Climate-Resilient Design of Covered Historic Courtyards in Mediterranean Climates: The Role of Roof Geometry and Passive Strategies Under Future Scenarios

Abstract

Covered courtyards are increasingly being adopted as a passive strategy for the climatic rehabilitation and adaptive reuse of historic buildings. However, their thermal behaviour is strongly conditioned by roof geometry, local climate conditions, and future climate warming, aspects that have not yet been comparatively addressed within a climate resilience framework. This study evaluates the energy and thermal performance of three representative roof typologies for covered historic courtyards—glazed dome, glazed flat roof, and south-facing sawtooth roof—across two Mediterranean climates of contrasting severity (cold continental and warm–dry), considering both current and future climatic conditions (2050–2080). Additionally, two design approaches are compared: a baseline design (BD), based exclusively on geometric configuration and standard glazing, and an enhanced passive design (EPD), which incorporates improved glazing, controlled natural ventilation, and seasonal solar control. Dynamic simulations using EnergyPlus/DesignBuilder are employed to analyse heating and cooling demands, free-running thermal behaviour, overheating risk, and the climatic robustness of each solution. The results show that roof geometry constitutes the dominant factor governing the long-term thermal resilience of covered courtyards, particularly under future climate warming scenarios, while enhanced passive strategies significantly mitigate cooling demand and overheating in the most penalised typologies. The south-facing sawtooth roof consistently exhibits the highest climatic robustness under free-running conditions across the analysed scenarios, whereas the glazed dome and flat roof solutions display greater climatic sensitivity and benefit more substantially from the application of enhanced passive design strategies. Overall, the results provide quantitative design criteria to support resilient interventions in historic covered courtyards in Mediterranean climates under climate change.

1. Introduction

Courtyard architecture has historically constituted one of the most effective environmental devices for moderating indoor microclimates in warm and Mediterranean regions, due to its capacity to regulate solar radiation, promote natural ventilation, and dampen diurnal thermal fluctuations [1,2,3,4,5]. Its thermal performance depends primarily on geometric parameters—such as sky view factor, height-to-width ratio, and the degree of volumetric enclosure—as well as on the hygrothermal properties of the envelope components [6,7,8,9,10,11,12]. At the international level, numerous studies have demonstrated the effectiveness of traditional courtyards as a passive strategy to enhance indoor comfort and reduce cooling loads in arid, temperate, or humid climates [13,14,15]. However, most of this evidence focuses on open courtyards and on warm or temperate climates, while the behaviour of historic courtyards transformed into covered spaces in cold climates or in contexts with high thermal amplitude remains insufficiently documented [16,17,18,19,20,21,22].

In the context of the rehabilitation of built heritage, courtyards have undergone a growing transformation [23] through their enclosure with glazed systems, particularly since the late nineteenth century, with the aim of consolidating them as spaces for circulation, reception, or social interaction [24,25,26]. The introduction of translucent roofs transforms an outdoor microclimate into a semi-controlled indoor volume with a radically different thermal behaviour, characterised by increased solar gains and the greenhouse effect associated with glazing [18,21]. Although such roof systems may reduce thermal losses in winter and optimise access to natural daylight, they also present significant vulnerabilities related to overheating and increased cooling demand, even under moderate climatic conditions [19,27]. This risk is further intensified under future climate warming projections, particularly in Mediterranean regions, where a substantial increase in summer temperatures and in the intensity of extreme heat events is anticipated [28]. In this context, roof geometric configuration emerges as a key factor for modulating solar gains, overheating potential, and the thermal stability of the indoor space [16,29,30]. Recent studies on courtyards and atria further corroborate that form critically influences the balance between solar gains, natural ventilation, and overheating potential [1,20,31].

Despite this body of evidence, the literature still contains significant scientific gaps. Most studies focus on a single atrium or covered courtyard configuration, typically in contemporary buildings, without exploring comparative geometric variants or their behaviour under contrasting Mediterranean climates. Likewise, studies addressing future climate change scenarios (2050–2080) remain scarce and limited in scope [19,32]. Existing research rarely evaluates climate resilience metrics, such as thermal stability, free-running behaviour, or the sensitivity of cooling demand to progressive temperature increases [33]. According to recent reviews, the interaction between architectural form, glazing, natural ventilation, and local climate continues to be one of the least explored topics within the field of atria and covered courtyards [34,35]. Recent comparative studies focused on historic covered courtyards have shown that thermal response and cooling demand vary non-linearly with roof geometry; however, these analyses have so far been restricted to a single cold climatic context, without addressing robustness across contrasting Mediterranean climates or future warming scenarios.

In parallel, the growing attention to the adaptive reuse of architectural heritage highlights the need for passive strategies that reduce reliance on mechanical systems while enabling the reconciliation of conservation requirements, thermal comfort, and energy efficiency [36,37,38]. The integration of solar-control glazing, regulated natural ventilation, shading elements, and seasonal roof management has been proposed as an effective solution to limit overheating in atria and covered courtyards [39,40]. Nevertheless, the robustness of these strategies under future climatic conditions remains insufficiently explored. In particular, there is a lack of studies that jointly analyse the interaction between roof geometry, advanced passive strategies, and future climate evolution in historic covered courtyards.

In recent years, the enclosure or covering of traditional courtyards has become an increasingly frequent intervention in architectural practice across different geographical regions, particularly within adaptive reuse and heritage rehabilitation projects [35,41,42]. Such transformations are widely implemented to improve the functional performance of interior spaces, enhance environmental control, and extend the usability of courtyard areas throughout the year while maintaining their spatial and architectural identity. Previous studies have also shown that courtyard and atrium configurations play a significant role in regulating microclimates and influencing building energy demand, and that design variables such as courtyard geometry, orientation, envelope design, and roof systems strongly affect thermal performance [43]. Consequently, analysing the thermal and energy implications of alternative roof geometries becomes essential for supporting informed and sustainable design decisions in courtyard-based buildings [44].

In this context, the present study addresses a relevant research gap through a comparative evaluation of the thermal and energy performance of three roof typologies for historic covered courtyards: glazed dome (SD), glazed flat roof (FLAT), and south-facing sawtooth roof (SAW).

The analysis is conducted across two Mediterranean climates of differing severity—Csb (León) and Csa (Granada)—according to the Köppen–Geiger climate classification [33,45], and considers both current climatic conditions and future projections for 2050 and 2080. To this end, morphed climate files generated using widely validated methods are employed [46,47], enabling the assessment of the impact of climate warming on energy demand, overheating, and free-running thermal behaviour [48].

In addition, two design approaches are compared: a baseline design (BD), based on standard glazing and without additional passive strategies, and an enhanced passive design (EPD), which incorporates selective glazing, controlled natural ventilation, and seasonal solar control.

The present study aims to comparatively evaluate the thermal and energy performance of different roof geometries in historic covered courtyards, identifying which configurations exhibit greater robustness under current and future climatic conditions. In addition, the effectiveness of passive envelope strategies in reducing overheating and energy demand is analysed, particularly for those typologies that are initially more penalised from a thermal perspective. Finally, the evolution of these effects under future climate warming scenarios is examined, with the aim of identifying long-term trends and potential trade-offs in the energy balance, and of providing quantitative design criteria for the resilient rehabilitation of historic covered courtyards in Mediterranean climates.

Accordingly, this study addresses the following research questions:

• RQ1. How does roof geometry influence thermal and energy performance under current climatic conditions?
• RQ2. How effective are passive design strategies across different Mediterranean climates?
• RQ3. What is the impact of future climate scenarios on the resilience of covered courtyard roof typologies?

2. Materials and Methods

2.1. Case Study Geometry and Roof Typologies (SD/FLAT/SAW)

The methodological design of this study is structured to ensure comparability across scenarios and the scientific reproducibility of the results. All simulations are based on a common baseline model, keeping the opaque envelope, internal loads, usage schedules, and courtyard geometry constant, so that any observed variation in thermal or energy performance can be attributed exclusively to roof typology, the implemented passive strategy, or the climatic scenario. The three roof configurations are integrated through an equivalent geometric substitution, avoiding unintended form-related effects.

