Filename
stringlengths 22
64
| Paragraph
stringlengths 8
5.57k
|
|---|---|
Processed_First_Results_from_the_HI_Jodrell_All_Sky_Survey:_.txt | telescope pointing model, an improvement in the smooth- ness of the scanning and a concerted effort to reduce local (Jodrell-based) interference during the observing run. |
Processed_First_Results_from_the_HI_Jodrell_All_Sky_Survey:_.txt | The general method of detecting galaxies in the HIJASS data is as follows. An initial candidate galaxy list is formed by visually searching the datacubes. The visual display pro- gram KVIEW (Gooch 1995) is used to search through the data in three dimensions by displaying 2 axes and stepping through the third. Narrow velocity-width, bright galaxies are most easily seen when the data are displayed in R.A. versus Decl. and stepped through Velocity. Broad velocity-width, faint galaxies are more easily discovered when the datacube is displayed R.A. or Decl. versus Velocity. The selection cri- teria for this visual sample is that a detection must : (1) be easily visible above the noise (∼3σ in peak flux), (2) should have a spatial extent of greater than 1 pixel, and (3) be visi- ble over two or more velocity planes. Two lists are compiled using these criteria: one of ‘definite’ detections and one of ‘possible’ detections. The list of possible detections includes sources close to the selection limit or just below it but still considered possible sources. |
Processed_First_Results_from_the_HI_Jodrell_All_Sky_Survey:_.txt | A second candidate galaxy list is formed by running the automated finding algorithm POLYFIND, written by Robert Minchin and Jonathon Davies. POLYFIND de- termines the noise in non-masked regions of a hanning smoothed datacube and then looks for peaks at some user- defined level above the noise (typically set at 3σ). It then runs a series of matched filters over these identified peaks and a peak is noted as a potential source if a sufficiently good fit is obtained. The results from POLYFIND are then re-checked by eye using the same criteria as used for the eyeball method and lists of definite and possible detections made. |
Processed_First_Results_from_the_HI_Jodrell_All_Sky_Survey:_.txt | The two lists of definite detections are then amalga- mated and these objects included in our final sample of con- firmed HIJASS galaxies. The two lists of possible detections are also amalgamated. These objects are then re-observed with the Lovell telescope in single-beam mode using a band- width of 16 MHz. Those possible sources confirmed by this narrow-band follow-up are then added to the sample of con- firmed HIJASS sources. The possible sources found during the 2002 run have not yet been subjected to narrow-band follow-up. Hence, for the 2002 data, only the definite detec- tions found by the above selection methods have presently been included in the sample (69 sources). |
Processed_First_Results_from_the_HI_Jodrell_All_Sky_Survey:_.txt | is then used to generate a spatially integrated spectrum of the detection, using a box size based on the extent of the HI. The spectrum is generated using MBSPECT which also gives a measurement of the peak and integrated flux of each detection, as well as the 50 per cent and 20 per cent velocity- width, the rms noise and barycentric velocity. |
Processed_First_Results_from_the_HI_Jodrell_All_Sky_Survey:_.txt | Table 2 presents the derived parameters for the sample of confirmed HIJASS sources. Column 1 gives the HIJASS Name. Columns 2 and 3 give the Right Ascension (J2000) and Declination (J2000) from the IMFIT task. Columns 4-6 SV dV ), list the zeroth order moment (Integrated flux SInt= the peak flux (Speak), and the noise (rms dispersion around the baseline, σ); all as measured by MBSPECT. Columns 7- 9 give the first order moment (barycentric velocity V⊙), the velocity width at 20 per cent of the peak flux (∆V20), and the velocity width at 50 per cent of the peak flux (∆V50), all measured by MBSPECT in the radio frame and converted to cz. The error in integrated flux is calculated from the rms noise on the spectrum and the velocity extent of the source. Columns 10-13 contain the details of any counterpart to the HIJASS source, as listed in the NASA/IPAC Extragalac- tic Database (NED). Column 10 contains one of five possible classifications. If there is no object within NED which could be spatially coincident with the HIJASS source (defined as being within 6 arcmin), then Column 10 contains the clas- sification ‘PUG’ (i.e. Previously Uncatalogued Galaxy). If there is an object in NED which matches the HIJASS source in both position and space (defined as being within 6 arcmin and 100 km s−1) then this is listed as ‘ID’ (Identification). Those IDs which have been detected in HI for the first time by HIJASS are denoted by an asterix, i.e. ‘ID*’. If there is an object within NED which is spatially coincident with the HIJASS source (i.e. within 6 arcmin) but for which no redshift is listed in NED, then Column 10 contains the clas- sification ‘ASS’ (i.e. Association). In several cases, there is more than one galaxy within 6 arcmin: in these cases the classification ‘ASS*’ is used. For the ID, ID* and ASS clas- sifications, Column 11 lists the object within NED which appears to correspond to the HI detection. Column 12 lists the position offset (in arcmin) of the coordinates of the op- tical counterpart from the HI position. Column 12 lists (for the IDs and ID*s) the velocity offset of the barycentric ve- locity contained within NED from the barycentric velocity as measured by HIJASS. |
Processed_First_Results_from_the_HI_Jodrell_All_Sky_Survey:_.txt | There are currently 222 sources included in the sample of confirmed HIJASS sources. Of these, 170 (77 per cent) are clearly associated with a previously catalogued galaxy (clas- sification ID or ID*). However, 25 of these 170 objects have been detected in HI for the first time by HIJASS (classifica- tion ID*). For 4 of these 170 sources, HIJASS appears to be measuring HI from a pair or a small group of galaxies which lie at the redshift of the HI. |
Processed_First_Results_from_the_HI_Jodrell_All_Sky_Survey:_.txt | not be associated with the catalogued galaxy. 15 of these sources have only one possible optical counterpart within 6 arcmin of the HI position (classification ASS). We may be relatively confident about the optical identification of these sources. However, even for these sources there remains the possibility that the HI has been detected from an associated HI cloud (cf. Ryder et al. 2001) or a LSB companion. The other 8 of the 23 sources have more than one galaxy within 6 arcmin of the HIJASS position (classification ASS*). We intend to obtain accurate positions for all 23 of these sources using HI aperture synthesis observations, so as to unambigu- ously determine the optical counterpart of each source. |
