Porous cellular structures such as TPMS and ADMS can be tailored to control conductive, convective, and radiative heat transfer (Huang et al., 2024). Their high surface-area-to-volume ratios and tailored pore sizes allow for targeted thermal behaviour: smaller pores (<3 mm) restrict airflow and increase insulation. In comparaison, larger pores (>5 mm) promote convective exchange (Xuan et al., 2024). In previous field trials, structures of this kind tested as surrounding geometries for nesting tubes achieved modest reductions in peak internal-tube temperature T_nest compared to conventional nesting aids, yet they still largely mirrored ambient air temperature T_aᵢ_r trends. This outcome suggests that purely passive geometric modulation cannot fully decouple internal cavity temperatures from external thermal loads.

Assuming an optimized geometry for minimized heat transfer, the effective heat capacity of the structure governs this temporal delay, rather than preventing heat accumulation altogether. To overcome this limitation, an additional heat sink is required, for instance, adiabatic temperature reduction through evaporative cooling.

Evaporative cooling achieves this through the phase change of water from liquid to vapor. This process uses sensible heat, converting it into latent heat to break intermolecular forces, thereby lowering local temperatures while increasing humidity (Rupp and Gruber, 2021; Dutto et al., 2022; ASHRAE, 2024). Geometry plays a decisive role in modulating this phenomenon: it controls hydraulic behaviour, thermal mass distribution, and airflow interaction at wetted interfaces. In systems employing discrete droplet retention rather than continuous water films, irregular or graded morphologies can still enhance evaporation by increasing exposed surface area and promoting airflow disruption around droplets, thereby accelerating local heat and mass transfer (Kumar Dhamneya et al., 2017; Rupp and Gruber, 2021; Dutto et al., 2022). In graded or packed structures, increasing the effective surface area remains one of the primary strategies to boost evaporative performance (Jain and Hindoliya, 2011).

In porous geometries, convection must be considered alongside evaporation. Pore sizes of approximately 5 mm or larger ensure that airflow remains active, both increasing mass transport and enabling turbulence that promotes heat exchange, accelerating the evaporative process. Irregular topologies can induce local flow mixing and vortex shedding, enhancing convective heat and moisture exchange. However, these effects are highly context-dependent and cannot be resolved analytically, as they depend on coupled Navier–Stokes and energy equations, and are therefore estimated through empirical correlations or numerical models. Consequently, the optimisation of conjugate heat transfer in complex porous systems typically requires prototyping and iterative empirical validation to accurately calibrate performance, particularly at architectural scales (Kumar Dhamneya et al., 2017).

Although honeybees (Apis mellifera) are eusocial and not the intended occupants of the nesting aids designed in this study, their brood-thermoregulation strategies offer transferable physical principles relevant to evaporative cooling. A. mellifera maintains brood temperatures within a narrow optimal range of approximately 34 °C–36 °C, even under elevated ambient conditions. A key mechanism is the deposition of water droplets inside the hive, which cools the brood through evaporative phase change, aided by coordinated fanning behaviours that increase airflow across wetted surfaces. (Siefert et al., 2021).

Effective cooling in the hive relies not on bulk water pooling but on the controlled retention and spatial distribution of droplets on micro-scale surface features. This strategy maximises evaporative efficiency by keeping droplets exposed to airflow while preventing uncontrolled runoff (Siefert et al., 2021). Abstracted to a design context, this principle highlights the importance of pore morphologies capable of establishing droplets against gravity while sustaining convective exchange (Figure 2).

Natural systems rarely optimise a single parameter; instead, they reconcile competing demands across multiple scales. Biological organisms integrate material, structure, and form, tuning their properties locally to balance efficiency, stability, and environmental control (Knippers and Speck, 2012). In the architectural analogue explored here, pores in a facade component are therefore required to perform multiple roles: enhancing evaporative cooling by enabling airflow and surface exchange, stabilising droplets to regulate the pace of evaporation, and providing intrinsic stiffness that allows the component to remain self-supporting without auxiliary framing. This multifunctional behaviour mirrors biological systems in which material organisation and geometry together generate both mechanical stability and environmental responsiveness.