A representative covered courtyard typical of Mediterranean heritage architecture was modelled, with the courtyard geometry, opaque vertical envelope, and usage profiles kept invariant, so that differences between cases can be attributed solely to the roof typology and the passive strategies evaluated. Three common solutions for courtyard enclosure were analysed: glazed dome (SD), glazed flat roof (FLAT), and sawtooth roof with a south-facing vertical glazed surface (SAW).

These three geometric roof typologies commonly used in the enclosure of courtyards (dome, flat, and sawtooth) were analysed in order to isolate the effect of the upper geometry on the thermal balance and solar gains. The literature shows that enclosing a courtyard with a glazed roof can cause its behaviour to approximate that of an atrium, simultaneously modifying energy demand and thermal comfort, with high sensitivity to climate, glazing ratio, and the optical properties of the enclosure [30]. Likewise, in sawtooth roof configurations, the orientation and geometry of the glazed openings play a decisive role in governing solar gains and their seasonal control [18].

The reference geometric model derives from previous studies and experimental analyses conducted by the authors, based on the architectural interventions illustrated in Figure 1, carried out within knowledge-transfer activities related to heritage rehabilitation and adaptive reuse, with the vertical envelope, construction materials, occupancy, usage schedules, and internal load profiles kept constant in order to isolate the effects exclusively associated with roof geometry and the passive strategies evaluated.

The case study corresponds to an atrium (covered courtyard) with a rectangular plan, featuring an open courtyard area of 11.15 × 12.10 m and a total area (courtyard plus surrounding portico) of approximately 310 m². The space is enclosed by two storeys and presents a cornice height of approximately 10 m. The courtyard floor consists of continuous stone paving, with no vegetation or water features; these parameters are kept invariant in all cases to isolate the effect of the roof.

The three glazed roof typologies analysed, which represent recurrent solutions in contemporary interventions on historic courtyards, are those previously described:

• Triangulated semi-dome roof (SD): A curved glazed geometry characterised by high solar exposure and a uniform shape factor.
• Nearly flat glass roof (FLAT): A continuous horizontal glazed surface with high direct solar transmittance during summer conditions.
• South-facing sawtooth roof with vertical glazing (SAW): An alternating sequence of opaque surfaces and south-oriented vertical glazed elements, specifically designed to modulate seasonal solar gains.

Figure 2 shows a 3D view of the atrium model with the roof typology superimposed on a solar path dome (stereographic diagram), together with the results of a mean daily solar radiation analysis [40]. The blue dome represents the solar trajectory and azimuth directions, while the coloured arcs along the horizon indicate the solar path segments and their corresponding intensity. At the centre, the glazed footprint of the roof over the courtyard is visible, discretised into a mesh; the chromatic map over this mesh encodes the incident irradiation (kWh/m²) according to the colour scale, enabling the identification of areas with higher exposure and the shading effect projected by the surrounding building volumes.

The three configurations constitute exclusive variants of the same base space, implemented through direct roof substitution, thereby ensuring metric comparability across scenarios.

To assess the robustness of the simulation results, a sensitivity analysis was also conducted by varying two key modelling parameters: infiltration rate and internal heat gains. Infiltration was varied between 0.5 and 2 ACH around the base case value of 1 ACH, while internal heat gains were varied between 0 and 10 W·m⁻². The rest of the modelling assumptions remained unchanged. The results of this analysis are presented in Section 3.1.

2.2. Study Locations and Köppen–Geiger Climate Classification

The analysis is conducted across two contrasting Mediterranean climates in Spain, selected to represent differentiated thermal conditions within the Iberian Peninsula and to enable a comparative evaluation of the robustness of courtyard/atrium roof typologies under current and future climatic contexts, with potential extrapolation to regions with equivalent Köppen–Geiger classifications (see Section 5) [45]. The selected locations are León (Csb), a cool-summer Mediterranean climate with continental influence, characterised by cold winters, mild summers, and high daily thermal amplitude, and Granada (Csa), a warm–dry Mediterranean climate, with very hot summers, high summer solar radiation, and greater severity of overheating episodes.

Climatic classification is reported according to the Köppen–Geiger system to facilitate comparison with previous studies; this classification has been verified using the updated maps reported in [33,45], ensuring international taxonomic consistency. Current climatic conditions are represented using EnergyPlus EPW (Design Builder 7.3.1.003 Energy plus 9.4)weather files [49] for each location. In both cases, geometry, materials, internal loads, and operational profiles are kept constant, so that observed differences in energy demand and free-running thermal behaviour can be attributed exclusively to climate and roof typology.

2.3. Future Climate Files: SSP4.5, 2050/2080

Future climate weather files for 2050 and 2080 were generated from the baseline EPW using Future Weather Generator (FWG) software under the SSP4.5 emissions scenario [50,51]. The resulting EPW files represent projected mid- and late-century climatic conditions and were directly employed as boundary conditions in the simulations. The procedure is based on a “morphing” approach applied to the baseline file, adjusting meteorological variables to reflect projected changes while preserving the temporal structure of the original dataset [52,53,54].

To ensure comparability, three climatic horizons—2020 (baseline), 2050, and 2080—were simulated for Granada and León, with all non-climatic parameters (geometry, construction configurations, usage profiles, internal loads, infiltration rates, and control logics) kept identical. In this way, observed differences in heating and cooling demand, as well as in free-running comfort indicators, can be attributed exclusively to climatic conditions.

2.4. Design Variants: BD (Baseline Design) and EPD (Enhanced Passive Design)

Two construction configurations were defined in order to compare a reference case and an improved case under the same geometric and operational framework (see

Table 1. Comparative table defining the two scenarios: baseline case and enhanced case.

Parameter	BD (Baseline Design)	EPD (Enhanced Passive Design)
Glazing Type	Standard double glazing	Low-emissivity/selective double glazing
Ug (W/m2·K)	2.8	1.2
g (solar factor)	0.70	0.50
τv (visible transmittance)	0.75	0.70
Movable solar control	Not considered	Yes, with seasonal activation
Atrium natural ventilation	Fixed infiltration and minimal openings without control	Controlled natural ventilation
Opening criteria	Not applicable	Based on indoor–outdoor temperature. Opening when T_int > T_set and T_ext < T_int − ΔT, with ΔT ±+− 2–3 °C.
Night-time ventilation	Not modelled	Yes, scheduled night purge. Typical interval 22:00–08:00 subject to temperature criteria; automatic closure under unfavourable conditions.

). The baseline design (BD) represents the reference solution, maintaining the initial construction properties of the roof and without specific active passive strategies beyond the infiltration considered in the model. The enhanced passive design (EPD) incorporates a set of measures aimed at reducing overheating and improving thermal performance, while keeping all other system conditions (geometry, internal loads, and schedules) unchanged.

In the EPD configuration, three main modifications were implemented at the roof level: (i) improved glazing (lower thermal transmittance compared to BD); (ii) external shading applied to all roof glazing surfaces [55] (for the SD, FLAT, and SAW typologies), controlled by indoor air temperature ≥ 24 °C and operating 24/7; and (iii) controlled natural ventilation of 4 air changes per hour, in addition to infiltration, activated when T_in > T_out and T_in > 24 °C, also operating 24/7 [56]. This definition ensures that differences between BD and EPD are exclusively attributable to roof-related performance and strategies, thereby facilitating direct comparison across typologies and climatic scenarios.

The combination of BD/EPD × roof typology × climate × temporal scenario enables the assessment of both current performance and the climatic robustness of each solution.

2.5. Simulation Setup

Simulations were performed using EnergyPlus (via a modelling interface) for the three roof typologies and the two design scenarios (BD and EPD) across the two locations and the three defined climatic horizons. In order to decouple energy performance from passive thermal behaviour, two simulations were conducted for each case: (i) a conditioned simulation using an Ideal Loads Air System, aimed at estimating the annual specific heating and cooling demand; and (ii) a free-running simulation (HVAC deactivated) [57], intended to assess thermal comfort and the severity of thermal discomfort without mechanical assistance.

