Towards a Circular Economy in Urban Households: Spatial Challenges of Waste Collection Under Residential Growth in Warsaw

Abstract

The present article examines the relationship between the changing residential net floor area of residential units in Warsaw, driven by ongoing urban growth, and the spatial requirements for household municipal waste collection. Analyzing 20 years from 2003 to 2023 across 18 districts of Warsaw, this article examines how the interplay between building morphology, demographic structure, and municipal waste generation influences the spatial and infrastructural requirements for separate collection. The study panel regression and volume simulations were applied to assess these dynamics. The results demonstrate that the number of residents per unit is the strongest predictor of waste generation, while the effect of floor area is less robust but still relevant. Waste generation per unit increased by 20% during the study period, accompanied by a sevenfold rise in plastics and a nearly eightyfold increase in biowaste, which together impose growing spatial burdens on households and collection infrastructure. The study emphasizes the significance of integrating waste infrastructure planning with housing design, considering the urban areas’ heterogeneous and evolving nature (identified at the district level). In the transition to a circular economy, the results highlight the need for locally sensitive strategies that link everyday household waste management with systemic urban sustainability goals.

1. Introduction

The intensive waste generation and its growing environmental impact have made a shift in waste management essential. This necessity has become apparent because of the transformation of human settlements—from dispersed rural structures to highly concentrated urban forms—accompanied by population growth, economic development, and increased consumption [1,2].

Numerous studies examining the relationship between municipal solid waste (MSW) generation, income levels, and urbanization consistently demonstrate that waste is predominantly an urban phenomenon [3,4,5]. The correlation between population density and the amount of waste generated has been widely documented, showing that denser urban areas often produce higher volumes of municipal solid waste, directly affecting the planning and efficiency of waste management systems [4,6,7]. City dwellers are reported to generate approximately twice as much waste as rural households [8]. Latest findings show that MSW generation in Polish cities exhibits linear scaling concerning population size, meaning that larger urban centers generate disproportionately more waste per inhabitant [9].

The aforementioned factors necessitate new trends in city planning. One example is the compact city model, which promotes higher density, mixed-use urban areas, which can improve effective MSW collection by reducing travel distances and optimizing collection routes [10]. Higher population density and proximity to services in compact cities can facilitate better recycling participation and waste separation at the source [11]. However, poorly planned high-density areas may exacerbate waste accumulation and pose challenges for infrastructure capacity if waste generation outpaces collection [12].

The existing literature primarily focuses on the estimation of MSW generation in urban areas [13], including studies that investigate the influence of building morphology on the intensity of waste generation [14]. More recent studies on MSW generation and housing typologies increasingly adopt multi-dimensional frameworks, combining statistical analysis, GIS-based mapping, and integrated simulation methods to capture variability across space and time [15,16]. There is a visible emphasis on methodologies that can identify local variations and long-term changes related to urban development or decline [17,18]. In this context, waste generation data might indicate urban decay, as a factor of socioeconomic and physical shifts within the built environment [19]. Vetter-Gindele et al. [20] noted the relationship between building type and the quantity and composition of household municipal solid waste. The researchers suggested that the nature of the building may indicate the residents’ socioeconomic characteristics. Similarly, the study of Nistor et al. [21] suggests that demographic and urban form analysis is an important element of waste management research. The mentioned papers underscore the necessity to include social, economic, and spatial dimensions to implement effective separate waste collection systems and integrated waste management. Moreover, European cities’ research has shown that the effectiveness of selective waste collection depends largely on housing conditions, population density, and the availability of infrastructure adapted to local urban and social factors [22]. Therefore, when analyzing waste generation patterns in cities, it is essential to use models that can identify spatial differences and temporal changes. These models offer a more accurate understanding of the drivers of MSW generation [23].

Although many studies prioritize mass-based assessments, relatively few address the volumetric dimensions of waste [24]. Nevertheless, for effective MSW management, it is equally important to consider the spatial requirements of various waste types at the source collection stage (in-home storage), specifically within housing units or residential estates. A key challenge to implementing selective waste collection in households is the issue of limited space and maintaining domestic order. As demonstrated in the study by Pedersen and Manhice [25], the limited availability of space within residential apartments and residents’ aversion to storing visible waste fractions result in inadequate waste sorting practices. Waste, particularly bio-waste, is often regarded as unhygienic, esthetically displeasing, and a detriment to the representative character of the living space. Consequently, households frequently attempt to hide these items, and when this is not feasible, they forgo complete sorting.

From the perspective of household analysis, the number of inhabitants and the residential unit size are important factors in the context of waste collection possibilities [26,27]. In recent decades, there have been several studies regarding socioeconomic factors [28] and the impact of residential environment [29] on waste management. In Europe, the steady decline in household size has become a well-documented phenomenon [30]. The average household in the European Union comprises about 2.3 persons, with national values ranging from 1.9 in Finland and 2.0 in Germany, Denmark, and Sweden, up to 2.9 in Slovakia and 2.8 in Poland [31]. It has been shown that as the number of people in households increases, the amount of waste generated per person decreases [32]. This suggests a non-linear relationship between household size and waste output; however, it should be noted that the physical limitations of residential units impose significant constraints on the capacity for waste storage and separation. The average size of a residential unit also varies across Europe: Eurostat data indicates that the average is around 96 m², falling to around 85 m² in urban areas. Poland is slightly below the European standard, with an average residential unit size of 93 m² overall and around 75 m² in urban areas [22,33]. Overall, these observations indicate that demographic trends and dwelling size play a role in shaping the practical conditions for waste storage and separation within households, with relevance for the performance of selective collection systems.

