Comparative study for underlying concepts of land acquisition during post-disaster recovery in Italy, the United States and Japan

Relocation as a consequence of natural hazard disasters is a critical issue worldwide due to climate change, especially during the disaster recovery phase (Balachandran et al., 2022). Land acquisition (LA) is a planning technique used to mitigate disaster risks and minimize future losses (Smith et al., 2021). LA and downzoning have drawn increasing attention owing to the increased severity of disaster impacts induced by climate change and land degradation (Lacoere et al., 2023). Downzoning involves limiting development while reducing construction, as opposed to upzoning, which enhances development (Freemark, 2023). Past literature defines “downzoning” as “rezoning in the downward direction” (Lacoere et al., 2023, p. 191), and it is understood as a land use management approach to prevent development in low-density areas (Newburn and Ferris, 2016). Managed retreat, one form of downzoning related to LA, aims to minimize natural hazard risks by demolishing and removing structures and consequently renaturalizing high-risk lands (Institute for the Environment, 2017).

Downzoning harmonizes with managed retreat through the limitation of development on a specific scale. However, LA does not always connote downzoning; relocation projects can include new developments on higher ground, as in the case of recovery after the 2011 Great East Japan Earthquake and Tsunami (GEJET). Kondo and Lizarralde (2021) found that relocation sometimes produced another disaster risk due to upzoned relocation that created landslide risk at higher ground.

This study defined LA as a method of voluntary acquisition of lands for public purposes, by agreement, including payment by municipalities to landowners. The concept of LA is clearly distinguished from land expropriation, which is compulsory land acquisition with compensation for the purpose of carrying out public works projects.

The time required is a key issue for the processes of LA and relocation. LA for relocation often takes several years or even decades. As a result, some households choose self-driven relocation without waiting for the development of relocation sites in Japan (Kondo and Karatani, 2016). One buyout program in the United States took an average of five years, because the program did not allow application by individuals. Additional time was spent on recruitment, application and decision stages (Curran-Groome et al., 2021).

The trends of LA shift over time, and underlying concepts of LA also vary across different countries. Past trends of land use changes were mainly the result of top-down approaches, using land expropriation rather than LA. In the case of Japan, Suzuki (1990) points out that land expropriation was used as a system for acquiring land in connection with the development of roads and public facilities. However, while there was pressure for dynamic economic growth, and increasing traffic volume in the twentieth century, small landowners had accumulated in existing urban areas of large cities, making it difficult to adjust the rights of each party. This led to the amendment of a legal act in Japan, which protected the rights of landowners and made it possible to make intensive use of land through the construction of buildings.

Similarly, recovery after the 2011 Canterbury earthquake in New Zealand implemented a designated red zone indicating no infrastructure recovery. As a result, this intervention functioned as a compulsory land acquisition (Smith et al., 2021). From a geological perspective, Mach et al. (2019) pointed out that past trends of buyouts in the U.S. were in areas prone to river floods such as the Midwest, while observing recent trends in the U.S. shows an increasing number of buyouts in coastal areas facing risks induced by sea level rise.

From the perspective of disaster recovery and relocation studies, Costa et al. (2022) applied an agent-based model simulation to earthquake scenarios and suggested that place attachment was a key variable that determines if disaster survivors return. Paul et al. (2024) reviewed disaster recovery simulation studies and identified a gap in research about landowners’ decision-making about resettlement or relocation in disaster-impacted lands. While these studies provided critical insights into LA and relocation in disaster recovery, research gaps remain in terms of understanding perspectives on downzoning and time required for LA. This study considers the unanswered question: What are the underlying concepts of LA in the disaster recovery phase from a global perspective? This study qualitatively illustrates the LA program scheme for recovery in three countries – Italy, the United States and Japan – and identifies the underlying concepts distinctive to each country. As both long-term recovery outcomes and international comparisons are understudied, this study is also significant as it includes these aspects as part of distinguishing the varied concepts of LA in post-disaster recovery.

Suggesting a new concept provides a new theoretical framework and alternative approach in practice. In the age of climate change, land use management and land use changes for adaptation are critical agendas, including as solutions for shrinking space considering sea level rise (Haasnoot et al., 2021). Understanding trends and varied concepts of LA is essential for adaptive long-term solutions. This study therefore represents a unique contribution to research on land use changes, by addressing a research gap on long-term impacts of land use management in disaster recovery, as part of the clarification of LA concepts and international comparison of LA case studies.

This study employed a multiple case study approach that aimed to explore common and contrasting results and construct new theoretical insights (Yin, 2018). As understanding the underlying concepts of LA in different contexts is essential, this study seeks to identify a variety of LA models in different contexts to provide input to a multiagent-based model simulation for future works.