Processed_First_Results_from_the_HI_Jodrell_All_Sky_Survey:_.txt | There are then a further 29 sources (13 per cent of the whole sample) which do not lie within 6 arcmin of any previously catalogued galaxy (classification PUG). A study of the Digital Sky Survey (DSS) at the positions of these sources reveals an obvious and unambiguous optical coun- terpart in only 5 cases (J0327+67, J0721+74, J1720+71, J2016+75 and J2151+71). Fig. 2 presents the DSS im- ages of these five PUGs. Three of these objects (J0327+67, J2016+75, J2151+71) have a compact but relatively high surface brightness core but a low surface brightness disk. They may have been excluded from optical catalogues be- cause they were mistaken for stars. One object (J0721+74) is a highly inclined but relatively high surface brightness object, although very small (≃1 arcmin diameter). The fifth object (J1720+71) has a complex morphology and appears to be involved in some kind of interaction or merger. |
Processed_First_Results_from_the_HI_Jodrell_All_Sky_Survey:_.txt | A study of the DSS for the other 24 PUGs, reveals no unambiguous optical counterpart. In many cases there are several possible optical candidates within the positional un- certain of HIJASS. In several cases, however, no possible candidate can be seen. The optical counterparts of these sources must be of very low surface brightness. All of the PUGs will have accurate positions determined from aper- ture synthesis observations and will be the subject of deep optical follow-up work. |
Processed_First_Results_from_the_HI_Jodrell_All_Sky_Survey:_.txt | It is interesting to compare the number of PUGs found in the sample of confirmed HIJASS sources with those found in the Bright Galaxy Catalogue (BGC: Koribalski et al., in preparation). The BGC contains the 1000 brightest (in HI peak flux) sources in the whole of the HIPASS sample. 87 of these objects had not been previously catalogued, although 57 of these lie close (within 10 ◦) to the Galactic plane. Of the other 30 previously uncatalogued galaxies within the BGC, Ryan-Weber et al. (2002) found a single optical counterpart for 25 on the DSS. Whilst the relative number of previously uncatalogued galaxies within the BGC is much smaller than within the HIJASS sample (3 per cent compared to 13 per cent), most of the BGC objects can be unambiguously as- signed to an optical counterpart on the DSS, whilst most of the HIJASS PUGs cannot be. These differences are prob- ably primarily due to the different flux limits of the two samples. The faintest source in the BGC has a peak flux of 116 mJy. Only 6 of the 29 HIJASS PUGs have a peak flux larger than this. The much lower peak flux limit of HIJASS has produced a much larger fraction of PUGs compared to the BGC. Since these have generally lower HI flux, they are correspondingly harder to detect in optical data. |
Processed_First_Results_from_the_HI_Jodrell_All_Sky_Survey:_.txt | Figure 2. Digitial Sky Survey images of the 5 PUGs for which an obvious and umabiguious optical counterpart can be seen. Each image is 5 arcmin×5 arcmin. |
Processed_First_Results_from_the_HI_Jodrell_All_Sky_Survey:_.txt | ple. It presents the first redshift measurement of 52 galaxies (i.e. classes ASS, ASS*, PUG), 23 per cent of the whole sam- ple. Between 29 and 52 (13 and 23 per cent) of the objects within it have not been previously catalogued. It must also be noted that the ‘possible’ detections from the 2002 observ- ing run have not yet been followed up. Based on the results of previous narrow-band follow-up, this will probably lead to an additional 5-10 sources being added to the sample, many of them previously uncatalogued sources. |
Processed_First_Results_from_the_HI_Jodrell_All_Sky_Survey:_.txt | The main factor which determines the inclusion of a galaxy within the HIJASS sample ought to be its peak flux (rather than integrated flux). Eyeball searches are inevitably drawn to sources with larger peak fluxes. The POLYFIND au- tomated finding algorithm also initially looks for peaks in individual pixels. Fig. 3 is a histogram of the peak flux of every source in the sample of confirmed HIJASS sources. For a peak flux limited survey of a homogeneous distribution of galaxies we expect Nobj∝S−5/2 . The curve on Fig. 3 shows the best fit of this function to the observed distribution. This implies that our sample is complete to SPk ≃80 mJy. We noted above that the rms noise in HIJASS data shows con- siderable variation between cubes. In particular, the cubes from the 2002 run have considerably lower noise. The com- pleteness limit of 80 mJy corresponds to a 5σ detection in the pre-upgrade data. The 3σ detection limit for the post- upgrade data is at ≃39 mJy. Only 2 sources have a peak flux less than this. |
Processed_First_Results_from_the_HI_Jodrell_All_Sky_Survey:_.txt | Figure 3. Peak flux distribution of HIJASS sources. The curve shows the best-fit to the data of the function Nobj∝S , i.e. that expected for a peak flux limited sample of a homogeneous distribution of galaxies. |
Processed_First_Results_from_the_HI_Jodrell_All_Sky_Survey:_.txt | ple: we detect few galaxies at ∆V20<50 km s−1. This is be- cause there is a minimum believable velocity-width which a galaxy must have in order to be selected as a real source from our data. Sources with a narrower velocity-width will be mistaken for narrow band radio frequency interference. From the data it appears that this minimum believable velocity-width is around 4 channels wide for ∆V20, i.e. 20 =52.8 km s−1. This makes intuitive sense as it allows ∆Vlim 2 ‘high’ channels where the source is seen and believed and 2 ‘low’ channels where the flux is dropping off down to the 20 per cent level. The locus of this limit is drawn on Fig. 4 and constrains the data well. In Section 5 we consider the implications of this velocity-width limit for the completeness of HI-selected samples of galaxies. |
Processed_First_Results_from_the_HI_Jodrell_All_Sky_Survey:_.txt | Because the HIJASS sample is approximately peak flux limited, the detection of a galaxy of a given integrated flux, SInt, will (even in similar noise) be a function of its velocity- width: broader velocity-width galaxies of a given integrated flux have a lower peak flux and therefore are less likely to be detected than a narrower galaxy of the same integrated flux. This is clearly seen in Fig. 5, a plot of log(SInt) against ∆V20. From this can be seen a clear trend for the minimum detected integrated flux to increase with increasing velocity- width. |
Processed_First_Results_from_the_HI_Jodrell_All_Sky_Survey:_.txt | Figure 5. Plot of log (SInt) against ∆V20 for all the galaxies in the HIJASS sample. The long-dashed line is the locus of equation (3) for Slim pk =48 mJy (i.e. a 3σ detection from the pre-upgrade observing runs). The short-dashed line is the locus of equation (3) for Slim Pk =39 mJy (i.e. a 3σ detection from the post-upgrade observing run). |
Processed_First_Results_from_the_HI_Jodrell_All_Sky_Survey:_.txt | for the HIJASS sample. Hence, for a given peak flux limit, Slim Pk , the integrated flux limit, Slim Int is a function of ∆V20 via: Int ≃ 0.6 ∆V20 Slim Slim Pk The loci of this relationship for Slim Pk =48 mJy (i.e. a 3σ detection from the pre-upgrade observing runs) and for Slim Pk =39 mJy (i.e. a 3σ detection from the post-upgrade ob- serving run) are plotted on Fig. 5. These describe well the form of the observed cut-off in SInt as a function of ∆V20. |
