Modelling building ownership boundaries within BIM environment: A case study in Victoria, Australia

Computers, Environment and Urban Systems 61 (2017) 24–38

Contents lists available at ScienceDirect

Computers, Environment and Urban Systems journal homepage: www.elsevier.com/locate/ceus

Modelling building ownership boundaries within BIM environment: A case study in Victoria, Australia Behnam Atazadeh ⁎, Mohsen Kalantari, Abbas Rajabifard, Serene Ho The Centre for Spatial Data Infrastructures and Land Administration, Department of Infrastructure Engineering, The University of Melbourne, Victoria 3010, Australia

a r t i c l e

i n f o

Article history: Received 23 March 2016 Received in revised form 2 August 2016 Accepted 3 September 2016 Available online xxxx Keywords: Buildings Building Information Model Ownership boundary Subdivision plan IFC

a b s t r a c t In Australia, current practices for subdividing ownership spaces in buildings rely on two-dimensional (2D) analogue subdivision plans. As building subdivisions become more complex, these plans demonstrate limitations in communicating various boundaries that define complex ownership interests inside multi-storey buildings. With advances in three-dimensional (3D) digital information technologies, 3D digital models are increasingly being researched as a possible solution for improving the recording and representation of building ownership boundaries to overcome these communication challenges. This paper examines the feasibility of one such model – Building Information Models (BIM) – as it offers a collaborative, 3D, digital and semantically enriched data environment to support the integrated management of both physical and functional aspects of buildings over their lifecycle. Using a case study of current building subdivision practices in Victoria, this paper explores BIM's ability to model the boundaries of volumetric ownership spaces inside buildings. Through this process, relevant entities suitable for modelling building ownership boundaries are identified and proposed in the BIM standard. A prototype model is then implemented to showcase the practicality of the BIM environment for modelling building ownership boundaries. © 2016 Elsevier Ltd. All rights reserved.

1. Introduction 1.1. Background In Australia, land registration systems use subdivision plans to define the spatial dimensions of ownership interests associated with private and communal properties in buildings (Libbis, 2015). The objective of a building subdivision plan is to create “Good Title”: this means that the boundaries of private properties as well as common property areas must be unambiguously defined and owners should clearly understand their rights. Currently, subdivision plans rely on 2D-based representations to delimit 3D ownership rights which become inherently complicated in structurally and functionally complex multi-storey developments. The dominance of such buildings in urban areas infers communication challenges for a growing community of stakeholders in their interaction with existing 2D-based land registration systems. To mitigate such challenges, land registries are currently investigating the potential and viability of 3D digital data environments for managing complex land and property information (Rajabifard, Kalantari, & Williamson, 2012). ⁎ Corresponding author. E-mail addresses: [email protected] (B. Atazadeh), [email protected] (M. Kalantari), [email protected] (A. Rajabifard), [email protected] (S. Ho).

http://dx.doi.org/10.1016/j.compenvurbsys.2016.09.001 0198-9715/© 2016 Elsevier Ltd. All rights reserved.

These efforts have been stimulated by the significant progress in the area of 3D modelling over the last decade. This has resulted in the creation of 3D digital models for buildings, which in turn provides better visualization and perception of complex architectural as well as structural building elements (Hammoudi, 2011). Among 3D building models, Building Information Models (BIM) provides the most detailed 3D spatial and semantic information about every building element during the lifecycle of a building (NBIMS, 2012). BIM is a 3D digital data space for sharing building information to enable multi-disciplinary collaboration among various actors involved in the lifecycle of buildings (Eastman, Teicholz, Sacks, & Liston, 2011). Recent surveys indicate that the BIMbased paradigm brings more productivity gains and long term benefits in comparison with existing 2D-based practices in the Architecture, Engineering and Construction (AEC) firms around the world (McGraw-Hill Construction, 2012; McGraw-Hill Construction, 2014). 1.2. Aim and scope Using the subdivision practices of Victoria, Australia, as a case study, this paper aims to explore the feasibility of BIM for modelling and representing various types of building ownership boundaries. Currently, such boundaries are defined by cadastral surveyors by referencing building elements (e.g. walls and slabs) or through specifying virtual denotations, and recorded on analogue 2D plans. This investigation contributes to the general advancement of the use of BIM but more

B. Atazadeh et al. / Computers, Environment and Urban Systems 61 (2017) 24–38

importantly, provides the basis for realising digitally-based subdivision information and title registration transaction processes to improve information management and community understanding of ownership rights in complex developments. 2. Literature review 2.1. Communication challenges resulting from 2D subdivision plans Building subdivision plans utilize 2D floor plans as well as cross section diagrams to show boundaries associated with private and common property ownership spaces. This medium of communication efficiently represents the spatial extent of ownership spaces in simple structures such as single-storey buildings with relatively straightforward structural configurations. However, this practice is challenged by the requirements of multi-storey buildings, which is often not only endowed with complex architectural form, but also functional complexity with usage and access of building facilities differentiated among various groups of a building's inhabitants. For example, a multi-purpose building could have various access requirements to support retail, commercial and residential functions. 2D subdivision plans face challenges in communicating spatial complexity associated with boundaries of volumetric ownership spaces inside those buildings (Aien, Kalantari, Rajabifard, Williamson, & Wallace, 2013). These include: • Difficulty in representing the boundaries of ownership spaces with irregular forms. • Use of numerous 2D floor plans and cross-section diagrams to fully represent all ownership spaces in high-rise buildings: a large number of 2D diagrams results in difficulties in finding the right floor plan and cross section diagrams that, when viewed together, represents the horizontal and vertical boundaries for a specific private ownership space (Jazayeri, Rajabifard, & Kalantari, 2014). Consequently, interpretation of ownership boundaries for each apartment unit in high-rise developments is a time-consuming task. • 2D plans usually include textual notations which are difficult for owners, who have a limited background in understanding subdivision plans, to read and interpret accurately (Shojaei, Rajabifard, Kalantari, Bishop, & Aien, 2014). • Many of these buildings include commonly owned and managed properties (e.g. corridors, stairwells, gyms, etc.) and as illustrated in Fig. 1, representing boundaries of a common property spanning several stories is a very challenging task using 2D diagrams (Atazadeh, Kalantari, Rajabifard, Ho, & Ngo, 2016).