The analogy is further supported at the level of physiological thresholds. A. mellifera, moderate brood damage occurs around nest temperature T_nest = 36 °C, with lethal effects above T_nest = 40 °C (Buchmann, 2021; Siefert et al., 2021). Solitary cavity-nesting species show similar sensitivities: optimal development typically occurs at ≈ T_nest = 20 °C–25 °C, with metabolic stress beginning above ≈ T_nest = 35 °C and mortality rising sharply beyond T_nest = 40 °C (Radmacher and Strohm, 2011; Kierat et al., 2017; Ferrari and Polidori, 2022; Polidori et al., 2023; Melone et al., 2024). While absolute tolerances vary among species, these shared thresholds justify the transfer of water-based cooling principles at the level of thermal limits, without implying ecological or behavioural equivalence.

The water-management strategies of A. mellifera illustrate how liquid can be precisely deployed through controlled retention and exposure, balancing evaporation efficiency with stability. Translated to design, this principle could be extrapolated to porous structures for facade geometries in which pore size, curvature, and surface texture jointly determine droplet behaviour (Figure 2). Accordingly, the experiments presented here do not assess biological performance, nesting success, or species-specific behavioural responses. Instead, they evaluate the physical effectiveness of evaporative cooling–abstracted from honeybees’ thermoregulatory strategies–as a potential mechanism for moderating temperatures within artificial nesting cavities. The biological role model is thus used instrumentally to inform geometric and material design decisions, while detailed ecological variability among solitary bee species remains beyond the scope of this study.

This stage examined how different porous geometries retain water, as cooling performance depends not only on thermodynamic potential but also on the structure’s capacity to capture and distribute droplets. TPMS diamond with uniform pore size, and ADMS samples were tested in uniform (Ø15 mm pores) and graded (Ø5–25 mm pores) configurations previously evaluated for heat-transfer behaviour (Figure 3).

All models measured 200 mm × 200 mm × 150 mm. For each test, 600 mL of water was evenly applied to the top surface through a perforated container to simulate droplet distribution. After a fixed drainage period, the expelled water was collected and measured to determine the retained volume. Each geometry was tested three times under same conditions (T_aᵢ_r ≈ 15 °C) with identical water application rates. Retention values ranged from 23% to 30%, with the graded TPMS diamond showing the highest average retention (29.4%). This configuration was chosen for subsequent evaporative-cooling experiments. These repeated measurements were intended to assess relative retention capacity rather than statistical variance across populations of samples.

Two identical TPMS diamond-graded models were prepared as nesting aid prototypes, each featuring a central Ø 5 mm nesting tube aligned along a 200 mm axis. Fabricated with a bio-based PLA-Wood composite (30% Wood), selected based on previous experiments showing wild bees accepted nesting in it (Figure 1). One specimen was designated the water-supplied (sWS) sample and received 600 mL of water daily at 09:00 to enable evaporative cooling. In contrast, the second specimen served as the control (sC) sample, kept dry without water input (Figure 4).

To provide a performance benchmark against conventional nesting aids, three additional control samples with identical external dimensions and tube configurations were prepared: a solid wood block, an air-dried clay block, and a bundle of reed stems housed in a timber frame. Each contained a central Ø5 mm nesting tube, matching the geometry and sensor placement of sWS and sC samples.

The benchmark tests were carried out in parallel with the sWS and sC samples, enabling two distinct analyses:

Direct evaporative cooling effect–comparison between the temperature of the nesting tube (T_nest) in sWS and sC samples.

The samples were monitored for 6 days under various weather conditions (T_aᵢ_r, RH_aᵢ_r) to understand the behaviour of the evaporative cooling effect. T_nest and RH_nestwere monitored inside each nesting tube using probes positioned at the mid-length point.