Although the hazards and scale of the LA implementation varied among the cases, this study selected cases in three countries to identify differences in timeframes and land use changes in post-disaster recovery. The selection criteria for the three countries followed previous literature (Smith et al., 2021), including relationships to disaster management programs with different governance schemes, sharing similar hazard exposures and having LA programs. Italy was selected as an example of an LA model used in an interim phase spanning emergency response and permanent recovery, while the case in the United States exemplified land use change through LA. Japan was included as a country using LA on a large scale, due to the major role of LA in recovery after recent disasters.

Because the scope of this study is limited to the theoretical description and practice of LA rather than quantitative analysis or hypothesis testing, the methodology involved a literature review and summary of official reports to describe practical land use changes pre- and post–LA. Previous studies on LA focused on planning design and topics such as local capacities and public participation (Smith et al., 2021) or housing market effects (Hashida and Dundas, 2023). However, this study examined the timeframe of LA project implementation qualitatively, with aerial photos used as a complementary data source to confirm land use changes from pre-to post–LA. Although comparison is limited due to the different scales of the case study sites, based on resulting changes in land size, LA projects could be identified as downzoning or upzoning.

Before cross-country comparisons, this section introduces the background of each case study from Italy, the United States and Japan, respectively. In Italy’s long-term recovery process, LA was implemented as part of the emergency response to a disaster. The LA was driven by the Prime Minister’s declaration of a state of emergency and the initiative of the appointed Special Commissioner, who works with the affected regional government to consider emergency response policies, and proceeded with the LA and the development of provisional settlements through the enactment of decree laws and the commissioner’s ordinance (Alexander, 2010). Italian disaster recovery plans typically include revisions to existing plans, impose strong restrictions on built-up areas and avoid the consumption of new land (Nerozzi and Romani, 2014). This means that buildings are not relocated in the event of an earthquake because restoration, renovation and rehabilitation projects are undertaken in compliance with normal-time regulations. However, there have been land and building acquisitions, several buildings have been relocated to the suburbs, and the administration implemented partial LA to create public spaces between residential buildings (Mashiko, 2020). In Italy, LA is rarely carried out through projects directly related to recovery; rather it is implemented only when developing provisional settlements in emergency response.

Buyout programs in the United States are an incentive-based LA method in which the government purchases land in high flood-risk areas, encourages residents to relocate and renaturalizes residential land to mitigate repetitive losses and damage. The Federal Emergency Management Agency (FEMA) and the Department of Housing and Urban Development (HUD) can allocate funds to cover 75% of the budget for these projects. As an outcome of FEMA buyouts, 43,633 households relocated from flood-prone areas over 28 years starting in 1989 (Mach et al., 2019).

Recent growing interest in LA in Japan is due to its use post-tsunami, as one of the basic recovery measures after the GEJET was the promotion of a Group Relocation Program for Disaster Mitigation (hereafter, the Group Relocation Program) to mitigate potential tsunamis. In addition to the voluntary purchase of former residential land in tsunami-impacted areas, this program needed to purchase land, from voluntary sellers, in high-ground areas for the development of new relocation sites.

This case study focuses on programs related to the LA for provisional settlements organized in small towns affected by the 2012 Emilia-Romagna Earthquake. Nerozzi and Romani (2014) reported on emergency response activities by the Italian Civil Protection during the earthquake. Several studies on provisional settlements after earthquakes have focused on the CASE (Complessi Antisismici Sosteniibili ed Ecocopmatibili) project and MAP (Moduli Abitativi Provvisori) projects in L’Aquila, which was affected by the 2009 Abruzzo earthquake. Researchers have evaluated L’Aquila’s reconstruction approach, accumulating knowledge of its problems and ways to resolve them (Alexander, 2010, 2018; Ozerdem and Rufini, 2013; Mashiko et al., 2021; Valent, 2019). The CASE and MAP projects were decided through top-down decision-making, characterized by strong intervention from the central government. The CASE and MAP projects are appropriate for international comparison studies because of their implementation using LA, and programs related to provisional settlements developed in Emilia-Romagna are discussed in more detail.

The 2012 Emilia-Romagna earthquake resulted in 28 deaths, 300 injuries and 45,000 displaced people, with economic losses of EUR 12.2 billion (Franz, 2016). Following the state of emergency, the governor of the region was appointed as a Special Commissioner for 60 days, and commissioner ordinances were established to develop housing modules and public facilities (Franz, 2016). The locations of provisional settlements were selected considering the relationship between the historical city center and the urbanized area. Some municipalities classified provisional settlements as part of emergency response efforts, or as medium-to long-term recovery strategies (Nomura and Satoh, 2016).

