The making of cities after disasters: Strategic planning and the Central Italy temporary housing process

.


Introduction
Temporary Housing (TH) is an urban management challenge which calls for a better understanding of TH processes and connected regulatory frameworks to ensure it provides disaster-displaced communities with adequate housing and neighbourhood conditions.
Many researchers pose that TH management issues can be best addressed through pre-disaster strategic planning as shifting TH housing implementation from reactive to proactive, in principle, can help reduce costs, increase speed, improve cultural adequacy, and create public value (Abulnour, 2014;Da Silva, 2007;Edgington, 2022;Johnson, 2007;Krieger, 2019;Sukhwani et al., 2021).
Strategic planning must provide mechanisms for dealing better with change, including through forward thinking and capacity improvement.Pre-disaster TH plans should ensure that the necessary arrangements and resources are in place when their activation is needed, even when the context of the disaster is unknown.What requires establishing standard procedures for implementing TH provision protocols and addressing issues of e.g., organisational design, governance, spatial planning, logistics and stakeholders' coordination as part of disaster preparedness activities (Chang et al., 2010).Strategic plans should also cater for the analysis and assessment of TH operations to support the continuous improvement of future disaster recovery ventures (Sienkiewicz-Małyjurek, 2014).
Although the importance of strategic planning is widely recognised, and even formally codified in countries such as the USA, Japan and Italy, it is still unclear how it contributes to the delivery of quality TH assistance in disaster-stricken cities (Oggioni et al., 2019).
Existing research has mainly discussed the application of strategic planning to emergency management in conceptual terms (Choi, 2008).Although authors relate failures in post-disaster TH response (e.g., after hurricanes Katrina and Ike) to flaws in strategic planning at different governance levels (Sapat et al., 2011), more research is needed to create a solid evidence base for developing policy recommendations.
Only a few empirical studies describe the use and characteristics of strategic planning in emergency management and assess its overall impact on program quality and operational effectiveness (Manning, 2020).Research around TH public procurement is also limited (Tanaka, 2012) despite pre-set procurement is a key element of strategic planning, which can reduce corruption and enable setting design specifications that meet local needs (Johnson, 2007).
Since "best practices […] only emerge in situations where there is an effort made to glean lessons from experiences" (Greer & Trainor, 2021;p.321)this paper presents a 'catalyst' case study analysis in which the strategic TH plan of an exemplary case in Italy is scrutinised together with its outcomes.
The presented study reconstructs and audits the process of public procurement and delivery of TH, considering how this had been originally planned and was locally implemented post-disaster.It adopts an original post-mortem analysis approach based on business process modelling to map the prescriptive process, identify strategic planning flaws, and critically discuss links between TH planning, societal factors, and spatial transformations.Beyond providing insights plus practical recommendations for policy and practice in the analysed case, the paper suggests the opportunity to apply a similar analysis to study other strategic planning experiences in the future to support comparative research trends, which are currently emerging in the field (Edgington, 2022).The proposed approach can facilitate benchmarking and lessons learning in different contexts, which have a longstanding history of managing housing recovery after large-scale disasters through the use of TH.
The article is structured as follows: the research background is presented in Section 2. Section 3 illustrates the analysis methodology and Section 4 the conceptualised process.The analysis results are presented in Section 5. Their implications for strategic planning and policy in a broader context are outlined in Section 6, while conclusions can be found in Section 7.

Research background
Following an increase in the number and severity of disasters worldwide, The New Urban Agenda demands a better consideration of contingencies in urban planning (United Nations, 2017).This includes integrating Disaster Risk Reduction principles in direct TH delivery planning, by enhancing preparedness for effective TH response and by 'Building Back Better' in disaster recovery and reconstruction (UNDRR, 2015).
As a matter of fact, TH contributes to the construction of sociocultural relationships and to the formulation of conjectures on the future of a city, involving heterogeneous sustainability domains (Adeyeye & Albadra, 2008;Rotondo et al., 2020).Therefore, although it is critical to support a swift recovery after disasters, its poor planning can contribute to perpetuate urban vulnerabilities (Alexander, 2013;Boano, 2009;Bolin & Stanford, 1991), by widening social and spatial disparities (Contreras et al., 2017;Imperiale & Vanclay, 2019) and accelerating depopulation processes and economic impoverishment (Alexander, 1989).
Although these represent important contributions, knowledge of the critical decision-making components and processes which determine the success or failure of housing assistance programs is still limited (Fayazi & Lizarralde, 2018;Pezzica, Cutini, & Bleil de Souza, 2021) due to an insufficient understanding of post-disaster urban transition processes (Borsekova & Nijkamp, 2019).What requires further explorations on the process underpinning the making of cities after disasters and connected forms of inhabiting.