Fig. 1. An indicative example of a complicated common property area, adapted from (Atazadeh, Kalantari, Rajabifard, Ho, et al., 2016).

25

• Understanding of structural boundaries, which are defined by referencing building elements, could be problematic for inexpert stakeholders. For example, in Victoria, Australia, ownership boundaries can be defined by using one of three spatial relationships with its corresponding building element: the boundary touches either interior or exterior face of the building element, or it corresponds with the (virtual) median line of the building element (Libbis, 2015). These boundary types are often denoted by different types of lines in the plans. If a person does not have enough knowledge to understand the meaning of the lines delineated in the plan, he/she could misunderstand the location of the boundary. For instance, a non-specialist may misinterpret an interior ceiling boundary as a median one. • Finally, since the data environment of subdivision plans is analogue and static, it is impossible to view the boundaries of ownership spaces in an interactive mode (Döner et al., 2010; Stoter, van Oosterom, Ploeger, & Aalders, 2004). This drawback may affect interpretation and understanding of building ownership boundaries in complex situations. Moreover, an analogue data environment does not provide the capability to query, find and measure building ownership boundaries (Acharya, 2011).

To overcome the challenges mentioned above, there has been an increasing trend towards leveraging 3D digital building models in the cadastral domain. Some of these models comprise solely pure geometric information, such as shape, facade and texture, about building elements, while others also incorporate semantic information which would support a greater range of building information analysis (Kolbe, 2009). Open data models or standards lay the foundation for managing and exchanging 3D building information in an interoperable environment. Therefore, relevant 3D geospatial and BIM standards and their use for cadastral purposes will be briefly reviewed in Sections 2.2 and 2.3 respectively. 2.2. 3D geospatial standards and the cadastral domain Open Geospatial Consortium (OGC) has developed two standards in terms of 3D modelling of the built environment: CityGML and IndoorGML. The CityGML standard provides a multi-level representation for interoperable access and exchange of 3D urban information models (Groger, Kolbe, Nagel, & Hafele, 2012). This standard distinguishes five levels of detail (LoD), in which both spatial and semantic information about urban objects increases in higher LoDs. The CityGML standard is itself not specifically developed for cadastral purposes. However, the building model of this standard includes some entities which could potentially be used for modelling building ownership boundaries. For instance, an internal wall boundary can be modelled by “InteriorWallSurface” entity, while “OuterCeilingSurface” entity can be utilized for defining an external ceiling boundary. In addition, a few researchers have recently proposed cadastral extensions of CityGML standard using its Application Domain Extension (ADE) mechanism. One of the earliest investigations was by Dsilva (2009) who did a simple extension of a building model in CityGML with legal (cadastral) information from a Dutch jurisdiction. This extension comprised two parts. The first part includes legal attributes appended to the “_AbstractBuilding” class, such as building owner, the building registration number, the parcel number of the building, and the building type. The second part contains a new class called “_KadasterApartment” which is particularly used for managing legal information associated with each individual apartment, such as the ownership right, the apartment owner, the ownership type and the number of inhabitants within the apartment. However, Dsilva's investigation was not sufficient to semantically distinguish land parcels as well as various ownership arrangements - such as individually owned and communal properties defined inside buildings.

26

B. Atazadeh et al. / Computers, Environment and Urban Systems 61 (2017) 24–38

Therefore, Çağdaş (2013) developed a more comprehensive extension of CityGML, which encompassed entities used for modelling cadastral parcels and condominium units in the context of Turkish jurisdictions. This extension mainly proposed three new classes, namely the abstract “PropertyUnit” class and instantiable “CadastralParcel” and “CondominiumUnit” classes. The “PropertyUnit” class includes the ownership and taxation information required for any type of property unit. In addition to inheriting attributes of “PropertyUnit”, the “CadastraParcel” has its distinct attributes, such as the number, area, and value of the parcel, required for managing land parcels. Additionally, the spatial extent of “CadastralParcel” is defined through its inheritance from “LandUse” class. The “CondominiumUnit” class is used for modelling individually owned properties inside buildings and it has its specific legal attributes such as owner's name, taxpayer, party share, tax type and so on. This class has composition relationship with “Join Facility” and “Annex” classes, which are respectively used for modelling unlimited and limited commonly owned properties. Unlimited commonly owned properties are for the use and benefit of all owners, whereas limited ones must be used by a specific group of owners. The proposed extension by Çağdaş has been recently modified and adopted in OGC's land and infrastructure conceptual standard (Scarponcini, Gruler, Stubkjær, Axelsson, & Wikstrom, 2016). The IndoorGML standard has recently been developed for modelling indoor spatial information required for navigation purposes inside buildings (Lee et al., 2014). The basis of this standard is underpinned by two conceptual frameworks, namely Structured Space Model (SSM) and Multi-Layered Space Model (MLSM). SSM specifies the spatial layout of indoor environment, which is independent from its semantic interpretation. SSM defines two distinct types of spaces: primal and dual ones. The primal space defines the geometry and topology of indoor subdivided spaces. The dual space includes a connectivity graph induced from primal space by applying the Poincaré duality. This graph models adjacency relationships between indoor spaces and provides the navigation network within indoor environments. MLSM is a combination of multiple space layers of SSM, each of which specifies a distinct semantic interpretation of indoor environment for a specific application. For instance, an MLSM could include a topographic space layer, a Wi-Fi sensor space layer and an RFID sensor space layer. There could be two alternatives to investigate the use of IndoorGML in the cadastral domain (Zlatanova, Oosterom, Lee, Lic, & Lemmen, 2016). One option is to define a new cadastral space layer inside IndoorGML standard. The core module of IndoorGML includes “CellSpace” and “CellSpaceBoundary” entities which describe space objects and their boundaries respectively. These entities could potentially