The pWS and pC full-scale Facade components tested the applicability of the bio-inspired evaporative-cooling concept under real heatwave conditions during summer 2025. Each panel measured 200 mm × 150 mm × 1100 mm and was produced with wood–PLA composite containing 30% wood dust. The geometry employed a TPMS diamond–graded morphology, featuring Ø5 mm pores around the nesting tubes. These were optimized through preliminary droplet-retention tests and expanded gradually to Ø35 mm toward the perimeter to balance cooling efficiency with airflow (Figure 5).

The panels were conceived as a facade-mounted shader system. Although the porous geometry can capture rainwater, reliable thermal buffering during extreme events requires a controlled water supply (Figure 6). For this purpose, a drip-feed irrigation system was integrated. During testing in August 2025, the experiment ran continuously for 7 days, coinciding with heatwave episodes. The system began operation at 05:00 each day to pre-emptively counter rising temperatures. Daily water delivery ranged from 900 mL to 2500 mL with staggered doses of ≈300 mL triggered when embedded sensors detected T_nest (pWS) > 35 °C (high-risk threshold for B_nest) (Radmacher and Strohm, 2010; Polidori et al., 2023; Melone et al., 2024).

The pWS and pC models were installed on a real-building setup, in the third-floor southeast-facing loggia of a residential building in central Stuttgart, Germany, at a height of 1 m above the floor surface. The loggia is partially enclosed by glass panels, creating a semi-outdoor microclimate that amplifies solar heat gain and limits natural ventilation. The real-building setting was intentionally selected to evaluate the prototypes under realistic facade exposure conditions, ensuring that the results reflect actual environmental dynamics, including natural variations in solar radiation, temperature, and humidity. To obtain extreme summer conditions, the glass panels remained closed throughout the test, creating a heat-trap environment (Figures 6, 7).

Temperature and humidity were recorded inside the prototype at the mid-length of a Ø5 mm nesting tube (T_nest, RH_nest). To establish ambient reference conditions, two additional measurements were taken in the real-building setup: one inside the loggia enclosure to represent the enclosed-air microclimate (Tlog, RHlog), and one outside the loggia using a shielded weather sensor to represent the outdoor air temperature (T_aᵢ_r, RH_aᵢ_r).

The dataset was processed to calculate the mean, maximum, and minimum internal temperatures for each specimen. Temporal cooling profiles were constructed to identify patterns of diurnal variation. The duration of effective cooling was defined as the time span during which the temperature of the nesting tube in the water-supplied sample (T_nest (pWS)) remained lower than that in the control sample (T_nest (pC)).

To assess whether the cooling behaviour observed in the real-building setup could be generalized to other facade configurations and boundary conditions, a controlled validation test was conducted in a Clima Temperatur Systeme (CTS) C-Series climatic chamber. While the real-building setup experiment captured the behaviour of the Facade components within a fluctuating semi-enclosed microclimate, the chamber setup provided a stationary reference, isolating the physical process of evaporation from external drivers such as wind and solar radiation. This setup enabled a mechanistic validation of the temperature differential (ΔT), defined as the difference in internal-tube temperature between the control panel (T_nest (pC)) and the water-supplied panel (T_nest (pWS)). This comparison quantifies the magnitude of the evaporative cooling effect observed during the real-building test.

The same pWS and pC panels previously installed on site were tested within a controlled test volume (2 450 mm × 2 400 mm × 2 050 mm), ensuring uniform air distribution through closed-loop regulation of chamber-air temperature (T_aᵢ_r(ch)) and humidity (RH_aᵢ_r(ch)). Each specimen contained a central Ø 5 mm nesting tube equipped with an embedded temperature and humidity sensor, while an additional probe recorded ambient chamber conditions to verify the stability of the programmed cycles (Figure 8).

