Paradigm Shift in Evidence Based Policy Development for Education Buildings: Mitigation, Adaptation and School Environments Conducive to Learning

Keywords: net zero, education, sustainable buildings, climate change, cognitive performance

Introduction

Climate resilience is not simply about withstanding physical stressors but also ensuring that the building supports human health and wellbeing. This means prioritizing indoor environmental quality, such as maintaining proper air quality and comfortable thermal, lighting and acoustics conditions which is critical in school buildings where the cognitive and physical health of occupants is at stake. The climate resilience policies focusing on school buildings illustrate a multi-layered approach: European and national bodies set the strategic frameworks and funding priorities, while implementation happens locally with tailored solutions to address the unique challenges faced by schools across different regions. To support the national governments, we need a paradigm shift in policy making incorporating all climate resilience principles: (a) mitigation, (b) adaptation, and (c) indoor environmental quality conducive to learning. This paper presents a multifaceted approach to develop a research framework that blends the physical experimentation on the effect of indoor environmental quality on cognitive performance in a controlled environment with the well-defined school building stock models. What have we achieved so far at the UCL Institute for Environmental Design and Engineering in the last decade? Are we there yet?

Decarbonisation

When looking at school building stock modelling, the two main approaches, archetype-based modelling and one-by-one (or building-by-building) modelling - offer complementary strengths and trade-offs.

Archetype-Based Modelling: This approach groups school buildings into representative categories based on shared characteristics such as construction period, building type, and usage profiles. Instead of simulating every individual building, a few archetypes are developed that capture the “average” performance of each subgroup. In England, handful of schools were built before the industrial revolution in the 19th century when the proper mass education system was established. These schools are known as Victorian schools. Early 20th century brings the influence of German Bauhaus movement, followed by two distinctive post war periods including least loved prefabricated CLASP schools. In late 1970s, the first thermal regulatory requirements for schools were introduced and increasingly the schools’ built form becomes less prescribed. The advantages include: (a) lower data requirements & computation, (b) simplicity for policy analysis or broad assessments of regional or national school stocks, and (c) ease of communication as summarized results using archetypes are often easier to communicate to stakeholders who may not be familiar with technical details. To get into details of the archetype modelling see the article written by Dr Jie Dong.

The trade-off introduced by averaging and grouping, can obscure the unique characteristics of individual schools. Variability in energy performance, school design and orientation, surrounding built environment, local microclimates, or specific operational nuances may be lost in translation. This is especially critical when optimal retrofit measures are tested with aim to increase the climate resilience.

One-by-One (Building-by-Building) Modelling: In contrast, one-by-one modelling involves creating a detailed simulation for each school building. This method harnesses granular data ranging from specific construction details, building systems’ details, occupant density and patterns. The key features include: (a) increased precision, (b) potential for better accuracy, and on negative side (c) data and computational demands. Figure 1 presents the one-by-one framework called Modelling Platform for Schools (MPS).

Figure 1. Data Information Flow in Modelling Platform for Schools (MPS).

Modelling Platform for Schools (MPS) is an automated framework designed to generate dynamic thermal simulation models for individual school buildings. Developed by researchers at the UCL Institute for Environmental Design and Engineering, the MPS integrates diverse data sources such as building geometry, size, and the number of buildings within a school premise collected from databases like Edubase, the Condition Data Collection (CDC), and the Ordnance Survey (OC). The National Calculation Methodology (NCM) and the Display Energy Certificates (DEC) datasets are used to determine the occupancy patterns, occupant density, and the actual energy used intensity to validate the MPS. The MPS automatically creates 67,000 school buildings in over 22,000 schools in England with a different degree of success (traffic system is used to describe from perfect match to pay attention). A typical digital representation of schools in English school building stock are given in Figure 2.

Figure 2. Typical automatically generated school geometries.

