size of 2 kpc at a distance of 250 kpc. The distance of 250 kpc is a reasonable minimum distance for an unperturbed object at the edge of the MW; Grcevich & Putman (2009) find LG dwarf galaxies with neutral gas content 270 kpc away from either the MW or M31.² We would not expect to detect low mass galaxies with large gas reservoirs nearer to the MW due to interaction with the hot Galactic corona (Fukugita & Peebles 2006). Note that Leo T has an H i diameter of 0.6 kpc, and Leo P has an H i diameter of 1.2 kpc. The models of Sternberg et al. (2002) for gas in dark matter minihalos predict H i diameters up to 3 kpc with diameters less than 2 kpc a more common outcome. It should be noted that most UCHVCs are smaller than this criterion, with only six clouds having average H i diameters larger than 16' (see Figure 3). The S/N limit of 8 ensures reliability. This limit is higher than the general reliability limit of the ALFALFA survey data due to the different nature of the UCHVCs, including the strong potential for radio frequency interference (RFI) to masquerade as narrow-line sources. Confirmation observations of low S/N sources are ongoing, and in future work we will examine the reliability and completeness of the ALFALFA UCHVC catalog.

Given the abundance of HVC structures in the sky, the most important criterion for determining if a cloud is a good minihalo candidate is its isolation. Most of the known HVC structure is associated with Galactic processes, including accretion onto the MW; when considering clouds that could represent gas associated with dark matter halos, we wish to find objects distinct from existing HVC structures. In order to be considered a UCHVC, visual inspection must ensure that the cloud does not appear to be associated with a larger H i structure.

Our second isolation criterion is that the UCHVCs must be well separated from previously known HVC complexes. We compare the UCHVCs to the updated catalog of Wakker & van Woerden (1991; B. Wakker 2012, private communication, hereafter WvW). The WvW catalog includes 617 clouds, of which 393 are classified as belonging to 20 large complexes; the other clouds are classified into populations based on their spatial coordinates and velocity. For defining isolation, we only consider the WvW clouds which are part of a larger complex. The distance of a UCHVC from another cloud in degrees can be quantified via:

$$\begin{equation} D = \sqrt{ \theta ^2 + (f \delta v)^2}, \end{equation} \tag{ 1 }$$

where θ is the angular separation in degrees, δv is the velocity difference in km s⁻¹ between two clouds, and f is a conversion factor that parameterizes the significance we ascribe to the angular separation between two clouds versus their difference in velocity in determining whether they are associated with each other. Following Saul et al. (2012) and Peek et al. (2008), we adopt f = 05/km s⁻¹ as the weighting for the velocity separation for large scale HVC structure. Figure 4 illustrates our determination of the isolation criterion for deciding if the UCHVCs are separated from the WvW complexes. The isolation criterion was determined by comparing the separation of clouds within WvW complexes to the separation of LG galaxies from the nearest WvW cloud in a complex. The x-axis shows the distance to the nearest WvW cloud in a complex and the y-axis shows the fraction of objects whose closest neighbor is at that distance or closer (cumulative fraction). Ninety percent of WvW clouds in complexes are closer than 15° to their nearest neighbor in the complex; more than eighty percent of LG galaxies are located farther than 15° from the nearest WvW cloud in a complex. Hence we determine to use this value as our cutoff, shown by the dot-dashed line in Figure 4. We note that is a more generous criterion than that of Saul et al. (2012) and Peek et al. (2008) who adopt D = 25° as an isolation criterion; in Section 4.1 we examine this intermediate distance and determine it does not substantially affect our catalog.

Zoom In Zoom Out Reset image size

In addition, we institute a third isolation criterion based on HVC structure uncovered by ALFALFA. This structure is generally much smaller than previously known HVC structures; as can be seen in Figure 2 most α.40 HVCs are less than 1 deg in size while the sizes of the HVCs in the WvW catalog are several to tens of degrees.³ For this reason, we use f = 02/km s⁻¹ in Equation (1) when calculating isolation from HVC structure within the ALFALFA survey. The top panel of Figure 5 shows the final isolation criterion for UCHVCs and compares the UCHVCs to LG galaxies and the general HVC detections within the α.40 survey. We require that the UCHVCs have no more than three neighbors within D = 3°. This is a generous criterion as the LG galaxies have at most one neighbor within this distance. We wish to include all potential minihalo candidates and inspection indicates that allowing three neighbors includes all the sources that would be classified by eye as isolated. In the bottom panel of Figure 5 we explore the differences between the spring and fall populations of the UCHVCs. The fall sky appears to show more isolation on this scale with the UCHVCs having either one or no neighbors; in fact, this is a result of the prominent HVC structure in the fall sky. Clouds in the fall sky are either part of a larger structure or have no (or one) neighbors within D = 3°. Comparing to the general α.40 HVC population shows the prevalence of HVC structure in the fall sky with the fall HVCs generally having more neighbors than the spring HVCs.

Zoom In Zoom Out Reset image size

We note that with this criterion, only clouds with central velocities within 15 km s⁻¹ of the UCHVC can be considered as neighbors. Given that the median velocity width of the UCHVCs is 23 km s⁻¹, there is a possibility that this isolation criterion could leave our sources kinematically confused. Our first isolation criterion accounts for this through the examination of the UCHVCs for association with other clouds. In order to verify this, we examine the effect of changing the velocity weighting factor to f = 005/km s⁻¹. This expands the velocity selection to 60 km s⁻¹, almost three times the median FWHM of the clouds. We examine the number of clouds within 3° of the UCHVCs using this different value of f and find that the UCHVCs still have very few neighbors with this modified distance estimate. In fact, seventy-five percent of the UCHVCs still meet the criterion of three or fewer neighbors even when the expanded velocity space is considered. We examined the nine UCHVCs with more than five neighbors and note that three of them may possibly be kinematically associated with larger structure; we mark these UCHVCs in Table 1.

HVC structure often exists on scales much larger than 3°; while the UCHVCs are examined for obvious connection to larger structure and excluded in that case, we still wish to define a more isolated subsample. As the best subsample to represent H i sources associated with minihalo candidates, we define a "most isolated" subsample (MIS) of UCHVCs with no more than four neighbors within D = 10°. The top panel of Figure 6 shows the number of neighboring clouds within D = 10° for the UCHVCs, the MIS UCHVCs, LG galaxies, and α.40 HVCs. On this large scale, the MIS UCHVCs are generally more isolated than even the LG galaxies. We do note that the α.40 footprint means that we are not generally probing to a full 10° in all directions around a given cloud; increasing coverage of the ALFALFA survey may change the classification of a cloud in the future. In fact, two sources in the fall δ = +15° strip meet the MIS criteria but we exclude them from this subsample as determining isolation out to 10° for sources in an isolated 2° wide strip is problematic. We will revisit these two specific sources and the classification of the MIS UCHVCs in general with increased ALFALFA coverage in future work. In the bottom panel of Figure 6, we again examine the difference between the fall and spring populations. Here, the prominent HVC structure in the fall sky is apparent with many of the fall UCHVCs having a large number of neighbors out to a distance of 10°. There is also a strong difference evident between the UCHVC and α.40 HVC population with over half of the α.40 HVCs having more than 20 neighbors at D = 10°; this indicates the utility of our first isolation criterion of inspecting sources for connection to large scale structure.