First, a constant infiltration rate of 1 ACH was imposed in all simulations [58], both under free-running and conditioned operation, in order to maintain a uniform level of uncontrolled ventilation across alternatives. In addition, internal gains were represented through aggregated loads of 8 W/m² [59], corresponding to occupancy, equipment, and lighting, with a schedule of 100% between 08:00 and 20:00 and 20% during the remaining hours, identical in all cases, so as to avoid introducing operational differences that could mask the geometric and constructive effects of the roof [60]. Natural ventilation defined in the EPD scenario was incorporated as an additional rate of 4 air changes per hour, potentially available 24/7 and activated only when T_in > T_out and T_in > 24 °C, added to the base infiltration.

Furthermore, no nearby urban shading or local microclimatic effects were considered, and the reference building was modelled as an isolated case; this simplification allows performance differences to be attributed primarily to roof geometry and the implemented passive strategies. In the HVAC simulations, the Ideal Loads system operated with a setpoint range of 20–25 °C under continuous operation (24/7) to obtain theoretical thermal demands strictly attributable to the modelling assumptions and the building envelope (in kWh/m²·year), avoiding spurious effects derived from operating schedules [61].

The courtyard was modelled as an explicit thermal zone within the EnergyPlus model. Two operational modes were analysed. In the conditioned simulations, the courtyard was connected to the Ideal Loads system in order to evaluate its contribution to the overall heating and cooling demand. In addition, free-running simulations were performed with the HVAC systems deactivated, allowing the assessment of thermal behaviour and discomfort severity through accumulated degree-hours. This approach enables a consistent comparison between energy demand and passive thermal performance under identical geometric and boundary conditions.

2.6. Performance Indicators and Data Analysis

As described above, the experimental design combines three roof typologies (SD, FLAT, and SAW), two construction configurations (BD and EPD), two climates (León and Granada), and three climatic horizons (2020, 2050, and 2080), resulting in a total of 36 combinations. For each combination, two simulations were performed: (i) a conditioned simulation using Ideal Loads to estimate energy demand, and (ii) a free-running simulation [57] to characterise the intrinsic thermal behaviour without mechanical intervention, an approach widely employed to assess the potential of passive strategies and natural ventilation.

The results were analysed using three groups of performance indicators:

• Energy performance indicators (HVAC ON—Ideal Loads): Annual specific heating demand and annual specific cooling demand are reported in kWh/m²·year. Energy demand, rather than final energy consumption, is considered due to the idealised nature of the system, without explicit modelling of equipment or efficiencies [62].
• Comfort and discomfort indicators (HVAC OFF—free-running): Using a static comfort criterion of 20–25 °C and an evaluation period of 8760 h/year based on indoor air temperature, the following metrics are quantified: (i) hours within the comfort range, (ii) hours of cold discomfort, and (iii) hours of heat discomfort. A strict static comfort range (20–25 °C, 24/7) was intentionally adopted as a conservative indicator of thermal stabilisation in free-running conditions, enabling robust comparison across typologies, climates, and future scenarios. In addition, discomfort severity is assessed through accumulated degree-hours [63] relative to the limits of the comfort range.

The degree-hour indicators are defined as follows:

GH__heating = Σ_h max (T_h − 25, 0)  GH__cooling = Σ_h max (20 − T_h, 0)

where T_h is the hourly indoor air temperature (°C) and h represents each hour of the year.

• Climate resilience and robustness indicators: For each performance metric X, sensitivity to future climate conditions is evaluated through relative variations. The relative variation of each performance indicator under future climate scenarios is quantified using the following expressions:

$$\mathsf{\Delta }{\mathrm{X}}_{\left(2050\right)}={\displaystyle \frac{X\_2050-X\_2020}{X\_2020}}$$

$$\mathsf{\Delta }{\mathrm{X}}_{\left(2080\right)}={\displaystyle \frac{X\_2080-X\_2020}{X\_2020}}$$

where X represents the value of the analysed performance indicator (e.g., energy demand, hours in comfort, or degree-hours) under the baseline climate (2020) and future climate scenarios (2050 and 2080).

In addition, the stability of the relative ordering among typologies is analysed for the 2050 and 2080 scenarios (robustness of the demand- and discomfort-based “ranking”). These indicators make it possible to establish comparative criteria among geometric and operational alternatives.

3. Results

3.1. Sensitivity Analysis

To evaluate the robustness of the simulation results, a sensitivity analysis was conducted by varying two key modelling parameters: infiltration rate and internal heat gains. Infiltration was varied between 0.5 and 2 ACH around the base case value of 1 ACH, while internal heat gains were varied between 0 and 10 W·m⁻².

The analysis was performed for both climates considered in the study (León and Granada), maintaining the rest of the modelling assumptions unchanged. Heating and cooling demand variations were analysed relative to the base case in order to assess the stability of the results and the consistency of the typological trends observed in the previous sections.

Table 2. Sensitivity of annual heating and cooling demand to infiltration rate (0.5–2 ACH) for the analysed roof typologies (SD, FLAT and SAW) in the León and Granada climates.

Climate	Roof Typology	Demand	0.5 ACH	1 ACH (Base)	2 ACH
León (Csb)	SD	Heating (kWh·yr−1)	26,487	48,903	96,883
		Cooling (kWh·yr−1)	20,665	19,800	19,951
	FLAT	Heating (kWh·yr−1)	26,701	51,188	107,649
		Cooling (kWh·yr−1)	13,600	12,140	9492
	SAW	Heating (kWh·yr−1)	30,318	59,064	117,719
		Cooling (kWh·yr−1)	65	157	461
Granada (Csa)	SD	Heating (kWh·yr−1)	19,977	38,884	79,930
		Cooling (kWh·yr−1)	28,212	27,822	29,380
	FLAT	Heating (kWh·yr−1)	15,217	32,250	70,193
		Cooling (kWh·yr−1)	29,301	27,845	28,229
	SAW	Heating (kWh·yr−1)	18,929	38,540	79,209
		Cooling (kWh·yr−1)	2533	3872	6989

and

Table 3. Sensitivity of annual heating and cooling demand to internal heat gains (0–10 W·m⁻²) for the analysed roof typologies (SD, FLAT and SAW) in the León and Granada climates.

Climate	Roof Typology	Demand	0 W·m−2	5 W·m−2 (Base)	8 W·m−2	10 W·m−2
León	SD	Heating (kWh·yr−1)	50,034	51,138	48,903	47,457
		Cooling (kWh·yr−1)	16,367	18,462	19,800	20,731
	FLAT	Heating (kWh·yr−1)	57,943	54,479	51,188	50,355
		Cooling (kWh·yr−1)	9378	10,256	12,140	12,039
	SAW	Heating (kWh·yr−1)	68,558	62,537	59,064	56,816
		Cooling (kWh·yr−1)	10	79	157	234
Granada	SD	Heating (kWh·yr−1)	43,844	40,677	38,884	37,728
		Cooling (kWh·yr−1)	23,570	26,178	27,822	28,955
	FLAT	Heating (kWh·yr−1)	37,140	34,014	32,250	31,118
		Cooling (kWh·yr−1)	23,424	26,133	27,845	29,023
	SAW	Heating (kWh·yr−1)	45,719	41,130	38,540	36,896
		Cooling (kWh·yr−1)	2100	3159	3872	4380

Note: The base case corresponds to an infiltration rate of 1 ACH and internal heat gains of 5 W·m⁻². SD denotes spherical dome roof, FLAT denotes flat glazed roof, and SAW denotes south-facing sawtooth roof. Heating and cooling demands are reported as annual energy values obtained from EnergyPlus simulations. The sensitivity analysis evaluates the variation in energy demand resulting from plausible changes in infiltration rate and internal heat gains while maintaining all other modelling parameters constant.

summarise the sensitivity of heating and cooling demand to variations in infiltration rate and internal heat gains.

The results indicate that heating demand is particularly sensitive to infiltration rate, especially in the colder climate of León. Increasing infiltration from 1 to 2 ACH approximately doubles heating demand, whereas reducing it to 0.5 ACH reduces heating demand by nearly half. Internal heat gains have the opposite effect, decreasing heating demand while increasing cooling demand. However, the magnitude of this effect remains moderate compared to the influence of infiltration.

Importantly, despite these variations in absolute demand values, the relative performance ranking of the analysed roof typologies (SD, FLAT and SAW) remains unchanged across all tested parameter ranges.

