Evaluating the Significance of Criteria Contributing to Decision-Making on Brownfield Land Redevelopment Strategies in Urban Areas

Abstract

Brownfield land is one of the least exploited resources for urban development in a number of Eastern European countries. Establishing a rational strategy for redeveloping brownfields is an unambiguously complex task that requires considering a number of different economic, social, physical and environmental factors. The strategic decision-making has a long term impact on the quality of life, ecological balance and urban structure. Therefore, the paper is aimed at developing a comprehensive set of criteria that contribute to the redevelopment of brownfield land in urban areas. It focuses on six main development strategies that embrace creating residential, green, commercial, recreational activity and industrial areas or leaving land as a reserve. Geographic information system (GIS) tools are employed to collect the spatial information, obtain the initial set of criteria and derive the statistical data. Expert’s evaluations along with a statistical method of gauging the level of concordance of their opinion combined with Delphi method are used for determining significance of criteria within economic, social, physical (urbanistic) and environmental criteria groups. This study establishes the most significant criteria for implementing different scenarios of the brownfield land redevelopment in Vilnius, Lithuania. Developed framework will support the decision-making process in the brownfield land redevelopment aiding a sustainable urban planning.

1. Introduction

The reclamation of brownfield land, including old industrial and commercial areas, remains one of the priorities set by the EU policy aimed at gradually increasing density of population in urban areas. It has been estimated that approximately 500,000 hectares of brownfields suitable for reclamation were in Europe in 2005. Today, a large proportion of the brownfield land is still available for regeneration. It can be utilized for raising the economic attractiveness of cities to new investment, preserving urban identity, improving social climate and developing the prevention of natural elements in the city.

The brownfields can be defined as a land that has previously been developed, but is not in current use [1]. Derelict (abandoned) and vacant (not occupied) land can also be classified as the brownfield land. However, it should be noted that the definitions of the brownfield land vary significantly in various countries and are shaped by deindustrialization, urbanization, high density of population or other socio-economic factors [2]. Brownfields are also described as a territory that is affected by the previous use or the impact of the surrounding land and becomes unused or useless without further intervention [3]. Brownfield land provides a possibility of using free space and potential for additional urban development considering changes in the needs of the population [4]. Regeneration is the most commonly used way to exploit the potential of brownfields [5,6,7].

The research on brownfields provides a number of excellent examples of how such areas can be reclaimed thus achieving a new quality of the environment [8,9,10,11,12]. The reclamation of brownfields in Europe has been pursued through the effective integration of the concept of sustainable development into the EU projects such as Regeneration of European Sites in Cities and Urban Environments (RESCUE) [13,14] and Concerted Action on Brownfield and Economic Regeneration Network (CABERNET) [15]. The projects have been aimed at simplifying the procedure for new practical solutions seeking a sustainable development of the brownfield land. Notably, a public–private partnership (PPP) model has been increasingly successful for implementing projects on the redevelopment of brownfields [16,17,18]. A regeneration project in coastal area of Liverpool, UK, provides an excellent example of the partnership between public and private sectors [9,19]. In 2005, the EU and UK together with the private sector have paid a total of 560 million Euros for the redevelopment of the Waterfront area situated in the city. On the principle of the PPP, business and leisure complexes consisting of mixed-use areas have been built in brownfields and unused territories. About 2500 new jobs and an environmental aesthetic image have been created. Some of the old buildings have been renovated thus preserving the cultural heritage and city’s identity. A flood protection system has also been installed. The adoption of various environment-friendly solutions has resulted in a significantly decreased need for water, as well as a reduced air and water pollution in the Waterfront area of Liverpool.

The regeneration of territories and redevelopment of the brownfield land are progressively running through sustainable development and should integrally solve social, economic and environmental issues as well the problems of the physical environment [20,21,22,23]. However, the imbalance between the volumes of urban development objects and brownfields remains high, particularly, in many countries of the Eastern Europe [24,25,26]. After dissolution of Soviet bloc, Eastern European countries have experienced a sudden transition from central planning to the decentralized regulation of the market economy [27,28]. The need to reclaim unused urban brownfields, including military, industrial, and commercial buildings that do not perform their primary function, has significantly increased due to intensive economic processes, growing number of the population in big cities and the implementation of sustainable development policy [29,30,31,32]. However, due to a tight financial situation, the problem of brownfields in some Eastern European countries still remains a serious challenge. For example, in Czech Republic and Slovakia, the ongoing redevelopment processes of brownfields take place only in high-priority inner urban areas [24]. Moreover, many post-Soviet countries require methodology and strategies for brownfields redevelopment.