be extended with cadastral information to arrange ownership rights associated with indoor spaces. Another solution could be defining an external linkage between IndoorGML and existing 3D cadastral standards. For instance, Zlatanova et al. (2016) have recently investigated the possible synergies between IndoorGML and LADM (Land Administration Domain Model) and found that the concept of spatial units within LADM can be linked to “CellSpace” entity in IndoorGML. 2.3. BIM standard and the cadastral domain Each BIM platform utilizes a proprietary format for storing BIM models (Eastman et al., 2011). Therefore, BuildingSMART, an organization which promotes the use of BIM, developed the Industry Foundation Classes (IFC) standard as an open and vendor neutral data model to facilitate interoperability across multiple BIM platforms (ISO16739, 2013). The IFC standard is object oriented and uses a hierarchical spatial data model to manage building objects that have a geometric representation or spatial extent (see Fig. 2). “IfcProduct” is the most abstract super class in the spatial hierarchy of the IFC standard. It has two subclasses, namely “IfcElement” and “IfcSpatialElement” (Liebich, 2009). The former one is the super class of IFC entities, which is represented in yellow in Fig. 2, used for modelling physically existing elements such as building elements (IfcBuildingElement), and geographic elements (IfcGeographicElement). The latter one is the super class of two entities, which is shown in green in Fig. 2, defining the spatial arrangements within an IFC file. The first one is “IfcSpatialStructureElement” which is a generalization of the elements defining spatial structure of a building project. These elements include “IfcSite” (construction site of the project), “IfcBuilding” (buildings), IfcBuildingStorey (levels of a building), and IfcSpace (indoor spaces). The second entity is “IfcExternalSpatialStructureElement” which currently includes one specialized class named “IfcExternalSpatialElement”. This class models various types of outdoor spaces or regions. Due to the level of detail it manages, the potential value of BIM in the cadastral domain, and in particular the IFC data model, has become a focus of various investigations. Clemen and Gründig (2006) found that the IFC standard could be used for managing cadastral information in indoor environments; however, raw surveying measurements and observations needed to be processed before integration into BIM. The authors only provided a list of those IFC entities which needed to be considered in indoor cadastral application of IFC standard such as measurement units, observation nodes, observation edges, and observation groups. However, they did not investigate how indoor cadastral data can be accommodated within the spatial data model of IFC standard.

Fig. 2. Generalized data structure of the IFC standard. (For interpretation of the references to color in this figure, the reader is referred to the web version of this article.)

B. Atazadeh et al. / Computers, Environment and Urban Systems 61 (2017) 24–38

The urbanIT integration framework was another relevant research project which focussed on development and implementation of an integrated urban information model by coupling indoor spatial information, sourced from BIM, with outdoor spatial information which comes from geographic information systems (GIS) (Plume & Mitchell, 2011). In particular, the core of the urbanIT project was a proposed extension to the IFC standard for managing cadastral data both inside buildings as well as land parcels on the site of buildings (Barton, Marchant, Mitchell, Plume, & Rickwood, 2010). The proposed extension considered “IfcSpace” and “IfcZone” entities for modelling cadastral spaces inside buildings. In addition, the “IfcSite” entity was extended with cadastral attributes for modelling legal boundaries of 2D land parcels on a building site. Although this extension was a good trial for representing cadastral spaces within buildings, it did not investigate modelling the boundaries of these spaces within the IFC standard. Another investigation extended the Unified Building Models (UBM) with four boundary types, namely “Building Elements Surfaces”, “Digging Surfaces”, “Protecting Area Surfaces”, and “Real Estate Boundary Surfaces” (El-Mekawy & Östman, 2012). UBM was initially developed for facilitating data exchange from IFC to CityGML or vice versa (ElMekawy, 2010). In subsequent studies, El-Mekawy, Paasch, and Paulsson (2014, 2015) examined IFC, CityGML and LADM standards in terms of modelling various types of rights, restrictions, and responsibilities (RRRs) associated with 3D properties in Sweden. They found that both IFC and CityGML do not support legal aspects of 3D properties such as boundaries and RRRs, whereas LADM provides a framework for modelling only legal spaces. Therefore, they argued that the proposed extension to UBM would efficiently model 3D properties in comparison with those standards. Additionally, this extension could potentially provide the basis for facilitating interaction between BIM and cadastral domains (El-Mekawy et al., 2014). However, in this extension, the spatial relationships between boundaries and ownership spaces are not modelled – for instance, there is no support for how “Building Elements Surfaces” might reference walls or ceilings bounding a private apartment unit. Isikdag, Horhammer, Zlatanova, Kathmann, and Van Oosterom (2014) studied the impact of 3D digital building models, which are stored in semantically enriched data formats such as IFC and CityGML, as well as 3D cadastral models in valuating real estate properties in different jurisdictions such as Turkey, Netherlands and UK. They found that 3D digital data environments could potentially enhance valuation and taxation practices in the future; however, this improvement differs from one jurisdiction to another since each jurisdiction has its own requirements for valuating properties. More recently, Isikdag, Horhammer, Zlatanova, Kathmann, and Van Oosterom (2015) argued that a 3D digital building model equipped with “streetview” images would be able to describe environmental features in the vicinity of the building, which leads to improving representation of valuation information associated with apartment units of the building. Finally, Atazadeh, Kalantari, Rajabifard, Ho, et al. (2016) argued that the IFC standard does not support legal (cadastral) information and, therefore, they proposed an extension to the IFC standard for managing 3D cadastral entities. The proposed extension comprises legal property objects, which were envisaged as volumetric cadastral spaces, various types of legal documents, such as title, mortgage and caveats, as well as information about interest holders or owners. The realization of the extension in a prototype model highlighted that the BIM data environment has the potential capacity to manage cadastral information in high-rise and complex building structures (Atazadeh, Kalantari, Rajabifard, Champion, & Ho, 2016). However, this cadastral extension of IFC was predicated on those legal entities found from reviewing literature rather than examining subdivision practices in the real world. In addition, although this study found that the IFC standard could potentially provide the basis for managing cadastral boundaries in the BIM data environment, it did not investigate how these boundaries, particularly the range of cadastral

