ÇÕ.PDF

Transcription

1 통권 102 권 2019 년 9 월 ISSN September VOL. 102 국토연구 The Korea Spatial Planning Review 김동욱 개발제한구역지정의생태적효과 : 울산광역시를중심으로 이지원 김태형 서울시고령운전자교통사고특성분석 서성혁 정태열 해안곡 ( 谷 ) 계마을을대상으로한위요공간규모에관한연구 이진희 김현우 Analysis of Stormwater Runoff Reduction Effect Through Applying Low Impact Development Practices in a Flood Prone Area : Case of Incheon, South Korea 이종소 이상은 통신빅데이터를활용한국가하천친수지구의이용등급, 상세유형화및친수거점지구지정방법개발 이광현 7 개특 광역시의고령친화도와고령친화정책비교에관한연구 황종아 구자훈 서울시취약계층밀집지역분포와주거지특성분석 : 민간임대주택거주기초생활수급자를중심으로

2 국토연구는한국연구재단등재학술지입니다. 국토연구에실린논문의모든내용은필자개인의의견이며, 본연구원의공식견해가아님을밝힙니다. 편집위원회 위원장 이용우부원장 ( 국토연구원 / ywlee@krihs.re.kr) 박삼옥명예교수 ( 서울대학교 / parkso@snu.ac.kr) 위원 강호제 연구위원 ( 국토연구원 / hkang@krihs.re.kr) 변세일 연구위원 ( 국토연구원 / sibyun@krihs.re.kr) 권영섭 선임연구위원 ( 국토연구원 / yskwon@krihs.re.kr) 윤대식 교수 ( 영남대학교 / dsyun@yu.ac.kr) 김갑성 교수 ( 연세대학교 / kabsung@yonsei.ac.kr) 이승재 교수 ( 서울시립대학교 / sjlee@uos.ac.kr) 김경석 교수 ( 국립공주대학교 / gskim23@kongju.ac.kr) 이정록 교수 ( 전남대학교 / jrlee@jnu.ac.kr) 김동한 연구위원 ( 국토연구원 / dhkim@krihs.re.kr) 임동순 교수 ( 동의대학교 / dslim@deu.ac.kr) 김은란 연구위원 ( 국토연구원 / erkim@krihs.re.kr) 정성훈 교수 ( 강원대학교 / shjung@kangwon.ac.kr) 김종학 연구위원 ( 국토연구원 / jonghkim@krihs.re.kr) 하헌구 교수 ( 인하대학교 / hkha@inha.ac.kr) 김태진김홍배박미선 교수 ( 한국교통대학교 / tjkim@ut.ac.kr) 교수 ( 한양대학교 / hokim@hanyang.ac.kr) 연구위원 ( 국토연구원 / mspark@krihs.re.kr) Peter Batey Roberta Capello Professor (Univ. of Liverpool / P.W.J.Batey@liverpool.ac.uk) Onward Full Professor (Politecnico di Milano / Roberta.Capello@polimi.it) 박천규박태선 연구위원 ( 국토연구원 / cgpark@krihs.re.kr) 선임연구위원 ( 국토연구원 / tspark@gw.krihs.re.kr) Roger R. Stough Professor of Public Policy (George Mason Univ. / rstough@gmu.edu) 박환용 교수 ( 가천대학교 / hwanpark@gachon.ac.kr) 간사 한여정 책임전문원 ( 국토연구원 / yjhan@krihs.re.kr) w w w. k r i h s. r e. k r

3 국토연구 통권제 102 권 3 김동욱개발제한구역지정의생태적효과 : 울산광역시를중심으로 19 이지원 김태형서울시고령운전자교통사고특성분석 35 서성혁 정태열해안곡 ( 谷 ) 계마을을대상으로한위요공간규모에관한연구 49 이진희 김현우 Analysis of Stormwater Runoff Reduction Effect Through Applying Low Impact Development Practices in a Flood Prone Area : Case of Incheon, South Korea 69 이종소 이상은 통신빅데이터를활용한국가하천친수지구의이용등급, 상세유형화 및친수거점지구지정방법개발 83 이광현 7 개특 광역시의고령친화도와고령친화정책비교에관한연구 99 황종아 구자훈서울시취약계층밀집지역분포와주거지특성분석 : 민간임대주택거주기초생활수급자를중심으로

4

5 국토연구제 102 권 (2019.9): pp3~17 개발제한구역지정의생태적효과 : 울산광역시를중심으로 * Ecological Effects of Greenbelt Designation in Ulsan Metropolitan Area 김동욱 Kim Donguk ** Abstract Greenbelt was designated by Korean government to prevent indiscreet urban sprawling for 14 large cities in Importunate claims on deregulation of greenbelt system was continued in these days. But, ecological role of greenbelt as fundamental information for balanced decision making was not clarified sufficiently yet. In Ulsan metropolitan area, we clarified the roles of greenbelt system and prepared the sustainable management and improvement plans by analyzing ecological characteristics of vegetation established in Greenbelt and its surrounding area. Decrease of forest cover in greenbelt was lower than that of exterior and interior of greenbelt. Fragmentation patterns of forest cover corresponded with the pattern of forest cover. NDVI of greenbelt was higher than that of inner greenbelt and similar to that of exterior greenbelt. Surface temperature showed similar trend to that of NDVI. In consequence, greenbelt system has carried out well its primary roles. To mitigate negative effects from urbanization, we discussed management and improvement plans to promote ecological quality of greenbelt and urban landscape. Keywords: Greenbelt, NDVI, Surface Temperature, Urbanization, Landscape Change I. 서론 1. 연구의배경및목적도시화는세계적으로환경질과생태계패턴에많은영향을주고있다 (Piorr, Ravetz and Tosics 2011; Wu 2014; Salvati, Ranalli, Carlucci and Achille et al. 2017). 이러한영향은토지이용변화, 토지황폐화, 서식지파편화, 생물다양성감소및습지와야생동물의서식처파괴의원인이된다 (Grigorescu and Geacu 2017). 빠르게진행되는도시의확장으로도시화지역의기온이 상승하여열섬현상이일어나고 (Mirzaei, Haghighat, Nakhaie and Yagout et al. 2012), 그영향이대기수분함량의감소로이어져국지적인미기후변화가발생하고있다 (Souza, Santos Alvala and Nascimento 2016). 또한녹지지역의감소로산림분열을일으키며, 이러한단편화된숲의가장자리는광강도와바람에노출이되어가뭄스트레스를증가시킨다 (Murcia 1995; Oishi and Hiura 2017). 이러한변화로인해가뭄에민감한종들이사라지고, 외래종이증가하여생물다양성에부정적인영향을미친다 (Oishi and Hiura 2017). 전세계적으로도시가확장되고있고도시화가생 * 본논문은김동욱 (2013) 의석사학위논문 울산광역시개발제한구역의생태적효과평가에관한연구 를일부수정 보완한것임. ** 서울여자대학교환경생물학과박사과정 Ph.D. Candidate, Lab of Environmental, Seoul Women s Univ. xellosraw@naver.com 개발제한구역지정의생태적효과 3

6 물다양성과도시환경에적응하는종들에게미치는영향을연구하는것이생물다양성보전의핵심이다 (Kowarik and Lippe 2018). 또한, 어떤지역이생태적으로건전하게유지되기위해서는자연환경과인위적환경이발생시키는스트레스사이의기능적조화가필요하다 ( 이창석, 조용찬, 신현철, 이충화외 2006). 도시화의과정에서녹지공간은자연과의공간적연결성과다중기능성으로도심지역과구별될수있다 (Mabon and Shin 2018). 그리고보호지역은자연을유지하여생물다양성을보존하는데중요한역할을한다 (Sarmento and Berger 2017). 세계인구의절반이상인 54% 가도시에살고있으며 2050년까지 66% 이상으로증가할것으로예상된다 (United Nation 2014; Zinia 2018). 또한인간의활동이자연경관변화의근본적인원인을제공하며그로인한환경적변화는기후변화및생물다양성변화등으로나타나고, 도시화로인한자연의단절은생물다양성감소및멸종위기의동 식물들에게큰위협이된다 (Gbanie, Griffin and Thornton 2018). 따라서교란을받는지역의교란규모와강도의정보는생물다양성과환경보호뿐만아니라지속가능한발전에중요하게평가되고있다 (Thaden, Laborde, Guevara and Venegas-Barrera 2018). 우리나라또한 1960년대이후급속한산업화와함께도시가확장된결과인구가집중된도심부는녹지공간의확보가어려워환경의질이저하되었다 ( 최희준, 이정아, 손희정, 조동길외 2017). 도시팽창의과정에서녹지의가치에대한인식이부족하여대도시는물론중소도시에서도녹지손실로인한환경문제가심각하게나타났다 ( 박길용 2003). 우리나라는이러한문제를해결할목적으로 1971년개발제한구역을지정하였다. 이제도는무분별한도시확장을방지하고, 도시주변의자연보호를목적으로설계되었다 (An 2015). 초기에지정된개발제한구역은서울을비롯하여 14개도시에총 5,397km 2 이었다. 그러나현재는그중 7개도시 ( 서울, 대전, 대구, 부산, 울산, 광주, 창원 ) 에서만그제도가유지되고있다. 개발제한구역은새로운건물의건설이나임업, 특별여가활동금지등개발을직접적으로제약함으로써, 도시확장을봉쇄하는가장제한적인방법중하나로여겨진다 (Siedentop, Fina and Krehl 2016). 개발제한구역제도는 1950년영국런던주위에고리모양의녹지구역을지정하면서시작되었다. 유럽은개발제한구역이유럽대륙을가로질러형성되어있고, 페노스칸디아개발제한구역 (Fennoscandian Greenbelt), 중부유럽개발제한구역 (The Greenbelt Central Europe), 발칸개발제한구역 (The Balkan Greenbelt) 의세구역으로나누어관리하고있다 (Geidezis and Kreutz 2012). 유럽대륙의개발제한구역에 39개의국립공원이존재하고개발제한구역주변으로 25km의완충지대가있으며 3,200개이상의자연보존지역이있다 (Geidezis and Kreutz 2012). 이러한유럽의개발제한구역은향후수십년동안유럽의자연보호및생물다양성보호를위해중요한연구대상이될것이며대규모의생물다양성감소를막을수있는수단으로여겨진다 (Geidezis and Kreutz 2012). 2. 선행연구검토선행연구에는개발제한구역제도의문제점을제기한연구와개발제한구역의필요성을언급하는연구들이있고, 개발제한구역지정후시기별로도차이가나타난다. 개발제한구역에대해부정적인경우, 대도시의인구억제에기여하지못했고, 개발제한구역중녹지가아닌면적이많고개발제한구역으로인하여토지가부족하여지가가상승한다는등의문제점을제기 4 국토연구제 102 권 ( )

7 한연구가있다. 긍정적인경우, 개발제한구역에도시확산방지와쾌적효과, 환경적편익에긍정적영향을주었음을주장한다. 시기별연구를보면, 제도도입시기인 1970년대와 1980년대에는주로도시의확산이나난개발등도시문제를해결하기위한수단으로제도도입의필요성을지적했고, 1990년대들어서면서개발제한구역내주민의재산권제한과불편해소의관점에서제도개선에대한논의가증가했다 ( 이성원 2018). 양병이 (1992) 는개발제한구역이도시의무질서한팽창을억제하는긍정적인측면과기본목적과는달리대도시주변의도시들의광역화를초래하였고, 기존시가지의과밀을초래하여토지비용상승을초래하는부정적측면을지적하며개선방안을제시하였다. 최병선 (1993) 은개발제한구역이도시의무질서한확산방지와도시주변의자연환경보전이라는지정당시의목적은충분히달성하였지만개발제한구역안과밖의개발차이로인한주민들의불이익이증폭되고있다고지적했다. 김경환 (1998) 은개발제한구역제도를평가하여근본적인개선논의의근거를제시하였다. 평가의기준은첫째, 지정목적의효과적수행, 둘째, 지정목적의타당성, 셋째, 사회적인편익과비용, 넷째, 편익의수혜자와비용의부담주체에관하여평가하였고목적달성을위해큰사회적비용을초래한반면수혜자는분명치않다고지적하였다. 김재익, 여창환, 박선형 (2007) 은개발제한구역지정으로도시내부의개발가능지가줄어들면서개발제한구역외곽도시의발전이가속화되는추세라고주장했다. 박상규, 김창석 (2009) 은남양주시를대상으로개발제한구역해제가토지이용변화에미치는영향을연구하였고개발제한구역의해제유형보다해제대상지역의향후토지이용계획의내용에따라지표가달라진다는것을제시하였다. 이와같이이전의선행연구는개발제한구역의목적성취여부및부작용등에초점을맞춰토지이용의공간적분포가미친영향에관한연구가주를이루었다. 개발제한구역에관한생태학적연구사례로경관생태학및복원생태학관점에서생태학적평가를한연구 (Cho, Cho and Lee 2009), 그린벨트의생태계서비스가치평가연구 ( 류대호, 이동근 2013), 대기오염및지표온도를통한환경적평가연구 ( 김희재 2016) 등이있다. 우리가모방한유럽의개발제한구역은생태적인연구를통해얻은정보에기준하여체계적인국가및대륙의차원에서관리와연구가이루어지고있다. 반면에우리나라는개발제한구역지정후약 50년이지났지만개발제한구역지역의토지이용분포가미치는영향에대한연구가대부분을차지하고, 개발제한구역의생태를비롯하여그것이가져오는생태적인효과에대한연구가매우미흡한실정이다. 도시숲의효과적인복원및개선을위해개발제한구역에대한생태적구조및기능에관한연구는지속적으로이루어질필요가있다 (Cho, Cho and Lee 2009). 이를위해본논문은개발제한구역의목적중하나인도시의확장을억제하는역할에대하여우리나라최대공업도시인울산광역시를대상으로개발제한구역지정의효과를생태적관점에서평가하였다. II. 연구및방법 1. 조사지개황울산광역시는우리나라 7대도시중하나로면적이 1,060km 2 를차지하고, 인구는약 120만명으로우리나라의특별시와광역시중가장넓은지역이다 (<Figure 개발제한구역지정의생태적효과 5

8 Figure 1 _ Location Map of Ulsan Area 1> 참조 ). 울산은태화강을중심으로시가지가형성되었고방어진, 염포, 장생포항을중심으로도시가형성되어왔으며, 서쪽으로는경상남도밀양시, 경상북도청도군, 남쪽으로는부산광역시기장군, 경상남도양산시, 북쪽으로는경주시를접하고있다. 울산항을포함한울산본항, 미포항, 장생포항, 울산신항, 방어진항, 정자항등을중심으로해안지역은공업용수가풍부하고지반이경암질로되어있어공장건설은물론산업도시로발전할수있는천혜의입지조건을갖추고있다. 이러한지리적여건으로울산은지난 50년가까이우리나라경제성장을주도해온세계적인산업도시이며, 고속철도와역사개통에이어앞으로 20선석의신항만을울산항에추가건설하여새로운항만항만경제권을형성해나갈계획이다 ( 울산광역시청 ) 1). 2003년은 Landsat 7, 2016년은 Landsat 8 위성영상을사용하였다. 1975년, 1984년, 1993년, 2003년및 2016 년도의토지이용유형, NDVI(Normalised Difference Vegetation Index, 정규식생지수 ) 및지표온도는 ArcGIS 10.1과 ecognition Developer를이용하여분석하였다. 경관의변화패턴은 Fragstats 4.2를이용하여울산광역시개발제한구역의경관요소를활엽수림과침엽수림으로구분하여분석하였다. 분석한지수는면적분포의변화를알수있는패치수 (Number of Patch: NP), 패치의자연성을판단할수있는경관형태지수 (Landscape Shape Index: LSI), 패치의연결성및파편화를판단할수있는결합지수 (CHOESION Index) 를경관의변화를분석하였다. 개발제한구역지정에의한생태적효과는경관구조와기후완화기능차원에서공간적차이와시간적차이를통해검토하였다. 공간적차이는개발제한구역지역 (GB) 에서얻은결과와개발제한구역으로부터도시외곽으로설정한구역 (Exterior of Greenbelt: EGB) 과내부지역 (Interior of Greenbelt: IGB) 으로구분하여비교하였고, 시간적차이는 1910년부터 2016년까지 100년동안의변화를분석하였다. III. 결과 2. 경관의구조및질 1. 경관구조의공간적차이 1975년, 1984년, 1993년, 2003년및 2016년도의 Landsat 위성영상을분석하여개발제한구역지역의경관구조를분석했다. 1910년은조선총독부에서발행한고지형도를이용하여토지이용분류를하였다. 1975년은 Landsat 4, 1984년과 1993년은 Landsat 5, 개발제한구역지역 (GB) 은침엽수림이가장넓은면적을차지하였고, 낙엽활엽수림, 농경지순으로나타났다. 개발제한구역내부지역 (IGB) 은도시화지역이가장넓은면적을차지하였고, 침엽수림, 활엽수림및농경지가유사한면적으로나타났다. 개발제한구역 1) (2019 년 1 월 17 일검색 ). 6 국토연구제 102 권 ( )

9 Figure 2 _ Maps Showing the Land Cover Change in GB Around the Ulsan Metropolitan Area 외부지역(EGB)은 침엽수림과 활엽수림이 유사한 면 화단지의 조성으로 인해 도시화 면적이 증가하였다 적으로 넓은 면적을 차지하였고, 농경지, 도시화지역 (<Figure 3> 참조). 순으로 나타났다. 1910년부터 2016년까지 100년간 울산광역시의 2. 지표온도의 시간적 차이 경관구조를 분석한 결과를 그림으로 나타내면 <Figure 2>와 같다. 도시화 지역의 면적이 32.1km2 2 위성영상을 이용하여 공간별 지표온도를 분석한 결 에서 171km 로 증가함에 따라 가장 크게 변화한 것 과를 그림으로 나타내면 <Figure 4>와 같다. EGB와 으로 나타났다. 주거지역, 산업시설 및 공업지역이 IGB의 최고온도는 40 C 이상, GB의 최고온도는 밀집해 있는 IGB의 경우 지속적인 인구유입으로 도 37 C로 나타났다. 도심에서부터 개발제한구역 지역, 시화면적이 크게 증가하였다. GB의 경우 침엽수림 교외지역에 이르기까지 거리에 따른 온도 변화를 비 이 감소하고 활엽수림이 증가하는 경향을 보였다. 교한 결과, 지표온도는 도심에서 개발제한구역 지역 EGB의 경우 GB와 마찬가지로 침엽수림이 감소하 에 이르기까지 지속적으로 감소하는 경향을 나타냈 고 활엽수림이 증가하였고, 새로운 주거단지와 산업 고, 개발제한구역과 교외지역은 유사한 온도를 유지 개발제한구역 지정의 생태적 효과 7

10 Figure 3 _ Areal Changes of the Major Landscape Elements in Ulsan Figure 4 _ Maps Showing the Surface Temperature Change in GB Around the Ulsan Metropolitan Area 국토연구 제102권( )

11 Figure 5 _ Surface Temperature Diagram of Ulsan Metropolitan City Extracted Using Landsat TM Satellite Imagery 하는경향을보였다 (<Figure 5> 참조 ). 는경향을보였다 (<Figure 7> 참조 ). 3. 식생의변화 2) 경관의변화 1) 식생활력도변화 NDVI 에근거하여식생의활력도를평가한결과를그림으로나타내면 <Figure 6> 과같다. GB와 EGB는유사한수준의활력도가나타났고, IGB의경우대부분도시화지역으로이루어져식생의활력도가낮게나타났다. 이러한결과를도심에서부터개발제한구역지역, 교외지역에이르기까지거리에따른변화로비교한결과, 식생활력도지수는도심에서개발제한구역까지지속적으로증가하는경향을나타냈고, 개발제한구역에서교외지역까지는유사한상태를유지하 Fragstat을이용하여경관패치를분석한결과를그림으로나타내면 <Figure 8> 과같다. 패치의수는 EGB에서침엽수림과활엽수림이가장높게나타났다. LSI의활엽수림은 2003년이후증가했고침엽수림은 1910년이후크게감소했다. COHESION Index는 GB와 EGB 는같은수준을유지했고 IGB는크게감소했다. 3) 종다양성 IGB( 곰솔, 아까시나무군락 ), GB( 소나무, 굴참나무, 신갈나무군락 ), EGB( 굴참나무, 졸참나무군락 ) 조사지역 개발제한구역지정의생태적효과 9

12 Figure 6 _ Maps Showing the NDVI Change in GB around the Ulsan Metropolitan Area for about 20 Years since GB Designation Figure 7 _ NDVI Diagram of Ulsan Metropolitan City Extracted Using Landsat TM Satellite Imagery 10 국토연구 제102권( )

13 Figure 8 _ Change of Landscape Index of Forest Cover in IGB, EGB and GB of Greenbelt around Ulsan Metropolitan Broad-leaved Forest Needle-leaved Forest Figure 9 _ Species Rank-abundance Curves of IGB, EGB and Greenbelt in Ulsan 의종다양성을종순위- 우점도곡선을작성하여비교하였다 (<Figure 9> 참조 ). 분석결과, IGB에서종풍부도가극히낮게나타났고 GB 및 EGB 지역순으로높은경향을나타냈다. 4) 종조성 IGB, EGB 및 GB의종조성특성을밝히기위하여각구역에방형구를설치하여수집된자료를분석하였다 (<Figure 10> 참조 ). 서열화결과, II축상에서 IGB, EGB 및 GB의순서로배열되는경향이나타났 개발제한구역지정의생태적효과 11

14 Figure 10 _ Ordination of IGB, GB and EGB in Ulsan Area Different in the Degree of Artificial Disturbances 다. EGB 지소중왼쪽에치우친지소들은졸참나무와갈참나무의식피율이높게나타났고, 오른쪽에치우친지소들은신갈나무와굴참나무의식피율이높아 GB의지소들과뭉치는경향이나타났다. IGB의지소중왼쪽에치우친지소는아까시나무와억새의식피율이높게나타났고, 오른쪽에치우친지소는곰솔과사스래피나무의식피율이높게나타났다. 10.1cm 이상에서정규분포형을나타냈고, 그이하의계급에는출현하지않았다. 참나무류와때죽나무가 10.0cm 이하에서출현했다. 곰솔군락을이루는주요종의직경급별빈도분포에서 IGB에서 10.1cm 이상에서정규분포형을나타냈다. 10.0cm 이하의직경급에서곰솔은출현하지않았고참나무류가 7% 의빈도로낮게출현했다. GB에서 10.1cm 이상에서정규분포형을나타냈다. 10.0cm 이하에서참나무류가 50% 이상의높은빈도로출현했다. EGB에서 10.1cm 이상에서정규분포형으로나타났고, 10.0cm 이하에서참나무류가역 J형분포를나타냈다. 졸참나무군락을이루는주요종의직경급별빈도분포에서모든계급에서정규분포형을나타냈다. GB 에서 10.1cm 이상에서낮은빈도로출현했지만 5.0cm 이하의계급에서높은빈도로출현해유식물이계속보충되고있음을확인할수있었다. IV. 결론 5) 직경계급분포소나무, 곰솔, 졸참나무군락에서수집한매목조사데이터를기반으로주요교목성수종의흉고직경그래프를작성하였다 (<Figure 11> 참조 ). 소나무군락을이루는주요종의직경급별빈도분포에서소나무는 IGB에서 10.1cm 이상에서정규분포형을나타냈다. 10.0cm 이하에서소나무가 9% 의빈도로출현했고참나무류, 때죽나무및팥배나무가 45% 의높은빈도로출현했다. GB에서소나무는모든직경급에서출현했고 10.1cm 이상에서정규분포형을나타냈다. 10.0cm 이하에서참나무류가역 J형분포를보여유식물이보충되고있음을확인할수있었다. EGB 에서소나무는 이논문은인공위성영상을기반으로분석했고, 현장에서수집한식생자료가부족하여영상을활용한경관변화를분석하는데한계가있다. 그러나개발제한구역과같은보호지역에대한연구는지속가능한발전을목표로하는현시대의발전방향을위해서중요하다는점에서연구에의미가있다. 울산광역시는 1960년대초반경제개발 5개년계획수립과함께본격적으로발전이시작되었고, 도시화로인구가급격히증가하여토지의경관은농촌경관에서도시경관으로변하기시작하였다. 도시가급격히확장하면서도시공간구조는기능적인연계성과공간적인통합성이부족하게되었다 ( 최금식, 여홍구 2015). 12 국토연구제 102 권 ( )

15 Figure 11 _ Frequency Distribution of Diameter Classes of Major Tree Species of Forests in Ulsan 개발제한구역지정의생태적효과 13

16 이러한울산광역시에지정된개발제한구역의효과는다음의두가지로요약된다. 첫째, 경관적인요인으로울산광역시의개발제한구역은주거지역및공업지역이밀집해있는도심지역과농경지역인교외지역을분리시키면서도심의무분별한확장을억제하고있다. 경관변화를분석한결과 1970년이후패치의수가증가하여파편화의경향을보이지만패치의연결성을나타내는 COHESION Index 는 GB와 EGB에서같은수준을유지하고 IGB에서는지속적으로감소하였다. 이결과로 IGB에분포하는산림패치의연결성이약해지고분산되었다는것을알수있고, 이는도시화의면적이증가하고산림면적이감소한것을의미한다. 같은기간동안의 GB와 EGB의산림면적변화는시간의흐름에따라침엽수림이활엽수림으로변하는천이의과정으로나타났다. 즉, 개발제한구역이주거지역및공업지역의확장을 IGB 내로제한시키며교외지역에가해지는도시성장의압박을효과적으로억제하고있다. 둘째, 이논문에서수집된자료를바탕으로식생동태를분석한결과, 자연적인천이가이루어지고있음을확인하였다. 그러나개발제한구역을제외한군락에서때죽나무, 팥배나무등의출현빈도가높게나타났다. 이수종은인간간섭이빈번히발생하고오염이심한지역에번성하는종으로환경오염과인위적인간섭에노출되어있음을알수있다. 군락내종다양성은 IGB에비해 GB와 EGB에서높은수준으로나타난다. 종다양성이높은것은종들이서식하는환경이다양한것이고환경이다양하면유전자의다양성이높아져높은생물다양성을갖추게된다 ( 안지홍, 임치홍, 정성희, 김아름외 2016). 생물다양성이높은공간은생물간상호작용의다양성이높고, 인간과의관계의다양성또한높게나타난다. 따라서높은종다양성을유지하는개발제한구역과교외지역의관리에대 해지속적인연구가수행되어야할것으로판단된다. Dwivedi and Mohan(2018) 은도시의대형구조물과식생이많은국립공원및공원에서매우명확한온도구배차이를확인했다. 도시의열섬현상은토지이용및식생면적에의해결정된다고해석했다. 이논문에서수행한지표온도분석은울산광역시전체적인부분을확인하여온전한식생이많은개발제한구역, 교외지역및도심지역의경향을분석하였다. 시간이지나면서전체적인온도가상승하는경향을보였지만개발제한구역과교외지역은도심지역보다낮은온도를유지하는것을확인하였다. 즉, 식생이지표온도를조절하는역할을하는것을확인하고열섬현상이심한도시지역의문제를해결하는데중요한참고자료를제공할수있다. 참고문헌 1. 김경환 개발제한구역제도의평가와제도개선쟁점. 한국주택학회 6권, 2호 : Kim Gyunghwan An evaluation of greenbelt policy and major issues for reform. Korean Association for Housing Policy Studies 6, no.2: 김재익, 여창환, 박선형 개발제한구역의개발가능지분석과도시성장관리에대한시사점. 국토계획 42권, 3호 : Kim Jaeik, Yeo Changhwan and Park Sunhyung Evaluating the urban growth restriction function of the greenbelt and its implication for urban growth management. Journal of the Korea Planning Association 42, no.3: 김희재 수도권그린벨트의환경적효과분석 : 대기오염및지표온도를중심으로. 박사학위논문, 중앙대학교. Kim Heejae. An Analysis of the Greenbelt s Environmental Effect in the Seoul Metropolitan Area: Focusing on the ambient air pollution and land surface temperature. Ph.D. diss., Chung-Ang University. 4. 류대호, 이동근 수도권그린벨트지역의생태계서비 14 국토연구제 102 권 ( )

17 스가치평가연구. 국토계획 48권, no.3: Ryu Daeho and Lee Dongkun Evaluation on economic value of the greenbelt s ecosystem services in the Seoul metropolitan region. Journal of the Korea Planning Association 48, no.3: 박길용 지속가능한도시공원녹지정책. 한독사회과학논총 13권, 2호 : Park Gilyong The green policy for a sustainable urban park. Zeitschrift der Koreanisch-Deutschen Gesellschaft Fur Sozialwissenschaften 13, no.2: 박상규, 김창석 개발제한구역해제가토지이용변화에미치는영향 : 남양주시사례를중심으로. 국토연구 61권 : Park Sangkyu and Kim Changseok The influence of the relaxation of the greenbelt on urban land use change: Case of Namyangju-si. The Korea Spatial Planning Review 61: 안지홍, 임치홍, 정성희, 김아름, 이창석 생물다양성에대한기후변화의영향과그대책. 한국습지학회지 18권, 4호 : An Jihong, Lim Chihong, Jung Songhie, Kim Areum and Lee Changseok Effects of climate change on biodiversity and measures for them. Journal of Wetlands Research 18, no.4: 양병이 우리나라그린벨트관리의효율화를위한방안 : 서울시그린벨트를중심으로. 환경논총 30권 : Yang Byounge Methods for the effective management of Green Belt in Korea. Korean Journal of Public Administration 30: 울산광역시청. 울산소개 : 울산의오늘. kr/rep/uhnature (2019년 1월 17일검색 ). Ulsan City. About Ulsan. ure (accessed January 17, 2019). 10. 이광국 개발제한구역의생태적관리에관한연구. 건설환경연구 3집, 3호 : Lee KwangKug A study on ecological management of greenbelt areas. Korean Society of Construction and Environmental Engineers 3, no.3: 이성원 개발제한구역제도가도시확산방지에미친영향. 국토계획 53권, 2호 : Lee Sungwon The impacts of greenbelt policies on anti-sprawl. Journal of Korea Planning Association 53, no.2: 이창석, 조용찬, 신현철, 이충화, 이선미, 설은실, 오우석, 박성애외 국가장기생태연구장소로서구축된남산소나무림의생태적특성. Journal of Ecology and Environment 29권, 6호 : Lee Changseok, Cho Yongchan, Shin Hyuncheol, Lee Choonghwa, Lee Seonmi, Seol Eunsil and Oh Wooseok et al Ecological characteristics of Korean Res Pine (Pinus densiflora S. et Z.) forest in Mt. Nam as a long term ecological research (LTER) site. Journal of Ecology and Environment 29, no.6: 최금식, 여홍구 울산광역시도심의토지이용특성에관한연구. 도시행정학보 28권, 4호 : Choi Keumsik and Yuh Hongkoo A study on land use spatial structure characteristics of the City of Ulsan. Journal of The Korean Urban Management Association 28, no.4: 최병선 지역및도시개발의문제와대책 : 그린벨트관리제도의문제점과개선방향. 대한건축학회지 37권, 2호 : Choe Byungsun Problems and prospects of greenbelt menagement system in Korea. Review of Architecture and Building Science 37, no.2: 최희준, 이정아, 손희정, 조동길, 송영근 도시정원도입을위한고밀시가화지역내녹지네트워크구축가능성평가. 한국환경생태학회지 31권, 2호 : Choi Heejoon, Lee Junga, Sohn Heejung, Cho Donggil and Song Youngkeun. Feasibility of green network in a highly-dense urbanized area by introducing urban gardens. Korean Journal of Environment and Ecology 31, no.2: An Yehyun The effects of greenbelt policy in Seoul, Korea. 5/10/01/the-effects-of-greenbelt-policy-in-seoul-korea (accessed, January 17, 2019). 17. Cho Yongchan, Cho Hyunje and Lee Changseok Greenbelt systems play an important role in the prevention of landscape degradation due to urbanization. Journal of Ecology and Environment 32, no.3: Dwivedi, A and Mohan, B. K Impact of green roof on micro climate to reduce urban heat island. Remote Sensing Applications: Society and Envrionment 10: Gbanie, S. P., Griffin, A. L. and Thornton, A Impacts 개발제한구역지정의생태적효과 15

18 on the urban environment: Land cover change trajectories and landscape fragmentation in post-war western area, Sierra Leone. Remote Sensing 10, no.1: Geidezis, L. and Kreutz, M Greenbelt Europe - Structure of the initiative and significance for a pan European ecological network. Proceedings of the 1st GreenNet Conference, January 31, Erfurt. 21. Grigorescu, I. and Geacu, S The dunamics and conservation of forest ecosystems in Bucharest Metropolitan Area. Urban Forestry & Urban Greening 27: Kowarik, I. and von der Lippe, M Plant population success across urban ecosystem: A framewaok to inform biodiversity conservation in cities. Journal of Applied Ecology 55: Mabon, L. and Shin, W. Y What might just green enough urban development mean in the context of climate change adaptation? The case of urban greenspace planning in Taipei Metropolis, Taiwan. World Development 107: Mirzaei, P. A., Haghighat F., Nakhaie, A. A., Yagouti, A., Melissa, G., Keusseyan, R. and Coman, A Indoor thermal condition in urban heat island: Development of a predictive tool. Building and Environment 57: Murcia, C Edge effects in fragmented forestes: Implications for conservation. Trends in Ecology and Evolution 10: Oishi, Y. and Hiura, T Bryophytes as bioindicators of the atmospheric environment in urban-forest landscapes. Landscape and Urban Planning 167: Piorr, A., Ravetz, J. and Tosics, I Peri-urbanisation in Europe: Towards a European policy to sustain urban-rural futures. Frederiksberg: Forest & Landscape, University of Copenhagen. 28. Salvati, L., Ranalli, F., Carlucci, M., Achille, l., Ferrara, A. and Corona, P Forest and the city: A multivariate analysis of peri-urban forest land cover patterns in 283 European metropolitan areas. Ecological Indicators 73: Sarmento, W. M. and Berger, J Human visitation limits the utility of protected areas as ecological baselines. Biological Conservation 212: Souza, D. O., Santos Alvala, R. C. and Nascimento, M. G Urbanization effects on the microclimate of Manaus: A modeling study. Atmospheric Research 197: Siedentop, S., Fina, S. and Krehl, A Greenbelts in Germany s regional plans: An effective growth management policy? Landscape and Urban Planning 145: Thaden, J. J., Laborde, J., Guevara, S. and Venegas-Barrera, C. S Forest cover change in the Los Tuxtlas biosphere reserve and its future: The contribution of the 1998 protected natural area decree. Land Use Policy 72: United Nations(UN) World Urbanization Prospects. New York: United Nations. 34. Wu, J Urban ecology and sustainability: The state-of-the-science and future directions. Landscape and Urban Planning 125: Zinia, N. J. and McShane, P Ecosystem services management: An evaluation of green adaptations for urban development in Dhaka, Bangladesh. Landscape and Urban Planning 173: 논문접수일 : 심사시작일 : 심사완료일 : 국토연구제 102 권 ( )

19 요약 주제어 : 개발제한구역, 경관변화, 열섬현상, 대기정화 개발제한구역은대도시의무분별한도시확장을막기위해 1971년 14개의도시에지정되었다. 현재중 소도시의개발제한구역은해제되고 7개의대도시권중심으로개발제한구역이시행중이다. 개발제한구역제도의규제철폐에관한주장은최근에도지속되고있다. 그러나올바른의사결정을위한근본정보로서개발제한구역에관한생태학적역할은아직충분히연구되지않았다. 개발제한구역시스템의역할을명확히하고지속가능한보전과경영개선계획을수립하기위해한국의수도서울을포함한대도시지역의개발제한구역과그주변지역에구성되어있 는녹지의생태적특성을분석했다. 대부분대도시산림피복은감소했고개발제한구역보다개발제한구역의외부및내부의감소가더컸다. 산림피복의파편화패턴은산림피복의변화패턴과일치했다. 개발제한구역의 NDVI 는도심지보다높았으며외부지역과유사했다. 지표온도는 NDVI와비슷한경향을보였다. 결과적으로개발제한구역시스템은주요역할을잘수행했다. 도시화로인한부정적영향을줄이기위해개발제한구역과도시경관의생태적품질을향상시키기위한관리및개선계획에대한지속적인연구가필요하다. 개발제한구역지정의생태적효과 17

20

21 국토연구제 102 권 (2019.9): pp19~34 서울시고령운전자교통사고특성분석 * Examining the Characteristics of Traffic Accidents Involving Elderly Drivers in Seoul, South Korea 이지원 Lee Jiwon **, 김태형 Gim Tae-Hyoung Tommy *** Abstract This study aims to examine the characteristics of traffic accidents involving elderly drivers with regards to the urban structure in Seoul, South Korea. To this end, this research analyzed the traffic accident data of the year 2015 collected by the Korean Police Agency and Koroad(TAAS). The traffic accident records being collected in a spatial unit, spatial heterogeneity and non-dispersion are prone to occur. In order to prevent this, this study employs ordinary least squares regression(ols), a conventional spatial model for analyzing accident data, and geographically weighted regression(gwr), which is an alternative to the spatial econometrics model. It was revealed that the GWR model has a better goodness-of-fit than the OLS model. According to the GWR model, the population aged under 20 ratio and the industrial area ratio are negatively associated with traffic accidents in which elderly drivers are at fault. Meanwhile, employment density, commercial area ratio and the numbers of subway stations, bus stops, intersections, and traffic islands are positively associated with the traffic accidents involving elderly drivers. Keywords: Traffic Accident, Geographically Weighted Regression, Elderly Driver, Spatial Econometrics I. 서론도로교통공단의 교통사고현황 에따르면 2011~2015 년간전국에서발생한만 65세미만비고령운전자의교통사고사망자수는 4,594명에서 3,802명으로 17.2% 감소하였다. 반면, 고령운전자에의한교통사고사망자수는지난 5년간 605명에서 815명으로 34.7% 상승했고교통사고발생건수또한 2011년대비 2배가량증가하여비고령운전자의교통사고감소추세와대 조적이다 ( 도로교통공단교통사고분석시스템 ). 1) 우리나라는 1991년교통사고사망자수가 12,325 명을기록한이래로도로환경개선및교통안전정책수립등다방면에서노력한끝에교통사고사망자수를 1991년의 30% 수준 (2015년기준 ) 까지감소시킬수있었다. 그럼에도불구하고여전히우리나라가 OECD 회원국중교통사고사망률이상위권에머무는데는고령운전자의교통사고및사망자수증가가주요원인으로작용하였다고볼수있다. 2) * 이논문은 2019년대한민국교육부와한국연구재단의지원을받아수행된연구임 (NRF-2018S1A5A ). ** 서울대학교환경대학원박사과정 ( 제1저자 ) Ph.D. Candidate, Graduate School of Environmental Studies, Seoul National Univ. Primary Author aimeru2@snu.ac.kr *** 서울대학교환경대학원 협동과정조경학부교수및환경계획연구소부소장 ( 교신저자 ) Associate Prof., Graduate School of Environmental Studies and Interdisciplinary Program in Landscape Architecture and Associate Director, Environmental Planning Institute, Seoul National Univ. Corresponding Author taehyoung.gim@snu.ac.kr 서울시고령운전자교통사고특성분석 19