Processed_First_Results_from_the_HI_Jodrell_All_Sky_Survey:_.txt | It is worth noting that the fact that our sample is es- sentially peak flux limited has a dramatic effect on the pro- portion of broad velocity-width to narrow velocity-width galaxies included in the sample, compared to the propor- tion we would expect to find in an integrated flux limited sample. For the sake of illustration, we consider the fraction of galaxies with SInt >3 Jy km s−1 which will be included in the HIJASS sample as a function of velocity-width. The value of 3 Jy km s−1 corresponds to a galaxy of velocity- width 100 km s−1 with a peak flux of about 48 mJy (the 3σ limit for the pre-upgrade data). At ∆V20=100 km s−1, all galaxies with SInt >3 Jy km s−1 will be included in the HI- JASS sample. However, at broader velocity-widths, the inte- grated flux limit will increase [equation (3)] and the sample will contain a progressively smaller fraction of galaxies with SInt >3 Jy km s−1. |
Processed_First_Results_from_the_HI_Jodrell_All_Sky_Survey:_.txt | Figure 6. Plot of SInt against ∆V20.SPk for the galaxies in the sample of confirmed HIJASS sources. |
Processed_First_Results_from_the_HI_Jodrell_All_Sky_Survey:_.txt | Table 3. The fraction of galaxies with SInt >3 Jy km s−1 which will be missed from the HIJASS sample as a function of ∆V20. This assumes Slim Pk =48 mJy and that the distribution of galaxies in space at each ∆V20 is homogeneous. |
Processed_First_Results_from_the_HI_Jodrell_All_Sky_Survey:_.txt | At ∆V20=150 km s−1 only 68 per cent of galaxies with SInt>3 Jy km s−1 will be included in the HIJASS sample. At ∆V20 >300 km s−1 less than 20 per cent of galaxies with SInt>3 Jy km s−1 will be included. This is a particularly important selection effect since we might reasonable expect that broader velocity-width galaxies will tend to have higher HI masses (see e.g. Rao & Briggs 1993). Hence, compared to an integrated flux limited sample, our selection techniques may be significantly biased against the inclusion of higher HI mass galaxies. This effect will not bias an HIMF derived from the data provided that the selection effect is properly accounted for. It does, however, mean that the morpholog- ical mix of galaxies revealed by a blind HI survey is going to be biased towards narrow velocity-width dwarf galaxies and away from broad velocity-width giant galaxies. This bias needs to be born in mind when considering the relative pro- portions of the different morphologies of galaxies in an HI- selected sample. |
Processed_First_Results_from_the_HI_Jodrell_All_Sky_Survey:_.txt | Figure 7. Histogram showing the distribution of the offset be- tween HIJASS source position and the position of the identified optical counterpart. |
Processed_First_Results_from_the_HI_Jodrell_All_Sky_Survey:_.txt | The positional accuracy of HIJASS sources can be judged by considering the offset between the HIJASS positions and the positions listed in NED for those galaxies identified as being associated with each HIJASS source (i.e. the IDs and ID*s). Fig. 7 shows a histogram of these offsets. The majority of HIJASS sources (71 per cent) lie within 2.5 arcmin of the NED position, with only a very small fraction (7 per cent) lying beyond 4 arcmin. |
Processed_First_Results_from_the_HI_Jodrell_All_Sky_Survey:_.txt | Figure 8. (a) MHI distribution of the whole of the HIJASS sam- ple; (b) MHI distribution of the 77 galaxies which had not pre- viously been detected in HI (classes ID*, ASS, ASS*, PUG); (c) MHI distribution of the 52 galaxies with no previous redshift mea- surement (classes ASS, ASS*, PUG); (d) MHI distribution for the 29 PUGs. |
Processed_First_Results_from_the_HI_Jodrell_All_Sky_Survey:_.txt | tween 109.3–109.8 M⊙. All but 2 have integrated fluxes above 10 Jy km s−1. These would appear to mostly be ‘normal’ galaxies which had simply not been observed in HI prior to HIJASS. |
Processed_First_Results_from_the_HI_Jodrell_All_Sky_Survey:_.txt | Figure 9. (a) The cz distribution of the whole of the HIJASS sample; (b) the cz distribution of the 77 galaxies which had not previously been detected in HI (classes ID*, ASS, ASS*, PUG); (c) the cz distribution of the 52 galaxies with no previous redshift measurement (classes ASS, ASS*, PUG); (d) the cz distribution for the 29 PUGs. |
Processed_First_Results_from_the_HI_Jodrell_All_Sky_Survey:_.txt | per cent of the full sample lie in this mass range. A similar result was found by Ryan-Weber et al. (2002) for the 30 pre- viously uncatalogued galaxies in the BGC at |b|>10◦. They found the mass distribution of these 30 galaxies to peak at 108.7 M⊙, compared to a peak at 109.5 M⊙ for the whole of the BGC. This tendency for the newly discovered galaxies to have lower HI masses suggests a correlation between HI mass and optical detectability. |
Processed_First_Results_from_the_HI_Jodrell_All_Sky_Survey:_.txt | Fig. 9(a) shows the distribution of all the HIJASS galax- ies as a function of cz. The peak close to 0 km s−1 is due to galaxies in the Local Group and the M81 Group. Be- yond that there are prominent peaks at cz∼1200 km s−1 and 2500 km s−1. These are due to large-scale structure. Note that only a handful of galaxies have been identified beyond 4500 km s−1. This is mainly due to the increasing problems of RFI beyond this point and the difficulty of distinguishing any galaxies from interference. |
Processed_First_Results_from_the_HI_Jodrell_All_Sky_Survey:_.txt | Fig. 9(b) shows the cz distribution for all of those 77 galaxies which had not previously been detected in HI (Classes ID*, ASS, ASS*, PUG). This distribution is very similar to that of the whole sample. Fig. 9(c) shows the cz distribution for the 52 objects which had no previous redshift measurement (classes ASS, ASS*, PUG). Fig. 9(d) plots the HI mass distribution just for the 29 PUGs. All of these distributions show the same peaks at cz=1200 km s−1 and 2500 km s−1 as that seen in Fig 8(a). Clearly, the dis- tribution of new HI detections, new redshift measurements and newly catalogued galaxies follows that of the large-scale structure as revealed by the whole sample. This is partic- ularly interesting in regard to the previously uncatalogued galaxies. These have a HI mass distribution with a peak an order of magnitude lower than that of the whole sam- ple (see Fig. 8(d)) but the cz distribution is not skewed to nearer distances. |
Processed_First_Results_from_the_HI_Jodrell_All_Sky_Survey:_.txt | The relationship of HIJASS galaxies to large-scale structure is further explored in Fig. 10. This is a diagram showing the relationship of HIJASS galaxies (filled trian- gles) and all objects with redshifts in NED (open circles) in the complete R.A. strip between Decl.=70◦→78◦. The general association of HIJASS sources with the large-scale structure as delineated by the NED objects is clear. Most of those previously uncatalogued objects found by HIJASS lie in regions already populated with galaxies. Some of the structures which can be seen in the data include the group of galaxies at R.A.∼17h, V⊙∼1200 km s−1 (which in- cludes NGC 6217, NGC 6236, NGC 6248 and NGC 6395); the NGC 4291 group at R.A.∼13h, V⊙∼1600 km s−1; a group including UGC 03317, UGC 03343 and UGC 03403 at R.A.∼6h, V⊙∼1200 km s−1; and two prominent ‘walls’, at R.A.∼3h→5h, V⊙∼2600 km s−1 and at R.A.∼7h→10h, V⊙∼2300 km s−1. |