Two experimental scenarios were implemented. The first scenario reproduced the air temperature (T_log) and relative humidity (RHl_og) recorded inside the loggia of the real-building setup, representing the enclosed-air microclimate surrounding the facade components during the hottest field day. During this period, the outdoor air temperature (T_aᵢ_r) exceeded 30 °C and peaked at 36.9 °C; using seven discrete setpoints. As short-wave solar radiation could not be replicated, the pC specimen did not reach the maximum temperatures observed outdoors. The second scenario instead reproduced the internal-tube temperature of the control panel (T_nest (pC)) measured during the real-building test, representing the thermal conditions directly experienced within the nesting tube. This sequence used nineteen discrete temperature setpoints above 40 °C, corresponding to the lethal threshold for the bee brood (B_nest).

At the start of each cycle, the pWS panel received an initial 900 mL water dose through its integrated irrigation inlet to trigger evaporation. Subsequent water supply was automatically triggered whenever T_nest (pWS) > 35 °C, maintaining continuous evaporative activity during peak heat conditions.

Over six consecutive test days, the mean temperature inside the nesting tube of the water-supplied sample (T_nest (sWS, mean)) was 26.8 °C, compared to 30.0 °C in the dry control sample (T_nest (sC)). Here, T_nest denotes the internal-tube temperature measured at the midpoint of each 5 mm nesting cavity, while sWS and sC refer respectively to the water-supplied and control small-scale models. Maximum values reached T_nest (sWS) = 34.3 °C and T_nest (sC) = 44.9 °C, yielding an average temperature differential (ΔT = T_nest (sC) – T_nest (sWS)) of 3.16 K. Water application at 09:00 produced an immediate drop in T_nest (sWS), with peak differentials between 10:30 and 14:00 (ΔT = −13.9 K relative to T_nest (sC); −17.8 K relative to outdoor air temperature T_aᵢ_r). The cooling effect persisted through the afternoon and gradually diminished after 22:00, when both samples converged overnight (Figure 9).

Throughout the test, T_nest (sWS) < 30 °C, below the stress threshold for B_nest, while T_nest (sC) exceeded 40 °C for up to 4.5 h day⁻¹. Thermal imaging showed that the surface of the water-supplied sample was approximately 9 K cooler than the surface of the control sample (Figure 10).

Maximum instantaneous reductions of ≈10 K, mean daily cooling ranged from 2.15 K to 3.87 K. Strong cooling effects (ΔT >5 K) persisted for 1.5–6.25 h per day. Moderate cooling (ΔT >2 K) prevailed for most of the diurnal cycle.

A comparative test evaluated the sWS and sC models against conventional nesting materials–clay, wood, and reed stems–under identical environmental conditions (outdoor air temperature T_aᵢ_r and relative humidity RH_aᵢ_r) (Table 1). Here, T_aᵢ_r and RH_aᵢ_r refer to the air temperature and relative humidity measured near the samples during outdoor testing.

Even without evaporative cooling, the control sample sC maintained mean internal-tube temperature T_nest (sC, mean) = 29.0 °C, which was approximately 2.3–3.8 K cooler than the benchmarks (T_nest, mean = 32.0 °C–32.8 °C). Here, T_nest denotes the temperature measured at the centre of the 5 mm nesting tube inside each sample. When supplied with water, the water-supplied sample sWS achieved a mean internal-tube temperature T_nest (sWS, mean) = 26.0 °C and a maximum T_nest (sWS, max) = 34.3 °C, corresponding to relative temperature decreases of 14.9 K compared with reed stems, 11.6 K compared with clay, and 5.6 K compared with the dry control sC (Table 1). The daily temperature amplitude (ΔT_day = T_nest (max) – T_nest (min)) was lowest for sWS (16.5 °C), indicating that latent heat exchange moderated diurnal temperature fluctuations more effectively than in the other materials.

Duration analysis displayed: ΔT > 5 K persisted 4.0–6.2 h day⁻¹, and ΔT > 2 K up to 20–24 h day⁻¹. The internal-tube temperature in T_nest (sWS) never exceeded 40 °C, remaining below the lethal limit for bee brood (B_nest), whereas traditional materials remained above this lethal threshold for 2.5–5 h day⁻¹.