22 고령운전자는신체적 심리적노화로인하여인적요인이주행중미치는영향이크기때문에, 교통사고가발생했을때심각한인명피해로이어질가능성이높다 (Ferguson and Braver 2002; Braver and Trempel 2004; Boufous, Finch, Hayen and Williamson 2008; 최재성 2018). 따라서동일한조건에서발생한사고라도고령운전자는비고령운전자와는다른사고특성을보일수있어 (Langford and Koppel 2006), 고령운전자의교통사고발생과이로인한인명피해를감소시키기위해서는고령운전자의주행에영향을미치는공간에대한다양한요인을고려한분석이필요하다. 고령운전자는비고령운전자에비해서신체능력과심리적요인변화등인적요인이주행에미치는영향이상당하다. 예를들어, 고령운전자의시력은 30~40대에비해서 20% 이상떨어지며, 시력저하로인한고령운전자의가용시각장 (Useful Field of View) 감소는교통사고발생가능성을비고령운전자에비해 6배가량높이는것으로나타났다 (Ball, Beard, Roenker and Miller et al. 1993; Huisingh, Levitan, Irvin and MacLennan et al. 2017). 이에고령자의교통사고감소를위하여고령자의인적요인과주행환경 ( 기상상태, 도로유형, 노면상태등 ) 을중심으로교통사고특성 분석이이루어져왔다. 1) 2) 교통사고를선제적, 효과적으로대응하기위해서는주행환경을도로환경에국한하지않고도시의공간적특성을반영하는여러요인을투입하여도시계획적접근방식을도모할필요가있다. 이러한차원에서이논문은도시환경요인을포함한도시공간을구성하는요인들이고령운전자의교통사고에미치는영향을파악하고고령운전자의교통사고특성을규 명하고자하는데목적이있다. 한편과거교통사고분석에사용해온전통적인최소제곱법 (Ordinary Least Squares Estimation: OLS) 에근거한선형회귀기반의모수추정은교통사고자료와같이공간단위로수집되는자료를분석하는경우, 공간적이질성 (Spatial Heterogeneity), 비분산성 (Non-dispersion) 문제에맞닥뜨릴수있다. 이에본논문에서는공간적특성을반영하여고령운전자의교통사고발생요인을규명하는지리적가중회귀모형 (Geographically Weight Regression: GWR) 을최소제곱모형과함께투입하여연구를진행하였다. 연구의분석대상은 2015년서울시에서발생한고령운전자교통사고이며고령운전자는만 65세이상으로한정한다. II. 선행연구고찰 1. 고령운전자교통사고고령운전자에관련된연구는교통, 의학, 심리학등의분야에서주로연구되어왔다. 특히고령자의신체능력저하와심리적변화가주행중미치는영향을규명한연구가주를이루었다. 고령자의신체적노화는고령운전자의교통사고발생에큰영향을미치는것으로나타났으며, 고령자의시력저하, 가용시각장감소, 청력저하등으로고령운전자는주행중사각지대에위치한사람혹은물체를인지하는데어려움을겪게된다. 이는주행중예기치못한상황이발생했을때고령운전자의즉각적인대응을어렵게하여교통사고발생시심각한인명피해를초래할수있다 1) (2019년 8월 28일검색 ). 2) 도로교통공단 (2018) 에따르면, 2016년기준한국의교통사고사망자수는 10만명당 10명으로 OECD 회원국평균의 1.5배임. 65세이상인구 10만명당교통사고사망자수의경우, 25.6 명 ( 회원국중 1위 ) 으로 OECD 회원국평균인 8.5명보다약 3배가량높음. 20 국토연구제 102 권 ( )

23 ( 오주석, 이의용, 류준범, 이원영 2015; 이원영, 김기홍, 오주석 2015). 고령운전자의신체노화는정신적노화에영향을미친다. 특히고령운전자의인지반응시간을저하시켜주행중제한된시간내에즉각적인대응을요구하는상황에서교통사고발생가능성을높인다. 예를들어통행량이많으며복잡한상황이종종발생하는대형교차로나좌회전중고령운전자는충분한인지반응시간을갖지못하고 ( 후방차량의상향등및경적등으로인한 ) 심리적부담감으로인해무리하게좌회전을시도해서교통사고를일으킬가능성이있다 (Preusser, Williams, Ferguson and Ulmer et al. 1998; Zhang, Lindsay, Clarke and Robbins et al. 2000; 이순철 2006; Boufous, Finch, Hayen and Williamson 2008). 이순철 (2006) 은고령운전자가젊은운전자에비해서운전확신수준에차이가있으며, 이로인해좌회전시소요되는시간이현저하게길어진다고주장하였다. 이와관련하여장재민, 최재성, 김태형 (2017) 은고령운전자의연령이높아질수록운동신경저하로인해조향에어려움을느끼며, 인지반응시간이길어짐에따라자기판단에의한사고율이높게나타난다고밝힌바있다. 더불어고령운전자는주행중환경요인 ( 교통량, 주행속도, 도로유형, 토지용도등 ) 의영향을민감하게받아환경에대한심리적인부담감이상대적으로높다. 특히 Hakamies-Blomqvist and Henriksson(1999), Boufous, Finch, Hayen and Williamson(2008), 이순철 (2006) 은교차로와같이주행중고려할것이많은복잡한환경에서발생하는, 고령자에의한교통사고를고령운전자의주요교통사고특성으로보았다. 그외이상혁, 정우동, 우용한 (2012) 는신호기설치여부, 차도폭, 중앙분리대설치여부와같이도로환경이고령운전자의교통사고발생에유의미한영향을미침을밝힌바있다. Author Analysis Model Spatial Range Explanatory variable Oh, Lee, Ryu and Lee(2015) Jang, Choi and Gim(2017) Lee, Jeung and Woo(2012) Boufous, Finch, Hayen and Williamson(2008) Langford and Koppel(2006) Zhang, Lindsay, Clarke and Robbins et al.(2000) Table 1 _ Previous Studies on Traffic Accident Involving Elderly Drivers Cross-sectional Analysis Logistic Regression Model Ordered Logistic Regression Model Multivariate Linear Regression Model Frequency Analysis Logistic Regression Model South Korea (National Wide) South Korea (unit: Metropolitan City, Province) Daegu(Buk-gu, Nam-gu) New South Wales, Australia Australia Ontario, Canada Age, At-fault s Vehicle Type, Human Factors, Road Type, Accident Type Age of Driver, Road Type, Drink Driving, Accident Type Accident Type, Vehicle Type, Physical Demage, Protective Equipment, Speed, Drink Diving, Weather Condition, Road Surface Condition, Speed Limit, Road Width, Signal Light, Separating Facilities between Pedestrian Road-road Way, Median Strip Rurality, Location, Speed Limit, Road Surface Condition, Road Type, Signal Light, Light Condition, Vehicle Type, Accident Type Fatalities, Speed, Location, Intersection Type, Controls at Intersection, Light, Type of Vehicle, Vehicles in Crash, Restraint Use, Drink Driving, Crash Type, Prior Move, Point of Impact Age of Driver, Sex of Driver, Driver Condition, Driver Action, Use of Seat Belts, Ejection from Vehicle, Month, Day of Week, Hour of Day, Road Alignment, Roadway Configuration, Road Surface Condition, Speed Limit, Weather Condition, Light Condition, Crash Type, Left-turn Collision, Vehicle Type, Vehicle Manoeuvre 서울시고령운전자교통사고특성분석 21

24 고령운전자의경우, 인적요인이환경요인과상호작용하여교통사고가발생하는경우심각한인명피해로이어질가능성이높다 (Ferguson and Braver 2002; Braver and Trempel 2004; Boufous, Finch, Hayen and Williamson 2008; 최재성 2018). 그러나고령운전자의교통사고연구는주로특정지역전역을대상으로진행되었거나전국을대상으로진행되는경우가대부분이며, 대상지역의도시환경특성을고려한고령운전자교통사고연구는아직까지부족한실정이다 (<Table 1> 참조 ). 이는교통사고자료의특성상데이터양이방대하여사고발생지역의환경적특성을반영하는데한계가있기때문으로추정된다. 본논문은행정동을분석단위로설정하여보다미시적인범위에서도시구조가고령운전자의교통사고에미치는영향을규명한다는점에서선행연구와차별성을지닌다. 이후실증분석결과를통해서고령운전자의안전한주행환경조성과교통사고저감을위한정책마련에있어시사점을제공한다는점에서연구의의의를갖는다. 2. 교통사고연구방법론교통사고분석모형중종속변수가교통사고수, 사망자수, 사상자수등의빈도혹은사고율등의비율일때는교통사고분석기법으로선형회귀모형이나로지스틱회귀모형이기본적으로사용된다 (Resende and Benekohal 1997; Yan, Radwan and Abdel-Aty 2005). 이에더해특히빈도가가산자료 (Count Data) 의특성을보이면 (0의값이존재하고자연수만을가짐 ) 포아송 (Poisson) 회귀모형이나음이항 (Negative Binomial) 회귀모형이채택된다 (El-Basyouny and Sayed 2009; Noland and Quddus 2004; Siddiqui, Abdel-Aty and Choi 2012; 박민호, 노관섭, 김종민 2014). ( 여기서음 이항회귀모형은포아송모형의가정, 즉평균과분산이동일하다는가정이위배되는과분산의경우대안으로사용된다.) 한편이러한포아송류회귀모형은자료의공간상독립성을가정한다. 다시말해특정지점에서발생한교통사고는인접한교통사고와무관하다는것이다. 즉교통사고다발지점의존재를부정하는등공간적종속성을설명할수없다는한계점을지녀도로및환경요인특성을제대로반영하기어렵다는한계를내보인다 (Quddus 2008). 이러한배경에서공간계량모형 (Spatial Econometric Models) 은이웃효과를통제하기때문에공간적자기상관과미관측된이분산을적절하게처리할수있다는장점으로교통사고와같이공간단위로수집된자료를분석하는연구에적절한것으로제안되어왔다 (Anselin 1988; Quddus 2008). 전통적공간계량모형으로는공간시차모형 (Spatial Lag Model), 공간오차모형 (Spatial Error Model), 일반공간모형 (General Spatial Model), 지리적가중회귀모형등이있으며, 공간적상관성을고려하는공간적범위에따라서크게전역적모형 ( 공간시차모형, 공간오차모형, 일반공간모형 ) 과국지적모형 ( 지리적가중회귀모형 ) 으로구분된다 ( 이경아 2016). 또한유사한특성을갖는공간이서로정보를교환함으로서추정되는매개변수의불확실성을저감하도록고안된계층적베이지안모형 (Bayesian Hierarchical Model) 도사용된바있다 (Augero-Valverde and Jovanis 2006; Quddus 2008; Huang, Abdel-Aty and Darwiche 2010; Siddiqui, Abdel-Aty and Choi 2012). 3. 시사점도출본논문에서는고령운전자교통사고와교통사고연구방법론에대한선행연구고찰을진행했다. 본논문은고령운전자의교통사고관련문헌분석내용을기 22 국토연구제 102 권 ( )

25 반으로고령운전자가주행중영향을받을것으로추정되는도시환경요소를추출하여영향변수로설정하였다. 예를들어, 고령운전자에게즉각적인대응을요구하는교차로와교통섬과같은도로구조와보행자및차량통행량이많은상업지역, 버스정류장및지하철역변수등을독립변수로설정하여고령운전자의교통사고에미치는영향을규명하고자하였다. 교통사고자료는공간을단위로수집되기때문에공간적자기상관성 (Spatial Autocorrelation) 이존재할가능성이높다. 따라서선형회귀기반의모수추정통계기법을적용하는경우, 공간이질성혹은비분산성문제가발생할수있다. 이에본논문은공간적자기상관성을고려하여연구를수행하고자지리적가중모형과최소자승모형을실증분석의통계기법으로사용하였다. III. 연구방법설계 1. 지리적가중회귀모형 다 ( 조동기 2009). 위와같은배경에서교통사고자료와같이공간단위로수집되는데이터가갖는공간적자기상관성을고려하기위하여 1980년대이후공간계량모형이개발돼왔다. 특히 GWR 모형은 Fotheringham, Brunsdon and Charlton(2002) 이공간적이질성을명시적으로고려하기위해개발한것이다. 이모형은공간상에서로다른지역에위치한데이터에대하여각각의회귀모형을적용하는방식으로공간적이질성을탐색한다 ( 김동하, 강기연, 손소영 2016). 일반적인 OLS 모형은분석지역을대상으로전역적인효과를전제로하지만, GWR 모형은일종의국지적공간모형으로국지적통계량을사용하여공간관계를파악하기때문에세부지역별공간효과를보다잘반영할수있다는장점이있다. 따라서기존회귀모형으로파악하기어려운공간적이질성을표현하고해석하는것이가능하다. 본논문은아래에서전통적인 OLS 모형과공간적위치에따른차이를고려하는 GWR 모형을비교하여고령운전자의교통사고특성을규명하고자한다. 교통사고데이터분석에는주로 OLS 회귀분석이사용돼왔다. 그러나이는모든지역이동일한특성을가진다는전제를통해추정을수행하며이에공간적변이를고려하지못한다는한계를가진다 ( 박일수, 김은주, 홍성옥, 강성홍 2013). 전통적회귀분석은 Gauss-Markov 정리에따라오차의독립성, 등분산성가정을충족시켜야하지만교통사고자료와같이공간적으로인접한위치에서표출된사례일수록유사한값을가지는경향이있고, 따라서현실적으로해당가정을충족시키기어렵다 ( 이희연, 심재헌 2011; 박일수, 김은주, 홍성옥, 강성홍 2013). 이에 OLS 회귀모형분석을통해공간자료를분석하는경우공간적이질성및의존성문제로모수추정치의효율성이저하된 2. 연구범위본논문의대상지는서울시전역이며분석단위는교통사고분석의기초단위인행정동이다. 행정동은또한가구통행실태조사및교통수요분석에서교통분석존 (Traffic Analysis Zone: TAZ) 로사용된다. 고령운전자의교통사고특성을파악하기위하여경찰청과도로교통공단의교통사고분석시스템 (TAAS) 를통해 2015 년서울시교통사고원자료를수득하였다. 여기에서 2015년 1월 1일 ~12월 31일동안가해자가만 65세이상운전자인교통사고를추출하였으며, 가해차량은자전거, 특수차량등과같이일반적으로고령자가통행목적으로사용하지않는유형은제외하고승합 서울시고령운전자교통사고특성분석 23

26 차, 승용차, 이륜기, 원동기로한정하였다. 최종 424개행정동에서발생한고령운전자의교통사고 3,583건을분석에사용하였다. 3. 변수설정연구의종속변수는 2015년서울시행정동별고령운전자의교통사고발생건수이다. 본논문은실증분석을위해관련선행연구와서울시의특성을고려하여도시공간의인구사회경제적특성과도시환경특성을나타내는독립변수로모형을구성하였다. 인구사회경제적특성을반영하는변수는인구밀도, 만 65세이상비율, 20세미만인구비율, 전입인구비율, 기초생활수급자비율, 종사자밀도이며, 도시환경을나타내는변수로서도로연장, 학교, 대학교, 어린이보호구역, 아파트비율, 주거지역비율, 상업지역비율, 녹지지역비율, 공업지역비율, 토지이용복합도, 지하철역, 버스정류장, 교통섬, 교차로를투입하였다. 이중교통섬과교차로는통행량이많고주행중고려해야하는사항이많은복잡한주행환경을나타내는변수이며지하철역과버스정류장은대중교통접근성과대중교통이용객으로인한혼잡성이미치는영향을확인하고자변수로선정하였다. 토지이용복합도는행정동의용도지역이혼합된정도를나타내는 Shannon(1948) 이정의한엔트로피를의미하며계산에는 Cervero and Kockelman(1997) 이사용한식을사용하였다. 값이 1에가까운경우행정동내에다양한용도지역이균등한면적으로혼재함을의미한다. 실증분석을진행하기에앞서독립변수간상관관계를진행하였다. 그결과, 주거지역비율이인구밀도 (0.80) 와강한상관관계를보여해당변수를제외한 후분석을진행하였다. <Table 3> 은상관분석결과를보여주고있으며, 상관분석결과중유의확률이유의수준 0.1 이내인계수만표기하였다. Variable Traffic Accident Involving Elderly Drivers Population Density Population Over 64 Ratio Population Under 20 Ratio Migrant Population Ratio Y 1 X 1 Description Number of Traffic Accidents Involving Elderly Driver in TAZ Number of Residents (unit: thousands) / Area X 2 Number of Population Aged Over 64 Years Old / Population in TAZ X 3 X 4 Number of Population Under 20 Years Old / Population in TAZ Migrant Population / Population in TAZ Basic Livelihood Number of Basic Livelihood Recipients / Recipients Ratio X5 Population of TAZ Employee Density X 6 Number of Employees(unit: thousands) / Area Road Length X 7 Road Extension in TAZ (unit: km) Number of Elementary Schools, Middle School X 8 Schools, and High Schools in TAZ University X 9 Number of Universities in TAZ School Zone X 10 Number of School Zones in TAZ Apartment Ratio Residential Area Ratio Commercial Area Ratio (M) Green Area Ratio Industrial Area Ratio Mixed Land Use Table 2 _ Data Description X 11 Number of Apartment Complexes / Number of Houses in TAZ X 12 X 13 X 14 X 15 X 16 Residential Area / Area Commercial Area / Area ln ln Green Area / Area Industrial Area / Area P: Occupied area ratio of land use type i J: The number of land use types Subway Station X 17 Number of Subway Stations in TAZ Bus Stop X 18 Number of Bus Stops in TAZ Traffic Island X 19 Number of Traffic Islands in TAZ Intersection X 20 Number of Intersections in TAZ 24 국토연구제 102 권 ( )

27 Variable X 1 X 2 X 3 X 4 X 5 X 6 X 7 X 8 X 9 X 10 X 1 1 X *** 1 X **** 1 X *** *** 1 X *** *** ** X * *** 0.19 *** X ** X *** 0.13 *** X *** X *** *** X ** X *** X ** X *** 0.12 ** X * ** X *** X *** ** X *** X *** X *** * Table 3 _ Correlation 0.21 *** *** *** 0.19 *** *** * * 0.08 * 0.16 *** *** 0.09 * 0.19 *** ** *** ** 0.67 *** *** 0.31 *** 0.13 *** 0.40 *** 0.08 * 0.20 *** 0.17 *** * 0.09 * *** 0.21 *** 0.08 * 0.15 *** 0.08 * *** *** 0.14 *** 0.11 ** 0.13 *** *** 0.08 * 0.26 *** 0.27 *** Table 4 _ Descriptive Analysis Variable Min Max Mean S.D Y X X X X X X X X X X X X X X X X X X X X Variable X 11 X 12 X 13 X 14 X 15 X 16 X 17 X 18 X 19 X 20 X 11 1 X ** 1 1 X *** 1 X *** 0.18 *** 1 X ** 0.19 *** X *** 0.20 *** 0.24 *** 0.25 *** X ** *** 0.30 *** *** X * *** X *** 0.18 *** X *** 0.09 * *** 0.18 *** 0.19 *** 0.20 *** 0.16 *** *** 0.27 *** 0.16 *** 0.16 *** 0.25 *** 0.30 *** *** *** 0.56 *** Note: 1) p<0.1: *, p<0.05: **, p<0.01: *** 2) The correlation coefficients of variables at the 0.1 significance level only remained in the <Table 2>. 1 IV. 고령운전자교통사고특성분석 1. OLS 모형평가 OLS 모형의적합성을판단하기위하여 Gauss-Markov 검증을진행하였다 (<Table 5> 참조 ). 검증결과 Koenker(BP) 와 Jarque-Bera 통계값이유의하였다. 이는잔차가비정규분포를가짐을의미하며, 설명변수와종속변수사이에비고정성의관계가있음을나타낸다. 즉일부지역에서는특정설명변수가종속변수에중요한영향을미치지만, 그외특정지역에서는해당 서울시고령운전자교통사고특성분석 25

28 Table 5 _ Estimation of OLS Analysis Categories Value Number of Observation 424 Akaike s Information Criterion (AICc) Multiple R-Squared 0.56 Adjusted R-Squared 0.54 Joint F-Statistic Joint Wald Statistic Koenker (BP) Statistic Jarque-Bera Statistic Figure 1 _ Spatial Autocorrelation 변수의영향력이미미하다는것을의미한다. 이처럼 Koenker(BP) 통계가유의한경우 GWR을대안으로적용하는것이바람직하다. 또한 Moran s I(= 0.10) 에따르면고령운전자교통사고발생은공간적자기상관이있다고말할수있다. 구체적으로 <Figure 1> 에서와같이잔차가클러스터를형성하고있다. 2. GWR 모형분석도시고령운전자의교통사고발생위험요인을규명하고자 GWR 을활용하여분석을진행하였다. 독립변수간다중공선성이발생할경우, 결과값이도출되지않기때문에다중공선성을검토한후분석을진행하 였다. OLS 모형과 GWR 모형분석결과를비교하기위해서 OLS 분석에서유의한영향을미치는것으로나타난최종독립변수 { 유의확률 (p-value)<0.1 인변수 } 를투입하여도출된 OLS 분석결과를함께제시하였다 (<Table 6> 참조 ). 2015년기준서울시의행정동별사고특성분석결과는다음과같다. GWR 모형의설명력 (R²) 은 0.61로 OLS 모형의설명력 (0.55) 보다높게나타났다. 모형의적합도를나타내는 AICc값에대해 Fotheringham, Brunsdon and Charlton(2002) 은기준모형대비대안모형의 AIC 값의감소폭이 4 이하인경우, 모형간의유의미한차이가없다고보았다. 본논문에서는 2,493.64(OLS) 에서 2,488.68(GWR) 로감소하여 GWR 모형이 OLS 모형에비해서모형적합도측면에서도우수하다고볼수있다. GWR 분석결과에따르면 (<Table 6> 참조 ), 행정동별로투입변수가미치는영향력의크기와방향에서차이가나타났다. 공업지역비율과 20세미만인구비율은서울시전체지역에대해부 (-) 의효과를가졌다. 그외 ( 종사자밀도, 도로연장, 어린이보호구역, 상업지역비율, 지하철역, 버스정류장, 교통섬, 교차로 ) 는고령운전자의교통사고발생을증가시키는요인으로작용하였다. <Figure 2~9> 는 GWR 분석결과중행정동별고령운전자의교통사고발생위험요인중도시계획적함의를가진경우에대해회귀계수를지도화한것이다. 청색으로적색으로변할수록계수값이상승하여교통사고발생에해당위험요인이미치는영향력이크다는것을의미한다. 어린이보호구역은서울시행정동에교통사고발생을증가시키는요인으로작용하였다. 특히도봉구, 노원구, 성북구, 금천구일부행정동을중심으로고령운전자의교통사고발생에큰영향을미치는것으로도 26 국토연구제 102 권 ( )

29 Variables GWR Coefficients Mean Max Min S.D. Coef. t Under 20 Years Old Employee Density Road Length School Zone Commercial Area Ratio Industrial Area Ratio Subway Station Bus Stop Traffic Island Intersection Goodness-of-fit Table 6 _ Comparing OLS and GWR Models R ² = 0.61 Adjusted R ² = 0.56 AICc = 2, OLS R ² = 0.55 Adjusted R ² = 0.54 AICc = 2, VIF 출되었다. 전연령의교통사고를대상으로한박나영, 김태양, 박병호 (2017) 의연구에서는교통사고를감소시키는요인으로나타났으나, 고령운전자를대상으로한본논문에서는반대로교통사고를증가시켰다. 어린이보호구역은만 13세미만의어린이의안전한보행활동을위하여설치되는구역이다. 따라서교통안전 의식이낮은어린이들의돌발행동발생이잦은편이기에감속뿐만아니라주변을살피는등주의를기울여야하며주행차량의시야를방해하는불법주 정차차량에대한단속이요구된다 ( 중앙일보 2017a). 고령운전자는비고령자에비해가용시각장이좁으며돌발상황이발생했을때즉각적인대응이어려 Figure 2 _ School Zone Figure 3 _ Commercial Area Ratio 서울시고령운전자교통사고특성분석 27

30 Figure 4 _ Road Length Figure 5 _ Intersection Figure 6 _ Traffic Island Figure 7 _ Bus Stop 워 어린이 보호구역에서 사고 발생 가능성이 높다. 이 것으로 도출되었다. 상업지역은 주거지역과 함께 고 에 어린이 보호구역의 특징과 고령 운전자의 주행 특 령자의 주요 밀집 지역이자, 보행자 및 차량의 통행이 성을 고려한 대책 마련 필요성이 제기되며, 고령자 거 잦은 공간으로(이유진, 최명섭 2018) 주행 중 고려해 주 비율이 높은 지역의 어린이 보호구역을 대상으로 야 하는 사항이 많은 용도지역이다. 반면, 공업 지역 안전한 보행환경 조성을 위한 중점 관리가 요구된다. 비율은 고령 운전자의 교통사고 발생을 감소시키는 상업지역 비율은 서울시 행정동에 교통사고 발생 요인으로 작용하였다. 공업지역은 근로자들의 출퇴근 을 증가시키는 요인으로 작용하였다. 특히 관악구, 서 을 위한 목적 통행이 주를 이루는 공간이다. 상대적으 초구, 강북구, 도봉구, 강서구 일부 행정동을 중심으 로 경제 참여율이 낮은 고령 운전자의 노출 빈도가 로 고령 운전자의 교통사고 발생에 큰 영향을 미치는 낮기 때문에 공업지역에서 고령 운전자의 교통사고 28 국토연구 제102권( )

31 Figure 8 _ Subway Station Figure 9 _ Population Ratio Over 64 Ratio 발생가능성에영향을미친것으로보인다. 고령운전자의교통사고발생영향원인을구체적으로파악하기위해서는공업지역근로자연령구성비와공업지역의통행목적등을함께고려하여해석할필요가있다. 도시의공간적특성중도로환경을나타내는도로연장, 교통섬, 교차로변수는고령운전자의교통사고발생을증가시키는요인으로나타났다 (<Figure 4~6> 참조 ). 도로연장의경우, 서초구, 구로구, 관악구, 동작구를중심으로고령운전자의교통사고발생가능성증가에큰영향을미쳤다. 도로연장변수의영향력분포를보면상업지역비율변수와비슷한패턴을보이고있음을알수있다. 교차로변수는강남구, 서초구, 동작구, 용산구일대와관악구, 송파구, 성동구일부행정동을중심으로영향력이높았다. 실제로서울시의 2013년 ~2016년교통사고다발지점현황 ( 중앙일보 2017b) 에따르면지난 4년간교통사고가가장많이발생한곳은강남교보타워교차로였으며, 2위 ( 영등포교차로 ) 와 5위 ( 잠실교차로 ) 역시교차로였다. 교차로는여러방향에서 차량이진입하기때문에빠르고정확한인지반응및판단을요구한다. 같은맥락에서교통섬은원활한차량통행과보행자의안전을확보하기위하여교차로또는차도의분기점등에설치되는시설로서교통섬진입구간에서도로차선 ( 또는도로폭 ) 이감소하며, 보행자의통행이나타난다. 이와같은변화를인지하여즉각적으로대응하기어렵기때문에분석결과와같이고령운전자의교통사고발생가능성이증가하는것으로보인다. 교통섬은도봉구, 강북구, 성북구, 종로구, 서대문구, 중구, 은평구행정동에서특히고령운전자의교통사고발생에강한영향을미쳤다. 버스정류장은고령운전자의교통사고에양 (+) 의영향을미치는것으로도출되었다. 특히강남구, 송파구를중심으로교통사고발생에상대적으로강한영향을미치는것으로나타났다. 버스정류장에서교통사고가증가하는까닭은보행자및버스의통행으로인해고령운전자가시야확보에어려움을겪기때문으로보인다. 마지막으로지하철역은도로연장, 상업지역비율과유사하게서울시남부행정동을중심으로고령운 서울시고령운전자교통사고특성분석 29

32 전자의교통사고발생에미치는영향력이큰것으로나타났다. 특히서초구, 동작구, 영등포구, 관악구일대에큰영향을미쳤다. 지하철역은버스정류장과같이보행자의통행을유발하는시설이며, 주로대로변에위치하고있어교통사고위험을증가시킬가능성이높다. 선행연구에서도특히보행자-차량간사고위험을증가시키는것으로보고된바있다 ( 박승훈 2014; 이경아 2016). 상업지역비율, 교통섬, 지하철역변수는인구대비고령자비율 (<Figure 9> 참조 ) 이높은지역에서교통사고발생에미치는영향력이크게나타났다. 따라서상업지역비율, 교통섬, 지하철역의특징과해당지점에서실제로발생한고령운전자교통사고특성을고려하여맞춤식교통사고저감대책을마련할필요가있다. 덧붙여해당지점은유동인구가많기때문에보행환경안전제고를위한대책이함께강구되어야할것이다. 유의할점으로서, GWR 분석결과에서나타난것과같이도시구조특성을나타내는변수들이행정동별로영향력의크기와방향을달리가지고있기때문에가 Figure 10 _ Explanatory Power(R 2 ) 급적지역별로교통사고저감방안을마련하는것이바람직할수있다. 마지막으로행정동별 GWR 결정계수 (R 2 ) 는 <Figure 10> 과같다. 결정계수값이클수록본논문에서제시한설명변수들이고령운전자교통사고발생위험을잘설명한다는의미이다. 특히영등포구, 용산구, 마포구, 동작구, 구로구, 은평구, 양천구를중심으로서울시서부에위치한행정구역에서설명력이높게나타났다. V. 결론급속한고령화와맞물려고령운전자에의한교통사고발생빈도와치명률은지속적인증가세를보이고있다. 고령운전자교통사고방지를위해서는고령운전자의특성과주행이이루어지는공간에대한성격을고려하여교통사고발생요인을규명할필요가있다. 이에본논문은고령운전자의특성과주행에영향을미칠수있는도시환경요인과교통사고발생간의관계를규명하고자하였다. 특히교통사고와같은공간적특성에대한연구에서사용하는회귀분석과같은전통적인비공간통계기법은비분산성과공간적이질성문제를내포하고있다. 따라서본논문에서는관측치사이의거리를반영하여공간적으로다른위치에존재하는사례에각각다른계수를추정하여공간적비정상성을탐색할수있는지리가중회귀모형을이용하여도시고령운전자의교통사고특성을파악하고자하였다. 고령운전자및교통사고에관한문헌고찰을기반으로고령운전자의교통사고발생에영향을미칠것으로예측되는서울시 424개행정동의인구사회경제적특성과도시환경특성을나타내는설명변수를선정하였다. 최종모형분석결과, 종사자밀도, 도로연 30 국토연구제 102 권 ( )

33 장, 어린이보호구역, 상업지역비율, 지하철역, 버스정류장, 교통섬, 교차로는사고위험을증가시키는요인으로나타났다. 그중교차로의경우, 고령운전자교통사고발생에양 (+) 의영향을미치며, 이는 Garber and Srinivasan(1991), Preusser, Williams, Ferguson and Ulmer et al.(1998), Zhang, Lindsay, Clarke and Robbins et al.(2000), Langford and Koppel(2006) 의연구결과와일치한다. 특히고령운전자는복잡한환경에서심리적인부담을느끼며 ( 이순철 2006), 이에교차로나교통섬과같이빠른인지판단을요구하는주행환경에서교통사고발생이증가한다고파악하였다. 비고령운전자의경우, 운전자부주의운전에의한교통사고발생률이높으며환경요인보다는운전자의습관이나성격이교통사고발생에영향을미치는것으로나타났다 ( 김종회, 오주석, 이순철 2006; 이상혁외 2012). 고령운전자는비고령운전자에비해서주행중위험요소 ( 과속, 눈, 야간운전등 ) 에대한판별력이높아비고령운전자에비하여안전한주행을추구하는편이나 ( 김종회, 오주석, 이순철 2006), 신체적노화와인지능력하락으로인하여교통사고원인으로주요하게작용함을확인하였다. 따라서고령운전자의교통사고저감대책마련을위해서는고령운전자를대상으로운전자의특성을고려한교통사고연구가필요하며, 본논문은고령운전자를대상으로도시구조가교통사고발생에미치는영향을면밀히파악했다는점에서연구의의의를갖는다. 본논문에서일관되게밝힌것과같이계획가는행정동별로고령운전자의교통사고영향요인의회귀계수값이달리나타난다는점을고려하여지역별맞춤식의사고방지책을마련하는것이사고율감소에도움이될것으로판단된다. 예를들어서초구, 동작구, 관악구에서는상업지역비율이높은구간과지하철역이위치한장소를중심으로안전한보행환경을 조성하는것이요구된다. 본논문에서는고령운전자의교통사고와함께비고령운전자의교통사고특성을비교분석하여살펴보지못했다는점에서연구의한계를지닌다. 향후연구에서는비고령운전자와고령운전자의교통사고특성차이를고려한연구진행과고령운전자의교통사고를사고유형, 연령등을고려한면밀한분석이요망된다. 또한통계모형간분석결과를비교하는경우, 모형에따라유의하게나타나는변수가다르더라도각모형별설명력이가장높은모형을구축한후비교해야하며비교기준을 AICc 값으로설정하는등보다세심한연구진행이이루어져야할것이다. 참고문헌 1. 김동하, 강기연, 손소영 공간가중회귀모형을이용한서울시에너지소비에따른이산화탄소배출분석. 대한산업공학회지 42권, 2호 : Kim Dongha, Kang Kiyeon and Sohn Soyoung Spatial pattern analysis of CO2 emission in Seoul Metropolitan City based on a geographically weighted regression. Journal of the Korean Institute of Industrial Engineers 42, no.2: 김종회, 오주석, 이순철 운전행동결정요인의위반행동및사고에미치는영향. 한국심리학회지 19권, 3호 : Kim Jonghoi, Oh Juseok and Lee Soonchul The influences of driving behavior determinatns on traffic violations and accidents. The Korean Journal of Industrial and Organizational psychology 19, no.3: 도로교통공단교통사고분석시스템. (2019년 8월 28일검색 ). Koroad Traffic Accident Analysis System. or.kr (accessed August 28, 2019). 4. 도로교통공단 년 OECD 회원국교통사고비교 [2018년판]. alreportslist.do?menuid=web_kmp_ida_srs_otc (20 19년 8월 28일검색 ). 서울시고령운전자교통사고특성분석 31

34 Koroad Comparison of traffic accidents between OECD countries in StaticalReportsList.do?menuId=WEB_KMP_IDA_SRS_OTC (accessed August 28, 2019). 5. 박나영, 김태양, 박병호 국내교통사고밀도모형개발. 한국안전학회지 32권, 3호 : Park Nayoung, Kim Taeyang and Park Byungho Development of accident density model in Korea. Journal of the Korean Society of Safety 32, no.3: 박승훈 근린환경이보행자 차량충돌사고에미치는영향 : 북미워싱턴주시애틀도시를대상으로. 국토계획 49 권, 3호 : Park Seunghoon The effect of the neighborhood built environment on pedestrian vehicle collisions: Focused on the cases of the City of Seattle, Washington, U.S. Journal of Korea Planning Association 49, no.3: 박일수, 김은주, 홍성옥, 강성홍 지리적가중회귀모형을이용한지역별심정지발생요인에관한연구. 보건사회연구 33권, 3호 : Park Ilsu, Kim Eunju, Hong Seongok and Kang Sunghong A study on factors related with regional occurrence of cardiac arrest using geographically weighted regression. Health and Social Welfare Review 33, no.3: 오주석, 이의용, 류준범, 이원영 고령운전자의주요교통사고취약상황및인적요인분석. 교통연구 22권, 4호 : Oh Juseok, Lee Euiyong, Ryu Junbeom and Lee Wonyoung An analysis for main vulnerable situations and human errors of elderly Drivers Traffic Accidents. Transportation Research 22, no.4: 이경아 공간계량모형을활용한교통사고유형별발생특성분석 : 서울시를대상으로. 박사학위논문, 서울대학교. Rhee Kyungah Traffic Accident Analysis Using Spatial Econometrics: A case of Seoul. Ph.D. diss., Seoul National University. 10. 이상혁, 정우동, 우용한 고령운전자와비고령운전자의인적교통사고심각도비교분석. 한국ITS학회논문지 11 권, 6호 : Lee Sanghyuk, Jeung Woodong and Woo Yonghan Comparative analysis of elderly s and non-elderly s human traffic accident severity. The Journal of the Korea Institute of Intelligent Transport Systems 11, no.6: 이순철 고령운전자교통사고의심리적요인. 한국심리학회지 : 문화및사회문제 12권, 5호 : Lee Soonchul Psychological effects on elderly driver s traffic accidents. The Korean Journal of Culture and Social 12, no.5: 이유진, 최명섭 노인인구밀집지역의시공간적분포와결정요인분석 : 서울생활인구빅데이터의활용. 서울도시연구 19권, 4호 : Yi, Yoojin, Choi, Myongsub Determinants of the elderly s spatio-temporal concentration: Using bigdata of de facto population of Seoul. Seoul Studies 19, no.4: 이원영, 김기홍, 오주석 고위험군운전자의주요사고원인분석연구 : 고령운전자집단을중심으로. 원주 : 도로교통공단. Lee Wonyoung, Kim Kihong and Oh Juseok A Study on the Major Factor of High-risk Driver Groups Accidents: Focusing on elderly drivers. Wonju: KoROAD. 14. 이희연, 심재헌 GIS 지리정보학 : 이론과실습. 파주 : 법문사. Lee Heeyeon and Sim Jaeheon GIS Geography: Theory and practice. Paju: Bobmunsa. 15. 중앙일보. 2017a. 서울어린이보호구역교통사고 70% 이상, 도로횡단중발생. 9월 4일, (2019년 7월 27일검색 ). Joongang daily. 2017a. Over 70% of traffic accidents in Scho olzone in Seoul happened while crossing road. September 4, (accessed July 27, 2 019) b. 최근 4년간교통사고가장많았던곳은?... 강남교보타워사거리 > 영등포교차로 > 월암교순. 9월 29일, (2019 년 1월 31일검색 ) b. In the last four years, where was the most traffic accidents?... Gangnam Kyobo Tower intersection> Yeongdeoungpo interesection>weolam bridge. September 29, (accessed January 31, 2019) 17. 장재민, 최재성, 김태형 주행환경이고령운전자의교통사고심각성에미치는영향분석. 교통연구 24권, 1호 : Jang Jaemin, Choi Jaesung and Tae-Hyoung Tommy Gim Analyzing driving environment effects on severity of 32 국토연구제 102 권 ( )