Zoom In Zoom Out Reset image size

In Table 1 we present the UCHVCs; there are 59 sources total: 40 in the spring α.40 sky and 19 in the fall sky. Of the 59 UCHVCs, 17 are identified as being in the most isolated subsample, all of which are in the spring sky. The spring sky samples the outer regions of the LG where the expected density for dark matter halos may be lower but the environment is safer for gas-bearing minihalos than near the MW or M31. The fall sky samples the LG near M31 and includes the presence of a large amount of HVC structure, including the Magellanic Stream (MS; see Section 4.1 for a further discussion). We indicate those UCHVCs that are part of the original sample of UCHVCs discussed by G10 with a "G10" in the notes column and those UCHVCs that lie outside the area considered by G10 with an "O." Figure 7 shows maps of all the UCHVCs with contours in units of column density of H i (N_H i in atoms cm⁻²), representing the sum total of H i content along the line of sight; these plots represent the data from which all the parameters listed in Table 1 are derived. The minimum contour level is given in the figure and subsequent contour levels increase by factors of $\sqrt{2}$. We plot the contours in values of N_H i to demonstrate that the peak column density value is higher than the average value calculated later (see Section 3.4). However, we emphasize that since these clouds are barely resolved by the Arecibo beam, the column density contour values are only approximate and the average values are more robust; to accurately map the distribution of H i will require synthesis observations that provide a smaller beam. Column density values can be derived from the brightness temperature via:

$$\begin{equation} N_{{\rm H\,\scriptsize{I}}} = 1.823 \times 10^{18} \int T_B\, dv \, \, [{\rm cm}^{-2}]. \end{equation} \tag{ 2 }$$

In simple cases, the brightness temperature is related to the flux density at 21 cm via:

$$\begin{equation} T_B = \frac{606}{\theta ^2} S \end{equation} \tag{ 3 }$$

where θ is the (circular) beam in arcseconds and S the flux in mJy beam⁻¹.

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

The columns of the tables are as follows:

• 1.  
  Column 1: source name, in the traditional form for HVCs, obtained from the galactic coordinates at the nominal cloud center and the v_LSR of the cloud, e.g., HVC111.65-30.53-124 has l = 11165, b = −3053, and v_LSR= −124 km s⁻¹.
• 2.  
  Column 2: identification number in the Arecibo General Catalog (AGC), an internal database maintained by M.H. and R.G., included to ease cross–reference with our archival system and the α.40 catalog.
• 3.  
  Column 3: equatorial coordinates of the centroid, epoch J2000. Typical errors are less than 1'.
• 4.  
  Column 4: sequentially, we list heliocentric velocity, velocity in the local standard of rest frame (LSR; assumed solar motion of 20 km s⁻¹ toward l = 57°, b = 25°), velocity in the Galactic standard of rest frame (GSR; V_gsr = V_lsr + 225sin lcos b, with both velocities in km s⁻¹), and the velocity with respect to the LG reference frame from Karachentsev & Makarov (1996).
• 5.  
  Column 5: H i line FWHM (W₅₀), with estimated measurement error in brackets. The notes column indicates the method of measurement: a Gaussian fit or linear single peaks fit to the sides of the profile.
• 6.  
  Column 6: estimate of the cloud major and minor diameters, in arcminutes. Sizes are measured at approximately the level encircling half the total flux density. In many cases, the outer contours are more elongated than indicated by the ratio a × b. The half-power ellipses are also shown in the H i column density contour plots in Figure 7.
• 7.  
  Column 7: flux density integral (S₂₁), in Jy km s⁻¹.
• 8.  
  Column 8: signal–to-noise ratio (S/N) of the line, defined as

  $$\begin{equation} {\rm S/N} = \left(\frac{1000 S_{21}}{W_{50}} \right) \frac{w_{{\rm smo}}^{1/2}}{\sigma _{{\rm rms}}}, \end{equation} \tag{ 4 }$$

  where S₂₁ is the integrated flux density in Jy km s⁻¹, as listed in Column 7; the ratio 1000S₂₁/W₅₀ is the mean flux density across the feature in mJy; w_smo is W₅₀/(2 × 10), a smoothing width; and σ_rms is the rms noise figure across the spectrum measured in mJy. More details on the S/N calculation are available in Haynes et al. (2011).
• 9.  
  Column 9: the number of α.40 HVC neighbors within D = 3° (for f = 02/km s⁻¹)
• 10.  
  Column 10: the number of α.40 HVC neighbors within D = 10° (for f = 02/km s⁻¹)
• 11.  
  Column 11: notes column. For each source there is either a "g" or "p" indicating the method used (Gaussian or single peaks fit) to measure W₅₀. Sources considered by G10 are indicated with a "G10" in the notes column. Sources that are outside the footprint considered in G10 are marked with a "O." The UCHVCs that are also in the GALFA compact cloud catalog of Saul et al. (2012) are indicated with a "S12."

For completeness, we include in Table 2 the UCHVCs that were considered by G10 but do not meet the stricter selection criteria used here. The clouds from G10 can fail any of the criteria: S/N, isolation or v_LSR limits. The notes column indicates the reason a G10 cloud is not included here. The sources with S/N < 8 will be considered in future work when we extend the UCHVC catalog to lower S/N values after assessing reliability and completeness. In addition, we will extend the catalog to velocities including the Galactic hydrogen. It should be noted that the three sources that do not meet the isolation criteria only barely fail. Two sources have one and two more neighbors than allowed, respectively, and the third sources is excluded based on examination of large scale structure. These sources could still be good minihalo candidates.

