A cta A stronom ica 000,000{000 (0000)

P rinted 17 D ecem ber 2013

( LA TEX )

E xtinction in the G alaxy from surface brightnesses of E SO -LV galaxies:testing ’standard’extinction m aps

arXiv:astro-ph/0309751v1 27 Sep 2003

Jacek C holoniew ski,1 Edw in A .Valentijn2 1 A stronom 2K

icalO bservatory of W arsaw U niversity,A leje U jazdow skie 4, P L-00478 W arsaw , P oland, e-m ail: astro@ estym ator.com .pl apteyn Institute, P . O . B ox 800, N L-9700 AV G roningen, T he N etherlands, e-m ail: valentyn@ astro.rug.nl

17 D ecem ber 2013

A B ST R A C T

T he relative extinction in the G alaxy com puted w ith our new m ethod (C holoniew ski & Valentijn 1999,C V )iscom pared w ith three patterns:Schlegel,Finkbeiner & D avis (1998,SFD ), B urstein & H eiles (1978,B H ) and the cosecans law .It is show n that extinction ofSFD is m ore reliable then that ofB H since it stronger correlates w ith our new extinction. T he sm allest correlation coe cient have been obtained for the cosecans law .Linear regression analysis show that SFD overestim ate the extinction by a factor of1.4. O urresultsclearly indicate thatthere isnon-zero extinction atthe G alactic South poleand thattheextinction neartheG alacticequator(jbj< 40o)issigni cantly larger in the Southern hem isphere than in the N orthern. K ey w ords: dust, extinction - m ethods: statistical - G alaxy: general - galaxies: fundam entalparam eters

1

IN T R O D U C T IO N

W e have introduced in our previous paper (C holoniew ski & Valentijn 1999,hereafter C V ) a new m ethod for the determ ination of the extinction in our G alaxy. T he m ethod usessurface brightnessesofexternalgalaxies in the B and R bands as listed in T he Surface Photom etry C atalogue ofthe ESO -U ppsala G alaxies (Lauberts & Valentijn, 1989, hereafter ESO -LV ).T he rstdraftofthism ethod hasbeen published in our earlier paper (C holoniew ski& Valentijn 1991). T he m ain purpose of the present paper is to com pare our derived extinction values w ith recently published m aps ofextinction by Schlegel,Finkbeiner & D avis (1998),hereafter SFD , and w ith the frequently used m ap of B urstein & H eiles (1978), hereafter B H (see also B urstein & H eiles 1982).

2

T H E M ET H O D

O ur extinction determ ination (fully described in C V ) em ploys the surface brightnesses of externalgalaxies in the B and R bands: B , R .B asically, our m ethod produces the relative extinction com pared to an overallm ean -extinction w ith an unknown zero-point. T he form ula for the relative extinction (in B band) is sim ply a linear com bination of B and R : AB =

B

1

s R rs

c:

(1)

In order to use equation (1) one has to know three param eters:r,s and c. T he param eter r describes the ratio of extinction in the R and the B band (r = A R =A B ) and is assum ed to be constant.Its recent literature value is 0.61 (see C V for references) w hile w e derived in C V tw o new estim ates:0.62 and 0.64.A sa reasonable com prom ise w e adopt throughout this paper r = 0:62. T he inverse ofthe param eters describesthe slope ofthe linearrelation betw een surface brightnesses B and R .T he c param eter is introduced in order to m aintain zero-point issues. B oth param eters s and c depend on m orphological type T ,so they have been com puted (using equations 7 and 10 in C V ) separately for every m orphologicaltype. In this paper w e w illconsider severaldi erent subsam ples.Extinction w ithin every such subsam ple hasbeen com puted using the set of values s(T ) and c(T ) obtained from the sam e set ofdata.T he s(T ) and c(T ) coe cients for the m ost im portant tw o subsam ples used in ths paper are in Table 1. W e use in this paper the extinction in B band as described in equation (1) and denote it as A B (C V ).