35 elderly driver s traffic accidents. Transportation Research 24, no.1: 조동기 지역단위조사연구와공간정보의활용 : 지리정보시스템과지리적가중회귀분석을중심으로. 조사연구 10권, 3호 : Jo Donggi GIS and geographically weighted regression in the survey research of small areas. Survey Research 10, no.3: 최재성 초기 중기 후기고령운전자의사망자발생위험도분석과시사점. 한국안전학회지 33권, 1호 : Choi Jaesung Study on fatality risk of senior driver with aging classification. Journal of the Korean Society of Safety 33, no.1: Anselin, L Spatial Econometrics: Methods and models. Dordrecht: Springer Science & Business Media. 21. Augero-Valverde, J. and Jovanis, P. P Spatial analysis of fatal and injury crashes in Pennsylvania. Accident Analysis and Prevention 38, no.3: Ball, K., Beard, B., Roenker, D. L., Miller, R. L. and Griggs, D. S Age and visual search: Expanding the useful field of view. Journal of the Optical Society of America A 5, no.12: Boufous, S., Finch, C., Hayen, A. and Williamson, A The impact of environmental vehicle and driver characteristics on injury severity in older drivers hospitalized as a result of a traffic crash. Journal of Safety Research 39, no.1: Braver, E. R. and Trempel, R. E Are older drivers actually at higher risk of involvement in collisions resulting in deaths or non-fatal injuries among their passengers and other road users. Injury Prevention 10, no.1: Cervero, R. and Kockelman, K Travel demand and the 3Ds: Density, diversity, and design. Transportation Research Part D: Transport and Environment 2, no.3: El-Basyouny, K. and Sayed, T. A Urban arterial accident prediction models with spatial effects. Transportation Research Record: Journal of the Transportation Research Board 2102, no.1: Ferguson, L. C. and Braver, W Older driver involvements in police reported crashes and fatal crashes: Trends and projections. Injury Prevention 8, no.2: Fotheringham, A. S., Brunsdon, C. and Charlton, M Geographically Weighted Regression: The analysis of spatially varying relationships. New York: Wiley. 29. Garber, N. J. and Srinivasan, R Characteristics of accidents involving elderly drivers at intersections. Transportation Research Record 1325: Hakamies-Blomqvist, L. and Henriksson, P Cohort effect in older drivers accident type distribution: Are older drivers as old they used to be? Transportation Research Part F 2: Huang, H., Abdel-Aty, M. A. and Darwiche, A. L County-level crash risk analysis in Florida: Bayesian spatial modeling. Transportation Research Record 2148, no.1: Huisingh, C., Levitan, E.B., Irvin, M. R., MacLennan, P., Wadley, V. and Owsley, C Visual sensory and visual-cognitive function and rate of crash and near-crash involvement among older drivers using naturalistic driving data. Investigative Ophthalmology & Visual Science 58, no.7: Langford, J. and Koppel, S Epidemiology of older driver crashes Identifying older driver risk factors and exposure patterns. Transportation Research Part F 9, no.5: Noland R. B. and Quddus, M. A A spatially disaggregated analysis of road casualties in England. Accident Analysis and Prevention 36, no.6: Preusser, D. F., Williams, A. F., Ferguson, S. A., Ulmer, R. G. and Weinstein, H. B Fatal crash risk for older drivers at intersections. Accident Analysis and Prevention 30, no.2: Quddus, M. A Modelling area-wide count outcomes with spatial correlation and heterogeneity: An analysis of London crash data. Accident Analysis and Prevention 40, no.4: Resende, P. T. and Benekohal, R. F Effects of roadway section length on accident modeling. In Proceeding of Traffic Congestion and Traffic Safety in the 21st Century: Challenges, Innovations, and Opportunities. June 8, Illinois: ASCE. 38. Shannon, C. E A mathematical theory of communication. Bell System Technical Journal 27, no.3: Siddiqui, C., Abdel-Aty, M. A. and Choi, K. C 서울시고령운전자교통사고특성분석 33

36 Macroscopic spatial analysis of pedestrian and bicycle crashes. Accident Analysis and Prevention 45: Yan, X., Radwan, E. and Abdel-Aty, M Characteristics of rear-end accidents at signalized intersection using multiple logistic regression model. Accident Analysis and Prevention 37, no.6: Zhang, J., Lindsay, J., Clarke, K., Robbins, G. and Mao, Y Factors affecting the severity of motor vehicle traffic crashes involving elderly drivers in Ontario. Accident Analysis and Prevention 32, no.1: 논문접수일 : 심사시작일 : 심사완료일 : 요약 주제어 : 교통사고, 지리가중회귀모형, 고령운전자, 공간계량모형 고령운전자는신체적노화와심리적노화로인해비고령운전자에비해서주행중주변환경에보다민감하게영향을받는다. 본논문은서울시의 424개행정동 / 교통분석존을대상으로도시환경요인이고령운전자의교통사고발생에미치는영향을분석했다. 분석에는교통사고자료와같은공간자료의공간적특성을고려한지리가중회귀모형과전통적분석모형인최소자승회귀모형을비교해사용했다. 분석결과교통사고자료는공간적자기상관성이확인됐으며더불어지리가중회귀모형이최소자승회귀모형에비해서모형적합도와설명력이우수하게나타났다. 이에변수영향력을분석하는데지리가중회귀모형을사용했다. 분석결과, 고령운전자의교통사고발생에영향을미치는요인들은행정동별로차이를보였다. 그중종사자밀도, 어린이보호구역, 상업지역 비율, 지하철역, 버스정류장, 교통섬, 교차로는고령교통사고발생을증가시키는요인으로나타났다. 특히상업지역비율, 지하철역, 교통섬은고령자의거주비율이상대적으로높은지역에서사고발생에강한영향을미친다는것을확인했다. 고령운전자의사고위험은주로차량및보행자통행을유발하는상황이밀집된지역, 그리고인지판단과정에서충분한시간을요구하는교차로와교통섬과같은복잡한도로환경에서높게나타났다. 본논문은고령운전자의교통사고에도시환경요소를포함한다양한도시공간구성요소들이교통사고발생에영향을미치는것을확인했다. 고령운전자의교통사고를선제적으로방지하기위해서는지역별로고령운전자의교통사고영향요인을고려한물리적인환경개선이이루어져야할것이다. 34 국토연구제 102 권 ( )

37 국토연구제 102 권 (2019.9): pp35~48 해안곡 ( 谷 ) 계마을을대상으로한위요공간규모에관한연구 * A Study on the Size of Encirclement Space According to Coastal of Geographical Names and Terrain in Korea 서성혁 Seo Seonghyeok **, 정태열 Jung Taeyeol *** Abstract Recently, the harbor has lost its original landscape of the coast and its port-space identity due to the scale expansion through landfill and uniform landscape planning. It also made it difficult for people to access. And the purpose of this study is to provide a standard of spatial size for the development of hydrophilic space of harbor and comfort and familiarity to users. According to the method proposed in the study, the size of the surrounding space was 40,000~150,000m 2. In order to compare the spatial scale of the various natural villages, the result of the square conversion was 200~400m. This result is a length giving intimacy according to the theory of distance segmentation. It is difficult to simply define the attributes of the space that can be familiar with it. However, in a sense of space where humans perceive space of a certain size as one space, it is judged that distance is directly related to the recognition of scale. Therefore, if the results of the study are utilized, it will be helpful to create a traditional port space considering human scale and create an attractive landscape. Keywords: Gok( 谷 ) in Geographical Names, Surrounded Space, Space of Scale, Coast I. 서론 1. 연구의배경및목적우리나라는 3면이바다로둘러싸인지리적조건으로인해국토개발초기부터임해지향적개발이추진되면서무분별한연안개발이시행되었다. 또한세계화를향한연안국토축의개발및휴양, 위락시설에대 한요구가증가하면서개발압력이가속되고있는실정이다. 특히해양매립을포함한항만개발사업의비중이증가하여그규모또한거대해지고있는추세이다. 이러한양적인개발사업으로해안다운원풍경이사라지고있으며, 게다가획일화된경관계획으로항만공간의정체성을잃어가고있다. 이렇듯무분별한매립을통한규모의확장과획일화된경관계획으로는항만이사람들의선호공간으로발전할수있을것이 * 본논문은서성혁 (2019) 의석사학위논문 해안곡 ( 谷 ) 계지명의지형적위요및규모의관계에관한연구 의일부를수정 보완한것임. ** 경북대학교조경학과박사과정 ( 제1저자 ) Ph.D. Candidate, Dept of Landscape Architecture, Kyungpook Univ. Primary Author tjdgur0360@naver.com *** 경북대학교조경학과교수 ( 교신저자 ) Prof., Dept. of Landscape Architecture, Kyungpook Univ. Corresponding Author jungty@knu.ac.kr 해안곡 ( 谷 ) 계마을을대상으로한위요공간규모에관한연구 35

38 라보기어렵다. 따라서사람들이선호하는공간의적정규모즉, 공간감을부여하면서편안함과친숙함을느낄수있는적당한공간규모의기준치가있다면향후항만개발에유효할것으로판단된다. 한편, 사람들이선호하는공간이론으로 Appleton (1975) 의 Prospect and Refuge 이론은생물학이나행동과학과연관된모델실험을통해인간이생존을위해서적또는외부인에게보여지지않되, 인간은적또는외부인을볼수있는공간을선호한다고하였다. 또한사신사 ( 四神砂 ) 풍수이론에서도집터는산이전후 ( 前後 ) 좌우 ( 左右 ) 사방 ( 四方 ) 을감싸고있는곳이포근하고안락한느낌을준다고하여사람들이거주공간으로선호한다고하였다. 또한 Ushio Saitho, Masatoshi Takeshita and Kenji Ueshima(1992) 는워터프론트개발로불리는홍콩지구재개발성공사례나고대부터물가에사람들이모이던장소의공간규모의사례분석결과, 선박계류를위해수면을포함한위요된공간의규모가대략한변 500m 정방형을넘지않는다는것을확인했으며, 그의연구논문에서인간이외부공간에서방과같이아늑하고편안하게느끼는공간규모가 200~400m의정방형임을검증하였다. 이러한규모에서인간이편안함과친숙함을느낄수있다고단순히규정할수없는것은사실이지만, 이러한규모를하나의공간으로인식하는인간의공간감각과의미가연결된다면이수치는향후항만조성시친숙한환경을조성하는원단위적기준이될것으로판단된다. 이러한점에착안해본논문은우리나라항만및친수공간의경관계획시사람들에게친숙한공간조성을위한기초자료로활용하고자우리나라해안에서방과같은아늑함을느낄수있는곡 ( 谷 ) 계지명의마을을도출했다. 또한선행연구에서활용한위요공간의외곽선선정방법들을참고하여외곽에따라나타 나는공간에서실질적으로위요감을얻을수있는규모의기준을도출하고, 어떠한형태로위요된인상을주는지에대해살펴보고자한다. 2. 선행연구 1) 해외의위요공간규모연구사례일본의 Ushio Saitho, Masatoshi Takeshita and Kenji Ueshima(1992) 가연구한 지형적위요의인식과공간규모와의관계연구 는항만에서휴먼스케일을기준으로그규모를도출하기위한연구이다. 일본어로 무로 ( ムロ ) 라칭하고한자로 室 로쓰여지며, 고 ( 故 ) 지명이방과같은인식을주는지명을대상으로항만공간의규모에대해분석하였으며, 공간규모를산출하기위해위요공간대응범위의경계를선정하는방법을마을의중심선정, 외곽선경계선정, 개구부경계선정순으로진행하였다. 2) 국내의위요공간규모연구사례국내이진동, 김정규, 정태열, 김유일 (2010) 이연구한 곡 ( 谷 ) 계지명과지형적위요및공간규모의관계 에대한연구는위요로인식된공간규모를찾아내기위해둘러싸인감이강하게나타나는지명중곡계지명의마을을선정하였다. 우리나라내륙에위치한진안군의마을 270곳을대상으로곡계지명 33개를선정하고, 공간대응범위의경계를선정하는방법에따라공간의규모를도출하였다. 이에따른결과, 곡계지명의위요공간규모는 10만 ~25만m 2 으로나타났으며, 정방형환산하였을시거리가 300~500m의규모로나타났다. 36 국토연구제 102 권 ( )

39 3) 선행연구와의차별성우리나라는산지가대부분인지형으로내륙의곡 ( 谷 ) 계지명마을의수가상당히많다. 그렇기때문에내륙을대상으로한국내연구와같이일정범위에서연구할수밖에없는한계가있다. 또한내륙곡계마을은대부분지형적으로 4방면으로둘러싸여위요를느낄수있는반면, 해안곡계마을의경우대부분해안방면으로 1면만개방되어있다. 그럼에도불구하고위요를느끼는동시에개방감을느낄수있는우리나라의해안곡계마을을도출하는것에의의를두고, 또한그규모의기준은어느정도인가에대해분석하고자한다. 3. 연구의개요 1) 연구방법옛마을과인간의관계를살펴보면산악및구릉의외곽에의해하나의공간으로인식된위요공간의규모를찾아내기위해마을지명의지형적어휘를도출하고지형지명을공간인식에대한현상적이해를나타내는집단적표상으로파악하여분석하였다. 또한지명에대한문헌등으로부터지형지명중위요감이강하게나타나고있는곡 ( 谷 ) 계지명을가진해안마을을대상으로선정하였다. 마을지명과관련된산과하천, 해변등조상들이어떻게자연에입지하였으며, 어떠한지명으로마을 을개척하고, 주위환경과공간관계는어떻게이루어졌는가를파악하기위해우리나라지명에관한문헌과전통마을에대한선행연구를분석하였다. 1) 또한해안곡계지명마을의위요공간대응범위경계를선정하기위해지명과관련하여공간의규모를경사도에따라분석한국내연구이진동, 김정규, 정태열, 김유일 (2010) 과등고선에따라분석한해외연구 Ushio Saitho, Masatoshi Takeshita and Kenji Ueshima(1992) 를참고하여본논문에서활용하였다. 2) 연구대상지선정대상지를선정하기위해국토지리정보원에서제공하는우리나라행정지명이표기된 1/5,000 지도를통해해안에인접한곡 ( 谷 ) 계지명을확인하고, 지명사전을통해지명유래에관한정보를입수하여대상지의정보를정리하였으며, 검색되지않는지명은한국지명유래집및인터넷검색과대상지별각해당면 읍사무소등에유선으로문의하여정보를입수했다. 또한웹정보사이트구글 (Google) 에서제공하는지도와구 ( 舊 ) 지도를통해해안에인접한지명들을조사하고재확인하는작업을하였다. 이로써우리나라해안에인접한곡계지명의공간에대한전수조사를수행했다할수있으며, 본논문에서칭하는해안곡계지명의마을은 1면이해안방면으로개방된지형의공간을대상으로한다. 조사된 59개의대상지중선정조건 2) 에부합하고기존자연지형을유지하고있다고판단되는 52개의마을을선정하였다 (<Table 1> 참조 ). 1) 국토지리정보원 (2008); 노선화, 문병선, 김태영 (2007); 박재락 (2013); 신상섭 (1999); 이돈주 (2007); 이진동, 김정규, 정태열, 김유일 (2010); 이학동 (2003); 한필원, 민관식 (2016) 등. 2) 첫째, 마을지명중후부요소가곡 ( 谷 ), 골, 실 ( 室 ) 로명명되는곡 ( 谷 ) 계지명의마을. 둘째, 역사적으로지명이변천된경우는옛지명의어원이곡 ( 谷 ) 계지명의마을. 셋째, 지명의입지가도시화되어개발되었거나시가화로인해등고선복원이어려운수몰된지역은마을이존재하지않아도구 ( 舊 ) 지도에의해확인된마을. 넷째, 곡계지명중마을을중심으로 3면이산으로둘러싸여있으며, 1면은해안방면으로개방된마을. 다섯째, 곡계지명의지형적특징을띄지않거나지형지명관계가불명확한것은제외. 해안곡 ( 谷 ) 계마을을대상으로한위요공간규모에관한연구 37

40 Province Gangwon -do Gyeongang -do Table 1 _ Research Subject Geographical Name No. II. 곡 ( 谷 ) 계지명의특징 1. 곡 ( 谷 ) 계지명의유래 Province Geographical Name No. Sim-gok 1 Han-sil 27 Cho-gok 2 Ye-gok 28 Gi-gok 3 Ga-gok 29 Wit-gol 4 Hoe-gok 30 Bu-gok 5 Geomu-sil 31 No-sil 6 Gyeongsangdo Byeong-gok 32 Song-gok 7 Mangyang-gol 33 Sa-gok 8 Hang-gok 34 an-gol 9 Gol-jang 35 Dap-gok 10 Na-sil 36 Cheong-gok 11 Seok-gol 37 Du-mo-sil 12 Won-gok 38 Gan-gok 13 Su-gok 39 Dae-gok 14 Song-gok 40 Gok-chon 15 Seongju-gok 41 Geum-gok 16 Sowu-sil 42 Galmanggae-gol 17 Jeonla Wu-sil 43 Maligae-gol 18 -do Dae-gok 44 Keun-gol 19 Wol-gok 45 Dal-sil 20 Geum-gok 46 Wol-gok 21 Samma-gol 47 Uk-gok 22 Song-gok 48 Jeong-gok 23 Yong-gok 49 Ok-gok 24 Chung Jang-gok 50 Oksam-gol 25 cheong Seol-gok 51 Cheondae-gol 26 -do Gani-sil 52 우리나라지명은대체로그지역의주위자연환경즉산천, 초목, 암석고개등의이름을빌려작명되거나생활을향유하기위해제작된샘, 못, 다리, 성당등과 관련된것또는신앙의흔적인신당, 장승입석, 서원등에따라명명되는경우가많다. 이외에도지역의위치, 풍수지리, 형태등과관련된지명도많아그유래가다양하다. 또한한반도총면적의 70% 이상이산지로된지리적특성에맞게우리말지명에는산 (-뫼, -메, -미) 과골을소재로한지명이대부분이다. 곡 ( 谷 ) 은골짜기라고불리며길고좁은지형으로길이가너비보다는현저히길고그사이로하천이흐르는경우가많다. 우리의자연을산줄기와물줄기의관계로해석하는 대동여지도 나 산경표 는이두가지사이의요철 ( 凹凸 ) 관계를골짜기즉, 곡의형상으로보고있으며, 하천이나도랑이없는골짜기도풍수에서는물줄기로해석하고있다. 3) 예로부터중요한이동통로로이용되어온계곡은협곡 ( 峽谷 ) 이나빙식곡 ( 氷蝕谷 ) 등아름다운자연경관으로관광지를이루는곳이많다. 또한곡계마을의지명은과거선조들이마을을방 ( 房 ) 안에아름다운비단을연상케한다하여 나실 로부른것처럼곡계지명은 둘러싸임 또는 방과같은아늑함 을느끼며, 지형적으로위요된 Table 2 _ Origin of the Geographical Name Geographical Name Na-sil Gi-gok Origin of the Geographical Name Nagok-ri is also called Na-sil, and the village was called Seon-ra because she thought she was dancing in a silk suit. After exploring the plains of the south, it reminds her of a beautiful silk stream in her room. Around 390 (Korea s King Gongyang), Choe Bok-ha pioneered a cozy village in a large field and named it Muryeong, but later called it Gi-gok. Behind Mt. Pirae, a mountain stream extending west and east stretches north and south, gently surrounding Sim-gok the village. There is a village in a narrow, long valley that is sunk toward the beach, and there is a muzzle. Sim-gok is called a village in a deep valley. Source: (accessed July 1, 2019). 3) 지명어의한역화유형에관한설명은박병철 (2004) 에기초하고있음. 38 국토연구제 102 권 ( )

41 Figure 1 _ Old Map and Street View of Geographical Names Na-sil Source: (assessed July 1, 2019). 곳에마을이입지하였음을지명유래와구지도를통해확인하였다 (<Table 2> 및 <Figure 1> 참조 ). 2. 곡 ( 谷 ) 계지명의유형 Sim-gok 지명의유형은크게자연지명, 인공지명, 인연지명, 신앙지명으로분류할수있으며, 자연지명중에서 谷 의유형은곡, 동, 골, 실로대응된다 (<Table 3> 참조 ). Table 3 _ Type Classification of Geographical Name 3. 곡 ( 谷 ) 계지명의지형적특징우리나라의지명중산에관한지명이많이나타나는이유는국토대부분이산지라는자연적특성으로인한마을입지의지리적조건과산악숭배의원시신앙이라는두가지측면에서찾아볼수있다. 산지는거주지로서불리한점이많아, 보통거주지는평야지대에발달하는것이유리하다. 하지만우리나라는산지가많아마을이보통산지와평야의접촉지점인산록부와곡저부에많이입지하고있다. 산록부와곡저부는용수가풍부하고, 강풍을피할수있는지역이기때문이다. 4) 산줄기와물줄기의관계를바탕으로해석하는것이우리나라자연관의기본인데, 풍수지리에서말하는배산임수가명당으로여겨지는것으로볼때곡 ( 谷 ) 에입지한자연마을은우리고유의자연관과입지관을한눈에볼수있으며마을의일반적인모습을추출할수있는사례가될수있다고판단된다. 곡계지명마을은주위환경, 지형지세및지리적여건을고려하여입지하고있으며, 그마을의영역은위요감을느낄수있는특징을가지고있다고할수있다. Group of Geographical Name Type Kind of Geographical Name Classification Standard III. 곡 ( 谷 ) 계지명의위요공간분석 Natural Geographical Name 谷 -Gok, -Dong, -Gol, -Sil The name is derived from the meaning of Gok ( 谷 ), which is commonly found in the villages in the valley, and the land name of Go-eul which means Dong ( 洞 ) is made up of Gol. Sill( 室 ) is an old word meaning Gok ( 谷 ), which means a small village in a valley. Source: National Geographic Information Institute 위요공간대응범위선정 1) 마을중심선정마을중심은지형적시각환경이전개되어마을전체를내려다볼수있는곳에위치한종가로하였다. 이는시대적상황에입각하여입향조 ( 入鄕祖 ) 의사상과 4) 마을의입지특성에대한설명은홍경희 (1990) 에기초하여작성했으며, 부분적으로수정하였음. 해안곡 ( 谷 ) 계마을을대상으로한위요공간규모에관한연구 39

42 철학그리고입향목적을바탕으로마을의위치가결정되고나면종가가제일먼저자리를잡게된다. 또한종가의입지를고려하여주요건축물들이자리를잡으면서서서히마을이형성되어간다. 종가및주요건축물은현대와같은토목기술이발전하지못한상황에서최대한자연지형을활용하면서형성되었을것이다. 5) 마을중심인종가의선정은다음과같은방법으로하였다. 첫째, 각마을의면 읍사무소, 마을회관등에유선으로문의하여통해종가의위치를파악하였다. 둘째, 종가의위치파악이어려운경우대상지에해당되는구지도를활용하여위치를파악하였다. 셋째, 마을이존재하지않거나종가의위치파악이어려울경우우리나라전통방식 6) 을근거로하여마을을내려다볼수있으며, 마을공간의뒤에자리하여뒷산과근접한건물을종가로선정하였다. 2) 외곽선선정마을의외곽선은선행연구의방법을참고하여두가지방법으로선정하였다. 첫째, 등고선에따른방법 7) 으로 AutoCad 를활용하여마을의중심으로부터가장근접하고급변하는등고선을외곽선으로선정하였다. 둘째, 경사도에따른방법 8) 은일반적인전통마을이경사 9도이내에입지하고있는것과평지처럼건축 계획이가능한경사를 0~9도로두고있음을고려하여 ArcGIS 10.5를활용하여경사 9도이상으로변곡지점을연결하여외곽선을선정하였다. 3) 개구부경계선정지형적으로시각환경이전개되어마을전체를내려다볼수있는위치에있는마을중심으로부터해안방면으로바닥면이가장넓게보이는시각방향과외곽선이만나는양측접점을직선으로연결하는방법으로개구부의경계를결정하였다. 4) 위요공간대응범위및계측위요공간의대응범위선정과계측은국토지리정보원에서제공하는 1/5,000 지형도에위와같이마을의중심, 외곽선, 개구부경계선정과정을통해 AutoCAD 와 ArcGIS 10.5를활용하여 52개대상지의지명대응범위면적을계측하였다 (<Table 4> 참조 ). 2. 위요공간규모계측결과본논문에서는해안곡 ( 谷 ) 계지명마을의공간규모를측정하기위해등고선과경사도에의한두가지방법으로위요공간의범위를선정하고면적을계측하 5) 과거조상들의마을입지특성에대해신진동 (2008) 에서찾아볼수있음. 6) 조선시대성리학적종법성 ( 宗法性 ) 사회의마을공간에서앞뒤상대적위치에따라위계를달리부여하는전후 ( 前後 ) 개념으로, 위계가높을수록마을공간뒤에자리함. 7) 첫째, 지명을차지하는취락의중심을확정하기위해고지도를통해가능한오래된시점에서마을의중심부역할을하는사원, 우체국등을단서로하여마을의중심을선정함. 둘째, 마을의중심으로부터가장근접한등고선중급경사로판단되고, 평평한곳에대비하여물리적으로솟아오른지점이되는등고선을외곽선으로선정함. 셋째, 개구부경계선정은지형적으로열린부분의경계를선정하는과정으로마을중심으로부터개구부를향해외곽선과의양측접점을직선으로이어선정함. 8) 첫째, 지명을선정하기위해 -곡( 谷 ), -실 로명명되는마을을진안군행정지도를통해검색하고 1/5,000 수치지도를이용하여분석함. 둘째, 경사 9도이상으로변곡되는지점을연결한외곽선을결정함. 셋째, 마을중심에서개구부로가장넓게보이는시각방향과외곽선이만나는양측접점을직선으로연결하여개구부의경계를선정함. 40 국토연구제 102 권 ( )

43 Table 4 _ Boundary and Scope of Surrounding Space Geographical Name Sim-gok Division Boundary of Surrounded Space by Contour Lines Boundary of Surrounded Space by Slope Boundary of Geographical Name Scope of Geographical Name 였다. 그리고등고선과경사도에따른위요공간의면적계측결과의차이를알아보기위해 t-test를활용하여규모를비교하였다. 등고선과경사도에따른면적의차이를보면경사도에따른위요공간의규모가비교적크게계측되었으며, 면적값들의편차는등고선에따른규모가적게계측되었음을확인하였다. 그러나유의성검증결과 t=0.671, p=0.505로결론적으로두가지방법에따른 면적의차이는없음을알수있다 (<Table 5> 참조 ). 본논문에서는마을의형태와규모가매우다양함을고려하여비교적편차가적게계측되어일정한규모범위가나타나는등고선에따른위요공간의규모를제시하고자한다. 등고선에따른위요공간규모는최소 21,050m 2 에서최대 271,490m 2 의규모를보였으며, 평균 93,608m 2 로계측되었다 (<Table 6> 참조 ). Division Contour and Slope Table 5 _ Comparison of Areas According to the Method through Contour Line and Slope Paired Samples Statistics Division Average N Standard Deviation Std. Error Mean Contour Slope Paired Samples Test Mean Std. Deviation Paired Differences Std. Error Mean 95% Confidence Interval of the Difference Lower Upper t df Sig. (2-tailed) 해안곡 ( 谷 ) 계마을을대상으로한위요공간규모에관한연구 41

44 Table 6 _ Scale of Surrounding space of Geographical Names by Contour Line Method No. Area(m 2 ) A(m) No. Area(m 2 ) A(m) No. Area(m 2 ) A(m) No. Area(m 2 ) A(m) 1 56, , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , 또한등고선에따른공간규모는대상지 52개소중 46개소 (88%) 가 40,000~150,000m 2 에분포하고있으며, 이러한규모를위요공간의면적범위라고할수있다 (<Figure 2> 참조 ). 곡계지명은지형의특성상매우다양한형태를갖추고있기때문에공간의상호규모를 비교하기위해정방형으로환산하여그규모를비교하였다. 그결과한변의길이는최소 145m에서최대 521m로나타났고, 평균 295m의규모를보였으며, 위요공간의길이범위는 300m를중심으로 200~400m에포진하고있음을알수있다 (<Figure 3> 참조 ). Figure 2 _ Area Distribution of Surrounding Space by Contour Line Method 42 국토연구제 102 권 ( )

45 Figure 3 _ Square Length of Surrounding Space by Contour Line Method 3. 위요공간의길이와폭의관계 해안곡 ( 谷 ) 계지명의형태는매우다양한형태를가지고있으며, 대체로길이가긴형태를하고있음을확인하였다. 이러한지형특성을자세히파악하기위해 <Table 4> 와같이등고선에의해작도된지명대응범위형태의주연부를둘러싸고있는가장긴직사각형의길이와폭의규모를 Forman(1995) 의패치측정방법을참고하여다음과같이측정하였다 (<Figure 4> 참조 ). 그결과계측된길이는평균 497m, 길이가가장긴마을 31( 그무실 ) 은 1,275m, 길이가가장짧은마을 8 ( 사곡 ) 은 171m로나타났다. 계측된폭은평균 378m, 폭이가장긴마을 34( 항곡 ) 는 913m, 폭이가장짧은 Figure 4 _ Method of Measuring the Length and Width of the Surrounding Space 마을 33( 망양골 ) 은 69m로나타났다. 그리고폭에대한길이의비율크기순으로지명의번호를나열하여표로작성한결과마을 34( 항곡 ) 가가장낮은비율로폭이넓은형태임을알수있었으며, 가장높은비율의마을 33( 망양골 ) 은길이가긴형태임을알수있었다 (<Table 7> 참조 ). 또한계측된길이와폭의비율로나타낸결과평균비율은 1.52로나타났으며, 길이와폭의비율이 1.0 이상인마을은총 52개의지명중 36개로해안곡계지명의 70% 는대체로길이가긴형태임을알수있다 (<Figure 5> 참조 ). 이러한결과와같이곡계지명의길이가긴지형에서도아늑함을느낄수있었던것은인간의시각특성에기인한것으로판단된다. Yoshio Nakamura (2004) 는평행한면과수직하는면이같은길이라하더라도시각축의방향으로시선이펼쳐질수록지표면의투시적압축이많이발생하며, 이를형 ( 形 ) 의투시적압축이라하였다 (<Figure 6> 참조 ). 인간의시야에서펼쳐지는지상의풍경은다양한원근감을가지며이는양안시차 ( 兩眼視差 ) 에의해원거리에서시각적인깊이감각이발생한다. 이러한인간의시각적특성을적용한예는르노트르가설계한프랑스의보르비콩트 해안곡 ( 谷 ) 계마을을대상으로한위요공간규모에관한연구 43

46 Table 7 _ The Length and Width of the Surrounding Space Figure 5 _ The Length and Width of the Surrounding Space No. L(m) W(m) L/W No. L(m) W(m) L/W (Vaux le Vicomte) 정원에서도찾아볼수있다. <Figure 7> 과같이평면에서긴형태 ( 타원형 ) 를 인간의시점에서바라보게되면그림의우측과같이 압축된형태 ( 원형 ) 로인식된다. 이처럼인간은길이 ( 시각축 ) 에대한인식을더짧게느끼는특성이있다. 이러한특성은과거의우리선조들에게도적용되어 Figure 6 _ Perspective Compression of Form due to Visual Angle of Incidence Figure 7 _ Perspective Compression of Form Phenomenon(Vaux le Vicomte) 44 국토연구제 102 권 ( )

47 길이가긴해안곡계지형에서도종갓집에서바라보면시각축의길이가압축되어실제보다가깝게인식하여위요감이형성되었다고판단된다. 4. 해안권역별위요공간의규모 해안곡 ( 谷 ) 계지명의위요공간규모를도별과해안권역별로분류하여정리하였다. 먼저도별로분류하여공간규모의평균을도출한결과강원도곡계지명 의 6개평균규모는 88,781m 2, 경상도곡계지명 31개의평균규모는 85,017m 2, 전라도곡계지명 12개의평균규모는 96,761m 2, 충청도곡계지명 3개의평균규모는 151,285m 2 로나타났다. 도별로보면결과값으로는충청도가가장높으나, 대상지가 3개뿐이고, 52 개의마을중가장큰규모가포함되어있어객관성을담보하기는힘들다고생각된다. 따라서도별간규모차이는없다고할수있다. 다음은해안권역별로결과를살펴보면동해안의곡계지명은 15개로평균규 Table 8 _ Surrounding Space Scale in Coastal Areas Coast(unit) Province(unit) No. Average Area(m 2 ) The East Coast (15) The South Coast (30) The West Coast (7) Gangwon-do (6) Gyeongsang-do (31) Jeonla-do (12) Chungcheong-do (3) , ,39, 42-44, , 40, 41, 45 50, 51, 52 85, , ,285 98,780 81, ,247 모는 98,780m 2 로나타났으며, 남해안은 30개로평균규모가 81,073m 2, 서해안은 7개로평균규모가가장큰 151,285m 2 로나타났다. 이는우리나라가백두대간을중심으로동고서저의지형적특징을가지고있어서해안은보다많은평지가존재하기때문으로보인다 (<Table 8> 참조 ). IV. 요약및결론본논문은우리나라해안곡 ( 谷 ) 계마을에서아늑함을느낄수있는규모의기준을도출하기위해선행연구를참고하여등고선과경사도에따른규모를계측하였다. 그리고비교적일정한규모범위가나타나는등고선에따른방법으로위요공간의규모를제시하였다. 위요공간의규모를계측한결과, 대상지 88% 가면적값 40,000~150,000m 2 로면적의범위가다소크게나타났다. 이는다양한형태를가진마을의특징으로인한결과로, 이를고려해공간상호의규모를비교하기위해정방형으로환산하였다. 그결과규모의범위는 200~400m로나타났다. 다음으로해안곡계지명을가진마을의위요공간의특징을파악하기위해길이와폭을측정했다. 그결과폭에대한길이의비율의평균 1.52로, 전체마을중 70% 가길이가긴지형으로나타났다. 이처럼길이가긴지형에서도우리조상들이아늑함을느낄수있는이유를시선입사각으로인한형 ( 形 ) 의투시적압축현상으로보며, 인간이길이 ( 시각축 ) 에대한인식을더짧게느끼는특성에따라우리나라해안곡계지명의마을에서도길이 ( 시각축 ) 를가깝게인식하였다고판단하였다. 또한해안권역별위요공간규모분석을통해공간을인식할때지형적인영향을받는다는것을알수있었다. 우리나라전형적인산촌지형모델이라할수있는 해안곡 ( 谷 ) 계마을을대상으로한위요공간규모에관한연구 45

48 진안군 ( 내륙 ) 의곡계마을규모는 7만 ~25만m 2 로, 본논문의해안곡계마을규모에비해다소크게측정되었음을알수있다. 반면일본해안마을의규모비교결과는앞서서술한것과같이우리나라해안곡계마을과동일한범위를가지고있음을알수있다. 과거아시아대륙과일본이육지로연결되어있었다는점을고려한다면지리적인것뿐만아니라문자, 문화를비롯하여역사적으로도관계가있기때문인것으로보인다. 결론적으로우리조상들이해안에서마을을형성할때아늑하다고인식한공간규모는 200~400m로시 ( 視 ) 거리에따른이론에의거하여친밀한인상을줄수있는근경역에해당된다고할수있다. 단, 본논문에서정방형으로환산한방법은공간상호규모를비교하기위한것에불과하며, 정방형이라는형태가인식적인의미를가지는것은아니다. 또한이러한범위는경관적인관점에서실제공간의면적보다는공간의직경의길이, 즉거리가규모의인식에직접적으로영향을줄것으로판단된다. 사람들이해안을찾는이유중한가지로창 ( 窓 ) 과같은열린항만이중요한의미를갖는다고생각한다. 그러나창이존재하는공간을경험할때이러한공간이사람들에게머무르기도좋은환경이어야한다. 일반적으로사람들은넓게트인공간에서개방감과동시에아늑함을느끼기는쉽지않다. 즉개방감과함께아늑함을느낄수있는머무르기좋은공간은사람들이더욱선호할것이라판단된다. 이에본논문은개방감과아늑함을함께느낄수있는해안의곡계지명을도출한것에의의를둔다. 그리고이러한공간의규모에대한기준을본논문에서제시하였다. 이처럼규모를가지는항만및친수공간은한층친밀한인상을줄가능성이높다고판단되며, 항만의중심으로부터근경역에속하는지형과식생이이러한친밀한인 상을줄수있는요점이기때문에지형의변경이나식생제거를수반한개발에는정밀한검토가필요할것이다. 본논문은곡계마을을대상으로위요공간에대한 2차원적분석으로마치는데한계가있다. 차후연구에서는 3차원적공간구조분석과자연외부공간에서의위요경계의조건을규명하는데추가적인분석을통해향후아늑함을조성할수있는항만의친수공간계획에대한더욱구체적인방향을제시할수있을것으로판단된다. 참고문헌 1. 강릉시청. (2019년 7월 1일검색 ). Gangneung-si. (accessed July 1, 2019). 2. 국토지리정보원 한국지명유래집. 서울 : 국토지리정보원. National Geographic Information Institute (The) Original of Korean Geographical Names. Seoul: National Geographic Information Institute (2019 년 7월 1일검색 ).. (accessed July 1, 2019). 4. 노선화, 문병선, 김태영 곡 ( 谷 ) 계지명농촌마을의입지특성과공간구조에관한연구 : 충청북도괴산군을대상으로. 한국농촌건축학회논문집 9권, 2호 : Roh Seonhwa, Moon Byungsun and Kim Taiyoung A study on the location characteristics in rural area of valley category. Journal of The Korean Institute of Rural Architecture 9, no.2: 박병철 지명어의한역화유형에관한연구 : 제천지역지명자료분석을바탕으로. 구결연구 13권 : Park Byeongcherl A study on the type of Sino-Korean translation of toponym words: Focusing on the toponym words in Jecheon Area. Kugyol Studies 13: 박재락 종택마을입지의풍수적형식논리를적용한사례연구. 국학연구 22권 : 국토연구제 102 권 ( )