The paper is aimed at establishing the framework for supporting decision-making processes in the brownfield land redevelopment. The research was performed using data acquired in twenty districts (neighborhoods) of Vilnius, the capital city of Lithuania. It allowed determining the most significant criteria contributing to decision-making on brownfield land redevelopment strategies in urban areas. The obtained results will facilitate the decision-making process in the brownfield land redevelopment and assist the urban planning.

2. Methodology

2.1. Hierarchical System of Criteria

To determine the most significant criteria for implementing different scenarios of the brownfield land redevelopment, a comprehensive set of 152 criteria was established through the literature review [33,34,35,36,37,38,39,40]. As described previously [41], this initial set of criteria was used to develop the hierarchical system including economic, social, physical (urbanistic) and environmental criteria groups.

A hierarchical system of criteria used in this study allowed the following: (1) overcoming difficulties arising from using a sufficiently large set of criteria for multi-criteria analysis; (2) reducing the complexity and bias in eliciting weights of importance of criteria by experts; and (3) exploiting the flexibility and convenience of the tool of hierarchical structures.

Moreover, there are a number of other prominent features of hierarchical systems, which provide advantage whenever complexity is involved [42,43,44,45,46,47,48,49,50,51,52]. Hierarchical systems are built in blocks, which imply a faster speed of creating them. Higher levels of hierarchy have influence on the lower ones. Hierarchies are flexible, which means they can be modified in the creation process [53]. There are no formalized methods for building a hierarchical system. Usually, it is built using tradition, intuition, or structures of databases [54]. Hierarchical system can be deduced using literature or communication with experts of the related field [53].

In this study, an expert’s ranking in combination with a multiple criteria decision making (MCDM) method [41] was used to identify a final set of criteria.

2.2. Data Collection

The GIS technology was used to capture and digitize spatial data obtained for brownfield land in twenty districts of Vilnius city, as well as to combine and link up various data, including economic, social, physical and environmental indicators as described previously [55]. GIS data were then used for evaluation of each criterion from the final set of 18 criteria. As a result, a data set of 360 different multi-dimensional indicators was established. These indicators were then used for establishing criteria relative weights.

2.3. Relative Weights of Criteria

The task of establishing relative weights of criteria is a compulsory stage of any multiple criteria analysis. There are several approaches how to estimate weights of criteria by eliciting opinions from experts. The simplest and easiest to understand for experts would be using Likert scale of an appropriate number of grades. This approach unfortunately would hardly satisfy natural precision prerequisite, as vague weights would correspond to each grade [56]. At the other extreme, popular worldwide AHP (Analytic Hierarchy Process) method proposed by Saaty can be used. The latter method uses the 9-point scale, in which usually only 5 grades in fact are used [57]. In the study, having a relatively large number of criteria, this would be a serious limitation [58]. In addition, the AHP method can only be used by experts familiar with this method. Such method as UTA [59] attempts to resemble decision-maker’s global preferences omitting the stage of obtaining weights. It requires from each expert not only the evaluation of utilities induced by each value of each criterion, but also the estimation of differences between utilities of different projects. Its upgraded version UTASTAR uses group decision-support aggregation-disaggregation procedures for obtaining estimates of decision-makers’ preferences. It is a multiple stage, complicated process of reciprocal communication with experts, which again is a serious limitation in the case, when experts are chosen from the field other than operational research. Taking all above into consideration, a more favorable Delphi technique was chosen for working with a group of experts aiming to obtain consistent estimations [60].

In present study, the multiple criteria analysis was aimed at determining the most suitable redevelopment scenario T_i for each urban brownfield land. Therefore, relative weights of criteria were established for every brownfields redevelopment scenario T_1-6 by using expert opinions as described previously [41]. The experts were chosen by following strict selection criteria, requiring that each expert met at least one of the following requirements: (1) have three years of experience in spatial planning, economic, environmental protection, sociology and real estate management; (2) have three years of experience in the field of architecture and at least two designed and implemented projects; and/or (3) have three years of experience in policy making with respect to urban development, spatial planning and real estate market.

