Georeferenced data as a tool for monitoring the concentration of population in Finland in 1970–1998

Geo e e enced da a as a ool o moni o ing he concen a ion o popula ion in Finland in 1970–1998 JARMO RUSANEN, TOIVO MUILU, ALFRED COLPAERT AND ARVO NAUKKARINEN Rusanen, Ja mo, Toi o Muilu, Al ed Colpae & A o Naukka inen (2003). Geo e e enced da a as a ool o moni o ing he concen a ion o popula ion in Finland in 1970–1998. Fennia 181: 2, pp. 129–144. Helsinki. ISSN 0015- 0010. Changes in he dis ibu ion o popula ion in Finland o e he pe iod 1970– 1998 a e examined in e ms o a co-o dina e sys em o 1 x 1 km g id cells. The esul s indica e ha his sys em p o ides sui able a eal uni s o a a ie y o s a is ical and GIS me hods aimed a desc ibing and explaining he spa ial dis ibu ion o popula ion. The sys em is capable o yielding mo e de ailed in o ma ion han he e o o e on opics such as he concen a ion o popula ion in he u ban cen es o Finland – a p ocess ha has been going on since he beginning o he las cen u y, bu has slowed down no iceably in he ecen yea s. Ja mo Rusanen, Toi o Muilu, Al ed Colpae & A o Naukka inen, Depa - men o Geog aphy, PO BOX 3000, FIN 90014 Uni e si y o Oulu, Finland. E- mail: [email p o ec ed]. MS ecei ed 07 Janua y 2003. In oduc ion Demog aphic esea ch has a long adi ion wi h- in geog aphy and ep esen s a discipline ha is eminen ly sui ed o he s udy o he spa ial dis i- bu ion o popula ion and i s changes. The esul s o hese s udies a e equi ed o planning pu pos- es in many sec o s o he socie y, e.g. de e mina- ion o he capaci y and loca ion equi emen s o a ious public se ices a he na ional, egional o local le el. Al hough i is p ima ily he au ho - i ies ha ha e been in e es ed in he di ec ion o demog aphic ends, i is by no means a ma e o indi e ence o indi idual ci izens o amilies whe he hey li e in an a ea o expanding o di- minishing popula ion. The gene al image o an a ea is o a conside able ex en shaped by ends in popula ion, as well as wi h s uc u al changes o he a ea. The e is hus an ob ious need o s a- is ical da a ha a e bo h up o da e and geog aph- ically mo e de ailed han ea lie , and o p ecise in e p e a ions o such da a. This poses new chal- lenges no only o he p oduce s and use s o he s a is ics bu also o hose engaged in de elop- ing and applying analy ical and in e p e a i e me hodologies. The geog aphical examina ion o popula ion dis ibu ions and hei empo al changes has yp- ically elied on no ions such as spa ial concen- a ion and deconcen a ion (Lipshi z 1996; Bo gegå d e al. 1995), o u baniza ion and coun- e u baniza ion (Håkansson 2000), while mig a- ion esea ch has adop ed pe spec i es such as ha o he ‘ u na ound’ phenomenon (Long & Nucci 1997; Lewis 2000). The ac ual e ms in use a y no only acco ding o he heo e ical basis om which he esea ch se s ou bu also in ela- ion o he olume o mig a ion conce ned and i s empo al du a ion. Ou pic u e o he spa ial concen a ion and agglome a ion o popula ion in Finland has la ge- ly been shaped by he publica ions o Hus ich (1977) and Ales alo (1983). Mo e ecen ly, how- e e , he p esen au ho s ha e discussed changes in he se lemen s uc u e and he concen a ion o popula ion wi h he aid o geo e e enced da a and GIS me hods (see Naukka inen e al. 1993; Rusanen e al. 1997). As poin ed ou by e.g. Håkansson (2000), e- sea ch in o he dis ibu ion o popula ion can be g ea ly a ec ed by he employed geog aphical uni , so ha models ha apply a one spa ial le - 130 FENNIA 181: 2 (2003) Ja mo Rusanen, Toi o Muilu, Al ed Colpae and A o Naukka inen el do no necessa ily hold good a ano he . One pa icula ly di icul p oblem conce ns bounda y changes, which indica es ha census s a is ics o di e en yea s do no necessa ily apply o he same a eal uni s and he e o e, canno be com- pa ed easily (Howens ine 1993; Openshaw 1995; Rees 1998). The equi emen s placed on popula ion s a is- ics and he use ulness o he da a a ailable can a y g ea ly om one g oup o use s o ano he . Public au ho i ies, o ins ance, a e accus omed o using s a is ical means and de ini ions ex- p essed in e ms o adminis a i e a eas and de- duc ions made om hese when hey equi e in- o ma ion on which o base hei decisions. Re- sea che s, in con as , p e e o go back o he sou ces o he o iginal aw da a in o de o es hei hypo heses and cons