Table 2. UCHVCs from G10 that Fail UCHVC Criteria

Source	AGC	R.A.+ decl.	cz☉ VlsrVgsr VLG	W50(w)	a × b	S21	S/N	N3	N10	Reason
		(J2000)	(km s−1)	(km s−1)	(')	(Jy km s−1)
HVC244.51+53.41+160	208424	104850.1+050419	164 160 39 −18	19(3)	16 × 12	1.03	7	0	9	S/N
HVC249.03+57.58+178	219213	110813.6+055725	179 176 64 5	19(2)	12 × 9	0.67	7	0	8	S/N
HVC247.19+70.29+247	215418	114418.2+150509	246 247 177 129	30(10)	10 × 8	0.54	7	0	1	S/N
HVC290.37+66.23-115	227983	123116.7+035044	−118 −114 −199 −259	20(5)	6 × 4	0.44	9	0	3	Velocity
HVC298.30+72.91+185	226171	124557.2+100518	180 185 127 75	25(3)	5 × 4	0.57	9	5	21	Isolation
HVC299.62+67.65+326	227988	124619.1+044923	323 327 253 195	39(13)	14 × 7	0.76	6	0	0	S/N
HVC314.57+74.80+218	238626	130351.1+121223	211 218 176 127	36(13)	5 × 3	0.35	5	0	17	S/N
HVC 8.88+62.16+281	249538	143531.7+133126	269 282 298 264	18(6)	4 × 3	0.22	4	0	4	S/N
HVC 7.64+57.83-128	249248	144844.6+103510	−142 −128 −112 −147	22(1)	25 × 5	1.83	16	0	42	Isolation
HVC 15.11+45.54-148	258474	154035.2+074334	−163 −147 −106 −132	27(1)	7 × 5	0.68	9	4	19	Isolation

Download table as:  ASCIITypeset image

Figure 8 shows the distribution of measured properties for the α.40 UCHVCs and the most isolated subsample: integrated flux density (S₂₁), average angular diameter ($\bar{a} = \sqrt{a b}$), velocity FWHM (W₅₀), and v_LSR. The UCHVCs have integrated flux densities of ∼0.66–8.55 Jy km s⁻¹, with the vast majority having integrated flux densities below 3.5 Jy km s⁻¹ and a median flux density of 1.34 Jy km s⁻¹. The singly hatched histograms are the UCHVCs in the most isolated subsample. Note that the range of values for the MIS UCHVCs is similar to the larger UCHVC population, and the median values are essentially identical. The UCHVCs range in average diameter from essentially unresolved (∼4') to just over 20' in size, with the vast majority less than 16' in size and a median size of 10'. We note that there does appear to be a break in population based on size with UCHVCs clustered with H i diameters <16' in size and a tail of a population extending to larger sizes (including objects with H i diameters >30' not included in this work). We will explore this break in H i size in the HVC population in future work with a larger survey area. The W₅₀ values are centered around 15–30 km s⁻¹ with a few UCHVCs having widths extending up to 70 km s⁻¹; the median linewidth is 23 km s⁻¹. There are clouds whose velocities cluster near both v_LSR ±120 km s⁻¹, with a much stronger clustering of positive velocity clouds. However, when the MIS UCHVCs are considered, this clustering disappears. The vast majority of negative velocity clouds are also excluded from the MIS UCHVCs; the negative velocity clouds are predominantly in the fall sky, where large scale H i structure is much more prevalent, preventing the inclusion of any UCHVCs into the most isolated subsample.

Zoom In Zoom Out Reset image size

Given the observed properties of the UCHVCs, integrated flux density (S₂₁, Jy km s⁻¹), average angular diameter ($\bar{a} = \sqrt{a b}$, arcminutes) and velocity width (W₅₀, km s⁻¹), it is straightforward to derive some simple properties of the UCHVCs, modulo the unknown distance d (in Mpc), with the assumption that the clouds are optically thin. Sequentially, below we derive the mean atomic density, mean column density, H i mass, indicative dynamical mass within the H i extent, and H i diameter.

$$\begin{eqnarray} \bar{n}_{{\rm H\,\scriptsize{I}}} [\mathrm{atoms\ cm}^{-2}] & = & 0.74 S_{21} \bar{a}^{-3} d^{-1} \,\,\,\,{\rm cm}^{-3} \end{eqnarray} \tag{ 5 }$$

$$\begin{eqnarray} \; \;\, \bar{N}_{{\rm H\,\scriptsize{I}}} [\mathrm{atoms\ cm}^{-2}]& = & 4.4\times 10^{20} \bar{a}^{-2} S_{21} \,\,\,\,{\rm cm}^{-2} \end{eqnarray} \tag{ 6 }$$

$$\begin{eqnarray} \;\;\;\; M_{{\rm H\,\scriptsize{I}}} [M_\odot ] & = & 2.356\times 10^5 S_{21} d^2 \end{eqnarray} \tag{ 7 }$$

$$\begin{eqnarray} \; M_{{\rm dyn}} [M_\odot ] & = & 6.2\times 10^3 \bar{a} W_{50}^2 d \end{eqnarray} \tag{ 8 }$$

$$\begin{eqnarray} \;\, D_{{\rm H\,\scriptsize{I}}} [\mathrm{kpc}] & = & 0.29 \bar{a} d \qquad\quad \end{eqnarray} \tag{ 9 }$$

Of these derived properties, $\bar{N}_{{\rm H\,\scriptsize{I}}}$ is especially noteworthy as it does not depend on the distance. It should be noted that the column density values derived here are average values based on the global properties of the UCHVCs, in contrast to the approximation of spatially-resolved column density contours in Figure 7. Due to the large beam size of Arecibo, these values represent underestimates of the peak values of the clouds. We note that the dynamical mass is an indicative dynamical mass only. In addition to the uncertainty in the distance of the UCHVCs, the contribution to the linewidths of the UCHVCs from thermal broadening is unknown. For a range of reasonable temperatures, the thermal broadening can range from 16 to 21 km s⁻¹. For the clouds with the largest linewidths, the thermal broadening contribution (when accounted for in quadrature) may be negligible, while the narrowest clouds may be fully thermally supported. However, they could still have large-scale motions on the order of the thermal broadening, or less. For example, Leo P has a linewidth of 24 km s⁻¹ and a rotational velocity of 9 km s⁻¹, uncorrected for disk inclination (Giovanelli et al. 2013). To derive accurate dynamical masses will require higher resolution H i images in which evidence of large scale motions can be discerned (and, of course, distance information).