While considering the concept of provisional settlements, site selection and LA for each emergency facility were conducted by several commissioners. Emergency facilities can be divided into Edifici Pubblici Temporanei (EPT) and Prefabbricati Modulari (PM). The EPT included four types of facilities – municipal buildings, schools, gymnasiums and libraries. The PMs included three categories – demountable urban and rural prefabricated housing, prefabricated facilities for public service use and prefabricated facilities for educational use (Mashiko, 2024). For instance, site localization and LA for constructing removable prefabricated housing were determined by Commissioner Order No. 40/2012 on September 14, 2012. This order authorizes the acquisition of construction sites (14 plots in 10 municipalities), listed in consultation with local authorities, for public interest and urgency of work, and the implementation of infrastructure works.

In Mirandola, where the majority of prefabricated houses were built in urban areas, two provisional settlements were developed in the eastern and western parts. The eastern provisional settlement was classified for educational uses within the urbanized zone before the disaster; however, it was privately owned agricultural land (Plate 1). The Special Commissioner acquired agricultural land by using parcels as construction sites, which are now owned and managed by the local administration.

This provisional settlement consists of two prefabricated housing sites, two prefabricated schools, three temporary schools, one temporary library and one gymnasium. Table 1 lists the development processes of these facilities. For all the facilities, infrastructure development was initiated ahead of time. Once the site acquisition was completed, construction permits were obtained, and construction commenced and was completed. All the facilities except for the prefabricated housing sites are in continuous use. The two prefabricated housing sites were completely removed in early 2017. The site in lot 5 was vacant; positioned as a civil protection service site and reserved as land for emergency use. Lot no. 9d was converted into a bus terminal, transforming the provisional settlement into a center for education-related facilities and public transportation. Although this transformation of the provisional settlement was not planned from the beginning, it was consistent with the pre-disaster land use plan. When the construction of emergency facilities and the development of provisional settlements were conceived, site selection and LA was therefore conducted in ways that were consistent with and according to existing planning policy, contributing to the devastated area’s long-term recovery.

On Staten Island, New York City, New York, where Hurricane Sandy killed 60 people and caused 65 billion U.S. dollar damages (Institute for the Environment, 2017), many households from buyout promotion districts were relocated; however, some houses remained (Plates 2 and 3). McGhee et al. (2020) obtained new addresses for 323 households in one district (65% of buyout participants) and assessed the exposure and vulnerability of their homes and new residences. Although some households relocated within five miles (approximately 8 kilometers) of the disaster area, more than a few households moved to areas with similar risks and relocated to areas with a high poverty rate (McGhee et al., 2020).

Miao and Davlasheridze (2022) used publicly available project data from OpenFEMA to estimate buyout enforcement factors by county, using a double-hurdle model. The results show that counties with a low percentage of property taxes making up municipal revenue, that is, municipalities that would not suffer as much from population decline due to buyouts, are more likely to conduct buyouts (Miao and Davlasheridze, 2022). In addition, Elliott et al. (2020) used a logistic regression model to determine whether or not to implement a buyout program and showed a bias in program subsidies according to race, and Zavar and Fischer (2021) examined impacts of structural and historic racial discrimination.

Although buyouts aim to reduce the risk of inundation in flood-prone areas, they do not directly reduce the risk of residents who are moving out and carry the risk of racial discrimination. As a voluntary purchase of impacted/hazardous land, buyout programs have no control over post-buyout housing destinations. With selection of destinations left to the landowner, risks of resettlement to areas with a similar natural hazard disaster risk cannot be avoided. As a result, buyout programs contribute to risk reduction of the disaster area rather than of the victims.

The GEJET caused a death toll of 19,775 with an additional 2,550 missing as of 2023 and impacted more than 750 thousand housing units ^[1]. Relocation and related LA programs were the main countermeasures of recovery after the GEJET. The case study in Japan focuses on the (Small) Residential Area Improvement Program, Tsunami Recovery Base Development Program, Group Relocation Program for Disaster Mitigation and Fishery Community Disaster Mitigation Enhancement. As Sasaki (2014) points out, disaster-related public housing programs are not subject to LA in Japan, and land readjustment, a mechanism of replotting land for development, is fundamentally different from LA. Therefore, we excluded the Land Readjustment Program and the Public Housing Program from our analysis ^[2]. Table 2 lists the programs included in this study.

