*Article* **Innovative Remote Sensing Methodologies for Kenyan Land Tenure Mapping**

**Mila Koeva 1,\*, Claudia Stöcker 1, Sophie Crommelinck 1, Serene Ho 2, Malumbo Chipofya 3, Jan Sahib 3, Rohan Bennett 4,5, Jaap Zevenbergen 1, George Vosselman 1, Christiaan Lemmen 1,5, Joep Crompvoets 2, Ine Buntinx 2, Gordon Wayumba 6, Robert Wayumba 6, Peter Ochieng Odwe 6, George Ted Osewe 6, Beatrice Chika <sup>6</sup> and Valerie Pattyn <sup>7</sup>**


Received: 16 December 2019; Accepted: 10 January 2020; Published: 14 January 2020

**Abstract:** There exists a demand for effective land administration systems that can support the protection of unrecorded land rights, thereby assisting to reduce poverty and support national development—in alignment with target 1.4 of UN Sustainable Development Goals (SDGs). It is estimated that only 30% of the world's population has documented land rights recorded within a formal land administration system. In response, we developed, adapted, applied, and tested innovative remote sensing methodologies to support land rights mapping, including (1) a unique ontological analysis approach using smart sketch maps (SmartSkeMa); (2) unmanned aerial vehicle application (UAV); and (3) automatic boundary extraction (ABE) techniques, based on the acquired UAV images. To assess the applicability of the remote sensing methodologies several aspects were studied: (1) user needs, (2) the proposed methodologies responses to those needs, and (3) examine broader governance implications related to scaling the suggested approaches. The case location of Kajiado, Kenya is selected. A combination of quantitative and qualitative results resulted from fieldwork and workshops, taking into account both social and technical aspects. The results show that SmartSkeMa was potentially a versatile and community-responsive land data acquisition tool requiring little expertise to be used, UAVs were identified as having a high potential for creating up-to-date base maps able to support the current land administration system, and automatic boundary extraction is an effective method to demarcate physical and visible boundaries compared to traditional methodologies and manual delineation for land tenure mapping activities.

**Keywords:** fit-for-purpose; land tenure; land administration; cadastre; UAV; feature extraction; needs assessment

#### **1. Introduction**

The first goal of the sustainable development goals (SDGs)—target 1.4—set by the United Nations (UN) aims to deliver tenure security for all [1]. Strategies to support this goal rely in part on the development of land administration systems (LAS) that formalize land rights that support secure land markets, facilitate poverty reduction and support national development [2]. Broadly speaking, LAS can provide the infrastructure for implementing land-related policies and management strategies and maintain information about people and land involving different organizations, processes, and technologies [3].

Contemporary land administration incorporates the concept of cadastre and land registration, often with a specific focus on the security of land rights [4]. It conceptually fits within the broader land management paradigm [5] with its four land administration functions (land tenure, land value, land use, and land development), ultimately seeking to deliver sustainable development. These functions utilize an underlying land information infrastructure including reliable remote sensing data. It should be noted that in this paper cadastre is considered synonymous with land registry and land administration system.

In sub-Saharan Africa, and in the other developing regions, numerous activities for land tenure recording have been, and continue to be, initiated. For example, in alignment with the SDGs, the Global Land Tool Network (GLTN), an international network of partners setting a global agenda for the improvement of land management and tenure security, develops the so called Social Tenure Domain Model (STDM), a tool for registering formal, informal, group, or individual rights [6].

However, it is estimated that only 30% of the world's population has documented land rights and has access to a formal cadastral system [7,8]. Cadastral mapping is proven as the most expensive part of the land administration system [5]; therefore, there is a clear need for innovation for fast, accurate, and cost-effective land rights mapping. Existing approaches using traditional methods including field surveys often prove to be time-consuming, costly, and labor-intensive.