By automating the creation of thermal simulation models, MPS not only facilitates a comprehensive analysis of the existing school building stock but also supports policy decisions regarding energy efficiency upgrades and retrofits aimed at reducing carbon emissions. By treating each school individually, this approach can capture the heterogeneity inherent in a diverse school stock. This is particularly useful when planning tailored energy efficiency upgrades or when specific buildings require bespoke interventions. Because it considers the exact conditions and setups of each school, the approach can potentially provide a more accurate estimation of energy consumption, thermal performance, and retrofit impacts. The downside is that gathering comprehensive data for every building is resource intensive, and running simulations for hundreds or thousands of buildings can be computationally heavy.

The decision on which method to use often depends on the study’s goals, available data, and the level of detail required. In practice, if the data are available, the frameworks such as the MPS could be used in a hybrid mode: (a) archetype-based approach work well for national policy and broad strategy for decarbonisation building stock, and (b) : one-by-one modelling for tailored interventions and detailed analysis when the objective is to identify the specific characteristics and retrofit needs of individual school buildings. Despite, the archetype-based models are still predominantly used, the use versatility of the one-by-one building stock models, and increasing data availability will lead to their widespread use in future.

Cognitive Performance

As stated above the climate resilience strategies require the holistic approach beyond the energy use intensity and carbon emissions (operational and embodied). Multiple studies have demonstrated that both indoor temperature and ventilation significantly influence the cognitive performance of students. Research has indicated that there is an optimal temperature range—commonly around 19 - 23°C where students perform best on cognitive tasks. Deviations from this range upwards can lead to discomfort, lethargy and students might have difficulty focusing on complex tasks, which can diminish overall academic performance. Conversely, environments that are too cold can also disrupt attention and comfort, underscoring the need to strike a balance for optimal learning conditions. Poor ventilation often leads to the accumulation of CO₂ (as a proxy for ventilation rates) which has been correlated with symptoms such as headaches, drowsiness, and reduced concentration – all factors that can impair cognitive function (lot of academic papers do not differentiate CO₂ as a proxy for ventilation rates or as a pollutant on its own right; please read the paper produced by Dr Ben Jones for this issue who provide a criticism of the current built environment research in this field).

By optimizing thermal comfort and indoor air quality, schools can create a more conducive learning environment that supports robust cognitive performance, thereby enhancing overall academic outcomes.

What we need to know? Cognition refers to the mental processes related to knowledge which occurs in the brain, comprising attention, memory, learning, language, reasoning, problem solving, decision making, perception and a host of other vital processes. The measurement of cognition can be conducted through a range of methods, each differing in terms of their degree of objectivity and sensitivity.

It is essential to understand that unlike learning outcome which “belong” to an individual’s performance in tasks and/or subjects, cognitive performance offers a broader and more enduring quantitative assessment of one's cognitive ability, transcending learning contexts. To integrate the cognitive performance functions in the building stock models we need robust studies carried out in controlled environments, so we can link the outputs of the building simulations with the relevant cognitive performance functions. In my view the studies based on learning outcomes should not be coupled with the school building stock models, as learning is far away more complex phenomenon. A typical experimental process is shown in Figure 3.

The development of computer technology has led to the development of several computerised cognitive testing batteries. Computerised cognitive tests have been created and verified to target specific brain regions, offering numerous advantages compared to traditional pen and paper methods. In addition to teasing apart different cognitive domains, data collection can be automatic process, thereby minimizing the potential for errors and influence of administrator bias. Computers can enable precise recordings, such as highly accurate measurement of response latencies. The implementation of standardised assessment and the flexibility to adjust difficulty levels of computerized cognitive tests effectively minimize the floor and ceiling effects. While cognitive performance does not have standard "units" like physical measurements (e.g., temperature), speed (how quickly each pupil worked per unit of time) and accuracy (expressed as a percentage of possible errors) are typically used as the main indicators to characterize students’ performance in different cognitive tasks. They are often expressed as a relative value such as CPL (Cognitive Performance Loss, see the article written by Dr Jie Dong).