uc hei models. Fo his eason we will a emp o analyse and exam- ine demog aphic da a using a ious me hods in e ms o bo h adminis a i e and non-adminis a- i e a eal uni s, and o e alua e he subsequen esul s. Aims, da a and me hods We se ou he e o desc ibe 1) he changes in pop- ula ion ha ha e aken place wi hin he spa ial and se lemen s uc u e o Finland by means o 1 x 1 km g id cell da a, 2) he me hods a ailable o he analysis o popula ion da a, and 3) he ac- o s highligh ed by he g id cell me hod ha can- no be iden i ied when using adminis a i e a eal uni s. Finally, he use ulness o he esul ing in- o ma ion o he pu poses o demog aphic e- sea ch will be e alua ed. Geo e e enced da a on Finland a e a ailable a many spa ial le els. In p ac ice, any indi idual ci izen can be assigned a se o coo dina es de- sc ibing he geog aphical loca ion o he house in which he o she li es. The esul ing da a can be agg ega ed o ep esen highe -le el uni s, e.g. pos al dis ic s, pa s o municipali ies, whole municipali ies, g oups o municipali ies o he NUTS hie a chy as used by he EU. I mus be e- membe ed, howe e , ha he abo e-men ioned p oblem o compa abili y o e ime eme ges in he case o hese a eal uni s whene e adminis- a i e a eas a e combined o bounda ies a e al- e ed. On he o he hand, he ac ha he ini ial in o ma ion is a ailable a he indi idual le el is ega ded as a signi ican ad an age o GIS me h- ods (Ma in & Higgs 1997; see also Openshaw & Tu on 1996). Thus, he p incipal uni s employed he e a e he 1 x 1 km g id cells de ined in he da a p oduced by S a is ics Finland o he yea s 1970, 1980 and 1990–1998 (see Rusanen e al. 1997). These g id cells a e o ixed loca ion and a e independen o any changes in adminis a i e bounda ies. S a is- ical in o ma ion on human ac i i ies in Finland has been a ailable annually since 1987. Cu en me hods enable S a is ics Finland o e ise hei geo e e enced popula ion da a e y quickly, so ha igu es o he popula ion on 31s Decembe a e a ailable by he end o he ollowing Feb u- a y, oge he wi h s a is ics o he indi idual mu- nicipali ies. This annual in o ma ion is based on egis e s, so ha no sepa a e popula ion census is necessa y as in he USA, o example (see C ews 2000). The second basic a eal uni employed he e is he municipali y, o independen local go e n- men dis ic . These dis ic s ha e nume ous igh s and obliga ions ha de ine hem as adminis a- i e en i ies, including he igh o le y axes (Min- is y o he In e io 2000). In acco dance wi h he abo e p inciple, he popula ions o he munici- pali ies a e calcula ed he e om he g id cell ig- u es, and o his eason he igu es di e sligh ly om hose quo ed in he o icial s a is ics, as o ins ance in 1998, 1.42% o he na ional popula- ion could no be assigned coo dina es on he ba- sis o hei place o esidence. Howe e , his dis- c epancy can be assumed no o a ec he esul s. I should be no ed ha we a e no conce ned he e wi h s udying he po en ial ac o s esponsi- ble o demog aphic ends in pa icula a eas, e.g. bi h- a es, mo ali y o mig a ion – opics which ha e adi ionally been o he o e in pop- ula ion geog aphy (Ogden 1998; 1999). Ins ead, we ocus on desc ibing solely he concen a ion o popula ion. This is de ined he e in wo ways: as an inc ease in popula ion densi y pe squa e kilome e when using g id cell da a, and as an in- c easingly la ge p opo ion o he popula ion li - ing in a p og essi ely smalle numbe o munici- pali ies. Da a on popula ion changes in he 1 x 1 km g id cells and municipali ies a e analysed he e by a numbe o me hods, and assessmen s a e made as o whe he hese me hods yield simila o di- e gen esul s. The me hods es ed include he Gini Coe icien , which is mo e commonly used o analysing he spa ial dis ibu ion o income FENNIA 181: 2 (2003) 131 Geo e e enced da a as a ool o moni o ing he… a es (see Chak a o y 1994) a he han popula- ion (see B ad o d & Ken 1986; Lo ell-Smi h 1993). In addi ion, he concen a ion o popula- ion is desc ibed in e ms o mean alues and he Mo an and Gea y Indices used in e sion 8.0 o he A c/In o so wa e. Changes in popula ion a e also desc ibed by e e ence o deciles and by clas- si ica ion o he ‘spa ial demog aphic s uc u e’. Long- e m pa e ns a e isualized by means o a ank size model and he k iging in e pola ion me hod con ained in e sion 8.0 o A c/In o. The Hoo e Index