3

T H E SA M P L E S

For our analysis w e use surface brightnesses atthe radius of halftotalB lightin B and R bandsfrom ESO -LV .W eexclude from the sam ple those galaxies w hich have m orphological

2

J.C holoniewski,E.A .Valentijn

T able 1.C oe cients s(T ) and c(T ) used for com puting relative extinction in this paper according to equation 1 sam ple "A "

G eneralconditions:

sam ple "B "

T

s(T )

c(T )

s(T )

c(T )

-5 -4 -3 -2 -1 0 1 2 3 4 5 6 7 8 9 10

0.826 0.955 0.915 0.893 0.785 0.768 0.728 0.691 0.670 0.652 0.635 0.641 0.717 0.731 0.744 0.795

10.706 6.263 7.619 8.295 11.597 12.060 12.972 13.903 14.223 14.589 14.979 14.943 13.163 12.844 12.656 11.054

0.842 0.997 0.932 0.894 0.772 0.840 0.737 0.681 0.668 0.650 0.637 0.697 0.714 0.721 0.784 0.806

10.233 4.199 6.927 8.306 12.060 10.129 12.885 14.166 14.321 14.663 14.943 13.627 13.251 13.181 11.446 10.675

classi cationssuspected to beextinction dependent(m arked in ESO -LV w ith Tf lag equalto 4). W e also reject galaxies w hich have excessive (probably w rong) colours: B R sm aller than zero or greater than 2.1. Since w e focus in this paper on the com parison of our extinction w ith SFD and B H data, w e reject additionally those galaxies for w hich B H or SFD extinction values are not available. T he largest com plete subsam ple ofgalaxies ofESO -LV can bem adeby selecting galaxiesw hich havevisualapparent diam eter D org greater or equal to 60 arcsec. It contains, after applying the rejections described above,7974 galaxies (sam ple "A "). A s m entioned in C V , the com pleteness lim it of ESO LV galaxies is m orphologicaltype dependent and a universalcriterion,valid for every m orphologicaltype,is selecting galaxies w ith a visualdiam eter lim it larger than 100 arcsec. W e have used such m ore restricted sam ple for allthe com putations presented in C V . W e use it also here as sam ple "B " (after rejections described above). R esults described in Sections 4 and 5 show that m ore rigorousely de ned sam ple "B " producesslightly m ore accurate extinction per galaxy than the sam ple "A ".H ow everthe sam ple "A " is three tim es larger than the sam ple "B " w hat, at least in part,com pensate its slightly larger dispersion. Table 2 containsa sum m ary ofthe de nitionsofsam ple "A " and sam ple "B ". Since the ESO -LV galaxy catalogue covers the Southern sky ( < 17:5o ) our analysis refers to this part ofthe hem isphere.

4

T able 2.D e nitions ofthe galaxy sam ples.

A B (B H ) - present A B (S F D ) - present Tf lag 6 = 4 0 2:1 B R sam ple A D org