27

boundaries defined inside high-rise buildings, could be mapped onto their corresponding IFC entities. While the investigations reviewed above examine the potential value of BIM in the cadastral domain from different perspectives, no single study has yet explored the accommodation of various building ownership boundary types within the BIM environment. Thus, in the next section, the methodology for meshing building ownership boundaries into the BIM environment is developed. 3. Methodology In the cadastral domain, there are two main research paradigms, namely behavioural research and design science (Çağdaş & Stubkjær, 2011). Behavioural research focuses on the social science dimension of cadastral systems such as understanding institutional issues (Çağdaş & Stubkjær, 2009) and describes, explains and predicts existing human and organizational behaviour in the real world. On the other hand, design science concentrates on the information science dimension of cadastral systems (Çağdaş & Stubkjær, 2011). This research paradigm intends to improve the existing practices in the real world by using new or existing artefacts which support researchers' ability to overcome their problems (Johannesson & Perjons, 2014). The methodology of the paper is based on the design science paradigm since this research introduces the BIM environment as a new artefact to overcome communication challenges resulting from 2D building subdivision plans. Therefore, the methodology includes five main stages: 1. Explicate problem: The problem identified in this research is communication challenges of building (3D) ownership boundaries for non-specialists (see Section 2.1). This stage was done in the academic environment over the course of five months. 2. Define requirements: In this stage, the explicated problem is transformed into demands expected from the proposed artefact. In the context of this research, these “demands” are the ability to better communicate building ownership boundaries. More specifically, this requires the new artefact to capably represent the various types of building ownership boundaries. The adopted approach for identifying ownership boundaries is participant observation method which enables the researcher to gain a deep understanding of a phenomenon (Johannesson & Perjons, 2014). For this research, a placement in the industry environment, which is a surveying company specialized in subdividing ownership of building developments, was undertaken by the first author during a seven month period to directly observe how land surveyors delineate various types of building ownership boundaries (see Section 4.1). 3. Design and develop artefact: In this stage, the open BIM standard, i.e. IFC, was proposed as an artefact to address the identified problem and meet the defined requirements. This means that the existing data structure of IFC was first analysed to identify and propose relevant entities suitable for modelling building ownership boundaries. Subsequently, the boundaries identified in the previous stage were mapped onto their equivalent entity within the IFC standard (see Section 4.2). This stage took about two months in the academic environment. However, it should be noted that the first author had spent approximately six months to gain knowledge of BIM and IFC prior to undertaking the industry placement. 4. Demonstrate artefact: The demonstration provides an instance of the artefact to address the problem. Over the period of six months, a prototype BIM model was developed and implemented to prove the feasibility of the IFC standard for modelling building ownership boundaries. The initial BIM model was prepared in the academic environment but the correct delineation of ownership boundaries defined inside the BIM model was finalized by consulting the land surveyors in the company (see Section 5).

28

B. Atazadeh et al. / Computers, Environment and Urban Systems 61 (2017) 24–38

Fig. 3. Main stages of the methodology.

5. Evaluate artefact: The evaluation of the proposed artefact specifies to what extent it meets the requirements and alleviates the problem. In this research, the implemented BIM-based subdivision prototype was compared with its 2D plan version to show how representing

building ownership boundaries in a 3D digital data environment of BIM facilitates communication of these boundaries through greater clarity and interactivity of the data. The initial and theoretical evaluation of the artefact was done by the authors in the academic

Fig. 4. Taxonomy of building ownership boundaries.

B. Atazadeh et al. / Computers, Environment and Urban Systems 61 (2017) 24–38

29

Fig. 5. Interior boundaries, adapted from (LandVictoria, 2015).

environment. However, the final evaluation presented here was also consolidated according to land surveyors' opinions. The whole time spent for this stage was around four months (see Section 6). These stages are represented in the conceptual diagram shown in Fig. 3.

types of subdivisions in Victoria was used to develop a taxonomy of building ownership boundaries. The legal basis of defining building ownership boundaries in Victoria is underpinned by the Subdivision (Registrar's Requirements) Regulations (2011). In the cadastral domain, there is a dichotomy between general and fixed boundaries. The developed taxonomy is the expanded version reflecting this dichotomy (see Fig. 4).

4. Mapping building ownership boundaries into BIM 4.1. Boundaries on subdivision plans To investigate different types of boundaries delineated on subdivision plans, an industry placement with a land surveying company involved in subdividing large and complex multi-level developments in Victoria, Australia, was undertaken. In the field of built environment, industry placements provide the opportunity to understand current practices through investigating real world examples, which is complementary to the theoretical knowledge gained in an academic environment (Frank, 2005). During the placement period, a range of subdivision plans deemed to represent a comprehensive range of

4.1.1. General boundaries General boundaries are boundaries that are specified and observed based on real world, tangible spatial objects. In building subdivision plans, there are three main types of general boundaries: structural, ambulatory and projected. Structural boundaries are defined and measured by considering the building itself or a part of it. These boundaries can be classified into three categories: interior, median and exterior. Interior boundaries are delineated by the interior face of building elements (see Fig. 5). This means that all internal fixtures, coverings, and water proof membranes attached to walls, ceilings, and floors are part of the relevant private

Fig. 6. Median boundaries, adapted from (LandVictoria, 2015).

30

B. Atazadeh et al. / Computers, Environment and Urban Systems 61 (2017) 24–38

Fig. 7. Exterior boundaries, adapted from (LandVictoria, 2015).

Fig. 8. Ambulatory boundary in a subdivided development.

B. Atazadeh et al. / Computers, Environment and Urban Systems 61 (2017) 24–38

31

Fig. 9. Projections in subdivision plans, a) boundary delineation in cross section diagrams, b) boundaries labelled “P” in the first floor are defined by projection.

ownership space (LandVictoria, 2015). For walls, doors and windows, the internal surface is used to delineate interior boundary. For floor slabs, interior boundary is defined by referencing the upper surface, whereas, for ceiling slabs, this boundary is defined as the underside of the suspended ceiling. Median boundaries lie along the imaginary surface passing through the centreline of building elements. In other words, a median boundary is delineated through referencing a surface having the same distance from both exposed surfaces of a building element. This boundary is delineated to divide the ownership of its corresponding building element between the abutting ownership spaces. Fig. 6 represents an example of

median boundaries to illustrate how these boundaries are depicted and notated in subdivision plans. Exterior boundaries are delineated through referencing external surface of building elements. These boundaries indicate that the whole of relevant building elements are contained within the corresponding ownership space. For instance, as illustrated in Fig. 7, the boundary between Lot 1 and common property No. 1 should be interpreted that the wall, window and door are completely part of the Lot 1. As a specific type of general boundary (Williamson, Enemark, Wallace, & Rajabifard, 2010, p. 360), ambulatory boundaries are defined based on observing the movement of dynamic natural features such as

Fig. 10. Fixed boundaries.