Processed_First_Results_from_the_HI_Jodrell_All_Sky_Survey:_.txt | be very rare. For example, Kilborn et al.’s (2002) survey of 2400 deg2 of HIPASS data did not detect any galaxy this massive. Minchin (2001) presented results from sur- veying a 32 deg2 with the Parkes multibeam system to 12× the standard HIPASS exposure time. No galaxy with MHI >1010.6 M⊙ was detected. |
Processed_First_Results_from_the_HI_Jodrell_All_Sky_Survey:_.txt | Following the first stage of the Lovell telescope upgrade, the rms noise in HIJASS data is around 12-14 mJy beam−1. The typical noise in HIPASS data is ∼14 mJy beam−1, although there is considerable variation between HIPASS cubes with rms spanning the range 9-17 mJy beam−1. |
Processed_First_Results_from_the_HI_Jodrell_All_Sky_Survey:_.txt | The data at R.A.≃12h08m→12h34m, Decl.=22◦→30◦, observed during the 2002 run, was taken specifically because it provides a small overlap with the HIPASS survey between Decl.=22◦→25◦. Fig. 11 shows the integrated fluxes from HIPASS and HIJASS cubes for those galaxies which lie in the overlap region. The calibration between the surveys ap- pears robust. |
Processed_First_Results_from_the_HI_Jodrell_All_Sky_Survey:_.txt | As noted in Section 4.4, terrestrial-based RFI effectively limits the HIJASS bandpass to cz<4500 km s−1 whereas HIPASS can survey out to 12700 km s−1. However, only HI can be seen beyond 4500 km s−1 galaxies with MHI >∼M⋆ in HIPASS data. In fact less than 24 per cent of the HIPASS sample presented by Kilborn et al. (2002) lies be- yond 4500 km s−1. HIPASS therefore, has an advantage over HIJASS in that very massive sources can be detected over larger volumes. However, for galaxies with MHI <∼M ⋆ HI, both HIJASS and HIPASS are flux limited and HIJASS is poten- tially more sensitive. |
Processed_First_Results_from_the_HI_Jodrell_All_Sky_Survey:_.txt | The distribution function of neutral hydrogen masses among galaxies and intergalactic clouds (the HI mass function, HIMF) and, more generally, the neutral hydrogen density in the local Universe, ΩHI, are important inputs into models of cosmology and galaxy evolution. Prior to the advent of blind HI surveys, astronomers were restricted to constructing an HIMF by making HI measurements of optically selected samples of galaxies (see, e.g. Rao & Briggs 1993; Solanes, Giovanelli & Haynes 1996). |
Processed_First_Results_from_the_HI_Jodrell_All_Sky_Survey:_.txt | In recent years several authors have attempted to deter- mine the HIMF of the local Universe using an HI-selected sample of galaxies, with conflicting results. For example, us- ing the data from the the Arecibo HI Strip Survey (Sorar 1994), Zwaan et al. (1997) derived an HIMF with a shallow faint end slope (α=1.2) consistent with earlier HIMFs de- rived from optically selected samples. In contrast the HIMF derived from the Arecibo Slice survey (Schneider, Spitzak & Rosenberg 1998; Spitzak & Schnedier 1998) has an up- turn in its lowest mass bin, although this is due to only 2 galaxies in this bin. Recently, Rosenberg & Schneider (2002) have also reported a steep faint end slope (α≃1.5) to the HIMF they have derived from the Arecibo Dual-Beam Sur- vey (Rosenberg & Schneider 2000). HIPASS and HIJASS will provide much larger samples of galaxies and greatly im- prove the statistics of such determinations of the HIMF. |
Processed_First_Results_from_the_HI_Jodrell_All_Sky_Survey:_.txt | Figure 10. Distribution of R.A. vs cz of HIJASS galaxies in the range Decl.=70◦→78◦(filled triangles). Also shown (open circles) is the R.A. vs cz distribution of all galaxies within NED within this declination range and with measured redshifts. |
Processed_First_Results_from_the_HI_Jodrell_All_Sky_Survey:_.txt | However, previous studies based upon HI-selected samples of galaxies have tended to overlook the important effect that the inclination of a galaxy to the line of sight could have on its inclusion in such a sample. There are two factors to be considered. |
Processed_First_Results_from_the_HI_Jodrell_All_Sky_Survey:_.txt | tion effects on the composition of an HI-selected sample of galaxies and the implications this has for derivations of the HIMF and ΩHI. |
Processed_First_Results_from_the_HI_Jodrell_All_Sky_Survey:_.txt | Firstly, highly inclined galaxies may suffer from signifi- cant self-absorption. Studies of the HI emission from galax- ies have generally assumed that the HI line is optically thin in all circumstances. Relatively few authors (e.g. Epstein 1964a,b; Haynes & Giovanelli 1984) have addressed the is- sue of whether the HI emission from galaxies is actually op- tically thin in all galaxies. If this assumption is not valid for highly inclined galaxies, then the HI masses of such galaxies will have been underestimated. Some highly inclined galax- ies will be missed altogether from an HI-selected sample of galaxies, despite less inclined galaxies of the same HI mass being included. Both of these effects will lead to errors in the derived HIMF. |
Processed_First_Results_from_the_HI_Jodrell_All_Sky_Survey:_.txt | Secondly, as noted in Section 4.2, there is a minimum believable velocity-width, ∆Vlim 20 , which an object in a blind HI survey must have in order to be distinguishable from narrow-band radio frequency interference. For any given type of galaxy, the measured velocity-width will be narrower the more face-on the galaxy is. Hence, some galaxies with inclinations close to the line of sight could be missed. |
Processed_First_Results_from_the_HI_Jodrell_All_Sky_Survey:_.txt | where f(i) is the correction factor needed to correct the HI mass derived from the optically thin assumption to the ac- tual HI mass. Assuming this function is significant at all, then f(i) may vary for different morphological types and will increase as inclination angle increases. |
Processed_First_Results_from_the_HI_Jodrell_All_Sky_Survey:_.txt | integrated fluxes from HIJASS Figure 11. Comparison of and HIPASS for galaxies in the region 12h08m→12h34m, Decl.=22◦→25◦. |
Processed_First_Results_from_the_HI_Jodrell_All_Sky_Survey:_.txt | To find the expected observed distribution of galaxies as a function of inclination angle, φ(i), we have to multiply the intrinsic distribution of galaxies as a function of i, N(i), by the volume within which a galaxy at a given i can be observed, V(i), i.e. |
Processed_First_Results_from_the_HI_Jodrell_All_Sky_Survey:_.txt | The expected observed distribution of the HIJASS sam- ple as a function of inclination angle therefore depends on several other relationships: the distribution of galaxies as a function of MHI; the relationship (if any) between MHI and ∆Vo; the relationship between ∆Vo and ∆V20 and inclina- tion angle, i. In Section 5.3 we show that at large i (i.e. >50◦) this relationship can be simplified and used to study the ef- fect of HI self-absorption within the HIJASS sample. Firstly, in Section 5.2, we consider the effect of the velocity-width limit on the HIJASS sample at small inclination angles. |