60 arcsec

N = 7974

sam ple B D org

100 arcsec

N = 2450

by the so called one-w ay analysis ofthe variance theory (see e.g.Fish 1962).T hisaverage variance re ectsthe variations ofthe true extinction inside the squaresw ith size and the standard erroroftheextinction estim atorexpressed in equation (1). So,w hen the size of the squares tends to zero the average variance of the extinction should tend to the standard error ofthe extinction produced by our m ethod. Figure 1 show s the average variance ofextinction as a function of .T he m inim um value for w hich w e applied w as one degree.For sm aller values of a too large fraction of studied squares contain only one galaxy (such squares can not be taken into account in the one-w ay analysis of variance). A s w e expect, the average variance has m inim um for the sm allest applied value of (one degree). For sam ple "A " the standard error ofour extinction in A B is 0.43 m agnitude,w hile for sam ple "B " this is 0.40 m agnitude,w hich corresponds to a standard error in E (B V ) of approxim ately 0.10 m agnitude (for A B =E (B V ) = 4:3).Sam ple "B " produces relative extinction w ith slightly higher accuracy than sam ple "A ". B H used for the calibration oftheir extinction m ap B V colours of131 globular clusters and R R Lyrae stars.SFD used for the calibration oftheir extinction m ap B -R colours of 106 brightest cluster ellipticals and B -V colours of 389 ellipticalgalaxies w ith m easured M g2 index (505 objects in total). B oth B H and SFD report that their calibrators show a residualscatterin B -V ,w ith respectto calibration regression line,ofapproxim ately 0.03 m agnitude. O urextinction estim ator,w hen applied to photographically m easured surface brightnesses ofgalaxies in tw o bands as listed in the ESO -LV catalogue, has three tim es larger standard error than the calibrators used by B H and SFD . T his is an im portant disadvantage (at least as long as w e apply it to ESO -LV photom etrical data).B ut there is one im portant advantage of our extinction estim ator - it can be applied to m any m ore objects since w e can apply the m ethod,at present,to 7974 galaxies (sam ple "A ") from the ESO -LV catalogue.

TH E ACCU RACY

Equation (1) represents an estim ator ofthe foreground relative extinction for an individualexternalgalaxy.It is im portant to know its uncertainty (standard error). In order to obtain this,w e have divided the w hole sky into squares w ith size degrees and com puted the average variance of the extinction inside these squares using the form ula given

5

T H E C O R R E L A T IO N

W e have com puted the Pearson,Spearm an and K endallcorrelation coe cients(seePress,Teukolsky,Vetterling & Flannery 1992 for de nitions and softw are) betw een our extinc-

Extinction in the G alaxy from surface brightnessesofESO -LV galaxies:testing ’standard’extinction m aps T able 3.Pearson,Spearm an and K endallcorrelation coe cients betw een A B (C V ) extinction estim ate and three other extinction estim ates:SFD ,B H and csc(b). Pearson

Spearm an

K endall

0.234 0.219 0.188

0.158 0.148 0.127

0.304 0.282 0.252

0.208 0.192 0.170

m inim um value of A B (B H ) extinction.T his is in approxim ate agreem ent w ith SFD w ho discovered a sim ilar o set of 0.09 m agnitude. W e use in the regression analysis presented in the next Section the corrected B H ’s extinction: A B (B H )C = A B (B H )+ 0:12 instead ofA B (B H ) itself.

Sam ple "A " SFD BH csc(b)

0.276 0.247 0.202

7

Sam ple "B " SFD BH csc(b)

0.327 0.293 0.238

tion and extinction given by SFD and B H and for the cosecans law : A B = A 0 cscjbj;

(2)

w here b denotes G alactic latitude. T he com putation have been perform ed for sam ple "A " and sam ple "B " (see Table 3).T he param eters s(T ) and c(T ) have been com puted separately for every sam ple (see Table 1). A ll three correlation coe cients for both sam ples are the largest for SFD extinction and the sm allest for the cosecans law .T he extinction ofB H is alw ays betw een these tw o extrem e results. T he coe cients are generally higher for sam ple "B " than for sam ple "A " w hat suggests that sam ple "B " produces m ore accurate extinction than sam ple "A ". Since the sam ple "B " is a subsam ple ofthe sam ple "A " the correlation coe cients for "A " and "B " are not statistically independent.In order to produce a set ofstatistically independent correlations w e divide the sam ple "A " into six subsam ples: (i) 60arcsec D org < 80arcsec (N = 3876) (ii) 80arcsec D org < 100arcsec (N = 1648) (iii) 100arcsec D org < 120arcsec (N = 718) (iv) 120arcsec D org < 140arcsec (N = 647) (v) 140arcsec D org < 160arcsec (N = 298) (vi) 160arcsec D org (N = 787) , and com pute the param eters s(T ) and c(T ) separately for every subsam ple.A s a result ofthis procedure w e have six statistically independentcorrelation coe cients-see Figs 2, 3 and 4 forresults.A sbefore,the correlationsare the largest for SFD ,sm aller for B H and the sm allest for cosecans law .