32

B. Atazadeh et al. / Computers, Environment and Urban Systems 61 (2017) 24–38

coastlines and river borders (DELWP, 2014). In Fig. 8, the floor plan for the ground level of a development comprising four multi-storey buildings is represented and the southern boundary of one of the common properties (Common Property No. 1) is delineated by referencing the Yarra river. This boundary changes as the water level of the river lowers or rises. Gauge boards with in situ sensors can be used to provide real time values of water level as time series data sets. These time series values can be used to monitor the position of ambulatory boundaries over the lifecycle of the subdivided development. Projected boundaries are mainly used to define balconies and terraced areas of buildings. It is mainly delineated by extending structural boundaries in both horizontal as well as vertical directions. These boundaries are represented as thick broken lines in cross section diagrams (Fig. 9a). However, in floor plans, projected boundaries are delineated as thick continuous lines. The study showed that it was unusual to extend vertical building elements, such as walls, in horizontal directions in any condition. This means that floor plans can only show vertical projections (Fig. 9b). 4.1.2. Fixed boundaries Fixed boundaries are boundaries that are specified based on surveying measurements such as distance, angle and azimuth. In subdivision plans, this type of boundary is usually utilized to define the spatial extent of parking and storage areas in buildings. In Fig. 10, boundaries of a parking area in the basement level of a building are represented and it can be seen that most boundaries are delineated through their corresponding distances and azimuths which are annotated alongside the boundary lines. 4.2. Modelling building ownership boundaries in IFC standard The IFC standard includes hundreds of entities to support all data exchange requirements over the lifecycle of buildings. However, data exchanged between actors in each stage of a building project is based on a subset of IFC entities to meet a particular requirement. The matter of deciding which IFC entities are appropriate for specific requirements is

known as Information Delivery Manual (IDM) approach (Wix & Karlshoej, 2010). This paper is also based on IDM approach and we selected those IFC entities (as shown in Fig. 11) to be used for modelling building ownership boundaries in the cadastral domain. Among these entities, ‘IfcRelSpaceBoundary’ class plays a key role in modelling both geometry and semantic of boundaries associated with various ownership spaces. There are two main entities for representing volumetric spaces in IFC standard, namely “IfcSpace” for spaces inside buildings and “IfcExternalSpatialElement” for external spaces around buildings (see Fig. 11). First, geometric representation of ownership boundaries will be elucidated in Section 4.2.1. Subsequently, in Sections 4.2.2 and 4.2.3, it will be explained how each boundary type can be modelled by IFC entities as shown in Fig. 11. 4.2.1. Modelling geometry of boundaries in IFC Fig. 12 represents the entities used for modelling geometry. The most abstract entity for representing geometry is “IfcConnectionGeometry” which is associated to “IfcRelSpaceboundary” via ConnectionGeometry relationship. Since the building ownership boundaries must be represented as surfaces or faces, “IfcConnectionSurfaceGeometry” subclass can be used for defining the geometric connection between two ownership spaces. The IFC standard includes entities which define boundaries of spaces through connective relationships: SurfaceOnRelatingElement relationship specifies whether the connection is a surface (IfcSurface) or a face with an associated surface (IfcFaceSurface) via “IfcSurfaceOrFaceSurface” entity. Geometry and placement of the boundary surface are defined relative to its relating ownership space. Additionally, the optional “SurfaceOnRelatedElement” relationship provides the geometry and placement of the same surface within the local coordinate system of the related ownership space. 4.2.2. General boundaries in IFC For modelling structural boundaries, “IfcRelSpaceBoundary” entity should reference IfcBuildingElement class (or its subclasses) through “RelatedBuildingElement” relationship. In addition, the value of “IfcPhysicalOrVirtualEnum” attribute must be set to PHYSICAL since

Fig. 11. Semantic entities for modelling building ownership boundaries in IFC standard. (For interpretation of the references to color in this figure, the reader is referred to the web version of this article.)

B. Atazadeh et al. / Computers, Environment and Urban Systems 61 (2017) 24–38

33

example that illustrates how an interior wall boundary between two apartment units can be modelled using IFC entities. Although there is no physical manifestation for projected boundaries, these boundaries can be similarly modelled by referencing building elements since projected boundaries are delineated through extrusion of structural boundaries. For instance, the boundary surface referencing a wall in a balcony area can be extended to delineate its projection up to ceiling. Therefore, there does not appear to be a need to define a specific IFC entity for projected boundaries. For modelling ambulatory boundaries, “IfcRelSpaceBoundary” entity should reference “IfcGeographicElement” class, which is shown in red in Fig. 11, through “RelatedBuildingElement” relationship. “IfcGeographicElement” is the semantic entity for modelling geographic features such as rivers, lakes, trees and roads. Within this entity, the “PredefinedType” attribute specifies the type of geographic element. For modelling rivers and coastlines, this attribute should be set to “RIVER” and “COASTLINE” respectively.

Fig. 12. Relevant entities for modelling geometry of building ownership boundaries in IFC.

structural boundaries reference physically existent spatial objects. Both interior and exterior structural boundaries can be semantically distinguished from each other by specifying their corresponding enumerator from “InternalOrExternalBoundary” enumeration. There is currently no enumerator defined for median boundaries to differentiate them from other types of structural boundaries. Therefore, it is suggested that “InternalOrExternalBoundary” should be enriched with “Median” value. Semantic entities for building elements associated with structural boundaries are represented in blue in Fig. 11. Fig. 13 provides an

4.2.3. Fixed boundaries in IFC In order to model fixed boundaries, “IfcRelSpaceBoundary” entity should reference IfcVirtualElement class through “RelatedBuildingElement” relationship. This class is represented in grey in Fig. 11. Additionally, the value of “IfcPhysicalOrVirtualEnum” attribute must be set to VIRTUAL because surveying measurements are not tangible in the real world. Fig. 14 provides an example to illustrate how a fixed boundary between two car park areas can be modelled using IFC entities. 5. Prototype implementation To showcase the practicality of BIM environment, and in particular IFC standard, for modelling building ownership boundaries, a prototype model of a complex development, which comprises a multi-storey apartment and multi-level townhouses, located in Melbourne was developed. The reason for selecting this development was that various

Fig. 13. Mapping an interior wall boundary within IFC standard.