In total, twelve experts agreed to participate in the survey. Relative weights of criteria were determined within each group including economic, social, urbanistic and environmental. The maximum number of criteria per criteria group was five, making the task more feasible, since a smaller number of criteria required be comparing and evaluating by the expert. Experts were asked to fill in created proprietary forms in which they were required to state weights of criteria in percent. Overall, 12 experts have responded.

2.4. Non-Parametric Statistical Analysis

In order to assess agreement among experts in respect to criteria weights, the theory of Kendall was applied [61]. Initially, the magnitudes of criteria weights were ranked. Since each brownfields regeneration scenario is perceived in a different way, weights of the criteria were determined considering each scenario T₁–T₆ separately. Such ranks were denoted as e_ik, where i = 1, 2, ..., m is the index of criteria (in our case, m is equal to 4 or 5) while k = 1, 2, ..., r is the index of denoting experts (where r is the number of responded experts, 12 in our case). Kendall’s W was used in the chi-squared test statistics for gauging the level of concordance, which depends on the sum of squared deviations of all ranks e_ik by all experts.

$$e_i={\displaystyle \sum _{k=1}^r{e}_{ik}}$$

(1)

From the mean of such sums

$$\overline{e}=\frac{{\displaystyle \sum _{i=1}^m{e}_i}}{m}$$

(2)

Consequently, Kendall’s W equals the ratio between the sum S mentioned above, calculated by Formula (3), and its largest deviation, denoted by S_max, calculated by Formula (4). The latter sum is observed in the case of the absolute concordance of expert opinions in terms of ranks of importance of criteria.

$$S={\displaystyle \sum _{i=1}^m{\left(e_i-\overline{e}\right)}^2}$$

(3)

$$S_{\max }=\frac{r^2\times m\times \left(m^2-1\right)}{12}$$

(4)

Consequently,

$$W=\frac{S}{S_{\max }}=\frac{12\cdot S}{r^2\times m\times \left(m^2-1\right)}$$

(5)

Chi-squared test statistics for this variable is

$${\chi }^2=W\times r\times \left(m-1\right)=\frac{12\times S}{r\times m\times \left(m+1\right)}$$

(6)

The number of degrees of freedom υ = m − 1. For the test statistics, the level of significance α = 0.05 was chosen. Next, equal ranks within 6 sets of criteria were found. There were only two equal ranks at most. For the cases when ranks were equal, the following adjustment of the coefficient of concordance was applied [61].

$$W=\frac{12\times S}{r^2\times m\times \left(m^2-1\right)-r\times {\displaystyle {\sum }_{\varphi }\left(t_{\varphi }^3-t_{\varphi }\right)}}$$

(7)

where φ denotes the sets of equal ranks, and t_φ denotes the number of equal ranks within a set within φ.

Averages of weights elicited from experts, which were found to be concordant, were used in the followed analyses. The overall methodology pipeline is shown in Figure 1.

3. Results

3.1. Brownfield Land Redevelopment Scenarios

In order to build a framework that can support the decision-making on the brownfield land redevelopment in urban areas and to assist urban planning and development, this study aimed to establish what criteria are the most significant for redevelopment of brownfield land into the urban land of a different use. Whereas a number of models involving different types of the urban land use have been described previously [62,63], the following six scenarios can be distinguished for the redevelopment of brownfield land in urban areas (Figure 2a):

• redevelopment to a green area (T₁);
• redevelopment to a commercial area (T₂);
• redevelopment to a recreational area (T₃);
• redevelopment to an industrial area (T₄);
• redevelopment to a residential area (T₅); and
• leaving land as a city reserve (T₆).

The brownfield land redevelopment scenarios T_i were considered for twenty districts of Vilnius city (Figure 2b) [64]. Resulting scenarios may reflect the character of the urban area and the possible potential of the locality. Such brownfield land redevelopment opportunities can then be successfully used for developing partnerships between public and private capital applying the PPP principle [65].

3.2. Development of the Hierarchical System of Criteria: Case Study of Vilnius City

In order to determine the most suitable redevelopment scenario for each urban brownfield land, weights of criteria have to be evaluated for each scenario T_i establishing the most significant criteria. Therefore, an initial set of 152 criteria was established as described in Methodology. To reduce complexity, the study was confined to 48 criteria (selected set of criteria), and only the 18 highest ranked criteria (final set of criteria) (Figure 3) were used for further analyses.