6

C O R R E C T IO N O F T H E B H ’S E X T IN C T IO N Z E R O P O IN T

T he form ulae of B H produces for som e regions of the sky extinction less than zero (for the sam ples analyzed in this paper the m inim um value of A B (B H ) extinction is -0.12 m agnitude). In spite of B H ’s instructions to set these values to zero w e have actually used these negative values and found that A B (C V ) is for them signi cantly sm aller than for A B (B H ) 0 (see upper panel of Figs 7 and 8). T his m eans that the B H extinction (in the B band) w as underestim ated by 0.12 m agnitude - just the absolute value ofthe

T H E L IN E A R R E G R E SSIO N

In theidealcase therew ould belineardependencew ith slope equalto one betw een our relative extinction estim ate (C V ) and the extinction ofSFD and B H (corrected). Since our extinction values are relative, w ith an arbitrary zero-point,the constantterm ofthislineardependence should notbe equalto zero.T he constant,m ultiplied by 1, should be added to ourrelative extinction to transform it to the absolute extinction. W e have tted the straightlines(using the leastsquares m ethod) taking as independent variables extinction ofSFD and B H and as dependent variable our estim ate ofrelative extinction (C V ).W e have found the follow ing regression coe cients for sam ple "A ": A B (C V )= 0:662 ( 0:028)A B (SF D )

0:189 ( 0:010)

(3)

A B (C V )= 0:555 ( 0:026)A B (B H )C

0:176 ( 0:010)

(4)

A B (C V )= 0:741 ( 0:043)A B (SF D )

0:225 ( 0:016)

(5)

A B (C V )= 0:608 ( 0:040)A B (B H )C

0:206 ( 0:017)

(6)

and for sam ple "B ":

w here in brackets 1 errors are given. T he di erences betw een slopes and constant term s com puted for sam ple "A " and "B " are only m arginally larger than the com bined errors.In the forthcom ing w e use their averages. A graphical presentations of the regression lines are show n in Figs 5-8.In order to investigate w hether the postulate about linear dependence is valid,the regression lines are show n together w ith row data (low er panels) and w ith averages of A B (C V ) com puted for 0.05 m agnitude bins of A B (SF D ) and A B (B H ) (upper panels). T he slopesofthe regression linesare de nitely lessthan one:forSFD the slope iscirca 0.7 w hile forB H itiscirca 0.6. T he constantterm sforboth SFD and B H are approxim ately the sam e: -0.20 m agnitude. T his de nes the zero point of ourresults:adding 0.20 m agnitude to ourrelative extinction transform s it to the absolute extinction. W hen w e introduce our absolute extinction into equations 3 - 6,the constant term s in all4 equations gets close to zero,now setting the relation betw een our (absolute) extinction and SFD ’s and B H ’s extinction.T hus w e conclude: SFD overestim ate extinction by a factorof0:7 1 1:4 w hile B H by a factor of0:6 1 1:7. U p to now ve papers reported sim ilar results,nam ely: that SFD extinction is about 1.4 tim es too large. Stanek (1998b)obtained,using coloursoflow galactic latitude globular clusters, that the overestim ation factor is 1.35 (see also Stanek 1998a). A rce & G oodm an (1999a, 1999b) analyzed extinction in the Taurus region using four independent m ethods and found that the factor is 1.3 -1.5.G onzalez,Fruchter & D irsch (1999) obtain extinction in the sm all eld around G R B 970228. T hey found using tw o m ethods A V = 0:55 w hile A V (SF D ) = 0:78.T he ratio ofthese tw o