Processed_First_Results_from_the_HI_Jodrell_All_Sky_Survey:_.txt | Table 4. Illustration of the effect of the velocity-limit selection effect on the inclusion of galaxies within the HIJASS sample. Column 1 is a set of ∆Vo values. Column 2 lists the minimum inclination angle, imin, which a galaxy of each ∆Vo must have in order to be included in the sample (using eqn 14). Column 3 lists ζ, the fraction of galaxies which will be missed from the sample at each ∆Vo. |
Processed_First_Results_from_the_HI_Jodrell_All_Sky_Survey:_.txt | values above. As expected this effect gets progressively more serious for inherent narrow velocity-width objects: galaxies with ∆Vo<100 km s−1 cannot be seen with i<22◦; galaxies with ∆Vo<50 km s−1 are missed if i<49◦. |
Processed_First_Results_from_the_HI_Jodrell_All_Sky_Survey:_.txt | The derived values of ζ are listed in Column 3 of Table 4. Clearly, within HIJASS data, this selection effect becomes progressively more important at smaller ∆Vo. Whilst only 7% of galaxies have been missed at ∆Vo=100 km s−1, this number has risen to 35% at ∆Vo=50 km s−1. |
Processed_First_Results_from_the_HI_Jodrell_All_Sky_Survey:_.txt | As noted above, we cannot properly consider the ef- fect that the velocity-width cut-off may have on a derived HIMF without knowing whether there is a relationship be- tween ∆Vo and MHI and, if so, what form that relation- ship takes. Whilst one could argue about the precise rela- tionship between MHI and ∆Vo, previous studies suggest that the two are related such that more massive galaxies appear to have broader velocity-widths. For example, from HI measurements of an optically-selected sample of galax- ies, Rao & Briggs (1993) derived ∆V20=0.15M1/3 HI . From their HIDEEP sample, Minchin et al.(in preparation) have found that ∆V20=0.42M0.28 . However, such relations may HI be partly due to selection effects (see e.g. Minchin 2001). |
Processed_First_Results_from_the_HI_Jodrell_All_Sky_Survey:_.txt | Figure 12. Plot of log(∆Vo) against log(MHI) for those 186 HIJASS galaxies for which we have inclination angles. The solid line is the locus of the relationship ∆Vo=0.42M0.3 HI (equation 20). |
Processed_From_gas_to_galaxies.txt | The unsurpassed sensitivity and resolution of the Square Kilometer Array (SKA) will make it possible for the first time to probe the continuum emission of normal star forming galaxies out to the edges of the universe. This opens the possibility for routinely using the radio continuum emission from galaxies for cosmological research as it offers an independent probe of the evolution of the star formation density in the universe. In addition it offers the possibility to detect the first star forming objects and massive black holes. |
Processed_From_gas_to_galaxies.txt | In deep surveys SKA will be able to detect H I in emission out to redshifts of z ≈ 2.5 and hence be able to trace the conversion of gas into stars over an era where considerable evolution is taking place. Such surveys will be able to uniquely determine the respective importance of merging and accreting gas flows for galaxy formation over this redshift range (i.e. out to when the universe was only one third its present age). It is obvious that only SKA will able to see literally where and how gas is turned into stars. |
Processed_From_gas_to_galaxies.txt | More and more evidence supports the paradigm that that galaxies form hierarchicallly, present day galaxies have been built up over cos- mological time scales by the coalescence of many smaller, less massive galactic systems [4,57,65,82, 19,1]. For a particular galaxy this process is de- scribed by the so-called merger tree: the accumu- lation of dark matter mass over time into a final halo. These merger trees are extracted from nu- merical comological simulations and only describe the merging of dark matter haloes in a robust manner. |
Processed_From_gas_to_galaxies.txt | the typical shape of a galaxy merger tree has not been confirmed robustly from observations. Ob- servationally, the galaxy mass functions and the merger rates as a function of redshift are largely unknown. |
Processed_From_gas_to_galaxies.txt | In order to address these questions it is necessary to have a census of galaxies and their properties over a large range of environments and as large a possible range of cosmic time or redshift. While the stellar content of galaxies can be probed using optical and NIR imaging and spectroscopy using the best facilities on the ground and in space, the only way to probe the cold gas content of galax- ies is to use a radio telescope, either at very short wavelengths to measure molecular line emission (see also Blain et al. this volume on molecu- lar lines), or at long wavelengths to measure the H I emission using the 21-cm line. To do this out to redshifts of z = 3, since when considerable evo- lution may have taken place [83,37], requires a radio telescope tenfold larger (in collecting area) than any of the existing facilities. The SKA is such a telescope and the purpose of this chapter is to review the current state of affairs, from both the observational and theoretical points of view, and to outline how the SKA will be instrumental in answering the above questions to which we do not yet have firm answers. |
Processed_From_gas_to_galaxies.txt | optical or ultraviolet light which is affected by ex- tinction. The caveat is that the relation between the non-thermal radio emission and the star for- mation rate is assumed not to vary significantly with cosmic time. |
Processed_From_gas_to_galaxies.txt | The framework for the discussion in this chap- ter will be the hierarchical structure formation models of Baugh et al. (this volume), who use the technique of semi-analytic modelling to pre- dict the evolution of the HI mass function and the number of star forming radio galaxies, both signposts of evolution to be probed with SKA. This follows a widely accepted idea of how galax- ies form, based on the success of the Cold Dark Matter (ΛCDM) simulations. |
Processed_From_gas_to_galaxies.txt | Although ΛCDM has proven to be very suc- cesful in reproducing the WMAP results and re- sulting large scale structure, SKA will observe the baryonic component of the universe directly. Predictions of what we should observe are still wide open, as recent hydrodynamical simulations indicate that some of the basic assumptions of the semi-analytic models, i.e. that the gas gets heated quickly to the virial temperature, may not be correct. In the semi-analytic modelling the as- sumption is that the gas gets immediately shock heated to the virial temperature as it flows into the dark matter potential wells. Recently Bin- ney [10], Katz et al. [51] and Murali at al. [66] have pointed out that hydrodynamical simula- tions suggest that most of the gas is not heated to the virial temperature but accretes slowly in a cool phase from the surrounding large scale struc- ture filaments. The consequence is that galaxies grow through accretion of gas rather than merg- ing of dark matter halos. At redshifts beyond z ≈ 2 this cool mode of accretion appears to dom- inate over merging as the main process for galaxy formation. Because these two scenarios differ in the evolutionary scenario of the gas, SKA is the ultimate telescope for distinguishing these from one another. |