To verify whether the observed sensitivity patterns remain valid under future climatic conditions, an additional complementary test was performed for the León climate under the 2080 scenario (SD typology). The results show the same qualitative behaviour observed in the baseline scenario: heating demand remains highly sensitive to infiltration rate, while increased internal heat gains reduce heating demand and increase cooling demand. These results suggest that the sensitivity patterns identified for the current climate remain stable under projected future climatic conditions.

3.2. Free-Running Conditions: Hours in Comfort and Discomfort (20–25 °C)

Figure 3 presents the annual distribution of indoor thermal conditions, expressed as the number of hours out of a total of 8760 h, classified into comfort (20–25 °C), cold discomfort (<20 °C), and heat discomfort (>25 °C) for all combinations of climate (Granada and León), building configuration (BD: free-running operation; EPD: enhanced passive design), climatic scenario (2020, 2050, and 2080), and roof typology (SD: glazed dome; FLAT: flat glazed roof; SAW: sawtooth roof). In addition, the figure includes an outdoor climatic reference (EXT), derived from the dry-bulb temperature of the climatic files and classified using the same comfort thresholds. This reference does not represent the exact microclimatic behaviour of an open courtyard but provides a benchmark to interpret the thermal moderation effect introduced by the covered configurations.

3.2.1. Granada (Csa)

For the Granada climate, the thermal behaviour under free-running conditions shows a wide dispersion across configurations and scenarios, with overheating playing a dominant role in typologies with higher solar exposure. The outdoor reference indicates a predominance of cold discomfort (6567 h in 2020) with a relatively limited contribution of heat discomfort (1128 h). However, once the courtyard is covered, this thermal balance changes significantly: cold discomfort is substantially reduced compared with the outdoor reference, while heat discomfort increases.

Under BD conditions, overheating becomes the dominant component, particularly in the FLAT typology, where heat discomfort reaches maximum values of 3320 h in 2020 and 3972 h in 2080. By contrast, the SAW configuration shows lower sensitivity to overheating, with 1074 h of heat discomfort in 2020, a value even below the outdoor reference, although at the expense of a higher contribution of cold discomfort. The SD typology exhibits intermediate behaviour between these two extremes.

The introduction of EPD substantially modifies this thermal distribution. In all typologies, EPD increases the number of hours within the comfort range while significantly reducing heat discomfort. The highest comfort level is achieved in Granada–FLAT–EPD–2020, with 2944 h (33.6% of the year), representing an increase of 1141 h compared with the corresponding BD configuration. At the same time, heat discomfort in this typology decreases from 3320 h to 2190 h.

In the SAW typology, the EPD strategy reduces overheating to 698 h in 2020, a value lower than the outdoor reference (1128 h), highlighting the ability of this geometry, combined with advanced passive strategies, to limit excessive solar gains. Nevertheless, progressive climate warming partially reduces these benefits: under the 2080 scenario, heat discomfort in SAW–EPD increases to 1665 h, while cold discomfort continues to decrease.

3.2.2. Leon (Csb)

In León, the thermal pattern differs substantially from Granada. The outdoor reference is dominated by cold discomfort, reaching 7424 h in 2020, with almost no heat discomfort. In this colder climate, the courtyard enclosure acts as a significant thermal buffer, considerably reducing cold discomfort severity and increasing the number of comfort hours.

Under BD conditions, comfort levels in 2020 range from 1882 h in SAW to 2094 h in FLAT, while cold discomfort remains the dominant component, exceeding 5800 h in all cases. As the climatic horizon progresses, climate warming gradually reduces cold discomfort and increases comfort hours. For instance, in BD–SAW–2080, comfort reaches 2568 h, representing a significant increase compared with 2020.

The implementation of EPD leads to a further improvement in thermal conditions. In León–FLAT–EPD–2020, comfort reaches 3082 h (35.2% of the year), corresponding to an increase of 988 h compared with BD. This positive effect persists under future scenarios, although the gradual increase in summer temperatures leads to a moderate rise in heat discomfort.

Despite this increase, heat discomfort remains comparatively limited in León. The maximum value occurs in BD–FLAT–2080, with 2652 h of heat discomfort, while EPD–SAW–2020 records 0 h of heat discomfort, although with a high contribution of cold discomfort.

3.2.3. Overall Interpretation

The comparison with the outdoor climatic reference allows the role of courtyard enclosure to be interpreted more clearly. In both climates, the covered configurations substantially reduce cold discomfort hours relative to outdoor conditions, confirming the thermal buffering effect of the glazed enclosure. However, this moderation introduces a trade-off between cold and heat discomfort: the reduction in winter thermal penalties is accompanied by an increased risk of overheating, particularly in warm climates such as Granada.

The advanced passive strategies included in the EPD configuration mitigate this effect by redistributing hours from discomfort conditions—mainly overheating—towards the comfort range. Among the analysed typologies, the SAW geometry shows the most balanced behaviour under climate warming, limiting direct solar gains and maintaining a comparatively stable thermal response across scenarios.

3.3. Free-Running Conditions: Discomfort Severity (Degree-Hours)

Figure 4 shows the severity of thermal discomfort expressed as accumulated degree-hours (°C·h), associated with both overheating and underheating, for the same set of combinations of climate, design configuration, temporal scenario, and roof typology previously analysed. In addition, the figure includes an outdoor climatic reference (EXT) derived from the dry-bulb temperature of the climatic files and classified using the same comfort thresholds. Although this reference does not reproduce the specific microclimatic behaviour of an open courtyard, it provides a climatic benchmark that allows the thermal moderation effect introduced by the covered courtyard configurations to be interpreted relative to outdoor conditions.

Unlike analyses based exclusively on the number of hours outside the comfort range, the degree-hour indicator explicitly incorporates the magnitude of the thermal deviation relative to the defined limits (20–25 °C), allowing the quantification not only of how long discomfort occurs, but also of how severe it is. In this way, two cases with a similar number of hours outside comfort may exhibit very different thermal impacts depending on the intensity of the overheating or underheating experienced.

The use of degree-hours therefore constitutes a complementary and more sensitive indicator than hourly counts, particularly suitable for evaluating thermal behaviour under free-running conditions and for identifying configurations with high thermal peaks, even when their duration is limited. In the context of historic covered courtyards, this aspect is especially relevant, as certain typologies may concentrate short but intense episodes of overheating associated with direct solar gains.

The degree-hour results confirm and reinforce the trends observed in the analysis of comfort and discomfort hours, while providing an additional key insight into the thermal robustness of each solution. In particular, the comparison with the outdoor climatic reference makes it possible to identify the thermal buffering effect introduced by courtyard enclosure.

Accordingly, the combined analysis of hours and degree-hours provides a more comprehensive basis for comparison among roof typologies and passive configurations, enabling the evaluation of not only the frequency but also the severity of thermal discomfort, and reinforcing the interpretation of the results from a climate resilience perspective.

In Granada, under BD conditions, heat-related degree-hours increase markedly with the climatic horizon for the most solar-exposed typologies (SD and FLAT): in SD, values rise from 13,919.75 °C·h in 2020 to 20,091.03 °C·h in 2080, while in FLAT they increase from 16,677.60 °C·h to 23,709.78 °C·h over the same period. At the same time, cold-related degree-hours decrease, indicating a progressive shift in the thermal balance towards overheating as the climatic horizon advances. By contrast, the SAW typology consistently exhibits substantially lower heat-related degree-hours (for example, 1866.69 °C·h in 2020), albeit at the expense of high cold-related degree-hours in current and mid-century scenarios (on the order of 30,000–32,000 °C·h), in agreement with the hourly analysis showing a predominance of cold discomfort for this configuration.