In response, fit-for-purpose (FFP) land administration suggests technologies should be developed, adapted, selected, and applied to match the capacity and cost constraints of a specific context [4]. The main idea of the FFP approach is to ensure land tenure recording is delivered at scale on a regional and national level, rather than focusing on highly accurate solutions, with less coverage. Three main FFP characteristics are that the land administration systems should focus on the purpose, flexibility and upgradability. The concepts of FFP include principles that cover spatial, legal and institutional aspects on a country level. One of the key principles of the FFP approach is using "general" boundaries extracted by visual interpretation based on aerial images rather than "fixed" boundaries demarcated in the field and measured by a high precision Global Navigation Satellite System (GNSS) technology [9]. Some successful examples where the FFP approach was applied are Rwanda where a Land Tenure Regularisation Program (LTRP) was implemented, and Namibia and Ethiopia with their communal land registration and cadastral mapping [4]. To apply these principles of obtaining general boundaries cheaper and faster, there is a clear need for a new generation of tools and applications that are transparent and scalable [10].

In response, we are developing innovative, scalable methodologies, using remote sensing data and cadastral intelligence, based on fit-for purpose principles to respond to the continuum of land rights [11,12]. The aim of this paper is to assess user needs (in terms of land administration functions), and how the three remote sensing methodologies under development can meet these needs. It also considers how the adoption of these technologies may have governance implications. For the assessment we take into account the above mentioned land management paradigm [5] and FFP spatial and scalability requirements [4] for a case study located in Kenya. The combination of quantitative and qualitative results collected from fieldwork and workshops, taking into account both social (needs assessment and governance) and technical aspects (developed technologies), makes this paper a significant contribution.

The developed remote sensing methodologies for this study include (1) a unique ontological analysis approach using smart sketch maps, (2) unmanned aerial vehicles (UAV) for mapping procedures, and (3) Automatic Boundary Extraction (ABE), based on the acquired UAV images. The sketch maps mentioned above are hand-drawn either on a blank piece of paper or as annotations over existing spatial information, such as cartographic maps, aerial images, or other maps produced via community mapping. As people draw sketch maps based on observations and not based on measurements, the information is not georeferenced, but qualitative relations of the sketched information can usually be considered as correct (with respect to the background information) [13]. The usage of UAVs as cheap, affordable, and easy to use acquisition technology for obtaining high-resolution imagery is emerging for many applications [14,15]. Their applicability in the domain of land mapping was also explored extensively in a different contexts [16–21]. However, there is a lack of studies that evaluate the appropriateness of UAV technology considering the local context and the fit-for-purpose approach. The high-resolution images are usually used for manual delineation of visual boundaries with additional information attached including the ownership and value of the land [22]. However, manual delineation is time-consuming. To register unrecorded land rights more effective in terms of cost and time, innovative and scalable solutions were explored [23–25]. There are clear advantages in using ABE methodologies, therefore, new tools and techniques were developed for scaling up the mapping procedures in support of indirect cadastral surveying based on remotely sensed data [26,27].

The multidisciplinary nature of the current work, using different integrated approaches, and emerging remote sensing technologies, is novel and innovative to the land administration domain. The paper reports on the findings after fieldworks and workshops organized in Kenya, with the purpose of assessing the needs and end users' readiness, the applicability of the developed remote sensing methodologies, considering the needs and how they may affect the governance aspects. In the background section information related to the previous and current land administration system is explained and the study area is described. The overall methodology of the paper is explained in Section 3 and the concrete methods used for each of the assessed remote sensing methodologies are presented. The results are presented in section four, followed by critical discussion, conclusions, and suggested further steps.