The Italy case
Technical bottlenecks, delays, governance issues and corruption, as well as the long-term permanence of past TH schemes in many areas of Italy (Pezzica & Cutini, 2020) can be associated with a non-systematic consideration of space and time in TH planning (Forino & Carnelli, 2019).Italian TH programs have, in fact, frequently lacked an organic plan linking post-disaster recovery and reconstruction through development (Emidio di Treviri, 2018).Moreover, when a long-term vision was present (e.g., the 'New Messina', the 'city-territory' in Belice, the 'New Towns' in L'Aquila etc.), the political ambitions did not match the available resources and the local socio-economic context (Parrinello, 2020).This led to informal urban growth (in Messina), urban sprawl (in L'Aquila and in Belice) and to the decentralisation of several small historic towns (Irpinia and Molise), eventually generating forms of precarious urban living and physical degradation.
Within this context, applying later corrective planning strategies has often proved hard, due to high costs, planning constraints and conflicts with the occupants.So, pre-disaster TH procurement contracts were explored within the context of a broader reformation of the Italian civil protection system, to enable a prompt activation of resourcing agreements and multi-actor colloborative procedures.
A strategic agreement for the TH supply and delivery known as "Accordo Quadro", with validity 72 months (6 years), was first signed by the Italian National Department of Civil Protection (NCPD) and CONSIP in 2016 (CONSIP & NDCP, 2014), just before a first devastating earthquake hit an internal area of Central Italy in August 2016.The subsequent seismic swarm (magnitude 5.5-6.5)invested 139 municipalities and dozens of towns located in an area of 8000 km 2 in the central Apennines encompassing four different regions (Fig. 1).The disaster costed hundreds of human lives, the displacement of 49,844 people (mainly in hotels on the Adriatic coast) and 23 billion and 530 million euros of estimated material damages (NDCP, 2017), making the immediate activation of the strategic agreement necessary.7782 people applied for a temporary house -called S.A.E., an acronym for housing solution during an emergency -to remain close to their former homes.This required the construction of 228 TH sites (NDCP, 2018a), what was managed mainly by the relevant regional and municipal administrations -as the NDCP assumed an advisory and coordination role -together with the suppliers that had been previously contracted via the "Accordo Quadro".
While it is early to draw definitive conclusions regarding the outcomes of the Central Italy TH assistance program, some critical limitations of the "Accordo Quadro" have been already identified.Emidio di Treviri (2018) highlights flaws in evaluation criteria, priorities and consideration of technical elements and hence questions its suitability for supporting a timely and contextual TH assistance delivery.Oggioni et al. (2019) note delays and lock-ins in its post-disaster execution, linked to timing, technical, and managerial issues, which cast doubts on the suitability of the strategic plan for managing the response to a largescale disaster.Moreover, Forino and Carnelli (2019) and Rotondo et al. (2020) pose that many of the temporary settlements already built in Central Italy are likely to become an unplanned permanent presence within a context of growing socio-spatial fragmentation, depopulation, and inequality.
Since no model of the TH supply and delivery workflow existed, previous research mainly provides context-bound and/or subjective insights and cannot offer a comprehensive enough explanation for the identified flaws.To advance understanding of the benefits and pitfalls of strategic planning, the paper analyses the Central Italy TH supply and delivery as a business process, where the 'who', 'what', and 'when' are explicitly mapped ex-post, using a standard notation system.This enables shedding some light on what worked and what did not work, by linking specific process components to known issues and by pinpointing additional gaps emerging from the analysis of the model, which deserve attention and improvement.

Materials and methods
Since strategic planning for TH management is underexplored from an empirical perspective, a post-mortem case study analysis is needed to act as a catalyst for further studies.The Central Italy case was chosen for this research because this is where the first Italian strategic agreement for TH supply and delivery was enacted.This is also the latest major disaster in Italy which involved a large-scale deployment of prefabricated TH units (see Section 2.1).
The research design combines a formal mapping of the prescriptive 'as-planned' TH supply and delivery process with a qualitative examination of 'as-happened' deviations to analyse -without aiming to be to be exhaustive -critical activities and process components (Fig. 2).To this end, the study exploits business process modelling techniques from process management engineering (Section 3.1) and a meta-analysis of heterogeneous data sources (Section 3.2).
The practical value of transferring methods from business process management to improve disaster response has been discussed by several researchers (Hofmann et al., 2015;Nunavath & Prinz, 2015;Soini & Polancic, 2010), who have noted the common orientation of business and disaster response processes to pursue predetermined objectives.In particular, the analytic capacity of business process management can advance strategic planning through enhanced process automation, quality, and transparency (e.g., via process performance analysis, simulation and optimisation), by modelling activities and assigned responsible organisational units.Although business process models do not describe how coordination between actors is achieved (i.e., methods and tools adopted) and if human resources are overloaded, they explicitly represent data and dependencies between activities and translate multi-actor interactions into an ordered process choreography.What enables identifying issues and patterns plus improve accountability by increasing awareness of the process; allowing proposed changes to be discussed and agreed upon by stakeholders in advance.
Nonetheless, their application in TH management is still underexplored due to the challenges posed by rapid updating procedures which, differently from what happens in the industry, are not routine and consolidated and are executed by several operators unable to provide complete process information.As it happens in certain domains which traditionally operate in responsive mode, in TH management new process models are commonly generated in the field and possibly formalised only later.Strategic plans are also updated after disasters occur to respond to situational changes and environmental constraints.Therefore, some limitations affect the possibility to directly use off-the-shelf business process modelling tools for TH management.These include the difficulty of dealing with unplannable execution contexts (as understanding of the intended course of action for emergency management is often incomplete), a static representation of organisations, and process modelling tools having a fixed terminology and meta-models which could clash with those of humanitarian organisations, potentially leading to confusion and failure (Betke & Seifert, 2017;Peinel et al., 2012).Such limitations, however, do not apply to an ex-post analysis of the TH process.What requires a different approach to both process modelling (exclusive use of official data sources to extract the 'as-planned' workflow) and analysis (qualitative exploration based on all available data sources instead of quantitative simulations).In this paper, the identification of TH management planning issues is the result of a reflective process which encompasses the modelling and analysis of the TH process, where critical issues are associated to: • ad-hoc sub-processes not clearly formalised, which cannot be modelled.• blocking activities to which an undefined waiting time is assigned.
• chains of sequential activities where the disruption of one means the process' flow is interrupted.• loosely defined process components, which do not allow simulating the flow of activities.C. Pezzica et al.Despite this approach may appear to constraint the analysis to technical and procedural aspects, the way the model is conceived and analysed aims to overcome this limit.For instance, although several elements of the TH supply and delivery process could be evaluated individually and further deconstructed, the complexity of the process representation is kept at a manageable level so that its overall analysis can inform changes in future TH strategic agreements.This choice enables performing an assessment of the process, which -if not exhaustiveis at least more explanatory, as it manages to keep the study in focus.Additionally, it helps overcome difficulties in crystallising processes which are temporally constrained and hard to verify in full with stakeholders in order to enable criticism and ideas for improvement.The meta-analysis supports the modelling effort and enables a nuanced discussion of results via theory-informed interpretation, contextualisation, and triangulation.