Moreover, during the last several years, an emphasis has been placed on implementing Circular Economy (CE) principles within the framework of waste management strategies [34,35]. In line with the growing imperative to reduce the mass and volume of mixed (unsorted) MSW, there has been a marked shift toward the segregation of waste into distinct fractions—an essential step in facilitating the transition to a Circular Economy [36,37]. In this context, applying separate household-level waste collection is critical for adopting circular municipal waste management [38,39]. Simultaneously, the diversification of waste streams directly influences the scale and placement of collection infrastructure [40,41,42].

This paper focuses on the urban aspects of waste management, particularly the spatial implications of selective waste collection. While previous studies have examined the efficiency of selective collection systems, little attention has been paid to how the spatial requirements of these systems interact with the physical constraints of residential units.

This phenomenon is fundamental in contemporary urban transformations, such as the intensification of multi-family housing and the densification of city spaces. At the same time, the transition toward a circular economy has reinforced policy requirements for higher recycling rates and more rigorous waste separation, further intensifying the pressure on households to provide sufficient in-home space for sorting and temporary storage of waste streams.

The novelty of this study lies in reframing waste management from the household perspective, showing that the effectiveness of selective collection depends not only on behavioral or organizational factors but also on the compatibility between policy-driven collection requirements and the spatial realities of the built environment.

The paper’s structure is as follows: In Section 2, the materials and methods are presented, including the study context, data sources, and the applied statistical approach. The third section of the document presents the results of the analysis. Section 4 discusses the findings of the results, considering demographic change, housing typologies, and the transition to a circular economy. Section 5 concludes by highlighting the main results, implications for decision-makers, limitations, and directions for future research. The Appendix A contains a wide description of the study’s context.

2. Materials and Methods

2.1. Study Context

The study area is Warsaw, the capital of Poland, where intensive urban transformation has been observed over the past two decades, from 2003 to 2023 [43,44]. The focus is on changes in building density and residential structures at the district level, reflecting housing development trends [45,46,47]. The city is divided into 18 districts, each with diverse morphological, functional, and demographic characteristics [48]. Over the analyzed period, the population increased from 1.69 million in 2003 to over 1.86 million in 2023, de facto figure may be higher by around 180,000 unregistered residents [43,49].

Figure 1 presents the number of residents in each district of Warsaw in the years 2003 and 2023 [50].

Warsaw has an average population density of 3600 inh. km⁻², lower than many Western European cities due to its large green areas (over 30%) [51]. Density varies strongly between the districts: midtown areas such as Praga-Południe, Ochota, and Wola reach 7800–8300 inh. km⁻², while peripheral districts like Wawer, Wesoła, and Rembertów are less dense (1100–1300 inh. km⁻²). In recent years, the city center (Śródmieście) lost 28% of its population, while outer districts have grown rapidly—most notably Wilanów (+271%) and Białołęka (+144%) [52,53]. Projections to 2050 indicate continued strong growth in peripheral districts such as Bielany, Wilanów, and Ursus [54].

The analysis show differences in population density, household sizes, and residential housing type (single-family vs. multi-family). As illustrated in Figure 2, multi-family housing development areas are more concentrated west of the Vistula River [55]. A comprehensive description of the demographic and spatial characteristics of Warsaw is provided in Appendix A.1.

Figure 2. Location of residential areas in Warsaw, split into single-family housing and multi-family housing development areas [55].

Figure 2. Location of residential areas in Warsaw, split into single-family housing and multi-family housing development areas [55].

The paper links urban features with MSW generation in Warsaw, including the city’s selective waste collection system, developed in response to EU waste directives and broadened according to the Circular Economy framework [56,57,58,59,60]. A detailed description of legal changes in the MSW management system in Poland is provided in Appendix A.2.

In Poland waste management is regulated locally but aligned with the national system, involving at-source collection from residential buildings. MSW is divided into fractions: mixed waste, paper/cardboard, glass, plastics/metals, biowaste, green waste, and bulky waste. Collection rules differ by housing type, with single-family areas allowed longer storage times due to fewer households and larger lots, as shown in

Table 1. Frequency of MSW collection in single-family and multi-family residential buildings in Warsaw, with a minimum size of waste bags or containers.

	Single-Family Housing		Multi-Family Housing
	Collection Frequency	Container Size	Collection Frequency	Container Size
Mixed waste	every four weeks	120 L and more	twice a week	120 L and more; 1100 L are most common
Paper and cardboard	every four weeks	240 L	every week	120 L and more; 1100 L are most common
Glass	every four weeks	120 L	every four weeks	120 L and more;
Plastic and metal	every four weeks	240 L	every week	120 L and more; 1100 L are most common
Biowaste	every week	120 L	every week	120 L and more; 240 L are most common

[61].

In the yellow container/bag, plastic, multi-material packaging, and metals are collected as a single waste fraction. According to the study, metals account for 7.8% of the total mass of this fraction, and multi-material packaging accounts for 5.2% of the total mass. The remaining 87% consists of plastic waste [62].

The sitting of waste collection points/shelters in residential housing estates obey national regulations, which stipulate that the maximum distance to collection facilities equipped with recycling bins must not exceed 80 m in multi-family buildings [63]. Importantly, waste collection points must be accessible for people with disabilities, e.g., people with reduced mobility and vision impairment. The regulations do not specify minimum or maximum siting distances for single-family housing.