ha is equen ly used in s udies o popula ion concen a ion (see Bo gegå d e al. 1995; Long & Nucci 1997) is no employed he e, as he emphasis is on analysis a he g id cell le - el a he han a he municipali y le el. Decile analysis in ol es anking he da a uni s in size o de om la ges o smalles wi h espec o he s udied a iable, and di iding hem in o en equal g oups in e ms o popula ion desc ibed by he a iable, ega dless o whe he he a eal uni is a g id cell o a municipali y. Popula ion concen a ion as e lec ed in he Mo an and Gea y indices and he Gini coe icien Popula ion densi y in Finland in 1998 ela i e o land a ea was e y low by Eu opean s anda ds, only 16.9 inhabi an s pe squa e kilome e. I , howe e , we use he g id cell ma e ial o calcu- la e he mean densi y o inhabi ed cells only, we see ha he densi y has inc eased s eadily o e he pe iod examined he e, and is now app oach- ing 50 inhab./km2 (Table 1). The Mo an and Gea y me hods allow a spa ial au oco ela ion index o be de e mined o he g id cell ma e ial as ollows: when he Mo an In- dex is posi i e and he Gea y Index a ies in he ange 0–1 he dis ibu ion o popula ion can be desc ibed as ‘simila ’, ‘ egionalized’, ‘smoo h’ o ‘clus e ed’. I he Mo an Index is nega i e and he Gea y Index g ea e han 1, he co esponding e ms a e ‘dissimila ’, ‘con as ing’ and ‘checke - boa d’ ( o de ails, see A c/In o Use Manual 2000). In he p esen case, he g adual inc ease in he Mo an Index and he simul aneous de- c ease in he Gea y Index lend u he suppo o he imp ession gained om he p e ious able o a p ocess o consolida ion o he inhabi ed a ea a he han a sca e ing o dispe sal o popula ion (Table 2). One ypical echnique employed o desc ibe e enness o concen a ion in he dis ibu ion o popula ion is he Lo enz diag am (Ales alo 1983), o i s de i a i e, he Gini Coe icien (Fains ein 1996). Rega dless o whe he we use g id cells o municipali ies as he a eal uni s, he ob ained Gini Coe icien s indica e a concen a ion o popula- ion (Table 3), al hough he highe alue o he g id cell ma e ial as compa ed o he municipali- y le el also indica es he essen ially local na u e o his concen a ion. The ac ha he majo i y o he popula ion o Finland is loca ed wi hin a ela i ely small numbe o g id cells is indica i e o he low p opo ion o buil -up a eas wi hin he coun y’s o al se led a ea. Measu ed in bo h o he abo e ways, he a e o change was g ea es in he 1970’s, an obse a- ion ha con i ms he gene al imp ession ega d- ing he concen a ion o popula ion in Finland. I is consis en wi h he esul s o he examina ion by deciles published by Ales alo (1983). On he o he hand, analysis o he end o e he las wo decades in e ms o he Gini Coe icien p o ides de ian esul s o he wo se s o a eal uni s. The a e o popula ion concen a ion appa en ly has slowed down when assessed in e ms o he g id Table 1. Numbe o inhabi ed g id cells and mean popula ion densi y in Finland in 1970–1998 (Da a: S a is ics Finland). 1970 1980 1990 1992 1994 1996 1998 Numbe o inhabi ed g id cells 110477 104540 103242 103020 103036 103045 102873 Inhab./km241.3 44.1 47.8 48.4 48.9 49.1 49.4 Table 2. Mo an (I) and Gea y (C) Indices in 1970–1998 (Da a: S a is ics Finland). Yea 1970 1980 1990 1992 1994 1996 1998 I 0.543 0.531 0.568 0.570 0.581 0.590 0.597 C 0.427 0.435 0.400 0.394 0.384 0.377 0.370 132 FENNIA 181: 2 (2003) Ja mo Rusanen, Toi o Muilu, Al ed Colpae and A o Naukka inen cell da a, being only 1.1% in he 1990’s, whe e- as he igu es o he municipali ies indica e ac- cele a ion wi h a e minal a e o 3.5%. The esul ob ained om he g id cell da a may indica e a dec easing end in popula ion concen- a ion in he denses a eas, he pa e n being a - ibu able o he exclusion o uninhabi ed g id cells and he concen a ion o popula ion in a cons an ly dec easing numbe o cells and a mo e es ic ed geog aphical a ea. This end, which is la gely in e nal o indi idual municipali ies, ails o be e lec ed in he analyses based on munici- pali ies as he a eal uni s. Decile analysis Decile analysis employs di isions o he o al ma- e ial in o en h pa s. I can be ega ded as a lex- ible, non-gi en means o classi ica ion, and since he bounda ies o he deciles end o a y om one yea o he nex , he me hod is well sui ed o s udies o concen a ion o dispe sal, e.g. in pop- ula ion o incomes. Acco ding o Ales alo (1983), he