Business process modelling and notation
In practice, mapping the TH supply and delivery process as a business process means representing the former as a set of coordinated activities (i.e., units of work that require resources to be performed) executed by organisational units in interaction with partners, which jointly realise a certain goal by passing and creating physical and/or information objects.
Building an explicit representation of this process requires the use of a precise and flexible notation system such as the graphical modelling language known as Business Process Model and Notation 2.0 (BPMN 2.0): a leading standard (ISO 19510), used by most modelling tools for specifying business processes in a flowchart (O.M.G., 2014).
Besides activities, BPMN models events (start, end, boundary, intermediate catching and throwing), connectors and gateways controlling their flow (e.g., approval processes are modelled as event-based gateways).BPMN includes horizontal graphical components called "pools", which represent organisational units, each one undertaking a business process.The assignment of activities to resources is done by placing them in a pool and when a sequence flow crosses a pool (always via message flows) this indicates that a work is handed over to another actor.
A cloud-based service called ®Signavio, was used to create a digital model of the TH supply and delivery process.Among other things, this can perform automatic syntax checks and enable generating multiple process instances from one baseline process model.

Meta-analysis
A meta-analysis was used for providing theoretical grounding and geo-temporal contextualisation to the BPMN workflow reconstruction and subsequent qualitative assessment.
Initially, this required searching for official data sources to generate a sufficiently accurate unbiased description of the TH supply and delivery process 'as-planned' (Section 4), by combing through local regulations, rules, directives, besides conventions and adaptations.What was facilitated by the unprecedented availability of information about public administrations' activities, enforced by the legislative decree n.33 (Decreto Legislativo 14 Marzo, 2013) to favour control over the pursuit of institutional functions and the use of public resources.The core technical document used for modelling this prescriptive process is the Capitolato Tecnico D'Appalto (CONSIP & NDCP, 2014).In addition to this, the model was built considering all the relevant ad-hoc NDCP ordinances, i.e.Ocdpc n. 394, 405, 510, 518 and 538 (NDCP, 2016a, 2016b, 2018b, 2018c, 2018d), guidelines (NDCP, 2009) and integrative regulations (Presidenza del Consiglio dei Ministri, 2017), which were promulgated in due course to fill regulatory gaps 1 as a response to emerging issues in the execution phase.These comprise all official documents available at the national level.
Then the 'as-happened' workflow deviations were qualitatively audited using information from the National Associated Press Agency (ANSA, 2017a, 2017b) and the NDPC ( 2019), open geodata about all TH sites (NDCP, 2018a, 2020), plus secondary information (Goodwin, 2012) such as reports from 17 in-depth interviews and 98 questionnaires with stakeholders (Emidio di Treviri, 2018;Oggioni et al., 2019), among others (Chioni et al., 2021;Rotondo et al., 2020), to represent the widest case history and variety of problems.This, together with civic hacking data (TCI, 2016) and public life photographic observations (Gehl & Svarre, 2013) made by the authors during two field visits in 2019 and 2020 (Figs. 6 and 8), were reviewed to assist results' interpretation and triangulation in discussing gaps between the process 'as-planned' and 'as-happened' following a theoretical saturation protocol that applies to case-study research (Bryman, 2016). 1 For instance, the strategic agreement capped only the cost of the TH units, but the circular subsequently introduced reference prices for urbanisation works too, together with additional design constraints for TH layouts.