The city authorities have proposed minimum storage capacity guidelines for MSW [61], based on average daily per inh. generation rates. According to these recommendations, the estimated reference volumes are as follows: 2 L per person per day for paper waste; 2 L for metal and plastic; 0.4 L for glass; 0.35 L for biowaste (as part of municipal waste); and 4.5 L for mixed (unsorted) MSW. In the case of green waste, capacity requirements are to be determined individually, based on actual quantities generated and specific local conditions.

2.2. Data Sources

This study uses data from official statistical sources from 2003 to 2023, including the Central Statistical Office and reports from Warsaw City Hall [48,55,63,64,65,66,67,68]. The empirical material collected includes demographic and housing data aggregated at the district level and citywide information on the amount of municipal solid waste generated [50].

One of the key parameters used in the analysis is the average net floor area (NFA) of apartments, both overall and per inhabitant. This indicator was adopted as a basic measure of housing conditions and their spatial and temporal variation within the city. Choosing the most recognizable parameter for the comparison is problematic due to inconsistency in definitions between the countries [69]; therefore, the parameter used in this paper was chosen based upon the Polish residential floor area measurements used in the National Statistics, which is Useful (Net) Floor Area of a Dwelling [70], which is better recognized internationally as Floor Area of a Dwelling [71].

The analyses of the housing areas in the city were conducted according to the division made by the city planning department in the urban planning documents, that is, with the division into Gross Development Area in 2003 (GDA) [63] and Net Development Area in 2023 (NDA) [64]. The GDA of the residential function refers to the total area of land, including the investment lot with residential buildings and structures, internal roads, alleys, driveways, separated parking lots, open spaces (such as courtyards, green areas, and recreational zones), utility areas, and infrastructure [72]. NDA of the residential function refers to the total site, including local access roads, parking areas, footpaths, and local open space such as children’s play areas and amenity space [73].

The analyses were conducted regarding two residential areas—single-family and multi-family. A single-family residential area is a zone covered by detached, semi-detached, terraced (row) buildings, or clustered development, with a maximum of two housing units. Residential areas with buildings of three or more housing units are considered multi-family residential areas [74].

The analysis also included the MSW generation rate per inhabitant (expressed in kg per inh. per year) [75,76,77,78,79,80,81], which was used to measure the impact of waste generation on housing in individual districts.

Table 2. MSW generation rates for Warsaw for selected years.

Year	2003	2006	2007	2008	2013	2016	2017	2018	2019	2020	2021	2022	2023
MSW generation rate, kg inh−1	271	299	313	328	349	350	358	367	375	380	400	409	433

presents MSW generation rates for Warsaw for the years 2003–2023 [63,64,65,66,67,68].

Due to the lack of consistent statistical data on the share of municipal solid waste generated by households in all MSW stream before 2018, this study adopts an estimated average value for the entire analysis period. Based on official data available for 2018–2023, the household sector accounted for approximately 85.5% of total MSW generated annually [82]. This share reflects the quantity of waste produced by households; thus, the legal definition of municipal solid waste also includes waste from commercial, service, and institutional entities if its composition is similar to that of household waste [83].

The analysis presented in this article is limited to the selective collection of the basic MSW fractions, known in Poland as the five-container system [84,85]. This system encompasses mixed (unsorted) waste, paper and cardboard, plastic and metal, glass, and bio-waste. Analyzing the share of selective collection fractions was based on official data from Statistics Poland for 2010–2023, using the Municipal Infrastructure Reports [85]. Due to the lack of reliable data for 2003–2010, extrapolation was not applied, and the study period was limited to 2010–2023 to minimize uncertainty.

The above data served as a starting point for the analysis from the perspective of waste generation intensity. The dataset was supplemented with contextual variables reflecting spatial and structural conditions, such as building type (single-family and multi-family housing), population density, and changes resulting from implementing the selective waste collection system.

Integrating this data enabled a comprehensive analysis of the relationship between housing structure characteristics and MSW generation intensity, both temporally and spatially.

2.3. Statistical Methods

For each dataset, mean values were calculated and supplemented with the standard deviation (SD) and the coefficient of variation (CV) [86]. These descriptive statistics were used to evaluate the extent of temporal and spatial variability across the districts of Warsaw over the analyzed period.

A panel data regression framework was introduced to analyze the relationships between household waste generation and urban characteristics [87,88]. Using panel data enables the formulation of three primary model specifications: mixed regression, fixed effects, and random effects. These specifications reflect divergent assumptions regarding heterogeneity across cross-sectional units.

In this analysis, we used panel data for the years 2003–2023 for 18 districts of Warsaw. The dependent variable was the annual amount of household MSW generated per residential unit, measured in kilograms yearly. The analysis incorporated two explanatory variables: number of inhabitants per Warsaw district (X1) and average residential unit size in square meters (X2). Panel data regression enabled the investigation of temporal and spatial variations in household MSW generation patterns across Warsaw. The general panel regression model is presented in Formula (1):

$$Y_{it}={\alpha }_i+{\beta }_1{X1}_{it}+{\beta }_2{X2}_{it}+{\epsilon }_{it}$$

(1)

i—districts.

t—time.

${\alpha }_i$ unobserved, time-invariant district-specific effects.

To determine the estimation strategy for the panel regression model, the Durbin–Wu–Hausman specification test was used. The test evaluates whether the unobserved individual effects are correlated with the explanatory variables [87,88].