popula ion o Finland was ai ly e enly dis- ibu ed o e he deciles a he end o he 19 h cen u y, he ag a ian socie y o he ime showing li le agglome a ion o he popula ion in o u ban cen es. F om ha ime onwa ds, howe e , he popula ion g adually became concen a ed in a smalle numbe o municipali ies (Table 4), so ha by he 1970’s, hal o he coun y’s popula ion li ed in 52 municipali ies, 10.1% o hei o al numbe . The end con inued so ha in 1998, he co esponding igu e was 33 municipali ies, i.e. 7.3% o he o al. Pa o his e ec may be a ib- u ed o he amalgama ion o municipali ies wi h- in he sys em o local go e nmen , bu he p inci- pal ac o has wi hou doub been he ac ual con- cen a ion o popula ion wi hin a p og essi ely smalle numbe o owns and ci ies (Rusanen e al. 2000). Is his concen a ion isible in he g id cell da a? Fo answe ing his, a decile analysis com- pa able o ha in Table 4 is p esen ed in Table 5. The deciles indeed indica e a con inua ion o he concen a ion p ocess, so ha hal o he coun- y’s popula ion ( he popula ion o deciles 1–5) occupied a o al o 1541 km2 in 1970, 1167 km2 in 1980, 1294 km2 in 1990 and 1284 km2 in 1998. The 1970’s we e a pe iod o hea y popula- ion concen a ion, whe eas om he end o ha decade onwa ds Finland was a ec ed by a ‘ u n- a ound’ phenomenon, as no ed by se e al au ho s (Kauppinen 2000). This p ocess was expe ienced in many wes e n coun ies, en ailing abo e all a Table 3. Concen a ion o popula ion as shown by he Gini Coe icien s o he yea s s udied and pe cen age change o e he pe iod 1970–1998 (Da a: S a is ics Finland). Yea A eal uni 1970 1980 1990 1998 G id cell 0.78558 0.83450 0.85365 0.86324 Municipali y 0.59228 0.62145 0.62962 0.65186 Change in Gini Coe icien (%) 1970–1980 1980–1990 1990–1998 G id cell 6.2 2.3 1.1 Municipali y 4.9 1.3 3.5 Table 4. Concen a ion o popula ion by deciles o munici- pali ies in 1970, 1980, 1990 and 1998 (Da a: S a is ics Fin- land). Numbe o municipali ies Yea Decile 1970 1980 1990 1998 1 1 1 2 1 2 4 4 3 3 3 6 5 5 4 4 151010 9 5 26201816 6 38292723 7 50424038 8 66585756 9 97888788 10 212 204 206 214 To al 515 461 455 452 FENNIA 181: 2 (2003) 133 Geo e e enced da a as a ool o moni o ing he… decline in he popula i y o u ban a eas as li ing en i onmen s o amilies wi h child en. The inhabi ed a ea o Finland in 1998 amoun - ed o ca. 30.4% o he o al su ace a ea o 338 145 km2, wi h a consis en decline o his p opo - ion om one decade o he nex since 1970. The mos p onounced decline, almos 6000 km2, ook place du ing he 1970’s, al hough i has also been claimed ha he ma e ial o 1970 con ained some e o s in he coo dina es de e mined o in- di idual dwellings and ha he ac ual ne de- c ease was no necessa ily as g ea as his. I should also be no ed ha he end was a ela- i ely s eady o e he pe iod 1990–1998 ela i e o he 1980’s, al hough da a on he inal yea s o he decade a e s ill lacking. Changes may also be obse ed in he popula- ion densi ies o he deciles (Table 6). The high- es densi ies o all we e eco ded in deciles 1 and 2 in 1970, when Finland was expe iencing a p o- nounced mig a ion om he coun yside in o he owns and also ab oad, p ima ily o Sweden (see Kauppinen 2000). A ha ime he mean popula- ion densi y in he op decile was a ound 9700 inhab./km2, whe eas he co esponding igu e in 1998 was abou 6200. The maximum popula ion densi y was eached in Helsinki, he capi al. In he 1970 da a he densi y peaked a 29 234 in- hab./km2, whe eas he igu e o 1998 showed a educ ion o mo e han a hi d, 19 172 inhab./ km2. A numbe o ac o s can be iden i ied ha con- ibu ed o he educ ion o popula ion densi y in Finland o e he s udy pe iod. The mos signi i- can ac o was he ans e in he u ban cen es om dwellings o o he uses, mainly o ices and comme cial p emises. O he ac o s we e chang- es in he s uc u e o households, including educ- ions in he a e age amily size and he numbe o amilies wi h child en and an inc ease in he numbe o single-pe son households. I should be poin ed ou , howe e , ha he mean popula ion densi y o he wo mos densely inhabi ed deciles began o inc ease again in he 1990’s, pa ly on accoun o he e o s made o ill in he se lemen pa e n in buil -up a eas in acco dance wi h he p inciples o sus ainable de- elopmen . Ande sson (1988) ecognizes ou s ages in he de elopmen o owns in Finland: u baniza ion, subu baniza ion, disu baniza ion and eu baniza- ion. U baniza ion s age e e s o he g ow h o a cen al u ban nucleus, while he subu baniza ion s age in ol es a slowing down and e en ually ces- sa ion o his end. A he disu baniza ion s age, Table 5. Numbe s o inhabi ed g id cells by deciles in 1970, 1980, 1990 and 1998 (Da a: S a is ics Finland). 