Central Italy TH supply and delivery process
The TH process is initiated when the technicians of the local administration receive the request for a TH unit from a resident and is terminated when this is accepted by the requesting person.The modelled activities are grouped according to the following ten steps: 1. TH request logging.2. TH site proposal elaboration.3. TH site proposal assessment.4. Land acquisition and TH supply order submission. 5. TH production and layout proposal.6. TH site layout proposal assessment.7. Executive TH site project submission, approval, and TH production.8. Urbanisation works and foundation construction.9. TH supply.10.TH handover to the mayor.
Step 1: TH request logging.Once a TH request is received from a candidate, the local administration (with the support of the NDCP) approves or rejects it based on eligibility requirements for TH assistance and damage survey data.If the request is invalid, it is rejected, and the candidate informed.If the request is valid (i.e., the former house of the displaced person is in the red zone or declared inhabitable in an AeDES assessment2 with an outcome of E, or F if not quick to resolve,3 NDCP, 2016a), the applicant is informed, and the request is added to a waiting list.
Step 2: TH site proposal elaboration.After all the needs are assessed, the local administration's technicians elaborate the proposal for the candidate TH site(s), send it to the Regional Administration (RA), and wait for its approval or rejection from the regional offices.If the proposal is rejected, they elaborate a new proposal, otherwise they define the procedure for the TH site land acquisition in agreement with the RA.
Step 3: TH site proposal assessment.The RA receives the TH site proposal from the local administration and coordinates its evaluation.This involves assessing the hydro-geological situation of the proposed site(s) and verifying the need of an evaluation of environmental impact called "Valutazione di Impatto Ambientale" (VIA).If the VIA is necessary, they request it, and the required mitigation measures are proposed when appropriate.If the VIA is not needed, only the hydro-geological assessment is considered.If this is negative, a request is sent to the municipality for a new proposal.Before approving a proposal, the RA must evaluate possible associated direct and indirect costs (e.g., due to poor services' availability), including the absence of alternative options.When the proposal is approved, the decision is communicated to the municipality.After confirmation of receipt, the RA defines the land acquisition procedure, together with local authorities and with the support of the NDCP.
Step 4: Land acquisition and TH supply order submission.If the land is already public this step is omitted.If the land is private the expropriation procedure is initiated, and the land is acquired.If the owner opposes the decision of the public administration, he/she can appeal to the regional administrative court, "Tribunale Amministrativo Regionale" (TAR), with consequent potential delays.Once the land is acquired, the RA prepares the TH supply order, including the TH site's delivery report, and sends it to the supplier who had signed the binding strategic agreement.
Step 5: TH production and layout proposal.Once the order is received by the supplier, it can be accepted or not; in which case the RA issues the order to a different supplier and applies a penalty to the former. 4When accepting, the supplier signs the delivery report (which formally hands over the TH site to the supplier) and must submit the design of at least two masterplans with different TH arrangements within 5 natural days together with a weekly construction schedule for approval by the administration. 5In case of delays or errors the supplier pays a fine.6 Step 6: TH site layout proposal assessment.Once the proposed masterplans are received, the municipality, in agreement with the RA, either chooses one and sends the approval to the supplier, or requests amendments/integrations by fixing a new suitable term for the resubmission of the reviewed version.Special requirements and their relative costs need to be computed separately to facilitate the assessment by the administration.Once the layout is approved the communication is sent to the supplier.
Step 7: Executive TH site project submission, approval, and TH production.After receiving the layout approval, the supplier must submit to the RA the executive project of the urbanisation works and TH foundation systems -complete of the necessary environmental and geological surveys and of structures and plants for the TH units -within 20 natural days and wait for approval.After receiving the project, the RA approves it or requests amendments.When the project is approved, the RA prepares the tender for the urbanisation and foundation construction works, specifying delivery modes and time schedule.If the executive project evidences the need for ground consolidation and safety works, these are included in the primary urbanisation bid.The approval of the executive project enables the supplier to start producing the TH units.
Step 8: Urbanisation works and foundation construction.After winning the bid and accepting the job, a third-party constructor starts the execution of urbanisation works.The TH supplier must monitor them and check their compliance to the executive project.If mismatches are found, these need to be reported to the works' direction office (employed by the RA), otherwise the supplier is considered responsible for errors and connected delays.If no-cost amendments to the executive project are needed due to unforeseen circumstances related to the urbanisation works, the public administration requests these to the supplier, who must promptly provide the amendments.
Step 9: TH supply.After receiving communication on the handover of the TH site with the finalised preparation works from the public administration, the supplier has 30 consecutive days to deliver at least 50 % of the TH units ordered and 60 days to deliver all the rest. 7This involves bringing the units to the site, mounting and anchoring them to the foundations, linking them to the services (for up to 30 m from the perimeter of each unit), supplying the furnishings, communicating the end of the works, providing a certificate of plan conformity, and cleaning up.
Step 10: TH handover to the mayor.Once the TH units are handed over to the public administration, the RA employs a technician to perform a static structural audit and, if the outcome is positive, the administration sends the supplier a certificate attesting the installation has been completed and emits the invoice.Otherwise, the supplier needs to make up for the mistakes.After the invoice emission, the supply process ends, and the TH units' assignment procedure to end users begins.
The assignment procedure can be only guessed in its general lines as the documents used to build the model do not discuss this part of the process, which is usually managed by the local administration.The BPMN includes a reference to it, to highlight that potential issues and delays may arise whenever the public refuses to occupy the assigned TH units.From this point onwards, the local administration and the occupants need to collaborate to ensure that issues are promptly reported and resolved.However, this part falls out of the scope of this analysis, which ends with the hand-over of TH units to the mayor.

Modelling assumptions and choices
Building a comprehensive and explanatory model of this process (Annex A), required making the following assumptions and modelling choices: • The analysis focuses on candidate TH occupants, who send the request for a TH unit 8 to be approved by the public administration, 9 although a technician may do this on their behalf (Fig. 3).• The model does not consider re-assessments in case of aftershocks, as these are new instances of the same process.It represents one of the many parallel and subsequent cycles which occur after a disaster, and considers the data collected during the corresponding time window.• The model does not take into consideration cases where less than five TH units are built, which differ from the modelled scenario only in that required urbanisation works are reduced to the essential minimum (i.e., foundations, connection to services and road access).• The possibility to rent a plot of land from a private owner and return it in pristine conditions is excluded, because this option was ruled out by the 2017 ministerial circular (Presidenza del Consiglio dei Ministri, 2017), which stated that the construction of important urbanisation works would have made this an economically unviable option.• To improve clarity, the bidding process of urbanisation works appears exclusively managed by the RA, despite evidence that, in certain cases, this task was delegated to the local municipalities (Oggioni et al., 2019).• The TH site proposal assessment (Fig. 4) is modelled as a collapsed sub-process.Here, a choice was made to evaluate the need of a VIA only after receiving the results of the hydrogeological assessment, because passing the former is compulsory for the approval of a TH site proposal.• A collaborative pool called "Regional and local administration (with the support of the NDCP)" is created to separate the activities with a higher conflict potential and to avoid the duplication of tasks in the model, which would have made it less readable.

The process 'as-planned' and 'as-happened'
Fig. 5 shows the 10 steps in which the process described in Section 4 was subdivided and highlights in different colours where technical clashes have been identified.
Steps 'external' to the TH units' supply process, for which no reference timeframe was established in the strategic plan (Section 5.1), are flagged in red as they exert a high influence on TH delivery time.
Additional issues raising from uncontrolled ad-hoc procedures (e.g., deal with complaints from people and appeals from landowners to TAR), which are flagged in yellow, are linked to an ill consideration of critical process components (Section 5.2), i.e., the role of disaster-affected people and data requirements for process monitoring.

Critical process activities
This section examines the identified critical process activities: damage and TH need assessment; TH site selection, and land acquisition; TH site layout approval; management and execution of primary urbanisation works.