Table 3. Inferred Cloud Properties

Source	AGC	DH i	nH i	$\log \bar{N}_{{\rm H\,\scriptsize{I}}}$	log MH i	log Mdyn
		(kpc d)	(cm−3d−1)	(cm−2)	(M☉d2)	(M☉d)
HVC111.65-30.53-124	103417	5.8	−3.68	18.40	5.74	7.74
HVC123.11-33.67-176	102992	4.6	−3.62	18.35	5.48	7.64
HVC123.74-33.47-289	102994	1.6	−2.54	18.98	5.20	7.18
HVC126.85-46.66-310	749141	2.7	−3.12	18.62	5.28	7.48
HVC131.90-46.50-276	114574	2.2	−2.95	18.72	5.22	7.54
HVC137.90-31.73-327	114116	6.2	−3.53	18.58	5.97	8.19
HVC138.39-32.71-320	114117	4.5	−3.07	18.90	6.02	7.67
HVC154.00-29.03-141	122836	6.0	−3.25	18.85	6.21	7.97
HVC205.28+18.70+150★	174540	2.2	−2.46	19.19	5.69	7.40
HVC196.50+24.42+146	174763	3.8	−3.02	18.87	5.82	7.51
HVC196.09+24.74+166	174764	2.2	−2.94	18.71	5.19	7.43
HVC198.48+31.09+165	189054	4.6	−3.49	18.49	5.62	7.82
HVC204.88+44.86+147★	198511	2.0	−2.81	18.81	5.24	6.99
HVC234.33+51.28+143	208315	3.6	−2.72	19.15	6.07	7.49
HVC250.16+57.45+139	219214	1.6	−2.58	18.93	5.12	7.13
HVC252.98+60.17+142	219274	5.8	−3.11	18.97	6.30	7.96
HVC253.04+61.98+148	219276	3.7	−3.14	18.74	5.69	8.01
HVC255.76+61.49+181	219278	2.4	−2.91	18.78	5.33	7.21
HVC256.34+61.37+166	219279	3.2	−3.10	18.72	5.55	7.60
HVC245.26+69.53+217★	215417	2.8	−3.24	18.52	5.22	7.24
HVC277.25+65.14-140★	227977	1.5	−2.64	18.86	5.03	7.24
HVC274.68+74.70-123★	226067	1.3	−2.14	19.29	5.34	7.91
HVC290.19+70.86+204	226165	2.2	−2.83	18.83	5.33	7.32
HVC292.94+70.42+159	229344	4.4	−3.46	18.50	5.59	7.33
HVC295.19+72.63+225	226170	3.8	−3.41	18.48	5.44	7.80
HVC298.95+68.17+270★	227987	3.5	−2.62	19.23	6.12	7.70
HVC324.03+75.51+135	233763	1.8	−2.51	19.05	5.35	7.50
HVC320.95+72.32+185	233830	5.3	−3.69	18.35	5.60	7.78
HVC330.13+73.07+132	233831	1.2	−2.23	19.17	5.17	6.83
HVC326.91+65.25+316★	238713	3.1	−3.12	18.68	5.47	7.65
HVC 28.09+71.86-144★	249393	3.3	−3.25	18.58	5.42	8.11
HVC353.41+61.07+257★	249323	3.2	−3.12	18.69	5.50	7.43
HVC351.17+58.56+214★	249282	1.7	−2.29	19.26	5.53	7.77
HVC352.45+59.06+263★	249283	3.9	−3.46	18.44	5.42	7.92
HVC356.81+58.51+148★	249326	1.6	−2.54	18.99	5.22	7.70
HVC 5.58+52.07+163★	258459	3.0	−3.05	18.74	5.50	7.57
HVC 13.59+54.52+169★	258237	2.0	−2.55	19.08	5.50	7.36
HVC 13.60+54.23+179★	258241	2.9	−3.12	18.65	5.37	7.25
HVC 13.63+53.78+222★	258242	2.1	−2.84	18.79	5.22	7.29
HVC 26.11+45.88+163	257994	2.7	−2.72	19.02	5.68	7.48
HVC 26.01+45.52+161	257956	1.9	−2.40	19.19	5.56	7.41
HVC 29.55+43.88+175	268067	2.2	−2.51	19.15	5.65	7.81
HVC 28.07+43.42+150	268069	2.1	−2.62	19.00	5.43	7.57
HVC 28.47+43.13+177	268070	3.5	−3.22	18.64	5.54	7.48
HVC 28.03+41.54+127	268071	2.7	−2.62	19.13	5.80	8.35
HVC 28.66+40.38+125	268072	3.4	−2.85	19.00	5.87	8.11
HVC 19.13+35.24-123	268213	3.1	−2.77	19.04	5.82	7.28
HVC 27.86+38.25+124★	268074	2.8	−3.00	18.77	5.48	7.51
HVC 84.01-17.95-311	310851	5.4	−3.54	18.51	5.79	7.71
HVC 82.91-20.46-426	310865	2.4	−2.91	18.79	5.37	7.40
HVC 80.69-23.84-334	321318	3.7	−3.27	18.61	5.54	7.62
HVC 86.18-21.32-277	321455	2.8	−2.82	18.93	5.62	7.23
HVC 82.91-25.55-291	321320	2.7	−2.93	18.81	5.49	7.52
HVC 84.61-26.89-330	321351	3.5	−3.37	18.49	5.39	7.52
HVC 92.53-23.02-311	321457	3.8	−3.27	18.63	5.60	7.81
HVC 87.35-39.78-454	334256	2.7	−2.85	18.89	5.57	7.59
HVC 88.15-39.37-445	334257	2.0	−2.80	18.81	5.20	7.31
HVC108.98-31.85-328	333613	2.2	−3.02	18.63	5.11	7.23
HVC109.07-31.59-324	333494	2.7	−2.77	18.97	5.63	7.22

Note. ^★Part of the extremely isolated MIS subsample.

Download table as:  ASCIITypeset image

In Table 3, we summarize the inferred properties of the UCHVCs. The columns of the table are as follows:

• 1.  
  Columns 1 and 2: source ID as in Table 1.
• 2.  
  Column 3: H i diameter in kpc at d = 1 Mpc (Equation (9)).
• 3.  
  Column 4: log of the mean atomic H i density at d = 1 Mpc, in cm⁻³ (Equation (5)).
• 4.  
  Column 5: log of the mean H i column density, in cm⁻² (Equation (6)).
• 5.  
  Column 6: log of the H i mass at d = 1 Mpc, in solar units (Equation (7)).
• 6.  
  Column 7: log of the indicative dynamical mass within D_H i at d = 1 Mpc, in solar units (Equation (8)).

The H i masses, dynamical masses, mean atomic densities and mean column densities of the UCHVCs and the MIS UCHVCs are shown in Figure 9. At a distance of 1 Mpc, the H i masses are around ∼10⁵–10⁶ M_☉ and the dynamical masses are ∼10⁷–10⁸ M_☉. This would require the UCHVCs to have an ionized envelope of hydrogen or a substantial amount of dark matter in order to be self-gravitating. As discussed in Section 5, these median properties are a good match to the minihalo models of Sternberg et al. (2002). The median dynamical mass is 10^7.5 d_Mpc M_☉; this is close to the common mass scale of ∼10⁷M_☉ for the UFDs of Strigari et al. (2008).

Zoom In Zoom Out Reset image size

The HVC sky contains many large extended structures composed of multiple clouds. We explicitly require the UCHVCs to be isolated from the known large scale HVC structure of the WvW catalog. However, our isolation criterion for separation from WvW complexes is slightly relaxed in order to avoid excluding potential minihalo candidates. As can be seen in Figure 4, the distance to the nearest cloud within a WvW complex can extend to D = 25°. As we set our isolation criterion for UCHVCs to a separation of 15° from WvW clouds in complexes, we wish here to consider the possible association of the UCHVCs with WvW complexes. In Table 4 we list the UCHVCs that are less than 25° from a WvW complex. We note that only two UCHVCs in the fall sky (HVC86.18-21.32-277 and HVC87.35-39.78-454) are more than 25° from a complex in the WvW catalog; the other fall HVCs not listed in Table 4 are separated by less than 25° from clouds associated with the MS in the WvW catalog. Of the 40 spring UCHVCs, 7 are potentially associated with known large complexes, the majority of those being with the WA complex. While a few of the UCHVCs may be associated with known large complexes, the vast majority are not, as defined by our isolation criterion.