1970 1980 1990 1998 Inhab. km2 Decile N N N N in 1998 1 denses 47 66 84 82 4171–19172 2 se lemen 116 133 159 158 2633–4170 3 207 200 234 234 1795–2632 4 384 297 336 334 1257–1794 5 787 471 481 476 881–1256 6 1 856 825 751 721 536–880 7 4 992 1 928 1 420 1 299 267–535 8 11 290 6 816 4 635 3 813 70–266 9 spa ses 22 331 19 590 16 579 14 980 20–36 10 se lemen 68 462 74 209 78 561 80 770 1–19 Numbe o g id cells 110 472 104 535 103 240 102 873 Table 6. Popula ion densi y by deciles in 1970, 1980, 1990 and 1998 (Da a: S a is ics Finland). Inhab. km2 Decile 1970 1980 1990 1998 1 9703 6980 5876 6202 2 3932 3464 3104 3219 3 2203 2303 2109 2173 4 1188 1551 1469 1523 5 579 978 1026 1068 6 246 558 657 705 7 91 239 348 392 8 40 68 106 133 9 20243034 10 7666 134 FENNIA 181: 2 (2003) Ja mo Rusanen, Toi o Muilu, Al ed Colpae and A o Naukka inen he end is e e sed, un il he eu baniza ion s age ma ks new g ow h in he u ban nucleus. I is in e es ing o conside how hese s ages migh be e lec ed in an empi ical decile analysis. The u baniza ion s age seems o ha e aken place in Finland be o e 1970, as he popula ion o he mos densely inhabi ed cen al a eas (deciles 1 and 2) began o decline a e ha ime. Hence, he 1970’s and 1980’s may be assigned o he subu baniza ion and disu baniza ion s ag- es, as hese igu es in pa icula declined ma k- edly a i s and hen le elled ou somewha in he 1980’s, when he changes we e less p o- nounced in o he espec s, oo. The 1990’s ep e- sen he eu baniza ion s age, in which he popu- la ion o he mos densely inhabi ed a eas again s a ed o inc ease. I is impo an o bea in mind when e alua - ing hese indings ha he s ages a e dis inguished on he s eng h o only a single a iable. A mo e p ecise conside a ion would call o an in e nal analysis o he s uc u e o he en la ges ci ies, o example, and also o he in o ma ion ele an o he g ow h o u ban a eas, e.g. hei de elop- men and planning policies. Popula ion changes by deciles in 1990–1998 Dual ends in popula ion densi y and in he numbe o inhabi ed g id cells a e obse able du - ing he 1990’s, he cu -o poin being eached in 1993. The minimum a ea occupied by hal o he popula ion o Finland, i.e. deciles 1–5, inc eased nume ically o e ha ime (Table 7a), implying a sligh decline in popula ion densi y in he u ban nuclei and subu bs (Table 7b). F om 1993 on- wa ds, he end owa ds dense communi ies can be obse ed, al hough no qui e amoun ing o he igu es eco ded in 1970 and 1980. The mos spa sely popula ed u al a eas (decile 10) inc eased in numbe h oughou he 1990’s, whe eas hei mean popula ion densi y emained mo e o less s able. The inc ease may be a ibu - ed almos en i ely o educ ions in he popula ion o some g id cells p e iously con ained in decile 9, ep esen ing he u al a eas p ope , causing hei ans e o decile 10. The e e sal in he end in 1993 is p obably linked o he ac ha his was he wo s yea o he economic ecession in Finland, i.e. he pe i- Table 7. Changes in he numbe o g id cells (A) and popula ion densi y (B) in he 1990’s (Da a: S a is ics Finland). A. Numbe o g id cells T end Decile 1990 1991 1992 1993 1994 1995 1996 1997 1998 1990– 1994– 1990– 1993 1998 1998 1 84858686858483 8282+– 2 159 162 164 164 163 161 160 160 158 + – – 3 234 238 241 241 240 239 236 235 234 + – 0 4 336 339 342 344 343 339 338 336 334 + – – 5 481 484 487 488 485 482 480 477 476 + – – 6 751 751 750 751 744 736 728 724 721 0 – – 7 1420 1406 1400 1390 1364 1345 1329 1316 1299 – – – 8 4635 4500 4409 4332 4214 4096 4012 3915 3813 – – – 9 16579 16305 16105 15930 15746 15548 15380 15204 14980 – – – 10 78561 78815 79038 79314 79652 80007 80299 80583 80770 + + + B. Popula ion densi y inhab./km2 1 5876 5840 5802 5831 5923 6015 6100 6191 6202 – + + 2 3104 3064 3043 3058 3089 3138 3165 3173 3219 – + + 3 2109 2086 2071 2081 2098 2114 2145 2160 2173 – + + 4 1469 1464 1459 1458 1468 1491 1498 1511 1522 – + + 5 1026 1026 1025 1028 1038 1048 1055 1064 1068 + + + 6 657 661 665 668 677 687 695 701 705 + + + 7 348 353 356 361 369 