The improved thermal regulation of sWS is attributed to its graded pore architecture. Pores (Ø5 mm) around the nesting tube stabilised droplets and prolonged evaporation, while the outer (Ø25 mm) pores enhanced convective airflow. This spatial variation contributed to a reduction in peak temperature, indicating that the combined effects of geometry, material composition, and latent heat exchange can outperform dense (clay, wood) and fibrous (reed) references (Figure 11).

Across the seven-day test period, the internal-tube temperature of the water-supplied panel (T_nest (pWS)) remained consistently lower than that of the dry control panel (T_nest (pC)), with a temperature differential (ΔT = T_nest (pC) – T_nest (pWS)) of 5–8 K during peak hours. Here, T_nest refers to the temperature recorded inside the 5 mm-diameter nesting tube embedded in each facade prototype. The largest reductions occurred within the first hours after irrigation, and both temperature profiles converged overnight (Figure 12).

During the test, T_nest (pC) exceeded 40 °C, whereas T_nest (pWS) remained below 35 °C in all trials. Infrared imaging revealed cooler surface regions on pWS—particularly around droplet-retention zones with smaller pores—where surface temperatures were up to 8 K lower than those of the control panel (Figure 13).

Mean internal-tube temperatures were T_nest (pWS, mean) = 25.0 °C and T_nest (pC, mean) = 29.8 °C, confirming the cooling effect. Ambient references recorded within the real-building setup (loggia) showed enclosed-air temperature T_log, mean = 30.8 °C and relative humidity RH_log, mean = 38%, while outdoor-air temperature was T_aᵢ_r, mean = 26.4 °C with relative humidity RH_aᵢ_r, mean = 47%. These values indicate a higher thermal load and lower humidity inside the semi-enclosed real-building setup compared with exterior conditions (Figure 14; Table 2).

At peak load (13 August 2025), the maximum internal-tube temperature T_nest (pC,max) = 44.9 °C, RH_nest (pC) = 23%, while T_nest (pWS) = 33.3 °C, RH_nest (pWS) = 71%, resulting in a measured ΔT (pC–pWS) = 11.6 K. When the pC state (44.9 °C, 23% RH) is plotted on the Mollier diagram, it corresponds to a wet-bulb temperature Twet ≈26.1 °C, defining a theoretical wet-bulb depression of ≈18.9 K. The difference between the theoretical and measured cooling was approximately 7.3 K) indicating that the system achieved about 62% of the total adiabatic potential.

This deviation can be attributed to the fact that part of the latent heat absorbed during evaporation was used not only to cool the internal nesting cavity but also to reduce the temperature of the surrounding air layer within and around the porous geometry. In natural analogues, such as the evaporative regulation observed in Apis mellifera colonies, a similar distribution of cooling occurs: collected water droplets serve to stabilize both brood-cell temperature and hive-air conditions through combined convective and evaporative fluxes (Buchmann, 2021). In the tested prototypes, this shared cooling effect explains the observed ΔT within the tube relative to the full adiabatic potential, indicating that evaporated water contributes simultaneously to the microclimate of the nesting cavity and to localized air-mass cooling around the panel surface.

During the hottest days, T_nest (pWS) < 35 °C for 3–10 h day⁻¹, depending on RH_aᵢ_r and solar intensity (Table 3). Cumulative thermal exposure showed that pWS remained above 30 °C for t (>30 °C) = 23.5 h, compared with 79.0 h for pC, and above 35 °C for only 2.5 h versus 45.3 h in pC.

Daily profiles (Figure 14) exhibited consistent ΔT (pC–pWS) peaks and recovery trajectories, indicating that cooling was primarily governed by the water-supply cycle and radiative load. The facade-mounted pWS panel reduced T_nest by up to ΔT_max = 15 K compared to pC, while maintaining T_nest (pWS) < 35 °C, within the thermally safe range for B_nest (Table 3).

Two climatic-chamber test phases were conducted to validate the evaporative-cooling mechanism observed during the real-building setup experiment under controlled and stationary conditions. Scenario 1 reproduced the loggia-air temperature (T_log) and relative humidity (RH_log) profiles recorded on site, while Scenario 2 replicated the internal-tube temperature of the control panel (T_nest (pC)) to examine the temperature-dependent response of the system.