Processed_From_gas_to_galaxies.txt | When and how the first episodes of star forma- tion took place remains a key question of mod- ern cosmology. Much effort has been devoted to measuring the integrated cosmic star forma- tion rate (SFR) of the universe as a function of redshift or look-back time, initially at ultraviolet wavelengths [61], and more recently in the mil- limetre and radio regimes [12,38]. Unlike ultra- violet or optical light, radio continuum emission is unaffected by dust, which eliminates the need for uncertain extinction corrections. SKA will detect the radio continuum emission from very large numbers of AGN and star–forming galax- ies at redshifts z > 3. However, we caution that the conversion between a measured radio contin- uum flux and a derived a global SFR requires the adoption of canonical scaling relations [22] which remain valid out to at least z ∼1 [36,3] but, as we discuss below, may not be directly applicable to primordial star-forming systems at z > 3. |
Processed_From_gas_to_galaxies.txt | In the models considered here, the radio contin- uum emission from star–forming galaxies is com- puted using the spectro-photometric code GRASIL written by Silva et al.[79]. This code takes star– formation histories generated by GALFORM, includ- ing bursts of star formation, and produces a full spectral energy distribution, ranging from the far- UV to the radio. The radio emission includes both synchroton radiation by electrons acceler- ated in supernova remnants and free–free emis- sion from HII regions. The plots presented in this section are for a model very similar to that of Baugh et al. (this volume). Note that be- yond redshifts of z > 6 inverse compton losses against the cosmic microwave background will be- gin to affect the synchrotron emission. The non- relativistic, thermal emission is not affected [16] so that radio continuum emission still is a very good tracer of star formation at high redshifts, albeit in another regime. |
Processed_From_gas_to_galaxies.txt | Figure 1. Predicted 1.4 GHz flux densities for star–forming galaxies (with SFRs of 1, 10 and 100 M⊙/yr) at different redshifts. The horizon- tal line at 100 nJy represents the nominal SKA detection limit for an 8 hr integration. |
Processed_From_gas_to_galaxies.txt | Little is currently known about the properties of star–forming objects at redshifts above z∼ 5 but the next few years are likely to see rapid progress in this area. |
Processed_From_gas_to_galaxies.txt | Several recent results suggest substantial star formation at high redshift [5]. Studies of the stel- lar ‘fossil record’ of nearby galaxies for example [40]) imply that the star–formation rate in (e.g. the most massive galaxies increases with redshift and peaks at z=3 or higher. The recent detection of QSOs at redshifts with z>6 [28], in which the gas appears to have solar or above-solar metal- licity [69] suggests that the massive first stars in the Universe had already formed by z=8 or even earlier. Observations with the HST and ground–based optical/IR telescopes have begun to provide direct measurements of the number of star–forming galaxies at z ∼ 5–6 [14,98,47,81]. |
Processed_From_gas_to_galaxies.txt | These studies find typical surface densities of ∼ 100 deg−2 for star–forming galaxies at z ∼ 6. So it appears that the global star–formation rate at z ∼ 6 is roughly similar to that at z ∼ 3– 4, though with substantial uncertainities because the distribution of high–redshift galaxies appears to be clustered on scales comparable to the di- ameters of the individual fields studied to date [47]. |
Processed_From_gas_to_galaxies.txt | At even higher redshift, the detections of indi- vidual (gravitationally–lensed) galaxies at z ∼ 7 [58] and z = 10 [68] have recently been reported, but the density of star–forming galaxies at z > 7 is essentially unknown at present. |
Processed_From_gas_to_galaxies.txt | SKA will be a sensitive probe of high-redshift star formation through the detection of radio continuum emission from star–forming regions and the radio afterglows of gamma–ray bursts (GRBs), many of which are now known to arise from the supernova explosions of massive stars. |
Processed_From_gas_to_galaxies.txt | Models for dwarf galaxies forming in the vicin- ity of high–redshift QSOs imply likely star– formation rates of 30–60 M⊙/yr for these ob- jects [67]. Deep continuum surveys at 1.4 GHz with SKA will reach detection limits of 100 nJy in 8 hr, and as low as 10 nJy in longer integra- tions of ∼1000 hr [86]. This should allow objects with star–formation rates as low as 10 M⊙/yr to be detected at redshifts up to z∼4 (8 hr) and z∼11 (in 1000 hr) respectively (see Fig. 1). Indi- vidual globular clusters forming at high redshift may also be detectable by SKA, though there are currently few available predictions for the star– formation rates in such objects. |
Processed_From_gas_to_galaxies.txt | will be very high (∼200,000 deg−2), with “nor- mal”, starburst and active galaxies all contribut- ing over a wide range in redshift. Simulations by Hopkins et al. [46] imply that very high–redshift (z > 5) AGN and starburst galaxies will have sur- face densities of ∼1,000 deg−2 and ∼7,000 deg−2 respectively in surveys with a detection limit of 100 nJy at 1.4 GHz. |
Processed_From_gas_to_galaxies.txt | Although star–forming galaxies dominate the radio source counts at flux densities below about 1 mJy, active galaxies still account for a signif- icant fraction of the population at levels of a few µ Jy [39]. The problem of distinguishing star–forming galaxies from AGN is discussed by Garrett et al. [35]. Several parameters mea- surable by the SKA itself (including brightness temperature, morphology, time variability and possibly polarization) can be used, and Garrett et al.[35] show that radio continuum observations with high spatial resolution (∼ 25 mas) can reli- ably distinguish AGN from star forming galaxies at faint flux levels. This would require the SKA to have good sensitivity on baselines of at least 3000 km. High spatial resolution is also required because of the likely source density at flux levels of ∼10 nJy, which may be as high as 3 arcsec−2 [86]. |
Processed_From_gas_to_galaxies.txt | are likely to become increasingly common probes of high–redshift star formation up to (and per- haps even within) the epoch of reionization. |
Processed_From_gas_to_galaxies.txt | In this section, we examine the assumptions underlying the use of the radio continuum to measure SFR, and assess their validity at the high redshifts where SKA will be able to make a unique contribution. We also outline new sim- ulations which will enable us to better interpret the planned deep surveys. |
Processed_From_gas_to_galaxies.txt | Figure 2. 1.4 GHz radio luminosity (log) plot- ted against total far–infrared luminosity (taken from Bell 2003 [8]), showing the curvature at low luminosities. Normal star–forming galaxies are plotted as circles, intensely star–forming galaxies as stars (filled stars are ULIRGS, open stars are starbursts). Triangles are BCDs from a compari- son sample. The dotted and solid lines represent forward and bisector fits to the data. The thick dashed line shows the trend predicted by the final star–formation calibrations of Bell 2003 [8]. |