In León, the behaviour is clearly dominated by cold discomfort across all scenarios, particularly under the BD configuration, although a progressive evolution with the climatic horizon is observed. In BD–2020, cold-related degree-hours reach very high values, on the order of 30,000 °C·h (e.g., SD: 30,363.86 °C·h; FLAT: 26,186.70 °C·h; SAW: 25,012.98 °C·h), while the heat-related component is comparatively low (SD: 5804.73 °C·h; FLAT: 6302.35 °C·h; SAW: 1470.95 °C·h). As the climatic horizon advances towards 2080, cold-related degree-hours decrease consistently (e.g., FLAT: 26,186.70; SAW: 18,067.05 °C·h), while heat-related degree-hours increase appreciably (FLAT: 6302.35 °C·h; SAW: 11,764.22 °C·h), evidencing a gradual shift in the thermal balance towards warmer conditions, although without reversing the predominance of cold discomfort. The SAW typology systematically maintains the lowest levels of heat-related discomfort (e.g., 0.00 °C·h in 2020) but presents the highest levels of cold-related discomfort, especially under current and mid-century scenarios, which is consistent with the hourly analysis identifying this typology as the most penalised in cold climates. Overall, the degree-hour results confirm that, in León, future climate warming reduces the severity of winter discomfort and increases summer loads, although the thermal regime remains predominantly conditioned by cold, in contrast to the behaviour observed in Granada.

For the EPD configuration in Granada–FLAT, heat-related degree-hours decrease from 16,677.60 to 5456.00 °C·h in 2020 (−67%), from 21,336.52 to 8209.81 °C·h in 2050 (−62%), and from 23,709.78 to 9773.61 °C·h in 2080 (−59%). A consistent reduction in heat-related degree-hours is also observed in SD (e.g., from 13,919.75 to 10,746.42 °C·h in 2020), while in SAW the effect is more limited, reflecting the inherently lower absolute overheating load of this typology.

The EPD configuration in León reveals a pattern predominantly dominated by cold-related degree-hours and systematically attenuates overheating severity, albeit with a smaller magnitude than that observed in the warm climate, consistent with the predominance of cold discomfort. The most pronounced effect is again identified in the FLAT typology (León–FLAT), where heat-related degree-hours are reduced from 6302.35 to 0.00 °C·h in 2020 (−100%), from 8991.91 to 6029.91 °C·h in 2050 (−33%), and from 11,764.22 to 7524.35 °C·h in 2080 (−36%). In the SD typology, the reduction in the heat-related component is likewise appreciable (for example, from 5804.73 to 3919.21 °C·h in 2020), while in SAW the impact of EPD is necessarily limited, given that this typology exhibits very low or null absolute overheating loads. Nevertheless, it is confirmed that, where heat discomfort occurs, EPD shifts its severity towards lower values while preserving the typological hierarchy identified in the hourly analysis. Overall, these results confirm that, even in a cold-dominated climate, the adoption of advanced passive strategies contributes to moderating episodes of excessive temperature. Furthermore, these results reinforce the interpretation derived from the hourly analysis: courtyard enclosure reduces exposure to outdoor cold conditions but introduces a trade-off associated with overheating risk, particularly in warm climates, which can be mitigated through advanced passive design strategies.

In the present study, the determination of the comfort temperature is based on the following applicable regulatory references. First, Spanish regulations are considered, specifically the RITE—Regulation of Thermal Installations in Buildings (Royal Decree 178/2021), which establishes an operative temperature range of 21–23 °C to define thermal well-being in winter conditions, setting 21 °C as the reference value for the sizing of heating systems. Second, European regulations are taken into account through EN 16798-1:2019 Energy performance of buildings—Part 1 [64], which defines thermal comfort requirements and establishes Category I (high indoor environmental quality) winter conditions as 21 °C ± 2.0. Finally, in the United States, ASHRAE Standard 55 [65] defines the winter comfort range with an operative temperature of 20–24 °C, recommending a temperature of 21 °C for heating load calculations in residential, office, and educational buildings, as reported in the “Indoor Design Conditions” tables of the ASHRAE Handbook—Fundamentals [66].

The static comfort criterion of 20–25 °C, applied on a 24/7 basis, represents a strict condition for buildings operating under free-running conditions; therefore, the percentage of comfort hours should be interpreted as a demanding indicator of the degree of thermal stabilisation. On average, BD achieves 22.48% of hours within the comfort range, while EPD increases this value to 29.17%, corresponding to a gain of +6.70 percentage points. Complementarily, discomfort severity, quantified through total degree-hours (cold + heat), is reduced from approximately 29,559 (BD) to approximately 22,526 (EPD), i.e., a reduction of −23.8%. The breakdown by sign indicates that EPD markedly reduces the heat-related component (−45.6%), while the reduction in the cold-related component is more moderate (−12.8%), consistent with strategies primarily aimed at limiting overheating (shading and natural ventilation when T_in > 24 °C).

The evolution under future climate conditions reveals divergent behaviours depending on the location. In Granada, the percentage of hours within the 20–25 °C range under BD remains practically unchanged between 2020 and 2080, whereas under EPD it shows a slight decrease towards 2080, consistent with the increase in high outdoor temperature episodes and the inherent limitations of a strategy based on conditionally activated natural ventilation. In León, by contrast, the percentage of comfort hours increases towards 2080 under both BD and EPD, reflecting the moderation of winter conditions and the reduction in severely cold hours.

3.4. Heating and Cooling Energy Demand Results

Figure 5 and Figure 6 show the heating and cooling energy demand for the different locations, according to the construction systems defined for each climatic scenario (2020, 2050, and 2080) and the corresponding design approach (BD and EPD).

The temporal evolution shows the expected pattern under climate warming: heating demand progressively decreases, while cooling demand increases. However, the translation of these trends into the total annual energy balance is location-dependent. In Granada, total demand under BD remains practically constant between 2020 and 2080 (182–184 kWh/m²·year), as a result of compensation between the reduction in heating demand and the increase in cooling demand. By contrast, in León a net reduction in total demand is observed (BD: from 209.40 to 193.10 kWh/m²·year from 2020 to 2080), despite the relative increase in cooling demand, due to the marked decrease in heating demand in a climate historically dominated by winter loads

The EPD configuration maintains a robust reduction in total energy demand across climatic horizons and locations. Averaging the three roof typologies, in Granada EPD reduces total demand to 152.56 kWh/m²·year in 2020 and stabilises it at around 149 kWh/m²·year in 2050 and 2080, while in León it decreases from 185.51 kWh/m²·year in 2020 to 166.59 kWh/m²·year in 2080

3.5. Analysis of Energy Demand by Geographical Location Under Future Climate Scenarios

Figure 7 illustrates the behaviour of energy demand by geographical location as a function of future climate scenarios (2020, 2050, and 2080) and the corresponding design approach (BD and EPD) for each construction solution.

Heating energy demand shows a clear dependence on geographical location, future climate scenario (2020, 2050, and 2080), and design configuration (BD vs. EPD) for all construction solutions (SD, FLAT, and SAW).

In Granada, BD reductions from 2020 to 2080 amount to −24.7% in SD (from 125.42 to 94.43 kWh/m²·year), −26.6% in FLAT (from 101.92 to 74.86 kWh/m²·year), and −26.8% in SAW (from 124.32 to 90.95 kWh/m²·year). In León, the decrease is likewise consistent, with reductions of −19.0% in SD (from 194.39 to 157.75 kWh/m²·year), −18.2% in FLAT (from 165.12 to 135.04 kWh/m²·year), and −21.1% in SAW (from 190.53 to 150.43 kWh/m²·year). In absolute terms, León consistently exhibits higher heating demand than Granada across all scenarios and design configurations, reflecting the dominant influence of local climatic conditions.

The comparison between BD and EPD within each typology evidences a general reduction in heating demand with EPD, with moderate and stable magnitudes in SD and FLAT, and nearly negligible effects in SAW. In Granada, EPD reduces heating demand in SD by between −8.6% (2020) and −9.3% (2080), decreasing from 125.42 to 114.61 kWh/m²·year and from 94.43 to 85.62 kWh/m²·year, respectively. In FLAT, reductions range between −7.5% and −8.7% (for example, from 101.92 to 94.27 kWh/m²·year in 2020 and from 74.86 to 68.38 kWh/m²·year in 2080). In SAW, the reduction is small but stable (approximately −0.6%), with very similar values between configurations (for example, from 124.32 to 123.52 kWh/m²·year in 2020 and from 90.95 to 90.36 kWh/m²·year in 2080). In León, a similar decrease is observed in SD and FLAT (approximately −7.6% to −8.6%, depending on the scenario), whereas in SAW the BD–EPD difference remains limited (−0.7%). The only notable exception occurs in León–2080–SD, where EPD increases heating demand from 94.43 to 143.90 kWh/m²·year; apart from this case, EPD exhibits consistent reductions.