49 Park Jaerak An analysis of family head village s background of formation and its location space in terms of Feng Shui: Focused on regions as Yeongju, Mungyeong, Sangju, Uiseong and Cheongsong. Korean Studies 22: 신상섭 한국전통마을의공간구성과환경설계원칙 ( 상 ). 국토 215호, 서울 : 국토연구원. Sin Sangseop. Spatial composition and environmental design principles of Korean Traditional Village (Ⅰ). Planning and Policy 215, Seoul: Korea Research Institute For Human Settlements. 8. 신진동 조선시대전통마을의입지와공간특성에관한연구. 박사학위논문, 경원대학교. Sin Jindong. A Study on Locational and Spatial Characteristics in Traditional Villages of Chosun Dynasty. Ph.D. diss., Kyungwon University. 9. 이돈주 땅이름 ( 지명 ) 의자료와우리말연구. 지명학 1권 : Lee Donju Data of place names and the study on the Korean language. Journal of the Place Name Society of Korea 1: 이진동, 김정규, 정태열, 김유일 지명과지형적위요 ( 圍繞 ) 및공간규모인식의관계 : 전북진안군을중심으로. 한국전통조경학회지 23권, 4호 : Lee Jindong, Kim Junggyu, Jung Taeyeol and Kim Yuil Correlation between the geographical-name and scale of space surrounded with hills: Focused on Jinan-gun in Jeollabukdo. Korea Institute of Traditional Landscape Architectire 23, no.4: 이학동 전통마을의분석과풍수지리이론을통해서본주거환경조성원리의탐색. 주거환경 1권, 1호 : Lee Hakdong Research on principles for making residential environment view from of analysis of Pungsu Theory. Journal of The Residential Environment Institute of Korea 1, no.1: 한필원, 민관식 지형공간정보를이용한전통마을의입지및공간구조분석방법. 한국지적정보학회지 18권, 2호 : Han Pilwon and Min Kwansik Location analysis method of a traditional village using geo-spatial information. Journal of The Korean Cadastre Information Association 18, no.2: 홍경희 촌락지리학. 서울 : 법문사. Hong Kyunghee Village Geography. Seoul: Bubmunsa. 14. Appleton, J The Experience of Landscape. Hoboken: Wiley. 15. Forman, R. T Land Mosaics: The Ecology of Landscapes and Regions. Cambridge: Cambridge University Press. 16. Ushio Saitho, Masatoshi Takeshita and Kenji Ueshima 地形的圍繞の認識と空間規模との關係に關する硏究. Report of Port and Harbour Research Institute 31, no.2: Ushio Saitho, Masatoshi Takeshita and Kenji Ueshima The intertwining of the recognition of terrain and the size of space. Report of Port and Harbour Research Institute 31, no.2: Yoshio Nakamura 風景を創る 環境美学への道. 東京 : 日本放送出版協会. Yoshio Nakamura Create a Landscape: The Road to Environmental Aesthetics. Tokyo: Japan Broadcasting Publishing Association. 논문접수일 : 심사시작일 : 심사완료일 : 해안곡 ( 谷 ) 계마을을대상으로한위요공간규모에관한연구 47

50 요약 주제어 : 곡 ( 谷 ) 계지명, 위요공간, 공간규모, 해안 최근항만은무분별한매립을통한규모의확장과획일화된경관계획으로해안의원풍경과그지역만의항만공간정체성을잃어가며, 도시구조와의결별로사람들의왕래를사라지게했다. 이에본논문에서는항만의친수공간발전을위해이용자를고려하여사람들이선호하는공간감을부여하고편안함및친숙함을느낄수있는규모의기준을제공하고자했다. 그결과본논문에서제시한방법에따른위요공간의규모는 40,000~150,000m 2 로나타났으며, 다양한곡 ( 谷 ) 계마을의형태를고려하여공간상호규모를비교 하기위해정방형으로환산한결과값인 200~400m가시 ( 視 ) 거리분할이론에의거하여친밀한인상을주는거리로나타났다. 친숙함을느낄수있는공간의속성을단순히규정하기에는어려움이있지만인간이일정규모의공간을하나의친숙한공간으로인식하는공간감각의의미와연결된다면거리가규모를인식하는데직접적인관계가있을것으로판단된다. 따라서본논문의결과를활용한다면휴먼스케일을고려한한국다운항만공간조성과매력적인경관창출에도움이될것으로판단된다. 48 국토연구제 102 권 ( )

51 국토연구제 102 권 (2019.9): pp49~67 Analysis of Stormwater Runoff Reduction Effect Through Applying Low Impact Development Practices in a Flood Prone Area : Case of Incheon, South Korea * 상습침수지역의 LID 적용을통한우수유출량저감에관한연구 : 인천광역시를중심으로 Lee Jinhee 이진희 **, Kim Hyunwoo 김현우 *** Abstract Flood damages are continuously growing in urban areas due to the increase of impervious surfaces and the occurrence of extreme rainfalls caused by climate change. While traditional drainage systems are still effective in controlling stormwater runoffs, they are inefficient in terms of costs and performance since they are only used for stormwater management. Low impact development (LID) practices, a comprehensive soft-engineering approach, provides multiple benefits in diverse aspects, and has been recently adopted as a key concept for sustainable stormwater management in many countries. By using the LIDMOD3 modelling tool, the present study simulates the hydrologic performance of LID practices in three different scenarios and explores the cost effectiveness of LID measures compared to the conventional sewer pipeline infrastructures. A flood-prone area in Bupyeong borough, Incheon, South Korea, was selected as the study site. The findings show that the district s total runoff volume and peak flow are reduced by 14-36% and 33-66%, respectively, in comparison with conditions without LID installation. Non-point source pollutant loads diminish by 15-36%, depending on the LID designed scenarios. Upon the consideration of construction costs, the results indicate that LID practices can be more efficient when they are partially developed rather than when various LID practices are applied to an entire area. The study concludes by suggesting policy implications as to how government agencies and local jurisdictions can more effectively adopt LID practices as part of sustainable stormwater management planning and link with future urban regeneration projects. Keywords: Stormwater Runoff, LID, Green Infrastructure, LIDMOD3, Flood, Incheon * This research was supported by the National Research Foundation of Korea (NRF) grant funded by the Korea government(msit; No.NRF-2019R1G1A ). ** 인천대학교일반대학원도시행정학과석사과정 ( 제1저자 ) M.A. Candidate, Dept. of Urban Policy & Administration, Incheon National Univ. Primary Author leejh610@inu.ac.kr *** 인천대학교도시행정학과조교수 ( 교신저자 ) Assistant Professor, Dept. of Urban Policy & Administration, Incheon National University Corresponding Author kimhw@inu.ac.kr Analysis of Stormwater Runoff Reduction Effect Through Applying Low Impact Development Practices in a Flood Prone Area 49

52 I. Introduction The average precipitation of Incheon, South Korea, for the last 10 years (2009~2018) was around 1,195mm (KMA Open Weather Data Portal 1) ), which was not significantly different from previous years. However, flood damages within urban areas kept increasing over the years due to the frequent occurrence of local torrential rainfalls, which increased the intensity of precipitation. Particularly, damages caused by inland inundation are rapidly increasing because of the limited capacity and storage of urban sewer systems. This eventually resulted to an overflow which exacerbated the flood damage, especially during the heavy rain (Lee 2013). I nland inundation is mostly affected by three major causes: sediment deposits within the sewage conduit, tangled underground facilities and sewer pipes, and overflow from the sewer line due to increased river / stream stage. Climate change, which altered the precipitation pattern and increased the occurrence of natural disasters, has aggravated these three causes all over the world (IPCC 2007). To properly adapt and minimize impacts from these climatological changes, environmentally friendly urban storage and drainage approaches are required in the long-term. The concept of low impact development (LID) techniques was first introduced in the mid-1990s by the Department of Environmental Resources of Prince George s County, Maryland in the U.S., to address solutions for sustainable and cost -effective stormwater management (USHUD 2003). LID refers to various practices regarding infiltration, detention, treatment, and flow control of stormwater close to its source in order to minimize runoff and prevent the deterioration of water quality (UACDC 2010; USEPA 2018). The overall goal is to mimic the site s natural hydrological functions as pre-developed conditions (Benedict and McMahon 2006). Currently, more than 30 techniques have been developed and adopted to collect and treat stormwater runoff. The USEPA broadened the concept of LID and included these practices for key green infrastructure planning (USEPA 2018). Since 2013, when the Ministry of Environment (ME) in South Korea adopted the LID manual for environmental impact analysis, a number of agencies in South Korea started to actively control non-point pollution sources through implementing LID practices (ME 2013). To recover the urban water circulation system, various countries are now using a similar concept of LID, while its terms vary: Water Sensitive Urban Design (WSUD) in Australia, Sustainable Urban Drainage (SUD) in England, Low Impact Developments Urban Design (LIDUD) in New Zealand, and Sound Water Cycle on National Planning (SWCNP) in Japan (Kang, Lee, Koo and Cho et al. 2011). China is also paying numerous attentions in green infrastructure planning by pursuing the concept of Sponge City. Since 2015, the government has been requiring 30 pilot sponge cities to retrofit at least 20% of lands serving a pervious function (Chan, Griffiths, Higgitt and Xu et al. 2018). While the terms were used differently in diverse countries, the replacement of impervious pavements with various eco-friendly alternatives, making places more resilient to floods and bringing back the natural water 1) (accessed on January 13, 2019). 50 국토연구제 102 권 ( )

53 circulation systems, is a common characteristic. Increase of impervious surfaces induces much more amounts of peak flow and reduces the time to peak, which eventually increases the total flood volume, decreases groundwater storage, and minimizes the baseflow of urban streams (Kim and Li 2016; Choi and Shin 2002). One of the major goals of LID is to capture the first 1-inch (25.4mm) of rainfall, which contains the most harmful substances that remain on the roads or impervious surfaces, in order to minimize the pollutants flowing into the sewerage systems. Although various drainage techniques have been continuously introduced to rapidly reduce the initial rainfall, the past sewerage systems have failed to manage the quality of stormwater outflows. Particularly, major limitation still exists during the heavy downpours with the frequent occurrence of back flow from sewer grates. LID practices are not only efficient in controlling and capturing the stormwater at the initial stage, but also effective in managing both quality and quantity of runoff (USEPA 2018; ME 2013). Several studies identified that the construction of LID practices is more cost-effective compared to retaining the traditional sewer lines and detention facilities (USHUD 2003). Considering the environmental and economic aspects of LID, more active technical and political efforts should be paid for restoring the hydrologic and ecological functions of urban watershed in the long-term. This study examines the reduction effect of LID practices in stormwater runoff by adopting the LIDMOD3. A district prone to floods in Bupyeong borough, Incheon, was selected for the study area. This paper is constructed according to the following sections. First, previous studies that used multiple modeling approaches to examine the impact of LID practices are reviewed and the gap of existing studies is identified. Second, research methodologies, including study area, modeling technique, scenario setting, and analyzing method are presented. Third, runoff reduction effects of LID, as well as installing costs for both traditional drainage systems and LID practices, are compared and analyzed. The study concludes by suggesting policy alternatives for the effective and sustainable management of stormwater runoffs. II. Previous Studies on Examining the Effects of LID Practices LID practices have been introduced to minimize the runoff and increase the water quality that flows into the existing sewer lines. Several studies have adopted different types of LID by using various rainfall-runoff simulation models, for estimating the runoff reduction effects (Rosa, Clausen and Dietz 2015). The US Environmental Protection Agency (USEPA) s Storm Water Management Model (SWMM) and System for Urban Stormwater Treatment and Analysis Integration (SUSTAIN) are the most typical simulation models that have been employed to examine the runoff reduction effect of LIDs (Gao, Wang, Hunag and Liu 2015; Jeon, Park, Park and Kim 2014; Luan, Fu, Song and Wang et al. 2017; Nam, Kim, Kim and Kim et al. 2017; Park, Yoo, Park and Yoon et al. 2008; Rossman 2010; Suh and Lee 2013; Jeon, Park and Lee 2013; Jang, Mun and Yang 2013). However, other tools, such as the Analysis of Stormwater Runoff Reduction Effect Through Applying Low Impact Development Practices in a Flood Prone Area 51

54 Long-Term Hydrologic Impact Assessment-Low Impact Development (L-THIA-LID) model (Liu, Bralts and Engel 2015), LIDMOD2&3 (Kim, Kim, Choi and Jeon 2011; Jeon, Yoon, Lee and Cho et al. 2015; Jeon and Seo 2018), Soil and Water Assessment Tool (Woo, Ryu, Moon and Jang et al. 2012), and the improved Soil Conservation Service Curve Number (SCS-CN model) (Cronshey, Roberts and Miller 1986; Roehr and Kong 2010) were also often used to measure the effects of LIDs in specific regions. By using the SWMM, Park, Yoo, Park and Yoon et al. (2008) applied porous pavements and green roofs in the Geo-du district (25.2ha) of Chuncheon, South Korea, and identified that runoffs have been reduced by 32.7% in 44.44mm (2-yr design storm) rainfall and 23.6% in 73.16mm (10-yr design storm) precipitation. With the same method, Suh and Lee (2013) simulated the effect of infiltration trench and chamber in three different cities residential areas in South Korea and found that stormwater runoff reduced from 19.4% to 23.7%, while infiltration amount increased by up to 20%. Luan, Fu, Song and Wang et al. (2017) simulated five different LID scenarios in the mountainous Fragrance Hills region of Beijing, China. All LID practices had a positive effect on runoff reduction, but the results showed that the concave greenbelt was the most effective in controlling the storm events with the runoff reduction percentages ranging from 5.2% to 57.3% for 1, 2, 5, and 10-year rainfall events, whilst the vegetative swale provided the least reduction effect, ranging from 0.3% to 3.0%. Jeon, Park and Lee (2013) applied porous pavements and rainwater detention tanks in Towoul stream basin, Changwon, and discovered that peak runoff reduced by 1.2~6.2% for porous pavements, while the reduction effect of detention tanks ranged from 2.1~11.7%, depending on the size of a tank. Jang, Mun and Yang (2013) also found that porous pavements near the Mokgam stream basin (5,569ha) minimized the maximum daily rainfall runoff up to 9.5%. Gao, Wang, Hunag and Liu (2015) used the SUSTAIN to find out the impact of five different best management practices (BMPs; green roof, wet pond, vegetated swale, porous pavement, bioretention basin) in an industrial site located in Ma anshan, Anhui Province, China, and discovered that approximately 41% of total runoff volume was reduced after running the simulation. Employing both SWMM and SUSTAIN, Nam, Kim, Kim and Kim et al. (2017) analyzed the runoff reduction amount, as well as the mitigation effect in relation to construction costs for different LID scenarios (mix of permeable pavement, infiltration ditch, and ecological detention pond) in Oncheon Creek Watershed in Busan, South Korea. The results showed that there were % of runoff volume reduction through LID measures, while infiltration ditch was most cost-effective. Jeon, Park, Park and Kim (2014) applied the SUSTAIN while evaluating four LID practices (e.g., bioretention, infiltration trench, rainwater harvesting, and porous pavement) near the Andong city bus terminal site. The study found that a site s surface runoffs have diminished by up to 14-28% on average during the 550 simulations. The LIDMOD model was often performed for measuring the runoff reduction effect of LID practices in South Korea. With the LIDMOD2, Kim, Kim, Choi and Jeon (2011) demonstrated that the bioretention was the most effective means of controlling runoff compared to the artificial wetland and wet pond in a residential district of Yongin, 52 국토연구제 102 권 ( )

55 South Korea. A study from Jeon, Yoon, Lee and Cho et al. (2015) revealed that the annual stormwater runoff volume was reduced by 23% while the annual infiltration rate increased by 23.1% when porous pavement, green roof, and bioretention was developed in a site near the Andong city hall, in South Korea. Jeon and Seo (2018) recently updated the previous LIDMOD2 into LIDMOD3 and used them for assessing the function of permeable pavement and bioretention in an apartment complex in Seoul, South Korea. The results of runoff analysis revealed that the discharge volume decreased by a maximum of 37% with the 64% increase of infiltration rate. Roehr and Kong (2010) used the improved SCS-CN model to examine the effect of green roofs and identified that the annual rooftop runoff was minimized by 29% in Vancouver, Canada, and 55% in Shanghai, China. To summarize the previous studies, simulation analyses that had relatively large study sites tended to construct porous pavements, which comparatively had less spatial restrictions. In addition, a majority of studies were likely to simulate LID practices, such as bioretention, permeable pavement, and green roof, while their runoff reduction effects were mostly 10% higher compared to other LID measures (Jeon and Jung 2019). However, most of the study sites were located in a relatively small municipality that did not have severe flooding issues in the past whereas only a limited number of LID techniques were introduced during the simulation processes. Moreover, the USEPA s SWMM and SUSTAIN could not sufficiently integrate Korea s geographical conditions while running the simulation. By employing the LIDMOD3 and choosing the flood prone area for the study site, this research attempts to fill the previous literature gaps and determine the practical runoff reduction effects of diverse LID practices. III. Research Methods 1. Study Area Bupyeong borough has the third highest impervious surface proportion (63.3%) in the entire 10 boroughs of Incheon (i.e., overall impervious surface ratio of Incheon: 19.9%; Kim and Park 2017). Among 31 flood prone zones, eight areas are located in Bupyeong and the number of accumulated flooded areas was 1,250 as of 2017, which was the highest among the overall boroughs. The study area is located within the Shipjung-1 district of Bupyeong borough in Incheon, with the total size of 42,690.6m 2 (see <Figure 1>). It mostly consists of single-family housings with some commercial buildings at the edge of streets and a small park on the southeast. Most of the land uses are classified as 2nd class general residential zone and overpopulated constraint district. The target area has historically suffered from floods due to it s high impervious ratio (97.3%) and low lying location. Since the slope near the study area is falling from the northeast to south-west side, runoff flows at the same direction, following the impervious surfaces. This has triggered more flood damages on the southern part of the study area. In addition, dwelling units in the study area stand close together and roads are fairly small, which prevent the construction of large-scale Analysis of Stormwater Runoff Reduction Effect Through Applying Low Impact Development Practices in a Flood Prone Area 53

56 Figure 1 _ Study Area Surrounding Environments and Slope Table 1 _ Land Cover and Size of the Study Area (unit: ) Classification Pervious Surface Land Cover Area Size by Each Hydrologic Soil Type Barren Impervious Surface Lawn Rooftop Driveway Walkway Parkinglot Playground Total Size Others B ,764.2 C , , , , , , , , ,016 1, , , ,632.8 D Total Size stormwater detention or mitigation facilities. Thus, the study area can be a suitable modeling site for investigating the implementation effect of small-scale, soft-engineering approaches, such as LID practices. <Table 1> shows the detailed land cover and size of the study site. 2. Scenario setting This study analyzed the runoff reduction effect of LIDs by applying three different LID scenarios in one of the flood prone areas in Bupyeong borough, Incheon. Each scenario is constructed as LID-friendly design (Scenario1), LID-partially adopted design (Scenario 2), and LID- conservative design (Scenario 3). Scenarios have adopted LID practices by balancing the three types of LID functions (vegetated facility, retention facility, and infiltration facility), 54 국토연구 제102권( )

57 Table 2 _ LID Techniques Adopted in Each Scenario Scenario LID Type Practice Installation Location Size(m 2 ) Vegetated Facility Rain Garden Walkways in the southwest part Retention Facility Bioretention Two bare lands Scenario1 Scenario2 Scenario3 Porous Pavement Walkways in the southwest part Infiltration Facility Infiltration Chamber Rooftops in the study area 2,135.0 Porous Pavement Entire roads, walkways within park, and parking lots 11,405.1 Vegetated Facility Rain Garden Walkways in the southwest part (within flood prone area) Retention Facility Bioretention Bare land located in the West side Porous Pavement Walkways in the southwest part Infiltration Facility Infiltration Chamber Rooftops within flood prone area 1,450.0 Porous Pavement Two large roads and walkways within park 3,660.4 Vegetated Facility Tree Box Filter 50 tree box filters in the southwest part Porous Pavement 50% of walkways in the southwest part Infiltration Facility Infiltration Chamber Rooftops within northwestern side (two blocks) Porous Pavement Central road (middle part) and walkways in the park 1,852.9 which have been listed as the key LID elements in the Korean ME and Korea Environment Corporation s LID Implementation Manual. Appropriate LID practices for multi-family housing complexes, roads, parking lots, and rooftops were determined through considering the previous projects that adopted LID facilities in practice. <Table 2> shows the type, installed facility, location, and size of LID techniques for each scenario. Scenario 1 adopted the most numbers of LID practices, while Scenario 2 concentrated the installation of LID facilities within the flood prone area. Scenario 3 included a limited number of LID techniques in certain places that could minimize the flood damage. Specifically, LID practices were installed based on the geographical characteristics of site and features of each measure. The overall size of LID facilities for each scenario was 35% (14,988m 2 ), 15% (6,263m 2 ), and 7% (2,915m 2 ), respectively, out of total drainage area. In Scenario 1, rain gardens and porous pavements replaced the existing walkways, which were located on the southwest part of the study site. Since the slope of this area flows from northeast to southwest, two practices, which have a high efficiency in capturing stormwater runoff within the limited space, are the most suitable approaches in this site. Bioretentions were created in two empty bare lands, where their locations allowed to efficiently collect the northeast side of stormwater and provide a rest area for residents. Every driveway, walkway within a park, and parking lot was designed as permeable pavers, while infiltration chambers were installed to receive runoffs from the rooftops. Since the major purpose of bioretention and infiltration chamber is to collect the first 25.4mm runoff, which contains the significant portions of non-point pollutant, the size of those facilities was calculated based on collecting the first inch (25.4mm) of stormwater (Jeon and Seo 2018; Prince George s County 1999). In Scenario 2, rain gardens Analysis of Stormwater Runoff Reduction Effect Through Applying Low Impact Development Practices in a Flood Prone Area 55

58 Figure 2 _ Applied LID Techniques for Each Designed Scenario Scenario 1 Scenario 2 Scenario 3 and infiltration chambers were preferentially implemented where they over-lap with the flood prone area. Two major driveways that pass through the study area were replaced to permeable pavement, whereas walkways within a park and located on the southwest part of the study area were transformed into porous pavers. Major driveways that penetrate the east-west side of site can be relatively installed with low construction cost and they can intersect the runoffs in phases at the middle. Installing porous pavements in narrow alleys are likely to be inefficient in economic and environmental aspects. Bioretention was only installed on the west side of the existing bare land. This is due to the slope, which falls from Ship-jung park induces the runoff in western area. Bioretention may capture some of these runoffs and delay the peak time. Lastly, Scenario 3 employed the least amount and type of LID facilities. Instead of installing rain gardens above the walkways in the southwest area, 50 tree box filters were implemented, following the main road. Porous pavements were only given to 50% of existing southwest side walkways, walkways within a park, and a driveway located on the center, which penetrates from east to west. Since the slope flows from the northeast to southwest side, infiltration chambers were mainly installed on two blocks of the northwest corner to cover the rooftop runoffs. <Figure 2> shows the locations and adopted LID techniques for three different scenarios. 3. Data Acquisition and Analysis The data in this study have been analyzed in two phases. First, LIDMOD3, a site evaluation tool (Tetra Tech 2005) that has been recently co-developed by the Korea National Institute of Environmental Research and Andong University (Jeon and Seo 2018) was used to assess the runoff reduction effect of LID practices. This modeling tool was continuously updated from 2010 in order to estimate the runoff depth, infiltration depth, peak flow, and annual loadings (e.g., BOD, SS, T-N, T-P). The significant update of its recent version (2018) includes the runoff curve 56 국토연구 제102권( )

59 Table 3 _ Data Sources number (CN) being re-established based on a more Data Source detailed Korea soil map, which enabled the calculation Building Footprint, Road, Korea Ministry of Land, of the direct amount of runoffs from specific land uses Land Use Infrastructure and Transport and LID practices (Jeon and Seo 2018). Since the Korea Rural Development Soil, Elevation Administration storage capacity of LID practices are calculated based Korea Meteorological Precipitation on the Ministry of Environment s LID Facility Design Administration Guideline and the infiltration capacity was computed on Flood Prone Area Incheon Metropolitan City Digital Elevation Mode the basis of domestic case study s soil drainage National Geographic Information (DEM) conditions, the simulation results of LIDMOD3 may estimate the runoff characteristics precisely as similar to other sophisticated simulation tools for Korea (Jeon and Seo 2018; National Institute of Environmental Research 2019). Because modelling tools, such as SWMM and SUSTAIN, estimate the runoff with the consideration of various drainage systems and land-use conditions, their assessments may work better in a complicated urban area than LIDMOD3, which mostly rely on hydrological soil group for the simulation. However, SWMM requires diverse variables (e.g., pervious level for each sub-basin, Manning s roughness coefficient, etc.), while LIDMOD3 has relatively simple modelling process with small number of variables. Thus, local governments may readily manage the impervious surfaces, prepare flood mitigation plans within a short-time, and concisely check the runoff reduction effect of LIDs during the environmental impact assessment. To run the model, basic data, such as building footprint, road, land use, soil, elevation, precipitation, and flood prone area were needed (see <Table 3>). By using the ArcGIS v10.6, locations, as well Figure 3 _ Monthly Precipitation in 2018 as sizes for each data were calculated. While building footprint, road, and land use data were acquired from the Korea Ministry of Land Infrastructure and Transport (MLIT), soil and elevation data were gathered from the Rural Development Administration. Other data, such as precipitation data for 2018 were obtained from the Korea Meteorological Administration (see <Figure 3>), flood prone area from the Incheon. Metropolitan City, and digital elevation model (DEM) data from the National Geographic Information Institute. Analysis of Stormwater Runoff Reduction Effect Through Applying Low Impact Development Practices in a Flood Prone Area 57

60 Second, costs for LID practices that have been applied in each scenario were mostly estimated using the LID Practices Costing Tool developed by the Toronto and Region Conservation Authority and University of Toronto (2013). The costs for tree box filter, however, were not allowable for the calculation in the tool. Thus, the cost was drawn by adopting the LID manuals from the Korea Land and Housing Institute (2015), as well as previous projects that have recently constructed the best management practices in South Korea. The cost efficiency of LID approaches was then shown by comparing the simulated construction costs of conventional drainage systems and LID practices. Two assumptions were made for installing the sewer pipes within the study area. First, it was hypothesized that all existing pipes (total length: 1,767.26m) would be replaced by 600mm-diameter pipes. This is because the current sewer pipes are old enough to be replaced by new ones. According to Lee, Kim, Park and Kim et al. (2016), a pipe diameter size less than 600mm was ineffective to control the urban runoff in South Korea. For this reason, sewer pipes that had a diameter of less than 600mm were replaced and calculated. Since sewer lines were generally replaced by stages, the entire pipeline replacement costs were not considered in this study, which may significantly increase the b/c ratio of LID practices. Target pipelines were determined by using the drainage system spatial data obtained from the City of Incheon. The installation costs for sewer pipe were calculated based on the Korean ME s Subsidy Compilation and Execution Management Guideline for Sewer System (ME 2018). The guideline provides a detailed average fee for sewer pipe per 1-meter, including construction, basic design, construction document, and supervision expenses. Depending on the pipe type (e.g., concrete, steel, plastic, and cast iron) and diameter, the estimated price was different. IV. Result and Discussion 1. Hydrological Performances of LID Practices Hydrological performances of the three LID development scenarios were simulated based on the uniform storm event. The annual rainfall of Incheon in 2018 was 1,134mm. The results show the runoff/infiltration depth, as well as peak flow rate and time, and annual loadings for three circumstances ( pre-development, development without LID, and development with LID ). Study area slope was determined by adopting the average slope, which was about 11.3%. Considering that there were no developments in the study area, runoff and infiltration depths were 280.0mm/yr and 364.6mm/yr, respectively (see <Table 4>). In current condition (development without LID), the runoff depths of each scenario were fairly different. Scenario 3 had the most amount of runoff and Scenario 1 produced the least runoff depth. Even though the runoff depths for each scenario were different, infiltration, peak flow and time to peak were same with the values of 140.9mm/yr, 8.14m 3 /sec. and 0.68 minute, respectively. With the application of LID practices, Scenario 1, which applied the most numbers of LID technologies, had 58 국토연구제 102 권 ( )

61 Table 4 _ Hydrological Characteristics of the Three Designed Scenarios Circumstance Scenario 1 Scenario 2 Scenario 3 Runoff Depth (mm/year) Pre-development 280 Development without LID Development with LID Infiltration Depth (mm/year) Pre-development Development without LID Development with LID Peak Flow (m 3 /second) Pre-development 1.29 Development without LID 8.14 Development with LID Time to Peak (minute) Pre-development 4.30 Development without LID 0.68 Development with LID the largest impact on reducing runoff depth (-36.2%) and increasing infiltration depth (329.9%). The peak flow was also diminished significantly from 8.14 to 2.80m 3 /sec., while the peak time was delayed for about 1.29 minute. The peak flow was also diminished significantly from 8.14 to 2.80m 3 /sec., while the peak time was delayed for about 1.29 minute. Even though the reduction impact was relatively lower than Scenario 1, the overall runoff depth and peakflow rate of Scenario 2 declined by about 28.5% and 51.1%, respectively, while the infiltration depth and time to peak increased by approximately 199.4% and 0.71 minute. The overall reduction effect of Scenario 3 was relatively low compared to other scenarios as the LID technologies were partially installed within the study site. Runoff depth decreased by only about 14.5% and the infiltration depth increased by 89.8%. While peak flow reduced by more than 2.68m 3 /sec, peak time was delayed for only about 20 seconds. <Figure 4> compares the unit hydrolograph for three scenarios. Considering that the study site s predevelopment runoff depth was 280mm/year, Scenario 1 s runoff increased only about 110mm/year, while the two other scenarios runoff exceeded more than two times the pre-development condition. Despite the fact that a significant portion of LID practices were planned in Scenario 1, the results demonstrate that the potential impact of LID techniques are fairly high and they can be a suitable option to replace the traditional Scenario 1 Scenario 2 Scenario 3 Figure 4 _ Unit Hydrolograph for Each Scenario Analysis of Stormwater Runoff Reduction Effect Through Applying Low Impact Development Practices in a Flood Prone Area 59

62 drainage systems (Ahiablame and Engel 2012; Martin-Mikle, Beurs, Julian and Mayer 2015). Typical low-rise residential complexes in South Korea lack green areas, with a majority of the space covered by impervious surfaces and surrounded by multiple small alleys. This physical feature had eventually promoted more inundations during the heavy rainfall due to the reverse flow from the traditional sewer system. Locations like the study site, where the old low-rise residential blocks are clustered, will be particularly appropriate to employ LID practices not only to have beneficial effect on the environmental side, but also on the economic (e.g., saving sewer infrastructure budget) and social sides (e.g., improving overall landscape without a massive re-development) (Li, Deng, Li and Li et al. 2017). Further, many regeneration projects are currently implemented in South Korea to revitalize the old neighborhoods without demolishing the previous structures entirely. Because the present sewer pipeline infrastructures are insufficient to control stormwater runoffs and spaces are limited for large-scale drainage pumps and facilities placement, LIDs can fit properly in these types of communities with the deployment of relatively small spaces while maintaining the neighborhood s own identity. 2. Storage Capacity of Each LID Practice Storage volume for each LID practice was computed by using the coefficient of BMPs in order to compare the efficiency of LID types. Since vegetated facilities and retention facilities generally had a high unit storage volume, reduction widths were not so significant compared to the infiltration facilities when its size shrunk in different scenarios (see <Table 5>). While the size of typical infiltration facility, such as porous pavement, was relatively larger than other practices in Scenarios 2 and 3, the storage volume was substantially low. Particularly, the size of porous pavements was about 10 times larger than tree box filters, but the storage capacity had no significant difference. Infiltration chambers also had low unit storage volume compared to other practices. This finding indicates that infiltration facilities are likely to be inefficient in capturing runoff flows, and thus, suggests that retention and vegetated facilities-centered Table 5 _ LID Storage Volume in Each Scenario Scenario LID Type Practice Scenario 1 Scenario 2 Scenario 3 Vegetated Facility Retention Facility Infiltration Facility Vegetated Facility Retention Facility Infiltration Facility Vegetated Facility Infiltration Facility Size (m 2 ) Storage Volume (m 3 ) Rain Garden Bioretention Porous Pavement Infiltration Chamber 12, , Rain Garden Bioretention Porous Pavement Infiltration Chamber Tree Box Filter Porous Pavement Infiltration Chamber 4, , , Note: Coefficient by LID Practices 1) Bioretention and tree box filter: m 3 /m 2 2) Porous pavement: m 3 /m 2 3) Infiltration chamber: m 3 /m 2 60 국토연구제 102 권 ( )

63 installation should be mainly conducted with the Table 6 _ Capital Costs of Implemented LID Practices infiltration facilities locating in a proper place. Scenario LID practice Size(m 2 ) Capital Cost($) Rain Garden ,206.6 Bioretention , Cost Efficiency of LID Practices Scenario Porous Pavement 12, ,097, Infiltration Chamber 2, ,558.4 The cost efficiency of LID practices was investigated Total Cost 1,263,970.0 by comparing the cost of conventional sewer system Rain Garden ,467.4 with LID techniques implemented in each scenario. Bioretention ,661.1 Regarding the costs for LID practices, each scenario s Scenario 2 Porous Pavement 4, ,663.0 capital costs were calculated based on the size of the Infiltration Chamber 1, ,713.3 Total Cost 550,504.9 LIDs (see <Table 6>). Given that the costing tool Tree Box Filter ,034.5 showed the 2010-dollar value for the pricing, an Scenario Porous Pavement 2, ,750.0 inflation rate of 11.32% was employed to change it into 3 Infiltration Chamber ,257.1 a 2018-dollar value. Among LID practices, porous pavement had the Total Cost 298,041.6 highest unit( m2 ) cost, followed by tree box filter, infiltration chamber, and bioretention. Rain garden relatively had a lower construction fee compared to other techniques. While Scenario 1 had the highest LID costs with $1,263,970, the total cost of Scenario 2 was only 43.5% of Scenario 1, albeit there was only an 8% difference in runoff reduction efficiency between both scenarios. This evidence demonstrates that the strategic installment of LID practices can be a better approach in the cost effect perspective rather than covering the entire area with LID controls. The number of LID techniques included in Scenario 3 was relatively small and the total capital cost was about $298,041, which was more than two times smaller than Scenario 2. A construction cost (per 1-meter) of 600mm-diameter sewer pipe was different based on the pipe materials and by the following order: steel, plastic, concrete, and cast iron (see <Table 7>). The overall installment price included the construction fee, as well as the basic design, construction document, and supervision expenses. In Assumption 1, the costs of replacing all the Table 7 _ Overall Construction Costs for Sewer Pipe existing pipe with steel was about $2.3 million, while in Each Assumption cast iron took only 62% ($1.45 million) of its cost. The Overall Construction Cost($) Pipe material Assumption 1 Assumption 2 length of sewer pipes under the 600mm-diameter Cast Iron 1,456, ,504.8 was 707.2m. Replacement fees for these pipelines Concrete 1,748, ,726.3 (Assumption 2) ranged from $0.6 to 1 million, depending Steel 2,361, ,009,323.1 on the materials. Plastic 1,899, ,586.7 <Figure 5> illustrates the percentage of LID costs Analysis of Stormwater Runoff Reduction Effect Through Applying Low Impact Development Practices in a Flood Prone Area 61

64 Figure 5 _ Proportion of LID Costs to Conventional Sewer Pipe Costs to conventional sewer pipe costs for three different scenarios. In Assumption 1 (replacement of entire sewer pipelines within the study area), all scenarios LID costs were lower than the pipe replacement costs. Particularly, LID costs were only 11.9~86.8% of pipe prices, with respect to the scenarios. Cast iron, the most commonly adopted pipe material in South Korea, had the cheapest price, with the costs of LIDs being even lower than their pipe costs by about 13.4% in Scenario 1. However, when the LID costs were compared with partially replacing pipe costs (Assumption 2), their costs in Scenario 1 were more expensive than with the pipeline replacement, no matter what the materials were. Except in the Scenario 1 case, the two other scenarios LID construction costs were relatively lower than the sewer pipeline costs, ranging from 27.7% to 88.9%. In sum, the establishment of LIDs was financially effective compared to the pipeline replacement, except in a very conservative development case (Scenario 1 in Assumption 2). Particularly, Scenarios 2 and 3 were cost effective in all assumptions. Considering the runoff reduction effects of LID scenarios, there were only 7.7% runoff depth difference between Scenarios 1 and 2, whereas the costs difference was about $713,465 (56%); that is, Scenario 2 can be more cost-effective in terms of runoff attenuation. A combination approach to conventional drainage systems and LID measures is necessary because one may not solely rely on LID practices in controlling the stormwater runoff; thus, Scenario 2 can be a practical approach in reality compared to Scenario 1. To fully address the cost-effectiveness of LID practices, life-cycle costs should be examined in future research. The findings from previous literatures revealed that the maintenance charges for LID practices are substantially low for long-term compared to fully re-constructing the whole pipeline infrastructure (Joksimovic and Alam 2014; Lee, Kim, Pak and Jang et al. 2010). From this view, LID measures can rise as a sustainable alternative in mitigating future flood damage. 62 국토연구제 102 권 ( )

65 V. Conclusions This study simulated the hydrologic performance of the three LID designed scenarios under an identical storm event using the LIDMOD3 in a flood prone area, the Ship-jung district in Incheon, South Korea. While the runoff reduction effect was fairly different in each scenario, LID installments attenuated stormwater runoff volume from 14% to 36% and increased infiltration depth by % in comparison with conditions prior to the LID. This result corresponds with the previous simulation studies, which demonstrated that the LID controls typically have a reduction effect by about 30%, even though the degrees and levels of LID practice were dissimilar in each case (Ahiablame and Engel 2012). With consideration to construction costs, a blend drainage system that combines both traditional pipeline infrastructures and LID measures should be constructed to more effectively manage the stormwater runoff. As Scenario 1 was revealed to have inefficiency in reducing the runoff compared to the costs invested for the construction, local planners and decision-makers should identify appropriate places where LID practices can function the most rather than adopting all areas using LID techniques. Albeit this study considered the slope and other geographical features of a site to install appropriate LID measures where floods could be controlled the most, future research should further conduct spatial analyses (i.e., location analysis), in order to identify optimal sizes and places for each distinct LID practice. One of the major concepts that the Korea ME is currently promoting for sustainable urban flood management is Smart Water City. The government is pursuing to upgrade the conventional sewer systems using various flood-resilient measures, such as up-to-date hard engineering approaches, LID techniques, and environmental land use policies. Until now, the pilot projects have been mostly given to new towns and cities, where there were no initial land uses, buildings, and infrastructures. Of course, LID practices may better perform and be well managed in a new planned city. However, existing cities are getting old and rapid industrialization during the s in South Korea made various disruptive asphalts and concretes with a massive amount of impervious surface in the inner city. This eventually limited green spaces in the city center. Inland Incheon, the 2nd highest impervious surface ratio city in South Korea, is experiencing significant damages from stormwater runoff and losing valuable ecosystem functions due to the conventional urban settings. LID approaches can be more appropriate in these hard-engineered infrastructure based cities by providing aesthetically pleasing places and enhancing the overall urban ecosystems. The study results, thus, may provide valuable insights to local decision-makers whilst redeveloping the old inner city and may successfully link these LID practices with the current urban regeneration projects for achieving sustainable urban water/stormwater management systems in the long-term. Although this study provides several noteworthy insights in demonstrating the potential reduction effects of various LID practices, some limitations need to be acknowledged. First, calibration for the flow modeling could not be conducted since there were no available monitoring data. By adopting monitoring devices at the site, calibration should Analysis of Stormwater Runoff Reduction Effect Through Applying Low Impact Development Practices in a Flood Prone Area 63