The null hypothesis (Formula (2)) posits an absence of correlation, thereby indicating the consistency and efficiency of the random effects estimator. The rejection of the null hypothesis implies that the fixed effects model is preferred, given the inconsistency of the random effects estimates.

$$\mathrm{H}0: {\alpha }_1={\alpha }_2=\cdots ={\alpha }_n$$

(2)

The main advantage of the applied panel regression model is the simultaneous analysis of temporal and spatial variability between districts for selected variables, supplemented by correlation analysis. The model’s limitations result from the small number of available explanatory variables (interdependent variables were excluded) and the simplifying assumptions of linearity and homogeneity on which the estimators are based. This approach provides a robust and transparent tool for examining the relationship between demographic and housing characteristics and municipal waste generation in Warsaw’s districts.

In addition to the panel regression, a simulation was conducted to evaluate the impact of increasing segregation levels on the volume of waste fractions collected at source. The analysis included basic fractions collected at the household level: paper and cardboard, glass, plastic, metal, bio-waste, and mixed (unsorted). The generation rate for each fraction was assessed based on data presented in

Table 3. Share of source-separated MSW fractions collected in Poland between 2010 and 2023.

Selectively Collected MSW	2010	2012	2013	2015	2017	2018	2019	2021	2022	2023
Paper and cardboard, (%)	1.7	1.9	1.7	2.2	1.9	2.2	2.7	4.2	4.1	4.2
Glass, (%)	2.1	2.9	2.8	3.9	3.9	4.0	4.5	6.0	5.9	5.5
Plastic and metal *, (%)	1.2	1.8	1.9	6.6	7.1	7.3	6.9	8.6	8.7	8.9
Biowaste, (%)	1.8	2.1	2.8	6.0	7.5	8.1	9.4	14.1	14.2	15.2
Selectively collected MSW (all fractions) *, (%)	8.5	10.5	11.3	23.3	27.0	28.9	31.2	41.6	39.7	40.6

* also fractions not included in the 5-bin collection scheme.

and statistical data for Poland on the share of waste fractions collected separately in 2010–2023 [82]. In this case, it was not possible to cover the whole analysis period of 2003–2010. These mass values were converted to volume using typical for Polish conditions waste bulk density values—mixed (unsorted) MSW—200 kg·m⁻³, biowaste—300 kg·m⁻³, glass—400 kg·m⁻³, paper and cardboard—80 kg·m⁻³, and plastic and metal—50 kg·m⁻³. These values are representative of Poland and are recommended as appropriate reference parameters in the design of waste storage spaces [84,89].

The annual volume for a given household was calculated using the following formula (3) [90]:

$$V_f={\displaystyle \frac{W_f\times n}{\rho }}$$

(3)

where

W_f—generation rate of waste fraction, kg inh⁻¹ year.

n—number of people in the household.

$\rho$—bulk density of waste fraction, kg m⁻³.

In Warsaw, the frequency of waste collection from residential properties varies depending on the fraction, ranging from twice a week (e.g., mixed waste, biowaste in the summer) to once every four weeks (e.g., glass). In municipal solid waste, mixed (residual) waste is the dominant fraction and is a reference point for calculating household waste removal frequency. Therefore, the analysis assumed storage space for waste removal once a week (52 times a year) [61].

3. Results

3.1. Residential Space Distribution

Time variability and spatial differentiation in the average NFA of residential units and the number of residents per household across Warsaw’s districts between 2003 and 2023 are presented in

Table A2. Descriptive statistics of the average net floor area and number of residents per residential unit in Warsaw districts in the years 2003–2023.

	Average NFA				Number of Residents
	Range	Mean	SD	CV (%)	Range	Mean	SD	CV (%)
Warsaw	55.8–59.0	58.1	±1.1	1.83	1.7–2.3	2.0	±0.2	9.20
Bemowo	59.9–62.2	61.3	±0.7	1.17	2.0–2.8	2.3	±0.2	10.27
Białołęka	65.1–69.0	66.8	±1.0	1.47	2.0–2.4	2.1	±0.2	7.48
Bielany	51.3–54.1	52.9	±0.9	1.69	1.8–2.4	2.1	±0.2	9.49
Mokotów	53.0–56.4	55.3	±1.0	1.82	1.6–2.2	1.8	±0.2	10.32
Ochota	49.7–51.4	50.7	±0.6	1.14	1.6–2.3	1.8	±0.2	10.39
Praga-Południe	51.6–53.0	52.4	±0.4	0.85	1.7–2.3	2.0	±0.2	10.01
Praga-Północ	44.2–45.2	44.7	±0.3	0.75	1.5–2.4	2.0	±0.3	13.20
Rembertów	71.5–77.5	75.4	±1.9	2.46	2.2–2.9	2.5	±0.2	7.43
Śródmieście	48.5–51.2	50.3	±0.9	1.88	1.3–1.9	1.6	±0.2	12.29
Targówek	53.1–54.6	54.0	±0.5	0.89	1.9–2.6	2.3	±0.2	9.18
Ursus	56.5–58.5	57.7	±0.7	1.13	1.9–2.4	2.1	±0.1	6.34
Ursynów	69.2–71.6	70.4	±1.0	1.42	1.9–2.5	2.2	±0.2	7.52
Wawer	90.3–102.5	97.9	±4.1	4.20	2.3–2.8	2.5	±0.2	6.81
Wesoła	108.4–112.2	110.4	±1.2	1.05	2.6–2.8	2.7	±0.0	1.76
Wilanów	88.6–108.3	98.4	±7.2	7.34	1.5–2.3	1.8	±0.2	12.00
Włochy	60.7–66.5	64.1	±1.4	2.26	1.7–2.3	2.0	±0.2	10.96
Wola	43.5–47.5	45.5	±1.2	2.58	1.4–2.2	1.7	±0.2	14.18
Żoliborz	53.1–53.3	54.9	±1.2	2.11	1.6–2.0	1.8	±0.1	8.44

(Appendix B). It indicates a gradually increasing average NFA across the city (5.4% increase), ranging from 55.8 m² to 58.8 m². The coefficient of variation calculated for this period for the entire city was relatively low (CV = 1.83%).