34

B. Atazadeh et al. / Computers, Environment and Urban Systems 61 (2017) 24–38

Fig. 14. Mapping a fixed boundary within IFC standard.

types of building ownership boundaries, except ambulatory ones, were defined in its different parts. Therefore, representation and communication of almost any ownership boundary type can be compared in both BIM environment and subdivision plans. To construct the BIM model, Autodesk Revit was used since it is one of the more well-known BIM authoring tools used in the AEC industry for 3D modelling of complex developments. First, 2D architectural floor plans for each level and cross-section diagrams in CAD format were imported into the Revit Environment. Subsequently, various 3D building elements, such as doors, walls and windows, were created by extruding their footprints in the 2D CAD plans. The subdivision plan of the development was then used to delineate different types of building ownership boundaries. The correctness of these boundaries were checked and verified by consulting the land surveyor who prepared the subdivision plan of the development. For structural boundaries, Revit has the capability to automatically create the boundary by checking the “Room bounding” attribute. In addition, the location of each structural boundary is set to interior, median or exterior as specified in the subdivision plan of the development. Fixed boundaries were modelled using “Room Separator” tool. The delineated ownership boundaries specify the 3D delimitation of various private ownership spaces, such as apartment units, parking areas and storages, as well as common property ownership spaces such as elevators, corridors, and lobbies. Each ownership space was automatically constructed by applying the “Room” tool in Revit once its constituting boundaries were defined. Autodesk Revit has its own native format for storing BIM and this does not show the delineated ownership boundaries in 3D view. Therefore, the BIM model was exported in IFC format and imported into Solibri Model Viewer (SMV) software, which provides 3D visualization of IFC files. SMV has the capability to clearly highlight and represent boundaries of ownership spaces. The summary of steps for developing the prototype model is outlined in Fig. 15. Various parts of the implemented BIM model are represented below to showcase the practicability of the BIM environment for managing building ownership boundaries.

Fig. 16 represents two apartment units (Nos. 203 and 204) within the second level of the apartment building, in which the interior boundary between the units were highlighted in both plan version and BIM

Fig. 15. Stages of implementing the BIM-based subdivision prototype.

B. Atazadeh et al. / Computers, Environment and Urban Systems 61 (2017) 24–38

35

Fig. 16. Interior boundary, a) plan version, b) in BIM environment.

Fig. 17. Median boundary, a) plan version, b) in BIM environment.

environment. In addition, some townhouses are represented in Fig. 17, which shows two abutting townhouses (Nos. 2 and 3) and the position of their common median boundary in both subdivision plan and BIM environment. As illustrated in Fig. 18, a typical example of external boundary for townhouse No. 4 is also highlighted in both subdivision plan and BIM environment.

For projected boundaries, the BIM environment can represent these automatically by extruding the structural boundary surfaces in both vertical and horizontal directions (see Fig. 19). Finally, Fig. 20 shows that how fixed boundaries associated with a car park area can be represented in both subdivision plans and the BIM environment.

Fig. 18. Exterior boundary, a) plan version, b) in BIM environment.

36

B. Atazadeh et al. / Computers, Environment and Urban Systems 61 (2017) 24–38

Fig. 19. Projection boundary, a) plan version, b) in BIM environment.

6. Discussion There are recognized limitations of current 2D-based practices in representing and communicating information about ownership spaces within multi-storey developments. These include complex representations that are challenging for users to comprehend, reliance of textual descriptions and drafting methods for representing different boundary types, and multiple pages of information that makes locating information about individual properties difficult. The aim of this paper was therefore to examine the feasibility of the BIM environment for modelling and representing building ownership boundaries by investigating entities within the open IFC standard. The BIM prototype developed in this paper demonstrates some clear advantages in overcoming these communication challenges. Visualizing these spaces and their boundaries in a 3D digital environment is easy to understand and supports the community's understanding of who owns what, and where. In particular, the ability to model connective relationships supports the practice of using a building's structure to define ownership spaces. Here, the various boundary outcomes resulted from the implemented BIM model are compared with their 2D plan version. • Interior structural boundaries: For a person without a specialist background in terms of understanding subdivision plans, it is hard to interpret the location of boundary in this case (see Fig. 16a). In contrast, it

can be intuitively perceived inside BIM environment that the boundaries of apartment units are delineated by the internal surfaces of the wall (see Fig. 16b). This would communicate that neither owner of the apartment units has any rights to the wall, which would subsequently be a part of the common property. • Median structural boundaries: A land surveyor might define a median boundary through the use of an “M” annotation alongside the boundary line. However, the owners of townhouses represented in Fig. 17a may have difficulty in terms of understanding this boundary. When represented in a BIM environment, this boundary is unambiguous, as shown in Fig. 17b. Each owner can clearly understand that their ownership interest is spatially extended up to the half of the wall on their side. • Exterior structural boundaries: In the notation section of the subdivision plan it is implicitly indicated that this boundary is an external one. However, the owner of townhouse No. 4 may not accurately interpret its location (see Fig. 18a). Mapping this boundary into BIM environment, as illustrated in Fig. 18b, could be more congruous with the visual perception of non-specialists. This would easily communicate that the building element associated with the boundary (in this case, the wall) completely belongs to townhouse No. 4. • Projected boundaries: In subdivision plans, these can be misinterpreted similar to structural boundaries since their location is dependent on the position of their corresponding structural

Fig. 20. Fixed boundary, a) plan version, b) in BIM environment.