The GIS data collected in Lithuania showed that the capital city, Vilnius, contains a brownfield land area of 10.9 km² (Figure 4), the major part of which (83%) is a vacant land. Twenty districts of Vilnius city, identified as important for redevelopment of brownfield land, were selected for case study. With the help of GIS technology, the data set of 360 different multi-dimensional indicators was created for 20 districts of the city providing data platform for the multiple criteria evaluation. All investigated indicators were attributed to a certain group of criteria C_j as in Figure 3. In the final set of criteria, each criteria group comprises of up to five criteria as follows: {E₁, …, E₄} ∈ C₁; {U₁, …, U₅} ∈ C₂; {S₁, …, S₅} ∈ C₃; {N₁, …, N₄} ∈ C₄. Altogether, they form list of criteria (

Table 1. Definitions of criteria.

Group of Economic Criteria	Group of Urbanistic Criteria
E1—Infrastructure investment E2—Cost for new real estate E3—Number of projects funded by EU E4—Number of workspaces	U1—Empty sites U2—Number of schools U3—State and average age of new constructions U4—Magnitude of new constructions U5—Distance to the city center
Group of Social Criteria	Group of Environmental Criteria
S1—The level of unemployment S2—The level of poverty S3—Household incomes S4—The level of public crimes S5—Access to educational institutions	N1—Soil contamination N2—Heavy industry pollution N3—Green areas N4—Transport pollution

) used for further expert evaluation and establishing the most significant criteria.

3.3. Establishing the Weights of Criteria

Most important criteria (Table 1) of different groups allow comprising facets that can influence the choice of scenario for brownfield land redevelopment from different perspectives. In this study, the mathematical model described below was used for deriving weights of 18 key criteria. This is a compulsory initial step required for applying multiple criteria evaluation.

As described in the Methodology, the relative weights of criteria were established using the Delphi technique involving a group of experts.

In order to establish the level of concordance of expert opinions for each scenario T₁–T₆ within all groups C₁–C₄ of criteria, calculations of the Kendall’s W along with the Chi-squared test statistics, were performed for sets of criteria within the groups and criteria groups (

Table 2. The values of the Kendall’s W and Chi-squared for the sets of criteria within groups, and criteria groups prior to corrections.

	W	χ2	No. of Objects	χ2cr	χ2 − χ2cr
T1
Groups	0.330	11.87	4	7.81	4.06
Economic	0.821 *	29.56 *	4	7.81	21.75
Urbanistic	0.361	17.33	5	9.49	7.84
Social	0.314	15.08	5	9.49	5.59
Environmental	0.337	12.13	4	7.81	4.32
T2
Groups	0.815	29.33	4	7.81	21.52
Economic	0.767	27.60	4	7.81	19.79
Urbanistic	0.301 *	14.43 *	5	9.49	4.94
Social	0.174	8.33	5	9.49	−1.16
Environmental	0.185	6.67	4	7.81	−1.14
T3
Groups	0.550 *	19.79 *	4	7.81	11.98
Economic	0.633	22.80	4	7.81	14.99
Urbanistic	0.443	21.25	5	9.49	11.76
Social	0.417 *	20.00 *	5	9.49	10.51
Environmental	0.715	25.73	4	7.81	17.92
T4
Groups	0.456	16.40	4	7.81	8.59
Economic	0.744	26.80	4	7.81	18.99
Urbanistic	0.663 *	31.81 *	5	9.49	22.32
Social	0.328	15.75	5	9.49	6.26
Environmental	0.604	21.73	4	7.81	13.92
T5
Groups	0.626	22.53	4	7.81	14.72
Economic	0.685	24.67	4	7.81	16.86
Urbanistic	0.191 *	9.15 *	5	9.49	−0.34
Social	0.344	16.50	5	9.49	7.01
Environmental	0.078	2.80	4	7.81	−5.01
T6
Groups	0.278	10.00	4	7.81	2.19
Economic	0.167	6.00	4	7.81	−1.81
Urbanistic	0.587	28.17	5	9.49	18.68
Social	0.198	9.50	5	9.49	0.01
Environmental	0.104	3.73	4	7.81	−4.08

* Adjusted Kendall’s W and Chi-squared value, whenever equal ranks are found in a set, are denoted with an asterisk.

).