4.1.1. Magellanic Stream

The MS is an extended H i structure first noted by Dieter (1965) and first associated with the Magellanic Clouds by Mathewson et al. (1974). The MS is generally associated with the disruption of the Magellanic Clouds as they interact with the MW, although the exact mechanisms responsible for the MS are an open area of research. The two main parts of the MS are the Leading Arm (LA), which consists of gas ahead of the Large Magellanic Cloud (LMC) and Small Magellanic Cloud (SMC) in their presumed orbits, and the tail, which consists of the trailing material. Recently, Nidever et al. (2010, hereafter N10) presented an extension of the MS, bringing it to over a 200° length in total. Given the extent of the MS, possible association with the MS must be considered when attempting to understand HVCs of any sort.

For the α.40 footprint, the fall sky overlaps the tail of the MS and the spring sky is near the known edge of the LA but not contiguous to it. N10 extended the known tail of the MS and pointed out its complexity (see their Figure 4), so we must be especially careful with UCHVCs in the fall sky. In Figure 10, we show the UCHVCs plotted on the 200° MS presented in N10. The coordinates are the MS-coordinate system of Nidever et al. (2008) based on fitting a great circle to the MS, where L_MS is the longitude along the MS and B_MS is the latitude above/below the MS. The UCHVCs are shown as large symbols (red in the online version) to increase their visibility; they are neither shown to physical scale nor do their colors match the shading of the MS. The top panel shows the H i column density of the MS (log N_H i in cm⁻²). The bottom panel is the total intensity of the MS integrated along B_MS (K deg).

Zoom In Zoom Out Reset image size

In the spring sky, the α.40 footprint approaches but does not overlap the LA of the MS. This lack of direct coverage of the MS makes it a challenge to answer the question: could the UCHVCs be connected to the LA? Future surveys directed at determining any possible continuation of the LA will be able to directly answer this question. Until then, the key to answering this question is determining whether the UCHVCs have compatible velocities to be an extension of the LA. Clearly, the large velocity spread of UCHVCs seen in the bottom panel of Figure 10 appears to be incompatible with all of the UCHVCs being associated with the LA. Examining models of the MS can provide insight into these questions. Connors et al. (2006) model the MS as a tidal structure via interaction with the MW and LMC; they predict that the LA extends to L_MS ∼150° with a velocity turnover starting from L_MS ∼ 60° at v_LSR ∼ 300 km s⁻¹ extending to ∼ − 150 km s⁻¹. In contrast, Besla et al. (2010) simulate a first passage of the Magellanic Clouds and find an MS that extends to L_MS ∼ 50° with a velocity increasing with L_MS from v_LSR ∼ 200 to 400 km s⁻¹. If the Connors et al. (2006) model correctly represents the history of the MS, then the clouds located at v_LSR < 0 km s⁻¹ could be associated with the LA of the MS. If the Besla et al. (2010) model is accurate, then the UCHVCs are generally at higher L_MS values than predicted by the model but a few of the positive velocity clouds with L_MS < 100° and the highest v_LSR values may be associated with the MS. For whichever model of the MS is chosen, some of the UCHVCs could be associated with the LA, but given the large spread in v_LSR of the UCHVCs, it is impossible to associate all of the UCHVCs with the LA.

In the fall sky, the α.40 footprint overlaps the extension of the MS detailed in N10. In Figure 11 we offer a zoomed in view focusing on the fall UCHVCs compared to the MS from N10. Here, there clearly appears to be strong overlap between the UCHVCs and the known MS system. The three clouds in the fall sky at v_LSR > −200 km s⁻¹ appear to be kinematically separated from the MS. Two other clouds at L_MS ∼ −100° appear to potentially be spatially separated from the MS but the apparent separation could easily be a result of the coverage of observations of the MS. However, it is still possible that some of these UCHVCs do indeed represent galaxies. Many of the UCHVCs that overlap with the MS are also in the direction of the M31 subgroup. Disentangling the gas of known galaxies at a similar velocity from the MS is a long standing problem; see Grcevich & Putman (2009) for illustrative examples. This is also illustrated in Figure 11, where several LG galaxies are spatially and kinematically coincident with the MS.

Zoom In Zoom Out Reset image size

Previous studies have uncovered a population of compact clouds associated with the Galactic halo (e.g., Lockman 2002; Lockman & Pidopryhora 2005; Stil et al. 2006; Stanimirović et al. 2006; Ford et al. 2010; Dedes & Kalberla 2010). While well separated from the Galactic hydrogen, these clouds typically have low v_LSR values, and they generally appear to be consistent with Galactic rotation. The Galactic halo clouds with the most extreme velocities of Stil et al. (2006) have v_LSR ranging from ≲100 km s⁻¹ to 165 km s⁻¹. The compact halo clouds also tend to be cold clouds, with the vast majority of reported clouds having W₅₀ < 10 km s⁻¹. Given these characteristics of the halo clouds, the UCHVCs appear as a distinct population. The UCHVCs appear to universally be warm clouds with linewidths greater than 15 km s⁻¹. In addition, many of the UCHVCs have substantial velocities (|v_LSR| > 200 km s⁻¹) that are difficult to account for in a Galactic halo model.

GALFA-HI is a survey of neutral hydrogen in the Galaxy which, like ALFALFA, uses the ALFA multi beam receiver on the Arecibo 305 m antenna. For GALFA-HI, the IF signal is sent to a different spectrometer than that used by ALFALFA and is restricted to a ∼7 MHz bandpass centered on 1420 MHz. As a result, the GALFA-HI survey has a velocity resolution of 0.184 km s⁻¹ and covers a velocity range of ±700 km s⁻¹. It should be noted that much of the GALFA data is taken commensally with the ALFALFA data through the TOGS program. Hence, comparison of the results of the two surveys provides a check on our signal processing approach. Begum et al. (2010) presented an initial catalog of compact clouds from the GALFA-HI survey, and Saul et al. (2012, hereafter S12) recently released a catalog of compact clouds for the full initial data release of the GALFA-HI survey. Herein we focus on the compact clouds of S12 as the most extensive catalog of the compact cloud population discovered in the GALFA-HI survey and examine how the UCHVCs of this work are related.