376 381 386 391 + + + 8 106 110 113 116 119 123 126 130 133 + + + 9 30303131323233 3334+++ 10 6,3 6,3 6,3 6,3 6,3 6,3 6,3 6,3 6,3 0 0 0 FENNIA 181: 2 (2003) 135 Geo e e enced da a as a ool o moni o ing he… od in which he unemploymen peaked. Simila - ly, housing p oduc ion declined s eadily, o each i s lowes ebb in 1994. This was ollowed by a pe iod o economic eco e y (SVT 1999), wi h a new s imula ion o building ac i i y, especially in he me opoli an a ea o Helsinki and o he la ge municipali ies. This end was accompanied by a ise in popula ion densi ies, as he housing capac- i y o he buil -up a eas inc eased while he land a eas conce ned emained mo e o less cons an . The inc eased housing densi y was hence e lec - ed in he popula ion densi y. The abo e obse a ion demons a es he usabil- i y o he g id cell da a as an aid in moni o ing changes in spa ial s uc u e, and po en ially achie ing de ailed explana ion. This me hod al- lows, o ins ance, obse ing small changes in popula ion wi hin a municipali y e y easily. The spa ial demog aphic s uc u e in 1970–1998 The epea abili y o he decile me hod allows i o be applied o any coun y o any a eal uni , and any esea che can a i e a he same esul s. I should be emembe ed, howe e , ha such ac- cu a e geo e e enced da a a e a ailable in selec - ed coun ies only. We will u n ou a en ion now o s uc u al ea u es o he spa ial dis ibu ion o popula ion and he changes de ec ed in he dis- ibu ion in he c oss-sec ional da a o 1970, 1980, 1990 and 1998, wi h pa icula e e ence o de elopmen s du ing he 1990’s. The classi i- ca ion employed may be e e ed o as he ‘spa- ial demog aphic s uc u e’, as i ep esen s an a - emp o desc ibe he ela ion be ween popula- ion densi y and se lemen s uc u e. Hence, i does no ake in o accoun he unc ional ele- men s no mally implied in he e m egional s uc- u e, e.g. dwellings, jobs and he aspec s o in a- s uc u e ha suppo hese. The concep has been used ea lie o classi ica ion pu poses by Räisänen e al. (1996) and Rusanen e al. (1997). We will i s conside he si ua ion o e he whole o Finland and subsequen ly concen a e he anal- ysis on one speci ic egion, Kainuu. The in e p e a ion p o ided he e is based on he assump ion ha he cha ac e o a g id cell can be deduced om he size o i s popula ion. The classi ica ion conce ned is no necessa ily appli- cable o all a eas o o all coun ies, and i has been cons uc ed knowing well ha he dis inc- ion be ween u al and u ban is by no means un- ambiguous (see Malinen e al. 1994; Be y e al. 2000). I is impossible o build a model ha would apply equally well o all coun ies and unde all condi ions. The ollowing classi ica ion o spa ial demog aphic s uc u e is used: Inhab./km2Elemen o spa ial demog aphic s uc u e 1–5 Sca e ed se lemen 6–20 Ru al a eas p ope 21–100 Ru al a eas wi h buil -up ea u es 101–1000 Build-up a eas and subu bs wi h mos ly p i a e housing Mo e han High- ise cen es and subu bs o 1000 majo ci ies Examina ion o he si ua ion in he yea s men- ioned abo e indica es ha he mos densely pop- ula ed a eas, wi h o e 1000 inhab./km2, g ew mos apidly in he 1970’s, he a e o g ow h di- minishing in he 1980’s and eaching i s slowes in he 1990’s (Fig. 1). The a eas o subu ban p i- a e housing depa ed om his pa e n some- wha , howe e , as hese unde wen hei g ea es popula ion g ow h in he 1980’s. By con as , he ansi ional ca ego y be ween u al and u ban condi ions, ha wi h densi ies o 21–100 inhab./ km2, declined in popula ion h oughou he pe i- od s udied he e, al hough mos ma kedly in he 1970’s and leas so in he 1990’s. The u al a eas p ope dec eased subs an ially in ex en , while he a eas o sca e ed se lemen inc eased somewha in bo h o al popula ion and ex en , la gely as a esul o con ac ions in he popula ion o g id squa es p e iously included in he ca ego y o u- al a eas p ope . The end in popula ion densi y o e he whole coun y in he 1990’s was pola ized. The popula- ion o he densely inhabi ed a eas inc eased s eadily h oughou he decade, amoun ing o a o al ise o 4.6%, o 110 000 pe sons, be ween 1990 and 1998 (Fig. 2). The subu ban p i a e housing popula ion g ew in a simila manne , wi h he excep ion o a small decline in 1997. The popula ion o he u ban- u al ansi ion zone wi h densi ies o 21–100 inhab./km2, declined om 1993 onwa ds, howe e , and simul