In the six sets of responses, the expert opinions appeared to be non-concordant (Table 2). The most divergent cases were presented to the same experts along with a summary of the results elicited from the group of experts, by following the Delphi method recommendations [66,67]. Therefore, the adjusted relative weights of criteria, as a remedy to the discrepancies in the expert opinion, were determined (

Table 3. Corrections in Scenario 2 by Expert 10 (social criteria).

	Criterion	Unemployment Rate	Poverty Rate	Total Household Income	Crime Index	Access to Schools and Pre-Schools
Before		27%	12%	8%	19%	35%
After		22%	17%	27%	19%	15%

,

Table 4. Corrections in Scenario 2 by Expert 11 (environmental criteria).

	Criterion	Soil Pollution	Pollution from Factories, etc.	Spread of Forests and Green Areas	Pollution from Transport
Before		30%	10%	35%	25%
After		10%	30%	35%	25%

,

Table 5. Corrections in Scenario 5 by Expert 11 (urbanistic criteria).

	Criterion	Empty Sites	Number of Schools	State and Average Age of New Constructions	Magnitude of New Constructions	Distance to the City Centre
Before		25%	30%	10%	15%	20%
After		10%	30%	25%	15%	20%

,

Table 6. Corrections in Scenario 5 by Expert 2 (environmental criteria).

	Criterion	Soil Pollution	Pollution from Factories, etc.	Spread of Forests and Green Areas	Pollution from Transport
Before		35%	10%	25%	30%
After		10%	35%	25%	30%

,

Table 7. Corrections in Scenario 5 by Expert 5 (environmental criteria).

	Criterion	Soil Pollution	Pollution from Factories, etc.	Spread of Forests and Green Areas	Pollution from Transport
Before		30%	11%	33%	26%
After		11%	30%	33%	26%

and

Table 8. Corrections in Scenario 6 by Expert 2 (economic criteria).

	Criterion	Investments in Infrastructure	New Construction Cost	Number of Undertaken EU Projects	Number of Work-Places
Before		15%	23%	27%	35%
After		35%	23%	15%	27%

).

Following the first round of Delphi-adjustment only, along with the feedback communicating the results obtained from the first round, the amended opinions of experts appeared to be concordant. Calculations of the adjusted Kendall’s W along with the Chi-squared test statistics for the new opinions of experts, for each scenario T₁–T₆ within all groups C₁–C₄ of criteria, and for the groups, revealed the results presented in

Table 9. The values of the Kendall’s W and Chi-squared for the sets of criteria within groups, and criteria groups after adjustment.

	W	χ2	No. of Objects	χ2cr	χ2 − χ2cr
T1
Groups	0.330	11.87	4	7.81	4.06
Economic	0.821 *	29.56 *	4	7.81	21.75
Urbanistic	0.361	17.33	5	9.49	7.84
Social	0.314	15.08	5	9.49	5.59
Environmental	0.337	12.13	4	7.81	4.32
T2
Groups	0.815	29.33	4	7.81	21.52
Economic	0.767	27.60	4	7.81	19.79
Urbanistic	0.301 *	14.43 *	5	9.49	4.94
Social	0.326	15.67	5	9.49	6.18
Environmental	0.274	9.87	4	7.81	2.06
T3
Groups	0.550 *	19.79 *	4	7.81	11.98
Economic	0.633	22.80	4	7.81	14.99
Urbanistic	0.443	21.25	5	9.49	11.76
Social	0.417 *	20.00 *	5	9.49	10.51
Environmental	0.715	25.73	4	7.81	17.92
T4
Groups	0.456	16.40	4	7.81	8.59
Economic	0.744	26.80	4	7.81	18.99
Urbanistic	0.663 *	31.81 *	5	9.49	22.32
Social	0.328	15.75	5	9.49	6.26
Environmental	0.604	21.73	4	7.81	13.92
T5
Groups	0.626	22.53	4	7.81	14.72
Economic	0.685	24.67	4	7.81	16.86
Urbanistic	0.263 *	12.67 *	5	9.49	3.18
Social	0.344	16.50	5	9.49	7.01
Environmental	0.337	12.13	4	7.81	4.32
T6
Groups	0.278	10.00	4	7.81	2.19
Economic	0.315	11.33	4	7.81	3.52
Urbanistic	0.587	28.17	5	9.49	18.68
Social	0.198	9.50	5	9.49	0.01
Environmental	0.332	11.97	4	7.81	4.16

* Adjusted Kendall’s W and Chi-squared value, whenever equal ranks are found in a set, are denoted with an asterisk.