For cooling demand (see Figure 8), it is confirmed that, in both climates, demand increases with the climatic horizon across all three typologies and for both construction configurations. In Granada (BD), the increase from 2020 to 2080 amounts to +41.3% in SD (from 89.75 to 126.85 kWh/m²·year), +42.5% in FLAT (from 93.84 to 133.70 kWh/m²·year), and +143.4% in SAW (from 12.49 to 30.40 kWh/m²·year, starting from low baseline values). In León (BD), the increase is likewise uniform and more pronounced in relative terms for SD and FLAT, +65.8% in SD (from 38.51 to 63.87 kWh/m²·year) and +71.0% in FLAT (from 39.16 to 66.95 kWh/m²·year), while SAW maintains very low values (from 0.50 to 5.61 kWh/m²·year). In absolute terms, Granada consistently exhibits higher cooling demand than León across all typologies and scenarios, and the typological ordering under BD remains stable: FLAT ≳ SD ≫ SAW.

The EPD configuration systematically reduces cooling demand relative to BD in all cases, with a particularly strong effect in the FLAT typology. In Granada, BD-to-EPD reductions in SD range between −17.4% (2020) and −13.3% (2080) (for example, from 89.75 to 74.15 kWh/m²·year in 2020 and from 126.85 to 109.99 kWh/m²·year in 2080), while in FLAT they reach −57.7% (2020), −52.8% (2050), and −50.9% (2080) (from 93.84 to 39.65 kWh/m²·year, from 119.43 to 56.43 kWh/m²·year, and from 133.70 to 65.65 kWh/m²·year, respectively). In SAW, the reduction is more moderate (approximately −5.7% to −8.5%, depending on the scenario; for example, in 2080 from 30.40 to 27.82 kWh/m²·year). In León, EPD also consistently reduces cooling demand in SD (−34.8% in 2020 and −20.6% in 2080, from 38.51 to 25.12 kWh/m²·year and from 63.87 to 50.74 kWh/m²·year) and, again, with greater intensity in FLAT (−71.8% in 2020 and −59.1% in 2080, from 39.16 to 11.04 kWh/m²·year and from 66.95 to 27.37 kWh/m²·year). In SAW, cooling demands remain close to zero (in 2080, from 5.61 to 4.93 kWh/m²·year).

Overall, it is observed that although future climate warming increases cooling demand across all typologies, the EPD configuration robustly mitigates these demands, achieving the greatest relative benefits in the FLAT typology and maintaining persistently low absolute values in SAW.

3.6. Analysis of Typological Performance Under Future Climate Scenarios

Depending on the construction solution, the results reveal significant differences (see Figure 9).

The FLAT typology (flat glass roof) exhibits the best relative performance of the EPD configuration. When comparing BD and EPD, total energy demand decreases systematically: in Granada by −31.6% (2020), −34.5% (2050), and −35.7% (2080); and in León by −19.9%, −23.6%, and −25.3%, respectively. This reduction is explained by a very pronounced decrease in cooling demand (up to −71.8% in León–2020), accompanied by moderate reductions in heating demand (approximately −7.5% to −8.7%). Under free-running conditions, FLAT shows the largest increase in hours within the 20–25 °C comfort range: between +9.3 and +13.0 percentage points in Granada and approximately +11.1 percentage points in León, together with notable reductions in total degree-hours (up to −46.5% in Granada). In aggregate terms, FLAT is the typology in which the EPD package delivers the greatest simultaneous benefit in both energy performance and thermal comfort.

The SD typology (glazed semi-dome) shows a consistent reduction in energy demand, but a high sensitivity to local shading control of openings and to the balance of solar gains. Total demand is reduced by approximately −11.6% to −12.7% across all scenarios and climates, with heating and cooling reductions generally on the order of −8% to −9% and −13% to −35%, respectively. Under free-running conditions, comfort increases by approximately 4–6 percentage points, and total degree-hours decrease by approximately −13% to −15%. However, a relevant trade-off emerges in León–2080–SD, where EPD markedly increases heating demand (+52.4%) while reducing cooling demand (−60.0%), while still maintaining an overall reduction in total demand (−12.0%). This result indicates that, under certain climatic conditions and with a high glazed fraction at roof level, an indoor-temperature threshold control (T_in ≥ 24 °C, 24/7) may limit useful solar gains during cold or mid-season periods, increasing heating loads, even though the annual balance remains favourable due to the strong reduction in cooling demand.

The SAW typology (south-facing sawtooth roof with vertical glazing) presents the smallest margin for energy improvement with EPD, particularly in León, where cooling demand under BD is already very low. Reductions in total demand range from approximately −1% to −3%, depending on the scenario and climate. However, in Granada, EPD provides clear improvements in passive thermal performance, with comfort increases of up to +11.8 percentage points and reductions in total degree-hours of up to −52.6% (Granada–2050–SAW). In León, the behaviour is more strongly conditioned by the predominance of cold discomfort under the 20–25 °C criterion. Indeed, in León–2020–SAW, EPD eliminates heat discomfort (100% reduction in heat-related hours) but slightly reduces the overall percentage of hours within the comfort range (−1.95 percentage points), suggesting a redistribution of time outside comfort towards the cold side. Consequently, in SAW the contribution of EPD manifests more clearly in the reduction in heat-related discomfort severity than in annual energy savings.

Overall, for the three analysed solutions, annual heating demand under BD decreases systematically as future climate scenarios progress (2020, 2050, and 2080) across all typologies and in both climates, maintaining a consistent pattern across the four series (G–BD, G–EPD, L–BD, and L–EPD). In Granada (BD), reductions from 2020 to 2080 amount to −24.7% in SD (from 125.42 to 94.43 kWh/m²·year), −26.6% in FLAT (from 101.92 to 74.86 kWh/m²·year), and −26.8% in SAW (from 124.32 to 90.95 kWh/m²·year). In León (BD), the same behaviour is observed, with decreases of −19.0% in SD (from 194.39 to 157.75 kWh/m²·year), −18.2% in FLAT (from 165.12 to 135.04 kWh/m²·year), and −21.1% in SAW (from 190.53 to 150.43 kWh/m²·year). In absolute terms, León exhibits higher heating demand than Granada across all scenarios and typologies, reflecting the greater influence of the cold season at this location.

Annual cooling demand results (kWh/m²·year) show a consistent increase with the climatic horizon (2020, 2050, and 2080) across the three roof typologies and for both locations, while also maintaining a stable pattern across the four represented series (G–BD, G–EPD, L–BD, and L–EPD). In Granada, cooling demand is high in the SD and FLAT typologies, with increases from 2020 to 2080 from 89.75 to 126.85 kWh/m²·year (+41.3%) in G–BD–SD and from 93.84 to 133.70 kWh/m²·year (+42.5%) in G–BD–FLAT; the SAW typology starts from lower values but grows with higher relative intensity (from 12.49 to 30.40 kWh/m²·year, +143.4%) in G–BD–SAW. In León, although absolute values in SD and FLAT are lower than in Granada, a progressive increase is also observed: L–BD–SD rises from 38.51 to 63.87 kWh/m²·year (+65.8%) and L–BD–FLAT from 39.16 to 66.95 kWh/m²·year (+71.0%); in L–BD–SAW, cooling demand is practically residual in 2020 (0.50 kWh/m²·year) and increases to 5.61 kWh/m²·year by 2080.

3.7. Variation in Energy Demand by Typology Under Future Climate Scenarios: Climate Resilience and Robustness Indicators

The quantification of relative energy savings constitutes a fundamental indicator for evaluating the performance of the enhanced passive design compared to the reference scenario, as it enables a normalised comparison across climates, construction typologies, and temporal horizons under both current and future climatic conditions.

Table 4. Absolute and relative changes in annual heating and cooling demand under future climate scenarios (2050 and 2080) relative to the current climate (2020). The background colours (yellow and blue) are the same as in Figure 9.