aneously he a ea con ac ed sligh ly. The ca ego y o u al a - eas p ope declined in bo h popula ion and o al a ea h oughou he s udied pe iod, whe eas sca - e ed se lemen ca ego y inc eased sligh ly in bo h o al popula ion and a ea. 136 FENNIA 181: 2 (2003) Ja mo Rusanen, Toi o Muilu, Al ed Colpae and A o Naukka inen Fig. 1. Popula ion changes in di e en pa s o he spa ial demog aphic s uc u e o e he whole o Finland in 1970–1998 (Da a: S a is ics Finland). Fig. 2. Popula ion changes in di e en pa s o he spa ial demog aphic s uc u e o e he whole o Finland in 1991–1998 (Da a: S a is ics Finland). The Kainuu egion in No he n Finland, selec - ed he e o mo e de ailed examina ion, is cha - ac e ized by economic and demog aphic eces- sion. I belongs o he Objec i e 1 EU suppo a - eas on accoun o i s low income le els and spa se se lemen . I s popula ion, which has been on he decline since he 1960, was 93 218 pe - sons in 1998, and he egion has consis en ly been one wi h pi o al unemploymen in he whole coun y. The demog aphic end in Kainuu du ing he 1990’s depa ed ma kedly om ha ob ained o he whole coun y (Fig. 3). The mos densely in- habi ed a eas expe ienced a popula ion decline om 1992 onwa ds, and he same end a ec ed he p i a e housing a eas om 1996 onwa ds. Co espondingly, he ansi ion zone and he u- al a eas p ope los popula ion h oughou he decade, as in he whole coun y, and he popula- ion inc ease in he a eas o sca e ed se lemen FENNIA 181: 2 (2003) 137 Geo e e enced da a as a ool o moni o ing he… appea s o ha e come o an end in 1998. This was p obably he i s occasion du ing he cen u y when he popula ion o Kainuu dec eased in e e- y single ca ego y o i s spa ial s uc u e. I is p ob- able ha demog aphic ends in he spa se popu- la ion densi y classes in Fig. 9 we e he same as in Kainuu. Elsewhe e hey ha e esembled hose o he whole coun y. Finland became a membe o he EU in 1995. Kainuu was an Objec i e 6 a ea in 1995–1999 and is cu en ly an Objec i e 1 a ea o he pe i- od 2000–2006. Based on he abo e esul s one can a gue ha he de elopmen measu es imple- men ed in he egion unde he EU p og ammes ha e ailed o imp o e he nega i e demog aphic end in all elemen s o he spa ial sys em. De ailed analyses o popula ion ends a e pos- sible only wi h he aid o geo e e enced da a ha allow p ecise loca ion o he popula ion uni s. A ‘sliding scale’ e alua ion o popula ion densi y based on local le el da a helps a oiding he al- lacies ha a ise when one examines he whole coun y o la ge agg ega e a eas (Ma in 1991). The un a elled nega i e demog aphic end in Kainuu se es as an example o he use o g id cell da a o moni o ing egional de elopmen . The ank size model Rank size models ep esen a well-es ablished me hod o geog aphical esea ch ha has been applied o he s udy o hie a chical s uc u es in a ious a eal uni s and changes aking place in hese s uc u es. The uni s employed o his pu - pose a e usually owns o o he adminis a i e o unc ional en i ies (see, e.g. B ad o d & Ken 1986; Das & Du 1993). The me hod in ol es a ang- ing he da a uni s in size o de and p esen ing he esul s in diag amma ic o m. We in end o apply he ank size app oach o he 1 x 1 km2 g id cells, and use he densi y classes o desc ibing he ypes o se lemen and any changes in popula- ion. Fo compa ison pu poses, he app oach was done a wo le els: he whole coun y and he Kainuu egion. The 452 municipali ies o Finland a ied in popula ion om a me e 100 up o 500 000 in- habi an s in 1998 (Fig. 4). Wi h he municipali y le el as he a eal uni o in e es , he ank size model indica es ha abou one hund ed la ges municipali ies showed an inc ease in popula ion o e he pe iod 1970–1998, while he o he s ex- pe ienced popula ion decline. A he g id cell le el, he ank size model ails o p o ide de ailed in o ma ion o spa ial demo- g aphic s uc u e as e icien ly he classi ica ion (Figs. 5 and 6). I does, howe e , se e well in highligh ing he decline in he o al numbe o in- habi ed g id cells o e he whole coun y, i.e. he con ac ion in he a ea o human se lemen . Fu - he mo e, i emphasizes ce ain majo u ning poin s such as he beginning o he decline in popula ion in he mos densely inhabi ed a eas Fig. 3. Popula ion changes in di e en pa s o he spa ial demog aphic s uc u e in Kainuu in 1991–1998 (Da a: S a is ics Finland). 