3

4

J.C holoniewski,E.A .Valentijn

num bersgive again a factor 1.4.O ur nding thatSFD overestim ate extinction is also supported by Ivans et al.(1999) and von B raun & M ateo (2001). Som e kind ofdoubtaboutthecorrectnessofthe calibration is also in originalSFD paper w here w e can nd as the com m ent to their Fig.6 the statem ent that "A slight trend in the residuals is evident for both B H and D IR B E/IR A S corrections, in the sense that the highest reddening values appear to be overestim ated". Inspecting Figs 5-8 dem onstrate a generallineardependence betw een our relative extinction estim ate and the extinction ofSFD and B H .H ow ever,som e m inor,but statistically signi cant,deviations are clearly present,especially for A B (SF D ) and A B (B H ) less than 0.5 m agnitude.

8

G A L A C T IC L A T IT U D E D E P E N D E N C E

U nfortunately,w e can not use our m ethod ofextinction determ ination,as applied to the ESO -LV galaxy catalogue,to produce a new high resolution extinction m ap and to com pare it w ith SFD and B H m aps. T his is because our data are too sparse (less than one galaxy per square degree) and their accuracy is too low ( (E (B V )) 0:10 per galaxy) m ainly because ofinstrinsic scatterofthe surface brightness ofgalaxies. B ut the quality ofour data is good enough to evaluate theirgalactic latitude dependence.W e show thisdependence for our absolute extinction (equalto the relative extinction, asde ned in equation 1,plus0.20 m agnitude)togetherw ith SFD ,corrected B H and thecosecanslow -see Fig.9.A llfour solutions have been averaged inside 5 degrees galactic latitude bins in the points w here ESO -LV galaxies are (sam ple "A ").W e use A o = 0:10 for the cosecans law (equation 3) to m atch as close as possible to our solution. Fig. 9 show s that our absolute extinction is alw ays greaterthan zero con rm ing thevalueofthezero-pointcom puted in Section 7. A sone can see in Fig.9 oursolution asw ellasSFD and B H m im ic quite w ell the classic cosecans law and indicate considerable extinction near the galactic pole. T he di erences betw een our results and SFD and B H are the largest near the galactic equator and at the northern hem isphere. Since w e nd that both SFD and the (corrected) B H extinction standards overestim ate G alactic extinction by a factor 1.4 and 1.7 respectively, w e show the sam e data as in Fig.9 but w ith SFD and corrected B H extinction divided by these factors (or,equivalently,m ultiplied by 0.7 and 0.6 respectively) - see Fig.10.T he better agreem ent ofour solution w ith the rescaled SFD and B H data is evident, especially near the G alactic equator (b 0),con rm ing the need for rescaling SFD and B H extinction.B ut even for the rescaled data statistically signi cantdi erencesbetw een our results and SFD and B H ,although sm aller, still exist.Especially our data exhibit a South - N orth asym m etry near the galactic equator (jbj< 40o ) w ith m ore extinction in the Southern G alactic hem isphere w hich is not visible in SFD and B H data.