The initial major differences to note between the catalog of S12 and the UCHVCs are additional selection criteria for the UCHVCs: the limited range of velocities considered and the strong isolation criteria. A vast majority of the compact clouds from S12 do not meet these additional criteria. S12 note several populations of clouds in their catalog which they classify by velocity, linewidth and isolation. They split between warm and cold clouds at a linewidth of 15 km s⁻¹, or at a temperature of ∼5000 K. It should be noted that while ALFALFA does not have the velocity resolution of the GALFA-HI survey, the velocity resolution of ∼10 km s⁻¹ is sufficient to distinguish warm from cold clouds; as can be seen in Figure 8, the UCHVCs are all warm clouds with linewidths greater than 15 km s⁻¹. S12 also split their clouds into low velocity and high velocity populations at |v_LSR| = 90 km s⁻¹. They find a few cold clouds with v_LSR > 90 km s⁻¹, but the vast majority of their cold clouds are at lower velocities and associated with the Galactic disk, a very distinct population from the ALFALFA UCHVCs. The populations from S12 of most relevance to this work are their HVC population (|v_LSR| > 90 km s⁻¹) and galaxy candidate population; both of these populations are generally composed of warm clouds. The difference between the HVC population and galaxy candidate population of S12 is that the galaxy candidates have an additional stringent isolation criterion (different from the isolation criteria used here) and hence are the population most directly comparable to the UCHVCs. In Figure 12, we compare the distribution of the UCHVCs to the compact clouds of S12 in galactic longitude versus v_LSR. In the second Galactic quadrant, the UCHVCs overlap with the HVCs of S12. This corresponds to the fall sky, and, as noted in the previous section, when considering a stricter isolation criterion for separation from larger HVC complexes akin to that used by S12, the fall UCHVCs cannot be considered isolated structures. In the first and fourth Galactic quadrants, the UCHVCs as a population appear separated from the compact clouds of S12. The positive velocity clouds in the first quadrant and the clouds (at both positive and negative velocities) in the fourth quadrant have no HVC population counterpart in the GALFA compact cloud catalog. Especially in the fourth quadrant, there are multiple clouds at substantial velocities (v_LSR > 200 km s⁻¹) that appear well separated from other clouds populations.

Zoom In Zoom Out Reset image size

As a check of our methodology and dataset, we also perform a direct comparison of the ALFALFA UCHVCs to the catalog of S12. First, we examine which of the S12 galaxy candidates appear in the α.40 catalog. S12 find 28 HVCs that they consider extremely isolated and which they classify as galaxy candidates. Of these, 10 are within the α.40 footprint. Two of the GALFA galaxy candidates are classified as extragalactic sources in α.40 (AGC191803 and AGC227874) and are clearly associated with optical counterparts; a third S12 galaxy candidate is associated with UGC 7753, a large barred spiral galaxy. Four of their galaxy candidates are within the ALFALFA data but have |v_LSR| < 120 km s⁻¹ and are not included in this work (one is included in the α.40 catalog, AGC238801). One of the galaxy candidates is also included here in the UCHVC catalog—HVC351.17+58.56+214. Two of the S12 galaxy candidates are not seen in the ALFALFA data; these are both lower S/N sources (S/N < 7) and one has extremely narrow velocity width (W₅₀ = 3.9 km s⁻¹).

Secondly, we can examine the UCHVCs for counterparts in the S12 catalog. 11 of the 59 UCHVCs are included in the GALFA compact cloud catalog, of which one (HVC351.17+58.56+214) is classified by S12 as a galaxy candidate; the other 10 are included in their HVC sample. Seventeen of the UCHVCs are not included in the data coverage of the GALFA DR1 release (D. Saul 2013, private communication); these sources are in the spring sky region of δ = 8°–16°, where GALFA DR1 has limited coverage because GALFA-HI observations started one year after the commencement of ALFALFA data taking and hence commensal data for that time period are missing. Of the thirty-one UCHVCs with GALFA coverage not contained within the catalog of S12, eight of these sources are found by the algorithm but discarded due to either failing the S12 criteria or data quality issues, such as noise spikes. Five are seen in the data but not found by the signal identification algorithm of S12. The last eighteen are not visible in the GALFA-HI data (D. Saul 2013, private communication). In Figure 13, we explore the differences in properties between the UCHVCs found in the dataset of GALFA-HI by the signal identification algorithm of the S12 (including sources discarded from the final catalog), the UCHVCs visible in the GALFA data but not identified by their automated algorithm, and the UCHVCs not visible in the GALFA data. Most strikingly, there is a bimodal distribution in the average column density with the UCHVCs not visible in the GALFA-HI data having the lowest average column densities. In addition, there is a velocity width effect; generally the UCHVCs identified within the GALFA dataset are the narrowest velocity width sources. In the bottom right panel of Figure 13, we focus on UCHVCs with integrated flux densities less than 3 Jy km s⁻¹ as the higher flux sources are all detected in the GALFA-HI data. Then, there are 18 UCHVCs with linewidths greater than 23 km s⁻¹, the median W₅₀ of the full sample. Of these, only three are identified in the GALFA-HI dataset and those still tend to be among the highest flux objects with integrated flux densities greater than 1.45 Jy km s⁻¹, above the median value of 1.34 Jy km s⁻¹. The UCHVCs that are identified within the GALFA dataset that have flux densities below the median value of the UCHVC sample also have linewidths narrower than the median value of the UCHVCs. This is a straightforward result of the different focus of the two surveys; the GALFA-HI data are designed to detect narrow velocity width H i features associated with Galactic hydrogen while the ALFALFA dataset is designed to detect extragalactic H i sources with wider linewidths. While we will address the completeness and reliability of the UCHVC catalog in future work, we note that six UCHVCs not included in the GALFA catalog have all been confirmed as real H i signals via confirmation observations with the Arecibo L-Band Wide receiver (E. A. K. Adams et al., in preparation). In addition, the UCHVCs presented here have strict S/N criteria so the likelihood that many of the UCHVCs are false detections is small. This demonstrates the utility of the ALFALFA dataset, detection algorithm presented here, and the source inspection.

Zoom In Zoom Out Reset image size

An LG origin for HVCs, or at least a subset of the HVC population, has been considered before. With the advent of large-scale, sensitive, blind H i surveys, interest was revived in HVCs as tracers of dark matter halos. Blitz et al. (1999) and Braun & Burton (1999) both postulated an LG origin for HVCs; Braun & Burton (1999) specifically proposed that compact HVCs (CHVCs), identified by their isolation and undisturbed spatial structure, were good candidates to represent dark matter halos throughout the LG. de Heij et al. (2002b) extracted a set of CHVCs from the Leiden/Dwingeloo Survey (LDS), and Putman et al. (2002) similarly presented a set of CHVCs from the HIPASS. Further work, both observational and theoretical, is needed since the discovery of the CHVC population suggests that they most likely represent a circumgalactic population. The properties of the CHVC population from the two catalogs are summarized in Table 5. Sequentially, this table lists: object class, distance (in kpc), H i angular diameter (in arcmin), H i diameter (in kpc), peak column density, W₅₀, integrated flux density, H i mass, and dynamical mass within the H i extent. de Heij et al. (2002a) showed that the properties of the CHVCs for the two datasets are the same when accounting for the better spatial resolution and sensitivity of HIPASS and the better velocity resolution of LDS.