Damage and TH need assessment
Damage assessment is a crucial element of strategic planning for TH management as it establishes who is entitled to apply for direct TH assistance.However, it presents challenges linked to: (i) area coverage (if the disaster hits a large territory, especially if the area is hard to reach from the outside and there are issues travelling around it); (ii) aftershocks; and (iii) availability of professional volunteers for the entire length of the emergency.In the Central Italy case this step presented difficulties linked to bureaucratic problems, the lack of technicians with the skills required to complete the AeDES 10 forms (Italian acronym for usability and damage in the seismic emergency, NDCP, 2009), and slowdowns due to the effects of the second seismic wave (Emidio di Treviri, 2018).Matters of urgency and the delays provoked by the inefficiency of the AeDES filing system as implemented at the beginning of the emergency, motivated the NDCP (2016b) to change the existing policy and adopt new forms, called FAST, which are quicker to fill out and require only little prior training to use.

TH site selection
The selection of TH sites also has a significant impact on TH delivery time.In Central Italy, around 6 months after the first quakes, the viability of the proposed TH areasas well as the possibility to minimise their number -was still under evaluation in 16 of the 24 affected municipalities in Abruzzo, in 23 (out of 57) municipalities in Lazio, while comparable delays were reported in the Marche and Umbria regions (Oggioni et al., 2019).In the municipality of Fiastra, the elaboration of the TH site proposal took around 2 months while the approval from the Marche region took another 3 months (Emidio di Treviri, 2018).An important barrier for prompt TH site selection, reportedly, was the lack of suitable (public) areas respondent to TH dimensional and soil safety requirements, such as in the case of Visso and Camerino, which in some cases made further technical assessments or land arrangement projects necessary, with consequent delays (Emidio di Treviri, 2018).
Arguably, connected issues could have been avoided or greatly reduced if preventive measures, such as pre-reserving candidate TH sites (Félix et al., 2015), had been put in place as part of local-level strategic planning.This should have also helped striking a better balance between priorities of landscape ecological preservation and the necessity to support a prompt and equitable disaster recovery.
8 This is not always the case as people are given the option to request other forms of housing assistance such as cash grants for house renting.This exclusive decision is modelled using a XOR Gateway, known in technical jargon as the exclusive data-based gateway, (diamond with an "X" in it).This component evaluates in order all outgoing sequence flows and enables only that which evaluates to 'true' to continue the process.
9 Checks on the formal correctness of the request are done by the public administration by looking at the submitted documentation and the damage assessment data.Following the check requests can either be approved or rejected.This checkpoint is modelled as an event-based Gateway (diamond with a pentagon), which enables deciding based on events.In the model, each outgoing sequence flow, which exits the gateway needs to be connected to an intermediate catching event (e.g., rejection/approval message received).
10 AeDES forms have been used in Italy since the 1997 earthquake in Umbria.
They were conceived to support a rapid detection of damage, the definition of emergency measures, and to assess the safety of surviving houses or facility buildings with ordinary structural typologies, i.e., masonry, reinforced concrete, steel, wood.

TH site layout approval
Further slowdowns can happen during TH site layout approval following conflicts between local stakeholders and national civil protection authorities, caused by roles' overlaps and competing perspectives.
In Central Italy, there was occasional dissent among the NDCP, the LA and the RA over TH layouts choices, which in the case of Torricella Sicura required the intervention of the head of the NDCP to solve the dispute (Oggioni et al., 2019).This is again a strategic planning issue as the "Accordo Quadro" indirectly limited the power of local administrations over TH layout decisions; going against the principle of "centralised planning and decentralised execution" (Choi, 2008;p.12).A better organisational design should have guided strategic planning in establishing the balance between short-term goals and long-term outcomes (Davidson et al., 2007).
The BPMN shows that the strategic plan also did not consider critical sustainability factors such as activities related to the long-term maintenance and future use of the TH units and sites.A better definition of typological specifications (Bologna, 2007;Pezzica & Cutini, 2021) and an early consideration of end-of-life scenarios for candidate TH sites (Pezzica et al., 2020), could have reduced conflicts while limiting the influence of a few individuals on key planning decisions.
Disputes over TH layout choices and connected delays are likely to have been further exacerbated by the installation of irreversible foundation systems, whose technical requirements were not stated in the strategic framework agreement.The lack of broader sustainability constraints in TH layouts (Hosseini et al., 2021) implied that TH suppliers opted for the most economically and technically convenient solutions, being one storey layouts (Fig. 6), with consequent high consumption of land, materials, and energy resources.Additionally, the criteria used to fix a new suitable term for the reviewed project resubmission in step 6 (TH layout proposal assessment) should have been clearly defined to prevent delays as, on average, the elapsed time in this phase more than doubled the NDPC target of 30 days.