.

In order to derive relative weights of criteria considering all 18 criteria listed in Table 1, the method of deriving weight of each criterion using both weights of the group and of each criterion within the group, as proposed by Podviezko [54], was applied. This method is appropriate to use in cases when hierarchical system of criteria is built. The weights of criteria groups are multiplied by the weights of criteria within each group as shown in Formula (8):

$${\omega }_i={\omega }_{i_k}\times {\omega }_k$$

(8)

where k is the index of groups, and i_k is the index of criteria within group C_k.

For each brownfield redevelopment scenario T₁–T₆, the weights of the criteria groups were calculated using Formula (8) (Figure 5). This allowed establishing the significance of each group of criteria in the case that the brownfield land is redeveloped based on particular scenario T_i. Then, the weights of each criterion within each criteria group were established (

Table 10. Weights of individual criteria for each brownfield scenario T_1–6 in each group of criteria (%).

	T1	T2	T3	T4	T5	T6
E1	34.33	33.25	33.08	30.50	29.83	31.92
E2	28.42	29.42	27.25	23.75	30.67	23.50
E3	18.58	16.50	21.08	16.67	18.92	20.08
E4	18.67	20.67	18.67	28.92	20.42	24.50
	Σ = 100%	Σ = 100%	Σ = 100%	Σ = 100%	Σ = 100%	Σ = 100%
U1	22.75	21.25	27.58	28.83	14.50	30.92
U2	13.92	11.75	22.33	12.67	23.67	14.75
U3	20.58	18.92	17.33	17.92	21.58	18.00
U4	17.25	25.42	14.50	23.75	20.33	22.17
U5	27.75	23.00	18.25	16.83	20.17	14.17
	Σ = 100%	Σ = 100%	Σ = 100%	Σ = 100%	Σ = 100%	Σ = 100%
S1	15.00	18.17	16.50	26.42	17.42	15.25
S2	18.42	19.83	15.08	20.92	18.42	21.42
S3	17.17	24.83	17.67	21.08	20.92	16.00
S4	25.17	22.67	25.92	17.00	16.58	23.17
S5	24.42	14.58	24.75	14.42	26.58	24.33
	Σ = 100%	Σ = 100%	Σ = 100%	Σ = 100%	Σ = 100%	Σ = 100%
N1	20.00	18.92	18.00	30.92	17.17	22.00
N2	19.33	28.75	31.92	29.33	30.83	20.83
N3	29.50	29.67	30.67	18.92	26.08	32.92
N4	31.25	22.67	19.42	20.83	25.75	24.33
	Σ = 100%	Σ = 100%	Σ = 100%	Σ = 100%	Σ = 100%	Σ = 100%

). This allowed concluding that the application of Formula (8) can assist significantly in calculating weights of criteria in a hierarchical system of criteria.

3.4. Establishing the Most Significant Criteria Contributing to Brownfield Land Redevelopment Strategies

In order to establish the most significant criteria contributing to the redevelopment scenarios of brownfield land, the weights of all criteria for each brownfield redevelopment scenario were derived as described in Methodology (Figure 6). For convenience of decision-making process, the calculated weights of individual criteria for each redevelopment task in each group of criteria are presented in this paper in the scalar distribution form. This comparison allows a decision-maker to assess the meaningfulness of each criterion in redevelopment processes while working out a solution for one of the problems T_i.

The results revealed the most significant criteria contributing to the redevelopment of brownfield land applying a particular scenario T_i. As a result, the following most significant criteria for each case of brownfield land redevelopment scenario were identified:

Scenario T₁: The criteria of the environmental group have a decisive impact, particularly criteria N₃ (green areas per inhabitant, ω_N3 = 0.089) and N₄ (pollution from transport, ω_N4 = 0.095).

Scenario T₂: The criteria of the economic group have a decisive impact, particularly criteria E₁ (investments in infrastructure, ω_E1 = 0.112) and E₂ (cost of new rental estate, ω_E2 = 0.099).

Scenario T₃: The criteria of social and urbanistic groups have a decisive impact, particularly criteria S₄ (crime index, ω_S4 = 0.083), S₅ (access to educational institutions, ω_S5 = 0.079) and U₁. (empty sites, ω_U1 = 0.076).