Sternberg et al. (2002) and Maloney & Putman (2003) independently modeled gas in dark matter halos to understand the CHVC population. Based on considerations of their astrophysical properties, both groups concluded that the best interpretation of the CHVCs was as circumgalactic objects at d ≲ 200 kpc. Sternberg et al. (2002) found that if the CHVCs were at d > 750 kpc, their dark matter halos were extremely underconcentrated. They found that at d ≲ 150 kpc, the CHVCs were consistent with being gas pressure confined in dark matter halos. In this scenario, the CHVCs represent the subhalos surrounding the MW from its hierarchical formation. Both pointed out that the gas of the CHVCs must be largely ionized, implying that the total mass of gas is much greater than the observed mass. If the CHVCs were at distances of 0.7–1 Mpc, extremely low dark-matter-to-gas ratios would then be required to match the observed linewidths of the CHVCs, and they would violate the ΛCDM mass–concentration relation. They argued that the CHVCs must be at d ≲ 200 kpc to match size and total dark matter constraints. More recent observational evidence also indicates that the CHVCs must be at circumgalactic distances. The H i masses of the CHVCs at LG distances of ∼1 Mpc are a few times 10⁷ M_☉, large enough that they should have been detected in surveys of other galaxy groups though they have not been (e.g., Pisano et al. 2004, 2007; Chynoweth et al. 2011a; Zwaan 2001; Braun & Burton 2001). In addition, higher resolution observations of CHVCs show clear ram pressure indicators in many cases, indicating that the CHVCs are located at circumgalactic distances (Westmeier et al. 2005b). Observations of potential CHVC analogs around M31 also point to a circumgalactic origin. Westmeier et al. (2005a) studied HVCs associated with M31 in high resolution; importantly, the association of these HVCs with M31 allows a distance constraint to be derived. As outlined in Table 5, the properties of the M31 HVCs are a good match to the properties of the CHVCs at d ∼ 150 kpc, indicating that the two samples are likely a similar population.

Multiple searches have been undertaken for minihalos around nearby galaxy groups (e.g., Zwaan 2001; Braun & Burton 2001; de Blok et al. 2002; Minchin et al. 2003; Barnes & de Blok 2004; Pisano et al. 2004, 2007, 2011; Chynoweth et al. 2009, 2011a, 2011b; Kovač et al. 2009; Irwin et al. 2009; Mihos et al. 2012). Generally, these surveys must choose between sensitivity and coverage area. Irwin et al. (2009) undertook a deep survey of the nearby isolated galaxy NGC 2903 sensitive to an H i mass of 2 × 10⁵ M_☉ and covering 150 kpc × 260 kpc. This survey was sensitive enough to (barely) detect a Leo T analog but given that the survey footprint only extends to ∼100 kpc in projected radius from the galaxy center, detection of an object at ≳400 kpc from the galaxy center would depend strongly on orientation. Irwin et al. (2009) did detect one minihalo with an H i mass of 2.6 × 10⁶ M_☉, a comparable stellar mass and a dynamical mass of ≳ 10⁸ M_☉. Chynoweth et al. (2011b) undertook a large (480 kpc × 1.2 Mpc; 87 × 213) survey centered on the region between the M81/M82 and NGC 2403 galaxy groups. Their survey had a mass detection limit of 3.2 × 10⁶ M_☉which is not deep enough to detect a Leo T analog. While their survey covers a large footprint, it is focused on the region between two connected galaxy groups and coverage of the outskirts of the galaxy groups is limited. They detect several massive H i clouds (M > 10⁶ M_☉) and determine that these clouds likely arise from tidal processes given their clustering near M81. Mihos et al. (2012) surveyed the M101 group over 1050 × 825 kpc (85 × 67) to a mass sensitivity of varying from 2 to 10 × 10⁶ M_☉ over their footprint. This footprint includes all objects out to ∼400 kpc from the central galaxy, regardless of orientation, but the survey is not sensitive enough to detect a Leo T analog. They do identify a new low surface brightness dwarf galaxy through an H i detection and a starless H i cloud with an H i mass of 1.2 × 10⁷ M_☉.

In considering the UCHVCs as gas-bearing minihalos in the LG, we first want to examine the context of the LG and ask what we may empirically expect a minihalo to look like. The population of the LG has increased substantially in the last few years with the discovery of the UFD satellites of the MW from automated stellar searches of the SDSS (Willman 2010) and targeted searches for satellites of M31 (e.g., Ibata et al. 2007; McConnachie et al. 2009). The UFDs have indicative dynamical masses within the baryon extent of 10⁶–10⁷ M_☉ and most likely inhabit dark matter halos that qualify them as minihalos. With the exception of Leo T and the recently discovered Leo P, the UFDs are located within the virial radius of the MW or M31 and have no detectable gas content.

Surveys of low mass galaxies in the field indicate that, with large scatter, dwarf galaxies tend to be gas-rich and can have atomic gas as their dominant baryon component (e.g., Geha et al. 2006; Schombert et al. 2001). Modulo the uncertainties in how astrophysical processes affect the baryon content of the lowest mass halos, one would naively expect the trend of high gas fraction to continue as lower mass galaxies are discovered. Leo T is the only UFD discovered through optical surveys that has neutral gas content; it is also the UFD that is most distant from the MW. The other UFDs are located within the virial radius of the MW or M31 and many show signs of tidal interaction with the MW (e.g., Sand et al. 2012). Grcevich & Putman (2009) find that morphological segregation is strong in the LG with dwarf galaxies within 270 kpc of the MW or Andromeda showing no evidence of neutral gas content. Leo T is on the edge of detectability for SDSS; were it located farther away, its stellar population would not have been detected (Kravtsov 2010). Taken together, these facts raise the possibility that more gas-rich UFDs are lurking in the LG with distances and stellar populations that would leave them undetected in SDSS.

Leo T serves as our prototype of what a gas-rich minihalo will look like; it has motivated our search for more minihalos and the discovery of Leo P. In Figure 14 we examine the H i properties of the LG galaxies and neighboring dwarf galaxies within 3 Mpc in comparison to Leo T and Leo P to infer what we may expect for future minihalo detections. The top panel of Figure 14 shows a histogram of the H i masses of dwarf galaxies within the LG and neighboring systems, taken from McConnachie (2012). Leo P and Leo T have some of the lowest H i masses in the LG and Local Volume (LV); we would expect previously undetected systems to have low H i masses. The bottom panel of Figure 14 illustrates the parameter space occupied by Leo T and Leo P in the LG and LV; they have low H i masses and low dynamical masses.

Zoom In Zoom Out Reset image size

In assessing the UCHVCs as minihalo candidates, we first consider if their astrophysical properties are consistent with the scenario. As mentioned above, the UCHVCs are a good match to the models of Sternberg et al. (2002). Importantly, the UCHVCs also overcome the objections that ruled out the CHVCs as minihalo candidates throughout the LG. As summarized in Table 5, the UCHVCs have H i masses typical of ∼10⁵ d² M_☉ and H i diameters of ∼2.9 d kpc. These smaller sizes and lower fluxes suggest that at distances of 1 Mpc, the physical properties of the UCHVCs are good matches to the CHVC properties at distances of ∼250 kpc. In this scenario, the CHVCs could represent subhalos within the MW and the UCHVCs represent isolated structures within the LG.