The descriptive statistics reveal significant differences between the city’s districts. The largest NFA recorded during the analyzed two decades was observed in Wesoła (110.4 m²), Wilanów (98.4 m²), and Wawer (97.9 m²), the districts with a high level of single-family buildings. The lowest values were recorded in Praga-Północ (44.7 m²), Wola (45.5 m²), and Śródmieście (50.3 m²), respectively.

From 2003 to 2023, most districts experienced a moderate increase in residential unit size, with selected districts gradually declining, including Wilanów and Bemowo.

Table A2 (Appendix B) also presents the data on the number of residents per housing unit in 2003–2023. The overall trend in Warsaw indicates a decline in household size, but the pace and scale of this transformation vary significantly between districts. The occupancy load of residential units fell from 2.31 people in 2003 to 1.73 in 2023 (a 25% decrease). The coefficient of variation (CV = 9.20%) shows moderate variability over time. The highest average occupancy rates were observed in Wesoła (mean = 2.7), Wawer, and Rembertów (2.5). These districts also exhibited relatively stable occupancy patterns: 1.76, 6.81, and 7.43, respectively. The lowest average number of inhabitants per residential unit was observed in central districts, including Śródmieście (mean = 1.6), Wola (1.7), and Żoliborz (1.8). As demonstrated in Table A2 (Appendix B), the Śródmieście district also exhibited the most significant decline, with a 33% decrease from 1.92 in 2003 to 1.27 in 2023. The most significant temporal variability was recorded in Wola (CV = 14.18%), Wilanów (CV = 12.00%), and Praga-Północ (CV = 13.20%).

3.2. Determinants of Household Waste Generation

Table A3. Descriptive statistics of the average MSW generation per residential unit in Warsaw districts in the years 2003–2023.

	Range	Mean	SD	CV (%)
Bemowo	734–888	782.7	±42.0	5.37
Białołęka	642–875	741.9	±58.1	7.83
Bielany	645–841	719.8	±49.3	6.85
Mokotów	581–727	631.6	±36.4	5.76
Ochota	578–727	634.0	±38.4	6.05
Praga-Południe	613–786	675.6	±43.7	6.47
Praga-Północ	616–829	686.3	±54.5	7.95
Rembertów	743–975	864.1	±56.1	6.50
Śródmieście	513–648	554.2	±32.9	5.94
Targówek	691–908	779.1	±49.9	6.41
Ursus	637–833	734.7	±50.8	6.91
Ursynów	664–853	758.1	±49.2	6.49
Wawer	759–1044	878.0	±76.1	8.67
Wesoła	723–1126	937.3	±107.2	11.43
Wilanów	499–810	616.0	±78.3	12.71
Włochy	611–774	675.0	±44.5	6.60
Wola	536–700	592.9	±41.7	7.03
Żoliborz	543–723	611.0	±55.7	9.12

(Appendix B) presents the time variability and spatial differentiation in MSW generation per residential unit across Warsaw’s districts between 2003 and 2023.

During the analyzed period (2003–2023), a systematic increase in the average mass of MSW generated per residential unit in Warsaw was observed. The average value for the entire city increased from 532 kg in 2003 to 637 kg in 2023 (20% increase). The most significant growth was observed during the periods 2009–2011 and 2017–2020. Considerable variations have been observed among the districts of Warsaw. As demonstrated in Table A3 (Appendix B), districts such as Wesoła (957 kg), Wawer (887 kg), and Białołęka (744 kg) have consistently generated the highest MSW mass per residential unit per year in 2023. The districts of Śródmieście (467 kg), Wola (512 kg), and Żoliborz (615 kg) exhibited some of the lowest rates.

The CV was observed to be highest in peripheral districts such as Wilanów (12.7%) and Wesoła (11.4%), while the lowest variability was observed in Bemowo (5.4%) and Mokotów (5.8%). The most significant change in annual MSW generation is observed in Wesoła, Wawer, and Białołęka, while in the city center (Śródmieście), it was practically zero. A higher mean value and greater dispersion are exhibited by peripheral districts such as Wesoła, Wawer, and Rembertów, suggesting higher levels of waste generation and year-to-year fluctuation. Central districts such as Śródmieście, Wola, and Ochota exhibit more stable and lower levels of MSW generation. The lowest average number of inhabitants per residential unit was observed in central districts, including Śródmieście (mean = 1.6), Wola (1.7), and Żoliborz (1.8). As demonstrated in Table A2 (Appendix B), the Śródmieście district also exhibited the most significant decline, with a 33% decrease from 1.92 in 2003 to 1.27 in 2023. The most significant temporal variability was recorded in Wola (CV = 14.18%), Wilanów (CV = 12.00%), and Praga-Północ (CV = 13.20%).