9

D ISC U SSIO N

O urextinction estim ator (introduced in C V )de nitely indicates that the new extinction m ap ofSFD is m ore realiable than the old one ofB H .T he historicalcosecanslaw isin this com petition on the last place.T hisresult hasbeen obtained using six separate,statistically independent,sets ofdata. T he superiority of the SFD extinction m ap over the B H m ap have been dem onstrated by com puting correlation coe cients w ith our extinction results. H ow ever, the correlation coe cients are not sensitive to the am plitude of variation ofthe inputdata nor its zero-point.In order allow for further, m ore speci c, com parisons w e have perform ed a linear regression analysis betw een our relative extinction and SFD ’s and B H ’s extinction. T his analysis provides us the zero-point of our extinction (equalto 0.20 m agnitude) w hich allow s usto transform ourrelative extinction into absolute extinction. W e have show ed that, in com parison to our absolute extinction,SFD overestim ate extinction by a factor of 1.4. T his is in agreem ent w ith ve other authors and considerably changes our view on the am plitude ofthe G alactic extinction. Superiority of SFD over B H reddening m ap,as w e report in this paper,does not m ean that the rst is an ideal result - w e have discovered som e signi cant di erences betw een ourextinction and SFD w hen analyzing theirG alactic latitude dependence even aftercorrecting SFD extinction by dividing it by 1.4 factor. O ne possible source of these differences is that SFD assum e that the extinction is strictly proportionalto the dust colum n density w hat need not be true - the extinction to dust ratio can vary across the sky and m ay also depend on the w avelength. O ur results are im portant for creating reliable extinction standards and dem onstrate the correctness and usefulness of our extinction estim ator. T hey can be also regarded as a stim ulus for applying our extinction estim ator to other,largerand m ore accurate galaxy catalogues,particularly thoose expected from new w ide eld im aging surveys, Future applications of our m ethod to other databases m ay produce m aps w ith a courser resolution.A tpresentthe m ethod applied to ESO -LV data is unable to generate the extinction m ap w ith both su cientresolution and accuracy. O ur present m ethod should be rather taken as new extinction calibrator. B ut using the m ethod to larger and m ore precise galaxy catalogues m ay result into stand-alone high resolution extinction m aps.

A C K N O W LE D G M E N T S T hepartofthisw ork w asdoneduring thestay oftheauthors at European Southern O bservatory (G arching, G erm any). T he authorsthanksA ndrisLaubertsforcollaboration about the project.

R EFER EN C ES A rce,H .G .,G oodm an,A .A .1999a,A pJ,512,L135 A rce,H .G .,G oodm an,A .A .1999b,A pJ,517,264 von B raun,K .,M ateo,M .2001,A J,121,1522 B urstein,D .,H eiles,C .1978,A pJ,225,40 (B H ) B urstein,D .,H eiles,C .1982,A J,87,1165

Extinction in the G alaxy from surface brightnessesofESO -LV galaxies:testing ’standard’extinction m aps C holoniew ski,J,V alentijn,E .A .1991,M essenger,63,1 C holoniew ski,J,V alentijn,E .A .2003,astro-ph/0309750 (C V ) Fisz, M . 1963, P robability T heory and M athem atical Statistics. John W iley & Sons,N ew Y ork,London G onzalez,R .A .,Fruchter,A .S.,D irsch,B .1999,A pJ,515,69 Ivans,I.I.,Sneden, C .,K raft,P.,Suntze , N .B .,Sm ith,V .V ., Langer,G .E .,Fulbright,J.P.1999,A J,118,1273 Lauberts,A .,V alentijn,E .A .1989,T he Surface P hotom etry C atalogue ofthe E SO {U ppsala G alaxies,E uropean Southern O bservatory,G arching P ress,W .H .,Teukolsky,S.A .,V etterling,W .T .,Flannery,B .P. 1992,N um ericalR ecipes,C am bridge U niversity P ress Schlegel,D .J.,Finkbeiner,D .P.,D avis M .1998,A pJ,500,525 (SFD ) Stanek,K .Z.1998a,astro-ph/9802093 Stanek,K .Z.1998b,astro-ph/9802307

5

6

J.C holoniewski,E.A .Valentijn

average σ(AB(CV)) [mag.]

.5

sample A

.45

sample B .4

0

.5

1

1.5

log(∆) [deg.]

F igure 1.T he average variance ofextinction A B (C V ) inside degrees squares on the sky as a function ofthe size ( ) ofthese squares. T he m inim um variance (at one degree:log( ) = 0) represent the standard deviation ofthe estim ator ofrelative extinction used in this paper.