The LG is a bound group of galaxies, hence studying the kinematics of the UCHVCs can help constrain their association with the LG. In Figure 15 we compare the motions of the UCHVCs to the LG. Following Courteau & van den Bergh (1999), we plot v_☉ versus the cosine of the angle from the LG apex. In general the UCHVCs show similar behavior to the motions of the LG galaxies, lending credence to the possibility that they trace LG dark matter halos. They do appear to have a higher velocity dispersion, similar to the nearby neighbor galaxies that are not bound to the LG. This may suggest that the UCHVCs are outlying systems, marginally bound to the LG.

Zoom In Zoom Out Reset image size

Finally, we offer a preliminary comparison of the UCHVCs to the Via Lactea II (VL) simulation of Diemand et al. (2008), a high resolution cosmological N-body simulation of a MW analog. We compare the spatial and kinematic distribution of the UCHVCs to the dark matter halos of the VL simulation to see if the hypothesis of UCHVCs as minihalos is consistent with theoretical predictions. We utilize the full volume of the simulation, which includes 20,048 halos that extend to more than 3 Mpc from the central MW analog halo. In addition to the central massive halo, there is a second massive halo which is a fortuitous analog to M31 (Teyssier et al. 2012). In our favored model, we place this second massive halo at the approximate location of M31 in order to most closely match the LG. We also use the original simulation coordinates plus five random orientations of the subhalos to demonstrate the importance of structure within the LG. After the coordinate transformations, we only consider the halos within the simulation that lie within the boundaries of the α.40 coverage and meet our velocity criterion.

In Figure 16 we show the distribution of galactic latitude and v_LSR for the UCHVCs and the VL subhalos. Due to the presence of large and complex HVC structure in the fall sky, we focus on the spring sky for our comparison. In the left column we show all the halos that match our selection criteria; in the right column we show only those halos located further than 250 kpc from the central massive halo to more closely approximate the halos we expect to be gas-bearing. The effects of structure are much more noticeable when only the most distant halos are considered; the different orientations show a much wider spread in the distribution of |b| in this case. The galactic latitude plot is especially important as it provides a quick test of whether the distribution of clouds is within the Galactic disk or a circumgalactic distribution. If the UCHVCs are associated with the Galactic disk, a flattened distribution of |b| values is expected compared to the case if the UCHVCs are distributed around the Galaxy. The UCHVCs and MIS UCHVCs have similar distributions for |b| and |v_LSR|. The favored orientation of the VL simulation appears to match well the distribution of |b| for the UCHVCs. The large differences in the cumulative distribution function (CDF) of |b| for the random orientations shows the importance of structure. The kinematics of the UCHVCs appear to be consistent with the VL simulation in all cases with the CDF of |v_LSR| matching well in all cases. While it is beyond the scope of this paper to do a full halo-population analysis, the rough analysis presented here shows that the UCHVCs agree reasonably well with the VL simulation.

Zoom In Zoom Out Reset image size

We can also use the VL simulations to provide a rough check of the numbers of halos expected. There are 40 UCHVCs in the spring sky, including 17 in the most isolated subsample. We compare to our favored orientation of the VL simulation, noting that it matches the spring sky in that we are looking into the outskirts of the simulation as the spring region of ALFALFA probes the outskirts of the LG. There are a total of 168 VL halos that meet our velocity criterion in the region of the simulation that matches the α.40 spring footprint. When limited to halos with distances from the central MW analog halo greater than 250 kpc, there are a total of 44 halos; 27 of these halos have M_tidal > 10⁷ M_☉. Given the roughness of our numbers the two populations appear to be consistent.

As galaxies, the UCHVCs would favor the outskirts of the LG, rather than the central regions, with distances of ∼500 kpc–1 Mpc. They would have H i masses of ∼10⁵ M_☉ with envelopes of warm ionized hydrogen with masses of ∼10⁶ M_☉. The indicative dynamical masses within the H i extent are ∼10⁷–10⁸ M_☉, and the total hosting halo masses are likely ≳ 10⁹ M_☉. While this hypothesis is attractive, it cannot be definitively proven until distance constraints are in place for the UCHVCs. Further work is necessary in order to constrain their distances as the ALFALFA H i detection carries no direct distance information. The kinematics of the UCHVCs are dominated by LG interactions, so the velocity cannot offer any insights to the distance. The detection of an optical counterpart can constrain the distance through studies of the stellar population. It is also possible to constrain the distance solely through H i by using synthesis imaging to determine the rotational velocity of the UCHVCs and constrain the distance through the baryonic Tully–Fisher relation (e.g., Giovanelli et al. 2013; McGaugh 2012). An alternative to confirming the distance of the UCHVCs directly is to detect UCHVC analogs around other nearby galaxy groups and use the association with the group to constrain the distance and properties of the clouds. Planned future H i surveys using phased-array-feeds will be able to robustly detect these objects.

Confirming that a subset of the UCHVCs are galaxies will offer many insights. The UCHVCs will increase the number of low-mass galaxies known in the Local Volume, decreasing the discrepancy between simulations and observations. In addition, the UCHVCs will trace the outskirts of the LG allowing the comparison between simulations and observations to be extended to a larger volume. The UCHVCs will also serve as isolated examples of the lowest mass galaxies, having not yet interacted substantially with the MW. The UCHVCs offer the potential to study star formation in extreme, low metallicity environments as the presence of gas means there is a possibility of star formation. In fact, Leo T has recently formed stars and Leo P has ongoing star formation with one H ii region. Abundance measurements of the H ii region in Leo P indicate that it is among the lowest metallicity systems known and blind H i surveys may prove to be a promising way to detect low luminosity, extremely metal deficient galaxies (Skillman et al. 2013).

The two confirmed low mass gas-rich galaxies in the Local Volume, Leo T and Leo P, both have high average column densities and small H i angular diameters, as can be seen in Figures 8 and 9. It may be reasonable to expect then that the most compact and highest column density UCHVCs are the best candidates to represent low-mass gas-rich galaxies. HVC274.68+74.70-123, HVC351.17+58.56+214, and HVC13.59+54.52+169 are in the most isolated subsample, have average angular diameters <7', and have $\bar{N}_{{\rm H\,\scriptsize{I}}} > 10^{19}\, \mathrm{cm}^{-2}$; we suggest that these are the best galaxy candidates in our sample. One of these candidates, HVC351.17+58.56+214 is also identified by the GALFA-HI survey as a good galaxy candidate. Notably, it is among the most compact clouds included in this catalog (7' × 5') and has one of the highest column densities ($\log \bar{N}_{{\rm H\,\scriptsize{I}}} = 19.3$). If we adopt a representative distance of 1 Mpc, it has an H i mass of 3.9 × 10⁵ M_☉ and an indicative dynamical mass within the H i extent of 2.1 × 10⁷M_☉.