The basic principle of the Residential Area Improvement Program is to improve densely populated, defective housing by expropriating land and constructing new houses or public housing. Historically, the program scheme was implemented for slum clearance before Second World War and was applied to recovery efforts following the Great Hanshin-Awaji Earthquake (Ando, 2004) and a village affected by the Chuetsu Earthquake (2004) (Sawada, 2015). However, the program has rarely been implemented, except for instances of liquefaction damage in Abiko City, Chiba Prefecture and several cities in the GEJET. In Abiko City, 53 houses were removed, and 11 single-family detached disaster recovery public houses were constructed as small-scale improved houses (Plates 4 and 5). One of the advantages of this program is that renters are included as beneficiaries, providing an increasing number of disaster survivors with the option of remaining.

With the GEJET as a turning point, the Act on the Development of Tsunami-Resistant Communities was enacted. The program aimed to rebuild entire communities through inland relocation from tsunami-impacted areas. It is possible to construct housing complexes and public facilities in newly developed areas within municipalities. However, only one project was approved in each municipality, limited by national government subsidies (Ministry of Land, Infrastructure, Transport and Tourism in Japan, 2012).

As of March 2016, five years after the disaster, this program has covered 343.1 ha in 25 areas nationwide. The total area of residential facilities is 73.8 ha (21.5%), with the largest area of 26.7 ha in one municipality. Plates 6 and 7 show the Shin-Yamashita Station project in Yamamoto Town, which relocated the damaged station approximately 1.1 km inland. The area around the new train station will be used as a new compact urban area, and an elementary school will be built alongside the commercial areas (Yamamoto Town, 2018). In addition to the 201 residential lots for sale, 346 detached recovery public housing units were developed. In summary, the Tsunami Recovery Base Development Program is an urban hub inland relocation method that aims to reduce the impact of tsunamis by transforming land use, such as farmland, or redeveloping the city center.

The Group Relocation Program began in 1961 based on the experience of severe floods and landslides, was officially enacted after the 1972 flood (Araki et al., 2023) and played a vital role in recovery from the 2004 Nigata-Chuetsu Earthquake. The use of this program contributed to the largest number of districts and households recovering from the GEJET (Maeda et al., 2020). In addition to homeowners who lived in tsunami-impacted areas, renters could also be beneficiaries since the program could be merged with the Disaster Recovery Public Housing Program. LAs have been implemented in disaster-impacted areas and to acquire land such as paddy fields, forests or areas halfway up the mountain, developed for new residential use as part of the Group Relocation Program. Despite advantages of this program, certain disadvantages of relocation have been reported. Tanaka (2011) surveyed both relocated households and households that stayed at their original location and observed passive reasons for relocation in both groups. These results imply that relocation decision-making in Japan is attributed to groups in the neighborhood rather than to individuals. In the case of the GEJET, it was observed that 10 years after the relocation, neither do local markets actively engage in real estate development, nor are public lands transformed by renaturalization. Thus, the utilization of vacant land remains a critical topic in Japan.

The Fishery Community Disaster Mitigation Enhancement Program (hereafter, Fishery Program) contributed to the recovery of fishing villages after the GEJET. Tomita (2021) categorized the recovery patterns of small- and medium-sized fishing villages into three types –Fishery Program only (Type A), Group Relocation Program only (Type B) and a combination of the two programs (Type C). Tomita (2021) concluded that many local communities selected Type C for recovery from the GEJET because it supported construction work for housing and handled drainage, irrigation water facilities for fisheries and community centers (Reconstruction Agency, 2013). For example, in Kamaishi City, the Group Relocation Program and Fishery Program were implemented in an integrated manner (Plates 8 and 9), and in Tanohata Village the Fishery Program was combined with the Public Housing Program (Plates 10 and 11).

Table 3 shows an overall summary of the downzoning and timeframes of projects in the three countries. From the perspective of downzoning, managed retreat and buyouts in the United States contribute to minimizing repetitive losses and downzoning high-risk and low-density areas. Relocation projects in Japan achieved reduced exposure to tsunami risks along with new developments. Table 4 summarizes the Group Relocation Program data for GEJET-impacted areas. A total of 27 million square meters were designated as relocation-enhanced areas due to tsunami impacts and future risks, with an additional 7 million square meters newly developed in the same municipalities in nearby areas on higher ground. However, it is important to note that not all enhanced relocation areas are equivalent to downzoning and the data do not include other programs. Nonetheless, significant land use changes were implemented to minimize risks and sustain communities with close-distance relocations. Hence LA implemented in Japan can be understood as pursuing safety and living in nearby neighborhoods rather than downzoning. While downzoning is a common goal for LA in U.S. and Japan, it is important to understand its differences.