Management and execution of primary urbanisation works
TH management problems related to primary urbanisation bids to be considered in strategic planning include bureaucratic slowdowns, poor management of construction works, and defective TH design and  C. Pezzica et al. planning.In the Central Italy case, the "Accordo Quadro" set deadlines on design but not on tendering.Indeed, no clear audit system was identified for urbanisation works.
The bids managed by the municipality of Accumoli, and that for the construction of the Zona industriale A TH site in Norcia, took more than one month (Oggioni et al., 2019), i.e., around four times longer than anticipated (NDCP, 2018a).In Borgo1 TH site in Arquata del Tronto, delays occurred in the execution phase due to mistakes in the planovolumetric plans.In other cases, ad-hoc agreements had to be signed last-minute with external agencies prior to commencing the construction works.The Autonomous Italian Road Company (ANAS) had to be consulted for building Pescara del Tronto TH site close to a motorway, while the Ministry of Cultural Heritage had to be involved for building in the Zona industriale A, where archaeological remains were found.Whereas the latter is an unforeseen circumstance, the former issue could have been anticipated and settled as part of strategic planning at the local level, which should cater for the formalisation of aid agreements with other jurisdictions (Manning, 2020).
The lack of coordination between external construction companies represented a barrier to the timely preparation of Zona industriale A TH site in Norcia, which was further slowed down by trial-and-error attempts to correct the design of a few TH units to enable access for disabled people (Oggioni et al., 2019).Additionally, a scarce consideration of urban services in strategic planning caused later issues with water drainage and sewage systems, besides spatial accessibility problems impacting public transport.Reducing the risk of increased hydraulic loads connected to the construction of the TH settlements required carrying out additional works in Camerino (NDCP, 2018b) and in Norcia (NDCP, 2018c).Similarly, the construction of a new control system for rainwater in Collemagrone2 TH site in Amatrice became necessary, as well as the expansion of the ring road of Accumoli for ensuring a safe road access to the TH sites (NDCP, 2018d).
While the examples above evidence operational problems generated by gaps in strategic planning applied to TH management, the following best managerial practices also emerged, which could be formalised to streamline the process in the future.In Campo 0 (Amatrice) and San Pellegrino (Norcia), the TH site location approval was done in parallel with the TH supply order, thanks to the efficient coordination of the responsible administration with the NDCP (Oggioni et al., 2019).Moreover, in several TH sites, the installation of the TH units was done while urbanisation works were still in progress (Fig. 7).This situation is modelled in Annex B using BPMN to show how the 'as-happened' variation has improved the 'as-planned' baseline process.
The results nonetheless indicate that, while executing more tasks in parallel can bring some benefits in TH management, this strategy can provide only limited advantages if the other issues which have been identified (Table 1) remain unaddressed.

Critical process components
This section examines the role of citizens in setting minimum urbanisation and housing requirements and in changing the regulatory framework (Section 5.2.1), besides issues related to process monitoring and management (Section 5.2.2).The inclusion of these aspects in the analysis enables reflecting on how TH assistance provision is influenced by preparedness activities (Nigg et al., 2006), governance/power issues (Valent, 2019) and disaster-activated community resilience mechanisms (Bolin & Stanford, 1991).

Community participation
The 'as-planned' BPMN shows several public and private actors interacting and making decisions throughout the TH supply and delivery process, but the displaced people appear to have a marginal role in this.Despite the model only captures the role officially assigned to citizens, which does not necessarily reflect how end-users impacted the process in reality, 11 it flags yet another core weakness of the strategic plan.Indeed, involving local communities in the process is known to lower the risk of unexpected or unsustainable planning outcomes (Jha et al., 2010;Platt & Drinkwater, 2016;Shariat Alavi et al., 2021) and, in reality, the relational and physical resources of local people helped compensate for the design flaws of the TH units; which often required maintenance and additional finishing works, such as applying a suitable protective varnish to the wooden floor of the front porch (Emidio di Treviri, 2018).
Citizens played an important role in informing requirements for building better TH units and sites, by reacting to managerial decisions (ministerial circular of the 16/1/2017, Presidenza del Consiglio dei Ministri, 2017) which, stepping back from original guidelines, limited urbanisation works to essential services only (e.g. by cutting on the provision of trees and street furniture).These generated inequalities between social spaces funded by the public and through private donations and contrasted with wishes of the local population for better collective structures and gathering spaces (Chioni et al., 2021).Citizens' aspirations, hence, culminated in some bottom-up initiatives (e.g., in Capricchia and Cossito, Amatrice), which strengthened local community bonds (Emidio di Treviri, 2018).In August 2020 the authors observed numerous physical adaptions made by the people to TH units and sites, including the: (i) use of fences to delimit private space; (ii) closure of the external porch to enlarge the houses; (iii) addition of tables in the common external spaces to trigger socialisation; (iv) use of available green areas for urban gardening (Fig. 8) to create food security, (Salaheldin Ismail Elsayed, 2018).What highlights the benefits of providing flexible/adaptable TH layouts (Davis, 1978;Wagemann, 2017) as this empowers communities to independently address their needs and enables coping better with uncertainty, since appropriate design specifications can be decided only once the context of the disaster becomes known (Johnson, 2007).
Citizens played an important role also from a policy perspective.Although TH should be delivered promptly to fulfil their role and prevent that squatter or improvised shelter are created to fill the housing gap (Bashawri et al., 2014;Chen et al., 2014;Davis, 1978;Félix et al., 2013), exposing people to new vulnerabilities, the Central Italy TH assistance process was excessively slow.The delays prompted individuals to buy TH solutions from the market and install them illegally on private plots; often in agricultural land and/or in unsafe locations, mostly devoid of urbanisation services.The case of a woman aged 95, who was evicted by public authorities from an unauthorised wooden temporary house located in a landscape protected area subjected to high hydrogeological risk in Moreggini di San Martino (Fiastra) made the news, as the woman then decided to occupy an old container from the 1997 Umbria and Marche earthquakes to remain close to her destroyed house (ANSA, 2017a(ANSA, , 2017b)).The case put pressure on the government to authorise the 'temporary' use of hundreds of private TH solutions in the seismic crater.Finally, a new law (n.89/2018,later refined in the decree law n.162/2019)established that private TH units could be used until the end of the reconstruction or until the assignment of an adequate public accommodation if they met certain conditions such as the reversibility of the TH construction.
While the situation demanded such a resolution, Tanaka (2012) warns that such inconsistencies between pre-disaster directives and 11 A systematic mapping of community inputs in the TH process would require dedicated studies possibly involving multiple field visits, engagement with local activists, and the application of appropriate social science research methodologies; what falls beyond of the scope of this paper.
C. Pezzica et al. post-disaster response leave room for discretionary policymaking which can hinder the attainment of strategic planning goals.
The formal exclusion of citizens from the Central Italy TH supply and delivery process can be linked to the following 'community clashes', whose consideration should inform the future specification of community-sensitive functional requirements for TH (Pezzica, Bleil de Souza, & Dunichkin, 2021): • Acceleration of depopulation trends, due to excessive delays in TH assistance delivery (Emidio di Treviri, 2018) and to a lack of services' accessibility (Rotondo et al., 2020); which resulted in TH sites being populated mainly by the elderly.• Dysfunctional units' design, due to a poor fit with the mountain context and limited adaptability.Specific issues emerged in relation to the possibility to open the entrance door from the inside; set roof inclinations and chimneys' heights suitable to prevent issues with the snow; piping and boilers' insulation.Additionally, people who were not completely autonomous suffered from a lack of space to house caregivers (Emidio di Treviri, 2018).• Economic and technical inconveniences related to thermal comfort and water availability, due to an inadequate design of building services.