In Scenario 1, chamber-air temperatures ranged from T_aᵢ_r(ch) = 29.9 °C–36.3 °C, with RH_aᵢ_r(ch) = 42–28%. The control specimen reached T_nest (pC) = 36.5 °C, RH_nest (pC) = 38%, whereas the water-supplied specimen peaked at T_nest (pWS) = 29.8 °C, RH_nest (pWS) = 76%, yielding a maximum temperature differential ΔT (pC–pWS) = 7 K.

To estimate the theoretical cooling limit, the psychrometric state corresponding to the control-panel condition (T_nest (pC) = 36.5 °C, RH_nest (pC) = 39%) was plotted on the Mollier diagram. The corresponding wet-bulb temperature (Twet) ≈ 24.6 °C, defining a theoretical wet-bulb depression ≈11.9 K. With the measured temperature differential ΔT (pC–pWS) = 7 K, the system achieved approximately 59% of the adiabatic cooling potential.

In Scenario 2, chamber-air temperatures ranged from T_aᵢ_r(ch) = 40.5 °C–44.9 °C and RH_aᵢ_r(ch) = 27–21%. The control panel recorded T_nest (pC) = 40.5 °C–44.9 °C and RH_nest (pC) = 30–39%, while the water-supplied panel maintained T_nest (pWS) = 31.7 °C–35.0 °C, yielding a mean temperature differential ΔT (pC–pWS, mean) = 10.2 K.

At the peak condition (T_nest (pC) = 44.9 °C, RH_nest (pC) = 32%), the psychrometric state corresponded to a wet-bulb temperature Twet ≈29.1 °C, defining a theoretical wet-bulb depression ≈15.8 K. The water-supplied panel measured T_nest (pWS) = 31.7 °C and RH_nest (pWS) = 84%, yielding ΔT (pC–pWS) = 13.2 K, approximately 84% of the theoretical adiabatic potential. When plotted on the Mollier diagram, the pC and pWS system states align closely along the adiabatic line; shifting from the pC to the pWS state follows the direction of adiabatic cooling, confirming evaporation under near-adiabatic conditions. Across both scenarios, T_nest (pWS) < 40 °C, remaining below the biological-stress threshold associated with increased larval mortality in B_nest, whereas the C panel exceeded this limit under high-temperature conditions (Table 4).

The climatic-chamber results provide a controlled verification of the evaporative-cooling process under stable conditions, thereby confirming adiabatic cooling. Unlike the outdoor tests, where wind, radiation, and humidity fluctuated, the chamber setup isolated the direct thermal effect of evaporation. A comparison between both settings indicates similar cooling behaviour but differing efficiencies. In the real-building setup, the mean temperature differential between pC and pWS was ΔT_mean = 4.8 K, with a peak of 11.6 K, corresponding to about 72% of the adiabatic cooling potential. In the chamber, Scenario 1 reached ≈60–65% efficiency due to humidity accumulation, which reduced the vapour-pressure gradient, whereas Scenario 2, under drier conditions, achieved ≈80–85%. Overall, the data confirm that water supply enables a consistent evaporative-cooling effect, providing a reliable benchmark for refining the system and applying it to other Facade configurations and environmental contexts.

The results align with the initial hypothesis that combining porous geometry with controlled evaporation can moderate microclimates within a conceptual architectural element. Across the experimental stages, the water-supplied configurations maintained internal temperatures within ranges compatible with the thermal requirements of cavity-nesting bees during summer heat events, even though some were only slightly under the threshold temperatures in the experimental set-up.

From a technical perspective, these findings indicate that evaporative cooling can extend the thermal buffering capacity of porous nesting structures beyond what is achievable through passive material inertia alone. Geometry and surface porosity thus emerge as operative parameters for regulating temperature through a passive, water-based mechanism.