66 be checked in future studies to increase the reliability, as well as the validity, of research. Second, the study area of this research was focused on a specific plot in order to clearly identify the effects of LID practices within a residential-centered area. For a more accurate investigation of runoff reduction, further research is suggested in simulating the model at the watershed-level; thus, it may better estimate the generated runoff with the consideration of various hydrological and geographical features that may impact on final outlet. Third, this study may not fully consider various designed storms in estimating the produced runoff. With the actual past precipitation data, diverse designed storms as well as rainfall characteristics (e.g., heavy/average/limited water year) should be applied and classified for further study to predict the runoff in multiple meteorological conditions. Fourth, green roofs have not been applied during the simulation due to the current conditions of housings within the study site. Since the runoff reduction effect of green roofs has been proven effective in many previous literatures, further research is needed to better understand the influence of green roofs on attenuating stormwater runoff. Fifth, albeit this study applied a LID modeling tool in an attempt to discover the stormwater reduction, further study should focus more closely on the diverse water quality aspects with the usage of other simulation tools in order to discover more positive impacts and cost-efficiency in implementing LID practices. Finally, LIDMOD3 was widely adopted for the runoff estimation, but its usage is still limited to Korea. The benefit of this tool, which allows simple operation for users, was lacking in terms of considering the complex hydrological factors while running the simulation. Thus, further research is in need to expand this approach with the employment of other verified LID modeling tools (i.e., SWMM-LID) that may fully consider the complexity of urban drainage system and other geographical factors. In addition, modeling approaches that may consider the runoff efficacy of both LID controls and traditional drainage systems should be developed in the near future for a better estimation. 참고문헌 1. Ahiablame, L. M. and Engel, B. A Effectiveness of low impact development practices: Literature review and suggestions for future research. Water Air and Soil Pollution 223, no.7: Benedict, M. A. and McMahon, E. T Green Infrastructure: Linking landscapes and communities. 2nd ed. Washington, D.C.: Island Press. 3. Chan, F., Griffiths, J., Higgitt, D., Xu, S., Zhu, F., Tang, Y. T., Xu, Y. and Thorne, C Sponge City in China-A breakthrough of planning and flood risk management in the urban context. Land Use Policy 76: Choi Jiyoung and Shin Changmin Management of Nonpoint Pollution by Reducing Storm Runoff. Seoul: Korea Environment Institute. 5. Cronshey, R. G., Roberts, R. T. and Miller, N Urban hydrology for small watersheds (TR-55 REV.). American Society of Civil Engineers 55: Gao, J., Wang, R., Hunag, J. and Liu, M Application of BMP to urban runoff control using SUSTAIN model: Case study in an industrial area. Ecological Modelling 318: IPCC Climate Change 2007: The physical science basis. Cambridge: Cambridge University Press. 8. Jang Yeongsun, Mun Sungho and Yang Sunglin An 64 국토연구제 102 권 ( )

67 analysis of flood mitigation effect applying to LID in Mokgamcheon watershed using SWMM model. International Journal Highway Engineering 15, no.3: Jeon Sangmi, Park Jaehyeon and Lee Sangho Analysis of flood reduction effect using decentralized rainwater retention facility in urban area. Crisisonomy 9, no.12: Jeon Jihong, Park Sangsun, Park Changi and Kim Taedong Evaluating implementation of EPA SUSTAIN model for cost-effective LID design. Journal of the Korean Society of Urban Environment 14, no.2: Jeon Jihong, Yoon Yeongjun, Lee Suenghyeon, Cho Suhyeon and Kwon Taeyong Analysis of reducing nonpoint source pollution by Low Impact Development (LID) application at Andong City Hall. In Proceedings of 2015 Joint Conference of Korean Society of Water & Wastewater Korean Society on Water Environment. August 27-28, Goyang: Kintex. 12. Jeon Jihong and Seo Sungchul LIDMOD3 development for design and evaluation of low impact development. Journal of Korean Society on Water Environment 34, no.4: Jeon Jinhong and Jung Kwangwook The effect of infiltration rate of foundation ground under the bioretention on the runoff reduction efficiency. Journal of Korean Society on Water Environment 35, no.1: Joksimovic, D. and Alam, Z Cost efficiency of low impact development (LID) stormwater management practices. Procedia Engineer 89: Kang Jungeun, Lee Myeongjin, Koo Yusung, Cho Kwangwoo and Lee Jaewook Urban Renewal Strategy for Adapting to Climate Change: Use of green infrastructure on flood mitigation. Seoul: Korea Environment Institute. 16. Kim Hyunwoo and Li, Ming-Han Sustainable stormwater management: Examining the role planning capacity in mitigating peak surface runoff. Sustainability 8, no.9: Kim Jungjin, Kim Taedong, Choi Donghyuk and Jeon Jihong Design of structural BMPs for Low Impact Development (LID) application and modeling its effect on reduction of runoff and nonpoint source pollution: Application of LIDMOD2. Journal of Korean Society on Water Quality 27, no.5: Kim Sungwoo and Park Kyungmoon Improvement Plan on the Environmental Competitiveness for Sustainable Water Circulation in Incheon. Incheon: Incheon Institute. 19. Korea Land and Housing Institute Environmental Technology Development Project Annual Report. Daejeon: Korea Land and Housung Institute. search?q=cache:_sdqih6a10j:lhi.lh.or.kr/common/include/ file_download.asp%3fcategory%3dmemberdata%26idx%3 D341+&cd=1&hl=en&ct=clnk&gl=kr (accessed March 11, 2019). 20. KMA Open Weather Data Portal. Precipitation Analysis. (accessed January 13, 2019). 21. Lee Jeongho Cause and countermeasure of inundation damage. The Magazine of the Korean Society of Hazard Mitigation 13, no.1: Lee, K., Kim, H., Pak, G., Jang, S., Kim, L., Yoo, C. and Yun, Z. et al Cost-effectiveness analysis of stormwater best management practices (BMPs) in urban watersheds. Desalination and Water Treatment 19: Lee Sangeun, Kim Changhun, Park Taesun, Kim Mieun, Kim Seulye, Lee Taesam and Kim Jinhyun Development of the Urban Flooding Risk Prevention System(Ⅰ). Anyang: Korea Research Institute for Human Settlements. 24. Li, J., Deng, C., Li, Y., Li, Y. and Song, J Comprehensive benefit evaluation system for low-impact development of urban stormwater management measures. Water Resources Management 31, no.15: Liu, Y., Bralts, V. F. and Engel, B. A Evaluating the effectiveness of management practices on hydrology and water quality at watershed scale with a rainfall-runoff model. Science of the Total Environment 511: Luan, Q., Fu, X., Song, C., Wang, H., Liu, J. and Wang, Y Runoff effect evaluation of LID through SWMM in typical mountainous, low-lying urban areas: A case study in China. Water 9, no.6: Martin-Mikle, C., Beurs, K., Julian, J. and Mayer, P Identifying priority sites for low impact development in a mixed-used watershed. Landscape and Urban Planning 140: Ministry of Environment(ME) Application Manual of Low Impact Development (LID) Method for Environmental Impact Assessment. Seoul: Ministry of Environment. Analysis of Stormwater Runoff Reduction Effect Through Applying Low Impact Development Practices in a Flood Prone Area 65

68 Subsidy Compilation and Execution Management Guideline for Sewer System. Sejong: Ministry of Environment. /web/index.do?menuid=65 (accessed January 14, 2019). 30. Nam Chilho, Kim Mieun, Kim Jaemoon, Kim Byungsung and Shin Hyunsuk A study on urban stormwater reduction effectiveness based on SUSTAIN and SWMM model. In Proceedings of Korean Society of Civil Engineers 2017 Convention. October 18-20, Busan: Bexco. 31. National Institute of Environmental Research Individual communication. raw data. 32. Park Junho, Yoo Younggu, Park Youngkon, Yoon Heetaek, Kim Jonggun, Park Younshik and Jeon Jihong et al Analysis of runoff reduction with LID adoption using the SWMM. Journal of Korean Society on Water Quality 24, no.6: Prince George s County Low-impact Development Design Strategies: An integrated design approach; Department of environmental resources. Prince George s County, MD: Programs and Planning Division. 34. Roehr, D. and Kong, Y Runoff reduction effects of green roofs in Vancouver, BC, Kelowna, BC, and Shanghai, P.R. China. Canadian Water Resources Journal 35, no.1: Rosa, D., Clausen, J. and Dietz, M Calibration and verification of SWMM for low impact development. Journal of the American Water Resources Association 51, no.3: Rossman, L. A Storm Water Management Model User's Manual Version 5.0. Cincinnati, OH: U.S. Environmental Protection Agency. 37. Suh Joohwan and Lee Inkyu The water circulation improvement of apartment complex by applying LID technologies: Focused on the application of infiltration facilities. Journal of the Korean Institute of Landscape Architecture 41, no.5: Tetra Tech, Inc Mecklenburg county site evaluation tool-user s manual and guidance. Pasadena, CA: Tetra Tech, Inc. /SET2005/Meck%20Co% 20SET%20%20-%20Model%20 Documentation%20-%20draft% %20v7.pdf (accessed December 12, 2018). 39. Toronto and Region Conservation Authority and University of Toronto LID practices costing tool. technologies.ca/home/urban-runoff-green-infrastructure/low-i mpact-development/low-impact-development-life-cycle-costs / (accessed January 10, 2019). 40. United States Environmental Protection Agency(USEPA) Urban runoff: Low impact development. nps/urban-runoff-low-impact-development (accessed January 14, 2019). 41. United States Department of Housing and Urban Development (USHUD) The Practice of Low Impact Development. Upper Marlboro, MD: NAHB Research Center Inc. gov/portal/publications/practlow ImpctDevel.pdf (accessed January 14, 2019). 42. University of Arkansas Community Design Center(UACDC) Low Impact Development: A design for urban areas. Fayetteville, USA: University of Arkansas Press. 43. Woo Wonhee, Ryu Jichul, Moon Jongpill, Jang Chunhwa, Kum Donghyuk, Kang Hyunwoo and Kim Kisung et al Development of coupled SWAT-SWMM to evaluation effects of LID on flow reduction in Complex landuse. Journal Korean Society on Water Environment 28, no.4: 논문접수일 : 심사시작일 : 심사완료일 : 국토연구제 102 권 ( )

69 요약 주제어 : 우수유출수, 저영향개발, 그린인프라, LIDMOD3, 홍수, 인천 불투수층의증가및기후변화로인한빈번한집중강우의발생은도시내홍수피해를지속적으로증가시키고있다. 전통적인배수시스템은우수유출수를조절하는데효과적이지만, 우수관리에만초점을두기에건설비용과성능적인측면을고려했을때비효율적이라할수있다. 저영향개발 (LID) 기법은다양한혜택을제공하는종합적인물관리기법으로, 지속가능한우수관리실천을위해현재많은국가에서이를도입하고있다. 본논문은 LIDMOD3 모델링툴을활용하여 LID기법의수문학적성능을세가지시나리오에모의적용하여분석하였다. 또한, LID기법의비용효과를하수도기반시설설치비용과의비교를통해분석하였다. 연구대상지는인천시부평구의상습 침수구역인십정동일대이다. 연구결과, 대상지의유출량은 LID를설치하지않았을경우와비교하여약 14~36% 감소하였고, 첨두유출량은약 33~66% 줄어들었다. 비점오염물질의경우각시나리오에따라약 15~36% 감소한것을발견하였다. 건설비용을고려했을때, 전지역에 LID를설치하는것보다는일부지역에부분적으로설치하면서하수도기반시설을함께융합하여개발하는것이효과적이라고나타났다. 본논문은정부기관및지자체가지속가능한우수관리계획의일환이자향후도시재생사업과도연계한체계적인물관리방안으로서더욱효과적인 LID 도입을모색하기위한정책대안을제공한다. Analysis of Stormwater Runoff Reduction Effect Through Applying Low Impact Development Practices in a Flood Prone Area 67

70

71 국토연구제 102 권 (2019.9): pp69~82 통신빅데이터를활용한국가하천친수지구의이용등급, 상세유형화및친수거점지구지정방법개발 * A Study on Classification and Characterization of Water-Friendly Space for the Smart River Space Management Using the Mobile Big Data 이종소 Lee Jongso **, 이상은 Lee Sangeun *** Abstract For national and local policies related to urban development and regeneration, welfare, culture and tourism, it is very important for river managers to designate and operate water-friendly river spaces in accordance with residents demands. Using mobile big data verified by the previous study (Lee, Lee and Choi 2019), this study aims to propose scientific ways to characterize water-friendly river spaces. First, it suggests clear criteria regarding how frequently a river space is visited by residents: river spaces are classified to five grades over the log-normal distribution in the number of visits per area. Second, it infers residents purposes to visit a river space: the purposes are characterized to five groups, which largely depends on whether a river space is used as a place for neighborhood-convenience facilities, or as one for culture and tourist facilities for regional revitalization. Third, this study suggests a step-wise method for determining water-friendly hub spaces deliberately using all results. This study concludes that river managers should apply more scientific ways to designate and operate water-friendly river spaces while monitoring residents visits closely with mobile big data. Keywords: Water-Friendly River Space, River Space Characterization, Mobile Big Data, River Management I. 서론 2007년 하천법 개정이후부터하천기본계획을수립할때주민들이하천공간을적정하게활용할수있도록친수지구를지정해오고있다. 관리청이친수지 구를지정하는행위는지역의건전한하천이용을지원하기위한행정적인근거를확보한다는의미를지닌다. 우선, 하천관리청은하천공사를통해광장, 산책로, 자전거도로, 경관시설등하천이용을위한기반을확충하며, 이러한친수시설의기능을유지하는데 * 본연구는 2018년에서 2020년까지국토연구원에서진행중인국토교통부의 통신빅데이터를활용한국가하천관리효율성제고방안연구 로부터연구비와기초자료를지원받았으며, 본논문의작성및투고에대해서는연구의감독기관과사전에협의되었음. ** 국토연구원책임연구원 ( 제1저자 ) Assistant Research Fellow, Korea Research Institute for Human Settlements Primary Author jslee@krihs.re.kr *** 국토연구원수자원 하천연구센터장 ( 교신저자 ) Director of Water Resources and River Research Center, Korea Research Institute for Human Settlements Corresponding Author selee@krihs.re.kr 통신빅데이터를활용한국가하천친수지구의이용등급, 상세유형화및친수거점지구지정방법개발 69

72 필요한유지 보수비의일부를관할지자체에게지원하고있다. 친수지구관리에는주민들의하천이용이중요한요소라할수있으므로관리청은지역주민들이요구하는이용객편의시설, 레저 문화시설등의설치규제도하천법령과점용허가세부기준의범위내에서완화하게된다 ( 국토교통부 2018). 문제는하천기본계획수립시지역주민의공간활용방식에대한객관적이고면밀한수요조사방법이마련되어있지않다는점이다. 이러한가운데 2014년부터는근린친수와친수거점으로테마를부여하는등으로특성화를시도하기때문에여러가지부작용이발견된다 ( 국토교통부 2015; 2016). 국가하천정비과정에서대규모로지정하여공원화한뒤이용객이저조하여친수지구의효용에대한사회적비판을받는다든지, 정비후그리오랜시간이소요되지않았음에도지역 주민요청에의해공간내시설을변경하는일이빈번하게발생하고있다. 친수거점지구를지정한뒤랜드마크조성을위해많은예산이투입되었으나실제로는배후지역주민위주의낮은이용도를보이는경우도있다. 하천기본계획을자주지정하거나변경할수없으므로사전수요예측을기초로친수지구를지정할수밖에없겠지만, 초기지정된친수지구는이른바 예정지 로간주하는것이바람직할것이다. 실제로하천을대규모로인공화하거나근린친수나친수거점으로특성화하는것은지역주민의하천이용방식, 즉, 이용도와공간활용의유형을면밀히관찰하면서조심스럽게판단하는것이필요하다. 수요예측의실패는주민복지와동떨어진방식으로하천을조성 관리하게되므로불필요한시설설치와부적절한규제완화로하천난개발등의부작용뿐만아니라, 장기적으로유지관리예산의효율성저하를초래하기때문이다. 본논문에서는이러한문제점들을인식하고통신빅데이터를하천공간에활용하여그해결책을강구하 고자하였다. 통신빅데이터는다양한분야에서활용되고있는데최근에들어맞춤형행정과시민의수요에따라공원의관리 ( 오장근, 허학영, 심규원, 김태근외 2017) 나교통의수요관리 ( 성지은, 박기량 2014; 김혜주 2017; 이광섭, 엄진기, 성명언, 유소영외 2017; AirSage 2017) 분야에서활발히이용되고있다. 또한지역의인구와인구밀도추정에관련된연구를살펴보면김종학, 고용석, 김준기, 김동한 (2014) 은통신빅데이터를활용하여인구의활동을시공간적의변화에따라분포시켜공간정책활용방안을제시하였으며, De Jonge, van Pelt and Roos(2012) 는통신빅데이터로주간인구밀도와통근자이동성을추정하여공식통계화의가능성을탐색하였다. Deville, Linard, Martin and Gilbert et al.(2014) 는위치데이터를셀단위로하여회귀분석을실시한뒤인구밀도를추정하는연구를수행하였으며, Douglas, Meyer, Ram and Rideout et al.(2015) 는독립변수를이동전화통화량과토지피복데이터로하여인구추정회귀분석을실시하였다. 이와같이통신빅데이터는다양한분야에서적용범위가증가하고있다. 본논문은친수지구지정후하천공간을체계적으로특성화하는방법에초점을두고다음의세가지목적을가지고수행되었다. 첫째, 친수지구의이용도를판단하기위해이용등급을부여하는기준을제시한다. 이용등급은상대적인이용객수에대한것으로친수지구운영의효용과직결되기때문에유지 보수비예산지원의차등화등에활용될수있을것이다. 둘째, 해당친수지구가지역주민의근거리에서도시공원이나생활체육시설로활용되는지, 아니면문화 관광시설과같이지역활성화를위한장소로활용되는지에대한유형을구분한다. 이러한유형구분은기반시설을조성하거나규제완화를검토할때장소의특수성에대한판단의기준을제공할수있을것이다. 70 국토연구제 102 권 ( )

73 셋째, 친수거점지구를지정하기위한체계적인방법을제안하고자한다. 친수거점지구는광역권의명소로레저, 문화등이용수요가풍부하고시설물의집중설치가필요한지역이므로하천의고유가치를훼손하지않는범위내에서최소한으로지정해야한다. 따라서친수거점지구를지정하는취지상이용객이상대적으로많고지역활성화를위해필요한곳인지를반드시확인하는것이바람직하다. 이에본논문에서는친수지구이용도조사를통해이용등급과공간활용의유형을고려하여단계적으로지정하는방법과기준을제안하고자한다. 본논문은이종소, 이상은, 최진영 (2019) 의후속연구에해당된다. 선행연구에서는국가하천에위치한 297개친수지구의이용도조사를위해통신빅데이터자료의유용성을검증하였으며, 통신빅데이터로부터추정가능한정보를토대로다양한이용지표를개발한바있다. 본논문은선행연구에서제시한이용지표를자료원으로활용하여하천관리정책과실무차원에서지역의하천이용수요를잘이해하고, 이로인해친수지구의계획, 유지관리, 인 허가에있어서합리적인결정을하는데에크게기여할수있을것으로예상된다. II. 방법론본논문에서는친수지구이용지표를활용하여친수지구이용등급및상세유형화를실시하고친수거점지구를지정하는방법을 <Figure 1> 과같이제시하고자한다. 1. 통신빅데이터를활용한친수지구이용지표산정이종소, 이상은, 최진영 (2019) 은국가하천 297개친수지구를대상으로통신빅데이터의활용성을검토하고이용지표를산정하여친수지구를이용하는정도와방식을객관적으로설명한바있다. 현장조사를통해직접계측한친수지구이용객수와통신빅데이터로산정한친수지구이용객수를동일날짜, 동일시간대에대해비교하여활용성을검토하였다. 이어서, <Table 1> 과같이친수지구의이용도조사를위해면적당이용객수, 하천이용첨두율, 월최대이용시기, 주중이용률, 배후지역이용빈도, 이용객의성비, 이용객의우점연령층, 1인당이동거리등 8가지이용지표를정의하고 2017년도기간에대해시범적으로지표값을산정하였다. 본논문에서는이종소, 이상은, 최진영 1. Estimation of Water-Friendly Space Indicators - Extraction of Population Data in Water-Friendly Space - Field Survey Verification - Definition and Estimation of Water-Friendly Space Indicators Figure 1 _ Flow Chart 2. Estimation of Water-Friendly Space Grade - Normality Test - Outlier Test - Criteria for Grade of Water-Friendly Space 3. Classification of Water-Friendly Space - Clustering Analysis -t-test - Estimation of Discriminant by Logistic Regression Analysis 4. Designation of Water-Friendly Space - HighWater-Friendly Space Grade - Characterization of Culture, Leisure, Tourism etc. 통신빅데이터를활용한국가하천친수지구의이용등급, 상세유형화및친수거점지구지정방법개발 71

74 Table 1 _ Definition of Water-Friendly Space Indicators Source Characteristics Indicator Definition Use in This Study Number of Visitors per Area(people/m 2 ) Number of Visitors per Water-friendly Space Area(people/m 2 ) Peak Visit Rate(%) Maximum Number of Visitors/Minimum Number of Visitors Spatial Use Maximum Use Month(-)* Month with the Most Visitors Weekly Visit Rate(%) Weekly Visitors/Weekend Visitors Residents Visit Frequency(times) Average Number of Visits of Residents Travel Distance per Visitors(km) Travel Distance per Visitors Visitors Gender Ratio(%) Ratio of Male/Female Visitors Maximum Age Group(-)* 10s or Less, 20s, 30s, 40s, 50s, 60s or More Source: Lee, Lee and Choi (2019) 에서제안한다양한이용지표가운데이용등급산정과친수거점지구판별에불필요한월최대이용시기, 이용객의우점연령층, 이용객의성비등을제외하고면적당이용객수, 하천이용첨두율, 주중이용률, 1인당이동거리등의네가지지표를자료원으로활용하였다. 2. 친수지구이용등급산정 1) 정규성검정 용하였다. Kolmogorov-Smirnov 검정은누적분포함수 (Cumulative Distribute Function) 의개념을이용하며귀무가설 는 표본의분포는정규성을만족한다, 대립가설 는 표본의분포는정규성을만족하지않는다 로정의된다. 여기서, 는가정된이론확률분포의누가확률분포이며, 는표본자료의누가확률분포이다. 통상검정결과가유의수준 5% 에서 p-value 가 0.05 이상이되면귀무가설을기각할수없기때문에정규성을만족한다고할수있다. 친수지구이용도를등급화하기위해전체이용객수, 면적당이용객수, 전체이용객수와면적당이용객수의조합등여러가지방법이존재한다. 이가운데, 이용객의방문이활발한정도를상대비교하기위해친수지구면적당이용객수를토대로등급화를시도하기로하였다. 2017년도 297개친수지구의면적당이용객수를히스토그램으로나타낸결과로그노멀 (lognormal) 확률밀도함수와유사한형태를관찰할수있었으므로로그노멀확률밀도함수에대한정규성을만족하는지를확인해보고자하였다. 먼저자료의정규성을검정하기위해표본수가많을때적합한검정인 Kolmogorov-Smirnov 검정을적 2) 이상치검정및친수지구이용등급기준산정자료의정규성검정과함께이상치검정을실시하였다. 자료의이상치 (Outlier) 는어떤집단으로부터추출한무작위표본에서다른값들에비해비정상적으로크거나작은값을말하며, 이상치가존재하면계산된통계량에편중된결과를초래하게된다. 본논문에서는 Tukey(1977) 가제안한이상치판정방식가운데제 1사분위수 ( ), 제3사분위수 ( ), 을기준으로 ± 를적용해면적당이용객수에대한이상치의상 하한을판정하였다. 다음으로는친수지구의이용등급기준을산정하기 72 국토연구제 102 권 ( )

75 위해면적당이용객수를기준으로다섯가지등급으로구분하고자하였다. 로그로변환된면적당이용객수의확률밀도함수를동일한면적의 5개의구간, 즉, 이용도가매우활발한 I등급부터이용도가매우저조한 V등급으로이용등급기준을구분하고다시면적당이용객수로환원하여등급별기준을제시하였다. 3. 친수지구상세유형화를위한판별식도출 1) 군집분석본논문에서는성격이유사한친수지구들을특정유형으로묶기위해군집분석을실시하고자하였다. 군집분석은데이터가어떤군집에속해있는지알수없는상태에서거리값을이용해가까운거리에있는것들끼리묶어분류하는비지도학습 (Unsupervised Learning) 종류중하나로기본원리는군집내응집도를최대화하고군집간분리도를최대화하여유사한속성을지닌대상끼리묶어분류하는것이다. 군집분석은크게계층적군집 (Hierarchical Clustering) 과분할적군집 (Partitional Clustering) 으로나뉘는데, 본논문에서는계층적군집의종류가운데 Ward 연결법을적용하고자하였다. Ward 연결법은군집내의오차제곱합 (Error Sum of Square) 에기초하여군집을수행하며, 병합된군집의오차제곱합은이전각군집의오차제곱합보다커지게되는데, 그증가량을최소로하는방향으로군집을형성해나가는방법으로다른방법에비해노이즈나이상치에덜민감한장점이있다 (Ward 1963). 자료를몇개의군집으로분류할것인가는군집분석에서매우중요한요소가운데하나이다. Charrad, Ghazzali, Boiteau and Niknafs(2014) 는최적의군집개수를찾아내기위해 Software R의패키지인 NbClust를개발하였다. NbClust는 30개의최적군집의수결정 방법과군집의수, 거리측정값, 군집방법등의모든조합결과를통해최적의군집개수를제안하며, 이에본논문에서는 Charrad, Ghazzali, Boiteau and Niknafs(2014) 가개발한 NbClust 를활용하여최적의군집개수를찾아내고자하였다. 군집분석을통해친수지구의유형을판단하는데는 <Table 1> 의 8가지이용지표가운데월단위첨두율, 주중이용률, 1인당이동거리총세가지를적용하였다. 또한전체이용객이매우적거나월평균이용객이 0~1명에불과한친수지구는이용객 1인에대한지표영향이지나치게커이상치를야기하므로해당친수지구를제외하고군집분석을실시하였다. 2) 독립표본 t 검정 군집분석을통해분류된유형간의특성을통계적으로분석하고특징을파악하기위해세가지지표각각에대해독립된두모집단평균차이검증에사용되는 t검정을실시하고자하였다. 특정유형과타유형간에이용지표에있어서차이가있을것이라는연구가설하에귀무가설 와대립가설 을설정하였다. 여기서 는 I번째 유형의 번째지표의평균값을의미하며, 는타유형들의 번째지표의평균값을의미한다. 또한군집분석결과에따라 는유형의개수를 는군집분석에사용된월단위첨두율, 주중이용률, 그리고 1인당이동거리등세가지이용지표를각각의미한다. 통상유의수준 5% 에서 p-value 가 0.05 이상이되면귀무가설을기각할수없게되므로차이가있다고할수없다. 3) 로지스틱회귀분석 로지스틱회귀분석 (Logistic Regression) 은회귀분석의 통신빅데이터를활용한국가하천친수지구의이용등급, 상세유형화및친수거점지구지정방법개발 73

76 한종류로서, 종속변수가이분화된범주형변수일경우주로사용된다. 종속변수를확률변수인로짓 (Logit) 으로변환하여선형적인다중회귀모형을구성한뒤모형의회귀계수등을산정하여종속변수와독립변수들간의관계를추정할수있다. 또한로지스틱회귀분석은다변량정상분포가정을요구하지않으므로이분화된범주형예측변수의사용이쉽다는장점이있기때문에본연구에적합하다. 로지스틱회귀방정식의회귀계수를추정할때에는, 최소자승법 (Least Square Estimation) 이아닌최대우도법 (Maximum Likelihood Method) 을활용한다. 이는준거변수가이분변수 (0과 1만존재 ) 일경우이분변수의특성상이분산의가능성이매우높기때문에최소자승법의가정인오차의동분산성을만족시키기어렵기때문이다. 또한준거변수값의빈도가 0일경우로짓의값이정의될수없어음의값이예측되거나부호자체가달라지는등추정값이불안정해지기때문이다 ( 박광배 1999; 2000; Mertler and Vannatta 2001). 개의설명변수 ( ) 에대하여 1을가질확률을 라고표시할때다음 < 식 1> 과같이계산된다. exp exp < 식 1> 여기서 는절편을의미하고, 들은회귀계수들을의미한다. 이때 는최대우도법에의해추정된다. 로지스틱회귀분석은정규성, 선형성, 등분산성의가정을필요로하지않으며, 연속형및범주형설명변수가모두포함된모형으로분석할수있다 ( 계묘진 2012). III. 결과및검토 1. 친수지구이용등급산정 lognormal 확률밀도함수로변환된면적당이용객수의자료는다음 <Figure 2> 의 (a) 히스토그램과밀도함수그리고 <Figure 2> 의 (b) Q-Q plot과같이정규성을보이고있음을확인하였다. 또한 Kolmogorov-Smirnov 검정결과, 양측검정에대한 p-value 는 0.38로유의수준 5% 에서귀무가설인 를기각할수없어정규성을만족한다고할수있다. 이상치검정을위해기초통계량을분석한결과, 1분위수 (Q 1 ) =-2.60, 중간값 (Median)=-0.86, 평균값 =-0.93, 3분위수 (Q 3 )=0.79로분석되었으며, 이상치의상 하한은 Q IQR=-7.69, Q IQR=5.87 로 <Figure 2> 의 (c) Box Plot과같이이상치가없는것으로나타났다. 친수지구의이용등급기준을산정하기위해검정 Figure 2 _ Statistical Chart of Lognormal Probability Density Function(people/m 2 ) (a) Histogram and Density Plot (b) Q-Q Plot (c) Box Plot 74 국토연구제 102 권 ( )

77 을마친로그로변환한면적당이용객수를동일한면적의 5개의구간으로이용등급기준을구분하고다시면적당이용객수로환원하였다. <Table 2> 와같 Table 2 _ Criteria for Grade of Water-friendly Space Grade Grade Criteria Ⅰ(Very High) (people/m 2 ) > 2.928/m 2 Ⅱ(High) 0.950/m 2 < (people/m 2 ) < 2.928/m 2 Ⅲ(Normal) 0.207/m 2 < (people/m 2 ) 0.950/m 2 Ⅳ(Low) 0.033/m 2 < (people/m 2 ) 0.207/m 2 Ⅴ(Very Low) (people/m 2 ) 0.033/m 2 이 2.928명 /m 2, 0.950명 /m 명 /m 2, 0.033명 /m 2 의분류기준을얻었으며, 등급별로약 59개의친수지구를포함하는것으로분석되었다. 수계별로볼때영산강수계의경우, 46개의친수지구가운데 15개 (32.6%) 지구가이용도가매우저조한 Ⅴ등급구간으로분류되었으며, 섬진강수계역시 5개 (29.4%) 의지구가 Ⅴ등급구간에분류된것을확인할수있었다 (<Table 3> 참조 ). 전체적으로낙동강중류와금호강, 한강합류점 (<Figure 3> 참조 ), 일부도시구간에있는친수지구들이이용도가매우활발한 Table 3 _ Grade of Water-friendly Space by River Basin Basin Grade Ⅰ Grade Ⅱ Grade Ⅲ Grade Ⅳ Grade Ⅴ Total Han River 16(22.9%) 14(20.0%) 12(17.1%) 13(18.6%) 15(21.4%) 70(100%) Nakdong River 22(23.4%) 13(13.8%) 23(24.5%) 20(21.3%) 16(17.0%) 94(100%) Geum River 11(15.7%) 22(31.4%) 13(18.6%) 15(21.4%) 9(12.9%) 70(100%) Yeongsan River 8(17.4%) 7(15.2%) 10(21.7%) 6(13.0%) 15(32.6%) 46(100%) Seomjin River 2(11.8%) 3(17.6%) 2(11.8%) 5(29.4%) 5(29.4%) 17(100%) Total Figure 3 _ Water-friendly Space Grade: Han River Basin Example 통신빅데이터를활용한국가하천친수지구의이용등급, 상세유형화및친수거점지구지정방법개발 75

78 반면, 그외많은구간에서는이용도가저조한경향을보이는것으로분석되었다. 2. 친수지구상세유형화를위한판별식도출 1) 친수지구상세유형화군집분석을위해표준화된세가지지표는 <Figure 4> 와같이서로간의상관성은매우낮은것으로분석되었다. 친수지구의유형을분류하기위해 Ward 연결법으로군집을분석하고 NbClust 를활용하여최적의군집개수를탐색한결과, 최적군집의개수는 <Figure 5> 와같이 5개로분석되었으며, 273개소친수지구 ( 세가지지표의이상치를보인친수지구 24개소제외 ) 를 5개의유형으로군집분석한결과는 <Figure 6> 과같이유형I, II, III, IV, V가각각 59개소 (21.6%), 78개소 (28.6%), 81개소 (29.7%), 41개소 (15%), 14개소 (5.1%) 인것으로나타났다. 군집분석을통해분류된유형간의특성을분석하기위해독립된두모집단평균차이검증에사용되는 t검정을실시한결과는 <Table 4~6> 과같다. 이를해석하면, 유형I은타유형에비해월단위첨두율이약간낮고주중이용률은높으며 1인당이동거리는매우짧은수준인것으로나타났다. 엑스포수상공원, 유등체육공원등과같이근거리에서주중에방문하는비중이높으며연중이용추이는완만한특징이관찰되었다. 유형II는타유형에비해월단위첨두율이매우낮고주중이용률은매우높은수준이며 1인당이동거리는높은수준인것으로나타났다. 주로운동장시설로활용되는친수지구가다수분포하였으며비교적원거리에서주중에방문하는비중이높고연중이용추이는완만한특징이관찰되었다. 유형III은타유형에비해월단위첨두율이낮고주중이용률도낮으며 1인당이동거리는짧은것으로나타났다. 즉, 비교적근거리에서주말에방문하는비중이높고연중이용추이는완만한특징이관찰되었다. 유형IV는타유형에비해월단위첨두율이높고주중이용률은낮으며 1인당이동거리가매우긴것으로나타났다. 두물지구, 평사리공원, 송호국민관광지등과같이주로원거리에서주말에집중적으로방문하며연중이용추이 Figure 4 _ Correlation Coefficients between Variables. Figure 5 _ Optimization of Number of Clusters 76 국토연구제 102 권 ( )

79 Figure 6 _ Result of Clustering Analysis: Location of Principal Components U and V Table 4 _ Result of Independent Sample t-test: Peak Visit Rate Test Object Independent Sample t-test 95% Confidence Interval Equal Variance t-value p-value(two Side) Mean Difference Lower Upper Assumption(Levene) Type I Others Type II Others Type III Others Type IV Others Type V Others Table 5 _ Result of Independent Sample t-test: Weekly Visit Rate Test Object Independent Sample t-test 95% Confidence Interval Equal Variance t-value p-value(two Side) Mean Difference Lower Upper Assumption(Levene) Type I Others Type II Others Type III Others Type IV Others Type V Others Table 6 _ Result of Independent Sample t-test: Travel Distance per Visitors Test Object Independent Sample t-test 95% Confidence Interval Equal Variance t-value p-value(two Side) Mean Difference Lower Upper Assumption(Levene) Type I Others Type II Others Type III Others Type IV Others Type V Others 통신빅데이터를활용한국가하천친수지구의이용등급, 상세유형화및친수거점지구지정방법개발 77

80 에서성수기의특징이관찰되었다. 마지막으로유형V 는타유형에비해월단위첨두율이매우높고주중이용율은낮으나 1인당이동거리는타유형과의차이가특별히두드러지지않는것으로나타났다. 이포보오토캠핑장, 대성1 친수공원등과같이근거리와원거리에서주말에방문하는비중이매우높고연중이용추이에서성수기의특징을보이는것으로관찰되었다. 이용지표에대한유형의통계적특징과유형에속한친수지구의성격을전체적으로종합해보면유형I, II, III은전반적으로도시공원또는체육시설의성격을지니고있어근린친수지구로간주가가능하며, 유형 IV, V는전반적으로주말을이용한문화 레저 생태교육시설또는지역행사장소로서의성격을지니고있어친수거점지구로간주할수있는것으로판단된다. 2) 로지스틱회귀분석을통한판별식도출친수지구의특성을분류하기위해서매번자료의표준화와군집분석을실시하는것은복잡하고매우비효율적이라할수있다. 이에본논문에서는군집분석을통해분류된근린친수지구의성격을갖는유형I, II, III과친수거점지구의성격을갖는유형 IV, V를이분화하기위해로지스틱회귀분석을실시하였다. 유형I, II, III을 0, 유형IV, V를 1 로이분화하여 종속변수를설정하였으며, 독립변수로는군집분석에서사용된월단위첨두율, 주중이용률, 1인당이동거리의원자료를사용하였다. 개발된로지스틱회귀식은 <Table 7> 과 < 식 2> 와같으며, 이면유형I, II, III, 즉, 근린친수지구에속하고 이면유형IV, V, 즉, 친수거점지구에속하는것으로해석할수있다. 위식의정확도는유형IV, V의 55개가운데 46개를구분해내었고유형I, II, III의 218개가운데 214개를구분해내 (46+214)/273=0.952 수준으로나타났다. 분류해내지못한친수지구의경우는대부분군집간경계부근에있어근린친수지구와친수거점지구의특징을명확히정의하기힘든것으로나타났다. 친수거점지구성격이있음에도근린친수로간주하게되는경우가다소발견된것이사실이다. 그러나친수거점지구를보수적으로판별하는것은난개발을방지하고예산을효율적으로활용하기위해친수거점지구를최소한으로지정해야한다는점을감안할때정확도에문제가있다고보기는힘들다. 3. 친수거점지구지정을위한단계별판단기준제안친수거점지구는하천의고유가치를훼손하지않는범위내에서최소한으로지정해야하며, 광역권의명 exp exp < 식 2> Table 7 _ Result of Discriminant by Logistic Regression Analysis Variable Estimate Std.Error z value Pr(> z ) Intercept Peak Visit Rate( ) Weekly Visit Rate( ) Travel Distance per Visitors( ) 국토연구제 102 권 ( )