In terms of timeframes, the cases in Italy had shorter durations compared to those in other countries. LA for provisional settlements in Italy aimed to maintain the functions of public sectors, such as education, and improve the quality of life during evacuation periods. Acquired land was transformed into public education uses or left vacant. Italy’s recovery strategy emphasized conserving the historic town center’s heritage by maintaining public sectors and integrating prefabricated housing. Consequently, public facilities in Italy do not distinguish between provisional (or temporary) and permanent phases. In the case of the U.S., even though the project required at least five years for administrative processes, downzoning in the hazardous area was not completely achieved due to decision-making by individual households, and as a result, vulnerability remains at the area especially for those who decided to stay. Such LA approaches, which respect individual decision-making and property rights, are totally different from land expropriation, while downzoning and retreat is not completed in a shorter time. On the other hand, in the case of Japan, LA projects took more than five years because of reconstruction using collective approaches. While such collective decision-making could achieve the goal of safety, it is questionable if the LA projects in Japan represent a truly bottom-up approach.

As mentioned, the ultimate goal of recovery in Italy was the conservation of the historic town center through provisional settlement and public facilities. Therefore, the underlying concept in Italy can be understood as “tradition/legacy oriented.” The tradition/legacy-oriented approach requires years or decades. However, it contributes to the temporary capacity until the recovery of the main urbanized area is complete. Such a concept of tradition/legacy orientation or seamlessness could be useful in Japan.

The concept behind LA in Japan is a “continuation” of community. Unlike buyouts in the United States, designated destinations for relocation programs encourage disaster survivors to stay within their city/town or neighborhood. In Japanese society, the continuation of the community is important as a basic unit of social welfare, as observed in previous studies (Tanaka, 2011; Sawada, 2015). Otsuyama (2023) also found that LA near the original address enhanced neighborhood revitalization, whereas LA far from the former neighborhood caused a stagnation in neighborhood activities after the 2011 tsunami. Masuo et al. (2022) argued for the importance of preserving impacted historical buildings because of the huge losses of such buildings in the post-disaster phase in Japan. These findings are evidence of the similarity of “tradition” approach (Italy) and “continuity” approach (Japan), which both try to support the survivors as well as their historical homes/communities.

As mentioned earlier, LA projects implemented after the GEJET in Japan with collective decision-making did not take a purely bottom-up approach, and projects restricted individual decision-making. Conversely, LA projects in the U.S. respected individual decision-making but did not fully achieve their goals in a certain period. These limitations show how LA concepts embody an inherent tension between collective or individual decision-making to achieve their goals. Methods and approaches for balancing collectiveness and respect for individual decision-making, as well as attention to the conservation of historical built environment, would be valuable topics for future research.

A land use management technique, LA has diverse underlying concepts – tradition/legacy-oriented in Italy, downzoning in the United States and continuation in Japan. A comparison of recovery after earthquakes in Italy and Japan, indicates how LA in Italy is mainly used for the conservation of historical town centers through provisional settlements. As continuation at a safer place is the primary aim of LA in Japan; relocation destinations are designed within community boundaries. LA in the United States mainly aims for renaturalization and to minimize repetitive losses and damage, transforming hazardous areas through downzoning. As underlying concepts about LA in the three different countries are distinctive but not exclusive, learning from other countries could benefit program designs and contribute to new perspectives on LA policies during the recovery phase.

As this is a qualitative comparative case study, it was limited to providing an overarching summary. While it contributes to the first step of developing an agent-based simulation model, additional quantitative analysis of LA and downzoning is still needed. For instance, this study handled projects at different scales, while the detailed calculation of land size to recognize upzoning or downzoning would be a benefit for deeper understanding. Furthermore, the representative case studies from three countries examined programs and projects after recent disasters. As each country has a long and distinct history of disasters and recovery, narrowly selected disaster events and a single case from each country may limit how representative and generalizable these findings are for each country’s approach to LA. Future research with an expanded number of case studies could contribute to deeper understanding in this regard.

In the future, agent-based simulations based upon the different concepts of LA could contribute to improved pre-disaster recovery planning. Drawing on the various approaches of LA, certain severely impacted urban areas might require provisional settlement near the city center, whereas downzoning would be an option in rural areas that suffer from depopulation. Based on the global understanding of LA concepts gained through this study, building a recovery simulation model is the next step forward to advance the understanding of LA, and develop practical applications to support future disaster affected areas.

The authors appreciate Dr Sachi Suzuki and anonymous reviewers for the valuable comments.