Processed_From_gas_to_galaxies.txt | on poorly understood physics of cosmic-ray con- finement and diffusion operating over 107−108 yr, radio continuum surveys have been used to con- strain co-moving SFR locally, and at redshifts of up to z∼1–2 [25,64,84,38]. |
Processed_From_gas_to_galaxies.txt | tributions is underestimated by a factor of five or more [59,5,48]. While rare locally (see [60]), low-metallicity low-mass galaxies might be much more frequent at early times and high redshifts, given the predictions of the hierarchical merger models (e.g. Baugh et al. (this volume), [19]. In- deed, such objects may represent the primordial “building blocks” – or “sub-galaxies” [71] – in hi- erarchical scenarios of galaxy formation. Conse- quently, at redshifts z > 3, it is possible that the canonical scaling of SFR with radio continuum does not apply. Such a systematic mis-estimate of SFR could strongly skew our interpretation of deep radio continuum surveys. |
Processed_From_gas_to_galaxies.txt | The slope inflection of the FIR-R correlation at low luminosities probably stems from a num- ber of factors involving both the radio continuum and the infrared SED. Moreover, there is some evidence that both the radio continuum and FIR underestimate the SFR at low luminosities [8]. |
Processed_From_gas_to_galaxies.txt | Figure 3. The integrated star–formation history of the Universe, from [40]. The SFR calculated from the “fossil record” in the Sloan Digital Sky Survey (SDSS) is shown by the eight large filled circles. The other symbols correspond to inde- pendent determinations using instantaneous mea- surements of the star formation rate at various wavelengths. |
Processed_From_gas_to_galaxies.txt | thermal radio emission is virtually absent [72]. The SFR estimated by assuming a purely ther- mal spectrum is roughly ten times higher than inferred from a “standard” mainly non-thermal spectrum [22]. |
Processed_From_gas_to_galaxies.txt | Moreover, optically thick HII regions are very likely to be prevalent in young starbursts [59,5,6]. This means that instead of a flat thermal spec- trum, we would see a rising or highly-absorbed spectrum at ∼ 1 GHz, which would again alter the standard assumptions. Such an effect would also be expected to vary with redshift. |
Processed_From_gas_to_galaxies.txt | Figure 4. Local (z < 0.1) star–formation density for galaxies with star–formation rates between 0.01 and 1000 M⊙ yr−1 [73]. The solid line is de- rived from a fit to the local radio luminosity func- tion for star–forming galaxies, while filled squares show values derived from optical measurements of the Hα emission line [34]. In the local universe, the most likely place for a star to form appears to be in a galaxy with a star–formation rate around 10 M⊙ yr−1. The integral under this plot is the zero-point of the ’Madau diagram’ which plots the integrated star–formation density at different redshifts. |
Processed_From_gas_to_galaxies.txt | sion must be produced by a different mechanism (e.g. evolved supernovae expanding in a dense medium [18,2]; young supernova remnants; etc.). In any case, the inescapable implication of the radio spectral differences as a function of lumi- nosity, metallicity, or age is that the conversion of radio continuum to SFR may change as a func- tion of redshift. While this is probably not an is- sue with current facility surveys which probe only a limited redshift regime, deep SKA surveys will be able to detect not only the most luminous ra- dio sources at each redshift, but also adequately sample the statistically dominant “normal” star- forming population out to redshifts of 10 or so. Changes in the conversion factors as a function of redshift need to be considered when mapping cosmic star formation. |
Processed_From_gas_to_galaxies.txt | The integrated star-formation history of the universe (the Madau diagram; see Fig. 3) is likely to be accurately mapped out fairly soon. This tells us how much star–formation occurred at a given cosmic epoch, but to understand galaxy In evolution we need to know more than this. particular, where and in what range of environ- ments, did star formation occur at each epoch? Did most stars form in galaxies undergoing “qui- escent”star formation, or in galaxies undergoing violent starbursts, and how does this vary with redshift? How important a role do mergers play in triggering star formation? In the local uni- verse, star formation occurs in a wide range of en- vironments, and in galaxies with star–formation rates ranging over at least five orders of magni- tude (Fig. 4). There is no reason to suppose that the high–redshift universe is any less diverse. |
Processed_From_gas_to_galaxies.txt | nous star burst galaxies at almost any redshift and has the great advantage of a much larger field of view and hence superior survey capabilities. Moreover, unlike ALMA, the sensitivity of SKA to star formation does not depend on the pres- ence of metals. Since the first stars must form in a primordial metal-free environment, molecu- lar clouds would not be expected to contain CO or other molecules beyond molecular hydrogen. Hence SKA will provide a much ”cleaner” less bi- ased view of high-redshift star formation. |
Processed_From_gas_to_galaxies.txt | SKA has the unique ability to trace the trans- formation of HI gas into stars (and into the en- riched hot gas which future X–ray missions will detect) in a wide range of environments and over a large fraction of cosmic time. The fact that HI observations trace the depth of a galaxy’s po- tential, and so can relate star–formation to halo mass as well as to luminosity, is also important because it allows direct tests of galaxy evolution models which deal with mass as well as light. |
Processed_From_gas_to_galaxies.txt | Before discussing how SKA will contribute to probing the gas content of the universe over cos- mic time we will briefly describe its capabilities in terms of sensitivity and resolution to provide a baseline for H I imaging projects. H I line observa- tions typically require a brightness sensitivity of a few Kelvin, which in practice limits the angular resolution. For SKA this implies the following: if one takes as a guideline a Tb sensitivity of ∼ 1 Kelvin in a 0.5′′ beam at 1.4 GHz (so baselines out to at most a few hundred km), then one can detect L∗ galaxies out to a redshift of z = 1.5 in a 12 hour integration. Although the current design goals require a field of view of one square degree at 1.4 GHz, a much wider field of view ap- pears technically feasible, and would provide an impressive gain in survey speed (see also Blake et al. this volume). |
Processed_From_gas_to_galaxies.txt | Figure 5. Comoving H I mass density in neutral gas contributed by damped Lyα absorbers plot- ted as a function of redshift for a cosmology with ΩM = 0.27, and ΩΛ = 0.73 [83]. The region Ωstar is the ±1σ range for the mass density in stars in nearby galaxies [33]. The point at z = 0 is the value inferred from 21 cm emission from local galaxies [103]. |