Scenario T₄: The criteria of almost all groups, except those of the social one, equally strongly determine this redevelopment scenario. Among the prevailing criteria, E₁ (investments in infrastructure, ω_E1 = 0.096), E₄ (number of work-places, ω_E4 = 0.091), U₂ (number of schools, ω_U2 = 0.076), U₅ (distance to the city centre, ω_U5 = 0.063) and N₂ (pollution from heavy industry, ω_N2 = 0.078) have the major impact.

Scenario T₅: The criteria of the social group have a decisive impact, particularly criteria S₅ (ω_S5 = 0.091) specifying accessibility to education and pre-school educational establishments.

Scenario T₆: This scenario is strongly affected by environmental and urbanistic criteria, among which criteria N₃ (green areas per inhabitant, ω_N3 = 0.102) and U₁ (empty sites per inhabitant, ω_U1 = 0.082) have the major impact.

To conclude, the average weights of criteria significance (ω_Mi) and standard deviation (d_i) were calculated. Subsequently, they were ranked (Figure 7) showing that, overall, E_a1 (investments in infrastructure; ω_avg.E1 = 0.079), N_a3 (green areas per inhabitant; ω_avg.N3 = 0.073), E_a2 (cost of new real estae; ω_avg.E2 = 0.067), U₁ (areas of empty sites per inhabitant; ω_avg.U1 = 0.065) and N₂ (pollution from heavy industry; ω_avg.N2 = 0.064) are the most influential criteria in making decisions on brownfield redevelopment.

4. Discussion and Conclusions

Brownfields represent specific challenges for the environment and adjacent community as it has been affected by former uses; is derelict or underused; requires intervention to bring it back to beneficial use; and may have real or perceived contamination problems [68,69]. Moreover, all brownfields sites vary concerning their unique characteristics, such as location, size, extent of potential contamination resulting from previous use, etc. As a result, diverse stakeholders have heterogeneous concerns regarding successful and sustainable brownfield land regeneration [25].

In order to deal with the complex decision-making processes, several multi-criteria decision analysis (MCDA) approaches and tools have been developed and increasingly applied in different fields, including the land-use context. Prioritization tools based on sustainability frameworks and MCDA allow assessing requalification options from different points of view, respecting the needs of multiple stakeholders [20,43,68,70]. Due to MCDA ability to combine heterogeneous inputs with cost/benefit information and stakeholder views and being recognized as suitable tool to support the ranking of regeneration alternatives based on the sustainability framework [21,22,23,70], the previously described MCDA method [41] in combination with the expert’s ranking was used to identify a final set of criteria in this study.

With this study we aimed to establish the framework of criteria for supporting decision-making processes in the brownfield land redevelopment. The research was performed using data acquired in twenty districts of Vilnius city. A complex structure of criteria was required for such a multifaceted task. The division of criteria into groups has proved to be the most helpful solution allowing both to cast the set of criteria and enabling experts to estimate the weights of criteria.

The paper proposes a new approach for evaluation of criteria importance. The method utilizes relative weights of criteria groups and relative weights of criteria within the groups for estimation of the weights of individual criteria for each brownfield redevelopment scenario. In particular, results revealed that the redevelopment of brownfields to the commercial area is primarily related to economic criteria (ω_E|T2,% = 31.58%), whereas redevelopment of brownfield land to residential areas is influenced by the social criteria (w_S|T5,% = 34.08%). Not surprisingly, the economic criteria has the greatest impact on brownfields redevelopment into industrial areas (ω_E|T4,% = 31.58%).

Notably, the most significant criteria contributing to the decision-making strategies for the redevelopment of brownfield land in urban areas were determined in this study. Not surprisingly, majority most important criteria for redevelopment to green, commercial, recreational, or residential areas were very relevant to the redevelopment strategy and were from the criteria groups such as environmental (green areas per inhabitant and pollution from transport), economic (investments in infrastructure and cost of new rental estate), social (crime index and access to educational institutions) and social (accessibility to education and pre-school educational establishments), respectively. Interestingly, results revealed that for redevelopment to the industrial area, criteria from three criteria groups including economic, urbanistic and environmental were found to be equally important.

Overall, the analysis of brownfield land redevelopment scenarios and evaluation of the criteria significance will assist in developing decision-making guidelines for various brownfield land redevelopment solutions.