About the individual effects, the number of inhabitants per district (X1) showed a strong and statistically significant positive relationship with MSW generation in households (B = 2.18, SE = 0.28, t = 7.68, p < 0.001). This suggests that, holding other factors constant, each additional inhabitant is associated with an increase of approximately 2.18 kg of MSW per residential unit per year.

The average residential unit size (X2) also exhibited a positive coefficient (B = 2.23, SE = 1.14, t = 1.95), but the effect was only marginally significant (p = 0.051). The 95% confidence interval (−0.01 to 4.47) indicates that the role of residential unit size in influencing waste generation remains uncertain.

The correlation between population size and average residential unit size was calculated for each of Warsaw’s 18 districts and the city as a whole (Figure 3).

For Warsaw, the correlation coefficient between population size and average unit size was −0.51, indicating a moderate negative association across the city during the studied period. At the district level, the results exhibited significant variation. A strong negative correlation was observed in Wilanów (−0.97), Włochy (−0.95), Ochota (−0.97), Praga-Północ (−0.94), and Śródmieście (−0.88). These findings indicate consistent statistical patterns across the years. In contrast, several districts exhibited positive correlations, including Rembertów (0.86), Wawer (0.85), Targówek (0.72), Żoliborz (0.59), and Wesoła (0.37).

An analysis of the mass of MSW generated per square meter of NFA of residential units in Warsaw from 2003 to 2023 reveals considerable variation between districts and within the analyzed period. As demonstrated in Tabel B3, the average intensity of waste generation per square meter of residential area at the Warsaw level increased from 11 kg m⁻² in 2003 to 13 kg m⁻² in 2023. However, the dynamics of change vary significantly depending on the characteristics of individual districts.

Central districts, including Śródmieście, Wola, and Praga-Południe, are characterized by higher levels of waste per unit area. The most significant increases in this regard were observed in 2009–2011 and following 2018. The average level of the mass of MSW generated per square meter in districts such as Mokotów, Ursynów, Bielany, and Żoliborz ranged between 11 and 13 kg m⁻² and was relatively stable. Central and inner-city districts such as Śródmieście, Praga-Północ, and Wola display higher median values and relatively narrow interquartile ranges, indicating consistently high generation per unit area. In contrast, peripheral districts such as Wesoła, Wilanów, or Wawer generally display lower median values and wider variability, where included. Outliers and higher dispersion are visible, especially in Rembertów and Praga-Północ, suggesting irregular waste patterns.

3.3. Mass and Volume of Waste Fractions Collected at Source

Table 3 presents the share of source-separated MSW collected in Poland between 2010 and 2023 [82].

An intensive growth in source-separated MSW in Poland was observed between 2010 and 2023, as shown in Table 3. In 2010, the rate of separate collection of MSW was only 8.5%. This percentage has indicated a consistent upward trend over the analyzed period, reaching 40.6% in 2023. The most significant increase was observed during the periods 2013–2015 (from 11.3% to 23.3%) and 2019–2021 (from 31.2% to 41.6%). Of the individual fractions, the most significant increase was for biowaste, from constituting 1.8% in 2010 to 15.2% in 2023. This acceleration was evident after 2015. Furthermore, plastic and metal waste increased from 1.2% in 2010 to 8.9% in 2023. Paper and cardboard, along with glass, presented more consistent growth, from 1.7% to 4.2%, and 2.1% to 5.5%, respectively. The aforementioned dynamics of change are based on statistical data regarding the mass of waste generated.

Figure 4 illustrates the mass of MSW basic fractions per capita (paper and cardboard, glass, plastics and metals, and biowaste) in the period 2010–2023.

As presented in Figure 4 in 2010, the values for all categories ranged from approximately 5 to 8 kg per inhabitant. By 2023, these values had increased to around 17 kg for paper and cardboard, 21 kg for glass, 33 kg for plastics and metals, and over 55 kg for biowaste.

Figure 5a,b compare the average mass and volume of MSW fractions generated per apartment in Warsaw in 2010 and 2023, showing a significant increase over the analyzed period.

In terms of mass (Figure 5a), the most significant increase was recorded for the biowaste fraction, increasing from 12.0 kg per housing unit in 2010 to over 96.8 kg in 2023. A considerable rise was also observed in plastic and metal, growing from 8.0 to over 56.7 kg. The mass of glass, paper and cardboard increased from 14.0 to 35.0 kg and from 11.4 to 26.7 kg, respectively. These values represent the average annual mass of waste generated per residential unit.

The dynamics mentioned above are confirmed by data on waste volume (Figure 5b), which highlights a significant increase in the volume of plastic and metal, rising from around 160 L in 2010 to over 1133 L per housing unit in 2023. Notable increases in volume were also observed for the biowaste, from 40 L to 322 L, and for paper and cardboard (from 142 L to 334 L). On the other hand, the glass volume increased relatively slightly, from 35 to 88 L annually.

4. Discussion

4.1. Waste Generation

The NFA spatial-temporal analysis in Warsaw indicates that housing conditions have improved over the last 20 years, consistent with broader trends in Europe [22]. On the other hand, there is a clear disproportion between the city’s districts, from the compact central units in Wola or Praga-Północ (apartments of approximately 45 m²) to the much larger suburban residential units in Wesoła or Wilanów (average size of over 98 m²). This diversity confirms earlier observations that aggregated averages mask local realities, as both urban form and housing typology significantly shape household waste generation [14,22].