Processed_From_gas_to_galaxies.txt | tions assume H0 = 71 km sec−1 Mpc−1, ΩM = 0.27, and ΩΛ = 0.73. The calculations followed the SKA strawman design figures: Arme/Tsys = 20000, 70% of the collecting area is within 100 km and assumed two polarizations. |
Processed_From_gas_to_galaxies.txt | It is clear from Table 1 that it is possible to de- tect large galaxies in H I emission beyond a red- shift of z = 1 fairly easily, though with limited resolution. For example a galaxy like the Milky Way can be detected out to a redshift of about 1 in a 12 hour integration. Its companions, the Magellanic Clouds, can still be seen out to red- shifts of a few tenth. Large spiral galaxies such as M 101 can be probed much farther, to z ≈ 2.5. |
Processed_From_gas_to_galaxies.txt | Figure 8. Atomic Hydrogen detected in Abell 2192 at z=0.1887. Upper panels: individual channel maps from the VLA datacube. The rms noise and Heliocentric velocity are noted above each panel. Lower left: Total H I map. The contours coincide with the position of an inclined barred late-type galaxy. Lower middle: contours from the three channel maps plotted on top of each other; solid: cz=56483 km/s, dashed: cz=56542 kms, dotted: cz=56600 km/s. Lower right: Hα velocity field of the optical counterpart obtained with the PMAS IFU spectrograph at the 3.5m telescope on Calar Alto. There is excellent correspondence between the H I and Hα recession velocities, confirming the H I detection. From Verheijen and Dwarakanath (in preparation). |
Processed_From_gas_to_galaxies.txt | is shown not only by the latest estimates of the change in comoving star formation density with redshift [40] but also by the most recent estimates of the evolution in gas density with redshift from damped Lyα studies [83,37]. |
Processed_From_gas_to_galaxies.txt | Damped Lyα studies [83] indicate that the co- moving H I mass density is a few times the present value beyond z = 1 and out to z = 3 as illustrated in Fig. 5. The evolution in comoving H I mass density as shown in Fig. 5 is still dependent on assumptions. A recent study of the column den- sity distribution function of H I from observations of nearby galaxies by Zwaan et al. [106] suggests that there is not as strong an evolution in comov- ing H I mass density as suggested in Fig. 5. SKA, however, will be able to measure the H I emission in galaxies back to redshifts of z ≈ 3 directly and will revolutionize this area of research. |
Processed_From_gas_to_galaxies.txt | which galaxies assemble over cosmic time. Since the evolution of comoving H I gas density as de- scribed above has not been directly measured, but has been derived from H I column density mea- surements of Damped Lyα absorbers it is crucial to be able to image the entire H I disks associated with these Damped Lyα absorbers. If the associ- ated H I masses are indeed of the order of M∗ H I = 6.2 × 109 M⊙ [104,105,106] then SKA can image these disks out to z = 1.4 in integration times of 12 hours. Such observations will be crucial for linking the existing column density measurements to the local H I density from observations of the emission. |
Processed_From_gas_to_galaxies.txt | by measuring the H I mass function over a range of redshifts and in different environments one can probe the effects of galaxy evolution. The mod- els of Baugh et al. (this volume) also indicate changes in the H I mass function (see Fig. 6) which SKA will be able to determine. |
Processed_From_gas_to_galaxies.txt | Out to z ≈ 0.3 one will be able to resolve a fair fraction of the galaxies so one can obtain rotation curves, mass distributions, and gas fractions for more than ≈ 104 galaxies between now and 3 Gyr ago. One then is in a unique position to trace the evolution of the ISM in galaxies over a substantial fraction of the age of the universe, from the era of strongest evolution and star formation activity until the present. In addition one would learn whether and how the evolution depends on the dark matter content and environment. |
Processed_From_gas_to_galaxies.txt | The hierarchical galaxy formation simulations suggest that most of the evolution in the HI mass function is in the mass range MH I = 4 × 108 – 4 × 109 M⊙and within the last seven billion years. Fig. 6 shows that at z = 1 one expects ∼ 3 times more galaxies in this mass range than at z = 0. This figure also illustrates that there is very little evolution in the high mass ( MH I > 4× 109 M⊙) end of the H I mass function. This is the range where the SKA is especially sensitive and can probe easily to at least z = 1, so that this prediction can be verified. |
Processed_From_gas_to_galaxies.txt | there is in galaxies. Since in this picture the size of the filaments that feed the galaxies grow from galaxy size at high redshift, to hundreds of kpc size structures feeding entire groups at low red- shifts, deep H I surveys should be able to verify this idea. If on the other hand mergers of pre- dominantly small systems are the main process one expects to find a very different distribution of H I structures with redshift. |
Processed_From_gas_to_galaxies.txt | In addition, as detailed in the previous section, one can use the continuum emission to probe the star formation rates of the detected galaxies, in- dependent of the effects of extinction using the radio continuum-FIR correlation [42,41,21,22,78] and test the model predictions. In addition, this information can be used to link the star forma- tion rates to the H I contents of galaxies as a func- tion of redshift and environment. The global star formation rates can be compared with the semi- analytic models and the cool gas accretion models in order to test the predictions. That environ- ment plays an important role is suggested by the analysis of Sloan Digital Sky Survey (SDSS) data [53,52,45,44] which shows a dependence of galaxy mass and star formation rate on environment. |
Processed_From_gas_to_galaxies.txt | Figure 6. The evolution of the H I mass function with redshift in the Baugh et al. (this volume) model. The points show the recent determina- tion using HIPASS galaxies by Zwaan et al.[106]. The curves show the model predictions at differ- ent redshifts (solid: z=0 ; dotted: z=1 ; dashed: z=3, dot-dashed: z=4). |
Processed_From_gas_to_galaxies.txt | provided from existing observations, both optical and H I imaging. Zaritsky [101] and Zaritsky & Rix [102] studied several tens of galaxies and as- cribed the coincidence of evidence for recent star formation from stellar population studies with measurable asymmetries in the galaxy shapes, to recent accretion events. Kinematic lopsidedness, observed in the 2 dimensional H I velocity fields of galaxies [90,85] has also been considered as a re- sult of recent minor mergers. Furthermore there are at least twenty examples of galaxies which in H I show either signs of interactions and/or have small companions [75,76,87]. This suggests that galaxies often reside in an environment where ma- terial for accretion is available. An example is given in Fig. 9 which shows a new H I image of the galaxy NGC 3359 which is accreting gas from a faint H I companion. |
Subsets and Splits