Process monitoring and management
Poor financial and time management of TH plans is widely recognised as a risk factor in post disaster recovery.However, the data gathered by the civil protection (NDCP, 2020; Annex C) to monitor progress in the TH process does not allow an effective control of intermediate deadlines in relation to project milestones.Projects' approval dates were not collated, making it difficult to check if the TH suppliers were able to meet their deadlines.In addition, no data was published about the duration of the initial situation assessment and about the time spent to select suitable TH sites.
Overall, the model evidences a lack of time constraints for most of the administrative decision-making process activities, with negative implications for time-management.This makes both the ideal timeframes for TH supply, mentioned in the "Accordo Quadro" and the actual time steps published by the NDPC, unsuitable proxies to assess and manage delays in TH provision and calls for better data collection and tracking.The complexity of the TH supply and delivery process makes analysing the causes of delays without the support of appropriate data and process models problematic, which obscures accountability and public support.

Discussion and global outlook
Although the use of strategic planning in TH management seems promising, the heterogeneity of its applications and the lack of   procedures and policy guidance, represent ongoing challenges in practice (Manning, 2020).As strategic plans can differ in scope and comprehensiveness, with implications for their outcomes, it is important to have mechanisms in place to reassess and revise strategies and plans (Bryson, 2018).In fact, while FEMA has engaged in strategic planning for post-disaster response since 2002, in early efforts, roles were unclear, inventory control was poor, TH procurement and siting agreements were lacking and procedures missing (Sapat et al., 2011).The presented post-mortem analysis tells us how effective the Italian bureaucratic apparatus has been in providing TH housing assistance after the 2016-2017 earthquakes.It confirms that the "Accordo Quadro" and connected regulatory instruments, were rather limited in scope and short-sighted; what ended up encouraging the implementation of onesize-fits-all solutions, as posed by (Emidio di Treviri, 2018).The study also evidences an enduring gap between urban planning and emergency management in Italy, reflected in the lack of pre-disaster strategies for the selection and design of TH sites.
In fact, while supra-regional contracting agreements were in place before the Central Italy earthquakes, flaws in strategic planningparticularly at the local level -severely affected the effectiveness of the TH response and the capacity of local communities to recover.Missing or faulty plans, policies for TH long term use, community participation protocols, inclusion of vulnerable populations needs, and TH design, were flagged.Thus, strategic plans did not align TH activities with a common vision of urban resilience, to establish best pathways to achieve recovery aims, overall quality and performance, and value creation.
The study also confirms that a better alignment of TH-related decisions is needed at the strategic (the plan), the managerial (the process) and the operational (the activities) levels (Pezzica & Cutini, 2021).Although data gaps prevent making exact estimates, available information suggest that approval times are likely to have been longer than design times.Furthermore, results point that the implemented managerial process does not include intermediate mitigation strategies, meaning delays are likely if anything falls out of the plan.Faulty or missing auditing protocols and inadequate preparedness measures at different governance levels (discussed in Section 5) indicate significant shortfalls in strategic planning, which caused operational bottlenecks and delays.
Table 2 pins down the issues which, instead of relating to single critical activities as in Table 1, cut across the whole TH supply and delivery process.It also suggests policy recommendations and mitigation strategies to be tested.The list can then be expanded to include lessons from other countries.A collation of results from different ex-post analyses could result in benchmarks forming an inventory of information useful to inform corrective planning actions critical to achieve strategic planning goals.What can mobilise the development and refinement of policies based on knowledge across jurisdictions (Zulean, 2014).Particularly the US and Japan, which also have strategic agreements in place for TH procurement and rely on a fixed design for the units.These, to the best of the authors' knowledge, do not have formal ex-post analyses of TH process instantiations.
There is, therefore, an opportunity to transfer the proposed postmortem evaluation and the use of BPMN to enable the examination of best practices in these contexts.For instance, the US includes pre-stock options, which are adopted to deliver a prompt TH assistance postdisaster (Perrucci & Baroud, 2021) and which could be transferred to the Italian case.Additionally, Japan is trialling new policies to take better into account residents' views and wellbeing in the planning of TH neighborhoods (Edgington, 2022); following critiques of its technocratic "plan-only" approach, justified by an alleged reduction of delivery time (Bris & Bendito, 2019).After the Tōhoku earthquake and tsunami some municipalities also started to develop new land parcels for housing recovery (Edgington, 2017).
Nevertheless, since contextualization determines best practice applicability (Abulnour, 2014), strategic planning proposals should cater for differences in each country's resources, regulations and governance structures.In the US, the supply and delivery of transportable TH units after disasters (manufactured homes and RVs/travel trailers) has been centrally managed by the federal government up until 2018, when the Disaster Recovery Reform Act first acknowledged the shared responsibilities of state/local authorities in TH response and recovery (Windle et al., 2019).Conversely, the general TH policy framework in Japan is determined at a national level, where funding is also allocated, while the prefectures (or delegated larger municipalities) select the location of TH sites in safe public-owned sites and ensure the delivery and installation of the TH units via the activation of different strategic agreements independently signed with the Japan Prefabricated construction suppliers and manufacturers Association; what can make standards vary across them (Bris & Bendito, 2019).
Strategic planning for TH management should, hence, be informed by an understanding of both technical aspects and local administrative culture.Additionally, indicators used for process monitoring and management should be qualified by a context-based reading of planning intent.
From a methodological point of view, linking business process modelling and a meta-analysis of available data sources was instrumental to reconstruct the Central Italy TH supply and delivery process and support a nuanced understanding of critical activities and components.
Although the BPMN did not assist the examination of issues arising from complex decision-making tasks (e.g., deciding the location, number, dimension, density, and layout of TH sites), following the identification of procedural/administrative bottlenecks, it can inform improvements in time and cost management (e.g., via tasks' parallelisation, optimisation of information flows).In future applications, if data about all intermediate steps is systematically collated and available to run process simulations, more advanced BPMN analytical capabilities could be exploited to, e.g., estimate material resources, total service and waiting times, number of waits (for activities managed by an operator who must be free to execute them), costs, number of actors etc. Deviations from the ideal process model could, in principle, be systematically mapped and comparatively studied by critically confronting planned and actual processes -what requires a separate modelling of specific process instances (as in Annex B) -to automatically spot inefficiencies and propose timely adjustments using process mining tools and engineering techniques.Furthermore, BPMN can help develop a workflow management system able to prevent technical problems, reduce community clashes, and speed up TH operational tasks.
Whenever building a model for each case is not feasible due to a lack of statistically significant information that allow to instruct the system, as in this study, useful information can nonetheless be extracted from the analysis of the tactical ad-hoc adjustments made to the prescriptive plan by humanitarian agencies.Additionally, similar efforts provide an input to digitalisation which facilities coordination and data collection towards process simulation.