At the UCL Institute for Environmental Design and Engineering we use the Behavioural Assessment and Research System (BARS), but in principle the advantages of cognitive test batteries (SMS, BARS, NATB) over traditional cognitive tests: (a) have proven validity in measuring cognitive performance in a variety of domains, (b) are often standardized, which can make them less susceptible to bias, (c) can be administered by personnel with limited training, and (d) cover a wider range of domains associated with education. Figure 3 proposes a set of cognitive performance tests suitable for learning environments (some simplifications might be possible, but more evidence is required). For detailed description see the article written by Dr Didong Chen in this issue.

Figure 3. Structure of a typical experimental protocol.

Based on our experience on evaluation of cognitive performance of students in controlled environments (based on the doctoral work carried out by Dr Riham Ahmed, 2016, Dr Didong Chen, 2025 and Dr Zeyu Zhao, 2025) to enable the integration into the school building stock simulation the experimental studies must:

1.    be well documented including the study design, ethics approval, execution, data analysis including access to the data created via open access platforms such as GitHub

2.    conduct a robust sample size calculation prior to the experiment to ensure enough statistical power of the research findings

3.    implement a validated cognitive test battery measuring a wide range of cognitive domains of importance to learning

4.    control for factors like age, gender, acute exercise, sleep, and caffeine which have long been associated with individuals’ cognitive functioning

5.    to account for confounding factors, ensuring that either temperature, ventilation rates and/or any other parameter is the sole independent variable. If this is not possible, investigations should strive to quantify and disclose the variability in confounding factors, enabling researchers to determine whether these factors are likely to affect the cognitive performance outcomes or not

6.    testing one individual alone in the chamber might be a better solution as one participant might be impacted by another when taking the cognitive test.

When comparing the results from prior research, different cognitive tests were used to measure cognitive performance, each test might be associated with several cognitive domains, which makes it difficult to reach a definite conclusion by comparing the results like-for-like based on individual tests. Interpreting the findings from an aggregated perspective is necessity before the academic community reaches consensus.

Techniques such as electroencephalography (EEG), heart rate monitoring and fMRI provide researchers with means to deeply understand and assess cognitive performance by evaluating physiological and psychological response mechanisms. Not applied to the built environment research too often, so more transdisciplinary approach might become a norm in future years. Although not directly required for integration in the building simulation models, this might further improve the robustness of the cognitive performance functions required.

Climate Change

Existing English schools will continue to operate for years in the future, many of which lack the consideration for climate change in terms of construction and operation. CIBSE (Chartered Institution of Building Services Engineers) provides specialized weather data files that are essential for climate-resilient building design in the UK. CIBSE offers future weather files based on UK climate projections (UKCIP09), covering (a) time periods: 2020s, 2050s, 2080s, (b) emissions scenarios: low, medium, high, and (c) percentiles: 10th (cooler), 50th (median), 90th (hotter). These are currently available for 14 UK locations and are structured to support simulations under different climate futures. For high-resolution (2.2 km) climate change weather data in the UK, CIBSE is testing the new set of weather files based on the UKCP18 local projections (2.2 km) which should be implemented in future.

Based on the archetype model of English school building stock developed at the UCL Institute for Environmental Design and Engineering, Dr Jie Dong estimated the frequency distributions of hourly Cognitive Performance Losses (CPLs) in English schools per region during non-heating school days (Figure 4). Note that schools in Southern England have minimal hours categorized as 'No loss' and 'No significant loss' in future climates both of which are much higher than those in the other two regions. Having this in mind all mitigation and adaptation policies should consider regional differences.

Figure 4. The frequency distributions of hourly CPLs within different levels in English schools per region during non-heating school days.

Finally, the Figure 5 shows the potential of air conditioning as an active adaptation measure in reducing cognitive performance loss of pupils in insulated schools. The median values of cognitive performance loss and the corresponding cooling loads at set point temperatures of 21°C and 25°C in schools across all three regions of England highlight significant increase in the cooling load in Southern region of England with climate change if cognitive performance loss is to be maintained at levels conducive to learning. Unfortunately, at the time of writing this paper the impact of decarbonisation on climate change and cognitive performance has now been analysed yet, but the architype modelling set up the trends and challenges that we will face on transition to net zero.