Clim.	CONF	Typol.	Heating D. 2020	Heating Demand 2050		Heating Demand 2080		Cooling D. 2020	Cooling Demand 2050		Cooling Demand 2080
			kWh/m2·yr	kWh/m2·yr	Δ%	kWh/m2·yr	Δ%	kWh/m2·yr	kWh/m2·yr	Δ%	kWh/m2·yr	Δ%
		SD	125.42	−20.98	−16.73	−30.99	−24.71	89.75	23.62	26.32	37.1	41.34
	BD	FLAT	101.92	−18.13	−17.79	−27.06	−26.55	93.84	25.59	27.27	39.86	42.48
Granada		SAW	124.32	−23.2	−18.66	−33.37	−26.84	12.49	11.06	88.55	17.91	143.39
(Csa)		SD	114.61	−19.74	−17.22	−28.98	−25.29	74.15	23.05	31.09	35.84	48.33
	EPD	FLAT	94.27	−17.62	−18.69	−25.89	−27.46	39.65	16.78	42.32	26	65.57
		SAW	123.52	−23.02	−18.64	−33.17	−26.85	11.49	10.72	93.3	16.33	142.12
		SD	194.39	−24.57	−12.64	−36.64	−18.85	38.51	15.94	41.39	25.36	65.85
	BD	FLAT	165.12	−20.18	−12.22	−30.08	−18.22	39.16	17.46	44.59	27.79	70.97
León		SAW	190.53	−27.8	−14.59	−40.11	−21.05	0.5	2.76	552	5.11	1022
(Csb)		SD	178.3	−23.32	−13.08	−34.39	−19.29	25.12	16.25	64.69	25.62	101.99
	EPD	FLAT	152.57	−20	−13.11	−29.08	−19.06	11.04	10.3	93.3	16.33	147.92
		SAW	189.12	−27.59	−14.59	−39.77	−21.03	0.38	2.43	639.47	4.55	1197.37

Note: For configurations with near-zero baseline cooling demand (e.g., SAW in León–2020), relative percentage changes should be interpreted with caution, as small absolute increases may translate into large relative variations.

presents the annual energy savings achieved with the enhanced passive design (EPD) relative to the baseline design (BD) for the studied climatic locations, according to roof typology and future climate scenarios (2020, 2050, and 2080).

In Granada, the enhanced passive design (EPD) leads to consistent reductions in heating demand across all analysed typologies and scenarios, with values ranging approximately between −17% and −19% in 2050 and between −25% and −27% in 2080 relative to the current climate. Regarding cooling demand, EPD delivers substantial savings, particularly in the FLAT typology, where reductions reach −57.7% under current climate conditions and increase to values on the order of −42% to −66% under future scenarios. As a result, total annual energy demand (calculated as the sum of heating and cooling) is significantly reduced in Granada, reaching a maximum reduction close to −35% in FLAT–2080, while reductions for the SD typology remain around −12–13% and remain limited (approximately −2–3%) in the SAW typology, where the impact of EPD is marginal.

In León, where heating dominates the annual energy balance, EPD produces moderate but stable reductions in heating demand, on the order of −12–15% in 2050 and −18–21% in 2080 for the SD and FLAT typologies. Although cooling demand under current climate conditions is practically negligible, future scenarios show a progressive emergence of cooling demand, with moderate absolute increases but high relative increases, particularly in the FLAT typology. In terms of total annual energy demand, EPD enables reductions ranging between −20% and −25% for FLAT, around −11–12% for SD, and below −2% for SAW. A specific exception is observed in León–2080–SD, where EPD leads to an increase in heating demand (+8.8%) associated with a reduction in useful solar gains; nevertheless, total annual energy demand continues to decrease due to the simultaneous reduction in cooling demand.

4. Discussion

The demand results obtained reveal a progressive increase in heat-related discomfort in the warm climate (Granada) and a partial attenuation of cold-related discomfort in the cold climate (León). This pattern is fully consistent with the projections of the IPCC AR6, which anticipate an increased risk of overheating in southern Europe and a moderation of winters in continental climates [67,68]. In Granada, the stability of total demand under BD reflects a compensation effect between heating and cooling uses, which is typical of warm Mediterranean climates [69]. In León, by contrast, the reduction in heating demand continues to dominate the annual balance, even in the presence of a relative increase in cooling demand, as reported by recent European studies [70,71,72]. The introduction of EPD consolidates significant and stable reductions in total energy demand at both locations, confirming the capacity of advanced passive design to mitigate the energy impact of climate change, in agreement with previous research on optimised building envelopes and controlled natural ventilation [73,74]. The BD–EPD comparison confirms that enhanced passive strategies significantly increase annual comfort hours, with relative increases exceeding 60% in Granada and 50% in León in selected cases, in line with recent studies highlighting the critical role of the building envelope and roof design in future thermal resilience [75,76]. Nevertheless, the persistence of high levels of discomfort in Granada towards 2080, even under EPD conditions, highlights the limitations of stand-alone passive solutions and reinforces the need for integrated approaches combining solar control, ventilation, adaptive comfort, and, where necessary, low-energy cooling systems.

The comparison with the outdoor climatic reference introduced in Figure 3 and Figure 4 provides additional insight into the thermal behaviour of covered courtyards. While courtyard enclosure substantially reduces exposure to cold outdoor conditions, particularly in the colder climate of León, it simultaneously introduces a trade-off associated with increased overheating risk, especially in warm climates such as Granada. This behaviour reflects the thermal buffering effect generated by the glazed enclosure, which moderates external temperature fluctuations but may also favour heat accumulation under high solar gains. These results highlight the importance of combining courtyard enclosure with appropriate passive strategies capable of limiting summer overheating while preserving the winter thermal benefits of courtyard enclosure.

The limitations discussed, based on the combined analysis of comfort hours and discomfort severity (degree-hours), confirm that the impact of climate change on buildings operating under free-running conditions is strongly dependent on local climate and on the effectiveness of passive strategies in modulating overheating. In Granada, the shift in the thermal balance towards heat-related discomfort and the limited improvement in comfort under future scenarios are consistent with studies warning about the growing insufficiency of natural ventilation as a stand-alone adaptive strategy in the face of more frequent and intense heatwaves [77,78,79]. By contrast, in León, the reduction in winter discomfort and the increase in comfort hours towards 2080 are in agreement with studies identifying net thermal benefits of warming in cold regions, provided that summer overheating remains controlled [80,81]. Within this framework, EPD significantly reduces discomfort severity, particularly the heat-related component, confirming the effectiveness of shading and controlled ventilation strategies described in previous studies [82]. The use of a strict static comfort criterion (20–25 °C, 24/7) should nevertheless be interpreted as a demanding indicator of thermal stability, given that adaptive comfort approaches allow for wider acceptability ranges (EN 16798-1; ASHRAE 55).

The evolution of energy demand across the climatic horizon (from 2020 to 2050 to 2080) highlights an appropriate typological response, in which climate acts as the primary driver, while roof typology modulates the magnitude of the response without substantially altering the relative hierarchy among solutions. Regarding the typological comparison within each climate, FLAT tends to exhibit the lowest heating demand in both Granada and León (under both BD and EPD conditions), whereas SD and SAW present higher and closely comparable levels. This hierarchy is maintained across the three temporal horizons, indicating that the effect of future climate manifests primarily as a global reduction in heating requirements, without substantially altering the relative ordering among typologies, except for the specific case identified for León–2080–SD under EPD.

The systematic reduction in heating demand and the concurrent increase in cooling demand reinforce growing summer vulnerability, particularly in roof typologies with high solar gain (SD and FLAT), as reported by recent studies on overheating risk [83,84]. Under the EPD configuration, the FLAT typology consistently demonstrates the greatest capacity to simultaneously mitigate both heating and cooling demand, suggesting that climate change amplifies pre-existing typological trends rather than introducing qualitatively new behaviours. However, the SAW typology exhibits greater passive thermal stability under free-running conditions, with lower overheating intensity and a more balanced distribution of discomfort hours. This contrast highlights the difference between overall energy optimisation and passive thermal robustness in courtyard roof design. Nevertheless, the trade-offs observed in highly glazed typologies in cold climates underline the need for climate-specific adaptive strategies, integrated with advanced thermal management criteria, in order to ensure future energy performance [85].