Beyond demonstrating thermal performance, the results suggest that porous geometry and controlled evaporation can be explored as technical strategies to moderate microclimatic conditions within biologically sensitive cavities. Within a more-than-human design perspective, environmental control is reframed not as optimization for human comfort, but as a set of measurable boundary conditions relevant to non-human physiological thresholds (Haraway, 2016).

In this context, the façade prototypes are treated as exploratory test elements rather than as redefinitions of architectural agency. They serve to examine whether material systems can simultaneously support structural, thermal, and habitat-related functions when evaluated against species-specific criteria (Knippers and Speck, 2012; Sattler and Österreicher, 2019). Within this constrained scope, the findings may be read as indicating that architectural surface elements can be evaluated against non-human thermal thresholds using performance logics comparable to those traditionally applied in human comfort research, or in purpose-built structures intended for domesticated animals (Polidori et al., 2023). In this sense, comfort and habitat for non-human species are treated not as ancillary outcomes, but as explicit design criteria assessed through measurable environmental parameters.

Accordingly, the technical investigations—measurements of temperature fluctuations, analysis of convective regimes, and controlled replication of climatic stressors—are employed here as methods for examining how material systems and environmental conditions interact in ways that are relevant to species-specific physiological thresholds. While the experiments remain limited to defined configurations and boundary conditions, this approach allows the material–environment interaction to be analysed in a manner that is comparable across human and non-human comfort frameworks, without extending the empirical claims beyond the tested scale.

By translating the physical process of evaporation into an architectural regenerative strategy, the study situates design within existing discourses of multispecies cohabitation (Haraway, 2016). It proposes that maintaining microclimatic stability for non-humans should be regarded as integral to the resilience of urban ecosystems rather than as an ancillary effect. The prototypes developed are not proposed as solutions but as exploratory models linking environmental physics with questions of coexistence and care of people and the environment alike.

The implications extend beyond the individual facade element to the scale of the urban fabric. If architectural surfaces were considered a distributed set of more-than-human resources, cities may be interpreted as interconnected systems in which façades, roofs, and shading devices collectively contribute to microclimatic regulation and ecological continuity (Buchholz et al., 2020). In this sense, architectural envelopes can be understood as infrastructural “stepping stones,” supporting both microclimatic regulation and spatial connectivity for non-human life within dense urban contexts.

Several limitations of the present study should be acknowledged. First, biological responses are interpreted using generalized thermal thresholds for cavity-nesting bee offspring; species-specific tolerances, behavioural adaptations, and nesting preferences may vary. Second, performance is climate-dependent: evaporative cooling efficiency is constrained by ambient humidity, water availability, and local microclimatic conditions, limiting direct transferability across geographic regions. Third, while evaporative façade systems for human comfort are often benchmarked using similar psychrometric metrics, direct comparisons of comfort outcomes are not appropriate here, as the biological thresholds and exposure conditions of non-human species differ fundamentally from those of human thermal comfort models. In this study, cross-domain benchmarking is therefore restricted to physical efficiency (e.g., fraction of adiabatic cooling potential), rather than experiential or comfort-based criteria.

Taken together, the findings position bio-inspired evaporative cooling as a transferable physical mechanism rather than as a prescriptive ecological solution. By drawing on thermoregulatory strategies observed in biological systems and translating them into quantifiable architectural performance metrics, the study aligns biomimetic design with established building-physics evaluation methods. Temperature differentials, cooling ratios, and fractions of adiabatic potential are therefore employed not as indicators of comfort in a human sense, but as analytical tools for assessing microclimatic suitability relative to species-specific thermal thresholds.

While the prototypes presented here are exploratory and limited to defined boundary conditions, they indicate how porous geometry combined with controlled evaporation can be evaluated as a technically viable strategy for moderating microclimatic extremes in biologically sensitive cavities. Future work is required to extend this investigation toward scalability, including the refinement of additive manufacturing workflows, material selection, and geometric optimisation to balance thermal performance, structural stability, water management, and production constraints. In parallel, broader validation across climatic contexts and urban configurations will be necessary to assess robustness and transferability.