144 FENNIA 181: 2 (2003) Ja mo Rusanen, Toi o Muilu, Al ed Colpae and A o Naukka inen Haa sen T, P G oo e & PPP Huigen (2003). Measu - ing age di e en ials in ep esen a ions o u ali y in The Ne he lands. Jou nal o Ru al S udies 19, 245–252. Håkansson J (2000). Changing popula ion dis ibu ion in Sweden. GERUM – Kul u geog a i 2000: 1. Howens ine E (1993). Measu ing demog aphic change: he spli ac p oblem. P o essional Ge- og aphe 45: 4, 425–430. Hus ich I (1977). Suomen “ku is umisen” auh i hil- jenee (The speed o he “sh inkage” o Finland is slackening). Te a 89: 4, 214–221. Kauppinen J (2000). Muu oliike Suomessa uosina 1989–1994 koo dinaa ipohjais en paikka ie ojen pe us eella (English summa y: Mig a ion o pop- ula ion in Finland in 1989–1994, examined om coo dina e-based GIS da a). 155 p. Sii olaisuus- ins i uu in u kimuksia A 22. Lang o d M & DJ Unwin (1994). Gene a ing and mapping popula ion densi y su aces wi hin a geog aphical in o ma ion sys em. The Ca o- g aphic Jou nal 31, June, 21–26. Lewis G (2000). Changing places in a u al wo ld. The popula ion u na ound in pe spec i e. Geog- aphy 85: 2, 157–165. Lipshi z G (1996). Spa ial concen a ion and decon- cen a ion o popula ion: Is ael as a case s udy. Geo o um 27: 1, 87–96. Long L & A Nucci (1997). The Hoo e Index o pop- ula ion concen a ion: A co ec ion and upda e. P o essional Geog aphe 49: 4, 431–440. Lowell-Smi h EG (1993). Regional and in ame o- poli an di e ences in he loca ion o ees and- ing ambula o y su ge y cen e s. P o essional Ge- og aphe 45: 4, 398–407. Malinen P, R Ke änen & H Ke änen (1994). Ru al a ea ypology in Finland – a ool o u al policy. 78 p. Uni e si y o Oulu, Resea ch Ins i u e o No he n Finland. Resea ch epo s 123. Ma in D (1991). Geog aphic in o ma ion sys ems and hei socioeconomic applica ions. 182 p. Rou ledge, New Yo k and London. Ma in D & G Higgs (1997). Popula ion geo e e enc- ing in England and Wales: basic spa ial uni s e- conside ed. En i onmen and Planning A 29: 2, 333–347. Ma in R (2000). In memo y o maps. T ansac ions o he B i ish Geog aphe s, New Se ies 25: 1, 3–5. Minis y o he En i onmen (2000). <h p://www. yh. i/ajankoh / iedo e/ym/ ied99/ ym12299.h m>. 15.12.2000 Minis y o he In e io (2000). <h p://www: in e min. i/eng/municipal/municipali ies.h ml>. 15.12.2000 Muilu T & J Rusanen (2003). Ru al de ini ions and sho - e m dynamics in u al a eas o Finland in 1989–1997. Submi ed manusc ip . Naukka inen A, J Rusanen & A Colpae (1993). The u u e o u al se lemen s uc u e in Finland. Geog aphy Resea ch Fo um 13, 61–70. Ogden PE (1998). Popula ion geog aphy. P og ess in Human Geog aphy 22: 1, 105–114. Ogden PE (1999). Popula ion geog aphy. P og ess in Human Geog aphy 23: 4, 617–631. Openshaw S (1995). The u u e o he census. In Openshaw S (ed). Census Use ’s Handbook , 389– 411. GeoIn o ma ion In e na ional, Glasgow. Openshaw S & I Tu on (1996). New oppo uni ies o geog aphical census analysis using indi idu- al le el da a. A ea 28: 2, 167–176. Räisänen S, J Rusanen & A Naukka inen (1996). So- cio-economic g id da a and GIS o analyzing he changes in he Finnish coun yside. Second Join Eu opean Con e ence & Exhibi ion on Geog aph- ical In o ma ion, Ma ch 27–29, 1996. Ba celona, Spain. Geog aphical In o ma ion. F om Resea ch o Applica ion h ough Coope a ion Vol 1, 651– 660. Rees PH (1998). Wha do you wan om he 2001 Census? Resul s o an ESRC/JISC su ey o use s iews. En i onmen and Planning A 30, 1775– 1796. Rusanen J, A Naukka inen, A Colpe & T Muilu (1997). Di e ences in he spa ial s uc u e o he popula ion be ween Finland and Sweden in 1995 – a GIS iewpoin . 46 p. S a is ics Finland Re- sea ch Repo s 221. Rusanen J, T Muilu, A Colpae & A Naukka inen (2000). Asu un Suomen äes ömää ä yli ää 50 asukas a neliökilome illä. Tie oaika 5/2000, 14– 17. Sil as i T (2002). Maaseudun muu u a me ki ykse – uo annon ilas a kulu uksen koh eeksi. (Chang- ing meanings o he coun yside – om a p oduc- ion space o an objec o consump ion.) Yh- dyskun asuunni elu (Communi y Planning) 40, 11–25. Wheele J O (1998). Mapphobia in Geog aphy? 1980–1996. U ban Geog aphy 19: 1, 1–5. SVT (Suomen i allinen ilas o) (OSF = O icial S a- is ics o Finland) (1999). Raken aminen ja asu- minen. Vuosiki ja 1999. (Cons uc ion and Hous- ing. Yea book 1999). Raken aminen 1999: 20. 226 p. Tilas okeskus. S a is ics Finland.