81 Figure 7 _ Decision Tree for Classification of Water-friendly Space 소로레저 문화등의성격을갖고이용수요가풍부해야한다. 그러기위해서는현재처럼친수거점지구를객관적이고명확한기준없이먼저지정하기보다는친수지구를조성한뒤중장기적인모니터링을실시하여친수거점지구인지를세분화해야할것이다. 중장기적인모니터링실시후본논문의결과를활용하여로지스틱회귀분석을통해유형IV, V에속하는지를 < 식 2> 을통해판단한뒤, <Table 2> 에따라해당친수지구의이용등급이 I 또는 II 등급에속하게되면주말을이용한문화 레저 생태교육시설또는지역행사장소로서의성격을지니고있다고판단할수있을것이다. 즉로지스틱판별식의결과가 0.5 이상이고면적당이용객수가 0.950명 /m 2 이상이된다면향후친수거점지구지정을고려해볼수있을것이다. IⅤ. 결론본논문은친수지구를지정한이후하천공간을특성화하는방법에초점을두고이용도를판단하기위한이용등급산정, 이용의특성을분류하기위한상세유형화그리고친수거점지구지정을위한체계적인방 법을제안하고자하였으며, 연구결과는다음과같이요약할수있다. 첫째, 친수지구의이용등급을산정하기위해면적당이용객수의지표를 lognormal 확률밀도함수로변환하고정규성과이상치검정을실시하였다. 그결과등급별로약 59개의친수지구를포함하는것으로분석되었는데영산강과섬진강수계에이용등급이저조한친수지구가많았으며, 전체적으로낙동강중류와금호강, 한강합류점및일부도시구간에있는친수지구들의이용이활발한것으로분석되었다. 둘째, 친수지구상세유형화를위해군집분석을실시하여총 5개의유형으로분류하였으며, 통계검정을통해각유형간의특징을분석하였다. 또한향후친수지구의성격을분류하기위해매번표준화및군집분석을실시하는것은복잡하고비효율적일수있기때문에실무에서간편하게활용될수있도록로지스틱회귀분석을통해근린친수지구와친수거점지구의유형을이분화하였다. 유형I, II, III은전반적으로도시공원또는체육시설의성격을지니고있어근린친수지구로간주가가능한것으로나타났으며, 유형IV, V는전반적으로주말을이용한문화 레저 생태교육 통신빅데이터를활용한국가하천친수지구의이용등급, 상세유형화및친수거점지구지정방법개발 79

82 시설또는지역행사장소로서의성격을지니고있어친수거점지구로간주가가능한것으로나타났다. 로지스틱회귀분석을통해근린친수지구와친수거점지구를분류할때 95.2% 수준의정확도높은판별식을도출할수있었으며, 이판별식은다소보수적으로친수거점지구를분류한다는점을확인할수있었다. 셋째, 본논문에서수행한연구결과들을종합하여단계적으로친수지구를지정하는방법과기준을제안하였다. 친수거점지구는이용객이상대적으로많고광역권의명소로레저 문화등의성격을지니고있기때문에상세유형결과의유형IV와 V에속하며, 이용등급이 I, II등급에속하게된다면친수거점지구의지정을고려해볼수있을것이다. 중장기적인모니터링실시후본논문의결과를활용하여로지스틱판별식의결과가 0.5 이상으로유형IV, V에속하는지는판단한뒤, 면적당이용객수가 0.950명 /m 2 이상이되어이용등급이 I 또는 II 등급에속하게되면주말을이용한문화 레저 생태교육시설또는지역행사장소로서의성격을지니고있다고판단할수있을것이다. 2007년 하천법 개정이후지역주민의공간활용방식에대한객관적이고면밀한수요조사방법이정립되어있지않은상태에서국가하천정비시대규모로친수거점지구로지정하여공원화하였다. 그결과, 몇몇지구에대해서는이용객저조에따른친수지구의효용에대한사회적비판, 하천난개발의부작용, 유지관리예산의효율성저하등의부작용을초래해왔다. 이에근린친수지구나친수거점지구로특성화하는것은지역주민의하천이용방식과공간활용의유형을면밀히관찰하면서조심스럽게판단하는것이필요할것으로생각되며, 먼저세분화하여지정하기보다는친수지구를지정한이후중장기적인모니터링과객관적인분석을통해근린친수지구와친수거점지구로지정해야할것이다. 본논문의이용등급결과를통해친수지구의효율적인운영과유지 보수비예산지원차등화등에활용될수있을것이며, 상세유형화및친수거점지구지정방법은친수지구계획, 기반시설조성, 인 허가에있어합리적인결정을하는데크게기여할수있을것으로판단된다. 참고문헌 1. 계묘진 Lasso 를기반으로한로지스틱회귀모형연구. 석사학위논문, 계명대학교. Kye Myojin A Study on Logistic Regression Model based on Lasso. M.D. diss., Keimyung University. 2. 국토교통부 대강수변친수공간관리개선방안. 세종 : 국토교통부. Ministry of Land, Infrastructure and Transport(MOLIT) Improvement of the Management of Water-friendly Space in 4 Rivers. Sejong: MOLIT 하천유지관리평가및개선방안연구. 세종 : 국토교통부 Study on River Maintenance Evaluation and Improvement Plan. Sejong: MOLIT 하천점용제도개선및표준매뉴얼개발연구 : 세종 : 국토교통부 A Study on the Improvement of River Space Use Permission and Development of Standard Manual. Sejong: MOLIT. 5. 김종학, 고용석, 김준기, 김동한 스마트셀기반활동인구의공간정책활용방안연구. 세종 : 국토연구원. Kim Jonghak, Ko Yongseok, Kim Joonki and Kim Donghan The Application of Smart Cell in Space Policy. Sejong: Korea Research Institute for Human Settlements. 6. 김혜주 통신유동인구빅데이터로교통혁명시대를열다. 월간교통 229호, 세종 : 한국교통연구원. Kim Hyeju Opening the traffic revolution in communication big data. Monthly KOTI Magazine on Transport 229, Sejong: The Korea Transport Institute. 7. 박광배 변량분석과회귀분석. 서울 : 학지사. Park Kwangbae Analysis of Variance and Regression. 80 국토연구제 102 권 ( )

83 Seoul: Hakjisa 다변량분석. 서울 : 학지사 Multivariate Analysis. Seoul: Hakjisa 9. 성지은, 박기량 빅데이터를활용한정책사례분석과시사점. 과학기술정책 24권, 2호 : Seong Jieun and Park Kiryang Policy case analysis and implications using big data. Science and Technology Policy 24, no.2: 오장근, 허학영, 심규원, 김태근, 최진영 통신빅데이터를활용한국립공원탐방객실태조사및이용패턴분석검증연구. 원주 : 국립공원관리공단. Oh Jangeun, Heo Hakyoung, Sim Gyuwon, Kim Taegeun and Choi Jinyoung A Study on National Park Visitor Survey and Pattern Analysis Using Mobile Big Data. Wonju: Korea National Park Service. 11. 이광섭, 엄진기, 성명언, 유소영, 민재홍, 이준 통신량빅데이터활용성에관한연구 : 의정부시내 외부활동및이동패턴과도시철도망연계성을중심으로. 국토계획 52권, 5 호 : Lee Kwangsub, Eom Jinki, Seong Myeongeon, You Soyoung, Min Jaehong and Lee Jun Usability of mobile phone big data: Focusing on the activity and mobility patterns and urban railway network in the Uijeongbu City. Journal of Korea Planning Association 52, no.5: 이종소, 이상은, 최진영 국가하천친수지구공간관리를위한통신빅데이터활용성검토 : 자료검증과이용지표선정. 국토연구 101권 : Lee Jongso, Lee Sangeun and Choi Jinyoung Using the mobile big data for the smart river space management: Data validation and water-friendly space indicators The Korea Spatial Planning Review 101: AirSage Innovations in data collection: The power of where and when. tes/47/2017/06/airsagetnmug.pdf (accessed May 22, 2019). 14. Charrad, M., Ghazzali, N., Boiteau, V. and Niknafs, A NbClust: An R package for determining the relevant number of clusters in a data set. Journal of Statistical Software 61, no.6: De Jonge, E., van Pelt, M. and Roos, M Time patterns, geospatial clustering and mobility statistics based on mobile phone network data. In Proceedings of The Federal Committee on Statistical Methodology Research Conference. January 10-12, Washington D.C.: Washington Convention Center. 16. Deville, P., Linard, C., Martin, S., Gilbert, M., Stevens, F. R., Gaughan, A. E. and Blondela, V. D. et al Dynamic population mapping using mobile phone data. Proceedings of the National Academy of Sciences 111, no.45: Douglass, R. W., Meyer, D. A., Ram, M., Rideout, D. and Song, D High resolution population estimates from telecommunications data. EPJ Data Science 4, no.1: Mertler, C. A. and Vannatta, R. A Advanced and Multivariate Statistical Methods: Practical application and interpretation. Los Angeles: Pyrczak Publishing. 19. Tukey, J. W Exploratory Data Analysis. Boston: Addison-Wesley. 20. Ward Jr., J. H Hierarchical grouping to optimize an objective function. Journal of the American Statistical Association 58: 논문접수일 : 심사시작일 : 심사완료일 : 통신빅데이터를활용한국가하천친수지구의이용등급, 상세유형화및친수거점지구지정방법개발 81

84 요약 주제어 : 친수지구, 친수지구유형화, 모바일통신빅데이터, 하천관리 도시의개발, 재생, 복지, 문화관광등의지역정책과하천이용의수요에맞춰하천공간을지정하고운영하는것이한층더중요해졌다. 이에본논문은체계적이고객관적으로하천공간을지정 운영하기위해선행연구에서검증된통신빅데이터이용지표를사용하여친수지구를특성화하는방법을제안하고자한다. 첫째, 친수지구의이용도를판단하기위해면적당이용객수를로그노멀분포로변환하여 5등급으로제시하였다. 둘째, 여러가지통계분석을통해친수지구를 5개의유형으로구분하고특징을분석하 였으며, 친수지구가근린친수와친수거점의어디에해당하는지구분하였다. 셋째, 이용등급과공간활용의유형을고려하여친수지구를단계적으로지정하는방법과기준을제안하였다. 향후중장기적인모니터링과객관적인분석을통해근린친수지구와친수거점지구로지정해야할것이며, 본논문의결과는친수지구의계획, 유지관리, 인 허가에있어서합리적인결정을하는데크게기여할수있을것으로예상된다. 82 국토연구제 102 권 ( )

85 국토연구제 102 권 (2019.9): pp83~ 개특 광역시의고령친화도와고령친화정책비교에관한연구 * A Study on the Comparision between Age-Friendly Index and Age-Friendly Policy in 7 Metropolitan Cities 이광현 Lee Kwanghyun ** Abstract This study developed and analyzed the age-friendly city index which is composed of objective indicators including physical environment considering the situation in Korea and identified and analyzed the age-friendly indexes of 7 cities in the metropolitan area. The results of this study are expected to help local governments prioritize age-friendly city policies. Although it is difficult to make a comprehensive comparison because of the geographical characteristics and the local environment, it is necessary to analyze the results of the indices in each region to find causes and remedies. In addition, after analyzing the current status of age-friendly policies in each region, there has been a comparative analysis with the current index and a diagnosis about the direction of the current progress. Of course, not all regions need to establish and implement the same age-friendly policies, but equal attention and efforts must be made in the 4 areas. Keywords: Age-Friendly Indicator, Age-Friendly Index, Age-Friendly Policy, Age-Friendly City, Physical Environment I. 서론 1. 연구의배경및목적우리나라는최근급격하게진행되고있는인구고령화로인한변화를맞이할준비가필요하다. 이러한인구구성의변화는우리나라인구의 14.6%(712만 5천여명 ) 을차지하고있는베이비붐세대가점차노년기로접어들기때문이며 ( 김미령, 권순재, 박충선 2012, ), 2018년 65세이상인구가차지하는구성비 는 14.3% 로지속적으로증가하여 2060년에는 41% 가될것으로전망되고있다 ( 통계청 2018). 또한고령자고용률은 30.6% 로여전히상당수고령자들이일을하고있으며, 고령자중 61.2% 는경제적인이유또는일하는즐거움때문에장래에일하기를원한다. 앞으로고령자의인구비율과사회 경제적참여는더욱늘어날것이며, 그들을위한물리적 사회적 정책적변화와개선이필요한상황이다 (Lee and Kim 2019). 본논문은우리나라의실정에맞으면서보다객관적인방식으로개발된고령친화도시지표들을활용하 * 본논문은이광현 (2019) 의박사학위논문인 고령친화도시지표개발및평가를통한정책적함의연구 의일부를수정 보완한것임. ** 경일대학교건축학과조교수 Assistant Prof., Dept. of Architecture, Kyungil Univ. kh.lee@kiu.kr 7 개특 광역시의고령친화도와고령친화정책비교에관한연구 83

86 여 7개특 광역시의고령친화지수를확인하고각시의현황을분석하고자한다. 그리고분석결과와함께각지방자치단체가시행하고있는고령친화정책및사업을비교 분석하여현재각지방자치단체에서시행중인고령친화정책및사업이올바르게나아가고있는지를확인하고개선방향을제시하고자한다. 본논문에서도출된결과는앞으로지방자치단체들이바람직한고령친화정책을수립하는데현실적인도움이될것이다. 2. 연구의범위와방법연구의시간적범위는 세계보건기구 (WHO) 고령친화도시의가이드라인 2007년자료를시작으로 2019 년까지를범위로하였다. 지표에적용가능한데이터중국내주요기관의자료는 2017년, 2018년을기준연도로하였고, 2019년자료가가능한것은지표에반영하였다. 그리고고령친화정책자료는 2010년자료부터 2019년최근자료를조사하여반영하였다. 고령친화도시지표로활용된자료들은노인실태조사 ( 정경희, 오영희, 강은나, 김경래외 2017), 노인복지시설현황 ( 보건복지부 2019) 등보건복지부산하기관에서의조사결과를활용하였고, 시 도사회보장계획, 통계청, 행정자치부, 시 군통계연보, 한국감정원부동산통계정보시스템, 한국노인인력개발원의최근통계 조사자료들을수집하여활용하였다. 일부조사중이거나최근자료구득이불가한자료는 2016 년이전데이터를활용하였다. 고령친화정책자료는각시의홈페이지에게시되어있는부서별주요업무, 신문기사, 노인복지계획, 어르신종합계획과저출 산 고령사회정책등 2019년최근자료까지활용하였다. 1) 공간적범위는 7개특 광역시를분석단위로하였고, 세종시는일부지표의자료구득이불가하여조사범위에서제외하였다. 연구방법은 4단계로구성된다. 첫째, 이광현, 김세용 (2017) 의 고령친화도시지표개발에관한연구 의지표선정객관성을보완하여지표를수정하였다. 둘째, 수정한지표에국내주요기관의최신데이터를적용하여 7개시의고령친화도를평가및분석하였다. 셋째, 앞서나온각지역의고령친화도와현재추진중인고령친화정책을비교 분석하였다. 마지막으로본논문의결론을도출하였다. II. 이론적고찰 1. 고령친화도시의개념 <Table 1> 의선행연구와같이고령친화도시는단순히노령친화적인도시에초점을맞추는것이아니라연령에상관없이누구나살기좋은환경을만들기위한것이며, 주민이평생동안육체적, 심리적으로안전하고건강하게살수있는도시에목표를맞추어야한다. 그러기위해서는물리적인환경은다양한연령대가이용하는데불편이없고건강하게살수있는환경으로조성되어야하며, 건강 복지 사회 경제적부문은고령자가노화로인해신체적, 정신적으로어려움이있는부분을사전에준비, 지원하여고령자가사회의구성원으로함께지낼수있도록해야한다. 고령친화도시는고령자가평생을살던곳에서지속 1) 광주시의회 (2019), 대구시저출산 고령사회과 (2010), 대전시노인복지과 (2019), 보건복지부 (2018), 서울시복지건강실 (2012), 한국노인인력개발원 (2018), 행정안전부 (2018) 및광주 대구 대전 부산 서울 울산 인천시홈페이지, 국가통계포털 ( 통계지리정보서비스 ( 한국감정원부동산통계정보시스템 ( 을참고함. 84 국토연구제 102 권 ( )

87 Table 1 _ Definition of Age-friendly City WHO (2007) Seoul Welfare Foundation (2010) Kim (2010) Park (2016) Source: Park 2016, 17. 적으로불편없이살수있는환경이갖추어진곳이며, 이런환경은물리적환경과건강 복지 사회 경제적인환경이조화롭게균형을맞추었을때완성될수있다. 2. 고령친화지표 Definition A city where it is not convenient for you to get older, a city where everyone of all ages can live, and a city where elderly people can actively participate in society so that they can live a healthy old age. A city that recognizes the characteristics of the elderly population and the diversity of needs, respects the choice and decision-making of the elderly, and considers it not to be excluded as an age in the whole living environment. A city that aims to create a good environment for all citizens, not just an elderly-friendly city for the elderly, and a city that flexibly copes with the changing needs of citizens as they age. A community where everyone in an individual's community secures autonomy and independence and supports active life in a variety of physical, social, economic and cultural areas. 고령친화지표에관한선행연구는 <Table 2> 와같다. 대부분의국내선행연구의고령친화지표들은 WHO (2007) 의가이드라인을기반으로개발되었으나, 주관적인의견을묻는방식으로설계되어지속적인비교가어렵고객관성이부족하다. 본논문은고령친화도시를조성하기위한물리적환경과건강 복지 사회 경제적영역을모두포함하면서, 객관적인자료를통해고령친화지수를도출할수있는이광현, 김세용 (2017) 의 고령친화도시지표개발에관한연구 의지표를기준으로하였다. 이광현, 김세용 (2017) 의고령친화지표는 WHO(2007) 의 8대 Table 2 _ Age-friendly Indicator Studies 영역을물리적환경, 건강 복지, 사회와경제영역의 4 대영역으로재조정하였고, 12 개하위영역과총 52 개지표로구성되어있다. 하지만임의적인 1차지표선정으로이에대한객관성이부족하여보완할필요가있다. III. 지표보완및지수분석 1. 지표수정 Precedent Research Global Age-friendly Cities: A Guide (WHO 2007) Best Cities for Successful Aging (Chatterjee, King and Irving 2014) Age-Friendly Cities Performance Assessment Indicators System Validation(Pinheiro, Diogo, Gois and Paul 2015) Assessment of the City s Elder-Friendliness in Seoul (Kim, Kim and Kim 2008) Study on the Development of Age-friendly Cities in Gyeonggi-do(Kim, Nam and Ko 2011) A Study on the Evaluation of Aging in Busan (Lee and Lee 2011) Development of Detailed Guidelines for the Design of Health and Welfare Models for the Elderly Area(Jeong, Lee, Lee and Yoo 2008) Developing Indicators for Building Elder-Friendly Communities in Korea (Kim, Chang, Oh and Choi 2014) Development of Age-friendly City Indicators in South Korea (Lee and Kim 2017) Source: Lee and Kim 2017, ) 지표선정객관화 No. of Area and Item 8 areas/ 84 items 8 areas/ 84 items 3 areas/ 10 sub-areas/ 36 items 10 areas/ 49 items 8 areas/ 61 items 10 areas/ 59 items 5 areas/ 15 sub-areas/ 34 items 4 areas/ 10 sub-areas/ 29 items 4 areas/ 12 sub-areas/ 52 items 이광현, 김세용 (2017) 의지표선정의객관성을높이기위해국내 5개와해외 3개의선행연구들중최소 2개 7 개특 광역시의고령친화도와고령친화정책비교에관한연구 85

88 의연구에서중복선정된지표를기준으로 1차지표를선별하였다. 지표선정에사용된 8개의선행연구는 <Table 2> 와같다. 그리고국내의다양한자료중고령자통계 ( 통계청 2018), 노인실태조사 ( 정경희, 오영희, 강은나, 김경래외 2017), 노인일자리통계동향 ( 한국노인인력개발원 2018) 와행정안전부, 보건복지부, 한국감정원, 한국노인인력개발원, 통계청등다양한기관에서자료수집이가능한지표를기준으로 2차선별하였다. 2) 최종지표선정 이광현, 김세용 (2017) 설문조사의지표선정조건인가중치평균 0.6 이상을만족하는지표가최종으로선정되었고, 최종지표는 <Table 3> 과같이기존 52개에서 42개 ( 물리적환경18개, 건강 / 복지 10개, 사회 6 개, 경제 8개 ) 로축소되었다. Table 3 _ Number of Revised Indicators by Area Area Lee and Kim(2017) Revised Indicator Physical Environment 22 items 18 items Health/Welfare 11 items 10 items Social 9 items 6 items Economic 10 items 8 items Total 52 items 42 items 2. 지표의표준화 선정된물리적환경, 건강 복지, 사회, 경제 4대영역의총 42개지표조사값들은비율과절대치들이혼재되어있어지수화하기에는적합하지않아, 지표의지역별비교를위해표준화과정을거쳤고, 각지표의표준화값은 -2~+2 의범위에속한다. 3. 지역별지수분석스파이더그램을통해시별고령친화도를한눈에볼수있도록하고자했으나 42개지수를한스파이더그램에서확인하기는어려워, <Figure 1> 과같이중분류별로합산하여최솟값과최댓값를원의구성요소로설정하고지수의범위는 +8~-8 로설정해원의바깥쪽에가까울수록고령친화적상태임을표시하였다. 특히범죄율, 안전사고, 대기오염도, 노인학대, 기초생활보장수급등의비친화적지표는부호를변환하였고, 별표 (*) 로 <Table 4> 와같이표기해구분하였다. 부문별 7개도시의고령친화지수는 <Table 5> 와같다. 서울은건강 / 복지와사회영역에서높은지수를보이고있지만, 경제부분에서는가장낮은지수를보였다. 서울은일자리지원관련지수가매우낮기때문인데이에대한정책적노력을기울여야할것으로보인다. 부산은 7개시중에가장고령인구비율이높지만모든영역에서지수가낮아전반적인고령화정책에대한보완이필요할것으로판단된다. 대구는부산다음으로고령인구비율이높지만경제영역에서지수가낮게나타나므로이를정책에반영하여개선해야할필요가있다. 인천은건강 / 복지영역이다소낮고사회영역의지수가 7개시중가장낮아이를보완해야할것으로보인다. 광주는물리적환경영역의지수가 7개시중가장낮아이를보완한다면고령친화도시를조성하는데큰문제가없을것으로판단된다. 대전은모든영역에서높은지수를보이고있으며사회와건강 / 복지영역은가장높은지수를나타내, 7개시중가장고령친화적인도시라고할수있다. 마지막으로울산은건강 / 복지영역이 7개시중에가장낮은지수를보였다. 울산은아직고령인구비율이가장낮은편이긴하지만앞으로증가할고령인구에대한준비를해야할것으로보인다. 86 국토연구제 102 권 ( )

89 Area Sub-area Indicator Measurement and Coding Weighted Value Urban Park Space Area of urban park per 1,000 people 0.83 Bicycle Track Extension of bicycle tracks per unit area 0.77 Outdoor Negligent Accidents* Negligent accident experience over 65 years old 0.78 Spaces Crime Rate* Number of crimes per 1,000 people 0.72 Air Pollution* Fine dust in the atmosphere 0.77 Buildings Accommodation Accessibility ratio at buildings with accessibility obligation 0.83 Public Toilet Number of public toilets per 1,000 people 0.73 Transportation Fare Satisfaction Survey 0.69 Kindness Satisfaction Survey (Kindness of bus drivers and station personnel) 0.72 Facilities for the Mobility Impaired Satisfaction Survey 0.81 Physical Congestion Satisfaction Survey (Public transportation congestion) 0.63 Environment Transportation Low-floor Bus Rate of low-floor buses 0.74 Elderly Traffic Accident* Number of elderly traffic accidents per 1,000 people over age Preference of Public Transportation Usage rate of public transportation when outside by people over age Homeownership Rate Homeownership rate of people over age House Price* Average home sales price 0.68 Housing Percentage of residences of people over age 65 that have elderly Convenience of Living care equipment 0.75 Housing Welfare Facility for the Number of h ousing welfare facilities for Elderly the elderly per 10,000 people over age Number of Hospitals Number of hospitals per 1,000 people 0.75 Number of Doctors Number of doctors per 1,000 people 0.74 Medical Services Satisfaction with medical services 0.81 Health Medical Welfare Facility Number of medical welfare facilities for the elderly per 1,000 people for the Elderly over age Health/ Welfare Social Economic Welfare Social Participation Respect Education/ Information Economic Level Employment Help Services General Health Condition Percentage of people over 65 who think they are reasonably healthy 0.70 Checkup Rate Checkup rate of people over age Exercise Rate Exercise rate of people over age Social Welfare Budget Social welfare budget ratio by city and province 0.75 Number of Government Employees Number of welfare recipients per government employee in charge of welfare for the elderly 0.65 Residential Care Welfare Facilities Number of residential care welfare facilities for the elderly per 1000 people for the Dlderly over age Group Participation Rate Group participation rate of people over age Relationship with the Community Percentage of positive relationships with friends and community of people over age Use of Senior Citizen Community Center Utilization rate of senior citizen community center over age Elder Abuse* Abuse reporting rate for elder abuse per 1,000 people over age Contact with Family Frequency of visiting to and from a non-living child 0.76 Source: Lee and Kim 2017, 57-58; Lee and Kim 2019, 6-7. Table 4 _ Age-friendly Indicators with Weighted Value Lifelong Education Number of lifelong education facilities per 10,000 people over age Economic Condition Rate of economy condition more than quartile of people over age Public Assistance* Recipient rate of national basic livelihood security of people over age Employment Rate Percentage of the elderly with jobs over age Jobs Created Number of jobs created per 1,000 people over age Financial Investment Money invested per person over age 65 (KRW 1,000) 0.66 Participants Number of participants over age 65 for elderly jobs 0.65 Supporting Organization Number of local government organizations supporting elderly jobs per 100,000 people over age Job Fulfillment Institution Number of institutions fulfilling elderly jobs per 10,000 people over age 65 (Excluding municipalities) 개특 광역시의고령친화도와고령친화정책비교에관한연구 87

90 Figure 1 _ Age-friendly Spidergram by City Seoul Busan Daegu Incheon Gwangju Daejeon Ulsan Area Sub-area Seoul Busan Daegu Incheon Gwangju Daejeon Ulsan Outdoor Spaces Buildings Physical Transportation Environment Housing Sum Health/ Welfare Social Economic Table 5 _ Age-friendly Index by Area in 7 Cities Health Welfare Sum Social Participation Education/Information Respect Sum Economic Level Employment Help Services Sum 국토연구제 102 권 ( )

91 IV. 고령친화지수와정책비교 1. 정책비교방법본논문은 7개특 광역시의고령친화정책, 시행계획과예산등다양한고령친화자료를분석후고령친화지수가평균보다낮은지표에대해각지방자치단체에서개선가능한정책을반영하고있는지확인하였다. 2. 시별분석결과 1) 서울서울의고령친화도, 서울어르신종합계획분야별추진계획 ( 서울시복지건강실 2012) 과서울복지본부예산을비교한결과는 <Table 6> 과같다. 서울은다른시와다르게물리적인환경영역에대한개선도 어르신종합계획 에포함되어있다. 특히서울은유니버설디자인가이드라인을공공시설물에적용하여안전사고, 노인교통사고를방지하고자하고있으며, 고령자를위한임대주택과독거어르신지원주택공급등의사업등이정책에반영되어있다. 이외의물리적환경영역의지표는타부서와의협업이이루어져야만하는것으로부서간의역할정립이필요할것이다. 건강 / 복지영역에는 건강한노후 정책으로재가노인복지시설과노인의료복지시설등의정책이포함되어있으나, 건강관리 유지를위한건강교육정책의반영이필요하다. 사회영역에서는경로당활성화를위한다양한사업이정책에반영되어있으나, 다양한단체활동참여가가능한프로그램을추가할필요가있다. 경제영역에서 맞춤형일자리정책 이시행되고 Table 6 _ Comparison between Policy and Index below Average in Seoul 있지만, 공공일자리에치중하고있어민간기업의참여가필요한것으로보인다. 전반적인경제영역의지수가낮아지속적인재정투입과프로그램개발이필요하다. 2) 부산 Area Indicator Index Reflection Physical Environment Health/ Welfare Social Economic Negligent Accidents Crime Rate Accommodation Public Toilet Congestion Homeownership Rate House Price Housing Welfare Facility for the Elderly Medical Services Medical Welfare Facility for the Elderly Social Welfare Budget Residential Care Welfare Facilities for the Dlderly Group Participation Rate Use of Senior Citizen Community Center Contact with Family Economic Condition Employment Rate Jobs Created Financial Investment Participants Supporting Organization Job Fulfillment Institution 부산의고령친화도, 부산시고령친화도시조성을위한노인복지실행계획 ( 부산복지개발원 2016) 과예산 7 개특 광역시의고령친화도와고령친화정책비교에관한연구 89

92 실행 ( 부산광역시홈페이지 2) ) 을비교한결과는 <Table 7> 과같다. 노인복지실행계획에는물리적환경영역과관련된사항이없으며, 해당지표에대해담당부서에서독립적으로관리하고있다. 외부환경, 교통과주택에관한사항은노인의생활과밀접한관계가 Table 7 _ Comparison between Policy and Index below Average in Busan Area Indicator Index Reflection Physical Environment Health/ Welfare Social Economic Urban Park Space Bicycle Track Crime Rate Air Pollution Public Toilet Transportation Fare Low-floor Bus Convenience of Living Housing Welfare Facility for the Elderly Medical Welfare Facility for the Elderly General Health Condition Exercise Rate Residential Care Welfare Facilities for the Dlderly Group Participation Rate Relationship with the Community Use of Senior Citizen Community Center Contact with Family Lifelong Education Economic Condition Public Assistance Employment Rate Jobs Created Financial Investment Participants 있으므로해당부서와지속적인협업을할필요가있다. 건강 / 복지영역은 노인요양시설확충 사업에노인의료복지시설확충, 재가노인지원서비스센터운영지원등의일부지표가정책에반영되어있다. 다만평소에운동을할수있는공간을마련하고정기적인건강교육및관리를하는사업이병행된다면바람직할것이다. 사회영역은 지역밀착형노인복지인프라구축 정책에경로당지원이포함되어있고, 노인대학및노인교실운영도포함되어있다. 다만, 주변이웃과지역사회와의관계를유지할단체참여프로그램의보완이필요하다. 경제영역에서는 건강하고활기찬노후생활보장 정책에서노인일자리추진, 고령인력종합관리센터운영, 노인취업교육센터운영등의사업을추진하고있다. 그러나전반적인지수가낮아정책의실효성도동시에높여야할것이다. 3) 대구대구의고령친화도, 대구고령사회시행계획 ( 대구시저출산고령사회과 2010), 세부예산안과어르신복지과의업무 ( 대구광역시홈페이지 3) ) 를비교한결과는 <Table 8> 과같다. 물리적환경영역의정책은대부분의다른시와같이고령사회시행계획에서제외되어있다. 정책을담당하고있는어르신복지과에서는물리적환경영역에대해타부서의정책및진행사항을모니터링및협업할필요가있다. 다만, 고령자를배려한대중교통이용환경조성 사업이고령사회시행계획에포함되어대구의저상버스지수는양호한것으로나타났다. 건강 / 복지영역은현재노인요양및응급구조를위 2) (2019년 8월 5일검색 ). 3) (2019년 8월 5일검색 ). 90 국토연구제 102 권 ( )

93 Table 8 _ Comparison between Policy and Index below Average in Daegu Area Indicator Index Reflection Physical Environment Health/ Welfare Social Economic Urban Park Space Bicycle Track Negligent Accidents Kindness Elderly Traffic Accident Convenience of Living Housing Welfare Facility for the Elderly Checkup Rate Social Welfare Budget Number of Government Employees Residential Care Welfare Facilities for the Elderly Use of Senior Citizen Community Center Lifelong Education Public Assistance Employment Rate Jobs Created Financial Investment Supporting Organization Job Fulfillment Institution Table 9 _ Comparison between Policy and Index below Average in Incheon Area Indicator Index Reflection Physical Environment Health/ Welfare Social Economic Accommodation Transportation Fare Kindness Facilities for the Mobility Impaired Congestion Low-floor Bus Number of Hospitals Medical Services General Health Condition Social Welfare Budget Number of Government Employees Residential Care Welfare Facilities for the Elderly Group Participation Rate Use of Senior Citizen Community Center Elder Abuse Lifelong Education Economic Condition Public Assistance Job Fulfillment Institution 한인프라확충사업, 재가서비스활성화, 건강검진수검률제고등의사업이정책에반영되어있다. 사회영역은여가문화활성화를위한사업이시행되고있으나빈도수나참여인원이미비하여보다적극적인프로그램개발이필요한것으로판단된다. 대구는다른시에비해경로당이용률이낮아경로당지원에대한사업과평생교육에대한지원이필요할것으로보인다. 경제영역은노인일자리사업지원을확대하고전담기관을운영중에있어이미정책에반 영되어있으나, 지속적인재정투입과민간기업의참여를확대할필요가있다. 4) 인천인천의고령친화도와 인천노인정책과업무 ( 인천광역시홈페이지 4) ) 의비교결과는 <Table 9> 와같다. 인천은개별적인사업위주의노인정책을시행하고있어큰비전성립과함께추진전략을세워효과적으 4) (2019 년 8 월 5 일검색 ). 7 개특 광역시의고령친화도와고령친화정책비교에관한연구 91

94 로사업을진행할필요성이있어보인다. 현재 인천고령사회대응센터 에서 인천시고령사회대응을위한중장기계획수립 (2019~2023) 을하고있다. 계획이수립되면중장기적계획과전략이도출될것으로보인다. 물리적환경영역은다른시와같이타부서와의공동지표관리가필요하며, 특히대중교통담당부서와개선정책을모색할필요가있다. 건강 / 복지영역은평소에건강유지를위한프로그램을노인복지시설과연계하여지원할필요가있다. 사회영역은다른시와다르게경로당지원사업이제외되어있어, 경로당지원사업을검토할필요가있다. 또한고령자가다양한단체활동이가능하도록프로그램개발및단체지원사업을정책에반영하여낮은사회영역의지수를높이도록해야할것이다. 경제영역은노인정책과의담당업무에서는빠져있지만, 최근 인천형적합노인일자리사업 을공모하였다. 지속적으로공공과민간에서노인일자리지원에노력해야하며이를노인정책과에서관리해야한다. 5) 광주광주의고령친화도, 광주시노인복지정책 ( 광주광역시홈페이지 5) ), 사회복지분야예산과의비교결과는 <Table 10> 과같다. 물리적환경영역에서는 어르신이행복한빚고을 효령노인타운운영지원사업 이일부지표에반영되었으나, 대부분의지표는다른시와같이타부서와의협업을통한개선이필요하다. 또한광주는물리적환경영역의지수가 7개시중가장낮아전반적인지표에대한개선방향을관련부서와협업하여수립해야 Table 10 _ Comparison between Policy and Index below Average in Gwangju Area Indicator Index Reflection Physical Environment Health/ Welfare Urban Park Space Crime Rate Accommodation Kindness Facilities for the Mobility Impaired Low-floor Bus Elderly Traffic Accident Preference of Public Transportation Homeownership Rate Convenience of Living Housing Welfare Facility for the Elderly Medical Welfare Facility for the Elderly Exercise Rate Number of Government Employees Relationship with the Community Social Elder Abuse Contact with Family Economic Public Assistance 할것이다. 특히대중교통과와협력하여대중교통의개선방향을우선적으로모색할필요가있다. 건강 / 복지영역은대체적으로지수가양호하지만노인의료복지시설확충사업을추가할필요가있고, 고령자가운동할수있는환경조성과교육정책을추가할필요가있다. 사회영역은지수가낮은지표에대해정책이잘반영되어있으며, 경제영역에서노인일자리지원이잘되고있는것으로보인다. 5) (2019 년 8 월 5 일검색 ). 92 국토연구제 102 권 ( )

95 6) 대전 7) 울산 대전의고령친화도, 2019년도노인복지주요사업 ( 대전시노인복지과 2019) 과노인복지과예산 ( 대전광역시홈페이지 6) ) 의비교결과는 <Table 11> 과같다. 대전은모든영역에서평균이상의지수를나타내고있어, 일부지표에대한관리만한다면가장고령친화적인도시가될것이다. 물리적환경영역은타부서와긴밀히협업하여개선해야할것이며, 건강 / 복지영역은현재의지수가높으나지속적으로유지될수있도록관리해야할것이다. 사회영역도 가족왕래 지표를제외하고모두평균이상으로고령친화적인환경을잘조성하고있는것으로보인다. 다만, 효에대한교육과같은 가족왕래 지표가개선될수있는방안을고려할필요가있다. 경제영역은노인일자리사업관련지수가대부분높지만 경제활동참가율 지표를개선할수있는정책을검토할필요가있다. Table 11 _ Comparison between Policy and Index below Average in Daejeon Area Indicator Index Reflection Physical Environment Negligent Accidents Air Pollution Elderly Traffic Accident Preference of Public Transportation Homeownership Rate Convenience of Living Medical Services Health/ Welfare Number of Government Employees Social Contact with Family Economic Employment Rate 울산의고령친화도, 울산노인복지정책 ( 울산광역시홈페이지 7) ) 과사회복지예산과의비교결과는 <Table 12> 와같다. 물리적환경영역중건물영역의지수 Table 12 _ Comparison between Policy and Index below Average in Ulsan Area Indicator Index Reflection Physical Environment Health/ Welfare Bicycle Track Transportation Fare Kindness Facilities for the Mobility Impaired Congestion Low-floor Bus Elderly Traffic Accident Preference of Public Transportation Housing Welfare Facility for the Elderly - Number of Hospitals Number of Doctors Medical Services Medical Welfare Facility for the Elderly General Health Condition Checkup Rate Exercise Rate Social Welfare Budget Number of Government Employees Residential Care Welfare Facilities for the Elderly Relationship with the Community Social Elder Abuse Lifelong Education Economic Economic Condition ) (2019 년 8월 5일검색 ). 7) (2019년 8월 5일검색 ). 7 개특 광역시의고령친화도와고령친화정책비교에관한연구 93