Figure 5. The comparison of cognitive performance loss and corresponding cooling load at the set-point temperature of 21°C and 25°C.

Delivering a Paradigm Shift in Evidence Based Policies

The climate resilience and adaptation have gained interest not only from building experts, but also from educationalists and policymakers in the education sector. To support the national governments, we need a paradigm shift in policy making incorporating all climate resilience principles: (a) mitigation, (b) adaptation, and (c) indoor environmental quality conducive to learning. One-by-one modelling approach is versatile to cater for both, tailored interventions and detailed analysis when the objective is to identify the specific characteristics and retrofit needs of individual school buildings, and if aggregated at the national and regional levels to form basis for the evidence-based policy making and the strategy development.

However, stakeholders without building expertise may struggle to fully comprehend the outcomes of building performance assessments based on traditional engineering key performance indicators such ventilation rates, operative, air or radiant temperature, and the implications of these outcomes on cognitive performance of children in classrooms. Using cognitive performance loss as the knowledge performance indicator to describe the climate resilience of schools offers a means of using language that can be understood by non-building experts, facilitating the interpretation and communication of research findings to the intended audience.

Acknowledgments

This think piece is based on synthesis of work carried out by a large team of my colleagues and doctoral researchers in the last 15 years. Special thanks to Dr Jie Dong, Dr Duncan Grassie, Dr Didong Chen, Dr Zeyu Zhao, Dr Riham Ahmed, Dr Greig Paterson, Dr Sung Min Hong who were awarded their PhDs on the various aspects of the work on education buildings used to develop the ideas presented here. Also, many thanks to the core team of academics and postdoctoral researchers that I have worked with on these topics especially the core team members: Dr Ivan Korolija, Dr Yair Schwartz and Daniel Godoy Shimizu.

Bibliography

[1]     Zhao, Zeyu; Bagkeris, Emmanouil; Mumovic, Dejan; (2025) The combined impacts of indoor temperature and total volatile organic compounds on cognitive performance of university students: A controlled exposure study. Science of The Total Environment, 966, Article 178652. 10.1016/j.scitotenv.2025.178652.

[2]     Schwartz, Yair; Korolija, Ivan; Godoy-Shimizu, Daniel; Hong, Sung-Min; Mavrogianni, Anna; Mumovic, Dejan; (2024) School building-stock climate resilience: evaluating London’s school stock overheating performance. Intelligent Buildings International pp. 1-22. 10.1080/17508975.2024.2410804.

[3]     Dong, J; Schwartz, Y; Korolija, I; Mumovic, D; (2024) Unintended consequences of English school stock energy-efficient retrofit on cognitive performance of children under climate change. Building and Environment, 249, Article 111107. 10.1016/j.buildenv.2023.111107.

[4]     Chen, Didong; Huebner, Gesche; Bagkeris, Emmanouil; Ucci, Marcella; Mumovic, Dejan; (2023) Effects of short-term exposure to moderate pure carbon dioxide levels on cognitive performance, health symptoms and perceived indoor environment quality. Building and Environment, 245, Article 110967. 10.1016/j.buildenv.2023.110967.

[5]     Dong, Jie; Schwartz, Yair; Korolija, Ivan; Mumovic, Dejan; (2023) The impact of climate change on cognitive performance of children in English school stock: A simulation study. Building and Environment, 243, Article 110607. 10.1016/j.buildenv.2023.110607.

[6]     Ahmed, Riham; Mumovic, Dejan; Bagkeris, Emmanouil; Ucci, Marcella; (2022) Combined effects of ventilation rates and indoor temperatures on cognitive performance of female higher education students in a hot climate. Indoor Air, 32 (2), Article e13004. 10.1111/ina.13004.