#### **2. Case Background and Study Area**

Kenyan urbanization to date has been one of imbalanced growth due to ad hoc identification of urban areas, resulting in skewed distribution and inequality in development [28]. This challenge of land governance has been found to be a significant factor in constraining inclusive prosperity more generally across Africa's urbanization phenomenon [29]. It is particularly evident in contested peri-urban lands emerging as a result of metropolitan sprawl across sub-Saharan Africa [30,31]. In response, one of the key strategies consistently advocated by the international development community is the establishment (or improvement) of a formal land market. Such a techno-economic orientation and focus on market-driven urbanization is evident in many contemporary studies of land tenure that continues to pay limited attention to underlying political aspects of tenure regimes [32]. Land tenure is inherently social and political, and in Kenya, land is also overtly cultural. Attention to the cultural aspects of land is particularly relevant in urbanization in Kenya as, first, a majority of Kenya's land resources are held under customary tenure systems, and second, having remained unrecognized by formal systems since colonial rule, indigenous groups have long borne the burden of Kenya's structural adjustments, which have resulted in dispossession and longstanding tenure insecurity [33].

Situated in East Africa, Kenya covers almost 600,000 km2 constituted of 47 counties with a population of almost 45 million [34]. Approximately 80% of Kenya's land is categorized as arid or semi-arid, with only 15% of this suitable—and fully used—for agricultural production [35]. Since 1963, the land administration in Kenya is under the Ministry of Lands, Housing and Urban Development. The organization structure is presented in [36]. Under colonial and post-independence governments,

Kenya has operated two land tenure systems simultaneously: statutory (based on English property law) and customary. The 2010 Constitution now recognizes customary tenure systems as Kenya's third type of legal tenure, but administrative implementation of this recognition remains in its infancy. The National Land Commission (NLC) is tasked with oversight for all planning processes in Kenya. In relation to FFP, Kenya is one of the countries that first introduced this approach in 1954. Under this major land reform program, land consolidation and systematic adjudication methods were used to determine the parcel boundaries in the rural parts of Kenya. These boundaries were identified, walked, and demarcated by the local inhabitants, based on aerial images, thus it was a participatory approach. As a result so called Preliminary Index Diagrams (PIDs) were produced, which were used to register the rural land parcels in Kenya for many years [37]. Generally, Kenyan cadaster consists of different types of maps, such as survey plans, field notes, registry index maps, aerial photographs, topo-cadastral maps, deed plans, and title deeds, sometimes with variety of names and accuracy [38]. However, most of them are in paper format and are kept in archives (Figure 1). There has been a research also on integrating the buildings into databased and adapting the existing land administration system according to the international ISO: 19152 Land Administration Domain (LADM) standard [36].

**Figure 1.** (**a**) Cadastral map of part of Kajiado; (**b**) archive of land titles.

Kajiado is selected as the case study site for this study as it is part of the Nairobi metropolitan region and is the site of multiple contests for land (Figure 2). Its proximity to Nairobi and Amboseli national parks has also led to increasing human–wildlife conflict being experienced in Kajiado. Currently, approximately 25% of the county's population (of more than 800,000) is urban, almost 50% live below the poverty line, and the population growth rate of 5.5% is higher than the national average [39].

The current land registry map is riddled with information errors stemming from inappropriately scaled maps (resulting in scale errors and boundary overlaps)—the continued use of which introduces further errors in the land registry—which makes it now difficult or impossible to fit new development plans on the original map base [35]. These information issues have resulted in administrative challenges.

**Figure 2.** Location of Kajiado County and neighboring counties.

#### **3. Materials and Methods**

Land management activities rely upon three main components [5]: (1) land policy, (2) land administration functions, and (3) land information infrastructure. In the current research, the major focus is the land administration functions including (1) land tenure, (2) land value, (3) land use, and (4) land development. Specifically, this research is mainly seeking to contribute to improving land tenure including cadastral surveys of determining spatial information on parcel boundaries.

First, user needs were identified. Second, the identified needs were incorporated into a broader assessment of the three remote sensing methodologies, made up of 10 criteria, taken selectively from the land management paradigm [5] and FFP requirements [4]. Third, the governance aspects in relation to the new developments are also analyzed. The entire evaluation methodology is visualized as a conceptual framework in the following Figure 3.

**Figure 3.** Conceptual evaluation framework.