96 가매우낮은편으로해당부서와협업하여지표관리가필요한실정이다. 건강 / 복지영역은모든지표가평균이하로기본적인의료시설확충과서비스를개선할수있도록정책수립이필요할것으로보인다. 또한건강검진과지속적인건강관리가가능하도록홍보, 교육과운동환경등을제공할필요가있다. 사회영역은노인학대예방과인권확보를위한프로그램운영사업이필요한것으로보인다. 경제영역은대부분지수가평균이상이지만 경제상태 지표가개선될수있는추가정책을모색할필요가있다. 3. 소결 본논문은 7개시의고령친화지수중평균보다낮은지표들과현재지방자치단체에서시행하고있는정책및예산과비교하였다. 물리적환경영역에대한것은 <Table 13> 과같이일부지역에서포함하고있는부분도있으나, 대부분의고령화정책이나노인복지정책에포함되지않아관련부서와의긴밀한협업과정보교환이필요할것으로보인다. 서울시고령친화정책에는다른도시에비해물리적 Table 13 _ Policy Reflection Status by Area in 7 Cities City Physical Environ-ment Health /Welfare Social Economic Seoul 3/8 1/4 2/3 7/7 Busan 0/9 2/4 3/5 6/6 Daegu 0/7 2/4 2/2 6/6 Incheon 0/6 1/6 1/4 3/3 Gwangju 1/11 0/3 2/3 1/1 Daejeon 6/6 0/2 0/1 1/1 Ulsan 0/9 2/10 1/3 1/1 Note: A/B=Number of Indicator Reflecting Policy/Number of Indicator below Average. 환경개선정책이포함되어있어타도시의모범이되나, 건강 / 복지영역의지표를개선할수있는추가적인정책반영이필요하다. 부산은물리적환경영역에서평균이하의지표가 9개로이를개선할수있는정책을타부서와협의하여반영할필요가있으며, 타영역에서도미반영된지표를개선할수있는정책수립이요구된다. 대구는물리적환경영역에서저상버스를도입하는정책이반영되어있어저상버스에관한지수가높았고, 타도시에비해물리적환경을제외한정책반영도가높은편이었다. 인천은 고령사회대응센터 를만들어현재중장기계획을수립하는중으로타도시에비해정책완성도는낮은편이며, 정책수립시사회영역의경로당지원정책반영이필요하다. 광주는물리적환경 18개지수중 11개가평균이하로이를개선할수있는정책을관련부서와협업하여수립할필요가있다. 건강 / 복지영역에서의미반영된지표에대해개선정책을추후반영할필요가있다. 대전은모든영역에서평균이상의지수를나타내고있으나, 평균이하의지표에대해서는정책반영도가낮아이를개선할정책을추가수립할필요가있다. 울산은물리적환경과건강 / 복지영역의평균이하지표에대한정책반영이필요하다. 특히건강 / 복지영역은 10개의모든지표가평균이하인데 2개의지표만정책에반영되어있어우선적으로검토해야할것이다. V. 결론본논문은크게 3가지단계로이루어진다. 첫째, 지표선정의객관성을높여고령친화지표를수정하였고, 둘째, 고령친화지표에지표값을적용하여우리나라 7 94 국토연구제 102 권 ( )

97 개특 광역시의고령친화도를분석하는연구와마지막으로고령친화지수와각지역에서시행중인고령친화정책과비교 분석하였다. 본논문의결론은다음과같다. 첫째, 기존연구의고령친화지표선정의객관성을높여, 7개특 광역시에적용할수있는 4개영역, 11개부문의총 42개지표로수정하였다. 둘째, 물리적영역중일상활동을위한편의시설제공, 이동교통수단의편의성과건강유지및관리를위한시설확보등은도시의기본구성요소이지만, 고령자에게는이런기본적인물리적환경이가장중요하다. 이러한사항은고령자가살던집과공동체에서안전하고자립적으로계속살고싶어하는 Aging in Place 와 Aging in Community 의개념과연계한노인복지정책의필요성을더욱보여주었다. 그럼에도불구하고물리적환경영역에대한정책은서울을제외하고고령친화정책에적극적으로포함되어있는시는없었다. 대구와광주같이 1개지표만연관있는곳도있었지만, 물리적환경영역의지표들은교통과, 공원녹지과, 경찰청, 주택과, 건설과등다른기관과부서에서주관하는지표이기때문에사실상포함하기는어려운실정이다. 그러므로각지표에관련된부서와정책수립시협의를하고, 각지표를지속적으로모니터링해야할것이다. 보다적극적인고령친화정책을위해서는전체관련부서와협의및업무추진이가능한별도의부서가필요할것으로보인다. 셋째, 지표에대한정책이이미잘수립되어시행중인부분도있지만, 지수가낮은영역과보완해야할지표들도있어각시의담당부서는추후정책수립시잘진행되고있는것과보완해야할지표에대해충분한조사와분석을하여야할것이다. 그리고모든영역에서균형있는고령친화가조성될수있도록노력을기울여야할것이다. 넷째, 대부분의지방자치단체의사회복지분야의예산은매년증가하고있으며, 노인복지정책의예산도비례하여증가하고있다. 하지만일부지방자치단체에서는노인복지정책예산의상당부분을노인기초연금지원에배정하고있어, 고령자가사회의한구성원으로의미있는삶을살수있도록사회참여도를높일수있는경제및사회 문화생활을장려하는정책에예산반영비율을높여야할것이다. 마지막으로고령자들이사회구성원으로일상생활을하는데불편함이없도록모든영역을개선하기위해지방자치단체에서는지금도많은노력을하고있다. 하지만지속적인분석과개선을통해고령화정책의완성도를높여야할것이다. VI. 연구의한계와의의본논문은 7개특 광역시에적용하여고령친화지수를도출하였다. 또한고령친화지수와현재시행중인고령친화정책과의비교 분석을통해각지역이보완해야할정책의방향을제시하였다. 하지만, 다음과같은한계가있어추후보완및추가연구할분야가남아있다. 자료구득의한계로 7개시를적용공간단위로제한하였으나, 주요기관으로부터통계자료의원자료 (Raw Data) 가공유된다면시 군 구까지연구범위가확대가능할것이다. 시 군 구로연구범위가확대되어중소도시등도시규모나지역의특성에대한분석이추가된다면더욱정확하고의미있는연구가될것으로기대된다. 그럼에도불구하고본논문은지방자치단체가타지역과비교하여해당지역의현재수준을파악할수있으며, 높은지수는계속유지하고낮은지수의지표들은개선함으로써고령친화도시로발전하는데도움이될수있을것이다. 7 개특 광역시의고령친화도와고령친화정책비교에관한연구 95

98 본논문에서도출된고령친화지수와시행중인정책과의비교 분석결과는각지방자치단체들이고령친화지수가낮은지표를보완하는사업에우선순위를부여하는데참고가될수있을뿐만아니라지역특성에맞는바람직한도시 관리정책과프로그램개발에도기여할수있을것이다. 참고문헌 1. 광주시의회 년제1회추가경정일반및기타특별회계세입 세출예산서. 광주 : 광주시의회. Gwangju City Council Additional Supplementary General and Other Special Accounting Revenue and Expenditure Budget. Gwangju: Gwangju City Council. 2. 광주광역시. oardview.do?seq=2&menuid=gwangju &board Id=BD_ (2019년 8월 5일검색 ). Gwangju City. /boardview.do?seq=2&menuid=gwangju &boa rdid=bd_ (accessed August 5, 2019). 3. 국가통계포털. (2019년 9월 3일검색 ). Korean Statistical Information Service. (accessed September 3, 2019). 4. 김미령, 권순재, 박충선 베이비붐세대의정보활용이정보화생활만족도에미치는영향. 노인복지연구 57권 : Kim Miryoeng, Kwon Sunjae and Park Chungsun Babyboomer s use of information technology and it s effect on the digital life satisfaction: The mediating effect of the self-mastery. Korean Journal of Gerontological Social Welfare 57: 김선자, 김경혜, 김현정 서울시고령친화도평가연구. 서울 : 서울연구원. Kim Sunja, Kim Kyunghae and Kim Hyunjung Assessment of the City s Elder-friendliness in Seoul. Seoul: Seoul Research Institute. 6. 김선자 서울의고령친화도시추진전략. SDI 정책리포트 64호, 서울 : 서울시정발전연구원. Kim Sunja Seoul s strategy for aging-friendly cities. SDI Policy Report 64, Seoul: Seoul Research Institute. 7. 김수영, 장수지, 오찬옥, 최상희 고령친화공동체구축을위한지표개발. 한국노년학 34권, 3호 : Kim Sooyoung, Chang Sujie, Oh Chanohk and Choi Sunghee Developing indicators for building elder-friendly communities in Korea. Journal of the Korean Gerontological Society 34, no.3: 김영혜, 남효정, 고주희 경기도고령친화도시조성연구. 수원 : 경기도가족여성연구원. Kim Younghae, Nam Hyojeong and Ko Juhee A Research on Age-friendly City in Gyeonggi Province. Suwon: Gyeonggido Family & Women s Research Institute. 9. 대구광역시. = (2019년 8월 5일검색 ). Daegu City. = (accessed August 5, 2019). 10. 대구시저출산 고령사회과 저출산 고령사회시행계획요약보고서. 대구 : 대구시저출산 고령사회과. Daegu Low Fertility and Aging Social Division Summary Report of Low Fertility and Aging Society Action Plan. Daegu: Daegu Low Fertility and Aging Social Division. 11. 대전광역시. dlist.do?boardid=budboard005&menuseq=349 (2019년 8월 5일검색 ). Daejeon City. ardlist.do?boardid=budboard005&menuseq=349 (accessed August 5, 2019). 12. 대전시노인복지과 년도노인복지주요사업. 대전 : 대전시노인복지과. Daejeon Senior Welfare Division Senior Welfare Projects in Daejeon: Daejeon Senior Welfare Division. 13. 박은희 대구고령친화도시조성기초연구. 대구 : 대구경북연구원. Park Eunhee Basic Research on Daegu Aging-friendly City Development. Daegu: Daegu Gyeongbuk Development Institute. 14. 보건복지부 보건복지통계연보. 세종 : 보건복지부. Ministry of Health and Welfare Statistical Yearbook. Sejong: Ministry of Health and Welfare 노인복지시설현황. 세종 : 보건복지부 Elderly Welfare Facilities. Sejong: Ministry of Health and Welfare. 96 국토연구제 102 권 ( )

99 16. 부산광역시. ge=&srchbegindt=&srchenddt=&srchkey=&srchtext= (2019년 8월 5일검색 ). Busan City. =&srchbegindt=&srchenddt=&srchkey=&srchtext= (accessed August 5, 2019). 17. 부산복지개발원 부산시고령친화도시조성을위한노인복지실행계획. 부산 : 부산복지개발원. Busan Social Welfare Development Institute Welfare Plan for the Elderly to Create an Aging-friendly City in Busan. Busan: BSWDI. 18. 서울시복지건강실 서울어르신종합계획. 서울 : 서울시복지건강실. Seoul Welfare and Health Office Seoul Senior Comprehensive Plan. Seoul: Seoul Welfare and Health Office. 19. 서울특별시. (2019년 8월 5일검색 ). Seoul City. (accessed August 5, 2019). 20. 울산광역시. (2019년 8월 5일검색 ). Ulsan City. (accessed August 5, 2019). 21. 이광현, 김세용 고령친화도시지표개발에관한연구 : 고령자입장에서바라본측정가능한물리적환경지표중심으로, 한국도시설계학회지 18권, 2호 : Lee Kwanghyun and Kim Seiyong Develop age-friendly city indicators. Journal of the Urban Design Institute of Korea 19, no.2: 이민홍, 이재정 부산시고령친화도평가연구. 부산 : 부산복지개발원. Lee Minhong and Lee Jaejeong A Study on the Evaluation of Aging Affinity for Busan Metropolitan City. Busan: Busan Social Welfare Development Institute. 23. 인천광역시 / (2019년 8월 5일검색 ). Inchoen City / (accessed August 5, 2019). 24. 정경희, 오영희, 강은나, 김경래, 이윤경, 오미애, 황남희외 노인실태조사. 세종 : 보건복지부. Jeong Kyeonghee, Oh Younghee, Kang Eunna, Kim Kyeongrae, Lee Yunkyeong, Oh Miae and Hwang Namhee et al Elderly Survey. Sejong: Ministry of Health and Welfare. 25. 정경희, 이윤경, 이소정, 유삼현 초고령지역의보건복지모형설계를위한세부지침개발. 서울 : 한국보건사회연구원. Jeong Kyeonghee, Lee Yoongyeong, Lee Sojeong and Yoo Samhyun Development of Detailed Guidelines for the Design of Health and Welfare Models for the Elderly. Seoul: Korea Institute for Health and Social Affairs. 26. 통계지리정보서비스. (2019년 9월 3일검색 ). Statistical Geographic Information Service. go.kr/publicsmodel (accessed September 3, 2019). 27. 통계청 고령자통계. 9월 27일, 보도자료. Statistics Korea Statistics of the elderly. September 27, Press release. 28. 한국감정원부동산통계정보시스템. /rone/resis/statistics/statisticsviewer.do (2019 년 8월 5일검색 ). Korea Appraisal Board R-ONE. e/resis/statistics/statisticsviewer.do (accessed August 5, 2019). 29. 한국노인인력개발원 노인일자리통계동향. 고양 : 한국노인인력개발원. Korea Labor Force Development Institute for the Aged Elderly Job Statistics Trend. Goyang: Korea Labor Force Development Institute for the aged. 30. 행정안전부 행정안전통계연보. 서울 : 행정안전부. Ministry of the Interior and Safety Statistical Yearbook. Seoul: Ministry of the Interior and Safety. 31. Chatterjee, A., King, J. and Irving, P Best Cities for Successful Aging. Santa Monica: Milken Institute. 32. Lee Kwanghyun and Kim Seiyong Development of age-friendly city indicators in South Korea. Urban Design International, Pinheiro, F. A., Diogo, M. T., Góis, J. E. and Paúl, C Age-friendly cities performance assessment indicators system validation. Psychology 6, no.5: World Health Organization(WHO) Global Age-friendly Cities: A guide. Geneve: World Health Organization. 7 개특 광역시의고령친화도와고령친화정책비교에관한연구 97

100 논문접수일 : 심사시작일 : 심사완료일 : 요약 주제어 : 고령친화지표, 고령친화도, 고령친화정책, 고령친화지수, 고령친화도시 우리나라는베이비붐세대가전체인구의 14.6% 를차지하고있다. 이들이곧노년기로접어들기때문에급격한인구고령화에대한대비가필요한상황이다. 지방자치단체들은각지역에대한고령화현황조사, 고령친화지표, 고령친화가이드라인에대한정책연구와실행계획을수립하고있다. 하지만기존연구들은사회 복지에관련된지표개발에만치중되어있고, 지표에대한값이고령자설문조사의결과에의존하고있어, 이것을기반으로정책수립시객관성과정확성에문제가있다. 본논문은한국의실정에맞으며물리적환경을포함한객관적지표로구성된고령친화도시지표를수정 보완하였다. 수정된지표들에국내주요기관의자료를기입하여 7개특 광역시의고령친화도를확인및분석하였다. 물론지리적특성 과지역환경의차이가있으므로일괄적으로비교하기는힘들지만각지역의지수를분석후원인과개선책을모색하는것은꼭필요한과정이다. 또한각지역의고령친화정책현황을파악한후고령친화지수와의비교분석을통해현재시행하고있는정책방향에대한진단을실시하였다. 모든지역이동일한고령친화정책을수립하고실행해야하는것은아니지만, 네가지영역에대한균등한관심과노력이이루어져야만할것이다. 마지막으로고령자들이사회구성원으로일상생활을하는데불편함이없도록모든영역을개선하기위해지방자치단체에서는지금도많은노력을하고있다. 하지만지속적인분석과개선을통해고령화정책의완성도를높여야할것이다. 98 국토연구제 102 권 ( )

101 국토연구제 102 권 (2019.9): pp99~116 서울시취약계층밀집지역분포와주거지특성분석 : 민간임대주택거주기초생활수급자를중심으로 Analyses on the Characteristics in the Distribution of the Concentration of the Vulnerable Groups and the Residential Areas in Seoul : Focusing on Residents of Private Rental Housing 황종아 Hwang Jong-A *, 구자훈 Koo Jahoon ** Abstract The aims of this study is to measure the concentration of the vulnerable groups and analyze the characteristics of the concentrated area using a spatial autocorrelation analysis. The study is focusing on the dwelling areas of vulnerable groups considering financial conditions. For this, the analysis targets were set as residents of private rental housing except those living in public rental housing in Seoul. The results showed that the spatial density of vulnerable groups was derived and shown to increase over time. In addition, the results of the spatial correlation is increasing between dense areas and deteriorated low-rise residential areas. Moreover, it was confirmed that the areas concentrated around the city center which are recognized as representative residences for vulnerable groups; these results show that both spatial polarization and economic polarization are occurring, which is an area of concern for our society. In housing policy, it is necessary to supply rental housing that satisfies the local space needs of vulnerable groups. Also, it is necessary to secure sufficient affordable housing for vulnerable groups because low-cost housing can be demolished as a result of the urban redevelopment project, so they can be concentrated in deteriorated areas. Although based on traditional theories, the segregation of residences is inevitable because of socioeconomic differences and the allocation of resources. However, if these trends prevail, they can accelerate regional imbalances and cause social conflicts, which will incur policy efforts at both the urban and spatial level. Keywords: Concentration of the Vulnerable Groups, Residents of Private Rental Housing, Spatial Autocorrelation, Deteriorated Low-rise Residental Area, Spatial Polarization I. 서론 1. 연구의배경및목적 쇠퇴한주거지역은시간이지날수록중간소득층이유 출되고, 저소득층과취약계층이남게된다. 일단저소득층이나취약계층이거주하는지역으로전락하면, 주택및주거환경개선에대한자발적노력이나투자가줄어들게되고, 이는다시주거환경악화와부정적이미지강화로이어지는악순환으로이어진다 ( 서울연 * 한양대학교도시대학원박사수료 ( 제1저자 ) Ph.D. Candidate, Graduate School of Urban Studies, Hanyang Univ. Primary Author jonga0607@naver.com ** 한양대학교도시대학원교수 ( 교신저자 ) Prof., Graduate School of Urban Studies, Hanyang Univ. Corresponding Author jhkoo@hanyang.ac.kr 서울시취약계층밀집지역분포와주거지특성분석 99

102 구원 2016, 255). 이러한일련의과정이진행되면결국취약계층이특정지역으로밀집하게되고, 사회적계층간의공간적분리로이어질수있다. 이는지역간주거수준의양극화를확대할수있으며, 심각한공간적격리를일으키고, 거주민의사회적배제등과같은사회적갈등을증폭시킬수있다는점에서경계해야하는현상이다 ( 김걸 2017, 62; 양승호 2014, 20). 많은도시연구에서빈곤층이특정지역에군집하는현상은많은사회경제적병리현상을야기할수있다고지적해왔다 (Duncan and Duncan 1955; Jargowsky 1997; Massey and Eggers 1990; Sampson, Raudenbush and Earls 1997). 자본주의사회에서소득수준에따른주거지분리현상은어쩔수없는자연스러운현상이며, 계층별입지과정은도시의물리적제약, 공공도시개발의결과이기는하지만기본적으로가구들의입지결정은각자의자유의지에의해결정되었다고할수있다 ( 주희선 2015, 2; 배순석, 전성제 2006, 192). 반면, 우리나라에서는빈곤층및저소득층의주거안정을위해공공임대주택이공급되었다. 대규모단지형임대주택은저소득층의집단주거지라는인식하에부정적인식을비롯하여대규모공급으로인한저소득층의집단화에따른낙인화, 사회적배제, 격리현상등의사회적문제를야기한다는평가를받아왔다 ( 윤인숙 2004, 11). 이러한사회적문제를완화하기위해임대주택의대안으로다가구매입을통한소규모로분산공급방안이시행되고있다. 이렇듯정부가공공임대주택을통해취약계층을수용하는정책을추 진하면서이들의주거입지는공공의결정에큰영향을받았다고볼수있다 ( 배순석, 전성제 2006, 193). 이러한임대주택의입지와외부효과에관한연구는영구임대주택이나다가구매입임대주택등다양한공공부문의임대주택의유형을통해다뤄졌다. 그러나경제적여건을반영한민간부문의임대주택에관한연구는아직미진한상황이다. 이에본논문에서는사회적계층간의주거지분리현상을민간임대주택에거주하는취약계층을대상으로공간적자기상관분석을통해그수준의변화를계량적으로측정하고자한다. 이때, 공공정책에의해집단화된공공임대아파트단지와매입임대주택에거주하는임대주택거주자를제외함으로써, 취약계층이경제적여건에의해선택한주거지를중심으로공간적밀집현상을분석하고자한다. 또한취약계층의대표적인주거지로인식되고있는쇠퇴한주거지역과의공간상관관계를시계열적으로분석함으로써, 취약계층주거지의특성및공간분포변화를분석하고자한다. 2. 연구의범위및방법연구의분석대상은사회경제적으로취약한계층이라볼수있는국민기초생활보장수급자를대상으로하였다. 공간적분석범위는서울시를대상으로해당통계자료가생산 공개되는단위인행정동을기준으로하였다. 국민기초생활보장수급자통계가최초로공개된시점은 2006년이지만누락된정보가많고행정경계변화로 MAUP 1) 문제를발생시킬수있어행정동 1) MAUP(Modifiable Areal Unit Problem) 는 공간단위수정가능성의문제 또는 가변적지역단위문제 등으로번역할수있으며, 1979 년 Openshaw 와 Taylor 에의해생성된용어임. 연구에이용되는공간단위의선택은기본적으로작위적 ( 공간단위의수정가능성 ) 이며연구결과는이용된공간단위의크기에영향을받아분석결과가달라질수있기때문에 ( 김영표, 임은선 2004), 행정구역개편과같이공간단위가시간의흐름에따라변화하여속성자료와공간자료의불일치문제가발생할수있음 ( 오충원 2002, 191). 100 국토연구제 102 권 ( )

103 의수가 424개로같은두시점인 2010년과 2015년의변화를분석하였다. 본논문에서는기초생활수급자를대상으로하되, 영구임대, 50년임대, 국민임대, 매입임대주택등공공에서공급한임대주택에거주하는수급자를제외하였다. 이를통해공공정책에의한영향을배제하고자하였으며, 민간임대주택에거주하는기초생활수급자를분석대상으로설정하고, 공간자기상관분석을통해밀집현상을분석하였다. 또한, 취약계층이쇠퇴한주거지역으로밀집되고있는지를파악하기위해 30 년이상경과한단독 다가구, 다세대, 연립주택이밀집한노후저층주거지와의시계열공간자기상관분석을통해변화분포를분석하였다. II. 이론적고찰 1. 도시공간의분리현상도시내에서의계층별주거의입지에관해서는고전적입지모델인생태학적관점과신고전주의경제학적관점, 관리주의적관점으로나누어볼수있다 ( 배순석, 전성제 2006, 192; 주희선 2015, 13; 양승호 2014, 10). 계층간의주거지분리에대해생태학적관점에서 Park(1921) 는주거지분리는도시가성장하는과정에서자연스럽게나타나는진화과정으로보았으며, Burgess(1924) 는중심지모형을통해사회, 경제적격차로인해토지이용에따른편차가발생하고, 도심을중심으로한동심원형태로지역이분화되어계층간주거지분화가발생한다고보았다 ( 김혜진 2017, 39). 신고전주의경제학적관점에서는도심으로부터주거지까지의교통비용에의해지대가결정되므로저소득층일수록고밀화된도심과도심인근에거주하고, 고소득층일수록도시외곽의넓은주택에거주한다고 설명하였다 (Miron 1982; 배순석, 전성제 2016, 193). 앞서살펴본도시생태학적관점과신고전주의경제학적관점이주택수요측면에서의접근이었다면, 관리주의적관점은주택을포함한도시자원자체가불균등하며, 이를분배하는과정에서도시관리자들이정책개입을통해중요한자원분배자의역할을하게되며, 주거지분리에도영향력을미친다고보았다 ( 윤인진 1998; 양승호 2014, 12). 2. 선행연구검토및연구의차별성사회 경제적취약계층분포에관한연구는크게공간적분포패턴과분리정도를측정하는연구와지역간격차를분석한연구가있다. 김걸 (2017) 은서울시를대상으로빈곤층의공간분포패턴변화분석을하였으며, 이를위해국민기초생활보장수급자와가구수통계를활용하였다. 국지적공간자기상관분석을통해주거지격리현상이발생하고있고빈곤층의비율이계속높게유지되고있음을밝혀빈곤이 대물림 되고있다고설명하였다. 박윤환 (2013) 은전국 251개시군구의대상으로빈곤층주거지의분리정도를측정하였으며, 국민기초생활보장수급자통계자료를활용하였다. 균일성, 고립성, 군집성, 집중성차원에서개발된지표를활용하여분리정도를측정하고그결과로지도와도표를통해공간적집중현상을시각적으로보여주었다. 지역간불균형및격차를나타내는연구중배순석, 전성제 (2006) 는서울시자치구를대상으로비유사성지표와엔트로피지표를활용하여공공임대주택, 연건평 9평미만주택, 건축후 30년이상경과한노후주택, 단독화장실부엌시설미비주거의네가지요소를통해저소득층주거의입지와공간적분리패턴을분석하였다. 송시형 (2009) 은전국의주택보급률과 서울시취약계층밀집지역분포와주거지특성분석 101

104 소득계층별점유형태, 부동산불평등지니계수를통해빈부격차를설명하였으며, 서울시자치구별전문직종종사자, 교육수준, 주택가격상승분을비교하여지역별격차를설명하였다. 김은래, 이명훈 (2004) 또한지역의불균형과격차현상의원인을분석하기위해서울시의인구, 활동, 시설및토지등에대한도시공간특성을살펴보았다. 그외에도다양한지표를활용하여취약지역을추출한연구로이명호, 유선철, 안종욱, 신동빈 (2016) 은 GIS 공간분석을활용하여취약지역을분석하였으며, 김형용, 최진무 (2012) 는취약근린지수의공간분포에대해행정동과집계구를비교 분석하였다. 살펴본바와같이기존연구는주로빈곤층또는저소득계층의통계정보를통해취약계층의공간적집중현상을분석하거나열악한수준의차이를분석하는등사회, 경제, 교육수준의격차를통한지역간불균형에초점을맞추었다. 본논문은서울시국민기초생활보장수급자를대상으로공간적분포변화를분석한다는점에서김걸 (2017) 의연구와가장유사하다. 이때, 기존연구는전체기초생활보장수급자를대상으로분석하였기때문에임대주택에거주하는수급자를모두포함하게된다. 이때문에국지적 Moran s I 분석결과, 밀집된지역의패턴이영구임대주택이입지한지역과거의동일하게나타났다. 영구임대는대규모아파트단지로공급되며, 하나의행정동에 2~3개의단지가입지한경우, 통계수치상다른지역에비해국민기초생활보장수급자수의편차가매우크게발생하면서영구임대아파트단지가거의일치되는공간적분포가나타나게될수밖에없다. 하지만이와같은분석결과는공간적입지를결정하는데정책의영향을크게받았다고불수있어, 자발적선택에의해결정된주거지분리및주거지특성을파악하기에는부족함이있다. 따라서본논문에서는임대주택거주자를제외한기초생활수급자를중심으로그들의경제적여건이반영된주거지의공간적분포와밀집현상을파악하고자하며, 쇠퇴한주거지역과의관계를통해취약계층의주거지특성을파악해보고자한다. 이때, 시계열공간자기상관분석을통해변화정도를계량화함으로써기존연구의한계를극복하고자한다. III. 연구방법및데이터 1. 공간자기상관을통한취약계층밀집지역측정방법공간적자기상관성을측정하는방법은 Moran(1950) 이고안한 Moran s I 지수를이용한자기상관성측정방법이가장일반적으로사용된다 ( 최열, 이재송 2014, 253). 이지수는지역간인접성을나타내는공간가중행렬과인접지역간속성데이터의유사성 (Similarity) 을측정하며, 인접한지역이비슷한특성이있을때는 (+) 값을, 다른특성이있으면 (-) 값을가지게된다. 단, Moran s I 지수는해당지역내통계량의공간적자기상관성을진단할수있으나, 이것만으로는분석대상지역내의공간적분포패턴을확인할수없다. 이와같은문제를최소화하기위해 Anselin(1995) 은국지적공간자기상관 (Local Indicator of Spatial Association: LISA) 분석기법을제시하였다. 국지적모란지수 (Local Moran s I) 는공간적연관성을네가지 2) 로구분하여 2) Local Moran s I 구분 : 1 HH(High-High): 높은값이높은값과인접해있음 2 LL(Low-Low): 낮은값이낮은값과인접해있음 3 LH(Low-High): 낮은값이높은값과인접해있음 4 HL(High-Low): 높은값이낮은값과인접해있음. 국지적인차원에서공간적군집은주로 HH, LL 유형을말하며, LH, HL 은공간적이례지역이라고볼수있음 ( 이희연, 노승철 2012). 102 국토연구제 102 권 ( )

105 시각화하여나타낼수있으며, 이를통해공간적군집의유의성과온열지점을찾을수있다. 두개의변수의관계를나타내는공간분석일경우, 이변량공간자기상관분석 (Bivariate Moran s I) 이시도된다. 이변량공간자기상관분석은기본적으로국지적공간상관성지표인 LISA에서이변량으로모델을확대적용한것이다 (Lee 2001). Lee(2001) 은이변량공간연관성을측정하기위해비공간상관계수인 Pearson s r과공간자기상관계수인 Moran s I 분석방법을결합하여비공간과공간을모두고려하는지수를개발하였다. 이는이변량의공간군집과이변량연관성에의한변수들간의유사성또는비유사성을측정하는것이며, 이변량공간상관관계모델 식은 < 식 1> 과같이정의된다 ( 황효정, 최현우, 김태림 2008; 김영호 2011). < 식 1> 본논문에서는전역적공간자기상관분석을실시하여 Moran s I 지수를통해서울시취약계층밀집정도와그변화를측정하고자하며, 국지적공간자기상관분석통해 HH유형의지역을중심으로공간적변화패턴을분석하고자한다. 또한취약계층주거지역과노후저층주거지역과의관계를이변량공간자기상관분석을통해계량화하고시기별변화의차이를분석하고자한다. 2. 변수의설정및데이터구축 우리나라에서는정확한소득자료를획득하기어려우므로사회 경제적계층간의공간적패턴을분석하는연구는본질적인한계를갖고있다. 하지만정부의공 적부조대상인국민기초생활보장수급자집단은제한적이나마빈곤의공간적집중을측정하고그현황을살펴보는데있어서매우중요한의미를지닌다 ( 박윤환 2013, 489). 국민기초생활보장수급자는 국민기초생활보장법 에따라명시된조건에미달될정도로생계가곤란한저소득층으로, 정부에서생계, 의료, 주거, 교육, 기타현물지원을받은사람이며, 소득인정액이중위소득 30~50% 이하로최저생계비에미치지는못하는취약계층을대표하는지표이다. 김걸 (2017) 과박윤환 (2013) 은국민기초생활보장수급자를빈곤층으로정의하고사회계층간의공간적분리현상을설명하는지표로활용하였으며, 조준혁, 이영성, 정해영, 곽태우 (2014) 는지역의쇠퇴정도를파악하기위해국민기초생활보장수급자수를인구 사회의쇠퇴지표로활용하기도하였다. 이처럼국민기초생활보장수급자통계자료는저소득 빈곤층을나타내며, 도시내의지역쇠퇴를진단하는대표적인지표로활용된다. 이에본논문에서는국민기초생활보장수급자를사회계층간의차이를나타낼수있는취약계층으로정의하고분석대상으로활용하였다. 분석에활용하기위한데이터구축에앞서공간적분포를파악하기위해국민기초생활보장수급자의통계량을 3가지차원으로구분하여살펴보았다. 먼저, 총량분포를보는방법으로취약계층의거주가절대적으로많은지역을파악할수있다. 하지만이는지역의거주인구가많을경우, 취약계층이밀집되어있다고보기힘들다. 둘째, 행정동면적대비취약계층의거주비율로공간적밀도를파악할수있다. 이는인구분포의밀집정도를파악하는가장간편하면서보편적인방법이다. 이경우에는공동주택과같이공급된취약계층주거지는높게나타나므로, 저층주택이주로입지한지역의취약계층의주거지와의편차가크게나타날수있다. 마지막으로지역에거주하는 서울시취약계층밀집지역분포와주거지특성분석 103

106 인구대비취약계층인구비율로공간적으로분포된취약계층의정도를파악할수있다. 이는지역의인구규모대비취약계층의인구비율을나타냄으로써, 다른계층대비취약계층의밀집정도를나타낼수있다. 이에본논문에서는사회적인구의계층간차이를반영할수있도록인구규모가반영된취약계층의밀집정도를분석데이터로설정하였으며, 주택의거주단위를산정할수있는가구수및세대수 3) 비율을활용하였다. 본논문의주요대상은국민기초생활보장수급자중공공부문의임대주택에거주하는수급자를제외함으로써공공의정책적영향을배제하고일반주거지및민간임대주택에거주하는수급자를대상으로한다. 이를위해공공임대주택에거주하는수급자세대수를산출할필요가있다. 국토교통부통계누리 4) 에의한각연도임대주택재고현황에따르면, 서울시에공급된공공부문임대주택은 2010년기준으로총 159,611세대이며, 2015년기준으로총 200,227세대로나타났다. 5) 공급주체별로구분하면 2010년기준 LH가 22.3%, SH가 77.7% 로나타났으며, 2015년에는 LH가 21.3%, SH가 78.6% 를차지하는것으로나타났다. 임대주택의유형별재고량을보면, 대부분의영구임대및공공임대주택이 2010년이전에공급되었음을알수있고, 2010~2015년사이공급된다양한임대주택재고량의증가는저소득층위주에서다양한계층으로확대되고있음을보여준다 (<Table 1>, <Figure 1> 참조 ) Year Total LH SH Total LH SH Permanent Lease 45,998 23,628 22,370 46,592 23,920 22, Yeares Public Rental Housing Redevelopme nt/ Renewal Table 1 _ Public Rental Housing in Seoul 20,269 2,837 17,432 20,191 2,759 17,432 Note Including Number of Redevelopment Housing by LH 51,802-51,802 58,763 58,763 Local Government National Rental Construction 14, ,347 22,636 1,908 20,728 Housing Purchase 11,127 8,315 2,812 21,423 11,054 10,369 Happily Housing Etc. 15, ,244 30,622 2,962 26,613 (unit: households, %) 35, ,007 42, ,384 Total 159, , % 77.7% 21.3% 78.6% Note: Etc.(10 years, 5 years, Employees, Long-term Rental Housing). Source: (accessed September 13, 2019). 3) 국민기초생활보장수급대상은가구단위로산출되며임대주택공급은세대단위로산출되기때문에동일한단위로가정하고산출함. 4) (2019년 9월 13일검색 ). 5) 국토교통부통계누리임대주택재고현황기준으로 2010년공공부문임대주택총세대수는총 159,611세대이며, 2015년총 235,451세대임. 그러나 2010년세대수는전세임대 (92,158세대) 를제외한수치이며, 2015년세대수는전세임대 (35,224세대) 를포함한수치임. 전세임대는금전적지원으로민간임대주택에거주하는방식이므로대상에서제외하였으며, 50년임대주택중지자체공급대상은 SH 공급총세대수에포함하여산출함. 104 국토연구제 102 권 ( )

107 Figure 1 _ Public Rental Housing in 2015 Source: Modifed. 임대주택거주수급자파악을위해임대주택의세대별공급대상정보를구축하였다. 먼저, SH 공급대상의경우, 임대주택유형별, 세대별자격정보와매입임대의입주시기및자격정보 6) 를통해임대주택유형별, 시기별수급자를추출하였다. LH 공급대상은마이홈포털내의임대주택유형별공급세대정보를기반으로하였으며, 매입임대의경우, 해당필지별토지소유권이전정보를통해매입시기를파악하였다. 이때, LH 임대주택세대별입주자자격요건을파악할 수없었으나, SH의공급량이전체임대주택이 77.7~78.6% 를차지하고있고, 동일유형의임대주택일경우, 수급자비율이유사한점을감안하여 SH 공급세대별입주자자격요건의비율을적용하여산출하였다. 단, 유형별공급된시기에따라수급자비율 7) 의차이가있어, 이를적용하여산출하였다. 취약계층과쇠퇴한주거지역과의공간적상관성을분석하기위한지표로는주거지의물리적쇠퇴를나타내는대표적인지표인 30년이상경과한건축물비 6) SH 임대주택관리및 SH 매입임대주택자료 ( 내부자료 ) 를활용 (SH공사 2016a; 2016b) 하였으며, 세대별신청자격기준으로 기초생활보장수급자 를구분함. 7) 임대주택유형별수급자비율은영구임대 (2000~2010년: 81.9%, 2011~2015년 : 73.3%), 매입임대 (2000~2010년: 42.98%, 2011~2015년 : 36.90%) 로시기별차이가나타났으며, 그외임대주택의경우큰차이가없었음 (50년임대 1.4%, 재개발 10.1%, 국민임대 0.2%, 행복주택 8.0%, 평균 ). 서울시취약계층밀집지역분포와주거지특성분석 105