The use of relative energy savings (%) as the primary indicator is consistent with the recent literature based on dynamic simulations and future climate scenarios, where this approach is systematically employed to compare the effectiveness of passive envelope strategies and solar gain control measures against a baseline case, minimising the influence of initial conditions and strengthening the robustness of the results [86,87]. The findings indicate that EPD is more effective in typologies with higher solar exposure, particularly in flat roofs, and that its energy-saving potential increases under warmer future climates, mainly through the mitigation of cooling demand, while maintaining a stable hierarchy among roof typologies [75].

From a broader discussion perspective, the observed patterns confirm that the effectiveness of the enhanced passive design is primarily governed by the interaction between solar exposure, roof typology, and climatic regime, and that this interaction becomes increasingly relevant under future warming scenarios. The superior performance of typologies with high solar exposure combined with effective solar control is consistent with previous studies identifying the building envelope as a key element for decoupling the increase in cooling demand associated with climate change, without significantly penalising the annual energy balance [87]. In cold climates, the emergence of trade-offs between the reduction in summer solar gains and their utilisation during winter reinforces the need to assess passive strategies using relative indicators capable of capturing opposing effects in heating and cooling demand within a robust comparative framework. In this regard, the results support the use of relative energy savings as the primary metric for discussing the resilience and transferability of passive envelope solutions under future climate scenarios [74,85], in line with established practice in the high-impact literature.

From a climatic perspective, the two case study locations were selected to represent contrasting thermal contexts within the Köppen classification framework, allowing the assessment of typological responses under both warm and cooler Mediterranean conditions. Granada corresponds to a Csb Köppen–Geiger climate (warm-summer Mediterranean), widely distributed across southern Europe, including southern France, northern and central Italy, the Adriatic coastal belt and western Greece [33,45]. León represents a cooler inland Mediterranean transition climate influenced by continental conditions, characterised by colder winters and higher heating demand while maintaining Mediterranean precipitation seasonality. Similar climatic conditions occur in elevated Mediterranean interiors across the Iberian Peninsula, southern France, parts of Italy and the western Balkans. Within this framework, the Köppen–Geiger classification provides a robust basis for identifying climatically comparable regions and supporting the transferability of building performance analyses [33,88]. Although derived from two Spanish case studies, the observed behavioural patterns, particularly the stability of typological hierarchies and the increasing relevance of solar control strategies under warmer scenarios, may therefore be interpreted as indicative trends for comparable Mediterranean climates. Nevertheless, the limited number of analysed locations remains a methodological constraint, and further research incorporating additional Mediterranean and semi-arid subclimates (e.g., southern Italy, Greece, North Africa and the eastern Mediterranean) would help to further validate the broader applicability of the proposed design strategies for historic covered courtyards.

5. Conclusions

The main conclusions of the study are as follows:

Heating demand decreases progressively from 2020 to 2050 and 2080, indicating a systematic reduction associated with future climate warming.

Heating demand in the colder climate of León is consistently higher than in the warmer climate of Granada for all typologies and scenarios, confirming the greater winter severity of the location and the higher relative contribution of heating to the annual energy balance. Nevertheless, in both locations, heating demand decreases progressively from 2020 to 2050 and 2080, reflecting the systematic effect of future climate warming. In parallel, the enhanced passive design (EPD) results in lower heating demand than the baseline design (BD) across all scenarios and construction solutions, confirming the effectiveness of the design improvement under both current and future climates. The absolute impact of EPD is more pronounced in León due to its higher initial demand levels.

RQ1: Roof geometry is confirmed as a dominant factor in the thermal robustness of covered courtyards. Typologies with asymmetric configurations or inherent self-shading capacity, such as the south-facing sawtooth roof (SAW), exhibit more stable thermal behaviour under free-running conditions, whereas glazed dome and flat glazed roof solutions show greater sensitivity to climatic severity and solar gains.

Regarding construction solutions, the FLAT configuration delivers the largest absolute and relative energy benefits by combining very strong reductions in overheating with moderate decreases in heating demand.

RQ2: Enhanced passive design strategies significantly mitigate overheating risk and cooling demand, particularly in geometries that are intrinsically more vulnerable. The effectiveness of these measures depends on roof typology, with the flat glazed roof (FLAT) showing the greatest benefits from the incorporation of solar control, improved glazing, and controlled natural ventilation.

This is followed by SD and SAW, a ranking that remains stable across all locations, climate scenarios, and design configurations, with these typologies maintaining higher and relatively similar demand levels and preserving their hierarchy under future scenarios. While climate and location determine the absolute magnitude of heating demand, the response of construction solutions to the implementation of EPD differs substantially among typologies, both in terms of energy demand and comfort. This confirms that roof geometry and upper-envelope configuration strongly condition the relative weight of solar gains, heat losses, and dissipation capacity. Moreover, EPD strategies provide a consistent mitigation effect regardless of the temporal horizon considered. This stability suggests that, under SSP4.5, the measures implemented in EPD are effective in containing the growth of cooling loads without introducing systematic increases in heating demand across the overall set of cases, albeit with typology-specific exceptions.

RQ3: Climate change scenarios intensify the thermal penalties associated with unfavourable geometries, increasing cooling demand and discomfort severity, particularly in warm climates. Nevertheless, typologies with higher intrinsic robustness maintain comparatively more stable performance over time, highlighting the key role of geometric configuration in long-term climate resilience. In cold Mediterranean climates, the application of passive strategies may introduce seasonal trade-offs between summer overheating mitigation and winter solar gains. In specific cases, this may result in moderate increases in heating demand; however, the overall annual energy balance remains favourable due to the simultaneous reduction in cooling demand.

The use of degree-hours as an intermediate variable facilitates the identification of trends and establishes a direct link between thermal comfort and energy consumption, demonstrating that EPD strategies not only improve comfort but also generate quantifiable and comparable energy reductions across climates, typologies, and future scenarios. By contrast, BD amplifies the impact of climate change on energy consumption in warm climates, while in cold climates future warming partially reduces total demand. In addition, location remains the dominant factor in absolute energy consumption, but roof typology decisively modulates the magnitude of achievable savings through passive strategies.

The comparison with outdoor climatic conditions highlights the thermal buffering capacity generated by courtyard enclosure. Although this configuration substantially reduces cold-related discomfort, it may simultaneously increase the risk of overheating, particularly in warm climates. These results underline the need for careful design of enclosure systems, where roof geometry and passive control strategies play a crucial role in moderating solar gains while preserving the thermal stability that characterises covered courtyard spaces.

Overall, the results highlight the relevance of courtyard enclosure design as a key factor influencing indoor environmental conditions in heritage buildings located in Mediterranean climates. By demonstrating the impact of roof geometry and enhanced passive design (EPD) strategies on thermal behaviour, the study contributes to a better understanding of how covered courtyards can balance thermal buffering benefits with the control of solar gains. These findings provide a useful basis for guiding future retrofit interventions in historic courtyard buildings and support the development of design approaches that reconcile heritage conservation with improved environmental performance, although further studies in semi-arid contexts are required to confirm the broader transferability of these findings.

From an applied perspective, the results allow the definition of a simplified rule-based framework for activating enhanced passive design (EPD) strategies in covered courtyards. As illustrated in Figure 10, the proposed framework translates the simulation results into a practical decision-support model based on continuous indoor–outdoor monitoring and the activation of passive measures according to thermal conditions and solar radiation levels. During warm periods, solar control devices are prioritised when outdoor temperatures exceed approximately 24–26 °C under high solar radiation in order to reduce overheating risk, particularly in configurations with large glazed surfaces. Complementarily, night-time ventilation is activated when indoor temperatures exceed outdoor conditions, enabling passive heat dissipation. Under cooler conditions typical of winter periods, ventilation rates are reduced to preserve useful solar gains. The framework also incorporates roof-typology adjustments, indicating stronger shading and ventilation strategies for flat glazed roofs (FLAT) and lower shading requirements for sawtooth roofs (SAW) due to their inherent self-shading capacity. Overall, this rule-based approach translates the simulation findings into an operational guideline that supports climate-responsive management and energy-efficient rehabilitation of historic covered courtyards.

The findings provide a practical basis for the design and retrofitting of covered courtyard systems, promoting climate-responsive passive strategies that enhance thermal performance while preserving the architectural character of historic buildings.