Conclusions
This study provided evidence that strategic planning alone does not ensure the success of post-disaster TH assistance programs and that careful process planning and management is needed.It presented a full description of the Central Italy TH supply and delivery process and explored the negative impacts of allowing room for discretionary decisions in this, demonstrating how flaws in strategic plans and their implementation post-disaster contribute to slow down disaster recovery and create new vulnerabilities.
Although in principle strategic planning can help limiting resources' exploitation and corruption and ensure compliance with adequate safety standards in TH provision, it must respond to citizens' needs to generate value.The analysis provided insights on how contextual factors and local communities influence the recovery trajectory of disaster-affected cities, even if they are not formally included in the decision-making process.Moreover, it suggested the need to develop better protocols and intermediate auditing/mitigation strategies to ensure that strategic plans can achieve their objectives.
To help discussing strategic planning and its outcomes, this paper transferred knowledge from the field of management engineering to a different and challenging context of application -i.e., TH managementwith a focus on the TH supply and delivery process.It used BPMN to reconstruct the workflow of the process 'as-planned' and qualitatively study both inherent flaws and 'as-happened' deviations to provide some practical recommendations for planning; an approach which can be reproduced in other contexts to support benchmarking.
The retrospective BPMN analysis enabled connecting the events that triggered the Central Italy TH supply and delivery process, its activities, the decisions that it involved, as well as the ordering of all these components, their outcomes, the responsible organisations and actors and, to an extent, their roles.In addition, it enabled pinpointing handovers points and where issues occurred in the post-disaster phase to highlight critical aspects to be improved.This was instrumental to find out about: (i) procedures requiring better structuring; (ii) missing or deficient intermediate and end-of-life assessments; (ii) activities prone to generate conflicts and/or delays; (iv) sustainability gaps.
The direct use of BPMN in pre-disaster strategic planning, however, presents important limitations and practical challenges that arise when the goal is to design well-structured processes because of missing contextual information, gaps in empirical data, and modelling uncertainties due to the contingent way in which decisions are made during emergencies.BPMN can nevertheless contribute to enhance the resilience of future strategic plans, and thus avoid TH assistance delays, by feeding information from past experiences (i.e., patterns of the schema, actors involved, process coordination system, time performance etc.) into new workflow and document management systems which support the TH supply and delivery process in real time, and monitoring systems which evaluate model performance and redesign.Furthermore, BPMN can foster the fast generation of novel hypothesis and simulate admissible instantiations to advance scenario-based planning.In combination with ICT and triangulation, BPMN would then support strategic planning and accelerate digital innovation in key urban emergency management processes such as TH, which, albeit unplanned, yet influence cities' structural changes well beyond the fast dynamics of emergency response.
Although some recommendations for policy and practice were advanced in Section 6, more empirical research is needed to test them.Future studies should also further clarify how strategic planning, and proactive TH management, can be improved to concretely enhance the capacity of cities to cope with housing crises and foster a developmental recovery after disasters.In addition to making the process execution more efficient, future research can examine specific parts of the process, where structured starting data are available and develop policies and practical guidance to adapt the TH response to the physical, environmental, social, and economic requirements of the disaster-affected territory; possibly exploring a shift of focus from specifications to performance requirements.

Declaration of competing interest
None.

Annex A
Central Italy TH supply and delivery process modelled using BPMN 2.0.

Annex B
C. Pezzica et al.Baseline and improved TH installation process with task parallelism, BPMN 2.0.C. Pezzica et al.

Fig. 7 .
Fig. 7. Boxplot, number of days elapsed in all TH sites to reach the milestones set by the NDPC.Own elaboration based on (NDCP, 2020) data.
of official TH data for the Central Italy seismic crater showing the total number of days elapsed between the process milestones.Original data source: Sito del Dipartimento della Prtezione Civile -Presidenza del Consiglio dei Ministri, distributed under Creative Commons Attribution 4.0 International Public License.

Table 1
Critical process activities, connected issues, and suggested mitigation strategies.

Table 2
Process planning flaws and proposed mitigation strategies.