108 율을적용하되, 기반시설등이열악한주거지역을나타내기위해단독 다가구, 다세대, 연립주택을대상으로하는노후저층주택비율을산정하였다. 활용된지표의산출식은 <Table 2> 와같다. IV. 취약계층거주밀집특성 1. 취약계층거주분포현황 2010년기준, 서울시기초생활보장수급자가구 8) 는 123,858호이다. 이중 39,839가구는임대주택에거주하고있으며이는전체수급가구중 32.34% 에해당한다. 2015년기준으로기초생활보장수급자가구는총 163,107호로전체수급가구중 31.02% 인 50,595가구 가임대주택에거주하고있는것으로나타났다 (<Table 3> 참조 ). 임대주택의공급재고량이증가하고임대주택에거주하는수급자가구도늘어나고있다. 하지만기초생활수급자가구가계속증가함에따라 2010년, 2015 년의민간임대주택에거주하는수급자는각각 67.7%, 69% 에해당하며, 이들은여전히경제적여건에맞는쇠퇴한주거지역에거주할것으로예측된다. 수급자의공간적분포와변화를파악하기위해행정동별통계수치를시각화하였다. 공간분포의시각화는구간값의설정에따라다르게인식될수있으며, 주로사용되는 Natural Breaks 방식이나등간격, 등도수분포는구분값을주관적으로설정해야하는한계를지니고있다. 이에본논문에서는구분값을전체 Table 2 _ Index and Calculation Formula Classification Calculation Formula Rate of Vulnerable Groups Total Vulnerable Groups / Total People Administrative District 100 (Total Vulnerable Groups Households Total Public Rental Housing Households) / Rate of Residents of Private Rental Housing Total Households 100 Rate of Deteriorated Low-rise Residental Area Over 30 year and Low-rise Housing / Total Households 100 Source: Year Household (person) A 123,196 (201,871) 163,107 (246,580) Table 3 _ Statistics of Variables (unit: households, %) Vulnerable Groups Public Rental Housing Residents of Private Rental Housing Rate of Residents of Public Rental Housing B/A*100 Total Vulnerable Groups households B Households (A-B) Rate of Residents of Private Rental Housing (A-B)/A* % 117,111* 39,839 83, % 31.02% 168,925** 50, , % Note: * 2010: Permanent Lease(40,946), Public Rental Housing(17,195), Redevelopment/renewal(38,754), National Rental Housing_ Construction(8,753), Purchase(11,463). ** 2015: Permanent Lease(44,124), Public Rental Housing(20,043), Redevelopment/renewal(59,501), National Rental Housing_ Construction(23,439), Purchase(21,818). 8) 서울시국민기초생활보장수급자 ( 동별 ) 통계자료중시설수급자와기타로분류되어행정동구분이불가능한대상은제외함 년의광진구, 양천구, 영등포구, 중랑구의행정동별통계량이누락되어구별총량을기준으로 2011 년행정동별통계량비율을적용하여산출함. 106 국토연구제 102 권 ( )

109 값의범위내에서사분위수로구분하고데이터의중심값과양극단의값을파악할수있는 Box Plot 방식을공간분포에적용하여나타내었다. 이때, 연도별로적용할시해당연도의사분위수의범위값이바뀌게되어시기별증감의차이를확인할수없어, 2010년의사분위수를적용하되, 2015년은 2010년의기준으로적용하여시기별변화의차이를확인하였다. 먼저, 전체취약계층의거주비율분포를살펴보면, 2010년에가장높은지역은수서동으로나타났으며, 가양2동, 등촌3 동, 가양3동의강서구일대와번3동, 월계2 동일대의강북구, 도심의남영동과회현동, 종로 1,2,3,4가동일대등으로나타났다. 2015년에는기존분포패턴에서동북권과그일대를중심으로전체적으로높아진것을알수있다 (<Figure 2> 참조 ). 민간임대주택에거주하는취약계층분포비율은전체취약계층의분포패턴과유사하게나타나지만, 비율이가장높은곳은남영동과회현동으로나타났다. 이는 <Figure 1> 과같이임대주택공급현황분포에서알수있듯이, 도심권에공급된임대주택이거의없음을감안하면, 대부분민간임대주택에거주하는 Figure 2 _ Spatial Distribution of the Vulnerable Groups Vulnerable Groups in 2010 Vulnerable Groups in 2015 Figure 3 _ Spatial Distribution of the Residents of Private Rental Housing Residents of Private Rental Housing in 2010 Residents of Private Rental Housing in 2015 서울시취약계층밀집지역분포와주거지특성분석 107

110 Figure 4 _ Spatial Distribution of the Deteriorated Low-rise Residental Area Deteriorated Low-rise Residental Area in 2010 Deteriorated Low-rise Residental Area in 2015 취약계층인것으로파악할수있다. 이외에영등포동과가리봉동일대는공급된임대주택이거의없는것으로보아, 민간임대주택에거주하는수급자가많은지역인것으로파악된다. 반면, 전농3 동과월계3 동은공급된임대주택에거주하는수급대상을제외하였음에도불구하고여전히높은지역으로나타남으로써, 해당지역의민간임대주택에수급자가많이거주하고있음을예상해볼수있다. 전체적인분포패턴을살펴보면, 일부대규모임대주택공급지역을제외한지역을제외하고, 전체수급가구수의거주분포와민간임대주택에거주하는수급자분포패턴과매우유사하게나타나고있다. 노후저층주택비율이높은지역은대체로도심부에밀집되어있으며, 그중삼청동은 2015년기준노후 저층주택이 62.1% 로가장많은지역으로나타났다. 이외에도신길4 동 (37.9%), 가리봉동 (35.1%), 광희동 (32.9%), 장위동 (32.7%), 성북동 (30.7%) 등도심권에인접한동북권일대를중심으로노후저층주택비율이높은지역으로나타났다. 또한금천구, 영등포구, 관악구일대가대체로높은것을알수있으며, 2015년에는더욱증가하는것을알수있다. 분석에활용된기초통계는 <Table 4> 와같다. 2. 공간자기상관분석을통한취약계층밀집의변화분석앞서살펴본바와같이, 전체수급자와민간임대주택에거주하는수급자의비율의분포패턴은유사하게나타났으나, 통계적으로유의미한분포양상임을파 Variables Year Num Mean STD Min Max Rate of Vulnerable Groups Rate of Residents of Private Rental Housing Rate of Deteriorated Low-rise Residental Area Table 4 _ Descriptive Statistics (unit: %, percent) (25.803) (24.912) 국토연구제 102 권 ( )

111 악하기위해공간자기상관분석을실시하였다. 먼저, 민간임대주택에거주하는수급자의밀집분포를파악하기전에, 공간적비교를위해전체취약계층을대상으로밀집분포를분석하였다. 이때, 유의한공간적밀집분포확인을위해국지적공간자기상관분석을실시하였으며, 결과는 <Figure 5> 와같다. 전체수급자를대상으로한공간자기상관분석결과, HH유형으로나타난지역은앞서언급한바와같이, 강서구의가양2 동, 등촌3 동, 가양3 동일대와강북구의번3동, 월계2 동을비롯한일대, 강남의수서일대등대규모영구임대주택단지가입지한지역과거의일치되어나타나고있음을알수있으며, 2010년과 2015년의비교를통해서도매우유사한분포를나타낸다. 이는공공정책에의한대규모임대주택단지에의한영향이매우크게나타나고있음을의미한다. 이에본논문에서는공공임대주택영향을제외함으로써, 경제적여건에의해민간임대주택에거주하는취약계층을대상으로공간적밀집현상의발생과그정도를분석하였다. 먼저, 전역적공간자기상관분석결과, 2010년의 Moran s I 값은 0.126, 2015년에는 0.207로증가하는 Table 5 _ Univariate Moran s I Statistics of the Residents of Private Rental Housing Classification Moran s I Z-score P-value 것으로나타났으며, 모두정적인공간적자기상관을갖는것으로나타났다. 통계적으로도유의확률 수준에서 z-value 값이각각 4.643, 7.419로유의한결과를나타내었다 (<Table 5>, <Figure 6> 참조 ). 공간자기상관의정도를나타내는 Moran s I 지수가증가하고있는것으로보아, 민간임대주택에거주하는취약계층또한특정한지역을중심으로공간적으로밀집되어있으며, 그현상이더욱심화하고있는것으로해석할수있다. 국지적공간자기상관분석결과로나타난 <Figure 5> 와같이공간적분포패턴을살펴보면, HH유형의지역이 2010년에는중구와용산구경계부에위치한회현동과남영동일대를중심으로나타나며, 동서축을따라성북동의월곡2동및석관동일대로이어지고있다. 또한서남권역으로는영등포구의신길동일대를비롯하여금천구의독산동, 시흥동일대로이어 Figure 5 _ LISA Analysis of the Vulnerable Groups LISA Cluster Map in 2010 LISA Cluster Map in 2015 서울시취약계층밀집지역분포와주거지특성분석 109

112 Figure 6 _ Univariate Moran s I and LISA Analysis of the Residents of Private Rental Housing Univariate Moran's I in 2010 LISA Cluster Map in 2010 Univariate Moran s I in 2015 LISA Cluster Map in 2015 지고있음을알수있다. 2015년과비교해보면, 기존 HH 유형은도심권을중심으로동북권과서남권방향으로의대각선축의형태가유지되고있으며, 서남권지역의 HH유형은다소축소되고, 동북권의중랑구면목동일대에 HH유형으로추가되어나타나고있다. 이러한지역의분포패턴과임대주택유형별현황분포도인 <Figure 1> 과대조해보면서남권과도심권동북권에이르는대각선축으로임대주택공급분포가다른지역에비해다소부족한패턴임을파악할수있다. 반면, 취약계층의밀도가낮으며, 주변지역도낮은값으로나타나는 LL유형은기존의대량공급된임대주택단지로강북구월계동일대, 서남권의가양동과 등촌동일대를중심으로나타났으며, 이는임대주택거주수급자를제외한것이반영된것으로보인다. 이외의강남구와송파구일대의경우, 대규모재건축단지가입지한지역을중심으로 LL유형이나타나고있으며, 해당지역이높은주거비용이소요되는지역인만큼취약계층의거주가어려운지역임을나타내고있다. 3. 취약계층밀집과노후저층주택과의공간자기상관관계분석본논문의주된분석대상은민간임대주택에거주하 110 국토연구제 102 권 ( )

113 는취약계층으로서, 이들의공간적분포는전통적으로이들의주거지로여겨지는노후저층주택과의밀접한관계가예상된다. 이에따라본논문에서는통계적으로유의미한공간적상관성과분포특성을파악하기위해, 공간자기상관분석을두변수간의관계로확장한개념인이변량공간자기상관분석으로실시하였다. 먼저, 이변량전역적자기공간상관분석결과, <Table 6> 과같이취약계층주거지역과노후저층주거지와의공간적상관성을나타내는 Bivariate Moran s I 값은 2010년에는 0.157, 2015년에는 0.183으로다소증가하는것으로나타났다. 모두정적인공간적자기상관을갖는것을나타내었으며, 유의확률 수준에서 z-value 값이각각 7.359과 8.179로매우유의한결과를나타내었다. 이는, 민간임대주택에거주하는취약계층주거지와노후저층주거지와의공간적으로상관성이높아지고있음을의미한다. 이러한두변수의공간적자기상관성을지역분포로파악하기위해이변량국지적공간자기상관성을나타내는 Bivariate Moran s I LISA 분석을실행하였다. 분석결과는 <Figure 7> 과같다. 2010년의 HH유형으로나타난지역은중구, 종로구, 성북구와성동구일대에집중되어있고, 2015년의영등포및금천구일부지역이새로운 HH유형의지역으로나타났다. 또한, 도심과연계된동측으로의일부지역은 HH유형의지역에서제외되었는데, 이는왕십리뉴타운, 전농 Table 6 _ Bivariate Moran s I Statistics of the Residents of Private Rental Housing and Deteriorated Low-rise Residential Area Classification Moran s I Z-score P-value 답십리재정비촉진구역의사업추진에의한영향이반영된것으로보인다. 2010년과비교하여, 2015년에는전역적공간자기상관성을나타내는 Bivariate Moran s I 값이증가하고있음에도불구하고, HH유형의지역이 42개지역에서 38개지역으로축소된것으로나타났는데, 이는기존의낙후지역으로의취약계층의공간적집중이심화하고있는것으로해석할수있다. 이러한결과를통해취약계층은주택가격이저렴한노후 낙후된지역에거주하고있다는사회적통념에부합함을공간적통계분석을통해입증하였다. 또한해당지역이기존도심지역을중심으로나타나고있음을알수있다. 개발사업등으로기존의취약계층이부담가능한주택이철거되거나줄어들면, 이들은더욱낙후된지역으로밀집되며, 해당지역의취약계층밀집정도가더욱심화될수있는현상을설명할수있다. V. 결론최근우리사회는저성장기에들어서면서경제적양극화가심화되고있으며, 소득계층간의주거지분리에따른공간적양극화발생이우려되고있다. 이에본논문에서는사회적계층간의주거지분리에대한고전적이론을바탕으로공공의정책적개입에따른영향을배제하고경제적여건에따른자연발생적인취약계층의공간적밀집현상을분석하고자하였다. 공간분석단위는서울시행정동을대상으로하였으며, 취약계층을대표하는기초생활수급자를대상으로하되, 정책에의해집단화된공공임대아파트단지와매입임대주택에거주하는임대주택거주자를제외함으로써, 이들을민간임대주택에거주하는기초생활수급자로정의하고분석의주요대상으로설 서울시취약계층밀집지역분포와주거지특성분석 111

114 Figure 7 _ Bivariate Moran s I and LISA Analysis of the Residents of Private Rental Housing and Deteriorated Low-rise Residential Area Bivariate Moran s I in 2010 LISA Cluster Map in 2010 Bivariate Moran s I in 2015 LISA Cluster Map in 2015 정하였다. 또한이들의주거지로인식되고있는쇠퇴한주거지로의밀집과그변화를알아보기위해단독및다세대 다가구밀집지역을주거환경이열악한노후저층주거지로설정하고민간임대주택에거주하는기초생활수급자와의관계를분석하였다. 취약계층의공간적밀집현상을분석하기위한방법으로공간자기상관분석을통해그정도를측정하고 2010년과 2015년의비교를통해변화분포를살펴보았다. 또한취약계층거주지와노후저층주거지와의이변량공간자기상관분석을통해공간적상관성변화 를분석하였다. 먼저, 공공부문의임대주택에거주하는수급자는약 30% 내외로이를제외한민간임대주택에거주하는기초생활수급자는약 70% 에이르는것으로나타났다. 민간임대주택에거주하는기초생활수급자를대상으로그들의공간적밀집과그정도의변화를측정한결과, 공간적상관정도를나타내는전역적 Moran s I 값이 2010년은 0.126, 2015년에는 0.207로시간이지남에따라증가하는것으로나타났다. 또한, 국지적공간자기상관분석결과, 취약계층의밀집분포는도심부를중심으로동북권과서남권의대각선축으로 112 국토연구제 102 권 ( )

115 이어지는밀집분포를나타내었다. 취약계층거주지의특성을파악하기위해이들의대표적인주거지로인식되고있는노후저층주거지역과의이변량공간자기상관분석을실시한결과, 전역적공간자기상관분석인 Bivariate Moran s I 값은 2010 년에는 0.157, 2015년에는 0.183으로다소증가하는것으로나타났다. 이러한결과는취약계층주거지와노후저층주거지와의공간적상관성이점차증가하고있음을보여주며, 이에반해유의한 HH의유형의행정동이줄어들고있음을감안하면, 특정지역으로의밀집이더욱심화되고있는것으로이해될수있으며, 특히도심의개발사업이발생하는지역주변인것을확인할수있다. 이와같이민간임대주택에거주하는취약계층의공간적분포와주거지특성을분석한연구결과에따른시사점은다음과같다. 첫째, 서울시는사회계층간의주거지분리및취약계층의공간적밀집현상이발생하고있으며, 그현상이심화되고있다. 이는소득에따른경제적양극화가공간적으로반영되는것으로지역격차와특정지역에대한낙후이미지를고착화할수있다는점에서, 도시의공간적격차를완화하기위한거시적차원의계획이필요함을시사하고있다. 둘째, 민간임대주택에거주하는취약계층의공간적밀집분포는동북권과도심권, 그리고서남권을축으로나타나고있음을파악하였다. 해당지역은기존임대주택공급이많지않은지역으로, 이와같은기존의취약계층의지역적공간수요를파악하여, 지역맞춤형임대주택의공급이이루어질필요가있다. 셋째, 취약계층의밀집지역과노후저층주거지역의공간적상관성은점차증가하는것으로나타났다. 특히개발사업등에의해기존노후주택이철거 감소함으로써취약계층과저소득층이부담가능한주택이 사라지고, 더낙후된지역으로이동시키는결과를불러오게되어공간적분리를더욱강화할수있다. 따라서낙후지역을개발할경우, 기존취약계층이거주할수있는거주공간을확보할필요성이있다. 본논문의연구결과는일반적으로인식하고있는취약계층의주거지특성과크게다르지않으나기존연구에서는다루지않은, 민간임대주택에거주하는기초생활수급자를대상으로하였다. 이를통해공공의정책적영향을배제한경제적여건에따른취약계층의주거지선택과그선택으로인한밀집현상을규명하고자하였다. 또한, 노후저층주거지와의관계를공간분석을통해계량화하고그변화를측정함으로써, 사회통념적으로인식하고있는취약계층이쇠퇴지역으로의밀집현상을실증적분석을실행했다는점에서의의가있다. 최근우리사회는사회적 경제적양극화가심화되고있고, 이러한시기의취약계층의밀집과주거지분리현상은주거수준의양극화와공간적불균형을발생시킬수있으며계층간의갈등을증폭시킬가능성이있음을고려해볼때, 도시정책적으로대처방안이필요한사안임을보여준다. 마지막으로본논문의한계는다음과같다. 먼저, 구득가능한자료의한계로정확한수급자대상을추출하지못하였으며, 행정동단위의공간분석으로미시적인분석과해석은다소무리가있었다. 또한이론적배경에따른밀집현상을규명하는것에집중하여발생원인및인과관계는밝히지못하였다. 향후세밀한지역단위의분석을통해취약계층의공간적집중현상을분석하고, 모니터링하여변화를파악할필요가있으며, 인구 사회적특성과경제 물리 환경적측면등다양한여건을고려하여취약계층의밀집원인등에대한분석이이루어질필요가있다. 서울시취약계층밀집지역분포와주거지특성분석 113

116 참고문헌 1. 국가통계포털. 2010년, 2015년주택총조사, 건축연도별주택현황통계. o?menuid=m_01_01&vwcd=mt_ztitle&parmtabid= M_01_01#SelectStatsBoxDiv (2018년 3월 2일검색 ). Korean Statistical Information Service. 2010, 2015 Housing Census, Housing Units by Type of Housing Units, Total Floor Space and Year of Construction. tatisticslistindex.do?menuid=m_01_01&vwcd=mt_ztit LE&parmTabId=M_01_01#SelectStatsBoxDiv (accessed March 13, 2018). 2. 국토교통부통계누리. 2010년, 2015년임대주택재고현황. FormId=840&hDivEng=&month_yn= (2019년 9월 13일검색 ) Molit Statistics System. Public Rental Housing stock in 2010, Id =37&hFormId=840&hDivEng=&month_yn= (accessed Se ptember 13, 2019). 3. 김걸 서울시빈곤층의공간분포패턴변화분석. 한국도시지리학회지 20권, 1호 : Kim Geol An analysis on the changes in spatial distribution patterns of the poor classes in the city of Seoul. Journal of the Korean Urban Geographical Society 20, no.1: 김영표, 임은선 GIS기반공간분석방법론적용연구. 안양 : 국토연구원. Kim Yeongpyo and Lim Eunseon Applications of GIS-based Spatial Analysis Methodology. Anyang: Korean Reasearch Institute for Human Settlements. 5. 김영호 공간네트워크의이변량공간상관관계를이용한서울시자전거와버스대중교통의연계가능성분석. 한국도시지리학회지 14권, 3호 : Kim Yeongho A study about bike and bus connection using a bivariate measure of spatial correlation. Journal of the Korean Urban Geographical Society 14, no.3: 김은래, 이명훈 서울시의지역간도시공간특성에관한연구. 국토계획 39권, 5호 : Kim Eunrae and Lee Myeonghun A study on the regional urban spatial characteristics in Seoul. Journal of the Korean Planners Association 39, no.5: 김형용, 최진무 취약근린지수의공간적분포. 국토지리학회지 46권, 3호 : Kim Hyeongyong and Choi Jinmu Spatial distribution of neighborhood deprivation index for Seoul. The Geographical Journal of Korea 46, no.3: 김혜진 수도권외국인의공간적분포특성과거주지분리. 박사학위논문, 성신여자대학교. 10. Kim Hyejin Spatial Distribution of Foreign Immigrants and Their Residential Segregation in Seoul Metropolitan Area. Ph.D. diss., Sungshin University. 11. 마이홈포털. 기존임대주택찾기. /hws/portal/sch/selectrentalhouseinfolistview.do (2018 년 4 월 30일검색 ). My-home Portal. Search for existing rent house. myhome.go.kr/hws/portal/sch/selectrentalhouseinfolistview. do (accessed April 31, 2018). 12. 박윤환 빈곤층주거지의공간적집중에대한연구 : 기초생활수급인구를중심으로. 도시행정학보 26집, 4호 : Park Yunhwan A study of spatial concentration of the poor residence: Based on basic livelihood security recipients. Journal of the Korean Urban Management Association 26, no.4: 배순석, 전성제 서울시저소득계층주거의입지현황과공간적분리패턴에관한연구. 국토연구 51권, 4호 : Bae Sunseok and Jeon Seongje Analysis of the spatial distribution and segregation pattern of low-income housing in Seoul. The Korea Spatial Planning Review 51, no.4: 서울연구원 저성장시대서울의도시정책을말하다. 서울 : 한울아카데미. Seoul Institute New Paradigm for Seoul s Urban Policy in the Low Growth Era. Seoul: Hanul Academy. 15. 서울열린데이터광장. 서울특별시 2010년, 2015년국민기초생활보장수급자. do?infid=10113&srvtype=s&servicekind=2&currentpageno =1&searchValue=&searchKey=null (2019년 6월 5일검색 ). Seoul Open-data Platform. 2010, 2015 Recipient of national basic livelihood guarantees. datasetview.do?infid=10113&srvtype=s&servicekind=2& currentpageno=1&searchvalue=&searchkey=null (accessed 114 국토연구제 102 권 ( )

117 June 5, 2019) 서울특별시 2010년, 2015년주민등록인구. data.seoul.go.kr/datalist/datasetview.do?infid=10043&srvt ype=s&servicekind=2&currentpageno=1&searchvalue= &searchkey=null (2019년 6월 5일검색 ) , 2015 Resident Registration Population in seo ul &srvType=S&serviceKind=2&currentPageNo=1&searc hvalue=&searchkey=null (accessed June 5, 2019). 17. 송시형 주거지현황을통해본주거공간의양극화현상. 극동사회복지저널 5권 : Song Sihyeong Socioeconomic polarization and dwelling places in Korean society. Journal of Far East Social Welfare 5: 양승호 도시에서의소득수준에따른거주지분리가거주민의사회통합의식에미치는영향. 박사학위논문, 서울대학교. Yang Seungho Effects of Income Segregation in Urban Residential Areas on Residents Sense of Social Cohesion. Ph.D. diss., Seoul National University. 19. 오충원 가변적공간단위의문제를고려한지가변동의시공간분석. 한국공간정보학회지 10권, 2호 : Oh Chungweon Spatia-temporal analysis of land price variation considering modifiable area unit problem. Journal of the Spatial Information Research 10, no.2: 윤인숙 매입임대주택정책의필요성. 도시와빈곤 67 호, 서울 : 한국도시연구소. Yoon Insuk The necessity for purchased rental housing policy. Urbanity & Poverty 67, Seoul: Korean Center for City and Environment Research. 21. 윤인진 서울시의사회계층별거주지분화형태와사회적함의. 서울학연구 10권 : Yoon Injin A case study of Seoul: Residential differentiation by class and its social implications. Journal of the Seoul Studies 10: 이명호, 유선철, 안종욱, 신동빈 GIS 공간분석을활용한사회취약지역의분석에관한연구. 지형공간정보 24권, 4호 : Lee Myeongho, Yoo Seoncheol, Ahn Jonguk and Shin Dongbin A study on analysis of the social vulnerable areas using GIS spatial analysis: Focusing on local governments in Seoul Metropolis. Journal of Korean Society for Geospatial Information Science 24, no.4: 이희연, 노승철 고급통계분석론. 파주 : 법문사. Lee Heeyeon and Noh Seungchul Advanced Statistical Analysis. Paju: Bobmunsa. 24. 조준혁, 이영성, 정해영, 곽태우 쇠퇴한지역에사는사람들은건강도쇠퇴할까? 국토계획 49권, 6호 : Cho Junhyeok, Lee Yeongseong, Jung Haeyeong and Gwak Taewoo Does a health of people living in declined regions also decline? Journal of the Korean Planners Association 49, no.6: 주희선 주거이동으로인한저소득층밀집의외부효과에대한연구. 박사학위논문, 서울대학교. Joo Heesun A Study on the External Effects of Low-income Group Concentration by Residential Migration. Ph.D. diss., Seoul National University. 26. 최열, 이재송 공간회귀모형을이용한산업및용도지역특성과환경오염과의상관관계분석. 국토계획 49권, 1 호 : Choi Yeol and Lee Jaesong Correlates between industries and zoning characteristics and environmental pollution employing spatial regression model. Journal of the Korean Planners Association 49, no.1: 황효정, 최현우, 김태림 정량적인해양환경패턴분석을위한이변량공간연관성적용. 한국지리정보학회지 11권, 1호 : Hwang Hyojung, Choi Hyunwoo and Kim Tearim Application of bivariate spatial association for the quantitative marine environment pattern analysis. Journal of the Korean Association of Geographic Information Studies 11, no.1: SH공사. 2016a. 매입임대주택. 내부자료 Seoul Housing & Communities Corporation(SH). 2016a. Management register of purchase rent housing. Unpublished manuscript b. 임대주택관리. 내부자료 b. Management register of rent housing. Unpublished manuscript. 30. Anselin, L Local indicators of spatial association-lisa. Geographical Analysis 27, no.2: Burgess, E. W The growth of the city: An introduction to a research project. The American Sociological Review 18: 서울시취약계층밀집지역분포와주거지특성분석 115

118 32. Duncan, O. D. and Duncan, B A methodological analysis of segregation indexes. American Sociological Review 20, no.2: Jargowsky, P. A Poverty and Place: Ghettos, barrios and the American City. New York: Russell Sage Foundation. 34. Miron, J. R Housing in Postwar Canada: Demographic change, household formation and housing demand. Montreal: McGill-Queen s University Press. 35. Lee Sangil Developing a bivariate spatial association measure: An integration of Pearson s R and Moran s I. Journal of Geographical Systems 3, no.4: Massey, D. S. and Eggers, M. L The ecology of inequality: Minorities and the concentration of poverty American Journal of Sociology 95, no.5: Park, R. E. and Burgess, E. W Introduction to the Science of Sociology. Chicago: The University of Chicago Press. 38. Sampson, R. J., Raudenbush, S. W. and Earls, F Neighborhoods and violent crime: A multilevel study of collective efficacy. Science 277, no.5328: 논문접수일 : 심사시작일 : 심사완료일 : 요약 주제어 : 취약계층거주밀집지역, 민간임대주택거주자, 공간자기상관분석, 노후저층주거지역, 공간적양극화 본논문은서울시를대상으로공간자기상관성을고려하여취약계층의거주밀집과주거지역특성을분석하고자하였다. 본논문에서는경제적여건에따른취약계층의거주밀집현상을분석하기위해기초생활보장수급자중임대주택에거주하는수급자를제외한민간임대주택거주자를대상으로하였다. 분석결과, 취약계층의공간적밀집현상이발생하고있음과동시에, 시간이지남에따라밀집도가증가하는것으로나타났다. 또한이들의대표적인주거지로인식되고있는노후저층주거지역과의공간자기상관분석을실시한결과, 공간상관성이증가하고있음을확인하였으며, 도심을중심으로밀집지역이집중되고있음을확인하였다. 이러한결과는최근우리사회의 관심분야인경제적양극화현상과더불어공간적양극화가발생하고있음을보여준다. 주택정책을수립할때취약계층의지역적공간수요를파악하여임대주택을공급할필요가있으며, 정비사업시행으로인해저렴한주택이철거됨으로써이들을더욱낙후된지역으로밀집시킬수있으므로, 취약계층이거주할수있는부담가능한주택을충분히확보할필요성이있다. 비록전통적인이론에근거하여사회 경제적차이와자원분배에따른주거지분리는불가피하나이러한현상이지속될경우, 지역간불균형을가속시키고사회적갈등을유발할수있다는점에서이를완화하기위한도시공간차원의정책적노력이요구된다. 116 국토연구제 102 권 ( )

119 국토연구투고안내 ( 요약 ) 투고규정 1. 국토관련분야의연구논문을대상으로하며, 다른간행물에발표되지않은것이어야한다. 2. 투고자는본지의논문작성요령에따라작성된원고파일, 논문투고신청서와논문투고동의서를함께제출해야한다. 3. 원고는도표를포함하여국문의경우 A4용지 15매 ( 영문은 18매 ) 내외로작성한다. 4. 논문의분량및형식이본규정과상이한경우학술지편집위원회에서조정을요구할수있다. 5. 논문작성언어는한국어, 영어에한한다. 6. 논문은소정의논문작성요령에따라작성한다 ( 홈페이지참고 ). 7. 논문은수시접수하며, 접수일자는원고가편집위원회에도착된날로한다. 8. 타학술지와의복수투고를금한다. 논문심사규정 1. 심사위원은편집위원회에서위촉한다. 2. 편집위원회는접수된원고 1편당관련전문가 3인이상을심사위원으로위촉하고그결과를토대로편집위원회가게재여부를최종결정한다. 3. 논문심사의판정은심사위원의의견에따라 1차심사에서는게재가, 수정후게재, 수정후재심, 게재불가의네가지로하고심사의견이일치되지않을경우에는차악판정한다. 2차심사 ( 최종 ) 에서는게재가, 수정후게재, 게재불가의세가지로하고최악판정에따른다. 4. 게재불가판정시에는상당한이유와설명이제시되어야하며투고자는심사결과에대해 1회에한해서면으로이의를제기할수있다. 논문작성요령 1. 논문제목및초록 1) 논문제목은국문과영문을병기하며, 사사및논문특이사항은국문제목에각주로표기한다. 2) 저자명은국문과영문을병기하며각주에소속, 이메일등을국문과영문으로기재한다. 단, 저자가다수일경우제1저자및교신저자를별도표기한다. 3) 초록은국문과영문모두작성하며, 영문초록은첫페이지, 국문초록은마지막페이지에배치한다. 4) 영문초록은 150단어 (200 자원고지 3장 ) 가량, 국문초록은 400자 (200 자원고지 2장 ) 가량으로작성한다. 2. 사용어및표기 1) 국문원고의경우국문을원칙으로하며, 필요할경우한자및외래어는괄호안에병기한다. 2) 본문의번호전개는다음과같이한다. 장번호 : I. 절번호 : 1. 항번호 : 1) 목번호 : (1) 3) 표, 그림은모두영문으로표기하며, 참고문헌은원어 ( 국문, 영문을제외한외국어 ) 와영문을병기한다.

120 3. 도표및그림 1) 도표및그림은모두영문으로작성하며, 차례대로 Table 1 _ 제목, Figure 1 _ 제목등과같이 일련번호 체제로표시한다. 제목과내용은관사, 전치사를제외하고모든단어의첫글자를대문자로표기한다. 2) 모든제목은해당표및그림, 사진의상단에표기한다. 인용한도표및그림은반드시출처를밝히며, 각주는 Note:, 출처는 Source: 에이어서작성한다. 3) 표와그림은본문에서 1회이상언급되어야한다. 4. 참고주 ( 내주 ) 및참고문헌 1) 참고주는본문내주표기를원칙으로한다. 인용문에저자명이포함될경우 저자명 ( 연도, 인용면 ) 은 / 는 으로표기하고, 저자명이미포함될경우 ( 저자연도, 인용면 ) 으로표기한다. 2) 인용저자명은 4인까지기재할수있다. 예 : 김두진 (2016, 10) 은, 김두진, 송동길, 이상염, 송재일외 (2011, 25) 는 ( 김두진 2016, 10), ( 김두진, 송동길, 이상염, 송재일외 2011, 25) 3) 참고문헌은원어 ( 국문, 영문을제외한외국어 ) 와영문의병기를원칙으로한다. 4) 인용한문장에는반드시각주를달아야하며인용문헌은참고문헌에포함되어야한다. 5) 본문인용및참고문헌의기술방식은아래형식에따르되제시되지않은내용은 The Chicago Manual of Style (15th ed. 2003) 을준용한다. 6) 참고문헌작성형식 1 단행본저자명, 역할어 ( 저자가편자인경우 ). 발행연도. 표제 : 부표제 ( 영문은이탤릭 ). 역자나편저자. 판차. 권차. 발행지 : 발행처. 장희순, 방경식 부동산용어사전. 서울 : 부연사. Jang Heesoon and Bang Kyungsik Dictionary of Real Estate Terms. Seoul: Buyonsa. 2 보고서저자명, 역할어. 보고서발행연도. 보고서표제 ( 영문은이탤릭 ). 발행지 : 발행처. 이상준, 이백진, 이현주 한반도북방지역인프라개발계획과협력전망. 안양 : 국토연구원. Lee Sangjun, Lee Backjin and Lee Hyunju A Study on the Infrastructure Development Policies of the Northeast Asian Countries and Prospects about the Future of International Cooperation. Anyang: Korea Research Institute for Human Settlements. 3 학회지저자명. 발행연도. 논문명. 자료명 ( 발행단체명 )( 영문은이탤릭 ) 권, 호 ( 연월차 ): 수록면수. < 국문논문 > 주신하, 임승빈 농촌경관계획수립기준정립연구. 농촌계획 14권, 4호 : Joo Shinha and Im Seungbin A study for rural landscape planning criteria. Journal of Korean Society of Rural Planning 14, no.4: < 외국어논문 > 佐藤貢一, 八島信良, 中西三化, 安達洋 氷盤と固定式海洋構造物の動的相互作用に関する研究. 日本建築学会構造系論文集 78 卷, 688 号 : Sato Kouichi, Yashima Nobuyoshi, Nakanishi Mitsukazu and Adachi Hiromi Study on dynamic interaction between ice-floes and fixed offshore structures: Dynamic behavior of offshore structures by collision of ice floe. Journal of Structural and Construction Engineering 78, no.688:

121 4 학회발표자료저자명. 발행연도. 논문명. 자료집명 ( 영문은이탤릭 ): 수록면수. 저자명. 발행연도. 논문명. 자료출처명 ( 행사명 )( 영문은이탤릭 ). 발표일시, 장소. 박지영, 송상훈 라멘구조공동주택사례분석을통한비용절감형장수명주택의방향. 한국주거학회 2014 추계학술발표대회논문집 26권, 2호 : Park Jiyoung and Song Sanghoon. A case study on beam-column residential apartments for cost-effective long-life housing. Proceeding of Autumn Annual Conference of KHA , no.2: 일반잡지 신문기사저자명. 발행연도. 기사명. 매체명 ( 영문은이탤릭 ), 발행일, 수록면수. 매체명 ( 영문은이탤릭 ). 연도. 기사명. 게재월일, 수록면수. 박경남 KDI, 올해경제성장률 2.6% 로하향조정. 건설경제, 5월 25일. Park Kyoungnam KDI cuts Korea s 2016 GDP outlook to 2.6 percent. CNEWS, May 리포트및이슈페이퍼저자명. 발행연도. 표제. 매체명 ( 영문은이탤릭 ) 호 ( 또는발행월일 ). 발행지 : 발행처. 최재성 교통투자재원의안정적확보를위한해외주행세도입사례와정책적시사점. 국토정책Brief 599호. 세종 : 국토연구원. Choi Jaesung Overseas examples of mileage taxes to provide stable transportation investment resources and their policy implications. KRIHS Policy Brief no.599. Sejong: Korea Research Institute for Human Settlements. 7 학술논문저자명. 수여연도. 논문명 ( 영문은이탤릭 ). 학위명, 수여기관명. 김훈상 주택건설사업의토지와자본의최적투입시점및규모. 박사학위논문, 서울대학교. Kim Hoonsang The Optimal Development Timing and Density of Housing Construction. Ph.D. diss., Seoul National University. 8 전자자료저자명. 발행연도. 자료명 [ 자료유형 ]. 발행지 : 발행처. 인터넷주소 ( 검색연월일 ). 김태성 중국사뒷이야기 [ 전자책 ]. 서울 : 실천문학사. com/book/book_detail (2005 년 4월 16일검색 ). Kim Taesung The Behind Story of Chinese History. Seoul: Silcheon Mumhak. barobook.com/book/book_detail (accessed April 16, 2005). 9 법령법령명 ( 영문은이탤릭 ). 발행연도. 법률번호 (OO 월 OO일제정 / 개정 / 일부개정 ). 제OO조, 제OO항. 국토기본법 법률제10758호 (5월 30일개정 ). 제2조. Framework Act on The National Land. Act no.10758(wholly amended May 30). article 내부자료저자 ( 발행기관 ). 발행연도. 자료명. 내부자료 ( 미출판자료 ). 한국토지주택공사 오류지구시설공급계획. 내부자료. LH The facilities supply planning in Oryu area. raw data. 11 보도자료발행기관. 발행연도. 기사명, 월일. 보도자료. 국토교통부 공간정보목록표준화완료공간정보이용더쉬워진다, 3월 1일. 보도자료. Ministry of Land, Infrastructure and Transport Using spatial information will be easier, March 1. Press release. 상세한내용및규정전문은국토연구원홈페이지 ( 의국토연구논문모집메뉴를참고.

122 알림 국토연구 논문모집 한국연구재단등재학술지 국토연구 에서논문을모집하고있습니다. 투고된논문은엄격한심사를거쳐게재됩니다. 상세한논문작성방법및투고절차는국토연구원홈페이지 ( 를참고하십시오. 투고대상 : 관련전문가및대학원생 원고매수 : 200 자원고지 80매이내 (A4 15매이내 ) 논문내용 : 국토관련분야연구논문, 정책연구 접수방법 : 국토연구원홈페이지를통한투고신청또는이메일접수 접수시기 : 수시 ( 연 4회발간 ) 문 의 : 국토연구원홍보출판팀학술지담당 Tel) , ) journal@krihs.re.kr 국토연구 우수논문상안내 국토연구원은국토관련연구의질을제고하고연구의욕을진작시키고자, 우수논문을선정하여포상하고있습니다. 우수논문은선정위원회에서한해동안게재된논문을종합평가하여선정합니다. 대상 : 1년간 국토연구 에게재된논문 발표 : 국토연구원홈페이지 ( 에게시및개별통보 시상내역 : 최우수상 1편 ( 연구장려금 200만원 )/ 우수상 2편 ( 연구장려금 100만원 ) 국토연구 는국토관련연구를장려하기위해투고료및심사료를받지않으며, 논문게재자에게는별쇄본 20 부를드립니다.