B. Atazadeh et al. / Computers, Environment and Urban Systems 61 (2017) 24–38

boundaries (see Fig. 19a). However, modelling in a BIM environment can facilitate communication of these boundaries in the same way as structural boundaries (see Fig. 19b). • Fixed boundaries: Interpretation of fixed boundaries in subdivision plans can also be difficult for owners. For instance, in Fig. 20a, azimuth values annotated alongside boundaries of carpark of unit 304 may seem strange for the owner and she/he might get confused in terms of interpreting those values. Another issue is that owners may only look at the floor plan of their car park area and they think that the ownership space of their car park is up to the ceiling. However, in cross section diagrams it is sometimes indicated that the vertical extension of the car park is only up to two meters. The remaining space from height of 2 m to the ceiling is part of the common property area. This is more clearly communicated in a BIM environment, which can provide an integrated and interactive view of all ownership spaces (see Fig. 20b).

37

For future research, simulating ambulatory boundaries in a BIM environment will be investigated since the IFC standard provides entities for storing data related to regular and irregular time series. Other future work could include investigating the impact of a BIM model enriched with building ownership boundaries in determining various responsibilities associated with managing assets within buildings. The boundary types identified in this paper are specifically used for defining ownership spaces in the state of Victoria in Australia. However, they could potentially be similarly used for boundary types in other jurisdictions, as well as defining other types of functional areas or spaces such as those defined in recently developed property measurement standards. Consequently, another potential future research direction could be to investigate the role of a BIM environment and the IFC standard in facilitating communication of boundaries of area and volume types used for property measurements. Acknowledgment

Overall, it is demonstrated in the above comparisons that 2D plans of subdivision are challenging for owners in terms of their ability to accurately read and interpret information about ownership boundaries. Moreover, the ability of users to interact with the information presented in subdivision plans is very limited since the data environment is analogue and static. In complex developments, the ability to mentally conceptualise a full image of all boundaries of an ownership space requires reviewing multiple 2D images of the ownership space delineated in floor plans and cross-section diagrams. For interest holders of ownership spaces with complex shapes, joining these various 2D images in their mind to form a 3D image can be a cognitive challenge. On the other hand, the digital data environment of BIM is able to more effectively communicate boundaries as it provides 3D visual representation and structural information that orientates the viewer to the physical world. In addition, the interactivity of a digital environment would enable owners to perceive their ownership boundaries from various viewpoints by rotating, zooming and panning the BIM model. This suggests that lay users may find it easier to interpret and understand building ownership boundaries modelled in a BIM environment. Accurate understanding of ownership boundaries would result in resolving the ambiguities in determining the rights associated with different privately and commonly owned parts of buildings. While the BIM data environment provides significant benefits in terms of communicating complex ownership space boundaries common in the urban built environment, there are however still some technical issues for delineating ownership boundaries in a BIM environment. The most important issue is that BIM models, which are usually prepared by architects, represent the building in the design phase. Building ownership boundaries in the real world may be different to the boundaries delineated based on the design model of the building; therefore, there is a need to check the conformance of BIM model with its building in real world to ensure the integrity of the BIM model.

7. Conclusions and future directions This paper responds to the acknowledged communication challenges in 2D-based methods of representing building ownership spaces in complex structures. We argued that the BIM environment holds potential for alleviating these challenges. To support this argument, relevant geometric and semantic IFC entities for modelling complex building ownership boundaries in the BIM environment were identified. A prototype model was developed and implemented to show how a BIM environment would enhance the communication of various building ownership boundaries by improving on the visual representation. In particular, the paper has demonstrated that IFC can provide the foundation for managing different types of building ownership boundaries in a 3D BIM digital environment.

This work was supported by the Australian Research Council (ARC) [grant number LP110200178]. The authors appreciate the efforts of editorial office and constructive comments of anonymous reviewers. The first author would also like to extend his thanks to Mr. Tom Champion, Mr. Jeff Clarke, Mr. Alan Norman and Ms. Kate Warshall, who offered their time and support throughout his placement in Reeds Consulting Company. The authors emphasize that the views expressed in this article are the authors' alone. References Acharya, B. R. (2011). Prospects of 3D cadastre in Nepal. 2nd international workshop on 3D cadastres. Aien, A., Kalantari, M., Rajabifard, A., Williamson, I., & Wallace, J. (2013). Towards integration of 3D legal and physical objects in cadastral data models. Land Use Policy, 35, 140–154. http://dx.doi.org/10.1016/j.landusepol.2013.05.014. Atazadeh, B., Kalantari, M., Rajabifard, A., Champion, T., & Ho, S. (2016). Harnessing BIM for 3D digital management of stratified ownership rights in buildings. FIG working week 2016 recovery from disaster (Christchurch, New Zealand). Atazadeh, B., Kalantari, M., Rajabifard, A., Ho, S., & Ngo, T. (2016). Building information modelling for high-rise land administration. Transactions in GIS. Barton, J., Marchant, D., Mitchell, J., Plume, J., & Rickwood, P. (2010). A note on cadastre: UrbanIT research project. (Sydney). Çağdaş, V. (2013). An application domain extension to CityGML for immovable property taxation: A Turkish case study. International Journal of Applied Earth Observation and Geoinformation, 21, 545–555. http://dx.doi.org/10.1016/j.jag.2012.07.013. Çağdaş, V., & Stubkjær, E. (2009). Doctoral research on cadastral development. Land Use Policy, 26(4), 869–889. http://dx.doi.org/10.1016/j.landusepol.2008.10.012. Çağdaş, V., & Stubkjær, E. (2011). Design research for cadastral systems. Computers, Environment and Urban Systems, 35(1), 77–87. http://dx.doi.org/10.1016/j. compenvurbsys.2010.07.003. Clemen, C., & Gründig, L. (2006). The Industry Foundation Classes (IFC)—Ready for indoor cadastre? Proceedings of XXIII International FIG Congress, Munich, vol. 18. McGraw-Hill Construction (2012). The business value of BIM in North America: Multi-year trend analysis and user ratings (2007–2012). Bedford, MA: McGraw-Hill. McGraw-Hill Construction (2014). The business value of BIM in Australia and New Zealand: How building information modeling is transforming the design and construction industry. Bedford, MA: McGraw-Hill. DELWP (2014). Ambulatory boundaries. Retrieved June 6, 2016, from http://www.dtpli. vic.gov.au/property-and-land-titles/surveying/advice-and-guidelines-for-surveyors/ ambulatory-boundaries. Döner, F., Thompson, R., Stoter, J., Lemmen, C., Ploeger, H., van Oosterom, P., & Zlatanova, S. (2010). 4D cadastres: First analysis of legal, organizational, and technical impact—With a case study on utility networks. Land Use Policy, 27(4), 1068–1081. http://dx.doi.org/10.1016/j.landusepol.2010.02.003. Dsilva, M. G. (2009). A feasibility study on CityGML for cadastral purposes. Master's Thesis Eindhoven, The Netherlands: Eindhoven University of Technology. Eastman, C., Teicholz, P., Sacks, R., & Liston, K. (2011). BIM handbook: A guide to building information modeling for owners, managers, designers, engineers and contractors. John Wiley & Sons. El-Mekawy, M. (2010). Integrating BIM and GIS for 3D city modelling: The case of IFC and CityGML. KTH, KTH, Geoinformatics. Retrieved from http://urn.kb.se/resolve?urn= urn:nbn:se:kth:diva-28899. El-Mekawy, M., & Östman, A. (2012). Feasibility of building information models for 3D cadastre in unified city models. International Journal of E-Planning Research (IJEPR), 1(4), 35–58. El-Mekawy, M., Paasch, J., & Paulsson, J. (2014). Integration of 3D cadastre, 3D property formation and BIM in Sweden. 4th international workshop on 3D cadastres 9–11 November 2014, Dubai, United Arab Emirates (pp. 17–34).