The studies cited in the article indicate that demographic factors significantly influence waste generation trends [26,28]. The average household size in Warsaw decreased from 2.31 persons per unit in 2003 to 1.73 in 2023 (a 25% reduction). This value is significantly lower than the average for Poland (2.9) and the EU (2.3), reflecting the broader European trend of decreasing household numbers [21,32]. Studies show that as the number of people in a household decreases, people generate more waste per person [91], resulting in a higher total waste generation despite a lower total household production. Demographic change, particularly the spread of single-person and small households, represents a structural challenge for selective collection systems. The problem is deepened by physical constraints of apartments in dense districts, where a lack of storage capacity reduces the feasibility of multi-fraction separation [14,25]. This is unfavorable from the perspective of the CE principles and increases the demand for waste collection space in these residential units [92]. A significant problem and challenge in this perspective is the methodology for collecting data on the number of residents and the underestimation of the actual population of Warsaw in official statistics, by nearly 10% on a citywide scale [43,93]. This complicates infrastructure planning, leading to a systematic discrepancy between actual waste production and available collection capacity.

Regression analysis confirmed that the number of residents per unit was the most robust predictor of household waste production. The role of residential unit size is shown to be positive but less specific. Further studies should investigate this relationship, using extended panel data and additional control variables.

Correlation analysis between population increases and residential unit size also indicates that waste management challenges in Warsaw differ and are shaped by diverse spatial and demographic conditions. In densely populated districts, strong negative correlations reflect the dynamics of a compact city, where shrinking household space limits the opportunities for selective collection. Increasing NFA in outlying districts provides storage capacity in suburban areas, demonstrating positive correlations.

The highest average mass of MSW generated per residential unit was recorded in the peripheral districts of Warsaw, such as Wesoła, Wawer, and Białołęka. This trend can be attributed to the larger average household sizes in these areas and the domination of single-family housing, resulting in higher waste generation per unit (Table A3, Appendix B). These districts are also experiencing dynamic population growth, which reflects broader demographic changes and the city’s ongoing suburbanization process [45].

While descriptive statistics show the increase in MSW generated per capita (from 271 kg inh.⁻¹ in 2003 to 433 kg in 2023), the interpretive value is in understanding these dynamics as a product of policy shifts and infrastructural adaptation. They correspond to key stages in the transformation of Poland’s waste management system (Appendix A.2): the period of adaptation to EU membership (2002–2006), municipalities’ assumption of responsibility (2013–2016), and integration with the principles of the circular economy (since 2017). The most intensive growth occurred between 2008 and 2011 (over 65 kg inh.⁻¹), which may be the result of, among other things, infrastructure investments and the consumption increase following EU accession [94]. In parallel, the mean weight of waste per housing unit rose from 532 to 637 kg per year between 2003 and 2023, reflecting the escalating burden placed on households in terms of waste storage.

The changes in municipal waste generation observed in individual districts of Warsaw over the 20 years under review reflect the effects of the city’s spatial and demographic transformation and changes in the waste management system (Appendix A.2). It has significant consequences in the context of selective waste collection [95,96].

4.2. Separate Collection

The perspective on separate collection provides additional understanding of the transition of the MSW management system in Warsaw. While the share of separately collected waste rose from 8.5% in 2010 to 40.6% in 2023, the critical issue lies in how spatial and infrastructural settings conditioned this progress. Suburban districts such as Wilanów, Ursynów, Wesoła, and Wawer—characterized by larger units and lower occupancy rates—offer more favorable conditions for in-home separation. By contrast, dense central areas such as Wola, Praga-Północ, and Śródmieście, with small units and higher occupancy, face systemic barriers to storing multiple fractions, a challenge widely documented in the literature [14,97]. The problem is further exacerbated by compositional changes in the waste stream: the rapid growth of bio-waste (79-fold) and plastics (7-fold) has created mounting pressure on container capacity and collection frequency, reflecting similar findings in other European contexts [98,99]. This is of particular significance due to the low bulk density of this fraction, where plastic accounts for the largest share [79], which results in a substantial volume of waste. During the period under analysis (2010–2023), a significant increase in the volume of separately collected waste per household was observed. Figure 6 presents the weekly volume of waste generated in 2- and 4-person households in Warsaw in 2010 and 2023.

The composition of MSW in 2023 is characterized by a dominance of lightweight fractions, particularly plastic, which constitute the foremost proportion of the total volume of generated waste, despite their comparatively negligible weight. The supremacy of light fractions in MSW, particularly in developed countries, results primarily from the high share of packaging waste within the overall waste stream [100]. It is not always apparent in statistical data, which are mainly based on data on the mass of waste [101,102]. A dynamic increase in the volume of MSW generated in households, even with a small growth of residential units’ floor area (NFA), can considerably affect the internal organization within household living space. It also burdens the collection system, especially regarding capacity and collection frequency.

A comparison between the calculated annual volume of MSW (presented in

Table 4. Comparison of the weekly volume of selected MSW fractions collected separately per 4-person household in 2010 and 2023, and the city’s official storage capacity recommendations.

Fraction	Calculated Volume, (L)	Volume Recommended by Warsaw Authorities, (L)
Plastic and metal	50.2	56.2
Biowaste	14.3	9.8
Paper and cardboard	20.0	56.2
Glass	4.0	11.2
Mixed (unsorted) MSW	84.0	126.4

) and the city’s official guidelines for minimum storage capacities (Appendix A.2) indicates that the annual volumes recommended (estimated on the basis of daily per capita generation rates) are higher than the values derived from this study and incorporate a safety margin, to secure conditions for effective selective collection [61].