Extinction in the G alaxy from surface brightnessesofESO -LV galaxies:testing ’standard’extinction m aps

correlation coefficient with AB(CV)

.5

.4 Pearson

SFD

.3

BH csc(b)

.2

.1

0 50

100

150

200

Dorg

F igure 2. Pearson correlation coe cient betw een our A B (C V ) extinction and the extinction of SFD and B H and the cosecans law com puted for six subsam ples de ned using visualdiam eter Dorg (see Section 5 for details)

7

J.C holoniewski,E.A .Valentijn

.5

correlation coefficient with AB(CV)

8

.4 Spearman

SFD

.3

BH csc(b)

.2

.1

0 50

100

150

200

Dorg

F igure 3.T he sam e as Fig.2 but for Spearm an correlation coe cient.

Extinction in the G alaxy from surface brightnessesofESO -LV galaxies:testing ’standard’extinction m aps

correlation coefficient with AB(CV)

.5

.4 Kendall .3

SFD

.2

BH csc(b)

.1

0 50

100

150

200

Dorg

F igure 4.T he sam e as Fig.3 but for K endallcorrelation coe cient.

9

10

J.C holoniewski,E.A .Valentijn

sample A 1.5

N=7974 slope: 0.66

AB(CV)

1

.5

0

-.5 0

.5

1

1.5

2

1.5

2

AB(SFD)

AB(CV)

2

0

-2 0

.5

1 AB(SFD)

F igure 5. T he dependence betw een the extinction of SFD A B (S F D ) and our relative extinction A B (C V ). U pper panel show s the averaged data inside 0.05 m agnitude bins.E rrorbars represent standard deviation (1 ).Low er panelshow s raw data.T he data are taken from sam ple "A " (see text).D otted straight lines visible on both panels represent the least squares t.

Extinction in the G alaxy from surface brightnessesofESO -LV galaxies:testing ’standard’extinction m aps

sample B 1.5

N=2450 slope: 0.74

AB(CV)

1

.5

0

-.5 0

.5

1

1.5

2

1.5

2

AB(SFD)

AB(CV)

2

0

-2 0

.5

1 AB(SFD)

F igure 6.T he sam e as Fig.5 but for sam ple "B ".

11

J.C holoniewski,E.A .Valentijn

sample A 1.5

N=7974 slope: 0.56

AB(CV)

1

.5

0

-.5 0

.5

1

1.5

2

1.5

2

AB(BH)

2 AB(CV)

12

0

-2 0

.5

1 AB(BH)

F igure 7.T he sam e as Fig.5 but for extinction ofB H A B (B H ).

Extinction in the G alaxy from surface brightnessesofESO -LV galaxies:testing ’standard’extinction m aps

sample B 1.5

N=2450 slope: 0.61

AB(CV)

1

.5

0

-.5 0

.5

1

1.5

2

1.5

2

AB(BH)

AB(CV)

2

0

-2 0

.5

1 AB(BH)

F igure 8.T he sam e as Fig.7 but for sam ple "B ".

13

14

J.C holoniewski,E.A .Valentijn

.8 CV (this paper) SFD BH + 0.12

AB

.6

csc(b)

.4

.2

0

-80

-60

-40

-20

0

20

40

galactic latitude (b)

F igure 9. G alactic latitude dependence of our absolute extinction (represented by the solid broken line w ith 1 error bars) com pared w ith the extinction according to SFD , B H and the cosecans law . T he B H extinction have been corrected by adding 0.12 m agnitude constant.N ote north-south assym etry ofour extinction near the G alactic equator (jbj< 40o ) and non-zero extinction near the G alactic south pole (b 90).

Extinction in the G alaxy from surface brightnessesofESO -LV galaxies:testing ’standard’extinction m aps

.8 CV (this paper) SFD * 0.7 ( BH + 0.12 ) * 0.6

AB

.6

csc(b)

.4

.2

0

-80

-60

-40

-20

0

20

40

galactic latitude (b)

F igure 10. T he sam e as Fig.9 but w ith SFD extinction m ultiplied by the factor 0.7 and for corrected B H extinction m ultiplied by 0.6.

15