38

B. Atazadeh et al. / Computers, Environment and Urban Systems 61 (2017) 24–38

El-Mekawy, M. S. A., Paasch, J. M., & Paulsson, J. (2015). Integration of legal aspects in 3D cadastral systems. International Journal of E-Planning Research (IJEPR), 4(3), 47–71. Frank, A. I. (2005). What do students value in built environment education? Cebe Transactions, 2(3), 21–29. Groger, G., Kolbe, T. H., Nagel, C., & Hafele, K. H. (2012). OGC city geography markup language (CityGML) en-coding standard. Open geospatial consortium: Wayland, MA, USA. Hammoudi, K. (2011). Contributions to the 3D city modeling: 3D polyhedral building model reconstruction from aerial images and 3D facade modeling from terrestrial 3D point cloud and images. Université Paris-Est. Isikdag, U., Horhammer, M., Zlatanova, S., Kathmann, R., & Van Oosterom, P. J. M. (2014). Semantically rich 3D building and cadastral models for valuation. Proceedings 4th international workshop on 3D cadastres, 9–11 November 2014, Dubai, United Arab Emirates. International Federation of Surveyors (FIG). Isikdag, U., Horhammer, M., Zlatanova, S., Kathmann, R., & Van Oosterom, P. (2015). Utilizing 3D building and 3D cadastre geometries for better valuation of existing real estate. In FIG working week 2015: From the wisdom of the ages to the challenges of the modern world (pp. 1–18). Bulgaria: Sofia. ISO16739 (2013). Industry Foundation Classes (IFC) for data sharing in the construction and facility management industries. buildingSMART. Jazayeri, I., Rajabifard, A., & Kalantari, M. (2014). A geometric and semantic evaluation of 3D data sourcing methods for land and property information. Land Use Policy, 36, 219–230. http://dx.doi.org/10.1016/j.landusepol.2013.08.004. Johannesson, P., & Perjons, E. (2014). An introduction to design science. Springer Publishing Company, Incorporated. Kolbe, T. H. (2009). Representing and exchanging 3D city models with CityGML. 3D geoinformation sciences (pp. 15–31). Berlin Heidelberg: Springer. LandVictoria (2015). Building subdivision guidelines. (Melbourne). Lee, J., Li, K. J., Zlatanova, S., Kolbe, T. H., Nagel, C., & Becker, T. (2014). OGC® IndoorGML. Wayland, MA, open geospatial consortium document no. OGC.

Libbis, S. (2015). Subidivsions Victoria: The ultimate guide. Melbourne, Australia: Hybrid Publishers. Liebich, T. (2009). IFC 2x edition 3. Model implementation guide. version 2.0. AEC3 Ltd. NBIMS (2012). National BIM standard — United States™ version 2. Retrieved from http:// www.nationalbimstandard.org/nbims-us-v2/pdf/NBIMS-US2_aB.pdf. Plume, J., & Mitchell, J. (2011). An urban information framework to support planning, decision-making & urban design. In P. Leclercq, A. Heylighen, & G. Martin (Eds.), 14th international conference on computer aided architectural design — Designing together (pp. 653–666) (Liège). Rajabifard, A., Kalantari, M., & Williamson, I. P. (2012). Land and property information in 3D. FIG working week. Rome, Italy: International Federation of Surveyors (FIG). Scarponcini, P., Gruler, H. -C., Stubkjær, E., Axelsson, P., & Wikstrom, L. (2016). OGC® land and infrastructure conceptual model standard (LandInfra). Shojaei, D., Rajabifard, A., Kalantari, M., Bishop, I. D., & Aien, A. (2014). Design and development of a web-based 3D cadastral visualisation prototype. International Journal of Digital Earth, 1–20. http://dx.doi.org/10.1080/17538947.2014.902512. Stoter, J. E., van Oosterom, P. J. M., Ploeger, H. D., & Aalders, H. (2004). Conceptual 3D cadastral model applied in several countries. FIG working week. vol. 2004. Subdivision Regulations (2011). Subdivision (registrar's requirements) regulations. Australia. Retrieved from http://www.austlii.edu.au/cgi-bin/download.cgi/cgi-bin/ download.cgi/download/au/legis/vic/consol_reg/srr2011538.pdf. Williamson, I. P., Enemark, S., Wallace, J., & Rajabifard, A. (2010). Land administration for sustainable development. CA: ESRI Press Academic Redlands. Wix, J., & Karlshoej, J. (2010). Information delivery manual: Guide to components and development methods. BuildingSMART International. Retrieved from http://iug. buildingsmart.org/idms/development/IDMC_004_1_2.pdf. Zlatanova, S., Oosterom, P. J. M. V., Lee, J., Lic, K. -J., & Lemmen, C. H. J. (2016). Use of LADM spatial units to define IndoorGML spaces for navigation. Indoor 3D workshop.