As shown in the table above, the city’s recommendations include a safety margin. Their implementation is intended to provide appropriate conditions for effective selective waste collection at designated collection areas. Another aspect to consider is the practical application of these guidelines, for example, in designing collection points within residential estates. Furthermore, waste volumes may increase due to improper collection practices, such as failure to compress paper and cardboard before disposal [103,104].

Separate collection in Warsaw can only be understood in relation to rising source separation efficiency rates. Its strategy should combine ambitious CE-driven targets with the city’s spatial and demographic heterogeneity. In this context, this study identifies two challenges: (1) the dominance of voluminous fractions, which test the limits of household and collective storage systems, and (2) the mismatch between regulatory standards and the infrastructural capacity of specific districts. These challenges require transitioning toward multifunctional, modular, and intelligent infrastructure [105,106], integrated with governance mechanisms that ensure responsiveness to local conditions. Such an approach is consistent with urban design in the spirit of compact cities [2]. These solutions include garbage shelters, semi-underground containers [96], stationary or mobile collection points [106], and drop-off stations at commercial premises [107]. In such contexts, the door-to-door collection system [108] has been promoted to improve accessibility and participation.

Moreover, 88% of Warsaw’s residents live in multi-family housing (47% of the total residential area). On the other hand, single-family buildings account for over half of the residential area but accommodate only 12% of the residents [64]. The dominance of compact, multi-unit buildings in central and suburban districts (e.g., Mokotów, Praga-Południe, Ursynów, Bemowo, Wola) causes a growing dependence of residents on collective infrastructure solutions. This relationship requires the technical organization of the waste collection system but also has implications for social equity, as most of the city’s inhabitants rely on shared infrastructure whose quality and availability directly determine participation in selective collection.

However, as indicated in the literature, the availability of space-efficient collection infrastructure remains limited in Poland [85]. This infrastructure gap, particularly in districts characterized by the prevalence of multi-family housing, presents a significant challenge to achieving complete alignment with Circular Economy objectives. This parallels findings for other Central and Eastern European countries, where infrastructural adaptation lags regulatory expectations [51]. Spatial accessibility emerges as a pivotal factor influencing participation in selective waste collection. Several studies have indicated that optimal distances of 50–100 m are recommended between dwellings and collection points, depending on factors such as urban density and morphology [12,96]. The recommended maximum distance for basic collection points in Poland is 80 m [63]. This threshold can be exceeded for specialized waste types in dense or historically developed districts. In such cases, the distance is often measured in meters and found to be 300–500 m [109]. These longer distances to collection points have been demonstrated to have a detrimental effect on participation rates and to lead to an increase in contamination rates [110,111]. Spatial accessibility remains a key factor in the effectiveness of infrastructure supporting selective waste collection in households [12,96].

The general findings on household MSW generation in Warsaw and separate collection opportunities indicate that key challenges are shaped not only by general quantitative indicators but also by understanding demographic change, housing morphology, and infrastructural adaptation.

This research confirms European context findings that the efficiency of circular economy transitions in urban waste management depends on aligning regulatory ambitions with local spatial and demographic realities [14,22,96]. For Warsaw, policy should move beyond uniform targets and incorporate differentiated, district-specific approaches: capacity scaling in suburban zones experiencing dynamic population growth, and compact, modular, or innovative collection solutions in historic and high-density areas. Regulation and governance are key to improving utility performance, as transparency and accountability mechanisms encourage efficiency and higher service quality. Future approaches should integrate governance indicators and promote public-oriented management to ensure sustainable and equitable service delivery [112]. Such systems must also account for individual city districts’ heterogeneous spatial and demographic development [113,114].

5. Conclusions

The analysis considered how existing waste management strategies interact with the changing nature of urban housing, particularly in the context of district-level diversification in Warsaw, long-term trends in municipal solid waste generation, and the development of separate waste collection systems.

In Warsaw, the diversity of housing typologies—ranging from compact multi-family blocks in central districts to low-density single-family homes in peripheral areas—creates distinct spatial challenges for municipal waste management. Multi-family housing accommodates 88% of the city’s population. It necessitates integrated, space-efficient solutions adapted to limited indoor and shared infrastructure conditions, such as standardized in-building waste storage rooms, semi-underground containers, and micro-collection points within walking distance. Conversely, the expansion of single-family areas offers increased space but reduced population density, highlighting the necessity for context-sensitive strategies that integrate selective collection with waste prevention to support the city’s transition towards a circular economy.

The study results indicate the necessity of incorporating waste management considerations into urban planning and residential design. It is essential to harmonize technical waste requirements with spatial local contexts at the level of individual housing units. In addition to infrastructure changes, it is equally important to apply various incentives and regulations.

The present study is limited to the use of aggregated local residential-level data and the lack of consistent local datasets. In many areas, including Warsaw and its districts, data availability is limited, which restricts the possibility of conducting more detailed spatial analysis and may affect the generalizability of the results. Another limitation of this study is that variables such as income level and consumption patterns were not controlled, which makes it difficult to isolate the independent impact of household size on waste generation. In the next stage, the study should be extended to examine the identified waste generation trends concerning broader economic dynamics, including changes in GDP and household consumption patterns. Additionally, further analysis should investigate how fluctuations in real estate prices have influenced residential floor space over time, particularly in the context of growing urban density.

The outcomes of this study may be transferable to other urban contexts, particularly in Central and Eastern countries, where there is ongoing evolution in waste policy reform and urban densification. The applied methodology can be a replicable framework in comparative studies, provided local context and building typologies are considered.